Abstract
Diatoms are a crucial component of marine ecosystems, recognized for
their broad environmental adaptability and wide temperature tolerance.
However, the molecular mechanisms underlying their adaptability to
diverse temperatures are unknown. In this study, we discover that heat
shock transcription factors (HSFs) are potentially important for
thermal tolerance in diatoms. Our study focuses on PtHSF2, annotated as
HSF2 in Phaeodactylum tricornutum’s genome, which is ubiquitous in
diatoms. Overexpression of PtHSF2 markedly enhances thermal tolerance
and increases cell size; causes significant differential expression of
several genes, including cell division cycle protein 45-like
(PtCdc45-like), ATM (ataxia telangiectasia mutated), ATR (ataxia
telangiectasia and Rad3-related), light-harvesting complex protein 2
(Lhcx2), and fatty acid desaturase. Cleavage Under Targets and
Tagmentation (CUT&Tag) and CUT&Tag-qPCR analyses demonstrate that
PtHSF2 directly targets and upregulates PtCdc45-like and Lhcx2 while
downregulating ATP-binding cassette transporter. Functional validation
of PtCdc45-like shows that its overexpression results in larger cell
size, enhances antioxidant capacity, and improves cell survival at
elevated temperatures. Collectively, our findings elucidate the
molecular mechanism by which PtHSF2 mediates high-temperature tolerance
in diatoms and validate the functions of its target gene PtCdc45-like.
These results highlight the importance of HSFs in diatom temperature
adaptation and provide insights into temperature acclimation in
microalgae.
Subject terms: Marine biology, Ecological genetics, Evolutionary
biology
__________________________________________________________________
Diatoms are crucial marine microorganisms, but the molecular mechanisms
underlying their temperature tolerance remain unclear. This study
identifies the heat shock transcription factor PtHSF2 as a key
regulator of thermal tolerance and cell size plasticity in
Phaeodactylum tricornutum, revealing its role in gene regulation and
stress response.
Introduction
Global warming, primarily driven by the greenhouse effect, has led to
the oceans sequestering over 90% of the excess heat on Earth since
1970^[50]1. The oceans play a pivotal role in maintaining the planet’s
thermal balance, yet the rise in sea temperatures has had detrimental
impacts on marine ecosystems. These impacts include substantial coral
reef degradation^[51]2, declines in marine fauna populations^[52]3, and
reductions in marine biodiversity^[53]4. Marine phytoplankton, critical
to oceanic ecosystems, contribute nearly half of global primary
productivity^[54]5 and are essential in regulating Earth’s climate.
Notably, among various phytoplankton groups such as diatoms,
dinoflagellates, cyanobacteria, and coccolithophores, diatoms are
distinguished by their extensive temperature tolerance^[55]6,
highlighting their pivotal role in responding to climate change.
Diatoms, which are vital to Earth’s ecosystems and account for
approximately one-fifth of global carbon dioxide fixation
annually^[56]7, are at the forefront of research in the context of
global warming. The warming trend has been linked to potential
reductions in diatom abundance worldwide by at least 10%, due to
increased ocean stratification^[57]8. However, there are perspectives
suggesting that this stratification may boost diatom productivity and
organic carbon output^[58]9. A groundbreaking study combining genomic
and phenotypic analyses reveals rapid evolutionary adaptation of
diatoms to higher temperatures^[59]10, underscoring their crucial role
as marine phytoplankton in adapting to climate change. Therefore,
investigating the molecular mechanisms underlying diatom thermal
tolerance is not only of scientific importance but also key to
foreseeing and understanding shifts in future marine ecosystems.
Heat shock transcription factors (HSFs) are highly conserved
DNA-binding proteins in eukaryotes, playing a crucial role in
environmental adaptation by finely tuning gene transcription and
enhancing stress resistance^[60]11–[61]13. Phaeodactylum tricornutum, a
model diatom species widely distributed in coastal and inland waters,
exhibits remarkable cell plasticity, including size variation and
morphological changes among oval, fusiform, and triradiate
forms^[62]14. This unique plasticity, combined with its exceptional
environmental adaptability, enables P. tricornutum to thrive in highly
unstable conditions, such as estuaries and rock pools^[63]14. Notably,
P. tricornutum exhibits significant temperature
adaptability^[64]15,[65]16, surviving in natural environments from the
tropics to subarctic regions^[66]17, and growing in laboratory
conditions from 5 to 28 °C^[67]18,[68]19. However, the molecular
mechanisms underlying its thermotolerance remain enigmatic.
The effects of temperature on diatom growth and physiology are
well-studied^[69]20–[70]22. High-throughput sequencing technologies
have paved the way for investigating the molecular mechanisms of diatom
responses to temperature stress. Many studies have explored these
mechanisms using genomics^[71]10, transcriptomics^[72]23–[73]25, and
proteomics^[74]26, providing valuable insights into diatom thermal
tolerance. However, these studies primarily focus on omics data, with
limited functional analyses of key temperature-responsive genes,
especially HSFs. In this study, we investigated the functional
mechanism of a conserved diatom HSF in P. tricornutum, which may offer
new insights into the molecular mechanisms underlying thermal tolerance
in diatoms.
Our findings revealed a widespread presence of HSFs in diatoms. For
example, HSFs are the most abundant transcription factors (TFs) in
diatoms such as Thalassiosira weissflogii, Thalassiosira pseudonana,
and P. tricornutum. In particular, the P. tricornutum genome comprises
69 genes encoding HSFs^[75]27, leading us to hypothesize that these
HSFs are potentially significant for the environmental adaptation of P.
tricornutum, especially under high-temperature conditions. To test
this, we employed overexpression and RNA interference strategies to
analyze the function of PtHSF2 in P. tricornutum. Physiological
analysis of PtHSF2 transgenic lines at four temperatures (15 °C, 20 °C,
25 °C, and 30 °C) revealed that overexpression of PtHSF2 significantly
enhanced heat tolerance, accompanied by a pronounced increase in cell
size. Furthermore, RNA sequencing (RNA-seq), Cleavage Under Targets and
Tagmentation (CUT&Tag), and CUT&Tag-qPCR were used to investigate the
regulatory mechanism of PtHSF2, which was found to be involved in the
transcriptional regulation of several genes. We verified the function
of the target gene cell division cycle protein 45-like (PtCdc45-like)
and evaluated the thermal tolerance of PtCdc45-like transgenic strains
in terms of redox levels and cell survival rates. These results
elucidate the critical role of PtHSF2 in regulating heat tolerance and
demonstrate that PtCdc45-like is an important target mediating the
involvement of PtHSF2 in the response to high-temperature conditions.
Results
PtHSF2 enhanced P. tricornutum’s tolerance to 30 °C
Diatoms, widely distributed in marine ecosystems, are ubiquitous in
nearly all aquatic habitats^[76]28. Different diatom species exhibit
distinct thermal tolerance ranges. For instance, Chaetoceros
tenuissimus grows optimally between 25 to 30 °C^[77]29, while
Fragilariopsis cylindrus thrives at −1 to 10 °C^[78]30. T. pseudonana,
Fistulifera solaris, and P. tricornutum have thermal tolerance ranges
of 10–35 °C^[79]10,[80]31, 15–35 °C^[81]32,[82]33, and
5–28 °C^[83]18,[84]19, respectively. To elucidate the molecular
mechanisms underlying the thermal adaptability of diatoms, an analysis
of transcription factor abundance and proportion was conducted in 15
diatom species with sequenced genomes. Results revealed that HSFs are
the most abundant transcription factors family in diatoms, ranging from
tens to hundreds of genes (Supplementary Table [85]1). For example, T.
weissflogii harbors 393 HSF-encoding genes, constituting 53% of its
total transcription factors, while dinoflagellates, coccolithophores,
and green algae possess fewer than 10 HSFs, accounting for only 0–7.2%
of their transcription factors. Given the global distribution and
extensive temperature adaptability of diatoms, it was hypothesized that
HSFs play a crucial role in their environmental adaptation. Consistent
with previous reports^[86]27, we found that P. tricornutum possessed 69
HSF genes, representing the highest proportion (44.2%) among all
transcription factors. An analysis of transcriptome data from the
MMETSP project^[87]34 and the re-assembly data by Johnson et al.^[88]35
indicated that 38 HSF genes are actively expressed in P. tricornutum
(Supplementary Data [89]1). Additionally, we observed the expression of
HSF orthologues in other species (Supplementary Data [90]1). We
speculate that these HSFs may play significant regulatory roles under
specific stress conditions, including high temperature, light stress,
and other environmental stresses.
Phylogenetic analysis categorized the 69 HSFs into four major groups:
Group I, Group II, Group III, and Group IV (Supplementary Fig. [91]1A).
Except for Group II, the other groups were further subdivided into
multiple subclasses. Group IV HSFs displayed closer homology with HSFs
from Emiliania huxleyi. Overall, P. tricornutum HSFs were more closely
related to each other than to HSFs from Chlamydomonas reinhardtii or
Arabidopsis thaliana. Notably, PtHSF2 (Phatr3_J55070) showed the most
significant upregulation among the differentially expressed HSFs under
elevated temperature conditions (Supplementary Data [92]2), with its
transcript levels varying significantly with temperature changes (15 °C
to 30 °C), peaking at 25 °C, followed by 15 °C, 30 °C, and 20 °C
(Fig. [93]1A), suggesting that PtHSF2 may play a crucial role in
temperature response.
Fig. 1. Physiological and morphological characterization of PtHSF2 transgenic
lines.
[94]Fig. 1
[95]Open in a new tab
A qPCR analysis of temperature effects on PtHSF2 transcription levels
in wild-type cells. Expression levels were normalized to β-actin and
TBP (TATA box binding protein) using the comparative Ct method. B
Schematic map of the PtHSF2 overexpression and RNAi expression
cassettes. Both expression cassettes were controlled with fcpC promoter
(PfcpC) and fcpA terminator (TfcpA) of fucoxanthin chlorophyll a/c
binding protein (fcp) gene of P. tricornutum. An omega leader motif was
inserted in between PfcpC and the PtHSF2 gene to enhance the
translation. Cell density monitoring (C), cell morphology observation
(D), and cell volume measurement (E) of PtHSF2 transgenic lines at
15 °C, 20 °C, 25 °C and 30 °C. Scale bars, 5 μm. Cell volumes for 100
cells were measured in each experiment. F Analysis of Nile Red (NR)
fluorescence intensity (FI) at 30 °C. G Cell survival rate at 30 °C. H
Analysis of reactive oxygen species (ROS) levels at 30 °C. A, C, E–H
Error bars represent mean values ± SD for three independent
experiments. Statistical significance was determined using two-tailed
unpaired Student’s t-test. Significant difference is indicated at
P < 0.05 (*) or P < 0.01 (**) level. WT, wild type; PtHSF2-1, −2, −3,
three PtHSF2 overexpressing strains; siPtHSF2-1, −2, −3, three PtHSF2
RNAi strains. Source data are provided as a Source Data file.
Furthermore, the analysis of the taxonomic distribution of PtHSF2
homologs in the ocean revealed that 99% of the hits in the marine
EUK-SMAGs (Eukaryotes Single-Cell and Metagenome Assembled Genomes)
database were from diatoms (Bacillariophyta) (Supplementary
Fig. [96]1B), indicating that diatoms are the primary hosts of this
gene. This finding suggests that PtHSF2 is widely present in diatoms
and may play a crucial role in their biological functions or adaptation
mechanisms. Supplementary Fig. [97]1C shows that the abundance of
PtHSF2 homologs is highest at higher latitudes and relatively low in
tropical regions, which could be attributed to differing environmental
conditions and ecological niches in these areas. Although diatoms
exhibit broad temperature tolerance, they generally thrive better in
cooler waters^[98]36.
SMART (Simple Modular Architecture Research Tool) domain annotation
identified a conserved HSF domain in PtHSF2 between amino acids 89 and
188 (Supplementary Fig. [99]2A). Annotated as HSF2 in the P.
tricornutum genome, it was named PtHSF2. Phylogenetic analysis
demonstrated that genes with higher homology to PtHSF2 were derived
from other diatom species, such as F. solaris, Pseudo-nitzschia
multistriata, and T. pseudonana (Supplementary Fig. [100]2B). To assess
PtHSF2’ regulatory function in microalgae, overexpression and RNA
interference (RNAi) plasmids were separately constructed
(Fig. [101]1B). Through genomic polymerase chain reaction (PCR)
(Supplementary Fig. [102]2C), quantitative real-time PCR (qPCR)
(Supplementary Fig. [103]2D), and western blotting (Supplementary
Fig. [104]2E), three overexpression lines (PtHSF2-1, PtHSF2-2, and
PtHSF2-3) and three silenced lines (siPtHSF2-1, siPtHSF2-2, and
siPtHSF2-3) were successfully established. We measured PtHSF2 protein
levels in the PtHSF2 overexpression strain (PtHSF2-proH1) at four
temperatures (15 °C, 20 °C, 25 °C, and 30 °C) and found a significant
increase at elevated temperatures (25 °C and 30 °C) (Supplementary
Fig. [105]2F).
Alignment analysis of the coding regions of the remaining 68 HSF genes
from P. tricornutum with dsPtHSF2 revealed that only 8 HSF genes,
including Phatr3_J38294, Phatr3_J48667, Phatr3_J48572, Phatr3_J42824,
Phatr3_J45391, Phatr3_EG02624, Phatr3_J50136, and Phatr3_EG00212,
contained perfectly matching segments with dsPtHSF2, all of which were
shorter than 16 bp (Supplementary Fig. [106]3A). According to Chen et
al.^[107]37, such short matches are insufficient to cause off-target
effects. Furthermore, analysis of the relative transcription levels of
these 8 HSF genes showed no significant downregulation (Supplementary
Fig. [108]3B). These results indicate that RNAi targeting PtHSF2 did
not cause off-target effects among its paralogues.
To characterize the thermal tolerance of PtHSF2-transformed microalgae,
cells were cultured separately at 15 °C, 20 °C, 25 °C, and 30 °C. Cell
density monitoring revealed that PtHSF2 overexpression cell lines
exhibited lower growth rates at 15 °C, 20 °C, and 25 °C compared to WT
cells and the RNAi lines (Fig. [109]1C). While temperatures above 25 °C
impaired growth, and 30 °C exposure for over three days severely
inhibited it, PtHSF2 overexpression lines had significantly higher cell
density than WT cells at 30 °C (Fig. [110]1C). Confocal microscopy and
Nile red staining revealed increased cell size and neutral lipid
content in overexpression lines, particularly at 30 °C
(Fig. [111]1D–F). Additionally, lipid droplets were observed outside of
the PtHSF2-1 cells under 30 °C, potentially due to cell rupture or
active secretion (Fig. [112]1D). Previous studies have demonstrated
that microalgae actively secrete lipid droplets during periods of high
lipid synthesis^[113]38. Quantitative analysis showed that cell volume
in overexpression lines was 2.1–2.4 times larger than WT cells at
30 °C, while knockdown lines were significantly smaller (Fig. [114]1E).
Thermal stress often impairs photosystem II (PSII), leading to reactive
oxygen species (ROS) production and lipid peroxidation^[115]39,[116]40.
Therefore, ROS and lipid peroxides, such as malondialdehyde (MDA), an
abundant byproduct of lipid peroxidation^[117]41, serve as indicators
of oxidative stress. Analysis of cell survival rates demonstrated that
PtHSF2 overexpressing lines exhibited significantly higher survival
rates at 30 °C compared to WT cells (Fig. [118]1G). Moreover,
overexpression lines exhibited significantly lower ROS (Fig. [119]1H)
and MDA levels (Supplementary Fig. [120]4A) and higher antioxidant
enzyme activities, including superoxide dismutase (SOD) and catalase
(CAT) (Supplementary Fig. [121]4B, C). The lipid saturation index was
significantly lower in the overexpressing lines and higher in the RNAi
lines compared to the WT (Supplementary Fig. [122]4D). These results
suggest that PtHSF2 may mediate the cell’s thermal response by
regulating the expression of genes related to cell size, lipid
synthesis, and antioxidant defense.
Deciphering PtHSF2’s regulatory mechanism through integrated RNA-seq and
CUT&Tag analyses
We employed RNA-seq, CUT&Tag, and CUT&Tag-qPCR techniques to unravel
the functional mechanisms governed by PtHSF2. The number and quality of
sequencing reads from RNA-seq and CUT&Tag libraries were presented in
the Supplementary Data [123]3 and [124]4, highlighting the sequencing
depth and overall data quality across all samples. RNA-seq analysis
revealed 1912 significantly differentially expressed genes (DEGs) in
the PtHSF2-1 versus WT comparison, with 940 upregulated and 972
downregulated. In the PtHSF2-2 versus WT comparison, 2,404 DEGs were
identified, including 1,121 upregulated and 1,283 downregulated. Among
these, 1449 DEGs were shared between the two PtHSF2 overexpressing
lines, comprising 676 upregulated and 773 downregulated (Supplementary
Data [125]5). Notably, genes encoding Cdc45-like (Phatr3_J9344),
light-harvesting complex protein 2 (Lhcx2, Phatr3_EG02404), fatty acid
desaturase (FAD, Phatr3_J46275), as well as ataxia telangiectasia
mutated (ATM, Phatr3_J51040) and ataxia telangiectasia and Rad3-related
(ATR, Phatr3_J44708), were significantly upregulated, while an
ATP-binding cassette (ABC) transporter (Phatr3_J42881) containing two
ABC_ATPase domains, which shares high homology with F. solaris and
Cylindrotheca closterium, was significantly downregulated
(Supplementary Data [126]6–[127]8 and Supplementary Fig. [128]5A, B).
Additionally, the expression levels of 12 PtHSF2 homologs exhibited
significant changes, with 5 genes upregulated and 7 downregulated
(Supplementary Data [129]8).
Gene Ontology (GO) annotation and enrichment analysis of the 1449 DEGs
revealed that 30.9% participated in cellular processes under the
“biological process” category, with 15.0% upregulated and 15.9%
downregulated (Fig. [130]2A and Supplementary Data [131]9). In the
“molecular function” category, 29.4% of DEGs were associated with
catalytic activity, with 12.6% upregulated and 16.8% downregulated
(Fig. [132]2B). For the “cellular component” category, DEGs enriched in
“cell” and “cell part” terms both accounted for 36.7%, comprising 16.9%
upregulated and 19.8% downregulated genes (Fig. [133]2C). KEGG pathway
analysis unveiled significant enrichment of genes involved in cell
cycle regulation, antioxidant defense, fatty acids synthesis,
AMP-activated protein kinase (AMPK) signaling, and carbohydrate
metabolism (Supplementary Data [134]6). Among these, two mitochondrial
inner membrane proteins, MPV17 (Phatr3_J15341 and Phatr3_Jdraft223),
associated with antioxidant processes^[135]42, were significantly
upregulated (Supplementary Data [136]6). Collectively, these results
highlight that PtHSF2 overexpression profoundly influences genes linked
to cell cycle regulation and environmental information processing.
Fig. 2. Combined RNA-seq and CUT&Tag resolved the regulatory mechanism of
PtHSF2.
[137]Fig. 2
[138]Open in a new tab
Gene Ontology (GO) analysis (Level 2) of the top 5 significantly
enriched terms for differentially expressed genes (DEGs) between PtHSF2
overexpressing cells and WT cells during the stationary phase at 20 °C.
A Biological Process; B Molecular Function; C Cellular Component. The
x-axis label “Genes involved (%)” in panels A–C represents the
percentage of DEGs associated with a specific GO term within the
overlap between the two PtHSF2 overexpressing lines, relative to the
total of 1449 DEGs identified. D–F GO analysis of potential target
genes identified by CUT&Tag. The x-axis label “Genes involved (%)” in
panels D-F indicates the percentage of potential target genes
associated with a specific GO term among the 295 target genes
identified. G The overlap between the CUT&Tag and RNA-seq, as analyzed
in the Venn diagram. H Distribution of CUT&Tag reads around the
transcription start site (TSS). IgG, IgG CUT&Tag; PtHSF2, Flag CUT&Tag.
I The top three DNA binding motifs of PtHSF2 as revealed by CUT&Tag
analysis. CUT&Tag-qPCR was performed to validate the binding sites of
PtHSF2, including J the promoter of Cdc45-like; K the promoter of
Lhcx2; and L the promoter of ABC. The fold enrichment was normalized to
the promoters of β-actin and TBP. IgG (IgG CUT&Tag) and WT (WT CUT&Tag)
were used as the control groups. Relative transcription levels of
potential target genes (M) Cdc45-like, (N) Lhcx2, (O) ABC, and (P) FAD
were determined by qPCR and normalized to β-actin and TBP. J–P
Statistical significance was determined using two-tailed unpaired
Student’s t-test. Significant difference is indicated at P < 0.05 (*)
or P < 0.01 (**) level. Each value represents mean ± SD (n = 3
biological replicates). Source data are provided as a Source Data file.
CUT&Tag, a potent method for investigating transcription factor-DNA
interactions^[139]43, identified 295 potential PtHSF2 target genes
(Supplementary Data [140]10), which underwent GO annotation analysis
(Fig. [141]2D–F). These target genes were enriched in catalytic (22.4%)
and binding activities (31.2%), participating in intracellular
biological processes (32.9%) (Fig. [142]2D, E). Moreover, a substantial
proportion was associated with membrane components (19.3%) and cellular
components (38.3%) (Fig. [143]2F and Supplementary Data [144]10),
suggesting PtHSF2’s involvement in regulating cellular structures and
membrane composition.
To identify PtHSF2’s binding sites, CUT&Tag and CUT&Tag-qPCR analyses
were performed. The Venn diagram (Fig. [145]2G) revealed 35
intersecting genes between potential targets and DEGs (Supplementary
Data [146]11), comprising 18 upregulated and 17 downregulated genes,
including Cdc45-like (Phatr3_J9344), Lhcx2 (Phatr3_EG02404), and the
ABC transporter (Phatr3_J42881). The limited overlap between DEGs and
potential targets likely stems from several factors. First, TFs can
regulate gene expression both directly and indirectly^[147]44, with
some targets showing no detectable expression changes^[148]45–[149]47.
Second, while RNA-seq offers broad coverage, it lacks specificity,
whereas CUT&Tag provides high specificity by targeting proteins of
interest with specific antibodies^[150]43. Signal intensity analysis
around transcription start sites (TSS) showed a significant enrichment
peak at 8.25 for PtHSF2, indicating its primary binding sites at the
TSS regions, which were absent in IgG controls (Fig. [151]2H). Using
the MEME algorithm^[152]48, we identified the top three DNA-binding
motifs for PtHSF2 (Fig. [153]2I), with the most prominent motif being
the sequence “GGTTAGGGTTAGGGTTAGGGTTAGGGTTAG.” This sequence represents
a putative consensus motif, indicating a conserved recognition pattern
across genomic sites. CUT&Tag-qPCR further confirmed PtHSF2’s binding
to promoter regions of Cdc45-like (Phatr3_J9344), Lhcx2
(Phatr3_EG02404), and ABC (Phatr3_J42881) transporter genes in both
overexpressing lines (Fig. [154]2J–L).
TFs can positively or negatively regulate the transcription of target
genes^[155]38,[156]44,[157]49. To elucidate how PtHSF2 modulates its
targets, we analyzed transcription levels of potential target genes. In
PtHSF2-overexpressing cells, Cdc45-like and Lhcx2 were significantly
upregulated, while the ABC transporter was notably downregulated
(Fig. [158]2M–O). The upregulation of FAD (Phatr3_J46275) in
overexpressing lines and its downregulation in RNAi lines (Fig. [159]2P
and Supplementary Data [160]6) corresponded with changes in unsaturated
fatty acid content (Supplementary Fig. [161]4D). However, CUT&Tag did
not detect direct targeting of the FAD promoter, suggesting that PtHSF2
regulates FAD indirectly. Interestingly, the regulatory trends for
target genes at 30 °C (Supplementary Fig. [162]6A-F) were consistent
with those observed at 20 °C (Fig. [163]2M–P; Supplementary
Data [164]5), suggesting that PtHSF2’s regulatory mechanisms remain
stable across these temperature conditions.
Functional identification of PtCdc45-like
In the WT background, PtCdc45-like transcription increased from 15 °C
to 25 °C but declined at 30 °C compared to 25 °C (Fig. [165]3A),
suggesting its critical role in high-temperature responses. To
elucidate PtCdc45-like’s biological function, we employed
bioinformatics and molecular biology approaches. Bioinformatics
analysis revealed that PtCdc45-like coding region spans 2274 bp,
encoding a conserved Cdc45-like protein domain between amino acids 29
and 754 (Supplementary Fig. [166]7A). Phylogenetic analysis showed that
PtCdc45-like is evolutionarily conserved among diatoms (Supplementary
Fig. [167]7B).
Fig. 3. Functional validation of the potential target gene PtCdc45-like.
[168]Fig. 3
[169]Open in a new tab
A qPCR analysis of PtCdc45-like transcription levels in WT cells across
four temperatures (15 °C, 20 °C, 25 °C, 30 °C). β-actin and TBP were
used as reference genes. B Schematic representation of the PtCdc45-like
overexpression cassette and RNAi expression cassette to silence
PtCdc45-like. C Monitoring of cell number. D Cell volume analysis. Cell
volumes were measured for 100 cells in each experiment. E Observation
of cell morphology in 20 °C-cultured cells by confocal laser scanning
microscope. Scale bars, 5 μm. F Determination of cell dry weight in the
plateau phase. G Lipid saturation index analysis. H Maximum quantum
yield of photosystem II (PSII) as measured by Fv/Fm. Data are presented
as means ± SD from three independent experiments. Significance was
calculated by two-tailed unpaired Student’s t test. Differences were
considered significant at P < 0.05 (*) and P < 0.01 (**) (A, C, D,
F-H). Source data are provided as a Source Data file.
To characterize its function, overexpression and RNAi plasmids were
constructed (Fig. [170]3B). Putative positive transformants were
selected based on bleomycin resistance and validated through genomic
PCR, qPCR, and western blot analyses, resulting in six confirmed lines:
three overexpression lines (PtCdc45-like-1, −2, −3) and three RNAi
lines (siPtCdc45-like-1, −2, −3) (Supplementary Fig. [171]7C–E). qPCR
analysis revealed transcription levels of PtCdc45-like were 2.6- to
3.8-fold higher in the overexpression lines and reduced to 20.2–29.3%
of WT levels in RNAi lines (Supplementary Fig. [172]7D). Western
blotting confirmed specific protein expression (~86.9 kDa) in
transformants but not in WT cells (Supplementary Fig. [173]7E). Protein
level analysis across varying temperatures demonstrated a
temperature-dependent increase in PtCdc45-like protein levels in the
PtCdc45-like-proC1 overexpression strain (Supplementary Fig. [174]7F).
Overexpressing lines exhibited significantly reduced cell densities
compared to WT and RNAi lines at 15–25 °C, but notably higher densities
at 30 °C (Fig. [175]3C). Under normal growth conditions (20 °C),
overexpression lines displayed significantly larger cell volumes than
WT, while RNAi lines had smaller cells, with cell size increasing
further at higher temperatures (Fig. [176]3D, E). These results
highlight PtCdc45-like as a key regulator of cell size in P.
tricornutum. Notably, the total dry weight of cell pellets, normalized
by cell number, revealed that larger cell volumes correlated with
higher average cell dry weight (Fig. [177]3F).
Cell plasticity is influenced by membrane fluidity and cell wall
flexibility. The fusiform structure of P. tricornutum, lacking a
silicified cell wall^[178]50, may facilitate cell size plasticity.
Consistent with this, overexpression lines exhibited a lower lipid
saturation index, while RNAi lines had higher indices (Fig. [179]3G).
This suggests a positive correlation between cell size and the
proportion of unsaturated fatty acids, supported by the upregulation of
FAD (Phatr3_J46275) in overexpression lines and its downregulation in
RNAi lines (Supplementary Fig. [180]7G). These findings align with
those observed in PtHSF2 transgenic lines (Fig. [181]2P) and prior
studies linking desaturase activity to unsaturated fatty acid
synthesis^[182]51,[183]52. In microalgae, fatty acids primarily exist
as triacylglycerols (TAGs), phospholipids, and glycolipids, with
phospholipids and glycolipids serving as the primary components of
membranes^[184]53. Accordingly, we hypothesize that larger cells had a
higher proportion of unsaturated fatty acids in their cell membranes,
contributing to enhanced membrane fluidity.
Overexpression lines exhibited lower maximum quantum yield of
photosystem II (Fv/Fm) than WT, while RNAi lines showed no difference
(Fig. [185]3H). Interestingly, Lhcx2 (Phatr3_EG02404) transcription
decreased in overexpression lines but increased in RNAi lines
(Supplementary Fig. [186]7H), contrasting with its regulation in
PtHSF2-transformed lines (Fig. [187]2N).
Overexpression of both PtCdc45-like and PtHSF2 resulted in comparable
phenotypic changes, including increased cell size and altered lipid
composition. Furthermore, the ABC transporter (Phatr3_J42881) was
significantly downregulated in both PtCdc45-like and PtHSF2
overexpressing lines and upregulated in the RNAi lines (Fig. [188]2O
and Supplementary Fig. [189]7I). This transporter’s ABC_ATPase domain
(Supplementary Fig. [190]5A), associated with mitochondrial mRNA
processing^[191]54, suggests that its downregulation may impair
mitochondrial function, contributing to slower growth.
Cell cycle progression analysis revealed that overexpression lines
(PtHSF2-1 and PtCdc45-like-1) had higher proportions of cells in the S
and G2 phases, indicating accumulation or arrest, whereas RNAi lines
displayed distributions similar to WT cells (Supplementary
Fig. [192]8A). This is consistent with differential expression of cell
cycle-related genes in overexpression lines (Supplementary Data [193]6;
Supplementary Fig. [194]6E, F; Supplementary Fig. [195]8B;
Supplementary Fig. [196]9), which may contribute to the observed cell
accumulation or arrest during these phases. Among these genes, the key
kinases regulating cell cycle and replication checkpoints, ATM and
ATR^[197]55, were significantly upregulated, suggesting the activation
of cell cycle checkpoints. In addition to these cell cycle-related
genes, PtCdc45-like significantly altered the transcription levels of
FAD, Lhcx2, and ABC (Supplementary Fig. [198]7G–I). Specifically, its
effects on FAD and ABC were consistent with those of PtHSF2, whereas
its effect on Lhcx2 was opposite.
Mutual regulation between PtHSF2 and PtCdc45-like
To investigate the regulatory relationship between PtCdc45-like and the
transcriptional variability of FAD, Lhcx2, ABC, ATM, and ATR we first
examined PtHSF2 transcription in PtCdc45-like overexpression and RNAi
lines. The results revealed that PtHSF2 was upregulated in the
overexpression lines and downregulated in the RNAi lines
(Fig. [199]4A). Combined with our earlier observation that PtHSF2
activates PtCdc45-like transcription, this finding suggests the
existence of a positive feedback loop between PtHSF2 and PtCdc45-like.
Fig. 4. Mutual regulation between PtHSF2 and PtCdc45-like.
[200]Fig. 4
[201]Open in a new tab
A qPCR analysis of PtHSF2 transcript levels in PtCdc45-like transgenic
lines at 20 °C. Relative transcript levels of FAD (B), ABC (C), Lhcx2
(D), ATM (E), and ATR (F) in the Ho-Ci and Hi-Co lines at 20 °C.
PtHSF2-1, PtHSF2 overexpressing line 1; Ho-Ci-1/2/3, three cell lines
with PtHSF2 overexpression and PtCdc45-like knockdown; siPtHSF2-1,
PtHSF2 RNAi line 1; Hi-Co-1/2/3, three cell lines with PtCdc45-like
overexpression and PtHSF2 knockdown. β-actin and TBP were used as
reference genes. G Cell density at 20 °C and 30 °C. H Cell volume
analysis at 20 °C. In each experiment, the volume of 100 cells was
measured. The data are expressed as means ± SD, based on three
independent experiments. Statistical significance was determined using
two-tailed unpaired Student’s t test. Differences were considered
significant at P < 0.05 (*) and P < 0.01 (**) (A, D–H). Source data are
provided as a Source Data file.
To further elucidate the regulatory interactions, two types of
transgenic cell lines were constructed: Ho-Ci (PtHSF2 overexpression
combined with PtCdc45-like knockdown) and Hi-Co (PtHSF2 knockdown
combined with PtCdc45-like overexpression). Six lines—Ho-Ci-1, Ho-Ci-2,
Ho-Ci-3, Hi-Co-1, Hi-Co-2, and Hi-Co-3—were successfully validated
(Supplementary Fig. [202]10A–F). Next, we assessed FAD and ABC
transcription levels in these cell lines. The results indicated that
knocking down PtCdc45-like in the PtHSF2 overexpression background did
not affect FAD and ABC, and similarly, overexpressing PtCdc45-like in
the PtHSF2 knockdown background had no impact on these genes
(Fig. [203]4B, C). These findings suggest that the regulation of FAD
and ABC is predominantly mediated by PtHSF2, with the influence of
PtCdc45-like being exerted indirectly through its positive feedback on
PtHSF2.
In contrast, Lhcx2 transcription was significantly altered in both
Ho-Ci and Hi-Co lines, as knocking down PtCdc45-like in the PtHSF2
overexpression background increased Lhcx2 while overexpressing
PtCdc45-like in the PtHSF2 knockdown background decreased Lhcx2
(Fig. [204]4D), indicating that Lhcx2 transcription is influenced by
PtCdc45-like. Combined with the data from Fig. [205]2K, N, these
results suggest that Lhcx2 is co-regulated by PtHSF2 and PtCdc45-like,
with PtCdc45-like acting as a repressor. Whether this repression is
direct or indirect remains to be investigated. ATM and ATR were
markedly downregulated in the Ho-Ci lines but upregulated in the Hi-Co
lines (Fig. [206]4E, F). Moreover, both genes were significantly
upregulated in PtHSF2 and PtCdc45-like overexpression lines
(Supplementary Data [207]6; Supplementary Fig. [208]6E, F;
Supplementary Fig. [209]8B; Supplementary Fig. [210]9), suggesting that
PtHSF2 may regulate these genes via PtCdc45-like.
Under normal growth conditions (20 °C), knocking down PtCdc45-like in
the PtHSF2 overexpression background alleviated the growth retardation
(Fig. [211]4G) and cell enlargement (Fig. [212]4H) caused by PtHSF2
overexpression. However, under heat stress (30 °C), PtCdc45-like
knockdown significantly reduced cell density (Fig. [213]4G).
Conversely, overexpressing PtCdc45-like in the PtHSF2 knockdown
background yielded the opposite results (Fig. [214]4G, H). These
findings highlight PtCdc45-like as a critical mediator of
PtHSF2-regulated cell size and thermal response. Overall, PtHSF2
positively regulates PtCdc45-like and Lhcx2 while negatively regulating
ABC and indirectly regulating FAD, with a positive feedback loop
between PtHSF2 and PtCdc45-like, and PtCdc45-like acting as a repressor
of Lhcx2.
PtCdc45-like overexpression improved thermal tolerance
The optimum growth temperature for P. tricornutum (CCMP2561) under
laboratory conditions was 20 °C. Elevated temperatures negatively
impacted cellular growth, often leading to growth inhibition or cell
death. Fig. [215]5A demonstrates that the inhibition rates at 25 °C and
30 °C were significantly higher in WT cells compared to PtCdc45-like
overexpressing cells. Notably, PtCdc45-like overexpressing cells did
not show a growth advantage at lower temperatures (15 °C and 20 °C)
(Fig. [216]3C), suggesting that PtCdc45-like overexpression enhances
heat tolerance rather than cold tolerance. To further investigate the
thermal tolerance conferred by PtCdc45-like overexpression, we assessed
cell survival rates and redox levels under high-temperature conditions
(25 °C and 30 °C). Given the clear growth trend observed at 25 °C
(Fig. [217]3C), survival rate analysis was not performed at this
temperature. After 13 days of incubation at 30 °C, the survival rate of
WT cells was only 23.6%, whereas PtCdc45-like overexpression lines
(PtCdc45-like-1, PtCdc45-like-2, and PtCdc45-like-3) exhibited
significantly higher survival rates of 76.7, 72.5, and 65.0%,
respectively (Fig. [218]5B).
Fig. 5. Response of PtCdc45-like transgenic lines to elevated temperatures
(25 °C and 30 °C).
[219]Fig. 5
[220]Open in a new tab
A Growth inhibition rates of WT, PtCdc45-like overexpression line, and
PtCdc45-like RNAi line at different temperatures 15 °C, 25 °C, and
30 °C were assessed by comparing cell densities on day 13 at each
temperature to those at the optimal growth temperature (20 °C). B Cell
survival rate at 30 °C. C ROS content, D MDA content, E SOD activity,
and F CAT activity under high temperatures. Data in this figure are
shown as mean ± SD. Three independent experiments were performed.
Two-tailed unpaired t-test against WT. Significant difference is
indicated at P < 0.05 (*) or P < 0.01 (**) level (A–F). Source data are
provided as a Source Data file.
Furthermore, ROS (Fig. [221]5C) and MDA (Fig. [222]5D) levels were
markedly lower in PtCdc45-like overexpressing cells compared to WT
cells, indicating a diminished sensitivity to high-temperature stress.
Since oxygenic photosynthesis is a major source of ROS
production^[223]56, the relatively lower ROS levels observed in
PtCdc45-like overexpressing cells could be attributed to the
significant suppression of Fv/Fm (Fig. [224]3H) and the downregulation
of Lhcx2 transcription (Supplementary Fig. [225]7H). Despite no
significant increase in SOD levels in overexpressing cells at 30 °C
compared to WT cells (Fig. [226]5E), CAT levels were markedly higher in
the overexpressing cells (Fig. [227]5F). Conversely, the PtCdc45-like
RNAi lines showed no significant differences in antioxidant enzyme
levels compared to WT (Fig. [228]5D–F). Collectively, these data
provided compelling evidence that PtCdc45-like overexpression augmented
cellular antioxidant defenses, thereby improving cell tolerance to
elevated temperatures.
Both PtHSF2 and PtCdc45-like were identified as crucial contributors to
heat tolerance in P. tricornutum. However, it remains unclear whether
this mechanism is conserved across other diatom species. To
investigate, we subjected T. weissflogii to 20 °C, 30 °C, and 35 °C,
followed by transcriptome analysis (quality control details in
Supplementary Data [229]12). Compared to 20 °C, the PtHSF2 orthologue
in T. weissflogii (snap_masked-ctg119-processed-gene-4.14-mRNA-1, 48.5%
similarity) was upregulated 1.21-fold at 30 °C and 1.92-fold at 35 °C,
with the PtCdc45-like orthologue
(snap_masked-ctg1929-processed-gene-0.66-mRNA-1, 53.4% similarity) also
exhibiting an upregulation trend (Supplementary Data [230]13). These
results suggest that HSF2 may serve as a key regulator of heat
tolerance in diatoms and that this regulatory mechanism may be broadly
conserved across diatom species.
Discussion
HSFs have traditionally been known as key regulators of heat shock
protein (HSP) transcription in response to heat stress^[231]11.
However, emerging evidence reveals that HSFs also regulated the
transcription of non-HSP genes^[232]57,[233]58. For instance, mammalian
HSF1 targets tumor necrosis factor-α (TNFα)^[234]59 and interleukin-6
(IL-6)^[235]60, while in Caenorhabditis elegans, HSF-1 upregulates
genes such as acdh-1 (acyl-CoA dehydrogenase 1) and vitellogenin lipid
transporters genes (vit-1, vit-3, vit-4, and vit-5) under non-stress
conditions^[236]61. Similarly, Saccharomyces cerevisiae HSF activates
RPN4 (Regulatory particle non-ATPase subunit 4) and PDR3 (Pleiotropic
drug resistance 3)^[237]62. HSF2 also displays diverse regulatory
roles. In mammals, HSF2 modulates c-fos^[238]63, HIF-1α
(Hypoxia-inducible factor-1α)^[239]64, EFEMP1 (EGF-containing
fibulin-like extracellular matrix protein 1), and PROM2 (Prominin
2)^[240]65. In A. thaliana, HSF2 regulates genes such as APX2
(ascorbate peroxidase 2)^[241]66,[242]67, GolS1/S2 (galactinol synthase
1 and 2), and IPS2 (inositol 3-phosphate synthase 2)^[243]66. Among
these previously identified HSF2 targets, none were regulated by PtHSF2
in our study. Instead, we discovered that PtHSF2 uniquely regulates
PtCdc45-like, Lhcx2, and ABC transporter genes, alongside influencing
FAD, ATM, and ATR transcription (Fig. [244]2; Supplementary
Data [245]6; Fig. [246]6). These genes, unrelated to HSPs, underscore
PtHSF2’s broader functional role in regulating cell size and thermal
tolerance, thereby expanding the known functions of HSFs beyond
traditional paradigms.
Fig. 6. Schematic representation of the mechanistic role of PtHSF2 in P.
tricornutum.
[247]Fig. 6
[248]Open in a new tab
Characteristically, overexpression of PtHSF2 enhanced tolerance to high
temperatures and increased cell size. RNA-seq analysis and qPCR
validation revealed significant differential expression of key genes,
including PtCdc45-like, Lhcx2, ABC, FAD, ATM, and ATR. CUT&Tag and
CUT&Tag-qPCR analyses confirmed that PtCdc45-like, Lhcx2, and ABC are
directly targeted and regulated by PtHSF2. Functional validation of
PtCdc45-like demonstrated that its overexpression resulted in increased
cell size and enhanced thermal tolerance. Mechanistically, PtCdc45-like
likely exerts these effects by upregulating ATM and ATR expression,
thereby activating DNA replication and cell cycle checkpoints to
promote the repair of DNA damage caused by high temperatures.
Additionally, the activation of these checkpoints induced cell cycle
arrest, contributing to the observed enlargement of cells. Solid red
and green arrows denote the direct targets of PtHSF2, while dashed red
and green arrows indicate its indirect targets. Up- and down-regulated
genes are indicated in red and green, respectively. Gray arrows
indicated biological processes involved.
Cell size plasticity enables microorganisms to adapt to environmental
changes^[249]68–[250]70. In P. tricornutum, cell size has been shown to
increase with rising temperatures. For instance, earlier studies
reported a 10% increase in cell volume from 18 °C to 25 °C^[251]16,
while cells grown at 23–25 °C were larger than those at lower
temperatures^[252]71. Under nitrogen limitation, cell size
progressively increased with temperatures ranging from 20 °C to
30 °C^[253]72. Similar trends have been observed in other diatoms. In
T. pseudonana, temperatures above the thermal optimum reduced cell
growth but increased cell volume^[254]73. Studies on T.
pseudonana^[255]10 and Skeletonema marinoi^[256]23 further revealed
that cells adapted to higher temperatures generally exhibit larger
volumes. However, this behavior is not universal. Some diatoms, such as
Amphora coffeaeformis and Nitzschia ovalis, exhibit decreased cell
sizes with rising temperatures^[257]74, and similar reductions have
been noted in other species under elevated temperatures^[258]24. While
the literature and our findings (Fig. [259]1E and Fig. [260]3D)
indicate a modest increase in cell volume in response to elevated
temperature, there is currently no conclusive evidence to establish a
direct causal relationship between cell volume expansion and enhanced
thermal tolerance. The observed increase in cell volume likely
represents a general physiological response of P. tricornutum to
high-temperature stress, rather than a specific adaptive mechanism
directly contributing to thermal tolerance.
Our study linked PtHSF2 and PtCdc45-like to increased cell size and
thermal response in P. tricornutum (Fig. [261]6). In WT cells, PtHSF2
expression peaked at 25 °C, with lower levels at 15 °C, 30 °C, and
20 °C (Fig. [262]1A), while PtCdc45-like transcription increased from
15 °C to 25 °C but decreased at 30 °C compared to 25 °C (Fig. [263]3A).
Although their transcription levels did not continuously increase with
rising temperatures, their significant upregulation at elevated
temperatures (25 °C and 30 °C) compared to the optimal growth
temperature (20 °C), coupled with the temperature-dependent increase in
protein levels in overexpression strains (Supplementary Fig. [264]2F;
Supplementary Fig. [265]7F), indicates a robust heat-stress response.
In both bacteria and eukaryotes, intracellular protein concentrations
are generally correlated with the abundances of their corresponding
mRNAs, but the correlation is not particularly strong, with the squared
Pearson correlation coefficient (R²) typically around 0.40^[266]75.
This suggests that a decrease in transcription levels does not
necessarily result in a proportional decrease in protein abundance. In
addition, nonlinear transcriptional responses to environmental stimuli
are well-documented. For example, in yeast species exposed to
temperatures ranging from 10 °C to 46 °C, Hsp70 expression peaked
6–8 °C above the optimal growth temperature but declined at higher
temperatures^[267]76. In Heterosigma akashiwo, CO₂-responsive genes
such as SLC4 (SLC4 family bicarbonate transporter), ALAT_GGAT (alanine
aminotransferase), and CA (carbonic anhydrase) showed peak
transcription at 400–600 ppm CO₂, decreasing at higher
concentrations^[268]77. Similarly, in human embryonic kidney cells and
cochlear tissues under cisplatin-induced stress, stress-responsive
genes like Chop (C/EBP homologous protein) and DR5 (death receptor 5)
increased with rising cisplatin concentrations but declined beyond a
threshold^[269]78. In nucleus pulposus cells, TIMP-1 (tissue inhibitor
of metalloproteinases 1) expression initially decreased with mild
acidity (pH 7.4–7.1) but increased as pH dropped further (pH
7.1–6.5)^[270]79. These findings suggest that transcriptional responses
often plateau or decline once environmental stress exceeds a critical
threshold.
HSFs are well-established mediators of diverse abiotic stress
responses^[271]80, with individual HSFs often responding to multiple
stressors^[272]81. For instance, in wheat seedlings, TaHSF3 is
upregulated under both low and high temperatures, and its
overexpression in Arabidopsis has been shown to enhance tolerance to
extreme temperatures^[273]82. Similarly, our findings demonstrate that
PtHSF2 was significantly upregulated at both low (15 °C) and high
temperatures (25 °C and 30 °C) (Fig. [274]1A), underscoring its dual
role in cold and heat stress responses. Interestingly, PtCdc45-like
transcription was not upregulated at 15 °C (Fig. [275]3A), suggesting
the involvement of alternative regulatory pathways under cold stress,
potentially explaining the lack of corresponding cell size increase
despite elevated PtHSF2 levels.
Cell size in unicellular microalgae is influenced by environmental
factors and growth status. For example, iron deficiency decreases
diatom cell volume^[276]83,[277]84, while elevated temperatures tend to
increase cell volume^[278]23. Cell size is also closely linked to
growth and division^[279]85, regulated by the transcriptional
expression of specific genes^[280]85,[281]86. In our study, under
identical culture conditions (with and without temperature stress),
PtCdc45-like overexpression led to significantly larger cell volumes
compared to WT cells, while knock-down lines exhibited smaller cell
sizes (Fig. [282]3D, E). These findings indicate that PtCdc45-like is
involved in the regulation of cell size in P. tricornutum.
Cdc45 is a vital protein in eukaryotic cell cycles, forming the CMG
(Cdc45/Mcm2-7/GINS) complex with Mcm2-7 (Minichromosome Maintenance
proteins 2–7) and GINS (Go-Ichi-Ni-San), functioning as a helicase
during DNA replication in S phase^[283]87. Cell size homeostasis
involves a balance between growth and division^[284]85, with
disruptions to cell cycle stages impacting division processes^[285]88.
In this study, PtHSF2- and PtCdc45-like-overexpressing cells exhibited
reduced growth rates, larger volumes, and dual chloroplasts
(Fig. [286]1C–E; Fig. [287]3C–E). Chloroplast division in P.
tricornutum occurs during the S phase^[288]89. Hence, we speculated
that the overexpression of PtHSF2 and PtCdc45-like disrupted cell cycle
progression, leading to cell accumulation or arrest in the S or G2
phase. Cell cycle progression analyses further supported this
hypothesis (Supplementary Fig. [289]8A). Previous studies have shown
that Cdc45 overexpression induces DNA replication stress, leading to
cell cycle arrest and proliferation inhibition^[290]90,[291]91.
Mechanistically, replication stress caused by Cdc45 overexpression
activates the ATM/Chk2 (ataxia telangiectasia mutated/checkpoint kinase
2) signaling pathway^[292]90. ATM and ATR (ataxia telangiectasia and
Rad3-related) are essential kinases that regulate cell cycle and
replication checkpoints by phosphorylating checkpoint effector kinases
Chk2 and Chk1 in response to DNA damage^[293]55. These signaling
pathways are vital in initiating DNA repair mechanisms, ensuring that
damaged DNA is repaired before cell cycle progression resumes^[294]55.
Consistent with these reports, we observed significant upregulation of
cell cycle-related genes, including ATM and ATR (Supplementary
Data [295]6; Supplementary Fig. [296]6E, F; Supplementary Fig. [297]8B;
Supplementary Fig. [298]9), in overexpression lines. This upregulation
likely contributes to the observed cell cycle arrest and may facilitate
DNA repair under thermal stress, maintaining genomic stability. These
results indicate that the regulation of cell size and thermal responses
by PtCdc45-like may be mediated indirectly through its influence on DNA
replication and checkpoint activation (Fig. [299]6).
We found that the siPtHSF2 and siPtCdc45-like lines exhibited similar
high-temperature tolerance as the WT, rather than the expected
reduction in thermal tolerance (Figs. [300]1C and [301]3C). This
unexpected observation may be attributed to potential compensatory
mechanisms and residual gene expression. Thermal tolerance in P.
tricornutum is likely governed by a network of genes and pathways with
overlapping or redundant functions. While PtHSF2 and PtCdc45-like
contribute to thermal stress responses, other HSFs or regulatory
proteins may compensate for their partial loss of function. Notably, P.
tricornutum possesses 69 HSF genes, suggesting potential functional
redundancy or overlap. Additionally, RNAi-mediated knockdown reduced
PtHSF2 and PtCdc45-like expression by 72.7–84.5% and 70.3–79.5%,
respectively (Supplementary Fig. [302]2D, Supplementary Fig. [303]7D),
potentially retaining sufficient gene activity to sustain baseline
thermal tolerance and mask phenotypic differences.
Orthologue comparisons across species revealed distinct temperature
response mechanisms. The HSF2 orthologue ([304]XP_005770267.1) was
present in E. huxleyi (a coccolithophore) with 39.7% similarity, but no
Cdc45-like orthologue was detected. In C. reinhardtii (a green alga),
both HSF2 orthologue ([305]XP_001702071.2) and Cdc45-like orthologue
([306]XP_001696501.1) exhibited low similarity (HSF2, 35.3%;
Cdc45-like, 27.1%) (Supplementary Data [307]13). This may explain the
narrower thermal tolerance ranges of coccolithophores and green algae
compared to diatoms, underscoring the centrality of PtHSF2 and
PtCdc45-like in diatom thermal adaptability.
Cell growth and proliferation depend on efficient substance transport
and energy supply^[308]92. ABC transporters exist widely in various
organisms and mediate transmembrane transport through ATP
hydrolysis^[309]93. In our study, PtHSF2- and
PtCdc45-like-overexpressing cells exhibited slower growth rates at
20 °C, consistent with previous findings that larger diatom cells grow
slower^[310]94. One plausible explanation is that larger cells require
more energy and nutrients to sustain growth due to their increased
metabolic demands^[311]95,[312]96. The downregulation of ABC
transporters such as Phatr3_J42881, Phatr3_J44752, and
Phatr3_Jdraft1613 in these lines likely impaired nutrient transport,
further limiting growth rates (Fig. [313]2O, Supplementary Fig. [314]7I
and Supplementary Data [315]7). STRING analysis
([316]https://string-db.org/) indicated Phatr3_Jdraft1613 interacts
with Phatr3_J44752, acetyl-CoA carboxylases (1 and 2), and the KCNJ
(K^+ channel inward rectifier domain-containing protein) (Supplementary
Fig. [317]5C). These findings suggest that these ABC transporters may
play roles in lipid and potassium ion transport, with their reduced
expression potentially impairing material transport and limiting
growth^[318]97. However, their relationship with cell growth and
proliferation requires further investigation.
Notably, while FAD was differentially expressed in RNA-seq analysis
(Supplementary Data [319]6), it was not identified as a target of
PtHSF2 in CUT&Tag experiments. This limited overlap likely arises from
the methodological distinctions between RNA-seq and CUT&Tag. RNA-seq
provides comprehensive coverage of DEGs, offering insights into broad
transcriptional changes^[320]98, but it cannot identify direct
transcription factor-target interactions. In contrast, CUT&Tag
specifically detects TF binding sites^[321]43, offering high
specificity but limited coverage. Consequently, many DEGs identified in
RNA-seq may reflect indirect effects rather than direct PtHSF2 targets.
Additionally, TF binding often involves complex regulatory networks
with co-factors and secondary effects^[322]44,[323]47, and not all
binding sites are functionally active^[324]45,[325]47. This highlights
the complexity of gene regulation and the functional diversity of TF
binding sites.
Accumulating evidence shows that P. tricornutum exhibits extensive
temperature adaptability^[326]17–[327]19. Genomically, it possesses 69
HSF genes that may contribute to its heat response regulation.
Morphologically, P. tricornutum is a polymorphic microalga, with its
plasticity in both shape and size playing crucial roles in
environmental adaptation. Previous studies have reported a tendency for
P. tricornutum cells to enlarge as temperatures
rise^[328]16,[329]71,[330]72. Although HSFs’ function in heat stress
response is well established, their specific regulatory mechanisms and
relationship to size plasticity and environmental adaptability in P.
tricornutum remain underexplored. In this study, we identified PtHSF2
as a critical regulator of high-temperature responses, with
PtCdc45-like as its downstream target. Through this regulatory
interaction, PtHSF2 contributes to cell size modulation and thermal
tolerance. These findings advance our understanding of the mechanisms
underlying diatom thermal adaptation and shed new light on the
biological functions of HSFs in shaping environmental resilience.
Methods
Strain and culture conditions
Phaeodactylum tricornutum
(CCMP2561), with a fusiform morphotype, was obtained from the National
Center for Marine Algae and Microbiota (NCMA, USA), and cultured in
sterile f/2 medium (excluding Na[2]SiO[3] ⋅ 9H[2]O). It was cultivated
at 20 ± 1 °C in an artificial climate incubator (Ningbo, China),
provided with a 12:12 h light-dark cycle at an irradiance of 200 μmol
photons m^−2 s^−1. WT cells were cultured at four different
temperatures (15 °C, 20 °C, 25 °C, and 30 °C) for 7 days. Subsequently,
cells grown at 20 °C and 30 °C were subjected to transcriptome
sequencing, while cells from all four temperatures were used to analyze
the transcription levels of PtHSF2 and PtCdc45. Transgenic strains were
selected on solid medium with bleomycin. Surviving colonies were
subsequently picked and transferred to fresh f/2 liquid medium
containing bleomycin (85 μg/mL). Before experimentation, transgenic
strains underwent several growth cycles in bleomycin-free medium to
negate potential bleomycin antibiotic effects. All experiments were
performed according to strict aseptic practices.
Thalassiosira weissflogii
(CCMA-189) cultures were maintained under standard growth conditions in
f/2 medium at 20 °C with a 12:12 h light-dark cycle. This strain was
obtained from the Center for Collections of Marine Algae (CCMA), and
its cultivation temperature range was between 18 and 25 °C. For
high-temperature treatment, the cultures were divided into three
groups: control (20 °C), moderate heat stress (30 °C), and severe heat
stress (35 °C). Each group was incubated at the respective temperature
for 5 days, after which cells were collected for transcriptome
analysis.
Sequence alignment, phylogenetic analyses, biogeographic and taxonomic
distribution
Using the Plant Transcription Factor Database (PlantTFDB 4.0,
[331]http://planttfdb.cbi.pku.edu.cn/)^[332]99, we identified 69 HSFs
from the P. tricornutum protein library downloaded from the National
Center for Biotechnology Information database (NCBI,
[333]https://www.ncbi.nlm.nih.gov/). Protein libraries for C.
reinhardtii and E. huxleyi were downloaded from NCBI, and the A.
thaliana protein library was downloaded from UniProt
([334]https://www.uniprot.org/). We predicted HSFs in these libraries
using PlantTFDB 4.0 and extracted their gene IDs and amino acid
sequences. The 99 obtained HSF protein sequences (Source Data) were
aligned using MEGA 6.0 with ClustalW, and a phylogenetic tree was
constructed. The results of the alignments were presented in Source
Data file. The Neighbor-Joining method inferred evolutionary
relationships, with bootstrap values calculated from 1000 replicates to
ensure robustness and reliability of the tree topology. Evolutionary
distances were computed using the p-distance method.
To construct phylogenetic trees for PtHSF2 and its homologs, as well as
for PtCdc45-like and its homologs, we conducted BLASTP searches in the
NCBI database to identify the most similar homologs from various
species. The sequences were aligned and phylogenetic trees were
constructed using MEGA 6.0, following the described procedure. The
sequences and alignment results were included in Source Data file.
We utilized the Ocean Gene Atlas
([335]http://tara-oceans.mio.osupytheas.fr/ocean-gene-atlas/) to
explore the biogeographic and taxonomic distribution of PtHSF2 and its
homologs. The Ocean Gene Atlas is a web service designed for
investigating the biogeography of marine planktonic genes^[336]100. It
includes the marine Eukaryotes Single-Cell and Metagenome Assembled
Genomes (EUK-SMAGs) database, which comprises 713 non-redundant and
manually curated eukaryotic MAGs and SAGs, covering a total of 10
million genes. The dataset was constructed from 280 billion metagenomic
reads obtained from the Tara Oceans project, covering diverse marine
environments such as polar, temperate, and tropical sunlit oceans. We
searched the protein sequence of HSF (Phatr3_ J55070) in the EUK-SMAGs
database using BLASTP, and the homologs were obtained at a threshold of
1E^−20. The visualization results were directly downloaded from the
website.
Gene amplification, vector construction, and transformation
The full-length coding regions (excluding stop codons) of PtHSF2
(GenBank: [337]XM_002184371.1/Phatr3_J55070) and PtCdc45-like (GenBank:
[338]OK484425/Phatr3_J9344) were PCR-amplified from P. tricornutum cDNA
using primer pairs HSF2-OF/HSF2-OR and Cdc45-like-OF/Cdc45-like-OR,
respectively, as detailed in Supplementary Table [339]2. These
amplified products were individually cloned into the pHY21^[340]101
vector downstream of the PfcpC (promoter of fucoxanthin chlorophyll a/c
binding protein), pHY22 (driven by the HSF2 promoter), and pHY23
(driven by the Cdc45-like promoter) using the ClonExpress II One Step
Cloning Kit (Vazyme, China). The 2xFlag tag was appended to the
C-terminus of these genes to facilitate protein expression detection.
The resultant recombinant overexpression vectors, namely pHY21-PtHSF2,
pHY22-PtHSF2-proH, pHY21-PtCdc45-like, pHY23- PtCdc45-like-proC were
introduced into microalgae via electroporation using the Gene Pulser
Xcell system (Bio-Rad) as described previously^[341]102. Briefly, cells
were collected, washed thrice with 0.375 M sorbitol, and resuspended in
0.2 mL of 0.375 M sorbitol. Linearized plasmid (3 μg) and
heat-denatured salmon sperm DNA (30 μg) were added to the samples. The
samples were then transferred to cuvettes and incubated on ice for
10 min. The electroporation was performed with the following
parameters: 500 V field strength, 25 μF capacitance, and 400 Ω
resistance. Subsequently, the samples were transferred to 5 mL of
liquid medium and incubated in the dark for 24 h. Finally, the cells
were harvested and spread onto solid medium supplemented with
bleomycin.
PtHSF2 and PtCdc45-like were silenced by using RNA interference (RNAi),
and the RNAi vectors were constructed as described previously with
slight modifications^[342]103. Specifically, a fragment of the PtHSF2
(spanning 348 to 689 bp) was amplified from P. tricornutum DNA using
primers siHSF2-F1 and siHSF2-R1, as listed in Supplementary
Table [343]2. The inverted sequence (from 689 to 348 bp) was amplified
using primers siHSF2-F2 and siHSF2-R2. Given the absence of intron in
the PtHSF2, the first intron of the PtHSF (GenBank:
[344]XM_002177210.1) was amplified using primers siHSF-F3 and siHSF-R3
to facilitate hairpin loop formation. These fragments were subsequently
cloned into the pHY21 vector using the ClonExpress Ultra One Step
Cloning Kit (Vazyme, China). Similarly, fragments of the PtCdc45-like
(from 1 to 465 bp and its inverted sequence from 368 to 1 bp) were
amplified using primer pairs siCdc45-like-F1/R1 and siCdc45-like-F2/R2,
respectively, as listed in Supplementary Table [345]2. These fragments
were then cloned into the pHY21 vector in both sense and antisense
orientations. The resulting RNAi vectors (pHY21-siPtHSF2 and
pHY21-siPtCdc45-like) were introduced into microalgae cells via
electroporation^[346]102.
Following previous studies^[347]104, we introduced two plasmids
simultaneously into P. tricornutum cells to generate cell lines with
the desired genetic modifications. Specifically, plasmid pHY21-PtHSF2
and plasmid pHY21-siPtCdc45-like were co-electroporated into WT cells
to create the PtHSF2 overexpression and PtCdc45-like knockdown cell
line (Ho-Ci). Conversely, plasmid pHY21-siPtHSF2 and plasmid
pHY21-PtCdc45-like were co-transformed to generate the PtHSF2 knockdown
and PtCdc45-like overexpression cell line (Hi-Co). The electroporation
conditions, as well as the selection and cultivation of the transformed
microalgae, were conducted as previously described.
Evaluation of transformants by molecular approaches
Screening on solid media containing bleomycin yielded hundreds of
transformants for each plasmid. Ten transformants for each plasmid were
then randomly selected for further screening and validation.
Transformants that failed to grow in the antibiotic-containing liquid
media were excluded. False positives were identified and eliminated
through genomic PCR, and qPCR was subsequently employed to select
transformants with stable high or low expression of the target gene.
These selection processes ensured that the observed effects were
attributable to the introduced genes rather than random variation among
the transgenic lines. Specifically, to verify the integration of the
transgene into the host cells, genomic PCR was conducted using primers
Pt89a and Pt91r (Supplementary Table [348]2) that were designed to
amplify the backbone of the expression cassette. The PCR products were
separated by gel electrophoresis and purified using a Gel Extraction
Kit (Omega Bio-tek, USA). Subsequently, the purified products were
validated by sequencing analysis. Genomic PCR was performed to confirm
the integration of the silencing expression cassettes into the genome
using the forward primer Pt89a, which targets the vector backbone, and
the reverse primers siHSF2R4/siCdc45-like-R3, which target introns. The
relative transcript levels of target genes were assessed by
quantitative real-time PCR (qPCR). Total RNA was isolated from WT and
transgenic lines using a RNAiso Plus Kit (Takara, Japan). The isolated
RNA was reverse-transcribed using a HiScript II 1^st Strand cDNA
Synthesis Kit ( + gDNA wiper) (Vazyme, China) to generate first-strand
complementary DNA. The qPCR assay was conducted in 96-well plates with
a 20 μL reaction volume using the AceQ qPCR SYBR Green Master Mix Kit
(Vazyme, China) on the Bio-Rad CFX96 real-time PCR system. Each sample
was tested in triplicate, and expression levels were referenced against
the endogenous gene β-actin and TBP (TATA box binding protein), which
were selected based on their stable expression across different
conditions in P. tricornutum, as verified in previous
studies^[349]105,[350]106. To ensure consistent comparisons among
different strains, cells used for molecular validation were cultured
under standard conditions (20 °C) until they reached the exponential
growth phase. All primers and their sequences utilized in this study
are listed in Supplementary Table [351]2. According to the MIQE
guidelines^[352]107, the specificity of the primers was confirmed by a
single peak in the melt curve of the amplification products (Source
Data). The amplification efficiency of the primers, determined from
standard curves, ranged between 90 and 110% (Source Data). Ct values
for each qPCR reaction were provided in Source Data file. Relative
transcription levels were calculated using the comparative Ct
method^[353]108 and were normalized to the reference genes β-actin and
TBP. Total protein was extracted and denatured. Protein expression in
transgenic cells was probed by western blot using an anti-Flag antibody
(1:3000, Rabbit monoclonal, ab205606, Abcam, UK) targeting the
Flag-tagged PtHSF2 and PtCdc45-like.
Analysis of off-target effects
Previous work on dsRNA specificity^[354]37 has shown that dsRNAs
triggering RNAi require one of the following criteria: (i) >80%
sequence similarity with target genes; (ii) ≥16 bp contiguous fragments
that perfectly match the target gene or >26 bp fragments that almost
perfectly match the target gene with one or two mismatches (single
mismatches inserted between matching fragments of ≥5 bp or mismatched
couplets inserted between matching fragments of ≥8 bp). Therefore,
these parameters can be used to predict the off-target effects of
dsRNAs. The dsRNA of PtHSF2 (designated as dsPtHSF2) was aligned with
the coding regions of the remaining 68 HSF genes from P. tricornutum
(Source Data) using MEGA 6.0 with ClustalW. Based on the alignment
results (Source Data), 8 genes with perfectly matching segments to
dsPtHSF2 were selected for further validation of their relative
transcription levels using qPCR. Since no homologous gene of
PtCdc45-like was found in P. tricornutum and no segments with
continuous perfect matches of ≥5 bp in length were found with the dsRNA
targeting PtCdc45-like (designated as dsPtCdc45-like), further analysis
of the off-target effects of PtCdc45-like RNAi was not conducted.
Cell density, growth rate, inhibition rate, and survival rate analysis
We synchronized exponentially growing cells by subjecting them to an
extended dark period (20 h) and then transferred them to fresh medium
for normal culture (12 h light: 12 h dark). Referring to previous
studies^[355]10,[356]19, each strain was acclimated to the experimental
temperature. After a 7-day acclimation period, a second transfer was
performed, and measurements of various parameters commenced. To ensure
sufficient cell quantities for subsequent measurements and to avoid
growth differences due to initial density variations, the initial cell
density was set to 1 × 10^6 cells/mL. The cultures were then maintained
at 15 °C, 20 °C, 25 °C, and 30 °C, respectively. Cell density was
monitored using a light microscope and a Bright-Line hemocytometer, and
the data were plotted as growth curves. The growth inhibition rate at
15 °C, 25 °C, and 30 °C was calculated by comparing cell density with
that at 20 °C on day 13. Cell viability was assayed using 1.0% Trypan
blue^[357]109. Briefly, cell culture (1 mL) was collected and
centrifuged to separate the supernatant and cells. The supernatant was
discarded and the cells were resuspended with 100 μL of 1.0% Trypan
blue. The samples were incubated at room temperature for 3 h,
centrifuged to remove the supernatant, and washed twice with PBS
(Phosphate buffered saline) to remove the unbound dye. Finally, the
cells were resuspended with PBS and observed and counted under a
microscope.
Laser scanning confocal microscopic observation and morphology analysis
Cell morphology and oil bodies were visualized using a confocal laser
scanning microscope (LSM 880 with AiryScan, Carl Zeiss,
Germany)^[358]110. Specifically, cells from day 12, collected at
consistent culture times, were stained with Nile red (0.1 mg/mL in
acetone; Sigma-Aldrich, USA) and incubated in the dark for 10 min at
room temperature. Observations were made using a 100×/1.4-NA oil
objective lens, with excitation and emission wavelengths set at 530 nm
and 592 nm, respectively. Images of each sample were captured randomly.
The cell size parameters were quantified by Fiji ImageJ tool^[359]111.
Briefly, images with embedded scale bar were imported into the ImageJ
program. The length of the scale bar was measured by line tool. The
unit of length (in pixels) was set to micron, based on the actual
length of the scale bar. Subsequently, cell length and width were
measured using line tool. The geometry of P. tricornutum is half
parallelepiped and its volume was calculated according to the following
equation^[360]112: V = l * w * h / 2, where V is volume, l is cell
length, w is cell width, and h is height (equal to the width).
Neutral lipid content analysis
To detect the neutral lipid content, cells were stained with Nile red
(0.1 mg/mL in acetone) and the relative fluorescence intensity was
analyzed using a Multifunctional Microplate Reader (BioTek Synergy H1,
USA)^[361]110. Briefly, 10 μL Nile red was added to 1 mL of cell
culture, and the samples were mixed well and incubated in the dark for
20 min at room temperature. The samples were transferred to a 96-well
plate and the fluorescence intensity was detected at excitation and
emission wavelengths of 530 and 580 nm, respectively. The unstained
cell cultures were used as blank controls. Finally, the neutral lipid
content was analyzed by the relative fluorescence intensity of Nile
red.
Fatty acid composition analysis
To analyze the composition of the fatty acids, they were extracted and
methylated, and detected using a gas chromatography-mass spectrometry
(GC-MS)^[362]113 with slight modifications. 50 mL of cell culture was
harvested and the supernatant was discarded. Samples were transferred
to 15 mL tubes and 5 mL 2 M KOH (in methanol) was added. Cells were
broken by sonication and incubated at 75 °C for 10 min. Samples were
stratified at room temperature and the supernatant was transferred to
50 mL tubes. 5 mL 2 M KOH was added to samples and the above steps were
repeated twice. The supernatants were combined and an equal volume of
2 M HCl (in methanol) was added. Samples were incubated at 75 °C for
10 min, stratified at room temperature, and the supernatant was
transferred to new 50 ml tubes. Subsequently, hexane was added for
extraction of methylated fatty acids. 6 μL 10 mg/mL methyl
nonadecanoate (in hexane) was added to samples, and samples were
concentrated to 1.5 mL using a Termovap Sample Concentrator. The
methylated fatty acids were analyzed by GC-MS and identified leveraging
the National Bureau of Standards spectrum library. Lipid saturation
index was calculated using the following equation^[363]114:
[MATH: Saturationindex=Totalfattyacid−Una<
mi>turatedfattyacid
Amountoftotalfattyacid×100
:MATH]
.
Analysis of photosynthetic efficiency
The maximum photochemical efficiency of PSII (Fv/Fm) was measured as
described previously with slight modifications^[364]113. Briefly, 2 mL
cell culture was harvested and incubated in the dark for 10 min.
Samples were then transferred into a measuring cuvette, and chlorophyll
fluorescence parameters were determined with a PhytoPAM Phytoplankton
Analyzer (Walz, Germany). The Fv/Fm was calculated based on the
following formula:
[MATH: Fv/Fm=Fm−FoFm :MATH]
, where Fv is variable fluorescence, Fm is maximum fluorescence, and
F[o] is initial fluorescence.
Transcriptome sequencing and analyses
Research has indicated that while many transcription factors are
activated only under specific conditions, such as light stress and high
temperature^[365]115, and analyzing transcription profiles and DNA
binding sites in organisms exhibiting visible phenotypes due to
transcription factor overexpression is a common approach to studying
transcription factor mechanisms^[366]116. This method can be applied to
organisms cultured under normal culture conditions without requiring
specific condition induction^[367]116. In this study, we observed that
overexpression of PtHSF2 under normal culture conditions (20 °C) led to
phenotypic changes such as increased cell volume and slowed growth.
Therefore, we conducted RNA-seq and CUT&Tag experiments on cells
cultured at 20 °C. After incubation at 20 °C to stationary phase, total
RNA was extracted from PtHSF2 overexpressing cells and WT cells using
RNAiso reagent (Takara, Japan), with three parallel experiments each
cell line. Then, mRNA was enriched with Oligo (dT) beads and fragmented
by fragmentation buffer. The fragmented mRNAs were reverse transcribed
into cDNA using a random hexamer primer, and subsequently, the
second-strand cDNA was synthesized. The double-stranded cDNA was
purified with QiaQuick PCR extraction kit, and the purified cDNA
fragments were subjected to end repair, poly(A) addition and Illumina
sequencing adapters ligation. Next, the ligated samples were separated
by size via agarose gel electrophoresis and then enriched by PCR
amplification. Finally, the resultant products were sequenced using the
NovaSeq 6000 platform (Illumina, San Diego, CA, USA) at the Novogene
company (Peking, China). RNA extraction and transcriptome sequencing
for T. weissflogii were carried out following the same protocol.
Raw sequence data were assessed for quality using SeqKit tool.
Trimmomatic was used to trim low-quality bases and remove adapter
sequences, producing high-quality reads for subsequent analysis. The
cleaned reads were aligned to the corresponding reference genomes of P.
tricornutum and T. weissflogii using Bowtie 2. RPKM (Fragments Per
Kilobase per Million bases) values were calculated and log2-transformed
to assess gene expression levels. DEGs analysis was performed using
DESeq2. Genes with an adjusted |log₂(FC)| ≥ 1 and p-value < 0.05 were
considered significantly differentially expressed. Significant DEGs
were further verified by qPCR, and relative gene expression was
determined by comparative Ct method and normalized by reference genes
β-actin and TBP.
For GO term enrichment analysis, the SwissProt database downloaded from
NCBI was used as the reference for annotation. The diamond tool was
utilized with an e-value cutoff of 1e–5 to map the DEGs to the
SwissProt database. After mapping, UniProt IDs were obtained, and
corresponding GO IDs were retrieved from the idmapping.tb.gz file,
downloaded from UniProt. WEGO (Web Gene Ontology Annotation Plot,
[368]http://wego.genomics.cn) was then used to analyze and obtain Level
2 GO terms, enabling the identification of enriched biological
processes, molecular functions, and cellular components in our
datasets. Additionally, KEGG pathway enrichment analysis was conducted
using the Assign KO Tool
([369]https://www.kegg.jp/kegg/mapper/assign_ko.html).
CUT&Tag combined with electroporation
CUT&Tag was performed as previously described^[370]43 with
modifications, and Hyperactive Universal CUT&Tag Assay Kit for Illumina
(Vazyme, China) was used. In addition, since P. tricornutum has a cell
wall, proteins were delivered into the nucleus using
electroporation^[371]117,[372]118. In brief, PtHSF2-1, PtHSF2-2, and WT
cells were used in this experiment, with three biological replicates
and one technical replicate for each line. A 1:1 mixture of cell
samples from these two overexpression lines was employed as a negative
control. Cells were collected after incubation at 20 °C to stationary
phase. Cell pellets were washed thrice with wash buffer and resuspended
in 200 μL antibody buffer containing a 1:50 dilution of the primary
antibody (Rabbit Anti-Flag, ab205606, Abcam, UK). Non-specific IgG
(1:50, Rabbit IgG, ab172730, Abcam, UK) was added to the negative
control. After gentle vortexing, samples were transferred into cuvettes
and electroporated as described previously^[373]102. The
electroporation parameters were set as follows: 500 V field strength,
25 μF capacitance, and 400 Ω resistance. The samples were transferred
into 1.5 mL tubes and incubated at 4 °C overnight. Centrifugation was
performed to remove the supernatant, and cells were resuspended in
200 μL Dig-wash buffer pre-mixed with secondary antibody (1:50, Goat
Anti-Rabbit IgG H&L, ab6702, Abcam, UK). The mixture was transferred
into cuvettes, and the electroporation was performed as described
previously^[374]102. Samples were then transferred into 1.5 mL tubes
and incubated at room temperature for 1 h. Cells were washed thrice in
Dig-wash buffer and resuspended in 0.04 μM Hyperactive pA -Transposon
adapter complex (in Dig-300 buffer) with gentle vortexing.
Subsequently, samples were transferred into cuvettes for
electroporation^[375]102. Samples were transferred into 1.5 mL tubes,
and the tubes were placed on a rotator at room temperature for 1 h. The
supernatant was removed, and cells were washed thrice with Dig-300
buffer and resuspended in 50 μL Trueprep Tagment Buffer L. Samples were
incubated at 37 °C for 1 h to fragment the DNA. 5 μL 20 mg/mL
Proteinase K, 6 µL 10% SDS and 10 µL 0.5 M EDTA were added to samples.
The samples were thoroughly mixed and incubated for 10 min at 55 °C.
DNA was extracted by phenol-chloroform-isoamyl alcohol method.
A TruePrep Index Kit V2 for Illumina (Vazyme, China) was used for PCR
amplification. DNA fragments were used as templates, and universal i5
primer, uniquely barcoded i7 primers, and CUT&Tag Amplification Mix
were added. PCR was performed with the conditions below: 72 °C for
3 min, 95 °C for 3 min, 12 cycles of 98 °C for 10 s and 60 °C for 5 s,
72 °C for 1 min, and 4 °C hold. Amplification products were purified
using VAHTS DNA Clean Beads (Vazyme, China). For library pooling, 10 μL
of sample from each of the three libraries of the WT line were
homogenously mixed to form the sequencing library for WT. Similarly,
5 μL of sample from each of the six libraries of the two overexpression
lines were homogenously mixed to form the sequencing library for
PtHSF2. Additionally, 10 μL of sample from each of the three libraries
of the negative control were homogenously mixed to form the sequencing
library. These three libraries were sequenced on an Illumina NovaSeq
platform (BGI, Shenzhen, China).
Data processing and analysis
Paired-end reads were aligned using Bowtie2 version 2.2.5
([376]https://bowtie-bio.sourceforge.net/bowtie2/index.shtml). Peak
calling was executed with Model-based Analysis of ChIP-Seq 2
(MACS2)^[377]119. Notably, we constructed a promoter database for P.
tricornutum using the hiPromoter software (Patent No. ZL202110822106.X)
and used BLAST to pinpoint the reads in peaks to promoters. GO and KEGG
pathway enrichment analysis of potential target genes were conducted as
previously mentioned. Binding motifs of PtHSF2 were analyzed using
Multiple EM for Motif Elicitation (MEME)^[378]48.
CUT&Tag-qPCR for profiling the binding of PtHSF2
To further validate the results of CUT&Tag high-throughput sequencing,
CUT&Tag-qPCR analysis was performed on three potential target genes
(Cdc45-like, Lhcx2, and ABC) using unmerged libraries.
Promoter-specific primer pairs were designed according to the
principles of ChIP-qPCR assays. The primer pairs are listed in
Supplementary Table [379]2, with specificity confirmed by a single
peaks in melt curve analysis and amplification efficiencies of 90–110%
(Source Data). The Ct values obtained from qPCR using these primers are
shown in the Source Data file. The promoters of β-actin and TBP, which
are located far from the promoters of Cdc45-like, Lhcx2, and ABC, were
used as internal controls.
Comprehensive analysis of the expression of 69 P. tricornutum HSFs across
MMETSP datasets
To investigate the expression of 69 HSFs across various species, we
utilized RNA-seq datasets from the Marine Microbial Eukaryote
Transcriptome Sequencing Project (MMETSP)^[380]34. We downloaded the
reassembled data, including transcriptome assemblies and expression
quantification datasets ([381]https://zenodo.org/records/3247846),
provided by Johnson et al.^[382]35. To assess the expression levels of
HSFs in P. tricornutum, we performed a BLASTn search using 69 HSF
sequences with an e-value threshold of 1e–5. Additionally, HSF
orthologues in other species were identified using tBLASTn with the
same threshold. Expression levels were determined by integrating the
BLAST results with the expression quantification datasets from the
MMETSP reassembled transcriptome.
Redox levels
The content of reactive oxygen species (ROS) was determined by a ROS
Assay Kit (Beyotime, China). Briefly, 1 mL cell culture were harvested
and washed twice with PBS. Cells were resuspended in 1 mL 10 μM
fluorescent probe DCFH-DA and incubated at 37 °C for 20 min. Cells were
washed thrice in PBS to remove probes did not enter cells. 1 mL PBS was
added to resuspend cells, and fluorescence intensity was measured using
a microplate reader.
Lipid peroxidation was assessed using a Lipid Peroxidation MDA Assay
Kit (Beyotime, China). Firstly, total protein was prepared using a Cell
Lysis Buffer (Beyotime, China) and protein concentration was determined
by a BCA Protein Assay Kit (Beyotime, China). 0.2 mL protein sample was
mixed with 0.4 mL MDA detection solution and heated at 100 °C for
15 min. Subsequently, 0.2 mL supernatant was transferred into a 96-well
plate, and the absorbance was measured at 532 nm.
Catalase (CAT) activity was assayed using ammonium
molybdate-chromogenic method^[383]120 with a Catalase Assay Kit
(Solarbio, China). In short, 30 mL cell culture were harvested and
washed twice with PBS. Cells were resuspended in 1 mL extraction
solvent and sonicated. Centrifugation was performed to separate the
cells and supernatant. 20 μL supernatant was mixed with 100 μL H[2]O[2]
substrate and incubated at 25 °C for 10 min. 180 μL CAT assay buffer
was added to samples, and the mixture was left at room temperature for
10 min. Samples were transferred into 96-well plates, and the
absorbance was obtained at 405 nm.
Superoxide dismutase (SOD) activity was determined using a Total
Superoxide Dismutase Assay Kit with WST-8 (Beyotime, China). 50 mL cell
culture were collected and washed twice with PBS. 200 μL SOD sample
preparation solution was added, and samples were homogenized using a
tissue grinder. Subsequently, centrifugation was performed and the
supernatant was collected. 20 μL supernatant, 160 μL WST-8 working
solution, and 20 μL enzyme working solution were added to 96-well
plates. The samples were thoroughly mixed and incubated at 37 °C for
30 min. Absorbance was measured at 450 nm.
Flow cytometry
Flow cytometry analysis was performed to examine the DNA content and
determine the cell cycle distribution. Five microalgal strains (WT,
PtHSF2-1, siPtHSF2-1, PtCdc45-like-1, siPtCdc45-like-1) were used in
this experiment. Following a previously established method^[384]121
with minor modifications, we monitored the DNA content changes in
exponentially growing cells over 12 consecutive time points. Starting
at the beginning of the light period (0 h), we sampled 5 mL of culture
every 2 h. Cells were collected, washed thrice with 1 × PBS, and then
fixed overnight at 4 °C in 1 mL of 70% ethanol. Post-fixation, we added
500 µL of SYBR Green I working solution, diluted in PBS from a 20 ×
stock solution in DMSO, and incubated the samples at room temperature
for 15 min in the dark. Fluorescence was detected using a flow
cytometer with an excitation wavelength of 488 nm, an emission
wavelength of 520 nm, and a detection filter range of 510–530 nm.
Statistical analysis
Statistical analyses and visualization were performed using GraphPad
Prism (version 8.0). Data from at least three independent experimental
replicates were given and presented as means ± SD. Two-tailed unpaired
Student’s t test was used to assess the differences between two
experimental groups. Significant difference was indicated at P < 0.05
(*) or P < 0.01 (**) level.
Reporting summary
Further information on research design is available in the [385]Nature
Portfolio Reporting Summary linked to this article.
Supplementary information
[386]Supplementary information^ (2.6MB, pdf)
[387]Transparent Peer Review file^ (884.3KB, pdf)
[388]Reporting Summary^ (127.7KB, pdf)
[389]41467_2025_58547_MOESM4_ESM.pdf^ (113.1KB, pdf)
Description of Additional Supplementary Files
[390]Supplementary Data 1-13^ (7.9MB, zip)
Source data
[391]Source Data^ (22.4MB, zip)
Acknowledgements