Sei sulla pagina 1di 14

Research

Noncoding and coding transcriptome responses of a marine


diatom to phosphate fluctuations
Maria Helena Cruz de Carvalho1,2, Hai-Xi Sun1, Chris Bowler2 and Nam-Hai Chua1
1

Laboratory of Plant Molecular Biology, Rockefeller University, New York, NY 10065, USA; 2Institut de Biologie de lEcole Normale Superieure (IBENS), Centre National de la Recherche

Scientifique (CNRS) UMR 8197 INSERM U1024, 46 Rue dUlm, 75005 Paris, France

Summary
Author for correspondence:
Nam-Hai Chua
Tel: +1 212 327 8126
Email: chua@rockefeller.edu
Received: 29 April 2015
Accepted: 27 October 2015

New Phytologist (2015)


doi: 10.1111/nph.13787

Key words: diatoms, long noncoding


intergenic RNAs, Phaeodactylum
tricornutum, phosphate, signaling, strandspecific RNA-sequencing (ssRNA-Seq), transcription factors (TFs).

 Phosphorus (P) is an essential element to all living cells, yet fluctuations in P concentrations
are recurrent in the marine environment. Diatoms are amongst the most successful phytoplankton groups, adapting to and surviving periods of suboptimal conditions and resuming
growth as soon as nutrient concentrations permit. A knowledge of the molecular underpinnings of diatom ecological success is, however, still very incomplete.
 By strand-specific RNA sequencing, we analyzed the global transcriptome changes of the
diatom Phaeodactylum tricornutum in response to P fluctuations over a course of 8 d, defining five distinct physiological states.
 This study reports previously unidentified genes highly responsive to P stress in
P. tricornutum. Our data also uncover the complexity of the P. tricornutum P-responsive sensory and signaling system that combines bacterial two-component systems with more complex pathways reminiscent of metazoans. Finally, we identify a multitude of novel long
intergenic nonprotein coding RNAs (lincRNAs) specifically responsive to P depletion, suggesting putative regulatory roles in the regulation of P homeostasis.
 Our work provides additional molecular insights into the resilience of diatoms and their ecological success, and opens up novel routes to address and explore the function and regulatory
roles of P. tricornutum lincRNAs in the context of nutrient stress.

Introduction
As a fundamental element to all living cells, phosphorus (P) is
a building block of nucleic acids and membrane phospholipids.
Through its water-soluble and inorganic active forms (phosphate esters and orthophosphate, respectively), P is also
involved in cellular energy transfers, metabolic pathways and
protein activation. Life in the oceans can be subject to extreme
environments, depleted of the essential elements that sustain
phytoplankton growth, especially in the open ocean far from
land (Morel, 2003). Moreover, in coastal areas, P is often considered to be the potentially limiting element (Bauerfield et al.,
1990; Woodward & Owens, 1990). It is currently accepted
that P, like nitrogen (N), iron or light, may drive marine
microbial evolution and niche adaptation (Dyhrman et al.,
2007).
Diatoms are amongst the most ecologically successful phytoplankton groups living in the oceans. Their ecological relevance
is paramount, sustaining both marine and terrestrial ecosystems
by their role in the global carbon cycle, producing each year
the same amount of organic carbon as all of the terrestrial
rainforests combined (Nelson et al., 1995; Field et al., 1998).
The marine environment is subjected to recurrent fluctuations
2015 The Authors
New Phytologist 2015 New Phytologist Trust

in nutrient levels through the movement of tides and currents.


Diatoms have been shown to be poor competitors when phosphate is deficient (Egge, 1998). Yet, these organisms are able
to survive under such deprived conditions and, when nutrients
are present again in optimal concentrations, such as well-mixed
coastal and upwelling regions, they will often re-dominate
(Morel, 2003). Therefore, the ecological success of diatoms
also depends on their resilience and their capacity to respond
and maintain themselves in a state of survival under unfavorable conditions, which enables them to resume active cell division and growth as soon as favorable environmental conditions
resume. This capacity is ultimately underpinned by a sensory
and regulatory system that detects environmental changes and
directly (or indirectly) regulates genetic pathways, allowing
diatoms to optimize their metabolism to the changing environment.
The availability of diatom genome sequences, combined with
high-throughput sequencing techniques, provides a unique
opportunity to perform comparative genome-wide transcriptome
analysis of the molecular changes underpinning the diatoms
resilience and ecological success under a fluctuating environment.
Our experiments were designed to mimic inorganic phosphate
(Pi) fluctuations that occur in the marine environment in order
New Phytologist (2015) 1
www.newphytologist.com

New
Phytologist

2 Research

to identify genes responsive to these changes. Physiological and


global transcriptomic changes were assessed in the model pennate
diatom Phaeodactylum tricornutum during the early and late
stages of Pi depletion over the course of 8 d, as well as on Pi
resupply. These changes were compared with cultures growing in
replete medium.

Materials and Methods


Diatom culture conditions
Axenic cultures of Phaeodactylum tricornutum (Bohlin) strain
CCMP632 were obtained from the Center for the Culture of
Marine Phytoplankton (East Boothbay, ME, USA) and maintained under continuous shaking (100 rpm) in 250 ml of filtered
(0.22 lM) steam-sterilized artificial sea water (Sigma) supplemented with f/2 nutrients, elements and vitamins (Guillard &
Ryther, 1962), with the exception of silica (f/2-Si), in 1-l glass
flasks. Cultures were kept at 20C under cool white fluorescent
lights at 100 lmol m2 s1 with a 12-h photoperiod. For the Pi
fluctuation studies, equal aliquots of 4-d-old cultures from the
same batch culture were inoculated in parallel in 250 ml of fresh
f/2-Si medium (control conditions) and in 250 ml of fresh f/2-Si
medium without phosphate supplement (Pi depleted), and cultured in the same conditions as described earlier. Culture growth
was followed using a hematocytometer (Fisher Scientific,
Pittsburgh, PA, USA) and the growth rate was calculated using
the natural logarithm of the difference in cell density during the
first 4 d of growth. P-replete and P-depleted cultures were all
started with the same initial cell densities of c.
4.5 9 105 cells ml1. Pi resupplementation was performed on
4-d-old Pi-depleted cultures. Briefly, c. 4050 ml of 4-d-old Pidepleted cultures were pelleted by centrifugation for 10 min at
1500 g and resuspended in 250 ml of fresh f/2-Si medium (starting cell densities of c. 2 9 105 cells ml1). Cells (250 ml) were
harvested always at midday (6 h of light period) by vacuum filtration (0.22 lM) at different treatment time points, flash frozen
in liquid N2 and maintained at 80C until use. The filtered
medium was used to measure the Pi content employing a
SensoLyte MG Phosphate Assay Kit (AnaSpec, Fremont, CA,
USA) following the manufacturers instructions. Experiments
were performed using duplicates of Pi-depleted cultures that
were harvested in bulk. All the experiments were repeated at
least twice.

(Zeiss) with excitationemission wavelengths of 594 nm for


chlorophyll and 488 nm for Nile Red.
RNA extraction, library preparation and ssRNA-Seq
Total RNA was extracted from P. tricornutum flash-frozen cell
pellets using the Trizol method according to the manufacturers
instructions (Invitrogen). Extracted RNA was treated with
TurboDNAse I (Life Technologies AM2238, Carlsbad, CA,
USA) according to product instructions and the treated RNA was
purified using an RNeasy spin column (Qiagen) with 0.5 volumes
of ethanol, washed and then eluted with 50 ll of roomtemperature molecular-grade water (Qiagen). The quality of the
purified RNA was assessed using an Agilent 2100 Bioanalyzer
(Agilent, Santa Clara, CA, USA). cDNA libraries for ssRNA-Seq
were prepared using the Illumina TruSeq Stranded mRNA Sample Preparation Kit following the Low Sample (LS) Protocol
guidelines (Illumina, San Diego, CA, USA) from two independent experiments.
ssRNA-Seq data assembly and analysis
ssRNA-Seq reads were mapped to the P. tricornutum genome
(Phatr2) using TopHat (version 2.0.8) (Kim et al., 2013). The
mapped reads were assembled using CUFFLINKS (v.2.1.1) (Trapnell et al., 2013) with Phatr2 annotation as the reference
(Nordberg et al., 2014). The assembled transcripts of each
ssRNA-Seq sample were merged and annotated using CUFFCOMPARE (v.2.1.1) (Trapnell et al., 2013). The expression levels of
each gene were then calculated from the fragments per kilobase
of exons per million fragments mapped (FPKM) using CUFFDIFF
(v.2.1.1) (Trapnell et al., 2013). The Pearson correlation coefficients (PCCs) were calculated between the FPKM of two replicates of each five physiological states corresponding to two
independent experiments. As the PCCs were high (> 0.9), we
pooled the reads derived from the two replicates in order to
increase the depth of the reads for the increased detection of noncoding RNA transcripts. A two-fold variance in FPKM and a
P value (Fishers exact test) of < 0.05 were used as cutoffs to
define differentially expressed genes. The reads from both replicates have been deposited in the Gene Expression Omnibus
(GEO) database http://www.ncbi.nlm.nih.gov/geo (accession no.
GSE66997).
Quantitative real-time reverse transcription PCR

Confocal laser microscopy


Cells were sampled for confocal laser microscopy at the same
treatment time points as used for strand-specific RNA sequencing
(ssRNA-Seq). For the staining of lipids, 2 ll of a freshly prepared
Nile Red solution (250 lg ml1 in acetone) were added to a 1-ml
diatom cell suspension. After 30 min staining in the dark at room
temperature, cells were washed by pelleting for 10 min at 1500 g
and then resuspended in 500 ll of sterile seawater (Sigma). Confocal microscopy analysis of the dyed diatom cell suspension was
performed using an LSM 780 laser scanning confocal microscope
New Phytologist (2015)
www.newphytologist.com

Total RNA was isolated and purified from P. tricornutum cells as


described above. For cDNA synthesis, 500 ng of total RNA was
incubated with SuperScript III Reverse Transcriptase (Invitrogen)
according to the manufacturers instructions. For quantitative
reverse transcription polymerase chain reaction (RT-qPCR) analysis, cDNA was amplified using SYBR Premix ExTaq (Takara,
Madison, WI, USA) with specific primers (Supporting Information Table S2) picked from random genes from those most highly
expressed under Pi depletion. Primers were designed with the
PRIMER-BLAST program (http://www.ncbi.nlm.nih.gov/tools/
2015 The Authors
New Phytologist 2015 New Phytologist Trust

New
Phytologist
primer-blast/) defining a PCR amplicon size of < 180 bp and
Phaeodactylum tricornutum CCAP 1055/1 (taxid: 556484) as the
reference organism to check for primer pair specificity. Quantitative PCR conditions were set as follows: 95C for 10 s, followed
by 40 cycles of 95C for 5 s and 60C for 30 s, and a final cycle
of 95C for 10 s and 60C for 5 s. Data were collected and analyzed by a Bio-Rad CFX96 real-time system (BioRad, Hercules,
CA, USA). CDKA and HISTONE 4 mRNA levels were used for
normalization (Siaut et al., 2007).
Correlation analysis between transcription factors (TFs) and
putative target genes
To predict the gene targets of selected transiently expressed heat
shock factors (HSFs), the PCCs were calculated using FPKMs vs
all Pi-responsive protein coding genes as well as noncoding genes.
Only the positively correlated gene targets (PCC > 0 and
r2 0.6) were retained. As HSFs bind to heat shock elements
within promoter regions of their target genes (Nover et al.,
2001), only targets with the conserved sequence 50 GAAnnTTC30
in their promoter sequences (broadly defined as sequences 3 kb
upstream of the transcriptional start site) were considered. The
results of the correlation analysis were visualized using Cytoscape
(v2.8.3) (Smoot et al., 2011).
Identification of long intergenic nonprotein coding RNAs
(lincRNAs) and characterization of their genomic features
All assembled intergenic transcription units were collected as
lincRNA candidates. Candidates with a length of 200
nucleotides and a predicted open reading frame (ORF) of 100
amino acids were defined as lincRNAs (Liu et al., 2012). When
considering gene models of protein coding genes, the best gene
models obtained from the Phatr2 annotation were used; for
those of lincRNAs, the gene models with maximum intron numbers were used. The lengths of entire unspliced transcripts (including introns) were used to compare transcript length
distribution. For ORF predictions, the spliced transcripts were
used and sent to GenScan (Burge & Karlin, 1997). To compare
the length of stop codon-free sequences, the distances between
each stop codon of each gene in three reading frames were calculated, as described previously (Niazi & Valadkhan, 2012). Mfold
was used to calculate the free energy of the spliced transcripts
(Zuker, 2003). As longer transcripts have lower free energy, to
better compare the free energy of lincRNAs and mRNAs from
protein coding genes, we classified them into different groups
based on sequence length. This was also performed when comparing the free energy with the GC contents of lincRNAs and
mRNAs.
To find putative cis-regulation of lincRNAs on their neighboring genes, the PCCs were calculated using FPKMs of all the
responsive lincRNAs vs mRNAs encoded by their neighboring
upstream and downstream genes. Only those Pi-responsive
neighboring genes with r2 0.6 (positive or negative correlation)
were considered as putative gene targets of cis-regulatory
lincRNAs.
2015 The Authors
New Phytologist 2015 New Phytologist Trust

Research 3

Results and Discussion


Culture growth and transcriptomic responses under Pi
fluctuations
When P. tricornutum cultures were exposed to Pi depletion in the
medium (Fig. S1), their exponential growth ceased rapidly when
compared with control cultures (Fig. 1a,b). Control cultures continued to grow for 8 d, reaching a cell density of c.
4.8 9 106 cells ml1, whereas cultures without Pi grew for the
first 2 d, peaking at a maximum cell density of c.
1.3 9 106 cells ml1, and then transitioned to stationary growth
(Fig. 1a). This was accompanied by lipid accumulation that was
detected after 4 d of Pi depletion and further increased after 8 d
(Fig. 1c). Neutral lipid accumulation is a common response of
microalgae under unfavorable conditions, namely nutrient stress
(Fields et al., 2014). When Pi was resupplied to 4-d starved cultures, culture growth recovered to control rates after 4 d and lipid
bodies were no longer detected (Figs 1b,c, S2). These observations allowed us to define five distinct physiological states to be
used as sampling points for the comparative transcriptomic studies: control 4 d (early control), Pi 4 d (early Pi depletion),
control 8 d (late control), Pi 8 d (late Pi depletion) and recovery 4 d.
ssRNA-Seq data generated from the five physiological states
were assembled and mapped (Table S1), yielding 10 083 genes
corresponding to 97% of the P. tricornutum annotated protein
coding genome (genome.jgi-psf.org/Phatr2). Globally, 6436 protein coding genes were differentially expressed in response to Pi
fluctuations when compared with the control, with the relative
portion of up- and downregulated genes being approximately the
same (Fig. 2a). During early Pi depletion, the number of upregulated genes was slightly lower than those upregulated at late Pi
depletion and at recovery (1503, 1958 and 1991, respectively)
(Fig. 2a). The number of downregulated genes was the lowest for
recovery (1396) and the highest for late Pi stress (2368), with
early Pi stress showing the downregulation of 2032 genes
(Fig. 2a). The expression of 21 of the most responsive protein
coding genes during Pi depletion was further verified by RTqPCR. These included genes coding for Pi transporters, alkaline
phosphatases (AlkPases), heat shock proteins, HSFs and several
proteins of unknown function (Table S2). A good correlation
between transcript abundance was found between ssRNA-Seq
data and RT-qPCR (r = 0.8, P ? 0) validating the robustness of
ssRNA-Seq data (Figs 2b, S3).
The plethoric origin of the diatom P-responsive genes
A striking aspect uncovered by the sequencing of the diatom
genomes is the diverse origin of their genes. To a large extent, this
is the result of their complex evolutionary origin, making diatoms
not quite plants nor animals (Armbrust et al., 2004; Bowler et al.,
2008). In order to investigate the origins of the Pi differentially
expressed genes, we compared their protein sequences with those
in the National Center for Biotechnology Information nonredundant (NCBI nr) database using BLASTp. The criteria used to
New Phytologist (2015)
www.newphytologist.com

New
Phytologist

4 Research

DIC
(a)

(c)

C 4d

-Pi
+Pi

5
4
3
2

**

**

Pi 4d

Cell density (106 ml1)

0
0

Pi 8d

Time (d)

(b)
2.5

+Pi 4d

2
1.5

*
1
0.5

C 8d

Growth rate (loge)

NR + Chl

0
C 4d

Pi 4d

+Pi 4d

define genes that were present in one species were those published previously (Armbrust et al., 2004; Bowler et al., 2008):
30% sequence positives; 30% alignment coverage of either
the query or subject sequences; and BLAST e-value of < 1e-5. The
larger fraction of Pi differentially expressed genes (30%) corresponded to proteins common to plants and animals (Eukaryotes),
20% were core proteins and 19.4% were exclusively of plant
origin (Fig. 2c). Shared proteins between P. tricornutum and
Thalassiosira pseudonana constituted < 2% of the differentially
expressed genes, and those found exclusively in P. tricornutum
accounted for c. 16% (Fig. 2c). The latter are probable de novo
genes that have evolved since the two diatom lineages diverged c.
90 million yr ago (Bowler et al., 2008). Seventy-seven differentially expressed genes were exclusively found in bacteria and were
absent in all other organisms (Fig. 2c). Many of the bacterial
genes that responded to Pi fluctuation were related to metabolism
and redox processes. Interestingly, included in the Pi differentially expressed genes were 123 genes also present in viruses, some
of which were simultaneously present in plants (four), in plants
and bacteria (19), in animals and bacteria (five) and in all the
groups (62), suggesting a possible horizontal viralhost gene
exchange. These consisted largely of membrane proteins, including several putative signaling protein kinases and Pi transporters.
The presence of Pi transporter genes in several virus genomes
New Phytologist (2015)
www.newphytologist.com

Fig. 1 Phaeodactylum tricornutum responses


to inorganic phosphate (Pi) fluctuations. (a)
Time course of cultures grown in fully
supplemented medium (closed circles) and
Pi-depleted medium (open circles); and (b)
mean growth rates (natural logarithm, loge)
of the first 4 d of culture under fully replete
control conditions (C), Pi depletion (Pi) and
Pi resupply after 4 d of Pi depletion (+Pi).
Data represent  SD of two independent
experiments (with at least three biological
replicates each). Statistically significant
differences between treatment and control
cultures were assessed by a Students t-test:
**, P < 0.01; *, P < 0.05. (c) Confocal laser
microscopy of diatom cells in the five distinct
physiological states: C 4d, 4 d under control
conditions; Pi 4d, 4 d Pi depleted; Pi 8d,
8 d Pi depleted; +Pi 4d, 4 d Pi resupply after
4 d Pi depleted; C 8d, 8 d control conditions.
Cells were stained with Nile Red (NR) to
visualize lipid bodies (green color).
Chloroplasts can be visualized by the
chlorophyll autofluorescence (red color).
Bars, 10 lm. Chl, chlorophyll
autoflorescence; DIC, differential
interference contrast.

suggested the occurrence of viral manipulation of the infected


host capacity for Pi uptake (Monier et al., 2012). Overall, 80%
of the differentially expressed protein coding genes were shared
with plants and 50% shared with animals, amongst which a small
fraction (< 1%, 64 genes) was absent from plants. These included
annotated genes for metabolism, DNA methylation and a putatively secreted AlkPase (49678). Our data strongly support the
view that the physiology of the P-starved diatom response results
from a combination of genes of diverse origins, probably forming
original metabolic pathways, such as the recently dissected urea
cycle (Allen et al., 2011), enabling the resilience and success of
diatoms in a fluctuating environment.
Pi depletion and Pi resupply define distinct physiological
and transcriptomic profiles
In order to assess the function of the differentially expressed genes
under Pi fluctuation, we performed a search on the available gene
ontology (GO) terms of the annotated transcripts and analyzed
their relative over-representation in the five distinct physiological
states. During early Pi depletion, the upregulated protein coding
genes that were most significantly over-represented were putative
phosphate transporters, revealing an active molecular strategy to
increase the efficiency of Pi uptake (Fig. 3a). Other over 2015 The Authors
New Phytologist 2015 New Phytologist Trust

New
Phytologist

Research 5

(a)

(b)

(c)

Fig. 2 Genome-wide transcriptome analysis of Phaeodactylum tricornutum response to inorganic phosphate (Pi) fluctuations. (a) Venn diagrams
representing the up- and downregulated genes in response to progressive Pi depletion and resupply. Cutoff of two-fold difference in fragments per
kilobase of exons per million fragments mapped, P < 0.05, when compared with 4 d control cultures. Pi 4d, 4 d Pi depleted; Pi 8d, 8 d Pi depleted; +Pi
4d, 4 d Pi resupply after 4 d Pi depletion. (b) Correlation between strand-specific RNA-sequencing (ssRNA-Seq) and quantitative reverse transcription
polymerase chain reaction (RT-qPCR) in the detection of Pi-responsive Phaeodactylum tricornutum genes. The x-axis gives the log2 value of the fold
change detected by ssRNA-Seq and the y-axis gives the same value detected by RT-qPCR. Green dots represent differentially expressed (DE) genes
detected by both platforms; gray dots represent genes that do not change by more than two-fold in expression under Pi fluctuation. (c) Gene origins of the
6436 Pi DE protein coding genes. Phaeodactylum tricornutum protein sequences were compared with the National Center for Biotechnology Information
nonredundant (NCBI nr) database using BLASTp. The criteria used to define genes that were present in one species were as follows: 30% sequence
positives; 30% alignment coverage of either the query or subject sequences; and a BLAST e-value of < 1e-5. Core, proteins present in animals, plants,
bacteria; Unique, proteins only found in P. tricornutum; Diatom, proteins common to Thalassiosira pseudonana and P. tricornutum.

represented GO enrichments were for protein coding genes


belonging to several metabolic pathways involved in carbohydrate
metabolism and catabolism (Fig. 3a). This points to the occurrence of a general shift in carbon metabolism in response to Pi
depletion. The regulation of intermediate carbon metabolism has
been demonstrated previously to be essential for the growth optimization of N-starved P. tricornutum (Allen et al., 2011; Levitan
et al., 2015), which suggests that the same adaptive metabolic
processes also operate in response to P starvation. Under late Pi
depletion, cells were probably severely P starved because of the
prolonged Pi scarcity and possibly the consumption of internal
reserves. Other over-represented GO enrichments in the upregulated genes during this stage were ubiquitin cycle-associated
terms, and genes encoding components in intracellular signaling
cascades and chromatin organization (Fig. 3a). The downregulated genes most significantly over-represented during early Pi
2015 The Authors
New Phytologist 2015 New Phytologist Trust

depletion encoded proteins involved in the photosynthetic light


reactions, and transcripts for these genes declined further during
late Pi depletion (Fig. 3b). The over-representation of the downregulated genes relating to photosynthesis also occurred during
the late exponential phase when cell population growth was
entering into stationary phase (Fig. 1a). Reduction of the photosynthetic machinery by the downregulation of genes involved in
the light reactions in photosystem II (PSII) and photosystem I
(PSI) is ultimately beneficial to a cell being subjected to a nutritive imbalance by reducing the generation site of reactive oxygen
species, and hence oxidative stress. A reduction in photosynthetic
activity has been reported previously to occur in P. tricornutum
under P scarcity (Yang et al., 2014), as well as under N stress
(Yang et al., 2013; Levitan et al., 2015). Our data highlight the
existence of a stringent relationship between nutrient supply and
the genomic control of photosynthesis. When Pi was resupplied
New Phytologist (2015)
www.newphytologist.com

New
Phytologist

6 Research
(a)

(b)
GO enrichments odds rao

GO enrichments odds rao


+Pi 4d

Pi 8d

C 8d

+Pi 4d

Pi 8d

Phosphate transport
Nucleode-sugar metabolic process
Inorganic anion transport
Anion transport
Glycolysis
Glucose catabolic process
Monosaccharide catabolic process
Hexose catabolic process
Alcohol catabolic process
Carbohydrate catabolic process
Cellular carbohydrate catabolic process
Glucose metabolic process
Hexose metabolic process
Monosaccharide metabolic process
Cellular carbohydrate metabolic process
Alcohol metabolic process
Ubiquin cycle
Unsaturated fay acid biosynthec process
Unsaturated fay acid metabolic process
Phosphoenolpyruvate-dependent sugar phosphotransferase system
Regulaon of metabolic process
Intracellular signaling cascade
DNA replicaon iniaon
DNA packaging
Chroman assembly or disassembly
Chroman assembly
Chroman organizaon
Protein-DNA complex assembly
Nucleosome assembly
Nucleosome organizaon
Chromosome organizaon
Deoxyribonucleode metabolic process
Organelle organizaon
Regulaon of gene expression
Regulaon of macromolecule metabolic process
Establishement of localizaon
Localizaon
Gene expression
Transport
Nitrogen compound metabolic process
Nucleobase, nucleoside, nucleode and nucleic acid metabolic process

12.0
Pi 4d

2.0

12.0
Pi 4d

C 8d

2.0

Negave regulaon of catalyc acvity


Negave regulaon of molecular funcon
Photosynthesis
Photosynthesis, light harvesng
Photosynthesis, light reacon
Amine transport
Amino acid transport
Secondary metabolic process
Carboxylic acid transport
Organic acid transport
Porphirin biosynthec process
Tetrapyrrol biosynthec process
Tetrapyrrol metabolic process
Porphyrin metabolic process
Isoprenoid biosynthec process
Isoprenoid metabolic process
Response to oxidave stress
Branched chain family amino acid metabolic process
Inorganic anion transport
Anion transport
ncRNA processing
ncRNA metabolic process
RNA processing
tRNA metabolic process
RNA metabolic process
Cyclic nucleode metabolic process
Cyclic nucleode biosynthec process
Nucleoside monophosphate biosynthec process
Nucleoside monophosphate metabolic process
Heterocycle biosynthec process
Nucleoside phosphate metabolic process
Nucleode metabolic process
Cofactor biosynthec process
Cofactor metabolic process
Transcripon
Unsaturated fay acid metabolic process
Unsaturated fay acid biosynthec process
Fay acid biosynthec process
Fay acid metabolic process
Dicarboxylic acid metabolic process
Pigment biosynthec process
Lipid biosynthec process
Cellular lipid metabolic process

Fig. 3 Gene ontology (GO) enrichment analysis of (a) upregulated and (b) downregulated genes in the transcriptome response of Phaeodactylum
tricornutum to inorganic phosphate (Pi) fluctuations. The top 15 (highest odds ratio) enriched GO terms of the biological process category in each sample
are shown. The odds ratio was defined as the ratio of the proportion of a GO term in (a) upregulated and (b) downregulated genes to the proportion of
this GO term in all diatom genes. The larger the odds ratio, the higher the relative abundance of this GO term compared with background. Multiple-test
adjusted P < 0.05 was used to define statistical significance. Pi 4d, 4 d Pi depleted; Pi 8d, 8 d Pi depleted; +Pi 4d, 4 d Pi resupply after 4 d Pi depletion; C
8d, 8 d control conditions.

to the medium, the over-represented upregulated genes coded for


proteins involved in unsaturated fatty acid biosynthetic/
metabolic processes and DNA replication initiation and DNA
packaging (Fig. 3a); these proteins are involved in the resumption
of active cell growth under recovery (Fig. 1c). Furthermore, other
over-represented upregulated genes during recovery included
those putatively involved in epigenetic modifications, such as
chromatin and nucleosome assembly, DNA packaging and chromosome assembly (Fig. 3a), suggesting a new acclimated epigenomic state anticipating repeated Pi-depleted conditions.
Optimization of P scavenging under Pi depletion
Common strategies used by microorganisms and plants to optimize Pi scavenging in response to Pi scarcity include an increase
in the number of Pi transporters and/or replacement of lowaffinity transporters with higher affinity ones, and the production
of phosphatases to utilize organic P sources from the environment (Clark et al., 1998; Riegman et al., 2000; Vance et al.,
2003). In a recent report, five of the six annotated Pi transporter
genes in P. tricornutum were upregulated after 48 h of Pi depletion (Yang et al., 2014). Manual curation and analysis of conserved domains enabled us to identify several new putative Pi
New Phytologist (2015)
www.newphytologist.com

transporter genes in P. tricornutum, with the number increasing


to 24 (Fig. 4a). This is approximately the same number of Pi
transporters as annotated in Arabidopsis thaliana (The
Arabidopsis Information Ressource, TAIR, database), whereas
Chlamydomonas reinhardtii has 14 annotated putative Pi transporters (Grossman & Aksoy, 2015). Twelve of the Pi transporter
genes in P. tricornutum were highly upregulated under Pi depletion, with 11 being dramatically downregulated on Pi resupply
(Fig. 4a), revealing a tight regulation of their expression level by
environmental Pi concentration. These results suggest that
diatoms are equipped with a highly responsive Pi transporter system programmed to operate only when Pi is scarce. Five of the
highly induced Pi transporters have been annotated as Na+/Pi cotransporters, a type of transporter that occurs in phytoplankton,
yeast and vertebrates, but which is lacking in plants. Interestingly,
the five Na+/Pi co-transporters in P. tricornutum share more
sequence homology to mammalian renal Pi transporters than to
yeast, bacterial or green algae (Chlamydomonas) Na+/Pi cotransporters.
Also supporting the notion of optimized Pi uptake under Pi
depletion was the very high induction of five genes coding for
AlkPases, with three expressed > 1000-fold under late Pi stress
(Fig. 4a). Marine bacteria have been shown to have a large
2015 The Authors
New Phytologist 2015 New Phytologist Trust

New
Phytologist

Research 7

Log2 expression rao


0

Na+/Pi co-transporter
Na+/Pi co-transporter
Pi transporter
Pi transporter
Na+/Pi co-transporter
Pi transporter
Na+/Pi co-transporter
Na+/Pi co-transporter
Na+/Pi co-transporter
Pi transporter
HYP
HYP
Pi transporter
Na+/tricarboxylate and Pi transporter
HYP
HYP
Na+/tricarboxylate and Pi transporter
Mito Pi carrier protein
Mito Pi carrier protein
Mito Pi carrier protein
Mito Pi carrier protein
Mito Pi carrier protein
HYP
Na+/Pi co-transporter

39432
49678
47869
45959

Alkaline phosphatase
Alkaline phosphatase
HYP
Alkaline phosphatase

+Pi 4d

Pi 8d

Membrane lipid metabolism

47239
47666
39515
23830
47667
22315
21441
33266
40433
46692
19030
22279
1277
14922
43646
17265
22453
47954
462
11414
37916
8987
11866
49842

3.0

42872
43665
50356
42467
9619
46400
49771
12431
46908
8860
1000
1611
40261
43116
11390
48445
54168
42683
12884
14125
3262
36877
34555
47576
43099

Putave DGTA synthesis enzyme


Phosphadylinositol-specic phospholipase C (PI-PLC)
Sulfoquinovosyldiacylglycerol 2 (SQD2)
Sulfoquinovosyldiacylglycerol 2 (SQD2)
Monogalactosyldiacylglycerol synthase 2 (MGD2)
Phospholipase D (PLD)
Phosphadylinositol-specic phospholipase C (PI-PLC)
Phospholipase D (PLD)
Phosphadylinositol-specic phospholipase C (PI-PLC)
Phospholipid:diacylglycerol acyltransferase (PDAT)
Phosphadylinositol-specic phospholipase C (PI-PLC)
Phosphadylinositol-specic phospholipase C (PI-PLC)
Phosphadate phosphatase (PAP1)
Digalactosyl diacylglycerol synthase 1 (DGD1)
Digalactosyl diacylglycerol synthase 1 (DGD1)
Phosphadylinositol-specic phospholipase C (PI-PLC)
Monogalactosyldiacylglycerol synthase 1 (MGD1)
Phosphadylinositol-specic phospholipase C (PI-PLC)
Digalactosyl diacylglycerol synthase 1 (DGD1)
Monogalactosyldiacylglycerol synthase 1 (MGD1)
Phospholipid/glycerol acyltransferase (ACT1)
Phospholipase D (PLD)
Phospholipase D (PLD)
Phospholipase D (PLD)
Phospholipid/glycerol acyltransferase (ACT1)

PolyP
metabolism

C 8d

+Pi 4d

3.0

Pi 8d

Pi 4d

C 8d
Pi transporters
AlkPhases

(b)
3.0

Log2 expression rao

3.0

Pi 4d

(a)

50019
15281
48538
54257

VTC
VTC
VTC
VTC

Fig. 4 Heat maps of log2 fold expression changes under inorganic phosphate (Pi) fluctuation of putative (a) Pi transporters and alkaline phosphatase
(AlkPase) coding genes and (b) membrane lipid metabolism and polyphosphate metabolism (PolyP) coding genes in Phaeodactylum tricornutum. Numbers
correspond to Joint Genome Institute (JGI) protein identifiers. Pi 4d, 4 d Pi depleted; Pi 8d, 8 d Pi depleted; +Pi 4d, 4 d Pi resupply after 4 d Pi depletion;
C 8d, 8 d control conditions.

fraction of AlkPases secreted under Pi-deficient conditions (Luo


et al., 2009). In P. tricornutum, one of the AlkPases (49678) was
predicted to be secreted. This AlkPase has been biochemically
characterized previously (termed PtAPase) and has been shown to
be released to the medium in response to Pi depletion (Lin et al.,
2013). In T. pseudonana, several AlkPases have been shown to be
upregulated on Pi depletion in both the transcriptome and proteome (Dyhrman et al., 2012). Furthermore, the authors also
showed that AlkPase activity triggered by Pi depletion was mainly
surface associated with some intracellular location (Dyhrman
et al., 2012). This, together with our data, suggests a similar location of activity of P. tricornutum AlkPases.
Maintenance of P homeostasis in diatoms has common and
distinct characteristics compared with higher plants
The ability to economize cellular Pi demands by the replacement
of membrane phospholipids by sulfolipids is known to occur in
plants (Vance et al., 2003), in Chlamydomonas (Grossman &
Aksoy, 2015) and in phytoplankton (Van Mooy et al., 2009;
Martin et al., 2011; Dyhrman et al., 2012; Abida et al., 2015).
Under prolonged Pi-depleted conditions (13 d), a total disappearance of phospholipids, with a concomitant increase in diacylglyceryl-hydroxymethyl-N,N,N-trimethyl-b-alanine
(DGTA),
sulfoquinovosyldiacylglycerol (SQDG) and digalactosyldiacylglycerol (DGDG), has been reported recently in P. tricornutum
(Abida et al., 2015). Membrane phospholipids in P-starved Arabidopsis plants have been shown to be cleaved by phospholipases
C (Nakamura et al., 2005; Gaude et al., 2008) and D (CruzRamrez et al., 2006). In this work, two phosphatidylinositolspecific phospholipases C (PI-PLC) were specifically upregulated
2015 The Authors
New Phytologist 2015 New Phytologist Trust

on early Pi depletion, and three others were upregulated during


late Pi stress (Fig. 4b). Two genes encoding phospholipases D
were also upregulated throughout Pi depletion (Fig. 4b). This
supports the occurrence of membrane phospholipid degradation
under Pi depletion, similar to that which occurs in Arabidopsis.
It also shows that the process of membrane phospholipid degradation under Pi depletion is regulated at the transcriptional level.
Different from the observations in Arabidopsis, the polar lipid
that increases most strongly in P. tricornutum membranes under
Pi stress is DGTA, a betaine glycerolipid (Abida et al., 2015).
This is corroborated by our transcriptome data, where a putative
DGTA enzyme (ARF4, 42872) was expressed > 30-fold under
early Pi depletion and > 80-fold under late Pi depletion (Fig. 4b).
Two genes coding for sulfolipid synthases (SQD2) were also
highly induced during early and late Pi depletion (up to 13-fold
expression increase) and, to a lesser extent, a gene coding for a
monogalactosyldiacylglycerol synthase (MGD2) (Fig. 4b). These
data corroborate, at the transcriptional level, previous findings
(Yang et al., 2014; Abida et al., 2015). They also confirm that
membrane remodeling events in P. tricornutum are tightly regulated at the transcriptional level and are related directly to P
scarcity.
Neutral lipid accumulation is a common response of microalgae under unfavorable conditions, namely nutrient stress (Fields
et al., 2014), and, in the present study, P. tricornutum cells accumulated large lipid bodies after 8 d of Pi depletion (Fig. 2c). We
suggest that the lipids resulting from the remodeling of membranes were shifted in the form of triacylglycerol (TAG) to the
nascent lipid bodies. In Dunaliella salina, fatty acids typically
found in chloroplast membrane galactolipids have been detected
in storage lipids (Cho & Thompson, 1986). Membrane
New Phytologist (2015)
www.newphytologist.com

8 Research

phospholipids can enter the TAG metabolic pathway by direct


conversion to diacylglycerol (DAG) by phospholipid diacylglycerol acyltransferase (PDAT)-catalyzed transesterification. The
gene encoding PDAT, an acyl-CoA-independent pathway
enzyme for the production of TAGs, initially characterized in
yeast (Dahlqvist et al., 2000; Oelkers et al., 2000), has an
ortholog in A. thaliana (Stahl et al., 2004) and in C. reinhardtii
(Merchant et al., 2012). An ortholog of PDAT is also present in
P. tricornutum (8860), being specifically upregulated during early
and late Pi stress (Fig. 4b), further supporting a role for membrane phospholipid recycling to the TAG pathway during Pi
depletion. A recent report has shown that N depletion in
P. tricornutum leads to the remodeling of membranes, and that
the PDAT gene ortholog is concomitantly upregulated, supporting its involvement in the movement of membrane phospholipids
to the TAG pathway (Yang et al., 2013).
Polyphosphate (PolyP), the most important phosphate reserve
used on Pi depletion in yeast, is synthesized through the vacuolar
transporter chaperone (VTC) complex (Ogawa et al., 2000).
Unlike plants, diatoms have VTC orthologs, and these proteins
have been shown to increase Pi allocation to PolyP in response to
N deficiency (Perry, 1976) and Pi deficiency (Dyhrman et al.,
2012). Genes encoding VTCs appeared to be regulated, at least
partially, at the transcriptional level, as two genes coding for
putative VTC proteins were significantly upregulated in response
to Pi deficiency in T. pseudonana (Dyhrman et al., 2012). Four
genes encoding proteins with a VTC domain (IPR018966) were
detected in P. tricornutum, with two being highly upregulated in
response to Pi depletion and the other two in response to Pi
resupply (Fig. 4b), suggesting distinct roles in PolyP metabolism
in response to Pi concentration.
Long noncoding transcripts specifically induced in response
to Pi depletion
The P starvation response in plants involves the induction of several RNA transcripts with no protein coding potential, which
have been described as having a key role in the regulation of P
homeostasis. These noncoding RNAs include several precursors
of micro-RNAs (miRNAs), such as miR399 and miR827,
involved in the regulation of signaling pathways and Pi transport
under Pi depletion (Bari et al., 2006; Hsieh et al., 2009).
Although the function of miRNAs in diatoms remains to be verified (Lopez-Gomollon et al., 2014; Rogato et al., 2014), gene
silencing has been shown to occur (De Riso et al., 2009). In the
present work, two annotated miRNA precursors (pti-MIR5472
and pti-MIR5471; Huang et al., 2011) were significantly upregulated in response to Pi stress in P. tricornutum (Table S3),
although functional studies are needed to examine the relevance
of these transcripts in the diatom Pi response.
Pi starvation in plants also involves longer nonprotein coding
transcripts, such as INDUCED BY PHOSPHATE
STARVATION 1 (IPS1). This transcript has a small ORF, but
was later found to be active as an RNA molecule, target mimicking miR399 (Franco-Zorrilla et al., 2007). IPS1 is recognized by
miR399 as a target; however, as a result of mismatches in the
New Phytologist (2015)
www.newphytologist.com

New
Phytologist
recognition sequence, their binding does not lead to miR399assisted cleavage and remains stable, which results in the finetuning of Pi uptake (Franco-Zorrilla et al., 2007). We have thoroughly analyzed the P. tricornutum noncoding transcriptome and
have identified 1510 putative lincRNAs (Table S4), 202 of which
were specifically upregulated in response to Pi depletion and
downregulated when Pi was resupplied to the medium. The 1510
lincRNA sequences identified in this study were compared with
the 10 402 mRNAs obtained from Phatr2. Phaeodactylum
tricornutum lincRNAs are significantly shorter than mRNAs with
c. 90% at < 1000 nucleotides (Fig. 5a,b). Most of these diatom
lincRNAs (> 70%) lack ORFs with coding potential in any of the
three possible reading frames (Fig. 5c). Nonetheless, approximately 28% of the identified lincRNAs have the presence of a
very short putative ORF (< 100 amino acids) (Fig. 5c,d), which
could potentially be translated into a short peptide. The majority
of the P. tricornutum lincRNAs are intronless (90%) with only a
small fraction containing a single intron (Fig. 5e) of similar size
to the introns found in mRNAs (Fig. 5f). The exons of lincRNAs
are shorter than those of protein coding transcripts (Fig. S4a,c)
and the distance between stop codons and the longest stop
codon-free sequences are shorter in diatom lincRNAs than in
protein coding transcripts (Fig. S4d,e). LincRNAs also have
lower GC content and lower free energy compared with mRNAs
(Figs S4f, S5, S6). Interestingly, all of these genomic features have
also been detected in functional human lincRNAs (Niazi & Valadkhan, 2012), suggesting conserved features of these molecules
across kingdoms.
To search for lincRNAs with a similar function to that of IPS1
in higher plants, we analyzed whether the P. tricornutum Piresponsive lincRNAs had a predicted target mimicry function
(Wu et al., 2013). No target mimicry function could be predicted
in any of the Pi-responsive lincRNAs identified in this work.
Using recently publically available RNA-Seq reads (accession no.
SRP040703), we sought to investigate the expression changes of
the Pi-responsive lincRNAs under 48 h N depletion (Levitan
et al., 2015). Amongst the top 20 upregulated lincRNAs in Pi,
only two were significantly upregulated under N (two-fold difference and P < 0.05) (Fig. 6a), revealing very specific responses
of the noncoding transcriptome to stress conditions. The fact that
these diatom lincRNAs were specifically expressed in response to
Pi depletion (Figs 6a, S3) suggests that at least some of these transcripts could have regulatory roles in P. tricornutum.
LincRNAs have been shown to associate with chromatin
remodeling complexes and to affect gene expression by cis- and
trans-action (De Lucia & Dean, 2011; Nagano & Fraser, 2011;
Guttman & Rinn, 2012; Bonasio & Shiekhattar, 2014; Liu et al.,
2015). We investigated the coexpression correlation between the
Pi-responsive lincRNA genes and their neighboring genes to
explore putative cis-regulatory functions of the lincRNAs. Several
potential gene targets have been identified as having a correlation
(positive or negative) to lincRNAs (Fig. 6b; Table S5). These
results can be used to further investigate the function of diatom
lincRNAs in the context of Pi depletion, but also in the context
of other abiotic stresses that have putative common responsive
molecular modules, such as N stress.
2015 The Authors
New Phytologist 2015 New Phytologist Trust

New
Phytologist

Research 9

LincRNA

0.5

Frequency (%)

P value = 4.12e243

(b)

0.6

Length of unspliced
transcripts (nt)

(a)

mRNA

0.4
0.3
0.2
0.1

3500
3000
2500
2000
1500
1000
500

<
50 500
0
10 100
00 0
1
15 500
00

20 200
00 0

25 250
00 0
3
30 000
00

35 350
00 0

40 400
00 0
4
45 500
00
5
00
0
>
50
00

Length of unspliced transcripts (nt)


(c)

Length of predicted
ORFs (aa)

LincRNA
80

Frequency (%)

P value = 1.63e-124

(d)

100

mRNA
60
40
20

1000
900
800
700
600
500
400
300
200
100
0

0
0

>1

Number of predicted ORFs per transcript


(e)

(f)

P value 1

100

Frequency (%)

mRNA
60
40
20

200
150
100
50

0
0

Diatom TFs as coordinators of the Pi transcriptomic


response
The dynamic nature of the diatom transcriptomic response to Pi
fluctuation is ultimately controlled by a network of TFs, whose
activity changes in response to the varying signals, thereby coordinating the activation of target genes. We have detected the presence of 275 genes coding for putative TFs in P. tricornutum,
which include all the previously annotated TFs (Rayko et al.,
2010) and several other newly identified putative TFs (Table S6).
These TFs have been manually curated for the presence of DNA
binding domains using INTERPRO (http://www.ebi.ac.uk/Tools/
InterProScan/), although we cannot completely exclude the presence of false positives. In response to Pi fluctuation, 62.5% of the
2015 The Authors
New Phytologist 2015 New Phytologist Trust

Length of introns (nt)

LincRNA

80

Fig. 5 Genomic structure of Phaeodactylum


tricornutum long intergenic nonprotein
coding RNAs (lincRNAs). (a, b) Length
distribution in nucleotides (nt) of unspliced
lincRNAs and mRNAs; (c) number of
predicted open reading frames (ORFs) and
(d) length of putative ORFs in amino acids
(aa) considering the three reading frames of
lincRNAs and mRNAs; (e) intron number and
(f) length of introns in nucleotides (nt) in
lincRNAs and mRNAs.

>5

Number of introns

genes for putative TFs were differentially expressed, revealing a


very dynamic response. Approximately 32% of the upregulated
TFs throughout Pi depletion belong to the HSF family (Fig. 7a).
When Pi was resupplied to the medium, there was a switch
between the abundance of TF families, with the Myeloblastosis
family (MYB) being over-represented (Fig. 7a). These results support a role for HSFs in the present stress response, which is in
agreement with the function of this class of TFs in higher plants
and mammals (
Akerfelt et al., 2010). However, as in Arabidopsis
(Nover et al., 2001), the significance of the abundance of the
HSF family in diatoms remains to be resolved. In Chlamydomonas and in vascular plants, the key regulators of the Pi
response are TFs (PSR1 and PHR1, respectively) of the MYB
family with a coiled-coil domain (MYB-CC) (Wykoff et al.,
New Phytologist (2015)
www.newphytologist.com

New
Phytologist

10 Research

+Pi 4d

Pi 8d

C 8d

Pi 4d

+Pi 4d

Pi 8d

C 8d

Pi 4d

+Pi 4d

C 8d

Pi 8d

Log2 expression ratio

(b)

Pi 4d

(a)
Log2 expression ratio

+Pi 4d

Pi 8d

XLOC_006307
XLOC_005436
XLOC_003159
XLOC_008284
XLOC_005700
XLOC_004740
XLOC_015130
XLOC_010279
XLOC_013710
XLOC_013110
XLOC_005097
XLOC_008283
XLOC_012004
XLOC_009323
XLOC_012001
XLOC_005083
XLOC_004568
XLOC_003589
XLOC_008370
XLOC_006663

D
D
U

Upstream
genes

LincRNAs

1999; Rubio et al., 2001). In diatoms, there are 34 annotated


MYB family TFs, but none is MYB-CC with the characteristic
LHEQLE conserved motif, suggesting that this domain was
acquired after the separation of green and red algae from the
common ancestor.
Amongst the TF genes upregulated during early Pi depletion,
genes for two HSFs (HSF3.1b and HSF4.7a) returned to their
control levels during late Pi depletion (Fig. 7b). This was further
confirmed by RT-qPCR (Fig. S7), suggesting a role for these TFs
in the coordination of an early response-specific gene network.
Assuming that these TFs will bind to the conserved heat shock
element (HSE) present in the upstream promoter region of putative target genes (Nover et al., 2001; 
Akerfelt et al., 2010), we
generated a transcription activation network between these TFs
and the significantly coexpressed genes (Fig. 7c). Several signaling/sensing genes and other genes coding for other HSF and
MYB TFs were present in this early responsive network, as well
as several lincRNA genes (Table S7). This gene regulatory model
can be used to design new studies to further dissect the relationship between the two transiently expressed TFs and the coexpressed genes, to deepen our understanding of the molecular
networks underlying diatom early physiological adaptations to P
depletion.
Sensing and signaling of environmental Pi fluctuations
involves a singular mix-and-match of genes and pathways
Transmembrane or cell surface receptors in unicellular organisms, such as diatoms, are likely to play primordial roles in sensing environmental changes, and the resulting signals are then
New Phytologist (2015)
www.newphytologist.com

XLOC_003268
XLOC_000522
XLOC_014096
XLOC_012003
XLOC_005083
XLOC_012001
XLOC_009323
XLOC_007795
XLOC_008370
XLOC_012005
XLOC_004569
XLOC_003517
XLOC_003820
XLOC_003072
XLOC_006831
XLOC_001033
XLOC_015215
XLOC_013578
XLOC_012002
XLOC_011959
XLOC_000944
XLOC_009426
XLOC_010053
XLOC_013214
XLOC_014025
XLOC_002204
XLOC_014931
XLOC_002701
XLOC_010537
XLOC_003159
XLOC_015596
XLOC_001067
XLOC_010415
XLOC_009204
XLOC_011480
XLOC_015595
XLOC_006121
XLOC_002503
XLOC_010061
XLOC_014912
XLOC_000622
XLOC_010157
XLOC_012114
XLOC_009535

5.0

Pi 4d

C 8d

5.0

5.0

Downstream
genes

5.0

Fig. 6 Long intergenic nonprotein coding


RNAs (lincRNAs) involved in Phaeodactylum
tricornutum responses to inorganic
phosphate (Pi) depletion. (a) Heat map of the
top 20 upregulated lincRNAs (P < 0.05)
under early Pi depletion (4 d, this work) and
early N depletion (N) (2 d, from Levitan
et al., 2015; RNA sequencing (RNA-Seq)
reads obtained from accession no.
SRP040703). C 8d, 8 d control conditions;
Pi 4d, 4 d Pi depleted; Pi 8d, 8 d Pi
depleted; +Pi 4d, 4 d Pi resupply after 4 d Pi
depletion; U, significantly upregulated under
N (more than two-fold, P < 0.05); D,
significantly downregulated under N (more
than two-fold, P < 0.05). (b) Putative cisregulatory protein coding targets of Piresponsive lincRNAs (Supporting Information
Table S5). Neighboring upstream and/or
downstream putative cis-targets were
identified according to positive or negative
expression correlation (r2 0.6) to lincRNAs.

transmitted to intracellular regulatory pathways, enabling the


cell to respond, regulate and adjust its metabolism to maintain
cellular homeostasis. We detected 659 genes putatively encoding
sensing and signaling functions (Table S8), 42 of which were
predicted to be membrane localized and therefore to have a
putative receptor/sensing function. These receptors constitute a
collection of several representatives of the signaling pathways
typically described in metazoans (G protein-coupled receptors
(GPCRs), serine/threonine (Ser/Thr) receptor kinases and/or
tyrosine (Tyr) receptor kinases) and bacteria (histidine kinase
receptors). Of the total putative sensing/signaling genes, 65.5%
were differentially expressed under Pi fluctuation. Amongst the
membrane-localized putative protein kinases were five histidine
kinase (HK) domain-containing proteins, the gene of one of
which, with an extracellular PAS domain (a protein domain
functioning as a signal sensor), was upregulated in response to
Pi stress (45485). Membrane HKs are major sensors of environmental changes in bacteria, regulating the activity of a second
component, through a phosphorelay between the C-terminal
HK domain and the aspartate residue of the response regulator
(Laub & Goulian, 2007). Several annotated genes for Tyr and/
or Ser/Thr kinases were also upregulated in response to Pi depletion, with four such kinases being putatively membrane located.
Genes encoding GPCR signaling pathway components have
been documented to exist in diatoms (Port et al., 2013). GPCR
proteins, found in most eukaryotic organisms, are cell surface
receptors that play a major role in signal transduction, perception and response to the environment (Fredriksson & Schioth,
2005). Four genes coding for GPCRs of the rhodopsin/class A
family and two of the glutamate/class C were significantly
2015 The Authors
New Phytologist 2015 New Phytologist Trust

New
Phytologist

Research 11
(b)
Log2 expression rao

Pt_HSF1.1a
Pt_HSF4.3c
Pt_HSF4.3a
Pt_bZIP15
Pt_HSF3.1b
Pt_HSF4.4b
Pt_CCHH5
Pt_HSF1.2b
Pt_HSF4.7a

HSF, 18

ZnFinger, 6

+Pi 4d

other, 8

bHLH, 1
CXC, 2

3.0

C 8d

Hox, 2

3.0

Pi 8d

TFs upregulated -Pi 4d (50)

Pi 4d

(a)

Myb, 8
bZIP, 5

TFs upregulated -Pi 8d (47)

Fig. 7 Transcription factors (TFs) responsive


to inorganic phosphate (Pi) depletion (more
than two-fold differentially expressed,
P < 0.05) in Phaeodactylum tricornutum. (a)
Pie charts of the different classes of TFs
upregulated under early Pi depletion (Pi
4d), late Pi depletion (Pi 8d) and resupply
(+Pi 4d). (b) Heat maps of the top 10 most
upregulated TFs during Pi depletion. C 8d
corresponds to control cultures during late
exponential growth in full medium. (c)
Inferred transcription activation pathways
under early Pi depletion of the two
transiently expressed heat shock factors
(HSFs) (Supporting Information Table S7):
3.1b, HSF3.1b (45560); 4.7a, HSF4.7a
(1650). Blue dots correspond to protein
coding genes, red dots correspond to long
intergenic nonprotein coding RNAs
(lincRNAs).

CBF/NF2,
1
Hox, 2

other
,6

bHLH, 3
CXC, 1
ZnFinger, 9

Myb, 6
bZIP, 6

TFs upregulated in +Pi 4d (42)

3.1b 4.7a

HSF, 6
other, 10
E2F, 2

Myb, 12

bHLH, 1
CXC, 3
Zn-

upregulated during P starvation. Other signaling genes encoding


proteins downstream of these receptors were found to be upregulated by Pi depletion, including genes for several adenylate
cyclase and Ras small GTPase. These results suggest that the signaling pathways may be fully activated and presumably functional in response to P scarcity in P. tricornutum.
Conclusion
Genome-wide transcriptome-enabled studies have proven useful
in providing mechanistic insights into the complexity of phytoplankton responses and adaptations to recurrent environmental
challenges (Mock et al., 2008; Shrestha et al., 2012; Thamatrakoln et al., 2012; Ashworth et al., 2013; Levitan et al., 2015).
Our data show that diatoms have a multitude of processes and
pathways enabling survival and resilience under P scarcity that
combine features possibly derived from different sources during
their evolutionary history. These processes are involved not only
in the optimization of P scavenging, but also in reducing cellular P demand by the internal recycling of P sources. Globally,
these strategies seem to be general processes used by diatoms
when confronting nutrient limitations in the marine medium,
as similar strategies were detected in response to the depletion
of iron (Allen et al., 2008), cobalamin (Bertrand et al., 2012)
and N (Levitan et al., 2015). The apparent dynamic, ongoing,
putative horizontal transfer of critical genes between marine
2015 The Authors
New Phytologist 2015 New Phytologist Trust

(c)
HSF, 13

bZIP, 4

bacteria, viruses and this diatom (i.e. phosphate transporters,


membrane signaling receptors, metabolism- and redox-related
genes) probably complements and enhances its molecular capacity to successfully respond and adapt to fluctuating Pi environments, hence contributing to its ecological success in the global
oceans.
Perhaps the most striking feature revealed by this work was the
identification of a multitude of lincRNAs specifically expressed
under Pi depletion. LincRNAs are ubiquitous molecules which,
in other model systems (yeast, animals and plants), are beginning
to attract considerable attention because of their important regulatory roles during development and stress (De Lucia & Dean,
2011; Nagano & Fraser, 2011; Guttman & Rinn, 2012; Bonasio
& Shiekhattar, 2014; Liu et al., 2015). These findings provide a
glimpse into the complexity of the molecular regulatory program
in diatoms, and open up additional routes to explore the function
and evolutionary significance of the nonprotein coding transcriptome in these ecologically successful organisms.

Acknowledgements
We thank Connie Zhao and Bin Zhang (Genomics Resource
Center, Rockefeller University), as well as Kaye Thomas, Pablo
Ariel and Tao Tong (Bio-imaging Resource Center, Rockefeller
University), for technical support. We also thank Jun Liu and
Huan Wang for help with the bioinformatic analysis, and the
New Phytologist (2015)
www.newphytologist.com

12 Research

four anonymous reviewers for their constructive comments.


M.H.C.C. and part of this work were supported by the Seventh
Research Program of the European Union FP7/2007-2013 under
Marie Curie Grant PIOF-GA-301466.

Author contributions
M.H.C.C., H-X.S. and N-H.C. planned and designed the
research; M.H.C.C. and H-X.S. performed the research;
M.H.C.C., H-X.S. and N-H.C. analyzed the data; M.H.C.C.,
H-X.S., C.B. and N-H.C. wrote the manuscript.

References
Abida H, Dolch L-J, Mei C, Villanova V, Conte M, Block MA, Finazzi G,
Bastien O, Tirichine L, Bowler C et al. 2015. Membrane glycerolipid
remodeling triggered by nitrogen and phosphorus starvation in Phaeodactylum
tricornutum. Plant Physiology 167: 118136.

Akerfelt M, Morimoto RI, Sistonen L. 2010. Heat shock factors: integrators of
cell stress, development and lifespan. Nature Reviews Molecular Cell Biology 11:
545555.
Allen AE, Dupont CL, Obornk M, Horak A, Nunes-Nesi A, McCrow JP,
Zheng H, Johnson DA, Hu H, Fernie AR et al. 2011. Evolution and
metabolic significance of the urea cycle in photosynthetic diatoms. Nature 473:
203207.
Allen AE, Laroche J, Maheswari U, Lommer M, Schauer N, Lopez PJ, Finazzi
G, Fernie AR, Bowler C. 2008. Whole-cell response of the pennate diatom
Phaeodactylum tricornutum to iron starvation. Proceedings of the National
Academy of Sciences, USA 105: 1043810443.
Armbrust EV, Berges JA, Bowler C, Green BR, Martinez D, Putnam NH, Zhou
S, Allen AE, Apt KE, Bechner M et al. 2004. The genome of the diatom
Thalassiosira pseudonana: ecology, evolution, and metabolism. Science 306: 79
86.
Ashworth J, Coesel S, Lee A, Armbrust EV, Orellana MV, Baliga NS. 2013.
Genome-wide diel growth state transitions in the diatom Thalassiosira
pseudonana. Proceedings of the National Academy of Sciences, USA 110: 7518
7523.
Bari R, Datt Pant B, Stitt M, Scheible W-R. 2006. PHO2, microRNA399, and
PHR1 define a phosphate-signaling pathway in plants. Plant Physiology 141:
988999.
Bauerfield E, Hickle W, Niermann U, Westerhagen V. 1990. Phytoplankton
biomass and potential nutrient limitation of phytoplankton development in the
southeastern North Sea in spring 1985 and 1986. Netherlands Journal of Sea
Research 25: 131142.
Bertrand EM, Allen AE, Dupont CL, Norden-Krichmar TM, Bai J, Valas RE,
Saito MA. 2012. Influence of cobalamin scarcity on diatom molecular
physiology and identification of a cobalamin acquisition protein. Proceedings of
the National Academy of Sciences, USA 109: E1762E1771.
Bonasio R, Shiekhattar R. 2014. Regulation of transcription by long noncoding
RNAs. Annual Review of Genetics 48: 433455.
Bowler C, Allen AE, Badger JH, Grimwood J, Jabbari K, Kuo A, Maheswari U,
Martens C, Maumus F, Otillar RP et al. 2008. The Phaeodactylum genome
reveals the evolutionary history of diatom genomes. Nature 456: 239244.
Burge C, Karlin S. 1997. Prediction of complete gene structures in human
genomic DNA. Journal of Molecular Biology 268: 7894.
Cho SH, Thompson GA. 1986. Properties of a fatty acyl hydrolase preferentially
attacking monogalactosyldiacylglycerols in Dunaliella salina chloroplasts.
Biochimica et Biophysica Acta (BBA) 878: 353359.
Clark LL, Ingall ED, Benner R. 1998. Marine phosphorus is selectively
remineralized. Nature 393: 426.
Cruz-Ramrez A, Oropeza-Aburto A, Razo-Herna ndez F, Ramrez-Cha vez E,
Herrera-Estrella L. 2006. Phospholipase DZ2 plays an important role in
extraplastidic galactolipid biosynthesis and phosphate recycling in Arabidopsis
roots. Proceedings of the National Academy of Sciences, USA 103: 67656770.

New Phytologist (2015)


www.newphytologist.com

New
Phytologist
Dahlqvist A, Stahl U, Lenman M, Banas A, Lee M, Sandager L, Ronne H,
Stymne S. 2000. Phospholipid:diacylglycerol acyltransferase: an enzyme that
catalyzes the acyl-CoA-independent formation of triacylglycerol in yeast and
plants. Proceedings of the National Academy of Sciences, USA 97: 64876492.
De Lucia F, Dean C. 2011. Long non-coding RNAs and chromatin regulation.
Current Opinion in Plant Biology 14: 168173.
De Riso V, Raniello R, Maumus F, Rogato A, Bowler C, Falciatore A. 2009.
Gene silencing in the marine diatom Phaeodactylum tricornutum. Nucleic Acids
Research 37: e96.
Dyhrman S, Ammerman J, Van Mooy B. 2007. Microbes and the marine
phosphorus cycle. Oceanography 20: 110116.
Dyhrman ST, Jenkins BD, Rynearson TA, Saito MA, Mercier ML, Alexander
H, Whitney LP, Drzewianowski A, Bulygin VV, Bertrand EM et al. 2012.
The transcriptome and proteome of the diatom Thalassiosira pseudonana reveal
a diverse phosphorus stress response. PLoS ONE 7: e33768.
Egge JK. 1998. Are diatoms poor competitors at low phosphate concentrations?
Journal of Marine Systems 16: 191198.
Field C, Behrenfeld M, Randerson J, Falkowski P. 1998. Primary production of
the biosphere: integrating terrestrial and oceanic components. Science 281:
237240.
Fields MW, Hise A, Lohman EJ, Bell T, Gardner RD, Corredor L, Moll K,
Peyton BM, Characklis GW, Gerlach R. 2014. Sources and resources:
importance of nutrients, resource allocation, and ecology in microalgal
cultivation for lipid accumulation. Applied Microbiology and Biotechnology 98:
48054816.
Franco-Zorrilla JM, Valli A, Todesco M, Mateos I, Puga MI, Rubio-Somoza I,
Leyva A, Weigel D, Garca JA, Paz-Ares J. 2007. Target mimicry provides a
new mechanism for regulation of microRNA activity. Nature Genetics 39:
10331037.
Fredriksson R, Schioth HB. 2005. The Repertoire of G-protein-coupled
receptors in fully sequenced genomes. Molecular Pharmacology 67: 14141425.
Gaude N, Nakamura Y, Scheible W-R, Ohta H, Doermann P. 2008.
Phospholipase C5 (NPC5) is involved in galactolipid accumulation during
phosphate limitation in leaves of Arabidopsis. Plant Journal 56: 2839.
Grossman AR, Aksoy M. 2015. Algae in a phosphorus-limited landscape. Annual
Plant Reviews 48: 337374.
Guillard RR, Ryther JH. 1962. Studies of marine planktonic diatoms: i.
Cyclotella nana hustedt, and Detonula confervacea (cleve) gran. Canadian
Journal of Microbiology 8: 229239.
Guttman M, Rinn JL. 2012. Modular regulatory principles of large non-coding
RNAs. Nature 482: 339346.
Hsieh L-C, Lin S-I, Shih AC-C, Chen J-W, Lin W-Y, Tseng C-Y, Li W-H,
Chiou T-J. 2009. Uncovering small RNA-mediated responses to phosphate
deficiency in Arabidopsis by deep sequencing. Plant Physiology 151: 2120
2132.
Huang A, He L, Wang G. 2011. Identification and characterization of
microRNAs from Phaeodactylum tricornutum by high-throughput sequencing
and bioinformatics analysis. BMC Genomics 12: 337.
Kim D, Pertea G, Trapnell C, Pimentel H, Kelley R, Salzberg SL. 2013.
TopHat2: accurate alignment of transcriptomes in the presence of insertions,
deletions and gene fusions. Genome Biology 14: R36.
Laub MT, Goulian M. 2007. Specificity in two-component signal transduction
pathways. Annual Review of Genetics 41: 121145.
Levitan O, Dinamarca J, Zelzion E, Lun DS, Guerra LT, Kim MK, Kim J, Van
Mooy BAS, Bhattacharya D, Falkowski PG. 2015. Remodeling of intermediate
metabolism in the diatom Phaeodactylum tricornutum under nitrogen stress.
Proceedings of the National Academy of Sciences, USA 112: 412417.
Lin H-Y, Shih C-Y, Liu H-C, Chang J, Chen Y-L, Chen Y-R, Lin H-T, Chang
Y-Y, Hsu C-H, Lin H-J. 2013. Identification and characterization of an
extracellular alkaline phosphatase in the marine diatom Phaeodactylum
tricornutum. Marine Biotechnology 15: 425436.
Liu J, Jung C, Xu J, Wang H, Deng S, Bernad L, Arenas-Huertero C, Chua
NH. 2012. Genome-wide analysis uncovers regulation of long intergenic
noncoding RNAs in Arabidopsis. The Plant Cell 24: 43334345.
Liu J, Wang H, Chua N-H. 2015. Long noncoding RNA transcriptome of
plants. Plant Biotechnology Journal 13: 319328.

2015 The Authors


New Phytologist 2015 New Phytologist Trust

New
Phytologist
Lopez-Gomollon S, Beckers M, Rathjen T, Moxon S, Maumus F, Mohorianu I,
Moulton V, Dalmay T, Mock T. 2014. Global discovery and characterization
of small non-coding RNAs in marine microalgae. BMC Genomics 15: 697.
Luo H, Benner R, Long RA, Hu J. 2009. Subcellular localization of marine
bacterial alkaline phosphatases. Proceedings of the National Academy of Sciences,
USA 106: 2121921223.
Martin P, Van Mooy BAS, Heithoff A, Dyhrman ST. 2011. Phosphorus supply
drives rapid turnover of membrane phospholipids in the diatom Thalassiosira
pseudonana. The ISME Journal 5: 10571060.
Merchant SS, Kropat J, Liu B, Shaw J, Warakanont J. 2012. TAG, Youre it!
Chlamydomonas as a reference organism for understanding algal triacylglycerol
accumulation. Current Opinion in Biotechnology 23: 352363.
Mock T, Samanta MP, Iverson V, Berthiaume C, Robison M, Holtermann K,
Durkin C, Bondurant SS, Richmond K, Rodesch M et al. 2008. Wholegenome expression profiling of the marine diatom Thalassiosira pseudonana
identifies genes involved in silicon bioprocesses. Proceedings of the National
Academy of Sciences, USA 105: 15791584.
Monier A, Welsh RM, Gentemann C, Weinstock G, Sodergren E, Armbrust
EV, Eisen JA, Worden AZ. 2012. Phosphate transporters in marine
phytoplankton and their viruses: cross-domain commonalities in viralhost
gene exchanges. Environmental Microbiology 14: 162176.
Morel FMM. 2003. The biogeochemical cycles of trace metals in the oceans.
Science 300: 944947.
Nagano T, Fraser P. 2011. No-nonsense functions for long noncoding RNAs.
Cell 145: 178181.
Nakamura Y, Awai K, Masuda T, Yoshioka Y, Takamiya K, Ohta H. 2005. A
novel phosphatidylcholine-hydrolyzing phospholipase C induced by
phosphate starvation in Arabidopsis. The Journal of Biological Chemistry 280:
74697476.
Nelson DM, Treguer P, Brzezinski MA, Leynaert A, Queguiner B. 1995.
Production and dissolution of biogenic silica in the ocean revised global
estimates, comparison with regional data and relationship to biogenic
sedimentation. Global Biogeochemical Cycles 9: 359372.
Niazi F, Valadkhan S. 2012. Computational analysis of functional long
noncoding RNAs reveals lack of peptide-coding capacity and parallels with 3
UTRs. RNA 18: 825843.
Nordberg H, Cantor M, Dusheyko S, Hua S, Poliakov A, Shabalov I, Smirnova
T, Grigoriev IV, Dubchak I. 2014. The genome portal of the Department of
Energy Joint Genome Institute: 2014 updates. Nucleic Acids Research 42: D26
D31.
Nover L, Bjarti K, Kumar MS, Ganguli A, Scharf K-D. 2001. Arabidopsis and
the heat stress transcription factor world: how many heat stress transcription
factors do we need? Cell Stress and Chaperones 6: 177.
Oelkers P, Tinkelenberg A, Erdeniz N, Cromley D, Billheimer JT, Sturley SL.
2000. A lecithin cholesterol acyltransferase-like gene mediates diacylglycerol
esterification in yeast. The Journal of Biological Chemistry 275: 1560915612.
Ogawa N, DeRisi J, Brown PO. 2000. New components of a system for
phosphate accumulation and polyphosphate metabolism in Saccharomyces
cerevisiae revealed by genomic expression analysis. Molecular Biology of the Cell
11: 43094321.
Perry MJ. 1976. Phosphate utilization by an oceanic diatom in phosphoruslimited chemostat culture and in oligotrophic waters of central North-Pacific.
Limnology Oceanography 21: 88107.
Port JA, Parker MS, Kodner RB, Wallace JC, Armbrust EV, Faustman EM.
2013. Identification of G protein-coupled receptor signaling pathway proteins
in marine diatoms using comparative genomics. BMC Genomics 14: 503.
Rayko E, Maumus F, Maheswari U, Jabbari K, Bowler C. 2010. Transcription
factor families inferred from genome sequences of photosynthetic
stramenopiles. New Phytologist 188: 5266.
Riegman R, Stolte W, Noordeloos AAM, Slezak D. 2000. Nutrient uptake and
alkaline phosphatase (ec 3:1:3:1) activity of Emiliania huxleyi
(Prymnesiophyceae) during growth under N and P limitation in continuous
cultures. Journal of Phycology 36: 8796.
Rogato A, Richard H, Sarazin A, Voss BR, Navarro SC, Champeimont RL,
Navarro L, Carbone A, Hess WR, Falciatore A. 2014. The diversity of small
non-coding RNAs in the diatom Phaeodactylum tricornutum. BMC Genomics
15: 119.
2015 The Authors
New Phytologist 2015 New Phytologist Trust

Research 13
Rubio V, Linhares F, Solano R, Martn AC, Iglesias J, Leyva A, Paz-Ares J.
2001. A conserved MYB transcription factor involved in phosphate starvation
signaling both in vascular plants and in unicellular algae. Genes & Development
15: 21222133.
Shrestha RP, Tesson B, Norden-Krichmar T, Federowicz S, Hildebrand M,
Allen AE. 2012. Whole transcriptome analysis of the silicon response of the
diatom Thalassiosira pseudonana. BMC Genomics 13: 499.
Siaut M, Heijde M, Mangogna M, Montsant A, Coesel S, Allen A, Manfredonia
A, Falciatore A, Bowler C. 2007. Molecular toolbox for studying diatom
biology in Phaeodactylum tricornutum. Gene 406: 2335.
Smoot ME, Ono K, Ruscheinski J, Wang P-L, Ideker T. 2011. Cytoscape 2.8:
new features for data integration and network visualization. Bioinformatics 27:
431432.
Stahl U, Carlsson AS, Lenman M, Dahlqvist A, Huang B, Banas W, Banas A,
Stymne S. 2004. Cloning and functional characterization of a
phospholipid:diacylglycerol acyltransferase from Arabidopsis. Plant Physiology
135: 13241335.
Thamatrakoln K, Korenovska O, Niheu AK, Bidle KD. 2012. Whole-genome
expression analysis reveals a role for death-related genes in stress acclimation of
the diatom Thalassiosira pseudonana. Environmental Microbiology 14: 6781.
Trapnell C, Hendrickson DG, Sauvageau M, Goff L, Rinn JL, Pachter L. 2013.
Differential analysis of gene regulation at transcript resolution with RNA-seq.
Nature Biotechnology 31: 4653.
Van Mooy BAS, Fredricks HF, Pedler BE, Dyhrman ST, Karl DM, Koblzek M,
Lomas MW, Mincer TJ, Moore LR, Moutin T et al. 2009. Phytoplankton in
the ocean use non-phosphorus lipids in response to phosphorus scarcity. Nature
458: 6972.
Vance CP, Uhde-Stone C, Allan DL. 2003. Phosphorus acquisition and use:
critical adaptations by plants for securing a nonrenewable resource. New
Phytologist 157: 423447.
Woodward EMS, Owens NJP. 1990. Nutrient depletion studies in offshore
North Sea. Netherlands Journal of Sea Research 25: 5763.
Wu H-J, Wang Z-M, Wang M, Wang X-J. 2013. Widespread long noncoding
RNAs as endogenous target mimics for microRNAs in plants. Plant Physiology
161: 18751884.
Wykoff DD, Grossman AR, Weeks DP, Usuda H, Shimogawara K. 1999. Psr1,
a nuclear localized protein that regulates phosphorus metabolism in
Chlamydomonas. Proceedings of the National Academy of Sciences, USA 96:
1533615341.
Yang Z-K, Niu Y-F, Ma Y-H, Xue J, Zhang M-H, Yang W-D, Liu J-S, Lu S-H,
Guan Y, Li H-Y. 2013. Molecular and cellular mechanisms of neutral lipid
accumulation in diatom following nitrogen deprivation. Biotechnology for
Biofuels 6: 67.
Yang Z-K, Zheng J-W, Niu Y-F, Yang W-D, Liu J-S, Li H-Y. 2014. Systemslevel analysis of the metabolic responses of the diatom Phaeodactylum
tricornutum to phosphorus stress. Environmental Microbiology 16: 17931807.
Zuker M. 2003. Mfold web server for nucleic acid folding and hybridization
prediction. Nucleic Acids Research 31: 34063415.

Supporting Information
Additional supporting information may be found in the online
version of this article.
Fig. S1 Phosphate content in the growth medium of
Phaeodactylum tricornutum on the sampling points for strandspecific RNA sequencing (ssRNA-Seq).
Fig. S2 Growth curves of Phaeodactylum tricornutum cell cultures
under different culture conditions.
Fig. S3 Validation of strand-specific RNA sequencing (ssRNASeq) expression fold changes by quantitative reverse transcription

New Phytologist (2015)


www.newphytologist.com

14 Research

polymerase chain reaction (RT-qPCR) with regard to


Phaeodactylum tricornutum randomly selected protein coding and
long intergenic nonprotein coding RNA (lincRNA) genes
detected as being upregulated during inorganic phosphate (Pi)
depletion by RNA-Seq (more than two-fold, P < 0.05).
Fig. S4 Sequence analysis of Phaeodactylum tricornutum long
intergenic nonprotein coding RNAs (lincRNAs) and microRNAs (mRNAs).
Fig. S5 Distribution of predicted folding energies of transcripts
of different lengths across long intergenic nonprotein coding
RNAs (lincRNAs) and micro-RNAs (mRNAs) in Phaeodactylum
tricornutum.
Fig. S6 Correlation between the GC content and the predicted folding energies of transcripts in different length categories across long intergenic nonprotein coding RNAs
(lincRNAs) and micro-RNAs (mRNAs) in Phaeodactylum
tricornutum.
Fig. S7 Validation of RNA sequencing (RNA-Seq) data by quantitative reverse transcription polymerase chain reaction (RTqPCR) of the two transiently expressed heat shock factors (HSFs)
during inorganic phosphate (Pi) depletion in Phaeodactylum
tricornutum.
Table S1 Number of reads and mapped reads obtained by
strand-specific RNA sequencing (ssRNA-Seq) for the five physiological sampling points of Phaeodactylum tricornutum in response
to inorganic phosphate (Pi) fluctuations

New Phytologist (2015)


www.newphytologist.com

New
Phytologist
Table S2 Primers designed for quantitative reverse transcription
polymerase chain reaction (RT-qPCR) analysis in Phaeodactylum
tricornutum
Table S3 Expression variation under phosphate fluctuations of
Phaeodactylum tricornutum micro-RNAs (miRNAs)
Table S4 Expression variation under phosphate fluctuations of
Phaeodactylum tricornutum long intergenic nonprotein coding
RNA (lincRNA) candidates
Table S5 Putative cis-targets of phosphate depletion-responsive
long intergenic nonprotein coding RNAs (lincRNAs) in
Phaeodactylum tricornutum
Table S6 Expression variation of Phaeodactylum tricornutum
transcription factors under phosphate fluctuations
Table S7 Correlation networks between HSF3.1b and HSF4.7a
and their putative target genes under early phosphate depletion
in Phaeodactylum tricornutum
Table S8 Expression variation of Phaeodactylum tricornutum
sensing- and signaling-related genes under phosphate fluctuations
Please note: Wiley Blackwell are not responsible for the content
or functionality of any supporting information supplied by the
authors. Any queries (other than missing material) should be
directed to the New Phytologist Central Office.

2015 The Authors


New Phytologist 2015 New Phytologist Trust

Potrebbero piacerti anche