Differential Response of Phaeodactylum tricornutum and Cylindrotheca fusiformis to High Concentrations of Cu 2+ and Zn 2+

: Diatoms can be used as biosensors to assess aquatic environment quality, because they are widely distributed in almost all aquatic environments and show varied sensitivities toward heavy metal ions. The marine planktonic diatoms Phaeodactylum tricornutum ( P. tricornutum ) and Cylindrotheca fusiformis ( C. fusiformis ) are typical representatives of planktonic diatoms and benthic diatoms, respectively. C. fusiformis is very sensitive to changes in the concentration of heavy metal ions, and can be used as an indicator of the quality of the sedimental environment, while P. tricornutum can tolerate higher concentrations of heavy metal ions. To explore the potential difference in responses to heavy metal ions between planktonic and benthic diatoms, we compared the transcriptome of P. tricornutum and C. fusiformis under Cu 2+ and Zn 2+ treatment. The results indicated that P. tricornutum has several genes involved in ion transmembrane transport and ion homeostasis, which are signiﬁcantly downregulated under Cu 2+ and Zn 2+ treatment. However, this enrichment of ion transmembrane transport- and ion homeostasis-related genes was not observed in C. fusiformis under Cu 2+ and Zn 2+ treatment. Additionally, genes related to heavy metal ion stress response such as peroxiredoxin, peroxidase, catalase, glutathione metabolism, phytochelatin, oxidative stress and disulﬁde reductase, were upregulated in P. tricornutum under Cu 2+ and Zn 2+ treatment, whereas most of them were downregulated in C. fusiformis under Cu 2+ and Zn 2+ treatment. This difference in gene expression may be responsible for the difference in sensitivity to heavy metals between P. tricornutum and C. fusiformis . biosynthesis-related complex methionine


Introduction
The distribution and composition of biological communities are controlled or influenced by environmental variations such as disturbances, stressors, and biotic interactions and change in resources and hydraulic conditions [1]; therefore, such biological communities can be used as indicators of environmental conditions. Diatoms, for example, can be used as biosensors to assess aquatic environment quality, because diatoms are widely distributed in almost all aquatic environments [2], and different species of diatoms show varying sensitivities toward heavy metal ions [3]. Therefore, their species and distribution can be used as an indicator of the degree of heavy metal pollution in aquatic environments [4][5][6][7]. P. tricornutum and C. fusiformis were obtained from the Microalgae Culture Center at the Ocean University of China. For P. tricornutum and C. fusiformis, algal cells were cultured using sterilized artificial seawater supplemented with f/2 nutrients at 20 • C and with four times of f/2 nutrients (2f) at 25 • C, respectively [20]. All cultures were grown under a 12:12 dark:light cycle under cool white fluorescent light (approximately 100 µmol m −2 s −1 ). Cell growth was detected by measuring the absorbance at 730 nm using a UV/visible spectrophotometer (UV-1800, Shimadzu, Tokyo, Japan).
For treatment with high concentrations of Cu 2+ and Zn 2+ , P. tricornutum and C. fusiformis cells were treated with Cu 2+ or Zn 2+ at final concentrations of 30 µM and 60 µM. Control cells were cultured in f/2 (for P. tricornutum) or 2f medium (for C. fusiformis). Each treatment was performed in triplicate in 250 mL flasks. Cell growth was detected on days 0, 1, 3, 5 and 7.

Scanning Electron Microscope (SEM) and Energy Dispersive Spectroscopy (EDS) Analysis
For SEM-EDS analysis, P. tricornutum and C. fusiformis cells were treated with Cu 2+ at a final concentration of 5 µM (PTCu and CFCu) and Zn 2+ at a final concentration of 30 µM (PTZn and CFZn). Control cells were cultured in f/2 (for P. tricornutum, PTC) or 2f medium (for C. fusiformis, CFC). Each treatment was performed in triplicate in 2 L flasks. After 48 h, cell pellets were collected, further washed using distilled sea water, and centrifuged at 5000× g for 4 min. The pellets were fixed with 2.5% glutaraldehyde (4 • C) overnight and sequentially dehydrated for 15 min each in 30%, 50%, 70%, 80%, 90%, 100% and 100% EtOH, followed by CO 2 critical point drying. Dried cells were placed on a conductive silicone rubber plate and treated with Gold sputtering, then viewed under the SEM (Hitachi's TM4000 Plus, Hitachi Limited, Tokyo, Japan). EDS was performed with IXRF's TM4-EDS.

Transcriptomic Analysis
For transcriptomic analysis, P. tricornutum and C. fusiformis cells were treated with Cu 2+ at a final concentration of 5 µM (PTCu and CFCu) and Zn 2+ at a final concentration of 30 µM (PTZn and CFZn). Control cells were cultured in f/2 (for P. tricornutum, PTC) or 2f medium (for C. fusiformis, CFC). Each treatment was performed in triplicate in 2 L flasks. After 48 h, cell pellets were collected, further washed using distilled sea water, and centrifuged at 5000× g for 4 min. The pellets were frozen in liquid nitrogen and stored at −80 • C.
To generate clean reads, raw sequence reads were trimmed using SeqPrep (https: //github.com/jstjohn/SeqPrep accessed on 5 October 2016), and the quality of the raw reads was controlled using Sickle (https://github.com/najoshi/sickle accessed on 15 March 2015) with default parameters. The clean reads were annotated according to Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), Clusters of Orthologous Groups of proteins (COG), the NCBI non-redundant protein sequences database (NR), Swiss-Prot, and Pfam databases. The mapped reads were further normalized using the reads per kb per million methods for the identification of differentially expressed genes (DEGs). Abundant genes were quantified using RSEM (http://deweylab.biostat.wisc.edu/rsem/ accessed on 14 February 2020) [21]. Differential gene expression was determined using the "edgeR" package in R (http://www.bioconductor.org/packages/2.12/bioc/html/edgeR. html accessed on 9 May 2018), based on the following threshold parameters: log2 foldchange > 2 and p-value < 0.05 [22]. Functional annotation and enrichment analyses were performed and classified using the GO and KEGG databases. 2+ and Zn 2+ on Growth of P. tricornutum and C. fusiformis

Effects of Cu
To investigate the effect of Cu 2+ , Zn 2+ on P. tricornutum and C. fusiformis, we compared the growth rates of P. tricornutum and C. fusiformis at different concentrations (0, 30 and 60 µM Cu 2+ and 0, 30 and 60 µM Zn 2+ ). The results showed that 60 µM Cu 2+ significantly decreased the growth of P. tricornutum (p-value < 0.05), while the influence of 30 µM Cu 2+ was not significant (p-value > 0.05), although there was also a tendency to decrease compared with that of the control group (Figure 1a). Both 30 µM and 60 µM Cu 2+ significantly decreased the growth of C. fusiformis (Figure 1b). Neither 30 nor 60 µM Zn 2+ significantly influenced the growth of P. tricornutum (Figure 1a), while 60 µM Zn 2+ decreased the growth of C. fusiformis on day 5 ( Figure 1b). As both 30 µM and 60 µM Cu 2+ significantly decreased the growth of C. fusiformis, the concentration of Cu 2+ for the transcriptomic analysis was set to a lower level (5 µM). As 30 µM Zn 2+ did not significantly influence the growth of both P. tricornutum and C. fusiformis, the concentration of Zn 2+ for the transcriptomic analysis was set to 30 µM. decreased the growth of C. fusiformis (Figure 1b). Neither 30 nor 60 μM Zn 2+ significa influenced the growth of P. tricornutum (Figure 1a), while 60 μM Zn 2+ decreased growth of C. fusiformis on day 5 ( Figure 1b). As both 30 μM and 60 μM Cu 2+ significa decreased the growth of C. fusiformis, the concentration of Cu 2+ for the transcriptomic ysis was set to a lower level (5 μM). As 30 μM Zn 2+ did not significantly influenc growth of both P. tricornutum and C. fusiformis, the concentration of Zn 2+ for the scriptomic analysis was set to 30 μM.

Effects of Cu 2+ and Zn 2+ on Cell Morphology of P. tricornutum and C. fusiformis
To investigate the effect of Cu 2+ and Zn 2+ on cell morphology of P. tricornutum an fusiformis, we observed the cells with SEM. The results showed that both Cu 2+ and Zn 2 not significantly change the cell morphology of P. tricornutum (Figure 2a-c), while Cu 2+ and Zn 2+ significantly changed the cell morphology of C. fusiformis (Figure 2d-f).  2+ and Zn 2+ on Cell Morphology of P. tricornutum and C. fusiformis

Effects of Cu
To investigate the effect of Cu 2+ and Zn 2+ on cell morphology of P. tricornutum and C. fusiformis, we observed the cells with SEM. The results showed that both Cu 2+ and Zn 2+ did not significantly change the cell morphology of P. tricornutum (Figure 2a-c), while both Cu 2+ and Zn 2+ significantly changed the cell morphology of C. fusiformis (Figure 2d-f). This

Accumulation of Cu 2+ and Zn 2+ on Biosilica Shell of P. tricornutum and C. fusiformis
To investigate the accumulation of Cu 2+ and Zn 2+ on the biosilica shell of P. tricornutum and C. fusiformis, EDS was conducted to analyze the concentration of Si, Cu and Zn on the cell surface of P. tricornutum and C. fusiformis. The results showed that in the control group (containing 0.04 μM Cu 2+ and 0.08 μM Zn 2+ in the medium) of P. tricornutum, the content of Cu and Zn was 16.72% and 13.76% (Figure 3a), respectively. In the Cu 2+ group (containing 5 μM Cu 2+ and 0.08 μM Zn 2+ in the medium) of P. tricornutum, the content of Cu and Zn was 18.70% and 13.75% (Figure 3b), respectively. In the Zn 2+ group (containing 0.04 μM Cu 2+ and 30 μM Zn 2+ in the medium) of P. tricornutum (Figure 3c), the content of Cu and Zn was 6.53% and 17.76%, respectively. While in the control group (containing 0.16 μM Cu 2+ and 0.32 μM Zn 2+ in the medium) of C. fusiformis, the content of Cu and Zn was 0% (Figure 3d). In the Cu 2+ group (containing 5 μM Cu 2+ and 0.32 μM Zn 2+ in the medium) of C. fusiformis, the content of Cu and Zn was 23.53% and 14.58% (Figure 3e), respectively. In the Zn 2+ group (containing 0.16 μM Cu 2+ and 30 μM Zn 2+ in the medium) of C. fusiformis, the content of Cu and Zn was 9.92% and 20.81% (Figure 3f), respectively. These results indicated that both P. tricornutum and C. fusiformis accumulated Cu and Zn on the cell surface.

Accumulation of Cu 2+ and Zn 2+ on Biosilica Shell of P. tricornutum and C. fusiformis
To investigate the accumulation of Cu 2+ and Zn 2+ on the biosilica shell of P. tricornutum and C. fusiformis, EDS was conducted to analyze the concentration of Si, Cu and Zn on the cell surface of P. tricornutum and C. fusiformis. The results showed that in the control group (containing 0.04 µM Cu 2+ and 0.08 µM Zn 2+ in the medium) of P. tricornutum, the content of Cu and Zn was 16.72% and 13.76% (Figure 3a), respectively. In the Cu 2+ group (containing 5 µM Cu 2+ and 0.08 µM Zn 2+ in the medium) of P. tricornutum, the content of Cu and Zn was 18.70% and 13.75% (Figure 3b), respectively. In the Zn 2+ group (containing 0.04 µM Cu 2+ and 30 µM Zn 2+ in the medium) of P. tricornutum (Figure 3c), the content of Cu and Zn was 6.53% and 17.76%, respectively. While in the control group (containing 0.16 µM Cu 2+ and 0.32 µM Zn 2+ in the medium) of C. fusiformis, the content of Cu and Zn was 0% (Figure 3d). In the Cu 2+ group (containing 5 µM Cu 2+ and 0.32 µM Zn 2+ in the medium) of C. fusiformis, the content of Cu and Zn was 23.53% and 14.58% (Figure 3e), respectively. In the Zn 2+ group (containing 0.16 µM Cu 2+ and 30 µM Zn 2+ in the medium) of C. fusiformis, the content of Cu and Zn was 9.92% and 20.81% (Figure 3f), respectively. These results indicated that both P. tricornutum and C. fusiformis accumulated Cu and Zn on the cell surface.

Annotation of P. tricornutum Transcriptome
To investigate the potential effect of Cu 2+ and Zn 2+ on gene transcription in P. tricornutum, we analyzed the transcriptome of P. tricornutum exposed to 5 μM Cu 2+ (PTCu) and 30 μM Zn 2+ (PTZn) for 48 h, with control (PTC) with no treatment of heavy metals. An average of 46,069,016 raw reads and 45,693,261 clean reads were generated from the total RNA extracted from P. tricornutum. A total of 98.39% of the clean read bases had a Q-value ≥ 20, and 94.99% of the clean read bases had a Q-value ≥ 30 (Table S1). De novo assembly generated 10,754 unigenes, including 10,167 known genes and 587 new genes. Figure 4 shows the length distribution of unigenes in P. tricornutum.  To investigate the potential effect of Cu 2+ and Zn 2+ on gene transcription in P. tricornutum, we analyzed the transcriptome of P. tricornutum exposed to 5 µM Cu 2+ (PTCu) and 30 µM Zn 2+ (PTZn) for 48 h, with control (PTC) with no treatment of heavy metals. An average of 46,069,016 raw reads and 45,693,261 clean reads were generated from the total RNA extracted from P. tricornutum. A total of 98.39% of the clean read bases had a Q-value ≥ 20, and 94.99% of the clean read bases had a Q-value ≥ 30 (Table S1). De novo assembly generated 10,754 unigenes, including 10,167 known genes and 587 new genes. Figure 4 shows the length distribution of unigenes in P. tricornutum.

Annotation of P. tricornutum Transcriptome
To investigate the potential effect of Cu 2+ and Zn 2+ on gene transcription in P. tricornutum, we analyzed the transcriptome of P. tricornutum exposed to 5 μM Cu 2+ (PTCu) and 30 μM Zn 2+ (PTZn) for 48 h, with control (PTC) with no treatment of heavy metals. An average of 46,069,016 raw reads and 45,693,261 clean reads were generated from the total RNA extracted from P. tricornutum. A total of 98.39% of the clean read bases had a Q-value ≥ 20, and 94.99% of the clean read bases had a Q-value ≥ 30 (Table S1). De novo assembly generated 10,754 unigenes, including 10,167 known genes and 587 new genes. Figure 4 shows the length distribution of unigenes in P. tricornutum.  The acquired unigenes were annotated according to the GO, KEGG, COG, NR, Swiss-Prot, and Pfam databases. Of all the assembled unigenes, 82.28%, 46.02%, 72.75%, 99.19%, 56.1%, and 74.18% were annotated by GO, KEGG, COG, NR, Swiss-Prot, and Pfam, respectively ( Figure 5, Table S2). The acquired unigenes were annotated according to the GO, KEGG, COG, NR, Swiss-Prot, and Pfam databases. Of all the assembled unigenes, 82.28%, 46.02%, 72.75%, 99.19%, 56.1%, and 74.18% were annotated by GO, KEGG, COG, NR, Swiss-Prot, and Pfam, respectively ( Figure 5, Table S2).

Identification and Functional Enrichment Analysis of Different Express Genes (DEGs) in P. tricornutum upon Cu 2+ Treatment
Transcriptome analysis of DEGs in P. tricornutum exposed to 5 μM Cu 2+ was performed, using high-throughput RNA sequencing. A total of 2006 genes, including 1119 up-and 887 downregulated genes were detected to be significantly regulated (p < 0.05) under Cu 2+ treatment, with a 2-fold change in abundance considered as the criterion of biologically significant difference (Table S3). DEGs were classified into three main functional categories of GO terms: molecular function (MF), biological process (BP), and cellular component (CC; Figure 6). The GO enrichment analysis for upregulated genes is shown in Figure 6a, in which only 20 annotation categories with the most significantly enriched DEPs are shown. For BP, DEGs were assigned to 13 subcategories involved in photosynthesis, carbon metabolism, and energy metabolism, with the three most abundant clusters being 'protein-chromophore linkage', 'photosynthesis, light harvesting in photosystem I', and 'photosynthesis, light harvesting'. For CC, DEGs were classified into five subcategories involved in photosynthesis, 'thylakoid membrane', 'chloroplast thylakoid membrane', 'plastid thylakoid membrane', 'photosynthetic membrane', and 'light-harvesting complex'. In the MF category, DEGs were divided into the two subcategories 'chlorophyll-binding' and 'tetrapyrrole binding'. The GO enrichment analysis for downregulated genes is shown in Figure 6b, in which only 20 annotation categories with the most significantly enriched DEPs are shown. For BP, DEGs were assigned to 11 subcategories involved in metal ion homeostasis, cation homeostasis, and ion transport. For CC, DEGs were classified into four subcategories involved in the integral component of (plasma) membrane and intrinsic component of (plasma) membrane. In the MF category, the DEGs were divided into four subcategories involved in (inorganic) cation and inorganic molecular entity transmembrane transporter activity.

Identification and Functional Enrichment Analysis of Different Express Genes (DEGs) in P. tricornutum upon Cu 2+ Treatment
Transcriptome analysis of DEGs in P. tricornutum exposed to 5 µM Cu 2+ was performed, using high-throughput RNA sequencing. A total of 2006 genes, including 1119 up-and 887 downregulated genes were detected to be significantly regulated (p < 0.05) under Cu 2+ treatment, with a 2-fold change in abundance considered as the criterion of biologically significant difference (Table S3). DEGs were classified into three main functional categories of GO terms: molecular function (MF), biological process (BP), and cellular component (CC; Figure 6). The GO enrichment analysis for upregulated genes is shown in Figure 6a, in which only 20 annotation categories with the most significantly enriched DEPs are shown. For BP, DEGs were assigned to 13 subcategories involved in photosynthesis, carbon metabolism, and energy metabolism, with the three most abundant clusters being 'protein-chromophore linkage', 'photosynthesis, light harvesting in photosystem I', and 'photosynthesis, light harvesting'. For CC, DEGs were classified into five subcategories involved in photosynthesis, 'thylakoid membrane', 'chloroplast thylakoid membrane', 'plastid thylakoid membrane', 'photosynthetic membrane', and 'light-harvesting complex'. In the MF category, DEGs were divided into the two subcategories 'chlorophyll-binding' and 'tetrapyrrole binding'. The GO enrichment analysis for downregulated genes is shown in Figure 6b, in which only 20 annotation categories with the most significantly enriched DEPs are shown. For BP, DEGs were assigned to 11 subcategories involved in metal ion homeostasis, cation homeostasis, and ion transport. For CC, DEGs were classified into four subcategories involved in the integral component of (plasma) membrane and intrinsic component of (plasma) membrane. In the MF category, the DEGs were divided into four subcategories involved in (inorganic) cation and inorganic molecular entity transmembrane transporter activity.
Overall, 19 DEGs involved in heavy metal ion stress response are listed in Table 1. These genes were mainly related to antioxidants such as peroxiredoxin, peroxidase, catalase, glutathione metabolism, phytochelatin, oxidative stress, and disulfide reductase. Most (14 Overall, 19 DEGs involved in heavy metal ion stress response are listed in Table 1. These genes were mainly related to antioxidants such as peroxiredoxin, peroxidase, catalase, glutathione metabolism, phytochelatin, oxidative stress, and disulfide reductase. Most (14 out of 19) of these genes were upregulated, indicating their important roles in response to the high concentration of heavy metal ions.

Effects of Zn 2+ on Gene Transcription in P. tricornutum
Transcriptome analysis of differential gene expression in P. tricornutum exposed to 30 µM Zn 2+ was performed, using high-throughput RNA sequencing. A total of 4043 genes, including 2184 up-and 1859 downregulated genes were detected to be significantly regulated (p < 0.05) under Zn 2+ treatment (Table S4). The GO enrichment analysis for DEGs in P. tricornutum under Zn 2+ treatment is shown in Figure 7, in which only 20 annotation categories with the most significantly enriched DEGs are shown. The GO enrichment for DEGs in P. tricornutum under Zn 2+ treatment was similar to that under Cu 2+ treatment, in which the upregulated genes were mainly involved in photosynthesis (Figure 7a), whereas the downregulated genes were mainly involved in ion homeostasis, cation homeostasis, and ion transport (Figure 7b). regulated (p < 0.05) under Zn 2+ treatment (Table S4). The GO enrichment analysis for DEGs in P. tricornutum under Zn 2+ treatment is shown in Figure 7, in which only 20 annotation categories with the most significantly enriched DEGs are shown. The GO enrichment for DEGs in P. tricornutum under Zn 2+ treatment was similar to that under Cu 2+ treatment, in which the upregulated genes were mainly involved in photosynthesis (Figure 7a), whereas the downregulated genes were mainly involved in ion homeostasis, cation homeostasis, and ion transport (Figure 7b). Overall, 24 DEGs involved in heavy metal ion stress response are listed in Table 2. These genes were mainly related to antioxidants such as peroxidase, catalase, peroxiredoxin, glutathione metabolism, phytochelatin biosynthetic process, oxidative stress, mutase superoxide dismutase, and disulfide reductase. Most (17 out of 24) of these genes were upregulated, indicating their important roles in response to the high concentration of heavy metal ions.  Overall, 24 DEGs involved in heavy metal ion stress response are listed in Table 2. These genes were mainly related to antioxidants such as peroxidase, catalase, peroxiredoxin, glutathione metabolism, phytochelatin biosynthetic process, oxidative stress, mutase superoxide dismutase, and disulfide reductase. Most (17 out of 24) of these genes were upregulated, indicating their important roles in response to the high concentration of heavy metal ions.  To investigate the potential effect of Cu 2+ and Zn 2+ treatment on transcription in C. fusiformis, we analyzed the transcriptome of C. fusiformis exposed to 5 µM Cu 2+ (CFCu) and 30 µM Zn 2+ (CFZn) for 48 h, with no addition of heavy metal ions as the control (CFC). An average of 43,832,802 raw reads and 43,323,647 clean reads were generated from total RNA extracted from C. fusiformis. A total of 98.25% of the clean read bases had a Q-value ≥ 20, and 94.64% of the clean read bases had a Q-value ≥ 30 (Table S5). De novo assembly generated 26,146 unigenes. Figure 8 shows the length distribution of unigenes. The acquired unigenes were annotated according to GO, KEGG, COG, NR, Swiss-Prot, and Pfam databases. Of all the assembled unigenes, 36.72%, 35.9%, 56.07%, 38.88%, 39.87%, and 55.78% were annotated by GO, KEGG, COG, NR, Swiss-Prot, and Pfam, respectively ( Figure 9, Table S6). The acquired unigenes were annotated according to GO, KEGG, COG, NR, Swiss-Prot, and Pfam databases. Of all the assembled unigenes, 36.72%, 35.9%, 56.07%, 38.88%, 39.87%, and 55.78% were annotated by GO, KEGG, COG, NR, Swiss-Prot, and Pfam, respectively ( Figure 9, Table S6).

Effects of Cu 2+ on Gene Transcription in C. fusiformis
Transcriptome analysis of differential gene expression in C. fusiformis exposed to 5 μM Cu 2+ was performed, using high-throughput RNA sequencing. A total of 1133 genes, including 315 up-and 818 downregulated genes were detected to be significantly regulated (p < 0.05) under Cu 2+ treatment (Table S7). The GO enrichment analysis for upregulated genes is shown in Figure 10a, in which only 20 annotation categories with the most Figure 9. Functional annotation of unigenes in C. fusiformis.

Effects of Cu 2+ on Gene Transcription in C. fusiformis
Transcriptome analysis of differential gene expression in C. fusiformis exposed to 5 µM Cu 2+ was performed, using high-throughput RNA sequencing. A total of 1133 genes, including 315 up-and 818 downregulated genes were detected to be significantly regulated (p < 0.05) under Cu 2+ treatment (Table S7). The GO enrichment analysis for upregulated genes is shown in Figure 10a, in which only 20 annotation categories with the most significantly enriched DEPs are shown. For BP, DEGs were assigned to 17 subcategories involved in signal transduction, nucleotide biosynthetic, organophosphate biosynthetic, etc. For CC, DEGs were classified into 1 subcategory, the plasma membrane. In the MF category, the unigenes were divided into 2 subcategories, 3 ,5 -cyclic-nucleotide phosphodiesterase activity, and cyclic-nucleotide phosphodiesterase activity. The GO enrichment analysis for downregulated genes is shown in Figure 10b. For BP, no DEGs were enriched. For CC, DEGs were classified into 2 subcategories as an intrinsic component of the membrane and integral component of the membrane. In the MF category, the unigenes were divided into 3 subcategories, phospholipid transporter, glutamyl-tRNA reductase, and lipase activities.
phodiesterase activity, and cyclic-nucleotide phosphodiesterase activity. The GO enrichment analysis for downregulated genes is shown in Figure 10b. For BP, no DEGs were enriched. For CC, DEGs were classified into 2 subcategories as an intrinsic component of the membrane and integral component of the membrane. In the MF category, the unigenes were divided into 3 subcategories, phospholipid transporter, glutamyl-tRNA reductase, and lipase activities.  Table 3, including 1 peroxiredoxin, 1 glutathione synthetase, 1 glutathione S-transferase, 1 glutathione peroxidase, 1 hydroxyacylglutathione hydrolase, 1 deaminated glutathione amidase, and 1 peroxinectin. In total, 7 out of 8 genes were downregulated, indicating considerable differences between C. fusiformis and P. tricornutum in response to the high concentration of heavy metal ions.   Transcriptome analysis of differential gene expression in C. fusiformis exposed to 30 µM Zn 2+ was performed using high-throughput RNA sequencing. A total of 1900 genes, including 854 up-and 1046 downregulated genes were detected to be significantly regulated (p < 0.05) under Zn 2+ treatment (Table S8). The GO enrichment analysis for upregulated genes is shown in Figure 11a. For BP, DEGs were enriched in 1 subcategory of cellular modified amino acid metabolic process. For CC, DEGs were classified into 3 subcategories, 3-oxoacyl-[acyl-carrier-protein] synthase activity, arginase activity, and cullin family protein binding. The GO enrichment analysis for downregulated genes is shown in Figure 11b, in which only 20 annotation categories with the most significantly enriched DEPs are shown. For BP, DEGs were assigned to 8 subcategories involved in the regulation of biological quality, homeostasis, posttranslational modification (amino acid modification), organelle assembly, response to topologically incorrect protein, etc. For CC, DEGs were classified into 3 subcategories as an intrinsic component of membrane, an integral component of membrane, and endoplasmic reticulum lumen. In the MF category, the unigenes were divided into 9 subcategories involved in catalytic activity, ATPase activity, tubulin (cytoskeletal protein, calcium ion, microtubule) binding, primary active transmembrane transporter activity, and protein kinase activity.  Table 4, including 1 thioredoxin-like protein, 1 glutathione synthetase, 4 glutathione S-transferase, peroxiredoxin, 1 glutathionyl-hydroquinone reductase, 1 glutathione peroxidase, 2 phytochelatin biosynthesis-related genes, 1 light-harvesting complex stress-related protein, 1 thyroid peroxidase, 1 oxidative stress-related Abc1-like protein, 1 catalase-peroxidase, 1 methionine sulfoxide reductase, and 1 peroxinectin. Half of them were downregulated, and the rest were upregulated, which is different from the result under Cu 2+ treatment.   Table 4, including 1 thioredoxin-like protein, 1 glutathione synthetase, 4 glutathione S-transferase, peroxiredoxin, 1 glutathionylhydroquinone reductase, 1 glutathione peroxidase, 2 phytochelatin biosynthesis-related genes, 1 light-harvesting complex stress-related protein, 1 thyroid peroxidase, 1 oxidative stress-related Abc1-like protein, 1 catalase-peroxidase, 1 methionine sulfoxide reductase, and 1 peroxinectin. Half of them were downregulated, and the rest were upregulated, which is different from the result under Cu 2+ treatment.

Discussion
Cu 2+ and Zn 2+ are crucial micronutrients for diatoms. When Cu 2+ and Zn 2+ are present in an adequate amount, diatoms exhibit a stronger fitness and grow faster. Cu 2+ and Zn 2+ are components of many enzymes in algae cells. Cu 2+ is involved in the electron transport of photosynthesis by serving as a ligand of cytochrome oxidase and plastocyanin [23,24]. In addition, Cu 2+ is a component of Cu-tyroninase and multicopper oxidase, which is involved in Fe-deficiency response [25,26]. Zn 2+ is an important component for carbonic anhydrases, which are involved in CO 2 fixation, and zinc finger transcription factors, which are involved in gene transcription [27,28]. In addition, both Cu 2+ and Zn 2+ are important for Cu/Zn-SOD (superoxide dismutase) which is involved in anti-oxidation [29]. However, excess Cu 2+ or Zn 2+ will interfere with cellular physiology and biological processes, resulting in decreased cell growth and even death. Different types of cells have different types and amounts of enzymes, thus their demands for Cu 2+ and Zn 2+ are various. Meanwhile, as the shielding and permeation properties of cell membranes for heavy metal ions are different in various species, their tolerances to heavy metal ions are also various. In this study, the growth of both P. tricornutum and C. fusiformis was inhibited at 60 µM Cu 2+ , while 30 µM Cu 2+ decreased the growth of C. fusiformis, yet did not have significant effect on the growth of P. tricornutum (Figure 1). Neither 30 nor 60 µM Zn 2+ significantly influenced the growth of P. tricornutum (Figure 1a), while 60 µM Zn 2+ decreased the growth of C. fusiformis on day five (Figure 1b). This indicated that P. tricornutum and C. fusiformis show different sensitivities to Cu 2+ and Zn 2+ .
To explore the mechanism underlying the difference in susceptibility to heavy metals between P. tricornutum and C. fusiformis, transcriptomic analysis was conducted. Ion transport is reported to be a response mechanism to the high concentration of heavy metal ions. In this study, it has been shown that under high concentrations of both Cu 2+ and Zn 2+ , most DEGs involved in photosynthesis were upregulated, indicating the effect of both Cu 2+ and Zn 2+ on photosynthesis in P. tricornutum. Meanwhile, most genes downregulated in P. tricornutum under Cu 2+ treatment were involved in metal ion homeostasis and transmembrane ion transport. This indicated that ion homeostasis and transmembrane transport might be the main mechanisms for P. tricornutum to respond to high Cu 2+ concentrations. Moreover, this enrichment of downregulated genes in metal ion transport was observed in P. tricornutum under Zn 2+ treatment. However, the enrichment of downregulated genes in metal ion homeostasis-related genes did not occur under Zn 2+ treatment, indicating a different response mechanism for Zn 2+ to that for Cu 2+ .
Besides genes related to metal ion homeostasis and transmembrane ion transport, some other genes were previously reported to be involved in heavy metal stress response, including genes related to catalase, antioxidation, ascorbate metabolism, glutathione metabolism, phytochelatin, and oxidative stress [9]. These genes are listed in Tables 1-4. Most of these genes were upregulated in P. tricornutum under both Cu 2+ and Zn 2+ treatments; however, only a few were upregulated in C. fusiformis, indicating that the response of C. fusiformis to heavy metal ion stress is different from that of P. tricornutum. Moreover, the enrichment of DEGs in ion homeostasis and transmembrane transport-related genes was not observed in C. fusiformis either. It is reported that C. fusiformis is sensitive to heavy metal ions [4,6], whereas P. tricornutum is more tolerant to Cu 2+ stress [19]. We suspect that difference in gene expression might be one of the mechanism's responses to the difference in susceptibility to heavy metals between P. tricornutum and C. fusiformis.
In addition, since the metal toxicity for cells is more related to intracellular metal bioaccumulation than to the metal concentration in water, and the fact that both P. tricornutum and C. fusiformis are widely considered biofilm-producing mixed diatoms [30][31][32][33], the role of metal ion management of the biofilm should be considered when exploring the mechanism underlying the difference in susceptibility to heavy metals between P. tricornutum and C. fusiformis. Using SEM and EDS analysis, we found that both P. tricornutum and C. fusiformis accumulated Cu 2+ and Zn 2+ onto the biosilica shell. In future work it would be informative to determine the intracellular concentrations of Cu 2+ and Zn 2+ .

Conclusions
Transcriptome analysis of P. tricornutum and C. fusiformis under Cu 2+ and Zn 2+ treatments indicated that genes involved in metal ion homeostasis and transmembrane ion transport, and those related to catalase, antioxidation, ascorbate metabolism, glutathione metabolism, phytochelatin, and oxidative stress, might play important roles in the response of diatoms to heavy metal stress.