This article is a Plant Cell Advance Online Publication. The date of its first appearance online is the official date of publication. The article has been
edited and the authors have corrected proofs, but minor changes could be made before the final version is published. Posting this version online
reduces the time to publication by several weeks.
LARGE SCALE BIOLOGY ARTICLE
High-Resolution Profiling of a Synchronized Diurnal
Transcriptome from Chlamydomonas reinhardtii Reveals
Continuous Cell and Metabolic Differentiation
OPEN
James Matt Zones,a,b,1 Ian K. Blaby,c,1,2 Sabeeha S. Merchant,c,d and James G. Umena,3
a Donald
Danforth Plant Science Center, St. Louis, Missouri 63132
of Biological Sciences, University of California San Diego, La Jolla, California 92093
c Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90095
d Institute of Genomics and Proteomics, University of California, Los Angeles, California 90095
b Division
ORCID IDs: 0000-0001-8346-7869 (J.M.Z.); 0000-002-1631-3154 (I.K.B.); 0002-2594-509X (S.S.M.); 0000-0003-4094-9045 (J.G.U.)
The green alga Chlamydomonas reinhardtii is a useful model organism for investigating diverse biological processes, such as
photosynthesis and chloroplast biogenesis, flagella and basal body structure/function, cell growth and division, and many others.
We combined a highly synchronous photobioreactor culture system with frequent temporal sampling to characterize genome-wide
diurnal gene expression in Chlamydomonas. Over 80% of the measured transcriptome was expressed with strong periodicity,
forming 18 major clusters. Genes associated with complex structures and processes, including cell cycle control, flagella and basal
bodies, ribosome biogenesis, and energy metabolism, all had distinct signatures of coexpression with strong predictive value for
assigning and temporally ordering function. Importantly, the frequent sampling regime allowed us to discern meaningful fine-scale
phase differences between and within subgroups of genes and enabled the identification of a transiently expressed cluster of light
stress genes. Coexpression was further used both as a data-mining tool to classify and/or validate genes from other data sets
related to the cell cycle and to flagella and basal bodies and to assign isoforms of duplicated enzymes to their cognate pathways of
central carbon metabolism. Our diurnal coexpression data capture functional relationships established by dozens of prior studies
and are a valuable new resource for investigating a variety of biological processes in Chlamydomonas and other eukaryotes.
INTRODUCTION
Daily rhythms of biological activities are ubiquitous. For photosynthetic organisms, light and dark cycles are the primary drivers of
both metabolism and endogenous circadian clocks that allow
organisms to anticipate and adapt to alternating light and dark
intervals. Diurnal regulation in plants and algae occurs at multiple
levels, including transcription, translation, and posttranslation
(Thines and Harmon, 2011; Kinmonth-Schultz et al., 2013; Reddy
and Rey, 2014), and a large component appears to be transcriptional as evidenced by transcriptome studies. For example, in the
model plant Arabidopsis thaliana, 30 to 50% of genes show cyclic
diurnal expression patterns under specific conditions of constant
temperature (Covington et al., 2008; Michael et al., 2008). Studies on
whole plants may be limited by tissue heterogeneity that precludes
analyses of cell-type-specific processes unless they are synchronous across most or all tissues (Endo et al., 2014). Unicellular
1 These
authors contributed equally to this work.
address: Biology Department, Brookhaven National Laboratory, Upton, New York 11973.
3 Address correspondence to [email protected].
The author responsible for distribution of materials integral to the findings
presented in this article in accordance with the policy described in the
Instructions for Authors (www.plantcell.org) is: James G. Umen (jumen@
danforthcenter.org).
OPEN
Articles can be viewed online without a subscription.
www.plantcell.org/cgi/doi/10.1105/tpc.15.00498
2 Current
algae offer a complementary approach to understanding diurnal
transcriptome regulation as samples contain material from only
a single cell type, and mitotic cycles are often amenable to diurnally
induced synchronization (Noordally and Millar, 2015). In cyanobacteria, diatoms, and the Prasinophyte alga Ostreococcus tauri,
up to 80% of genes are periodically expressed in diurnal conditions
(Monnier et al., 2010; Johnson et al., 2011; Ashworth et al., 2013). In
the red alga Cyanidioschyzon merolae, a much smaller fraction of
genes (;7%) showed diurnal expression periodicity (Kanesaki
et al., 2012). These studies revealed a significant temporal segregation of diverse processes, such as photosynthetic metabolism, cell division, and cell behavior (Noordally and Millar, 2015).
For studies of rhythmic or periodic gene expression patterns, two
major factors limit the quality of data. The first is sampling density,
which determines the temporal resolution of differential gene expression phasing. The second limiting factor is the degree of culture
synchrony that influences the coherence of gene expression patterns related to specific processes such as cell division. While some
diurnal nuclear transcriptome data have been generated for
Chlamydomonas (Kucho et al., 2005; Panchy et al., 2014), to date
there are no deep-sequencing transcriptome studies of Chlamydomonas or other photosynthetic organisms that have combined
high-frequency sampling in cultures in which cell synchrony of the
entire population of cells was measured and optimized.
Research using the model green alga Chlamydomonas reinhardtii (herein referred to as Chlamydomonas) has advanced the
understanding of many biological processes, including flagella
The Plant Cell Preview, www.aspb.org ã 2015 American Society of Plant Biologists. All rights reserved.
1 of 27
2 of 27
The Plant Cell
and basal body structure and function, chloroplast biogenesis
and photosynthesis, metabolism, DNA methylation, and circadian
rhythms (Harris, 2001). Development of Chlamydomonas as
a model organism has been further accelerated by the availability
of molecular genetic tools and a sequenced genome (Merchant
et al., 2007; Umen and Olson, 2012; Jinkerson and Jonikas, 2015).
Despite its long history as a model organism, the majority of the
predicted proteins in Chlamydomonas have unknown functions,
including many that are conserved in other eukaryotic groups
(Merchant et al., 2007; Karpowicz et al., 2011).
As observed in other photosynthetic species, cell growth, cell
physiology, and mitotic reproduction are strongly influenced by light
and diurnal conditions in Chlamydomonas (Howell and Walker,
1977; Mittag et al., 2005; Matsuo and Ishiura, 2010; Mitchell et al.,
2014; Tirumani et al., 2014; Cross and Umen, 2015). One advantage
of using Chlamydomonas to investigate diurnal or periodic gene
expression is the ability to achieve a high degree of culture growth
and division synchrony under phototrophic growth conditions (Lien
and Knutsen, 1979). Strong diurnal synchrony is a natural outcome of
the multiple fission cell cycle that Chlamydomonas uses for mitotic
reproduction (Bišová and Zachleder, 2014; Cross and Umen, 2015).
In a multiple fission cycle, there is a prolonged G1 phase during which
cells can grow up to 10 or 20 times in mass before dividing. Following
G1, mother cells undergo n rapid, alternating rounds of DNA synthesis and mitosis (S/M phase) to produce 2n uniform-sized daughter
cells. Typically, 8 or 16 daughter cells are formed in rapidly growing
cultures. In a 12-h-light/12-h-dark regime, cell growth occurs during
the day and division occurs at the light-dark transition or early in the
dark phase, after which daughters are released and begin growing
again during the subsequent light phase.
Here, we have taken advantage of diurnal synchronization in
Chlamydomonas to investigate its transcriptome with 1- or 0.5-h
temporal resolution across the day-night cycle. The combination of
frequent time-point sampling and a high degree of culture synchrony allowed us not only to determine broad patterns of gene
expression but also to characterize in detail the temporal ordering of
important biological processes as they relate to gene expression
patterning, including cell cycle progression, basal body and flagella
assembly, ribosome biogenesis, photosynthetic complex formation, and central carbon metabolism. Expression clustering was also
employed as a functional classification tool to make and/or validate
predictions about protein localization and function. The data set we
produced is a powerful new discovery tool and resource for understanding fine-scale temporal patterning of gene expression that
occurs during the Chlamydomonas diurnal cycle and cell cycles. It
will also serve as a reference data set for comparative studies of
coregulated gene expression in algae and other eukaryotes.
RESULTS
The Majority of the Chlamydomonas Transcriptome Is
Differentially Expressed in Synchronous Diurnal
Growth Conditions
We optimized growth conditions and culture synchrony in order
to identify genes that exhibit periodic diurnal and/or cell cycleregulated expression patterns. Phototrophic cultures were grown
in a 12-h-light/12-h-dark diurnal cycle under turbidostatic control
to maintain uniform illumination, biomass density, and nutrient
levels during the light phase, during which individual cells grew in
mass by ;10-fold. Immediately following the dark transition the
cells divided synchronously on average three or four times to
produce 8 or 16 daughters. Cultures were independently sampled
over two 24-h periods at 1-h intervals beginning at ZT1 (Zeitgeber
time = 1 h following the onset of illumination). Additional samples
were taken at 30-min intervals between ZT11 and ZT15 during S/M
phase, for a total of 28 samples per time course, ending at ZT24
(Figure 1A). Culture synchrony was assessed by comparing cell
growth in replicate experiments (Figure 1B) and by monitoring
progression through two cell cycle transitions: Commitment
(reached after the minimum amount of growth necessary to
complete one division cycle) and S/M (Figure 1C). The average cell
density during the light phase decreased as the average cell size
increased so that uniform biomass density and illumination were
maintained throughout the growth phase (Figure 1B).
RNA was prepared from the samples and processed for highthroughput transcriptome sequencing to generate more than
three million uniquely mapping reads per sample (Supplemental
Table 1). The subset of genes that showed the most variable
expression levels over each diurnal cycle (coefficient of variation $
1.2) was used to calculate Pearson correlation coefficients of all
pairwise combinations between replicate samples (Figure 1D).
Replicates were highly correlated (R $ 0.97) and could be subdivided into three major groups based on intersample correlation
values: (1) light phase (ZT2-ZT12), (2) light-dark transition (ZT12.5ZT15, and (3) dark phase (ZT16-ZT1) (Figure 1D).
Of the 17,737 predicted genes in Chlamydomonas (Merchant
et al., 2007; Blaby et al., 2014), 15,325 had read abundances
above a minimal cutoff of 1.061 RPM (reads per million mapped),
which was chosen as an inclusion threshold based on reliability for
detecting differential expression (see Methods). Of these, 12,592
showed at least a 2-fold change from their mean value in one or
more time points across the diurnal cycle with a false discovery
threshold of 0.05 and a maximum abundance >1 RPKM (reads per
kilobase per million mapped). These genes, representing >81%
of the measurable transcriptome, were grouped into 18 major
clusters (c1 to c18) based on similarity of expression patterns
identified by the K-means algorithm (Soukas et al., 2000) implemented in the MeV software package (Saeed et al. 2003; see
Methods), with each cluster containing 250 to 1071 genes (Figure
2A; Supplemental Data Set 1).
We also identified genes with the least variable expression to
find those that might be useful for normalization controls. A total of
171 genes had profiles with <2-fold expression variation between
maximum and minimum RPKM (Supplemental Data Set 2) and
were further examined in 15 other published RNA-Seq experiments (González-Ballester et al., 2010; Castruita et al., 2011;
Fang et al., 2012; Urzica et al., 2012b; Malasarn et al., 2013;
Goodenough et al., 2014; Schmollinger et al., 2014) (Mn- and
Cd- experimental data; S.S. Merchant, unpublished data). We found
24 genes in our low-variance set with maximum fold expression
changes of <2.5 and seven genes with fold expression changes
of <2 in all other experiments. These stably expressed genes are
predicted to be reliable internal controls for gene expression
measurements under diverse conditions.
Chlamydomonas Diurnal Transcriptome
3 of 27
Figure 1. Experimental Design and Culture Metrics.
(A) Clock diagram showing experimental design. Diagram shows ZT time of light (ZT0 to ZT12) and dark (ZT12 to ZT24/0) periods of synchronized
Chlamydomonas culture with growth (G1), division (S/M), and resting (G0) stages marked by inner gray arrows. Cartoons show relative cell sizes and two
successive divisions by multiple fission. Cultures were sampled hourly with additional samples taken at 0.5-h intervals during cell division (brown lines).
(B) Cell size distributions from replicate cultures taken at 3-h intervals during the light phase.
(C) Plots showing fraction of each replicate culture that had passed Commitment and mitotic index at each time point.
(D) Heat map depicting correlation between replicates for the most variably expressed genes (coefficient of variation $ 1.2) during the diurnal cycle.
We compared our clustering method to another algorithm, JTK
(Jonckheere-Terpstra-Kendall) cycle, which was designed for
detecting rhythmic expression from transcriptome data (Hughes
et al., 2010). JTK cycle identified 12,534 rhythmic genes with
a stringent false discovery threshold (1E-10), 11,142 of which
overlapped with our periodic expression data set (Supplemental
Figure 1A). In addition, the phasing (i.e., LAG) assignments from
JTK cycle were well correlated with specific diurnal expression
clusters, meaning that, overall, both methods identified similar
periodic expression patterns (Supplemental Figure 1B). However,
we also noted a sizeable fraction of genes with either complex or
transient expression profiles that were either missed by JTK cycle
or given low-confidence scores despite their high reproducibility
between replicates and high expression amplitudes (maximum
expression/mean expression) (Supplemental Figure 1C). Conversely, the genes identified by JTK cycle that were not included in
our differential diurnal set were almost entirely those that cycle
reproducibly, but with low amplitude (Supplemental Figure 1D).
4 of 27
The Plant Cell
Figure 2. Coexpression and Functional Annotation of Genes with Diurnally Cycling Expression.
(A) Expression heat map of significantly differentially expressed genes. Expression of 12,592 genes during the diurnal cycle in 18 clusters (c1 to c18) is
plotted as the number of standard deviations (s) from the mean. Relative time, light and dark phases, and cell cycle stages are indicated above (Supplemental
Data Set 1).
(B) MapMan term enrichment for clusters c1 to c18. Data are plotted as z-scores with high scores indicating significant enrichment. Level 1 and level 2
classifications separated by a comma are shown for selected terms from Supplemental Data Set 3.
We therefore used our original 18 coexpression clusters as
a primary basis for further evaluation.
Functional Annotation of Expression Clusters
To detect functional specialization within the 18 expression clusters,
we tested for enrichment of MapMan annotation/ontology terms
(Thimm et al., 2004). Term enrichments found in different clusters
included protein synthesis (c2, c4, c5, and c12) photosynthesis (c4,
c6, c5, and c7), cell division (c10), DNA synthesis and chromatin
structure (c1, c9, and c10), abiotic stress (c2 and c7), and cell motility
(c11 to c13) (Figure 2B; Supplemental Data Set 3). Many of these
ontologies were pursued further using manually curated sets of
genes that provided more accurate and comprehensive bases for
analysis of coexpression.
Elucidation of Coordinate Expression Patterns and Phase
Relationships between Cell Cycle and Chloroplast
Division Genes
We extended previous studies of cell cycle gene expression
(Bisová et al., 2005; Fang et al., 2006) by first performing comprehensive annotation for 108 Chlamydomonas genes predicted
to be involved in cell cycle regulation, DNA replication, chromosome segregation, and chloroplast division (Supplemental Data
Set 4). Entry into S/M phase began just after ZT11 and was largely
complete by ZT15 (Figure 1C). Strikingly, nearly every one of the
cell cycle-related genes showed coordinated upregulation and
high peak amplitude just prior to or during S/M phase, with a strong
enrichment of replication genes in cluster c10 (Figures 2B and 3A
to 3D; Supplemental Figures 4 and 5 and Supplemental Data Set 4).
Although individual mother cells do not cycle through each round of
S phase and mitosis in synchrony (Fang et al., 2006), phasing of cell
cycle gene expression is informative because the first genes
expressed will correspond to the earliest cell cycle events (e.g.,
S phase), while genes that are expressed later will correspond
to subsequent events (e.g., mitosis or chloroplast division).
Forty-six DNA replication-related genes showed coordinated
expression, peaking at approximately ZT11 near the beginning
of S/M, and this coordinated expression serves as a useful
temporal landmark for other cell cycle events (Figures 3A to 3D;
Supplemental Figures 4 to 6). Cyclin-dependent kinases CDKA1
and CDKB1 are homologs of universally conserved or green
lineage-specific cell cycle regulators, respectively (Robbens et al.,
2005; Bisová et al., 2005; De Veylder et al., 2007; Cross and Umen,
2015). In Chlamydomonas, CDKA1 function is required to initiate
DNA replication, whereas CDKB1 function is required later for
mitotic progression (Tulin and Cross, 2014). CDKA1 expression
peaked at the time of DNA replication but its lowest expression
during G1 phase never dropped below 20% of its peak value
(Figure 3E), consistent with CDKA1 protein being detectable in G1
phase cells (Oldenhof et al., 2004). In contrast, CDKB1 expression
Chlamydomonas Diurnal Transcriptome
Figure 3. Coordinate Expression and Phasing of Cell Cycle Genes.
5 of 27
6 of 27
The Plant Cell
is near zero when cells are not dividing and peaked sharply at
ZT12, an hour after CDKA1 (Figures 3E; Supplemental Figure 7B).
Two putative mitotic/S phase cyclin genes, CYCA1 and CYCB1,
were expressed maximally at ZT13, just after CDKB1, while the
hybrid A-B cyclin gene CYCAB1 was expressed maximally at ZT9,
2 h prior to visible signs of cell division (Figures 3E; Supplemental
Figure 7B). These data suggest that CYCAB1 acts as an early
activator of cell division and/or S phase, while CYCA1 and CYCB1
act later as mitotic cyclins. Other conserved genes with predicted
roles during mitosis (WEE1, ESP1, and CKS1) were expressed in
a pattern similar to CDKB1 (Figures 3F; Supplemental Figure 7C
and Supplemental Data Set 4).
The anaphase promoting complex/cyclosome (APC/C) is an E3
ubiquitin ligase that triggers mitotic exit and also serves to
maintain low CDK activity in G1 phase (Morgan, 2007; Chang and
Barford, 2014). Genes encoding APC/C subunits were expressed
in a broad peak from ZT10 to ZT13 but were also expressed
detectably in G1 phase (Figures 3G; Supplemental Figure 9A) as
was the gene encoding the G1 phase APC/C activator CDH1. In
contrast, the gene for the predicted mitotic activator of APC/C,
CDC20, was expressed in a sharp peak around ZT13 with little or
no expression outside of S/M phase (Figures 3G; Supplemental
Figure 9B). These data match key findings in other species where
there is a highly specific mitotic requirement for APCCDC20 and
a broader requirement for APCCDH1 upon mitotic exit and during
G1 phase (Zachariae and Nasmyth, 1999).
Proteins of the retinoblastoma (RB)-related protein complex in
Chlamydomonas are present throughout the cell cycle where they
are required for cellular size control at Commitment and to regulate
cell division number during S/M phase (Umen and Goodenough,
2001; Bisová et al., 2005; Fang et al., 2006; Olson et al., 2010).
mRNAs for each subunit of the complex (MAT3 [RBR], E2F1, and
DP1) are detectable in all samples with peak expression for all three
occurring relatively early in the cell cycle at ZT9, consistent with
a role in initiation of S phase and/or earlier events such as Commitment (Figures 3H; Supplemental Figure 7A). Interestingly, a gene
encoding a E2F-related protein, E2FR1, which has degenerate DNA
binding and dimerization domains (Bisová et al., 2005), is expressed
prior to the genes encoding the core RB complex, suggesting that
this protein has a function that is distinct from that of the canonical
RB complex (Figures 3H; Supplemental Figure 7A).
Besides recapitulation of known or anticipated regulatory
patterns, our data provide a means for classifying groups of cell
cycle regulators that have been less extensively investigated in
Chlamydomonas. Whereas most Chlamydomonas cell cycle
genes are present in a single copy (Bisová et al., 2005), the D-cyclin
family has four paralogs (encoded by CYCD1-CYCD4) whose
origins are at least as old as the split between Chlamydomonas
and its multicellular relative Volvox carteri (Prochnik et al., 2010).
Each D cyclin has a distinct expression profile with peaks at ZT8
(CYCD4), ZT10-ZT11 (CYCD2), ZT12 (CYCD1), and ZT15 (CYCD3)
(Figure 3I; Supplemental Figure 7D). These divergent expression
patterns point toward D-cyclin subfunctionalization associated
with their expression peak phasing: CYCD4 is a candidate regulator of and/or marker for Commitment; CYCD2 is a candidate
S phase activator; CYCD1 and CYCD3 are candidate mitotic
activators; and the prolonged expression of CYCD3 and CYCD2
also suggests a postmitotic function for these genes.
Unlike land plants that have many individual chloroplasts per cell,
Chlamydomonas and most other unicellular Chlorophyte algae
have a single large chloroplast that must be physically partitioned
during cell division, a process that occurs before cytokinesis
(Goodenough, 1970; Gaffal et al., 1995) (Figure 4). Several genes
predicted to encode chloroplast division proteins in Chlamydomonas were previously shown to be expressed during S/M (Wang
et al., 2003; Adams et al., 2008; Hu et al., 2008; Miyagishima et al.,
2012) but were not sampled at high temporal resolution. We found
that known chloroplast division genes peak at ZT11-ZT12, consistent with chloroplast division preceding mitosis and cytokinesis,
whose cognate genes are expressed at ZT13 (Figures 3J and 4;
Supplemental Figure 10). A number of uncharacterized genes with
predicted chloroplast-targeting sequences (364 genes) and/or
membership in the GreenCut group encoding conserved plant/algal
proteins (55 genes) (Karpowicz et al., 2011) were coexpressed with
the known chloroplast division genes in clusters c10 or c11 and are
candidates for direct participation in chloroplast division or other
chloroplast-related processes that are coordinated with the cell
cycle (Supplemental Data Set 4).
c10 membership was significantly overrepresented (z-score =
26) within our curated group of cell cycle-related genes identified
primarily through homology searches (Figure 2B; Supplemental
Figure 11 and Supplemental Data Sets 3 and 4). We asked whether
membership in c10 also correlates with cell cycle function in genes
identified in an unbiased genetic screen for temperature-sensitive
lethal cell cycle mutants (Tulin and Cross, 2014). We found that
genes corresponding to the DIV class of mutants, whose defect
Figure 3. (continued).
Plots of relative expression during light and dark phases. Phases are indicated by white and gray shading with expression levels normalized to a maximum of
1. Replica-averaged expression profiles of individual genes (italicized labels) or averages for all members of functional groups (nonitalicized labels) within
each category are shown. Plots were smoothed as described in Methods. Additional expression data for cell cycle genes are in Supplemental Figures 4 to 11
and Supplemental Data Set 4.
(A) to (C) DNA replication-related genes and protein complexes.
(D) Structural maintenance of chromosomes (SMC) complexes.
(E) Predicted S phase and mitotic cyclins and CDKs.
(F) Additional cell cycle regulators.
(G) APC/C genes.
(H) RB tumor suppressor pathway genes.
(I) D-type cyclin genes.
(J) Chloroplast division genes.
Chlamydomonas Diurnal Transcriptome
7 of 27
Figure 4. Summary of Cell Cycle Gene Expression.
Graphical representation of expression phasing for genes or gene groups related to cell division and cell cycle control (ZT11 to ZT14) with darkest colored
shading indicating peak expression time. The lower cartoon depicts key stages of division and the relative timing of events within a single division cycle. The
backward arrow indicates optional iterations of S phase and mitosis. Additional expression data for cell cycle genes are in Supplemental Figures 4 to 11 and
Supplemental Data Set 4.
was in mitotic progression, were highly enriched for membership in c10 (z-score = 13) (Supplemental Figures 11 and 12 and
Supplemental Data Set 5). In contrast, genes for a second class of
mutants called GEX that showed a failure to exit G1 phase were
found in several clusters and only showed significant enrichment
in c1 (z-score = 3.76), which contains a large number of stressrelated and chromatin-related genes (Figure 2B; Supplemental
Figures 11 and 12 and Supplemental Data Sets 3 and 5). We
predict from these data that DIV genes will likely have functions
directly associated with cell cycle progression (e.g., DNA replication, mitosis, and cytokinesis), whereas GEX genes represent
a more pleiotropic set of functions than DIV genes, and their further
study may reveal new connections between the cell cycle and
stress pathways and/or chromatin dynamics.
Expression of Genes Encoding Flagella and Basal Body
Proteins Is Coordinated with the Cell Division Cycle
The flagella of Chlamydomonas are motility and sensory organelles that are homologous in structure and function to animal cilia
and to the cilia or flagella found in other eukaryotic taxa (CarvalhoSantos et al., 2011; Ostrowski et al., 2011). During the cell cycle,
flagella/cilia are resorbed or severed prior to S phase or mitosis,
thereby freeing basal bodies to act as centrioles during mitosis and
cytokinesis (Johnson and Porter, 1968; Coss, 1974; Plotnikova
et al., 2009; Parker et al., 2010; Kobayashi and Dynlacht, 2011).
Flagella are reformed upon mitotic exit (Wood et al., 2012) and also
play a role in daughter cell release as a secretory site for hatching
enzyme (Kubo et al., 2009). In our experiment, flagella resorption
occurred synchronously as cells entered S/M and flagella
reformation was scored upon hatching, by which time daughters
already had full length flagella (Supplemental Figure 13).
Although a selected set of genes relating to flagella function
were found to be upregulated just after cell division genes (Wood
et al., 2012), we wanted to conduct a comprehensive analysis of
basal body and flagella gene expression during the cell cycle. We
first curated a core set of 193 basal body and flagella genes (Basal
Body, Axoneme, IFT, and BBsome) that have been identified in
previous studies (Keller et al., 2005; Lechtreck et al., 2009;
Dutcher, 2009; Shiratsuchi et al., 2011; Bower et al., 2013) (Figures
5A to 5C; Supplemental Data Set 6). Clusters c11 to c14 are
significantly enriched for the MapMan ontology term “cell motility”
(z-score > 2.5)(Figure 2B; Supplemental Data Set 3) and contain
an overrepresentation of known flagella and basal body genes
(P < 0.02) (Figures 6A; Supplemental Figure 14). Remarkably, nearly
every flagella and basal body gene exhibited a similar expression
pattern with an abrupt rise in expression around ZT12 when S/M
was ongoing, peak expression around ZT13-ZT14, and tapering
expression during the remainder of the dark period (ZT15 to ZT24)
(Figures 5A to 5C; Supplemental Data Set 6). This expression
pattern matches the expected requirement for peak flagella
synthesis in postmitotic daughter cells. Also notable were the
expression patterns of several flagella and basal body genes that
differed from the majority (Figures 5A to 5C; Supplemental Data
Set 6) and correlate with additional functions for the gene products
not involving flagella, e.g., in the actin cytoskeleton (IDA5 encoding actin and PRO1 encoding profilin) and in stress responses
(HSP70A, DNAJ1, and HSP90A) (Schroda and Vallon, 2009).
Therefore, our data suggest that additional basal body/flagella
genes that are not part of clusters c11 to c14 (e.g., POC13/FMO11,
8 of 27
The Plant Cell
Figure 5. Flagella and Basal Body Gene Expression.
(A) to (C) Heat maps depict relative expression levels for flagella and basal body genes that are divided into functional subgroups. The maximum expression
level for each gene is set to 1. Cell cycle stages, diurnal cycle time, and light and dark phases are indicated above each heat map. Individual gene names are
on the right. Cluster membership of genes is shown to the left (gray bars); brown stars indicate nondifferentially expressed genes.
(A) Flagella axoneme genes and subgroups.
(B) Core basal body genes, POC genes, and BUG genes (described in the main text).
(C) Intraflagellar transport (IFT) and Bardet-Biedl syndrome protein complex (BBSome) genes. IFT genes are further divided into Complex A subunits,
Complex B subunits, anterograde motor subunits, and retrograde motor subunits.
Chlamydomonas Diurnal Transcriptome
POC17/PHB1, and CCT3)(Figure 5; Supplemental Data Set 6) may
have additional functions outside of these organelles and that
coexpression is a useful discriminatory filter for identifying such
genes.
Among the basal body genes (Figure 5B; Supplemental Data
Set 6), we noted significant differences in expression patterns and
cluster membership among three subgroups: core basal body
(those with validated functions), POC (proteome of centriole), and
BUG (basal body upregulated after deflagellation): The latter two
groups were identified in a basal body proteomic study, with POC
genes validated by homology with centriolar proteins of other
species and BUG genes validated by upregulated expression after
deflagellation (Keller et al., 2005). We observed that core basal body
and axonemal gene groups had distinct distributions among c11 to
c14 (P = 6.2 E-09, Fisher’s exact test), with core basal body genes
predominantly found in c11 (peak expression at ZT12) and axonemal genes predominantly found in c12 or c13 (peak expression at
ZT13 and ZT14) (Figures 5A and 5B; Supplemental Figures 14D and
14E). POC gene cluster membership was not detectably different
from core basal body gene cluster membership (P = 0.76, Fisher’s
exact test), whereas BUG gene cluster membership was different
from both axonemal and core basal body genes, but more similar to
the axonemal cluster distribution (P = 0.026, Fisher’s exact test) than
to the basal body cluster distribution (P = 8.2 E-05, Fisher’s exact
test) (Supplemental Figures 14D and 14E). These data are likely to
reflect functional differences and temporal ordering between
the core basal body and POC genes that are required during S/M
for basal body assembly, replication, and mitosis, and the BUG
genes that we predict will encode basal body proteins related to
postmitotic nucleation and assembly of flagella (e.g., transition zone
proteins). Indeed, many of the BUG genes are also present in the
flagella proteome, whereas no POC genes are in this subset
(Supplemental Data Set 7). Although the specific functions of
most POC genes are unknown, the POC3/CEP290 gene product
localizes mainly to the transition zone and is required for flagella
assembly, and therefore does not match our prediction for involvement in basal body assembly or replication (Craige et al., 2010).
Three additional large data sets of candidate flagella/basal body
protein coding genes were investigated based on their synchronous
expression profiles: genes encoding proteins of the flagella proteome (FAPs) (Pazour et al., 2005), CiliaCut genes that are conserved in ciliated species but missing from species without cilia or
flagella (Merchant et al., 2007), and an RNA-Seq-based transcriptome study of genes upregulated after deflagellation (Albee
et al., 2013). We asked whether our coexpression data would
discriminate between genes encoding proteins with core functions
in flagella or basal bodies versus those with additional functions, or
in the case of the deflagellation experiment, between genes whose
expression is triggered by the stress of deflagellation and genes that
are required for flagella biogenesis. First, we found that the uncharacterized members for all three large data sets described above
(FAP, CiliaCut, and deflagellation upregulated) are significantly
9 of 27
enriched for membership in c11 to c14 (Figure 6A; Supplemental
Figure 14C) and predict that the c11 to c14 subset of genes from
these data sets has functions directly related to flagella biogenesis.
We extended and validated this observation by examining
comembership of genes in the three data sets (FAP, CiliaCut, and
deflagellation upregulated) with the assumption that genes found
in two or three of the groups are more likely tied to flagellum/basal
body function than genes found in only one of the groups. Indeed,
we found that members of c11 to c14 were highly overrepresented
in the set of genes that belong to more than one data set (P = 1.5
E-27), while genes in other clusters were underrepresented in the
overlap set (P = 3.0 E-27) (Figure 6B; Supplemental Figure 15B).
In summary, our data demonstrate robust, coordinated expression
timing of basal body and flagella genes that is likely coupled to cell
cycle progression and can be used as a powerful filter for identifying
and classifying proteins that function in these organelles.
Ribosomal Protein Genes Are Expressed in
Compartment-Specific Temporal Patterns
Green organisms must coordinate the activity of three separate
protein translation systems located in the cytosol, chloroplast,
and mitochondria respectively. rRNAs are encoded by the nucleus and respective organellar genomes, but ribosomal protein
genes (RPGs) for all three compartments are nucleus-encoded. In
principle, RPG expression for the three compartments could be
globally coordinated and directly reflects patterns of diurnal cell
growth, but instead we found that RPGs were expressed in distinct
compartment-specific patterns (Figure 7; Supplemental Data Set
8). Cytosolic RPGs were expressed at high levels throughout the
diurnal cycle used in this experiment and as a group constitute
between 15 and 45% of all unique protein coding gene transcripts
at any given time (Supplemental Figure 16). While cytosolic RPGs
almost never dropped below half their peak expression values,
they showed a consistent increase during the dark period with
peak expression at around ZT15 when cells exited S/M and new
daughters were hatching (Figures 7A and 7D). Notably, this was
not a time when cells were actively growing. Mitochondrial RPGs
showed an almost mirror-image pattern compared with cytosolic
RPGs with highest expression in the light and reduced expression
in the dark when respiration is most active (Figures 7C and 7D).
Most strikingly, chloroplast RPGs showed a very strong periodic
expression pattern, with a sharp peak early in the light period at
ZT2 followed by a continuous decline through the remainder of the
light period and partial recovery during the dark period (Figures 7B
and 7D). These expression patterns suggest an unexpectedly
strong partitioning of ribosome biogenesis among the three
compartments, and they underscore the unanticipated coupling
of ribosome biogenesis to organelle-specific requirements rather
than cell growth as a whole.
Outside of the general patterns described above, we found that
several RPGs that encode plastid-specific ribosomal proteins (i.e.,
Figure 5. (continued).
(D) Diagram of flagella and basal body associated structures at anterior of cell with expanded cross section of an axoneme. Additional data on flagella and
basal body genes are in Supplemental Figures 13 and 14 and in Supplemental Data Sets 6 and 7.
10 of 27
The Plant Cell
PSRPs; Zerges and Hauser, 2009) are expressed out of phase with
the majority of chloroplast RPGs. These include PSRP-1, whose
expression peaks sharply and transiently at ZT13 but has very little
expression at ZT2 during peak expression for most chloroplast
RPGs. PSRP-1 is a homolog of the cyanobacterial gene lrt, whose
transcript also accumulates after the light-to-dark transition. lrt is
thought to stabilize ribosomal subunit association, possibly as
a mechanism for stress-regulated translational control (Tan et al.,
1994; Samartzidou and Widger, 1998), and the identification of
the same pattern in Chlamydomonas suggests a deeply rooted
common function for PSRP-1 in the green lineage. RAP38 and
RAP41 are related proteins in the GreenCut group of conserved
green lineage proteins (Karpowicz et al., 2011) that purify as
stoichiometric components of intact Chlamydomonas 70S
chloroplast ribosomes (Yamaguchi et al., 2003) and whose genes
have nonoverlapping diurnal expression patterns (Figure 7B).
Mutants of the Arabidopsis homologs of RAP38 and RAP41
(CSP41a and CSP41b, respectively), have defective rRNA processing and altered polysome profiles, indicating a role for this
conserved pair of proteins in rRNA processing and ribosome
biogenesis (Beligni and Mayfield, 2008). The noncanonical expression patterns we identified for RAP38 and RAP41 further
suggest that they may function differently than core ribosomal
proteins and possibly from each other.
In addition to RPGs, we examined expression patterns for other
translational machinery and for nucleus-encoded regulators of
chloroplast gene expression (Zerges and Hauser, 2009), some of
which also showed atypical expression compared with chloroplast RPGs and suggest dynamic changes in chloroplast translation profiles at specific times of the diurnal and/or growth and
division cycles (Figure 7B; Supplemental Data Set 8).
In summary, our data support a highly orchestrated pattern of
ribosome assembly and compartment-specific translational control. Coexpression profiling further organizes translation-related
genes into specific expression subgroups that will provide
guidance for future studies of protein biosynthesis.
Plastid and Mitochondrial Protein Complex Subunits Show
Phased Expression Patterns
Metabolism is driven by multisubunit protein complexes in
chloroplasts and mitochondria, which generate ATP and reductant
using either light energy (chloroplasts) or energy from catabolism
(mitochondria). Light-driven reactions in the chloroplast are
Figure 6. Flagella and Basal Body Gene Functional Classification.
(A) Cluster membership distribution of flagella gene groups. Fractional
compositions of the expressed transcriptome and flagella- or basal bodyrelated gene subsets including the known flagella and basal body genes
corresponding to those in Figures 5A to 5C, CiliaCut genes, FAP genes, and
deflagellation response genes. Enrichment for membership of each gene
set in c11 to c14 compared with the entire transcriptome is indicated by
asterisks showing significant P values.
(B) Venn diagrams show overlapping membership among indicated categories of flagella-related genes. Cilia Cut, FAP, deflagellation response genes
belonging to either c11 to c14 (upper diagram) or all other clusters (lower
diagram) were examined. The observed number of genes in a sector is shown
inside the sector in bold with the expected membership number and SD based
on a null model shown below each number in parentheses. Asterisks indicate
sectors whose observed membership deviates significantly from the expected value. Additional data on flagella and basal body genes are in
Supplemental Figures 13 to 15 and in Supplemental Data Sets 6 and 7.
Chlamydomonas Diurnal Transcriptome
11 of 27
Figure 7. Expression Patterns of Genes Encoding Chloroplast-, Mitochondria-, and Cytosol-Targeted Ribosomal Proteins.
(A) to (C) Heat maps depicting relative expression levels of RPGs. Relative expression levels of RPGs and translational regulators targeted to cytosol (A),
chloroplast (B), and mitochondria (C). The maximum expression level for each gene is set to 1. Cell cycle stages, diurnal cycle time, and light and dark
phases are indicated above each heat map. Gene names are on the right with known targets of chloroplast translational regulators in parentheses next to
the corresponding gene name. Brown stars indicate genes without significant differential expression (<2-fold deviation from mean as defined in
Methods).
(D) Graph of averaged compartment-specific RPG expression. Average absolute expression levels of large and small subunit RPGs during the diurnal
cycle for cytosolic (blue), mitochondrial (orange), and chloroplastic (green) subunits. Additional data for ribosomal protein gene expression and
translational regulator gene expression are in Supplemental Figure 16 and Supplemental Data Set 8.
12 of 27
The Plant Cell
mediated by photosystems I and II (PSI and PSII) and associated
light-harvesting complexes (LHCI and LHCII), along with cytochrome b6f and the plastid ATP synthase complex. In mitochondria, the electron transport chain has five major complexes
(I to V), including an ATP synthase complex that is encoded by
a different set of genes than is the chloroplast ATP synthase. We
examined gene expression patterns for the nucleus-encoded
subunits for each of these complexes to determine both whether
they are coordinately expressed and whether there are peak
expression phase differences between them.
Subunits of each photosynthetic complex show broad lightphase expression peaks and are coexpressed in similar patterns,
but they also display reproducible intercomplex phasing differences
(Figures 8; Supplemental Figure 17 and Supplemental Data Set 9).
The earliest expressed genes encode subunits of the ATP synthase
and b6f complexes whose mRNA abundances reach their peaks by
ZT4-ZT6 and whose basal expression is relatively high compared
with genes encoding photoactive complexes (Figure 8). Genes
encoding subunits of PSI, PSII, and LHCI all show near zero expression in the dark and >1000-fold increases in transcript abundance in the light, with peaks at ZT6-ZT8, representing a 2-h phase
delay compared with peak expression of genes encoding cytochrome b6f and ATP synthase subunits. Genes encoding LHCII
subunits showed a further phase delay with a peak from ZT8 to ZT10
(Figure 8; Supplemental Figure 17 and Supplemental Data Set 9).
The staggered phasing we observed with cytochrome b6f and ATP
synthase complexes expressed earliest may enable productive
photochemistry without delay upon PSI and PSII assembly, thereby
minimizing phototoxic side reactions.
Consistent with previous estimates of protein abundance (Oey et al.,
2013; Drop et al., 2014; Natali and Croce, 2015), individual LHC gene
mRNA abundances varied by several orders of magnitude compared
with the peak expression levels of genes for PSI, PSII, b6f, and ATP
synthase complexes, which generally remained within a twofold range
between genes (Figure 8; Supplemental Figure 17 and Supplemental
Data Set 9). We also observed that genes encoding assembly
factors for each photosynthetic complex were either coexpressed
with or expressed just before the genes for their target complexes
(Supplemental Figure 17 and Supplemental Data Set 9).
When grown photoautotrophically in a diurnal cycle, Chlamydomonas cells accumulate starch during the day through photosynthesis and catabolize it at night via glycolysis and respiration (Gfeller
and Gibbs, 1984; Klein, 1987; Thyssen et al., 2001; Ral et al., 2006).
The dark-phase respiratory activity associated with starch catabolism is reflected in the expression patterns of genes encoding subunits of mitochondrial electron transport/oxidative phosphorylation
complexes I to V. Transcripts for these proteins began accumulating
in the mid- to late-light period and peaked in the dark phase (ZT14
to ZT17) (Supplemental Figures 18 to 20 and Supplemental Data
Set 10). Thus, respiratory complex genes are expressed out of phase
with those of photosynthetic complexes and peak during the dark
when respiration is expected to be most active.
Coordination of Tetrapyrrole Pathway Gene Expression with
Light and Dark Phases
Chlorophyll and heme are the two major classes of tetrapyrroles in
photosynthetic cells and both play key roles in energy metabolism.
Genes for the common and chlorophyll-specific tetrapyrrole
biosynthetic enzymes are coexpressed, with transcript levels
increasing rapidly in the light and peaking at ;ZT6. In contrast,
HEM15, which encodes the ferrochelatase enzyme dedicated to
heme biosynthesis, shows increasing expression during the dark,
highest expression at the end of the dark period, and steadily
decreasing expression during the light period when the chlorophyll pathway is most highly expressed (Supplemental Figure 21
and Supplemental Data Set 11). This temporal partitioning between heme and chlorophyll pathway genes may have evolved as
a strategy to reduce competition between the two tetrapyrrole
biosynthetic pathways and to ensure that chlorophyll is the major
tetrapyrrole produced during the light phase (Supplemental Figure
21 and Supplemental Data Set 11).
Transient Stress Responses Occur at the
Dark-to-Light Transition
The abrupt dark-to-light transition in our experiment enabled us to
identify and refine a cluster of 280 genes, mostly derived from c1
and c2, whose expression is transiently induced in a sharp spike
at the first light time point, ZT1 (Figure 9A; Supplemental Data Set
12) (see Methods). Functional annotation of this cluster showed
enrichment for abiotic stress-related genes (Figure 9B) that include five heat shock protein (HSP) genes and a gene whose
Arabidopsis ortholog (At1g04130) encodes a HSP cochaperone
(Cre08.g375650) (Schroda and Vallon, 2009). The cluster also
included VTC2, which encodes GDP-L-galactose phosphorylase,
the enzyme that performs the first committed step in the SmirnoffWheeler pathway for ascorbate biosynthesis (Urzica et al., 2012a)
(Figure 9A). We hypothesize that many other genes in this cluster
are induced by the stress associated with the abrupt transition
from darkness into high-light growth conditions.
LHC-like genes in the ELIP, LHCSR, and PSBS families are also
associated with light stress responses (Heddad and Adamska,
2002; Hutin et al., 2003; Elrad and Grossman, 2004; Bonente et al.,
2008). ELIP1-ELIP5 (encoding homologs of early light inducible
proteins) have previously been described in Chlamydomonas
(Elrad and Grossman, 2004; Teramoto et al., 2004), and, based on
recent genome updates (Blaby et al., 2014), we identified five
additional Chlamydomonas ELIP genes, ELIP6 to ELIP10
(Supplemental Figure 22 and Supplemental Data Set 13). ELIP2,
ELIP3, ELIP4, ELIP9, and ELIP10 as well as one of two PSBS
paralogs, the GreenCut member PSBS2 (Cre01.g016750), are in
the light stress cluster. The other PSBS paralog, PSBS1 (Cre01.
g016600), showed a similar expression pattern as PSBS2
(Supplemental Data Set 1), but its expression maxima was below
the threshold used for inclusion in the cluster. The highly transient
light-stress-activated expression of PSBS genes discovered in
our study may explain their lack of reported expression in previous
experiments and also suggests that, contrary to previous reports
(Anwaruzzaman et al., 2004; Bonente et al., 2008; Peers et al.,
2009), the Chlamydomonas PSBS proteins may play a role in
nonphotochemical quenching (NPQ), the conversion and dissipation of excess excitation energy as heat.
Besides abiotic stress, other functional terms that are enriched
in the stress cluster include ABC transporters, protein and amino
acid biosynthetic genes, and protein translation (Figure 9B;
Chlamydomonas Diurnal Transcriptome
13 of 27
Figure 8. Coexpression and Phasing of Nucleus-Encoded Genes for Photosynthetic Complexes.
(A) to (F) Plots of relative expression of photosynthetic genes. Data are averaged for two replicates during light and dark phases (indicated by shading) with
expression levels normalized to a maximum of 1 corresponding to the peak expression level of each gene.
(A) LHCII.
14 of 27
The Plant Cell
Supplemental Data Set 14). This suggests that a concerted
metabolic rewiring might occur at the dark-to-light transition when
cells must rapidly respond to light stress, while also preparing for
a massive upregulation of their photosynthetic capacity in the
subsequent light period.
Central Carbon Metabolism Genes Show Pathway-Specific
Expression Patterns
Central carbon metabolism in photosynthetic cells must respond
to changes in growth conditions such as activity of photosynthetic
complexes and external nutrients, and these changes can be
reflected by coordinated changes in transcriptional networks for
metabolic genes (Wei et al., 2006). The transcript abundance of
genes encoding enzymes of the Calvin-Benson-Bassham (CBB)
cycle of CO2 fixation exhibited highly correlated expression patterns along with CP12, encoding a chaperone for GAP3, and the
gene for RUBISCO ACTIVASE (Figure 10; Supplemental Data Set
15). Peak expression for nearly all genes encoding enzymes of the
CBB cycle occurred between ZT5 and ZT8. An exception from this
pattern was RBCS2, one of the two Rubisco small subunit paralogs that was expressed continuously in our diurnal growth
conditions. Another exception was TPIC1, the single gene in
Chlamydomonas encoding triose phosphate isomerase, which is
expressed during the light and dark periods, possibly reflecting its
dual roles in the CBB cycle as well as in glycolysis during the dark
phase (Supplemental Data Set 15).
Because our cultures were grown photoautotrophically (i.e.,
without acetate or other organic carbon sources), the patterns we
observed for dark metabolism gene expression reflect a state
where organic carbon is derived solely from endogenous pools
such as starch or storage lipids. Genes encoding enzymes required for the tricarboxylic acid (TCA) cycle and acetate assimilation had less coherent expression patterns than those for the
CBB cycle but were still strongly biased toward peak expression
during the dark period (ZT12 to ZT24) when respiration dominates
energy metabolism (Figure 10; Supplemental Data Set 15). Transcripts for starch biosynthesis enzymes gradually increased during
the light period, whereas the expression of genes encoding starch
catabolic enzymes was restricted to the dark period when starch is
broken down (Supplemental Data Set 15).
Paralog Expression Clustering Predicts Pathway
Assignments for Duplicated Metabolic Enzymes
Several enzymes of central carbon metabolism in Chlamydomonas are encoded as multigene families in which individual
isoforms may be targeted to different subcellular compartments
and are associated with pathways operating at different times
during the diurnal cycle (Supplemental Data Set 15). Experimental localization of biochemical activities has been described for some enzymes (reviewed in Johnson and Alric,
2013), but the pathways within which many enzyme isoforms act
in Chlamydomonas remain unknown. This uncertainty impacts the
accuracy of metabolic models that try to simulate metabolism in
a compartmentalized cell (Dal’Molin et al., 2011).
We used expression profiles of “signature” genes that are dedicated to specific pathways (e.g., sedoheptulose-1,7-bisphosphatase
[SBP1] to the CBB cycle and isocitrate lyase [ICL1] to the
glyoxylate cycle) as a template for correlation measurements
that helped assign individual paralogs of duplicated genes to
their cognate pathways (Supplemental Data Set 15). For example,
citrate synthase activity is required for both the TCA cycle in
mitochondria and for the glyoxylate cycle, whose location is
presumably in microbodies (the algal equivalent of a plant peroxisome) (Hayashi et al., 2015). The expression profiles of two
citrate synthase paralogs encoded by CIS1 and CIS2 were
compared with those of the signature TCA cycle genes (SDH3 and
SDH4) and to signature glyoxylate cycle genes (MAS1 and ICL1)
using Pearson’s correlation. CIS2 expression was strongly correlated with the glyoxylate cycle genes (0.91) and more weakly
with TCA genes (0.54), whereas CIS1 expression was poorly
correlated with glyoxylate genes (20.33) and moderately correlated with TCA genes (0.60) (Supplemental Data Set 15). The
expression profile-based assignment of CIS2 to the glyoxylate
cycle and CIS1 to the TCA cycle was supported and validated
by data from Arabidopsis, in which the most similar homologs of
CIS2 (CSY4 and CSY5) form a clade of peroxisome-targeted
citrate synthases and the most similar homologs of CIS1 (CSY1 to
CSY3) form a clade of mitochondrial-targeted citrate synthases
(Pracharoenwattana et al., 2005). By applying a similar approach
and an expression correlation threshold of $0.6, we propose
pathway assignments for a total of 23 enzyme isoforms
(Supplemental Data Set 15).
Transcription Factor Gene Expression Patterns
We examined the expression profiles of 230 predicted and/or
functionally characterized DNA binding transcription factors (TFs)
(Pérez-Rodríguez et al., 2010; Jin et al., 2014) in our study and
from previously published transcriptome experiments (GonzálezBallester et al., 2010; Castruita et al., 2011; Fang et al., 2012;
Urzica et al., 2012b; Malasarn et al., 2013; Goodenough et al.,
2014; Schmollinger et al., 2014) (Mn- and Cd- experimental data;
Figure 8. (continued).
(B) LHCI.
(C) PSII.
(D) PSI.
(E) b6f complex.
(F) CF1-CFo. Complete protein names encoded by each gene are in Supplemental Data Set 9.
(G) Composite expression patterns for each complex shown in (A) to (F). Data are smoothened as indicated in the Methods.
(H) Schematic showing the subunit architecture of each complex with color corresponding to graph colors gene in (A) to (F). Chloroplast-encoded genes are
shown in pale gray. Additional data for photosynthetic complexes and assembly factors are in Supplemental Data Set 9.
Chlamydomonas Diurnal Transcriptome
15 of 27
Figure 9. Transient Stress Response at Dark-Light Transition.
(A) Transcript abundance for known stress response or ROS response genes (red) and GreenCut genes (green). The mean of all 290 light stress gene
expression profiles comprising the light stress cluster is plotted in black. Light-gray shading indicates the dark period. Complete protein names encoded by
each gene are in Supplemental Data Set 12.
(B) Functional characterization of the light stress cluster. MapMan ontology distributions with percentages of the total are shown in parentheses. Terms
highlighted with a star are significantly enriched (P value < 0.05). In total, the light stress cluster includes 13 GreenCut genes, several of which have unknown
functions (CGL150, CGL122, CPLD50, and CGL99) (Karpowicz et al., 2011; Heinnickel and Grossman, 2013). Additional data on the light stress cluster are in
Supplemental Data Set 12.
S.S. Merchant, unpublished data; Supplemental Data Set 16). We
compared levels of induction, expression amplitudes, and absolute expression maxima as a way of categorizing the regulatory
patterns of Chlamydomonas TFs. The mRNA abundances of 219
TFs were $1 RPKM in at least one time point of the time course,
and of these, 191 showed significant differential expression
patterns (Supplemental Figure 23 and Supplemental Data Set 16).
Eighty-six differentially expressed TFs were not significantly
upregulated in other experiments/conditions and are therefore
possible candidates for controlling diurnal and/or cell cycle related
gene expression. In contrast, 105 TFs showed differential expression in our experiment and significant upregulation in at least
one other experiment, with 28 of these showing greater upregulation in other data sets compared with their peak amplitude in our
data set (max_other/max_diurnal > 2) (Supplemental Data Set 16).
Ten TFs were upregulated following deflagellation (Albee et al.,
2013), with three of these also members of the core flagella clusters
(c11 to c14) and, therefore, candidates for controlling expression of
16 of 27
The Plant Cell
Figure 10. Expression of Central Carbon Metabolism Genes and Their Diurnal Phasing.
Metabolic pathways are shown for the CBB cycle (A), acetate metabolism (B), and the TCA cycle (C). Beside each pathway schematic are average expression
estimates of the two replicates normalized to peak expression and raw RPKM values for each of the two replicates for each gene of the pathway. Gray shading
indicates the dark period. The enzymes catalyzing each step are indicated with primary gene names beside each pathway arrow. Color coding (according to the
heat map) and clock icons indicate the time of peak expression. Black arrows and enzyme names in black are used for steps where different isoforms or enzyme
subunits are not expressed in the same pattern. Additional data on central carbon metabolism genes including protein names are in Supplemental Data Set 15.
flagella-associated genes (Supplemental Figure 23 and
Supplemental Data Set 16). In contrast, the other seven TFs
upregulated in response to deflagellation were also induced in
nutrient-limiting conditions, suggesting that they may be part of
a more general stress response that occurs after deflagellation.
Poorly Expressed Genes and Condition-Specific Expression
The expression estimates for ;2900 predicted genes in our data
did not rise above 1 RPKM at any time point (Supplemental Data
Set 17). It is likely that the products of these genes are required
either at low levels or only under specific conditions that are not
part of our diurnal regime (e.g., sexual reproduction and nutrient
limitation). Alternatively, some among the 2900 poorly expressed
genes may be pseudogenes. We compared expression estimates
for these 2900 genes to those from other published Chlamydomonas transcriptomes (Supplemental Data Set 17) and applied
a minimum expression estimate of 10 RPKM/FPKM as a threshold
for condition-specific expression. In doing so, we identified
a total of 460 genes that were significantly expressed in other
Chlamydomonas Diurnal Transcriptome
17 of 27
transcriptomes and are therefore strong candidates for being
condition-specific genes whose expression is not required for
vegetative proliferation (Supplemental Data Set 17).
DISCUSSION
High-Resolution Transcriptomics Reveal Temporal and
Functional Relationships among Biological Processes
By combining frequent time-point sampling and high biological
synchrony under carefully controlled diurnal growth conditions,
we were able to characterize in detail the transcription program of
the Chlamydomonas vegetative reproductive cycle with very fine
resolution. Using this system, we then investigated diverse biological processes that are diurnally regulated and/or cell cycle
regulated (Figure 11). Not surprisingly, we discovered a greater
extent of periodic gene expression than in previous studies
(Panchy et al., 2014), with ;80% of the detectable transcriptome
showing significant differential expression (Figure 2). Just as
importantly, we were also able to detect fine-scale phasing patterns that are likely to be meaningful indicators of relative timing
among different cellular and metabolic processes or even between sequential events related to the same process. For example, we observed a reproducible ordering of expression of
different cell cycle-related genes that matched their predicted
order of function (Figure 4). Similarly, we observed statistically
significant differences in expression phasing between core basal
body genes and flagella structural genes, a finding that is consistent with the temporal ordering of basal body duplication during
prophase followed by flagella biosynthesis after cells exit from
S/M (Figures 4 and 5) (Wood et al., 2012). Expression of genes for
photosynthetic membrane complexes showed phase lags between complexes that directly participate in photochemistry (PSI,
PSII, LHCI, and LHCII) versus the b6f and ATP synthase complexes
whose genes were transcribed earlier than those for PSI and PSII
(Figure 8). Moreover, frequent temporal sampling allowed us to
discover at least one transient expression cluster containing
stress-related genes with a sharp peak at ZT1 that would have
been difficult to detect with a sparser sampling regime (Figure 9).
Indeed, we observed in this transiently expressed cluster PSBS,
whose protein is implicated in light stress signaling and NPQ in
land plants, but whose function in Chlamydomonas NPQ has been
questioned due to an inability to detect PSBS mRNA under various
growth conditions (Bonente et al., 2008, 2012; Peers et al., 2009).
The relative contributions of different external and internal
regulatory inputs into the periodic expression profiles we observed remain to be determined. These inputs may include light or
dark, cell cycle controls (DNA replication, mitosis, cytokinesis,
basal body replication, flagella resorption, and regrowth), and the
circadian clock, some of which may be overlapping, as has been
observed in Chlamydomonas and other microalgae (Hwang and
Herrin, 1994; Serrano et al., 2009; Moulager et al., 2010; Kanesaki
et al., 2012; Swirsky Whitney et al., 2012). A genome-wide cDNA
microarray-based study in Chlamydomonas found that ;2 to 3%
of genes showed detectable circadian cycling under free-running
clock conditions, but this is only a small fraction of genes that we
found to be periodically expressed under a synchronous diurnal
Figure 11. Summary of Diurnal Gene Expression Patterns in Chlamydomonas.
Clock diagram depicting timing of diurnal expression for groups of genes
associated with indicated processes or structures represented by cartoons. Inner diagram depicts diurnal cycle with light and dark periods and
cell cycle phases. Peak expression timing for each group of genes is shown
as a gradient-filled arc, with higher relative expression represented by
a darker degree of shading and a thicker arc.
cycle (Kucho et al., 2005). Use of circadian (Matsuo et al., 2008) or
cell cycle mutants (Fang et al., 2006; Tulin and Cross, 2014) and
alteration of diurnal conditions may help begin to deconvolute
these contributions and provide a means of distinguishing how
cycling genes are controlled by each input.
Impact of Transcriptome Dynamics on Biological Processes
The mRNA composition of Chlamydomonas cells in our experiment was under constant flux (Figure 2). The widespread finding of
co-expression for genes governing related processes suggests
that most nuclear genes are under selection for coordinated
transcriptional responses at specific times in the cell cycle
and/or diurnal growth cycle. We did not assess organellar
mRNA levels in our experiment, but these are also known
to be periodically expressed and/or diurnally controlled
(Salvador et al., 1993; Hwang et al., 1996; Klein, 2008; Idoine
et al., 2014).
Flagella are normally made once per diurnal cycle after mitotic
exit and basal bodies are replicated during S phase. In these
cases, the highly coordinated expression of genes for the constituent proteins of these organelles matches a specific temporal
demand. Similarly, most cell division-related proteins are required
during S/M and could be detrimental if expressed at other times,
so tight temporal coupling between cell cycle mRNA and protein
production is also expected, though the magnitude of change we
observed for many genes was remarkably high. In cells with welldefined G1, S, G2, and M phases, there are separate transcriptional program for entry into S phase and M phase (Wittenberg and
18 of 27
The Plant Cell
Reed, 2005; Desvoyes et al., 2014), but in Chlamydomonas, the
rapid alternation between S and M phases likely necessitates one
large burst of transcription for both DNA replication and mitotic
genes. Nonetheless, fine-scale expression phasing of genes
whose products are predicted to participate in S phase and mitosis are not identical, with replication genes expressed prior to
mitotic genes (Figure 4). However, the largely overlapping temporal domains of expression for most cell cycle regulators suggest
that if protein abundance is modulated as cells alternate between
S and M phases (e.g., cyclin synthesis and degradation), the
modulation is likely to be posttranscriptional.
For proteins that are required for photosynthetic growth in the
light, such as photosystem proteins, CBB cycle enzymes, and
chlorophyll biosynthetic proteins, the production of their mRNAs
in the light phase can be rationalized by the demand for increasing chloroplast size and photosynthetic capacity as cells
grow. Indeed, our findings match well with direct measurements of
photosynthetic metabolites such as chlorophyll and starch from
synchronous cultures (Willamme et al., 2015).
On the other hand, RPGs from three different compartments
showed maximum expression at different times (Figure 7D). In the
case of mitochondrial RPGs, their expression roughly matches the
diurnal growth profile of cells. In contrast, chloroplast RPGs show
a far more restricted temporal domain of expression, suggesting
that most chloroplast ribosomes are produced early in the light
phase, perhaps in anticipation of demand for chloroplast translation during growth. However, this initial burst of production is not
maintained (Figure 7B). Interestingly, amino acid biosynthetic
genes are also upregulated early in the light period as observed by
corresponding ontology term enrichments for c4, c5, and c6
(Figure 2B) and the light stress cluster (Figure 9; Supplemental
Data Set 14), and this enrichment matches a pattern of free amino
acid accumulation early in the light phase (Willamme et al., 2015).
The coordinated upregulation of chloroplast protein biosynthetic
capacity at a specific time in the diurnal or cell cycle is of potential
utility in biotechnology applications that aim to enhance production of chloroplast encoded transgenic proteins (Specht et al.,
2010; Scaife et al., 2015).
The nearly uniform increase in cytosolic RPG expression at the
beginning of the dark period is the most difficult to understand
because it is out of phase with cell growth. A prior radiolabeling
study of cytoplasmic and chloroplast rRNA synthesis and incorporation into ribosomes from synchronous cultures showed
steady mass increases of rRNAs and ribosomes in the light phase
and higher rates of label incorporation for both compartments
during the light phase (Wilson and Chiang, 1977). Thus, RPG
mRNA accumulation does not appear to be directly coordinated
with rRNA synthesis and ribosome assembly. Ribosomal protein
genes can be regulated both transcriptionally and translationally
(Mager, 1988; Larson et al., 1991; Perry, 2007) and may also
participate in nonribosomal processes including transcriptional
and translational control of specific genes (Lindström, 2009). In
plants and algae, RPG regulation is less well studied (McIntosh
and Bonham-Smith, 2006), especially with respect to diurnal
cycles and the cell cycle, though Kucho et al. (2005) did find many
chloroplast RPGs in their circadian data set.
A diurnal study of the marine pico-alga O. tauri also found
evidence that cytosolic translation machinery might be made in
the dark and organelle machinery in the light, though the study
only observed a subset of genes and did not discriminate between mitochondrial and chloroplast genes (Monnier et al., 2010).
Finding similar dark-phase RPG expression profiles in a distantly
related green alga suggests that this expression pattern for cytosolic ribosomes may offer a selective advantage in photosynthetic
microalgae, and its underlying bases merit further investigation. In
yeast, cytosolic RPG biosynthesis is coupled to G1-S phase cell
cycle progression (Jorgensen et al., 2004; Bernstein et al., 2007;
Gómez-Herreros et al., 2013), though it remains to be determined if
there is any underlying mechanistic similarity between ribosome
biosynthesis and cell cycle control in yeast and algae. We also
note that at their expression peak, cytosolic RPGs comprise
;45% of all transcripts in Chlamydomonas (Supplemental
Figure 16), and their upregulation during the dark phase may
reflect the time when RPG expression competes the least with
other biosynthetic demands in the cell.
Diurnal metabolism is coupled directly to light conditions, and
we found here that genes related to light-regulated metabolic
processes were strongly regulated at the level of transcript
abundance, including genes for photosystem and light-harvesting
complex subunits, chlorophyll biosynthesis, CBB cycle genes,
and chloroplast translational machinery whose expression in
many cases changed by orders of magnitude between the light
and dark phases (Figures 7, 8, and 10; Supplemental Figures 16 to
20). Other areas of metabolism with periodic regulation of expression included respiration/TCA cycle, whose genes were more
highly expressed in the dark as expected given their role in starch
catabolism. The expression profiles of genes encoding enzymes
of other pathways, such as glycolysis/gluconeogenesis, were less
coherent, such as in the case of triosephosphate isomerase
(TPIC1), which participates in both anabolic and catabolic processes. In addition, paralogs for central carbon metabolism
genes, even those in the same pathway, frequently showed different expression profiles and transcript abundances, suggesting
functional partitioning between duplicates or absence of transcriptional control (e.g., RBCS1 and RBSC2; TAL1 and TAL2; and
ACS1 to ACS3) (Figure 10; Supplemental Data Set 15). The differential accumulation patterns and relative transcript abundances identified here for enzyme isoforms should help inform
metabolic modeling and engineering efforts that require enzyme
activity estimates for making accurate predictions, though
measurements of protein abundance will be essential to confirm
our predictions.
Our findings raise a more general question of how closely the
proteome matches that of the transcriptome in synchronous
growth conditions and why so many genes are expressed periodically. A global proteomic approach is likely to be informative on
this point, but to date has not been done for synchronized diurnal
cultures. Very few previous studies on diurnal regulated gene
expression in Chlamydomonas investigated the correlation between transcripts and proteins, and most focused on chloroplast
proteins and mRNAs that were not considered in our study (Herrin
et al., 1986; Eberhard et al., 2002; Lee and Herrin, 2002). Among
nuclear genes that have been examined, the highly abundant
Rubisco small subunit protein showed elevated abundance in the
light and decreased abundance in the dark (Recuenco-Muñoz
et al., 2015), similar to the combined mRNA abundance profiles we
Chlamydomonas Diurnal Transcriptome
measured for its two closely related paralogs, RBCS1 and RBCS2
(Figure 10A) (Goldschmidt-Clermont and Rahire, 1986). The mRNA
for intraflagellar transport protein IFT27 has a high-amplitude
diurnal expression profile with peak abundance just after cell
division and very little expression outside this time (Wood et al.,
2012) (Figure 5C; Supplemental Data Set 7). Unlike its mRNA,
IFT27 protein is present throughout the diurnal cycle but appears
to made only once per cell cycle in daughter cells during reflagellation and then becomes slowly diluted as cells grow
(Wood et al., 2012). In this case, the protein is more stable than the
mRNA, but the periodic mRNA expression still impacts the relative
abundance of IFT27 during the diurnal cycle and may even have
a global impact on the abundance of other IFT proteins (Qin et al.,
2007; Wood et al., 2012).
Using Transcriptomics of Synchronous Cultures as a Tool
for Functional Classification
Approximately 90% of the genes in Chlamydomonas (15,908/
17,737) have not been manually annotated with a primary gene
symbol, and only ;40% of genes have associated PFAM domains
(I.K. Blaby, unpublished data). Our study has revealed that genes
for related processes are frequently coexpressed in highly correlated patterns, thus providing a tool that can aid in assigning
functions to unknown genes. In each area where we investigated
a biological process, our coexpression data not only recapitulated
or validated functional information gathered over the past 50+
years (e.g., flagella genes and photosynthetic complexes), but in
many cases refined that information and generated new hypotheses about gene function. The utility of our data are exemplified by our analyses of flagella-related genes, for which we not
only showed that nearly every known flagella gene is coregulated
during the cell cycle, but also classified hundreds of putative
flagella and basal body protein coding genes that were identified in
large-scale experiments. This will help to narrow down these sets
to those likely to be involved directly in flagella and basal body
functions. Equally revealing were the few examples of core flagella
genes that were outliers and did not display the typical expression
patterns of flagella clusters 11 to 14 (Figure 5). In each case where
information is available, these outliers correspond to genes with
additional functions unrelated to flagella. A second striking example of potentially useful information deriving from coexpression
analyses came from chloroplast RPGs and translation factors.
These could be grouped into a majority class that were coexpressed with a peak early in the light phase peak and a few outliers
whose expression patterns suggest highly orchestrated and
specific changes in chloroplast translational regulation over the
course of the day-night cycle. Remarkably, our analysis identified
PSRP-1 as a gene whose transient upregulation at the light-dark
transition matches that of its distant ortholog in cyanobacteria
(Tan et al., 1994; Samartzidou and Widger, 1998), suggesting
ancient functional conservation or this ribosomal protein between
cyanobacteria and Viridiplantae.
Gene duplication and divergence are well established contributors to evolutionary diversification, and one such type of
divergence is in expression pattern between paralogs (Ohno,
1970; Conant and Wolfe, 2008). One of the most challenging
problems in metabolic modeling is the assignment of duplicate
19 of 27
enzymatic activities to specific pathways and subcellular compartments. Tools designed for predicting protein localization such
as Predalgo are valuable but imperfect and sometimes yield results that conflict with experimental data (Tardif et al., 2012). In the
absence of data on targeting for over two dozen proteins of central
carbon metabolism, coexpression proved to be useful in at least
some cases for assigning paralogs to compartments and pathways (Supplemental Data Set 15). It is interesting to note that for
some parts of central carbon metabolism, such as the CBB cycle,
the genes show highly correlated expression patterns, whereas in
other pathways, such as glycolysis and gluconeogenesis, the
expression patterns of many genes were discordant. This discordance highlights the limitations of using transcriptome data
where in some cases genes for enzymes that participate in the
same pathway show uncorrelated or even anticorrelated expression profiles (Figure 10). Such complex profiles could reflect
adaptive responses—perhaps resulting from integration with
other areas of metabolism that are poorly defined or understood—
or could simply reflect lack of selection for a coordinated transcriptional response because enzyme activity regulation is achieved at other levels.
Compilation of a Reference Data Set for Future Studies of
Diurnal and Cell Cycle Regulated Processes
One potential application for our data set is the construction of
transcriptional networks related to diurnal, cell cycle, and/or circadian regulation. Besides uncoupling the inputs corresponding
to the above three categories of regulation, a major challenge for
such work lies in dealing with highly dynamic data with little or no
information available on phase delays between TF expression and
target gene expression. The potential lack of correlation between
TF mRNA abundance, TF protein abundance, and TF transcriptional activity are additional challenges. Nonetheless, our finding
that nearly the entire set of DNA binding TFs in Chlamydomonas
showed strong and distinctly phased expression patterns suggests that at least some TFs and associated target genes are
controlled at the level of transcription. Moreover, we identified
a subset of TFs whose transcripts showed dynamic changes in our
specific diurnal conditions but not in other transcriptome experiments (Supplemental Figure 23 and Supplemental Data Set
16), making them prime candidates for controlling periodic gene
expression. A previous study to identify cis-regulatory elements
from diurnal transcripts of Chlamydomonas proved difficult to
compare with our study as culture synchrony in that study was not
measured, coregulated gene lists were not published, and the
“gold standard” genes used to validate clusters contained some
entries that were duplicated between clusters (Panchy et al.,
2014). Based on our findings, the identification of cis-regulatory
elements in Chlamydomonas may not be trivial because multiple
TFs are expressed in each cluster and each might have its own
distinct binding motif. Nonetheless, our data provide an improved
resource for attempting to find cis-regulatory elements and associate them with cognate TFs.
Our analyses of a high-resolution synchronous transcriptome
from Chlamydomonas covered several key areas of cell biology
and metabolism (Figure 10) but were by no means exhaustive.
There are many other cellular processes whose investigation will
20 of 27
The Plant Cell
be enabled by our data. A range of cell behavior and physiology
in Chlamydomonas is under cell cycle, diurnal, or circadian control
(Matsuo and Ishiura, 2010). Given the prevalence of transcript
cycling (>80% of expressed genes) and coexpression of
genes with related functions, it is likely that many processes in
Chlamydomonas have some periodic component that can investigated productively using data from our study.
METHODS
Culture Conditions
Strain CC-5152 MT- was derived from a cross between parental strains
21gr (CC-1690) and 6145c (CC-2895) and used for all experiments.
CC-5152 was inoculated from a single colony into a sterilized 400-mL
FMT150 photobioreactor (Photon Systems Instruments) containing HS
media (Harris, 1989). The culture was grown at 28°C in diurnal (12 h light,
12 h dark) conditions with equal fluences of red (630 nm) and blue (450 nm)
light (125 mE m22 s21 each) and maintained at a density of OD680 = 0.3 6
0.01 and volume of 400 mL by automated addition of fresh media and
removal of excess culture.
Sample Collection and RNA Preparation
Cultures were equilibrated in the photobioreactor for 3 to 4 d prior to
sampling. Sampling volumes were adjusted as described below to ensure
that equal biomass (pellet wet weight) was present in each sample. Pellet
biomasses were each equivalent to ;2 3 106 daughter cells. For light
phase samples, 30 mL was taken at each time point and culture volume was
restored to 400 mL by automated dilution as the remaining culture grew. For
dark phase time points, media were replaced after each sampling to
maintain a 400-mL culture volume, and progressively larger volumes were
removed at successive time points to compensate for culture dilution. A
small sample of culture from each time point was fixed in 0.2% (v/v)
glutaraldehyde and 0.005% (v/v) Tween 20 and used to measure cell size
and concentration using a Coulter Counter (Beckman Coulter; Multisizer 3).
Mitotic progression and passage through Commitment were scored as
previously described (Fang et al., 2006). Presence/absence of flagella was
scored microscopically using the same fixed samples as for scoring mitotic
progression. Replicates were performed one month apart from each other.
Collected cells were pelleted by centrifugation (30 s, 3220 RCF, 24°C) in
0.005% (v/v) Tween 20 and resuspended in 0.25 mL RNase-free water and
then mixed with 0.25 mL lysis buffer (50 mM Tris-HCl, pH 8.0, 200 mM NaCl,
20 mM EDTA, 2% SDS, and 1 mg/mL Proteinase K) at 70°C. Ten milliliters of
TRIzol (Invitrogen) was added, and samples were then immediately flash
frozen in liquid nitrogen and stored at 280°C. After all samples were
collected, RNA was prepared by thawing (10 min 24 °C) and transferring to
15-mL MaXtract HD tubes (Qiagen) followed by extraction with 2 mL of
chloroform. The aqueous phase was mixed with 1.5 volumes 100% ethanol
and applied to a MicroRNeasy column under vacuum (Qiagen). Washing,
DNase treatment, and elution were performed according to the manufacturer’s instructions (RNase-free DNase Set; Qiagen). The eluted RNA
concentration was measured using a Pearl nanophotometer (Implen) and
the quality of RNA evaluated by gel electrophoresis and a bioanalyzer
(Agilent RNA 6000 Nano Kit).
Library Preparation and Illumina Sequencing
Plate-based RNA-Seq with poly(A) selection sample prep was performed
on the Perkin-Elmer Sciclone NGS robotic liquid handling system, where
purified mRNA is converted into cDNA library templates of known strand
origin for sequencing on the Illumina Sequencer platform. The DynaBead
kit (Invitrogen) was used to select and purify poly(A)-containing mRNA from
5 µg of total RNA per sample. The mRNA was chemically fragmented with
103 fragmentation solution (Ambion) at 70°C for 3 min to generate fragments ranging in size between 250 to 300 bp. The fragmented RNA was
purified using AMpure SPRI beads (Agencourt) using a ratio of 160:100
beads volume:RNA sample volume. First-strand cDNA was synthesized
using SuperScript II reverse transcriptase (Invitrogen) and random hexamer
primers (MBI Fermentas) with actinomycin D in the master mix to inhibit
DNA-dependent DNA synthesis. First-strand synthesis thermocycler
conditions were 42°C for 50 min and an inactivation step at 70°C for 10 min.
This was followed by another purification using AMpure SPRI beads at
a ratio of 140:100 beads volume:cDNA volume. Second-strand cDNA
synthesis was performed using DNA Polymerase I (Invitrogen), RNase H
(Invitrogen), and a nucleotide mix containing dUTP (Roche). The secondstrand synthesis was done at 16°C for 60 min. The double-stranded cDNA
fragments were purified using AMpure SPRI beads at a ratio of 75:100
beads volume: cDNA volume followed by a second purification using
a bead:cDNA ratio of 140:100. cDNA fragments were end repaired,
phosphorylated, and A-tailed using a Kapa Biosystems kit followed by
ligation to Illumina barcoded sequencing adapters. AmpErase UNG (uracil
N-glycosylase; Applied Biosystems) was added to the double-stranded
cDNA library fragments and incubated at 37°C for 15 min to cleave and
degrade the strand containing dUTP. The single-stranded cDNA was then
enriched using 10 cycles of PCR with Illumina TruSeq primers and purified
using AMpure SPRI beads at a ratio of 90:100 beads volume:cDNA volume
to create the final cDNA library. Libraries were quantified using KAPA
Biosystem’s next-generation sequencing library qPCR kit using a Roche
LightCycler 480 real-time PCR instrument. The libraries were then prepared
for sequencing on the Illumina HiSeq sequencing platform using a TruSeq
paired-end cluster kit, v3, and Illumina’s cBot instrument to generate
clustered flow cells for sequencing. Sequencing of the flow cells was
performed on the Illumina HiSeq2000 sequencer using a TruSeq SBS
sequencing kit with 200 cycles, v3, following a 2 3 100 indexed run recipe.
Primary data are available at the NCBI Gene Expression Omnibus repository under accession number GSE71469.
Sequence Analyses and Identification of Genes with
Differential Expression
Quality control of RNA-Seq samples was performed on the raw paired-end
reads using Trimmomatic (Bolger et al., 2014) to remove contaminating
Illumina adapter sequences and low-quality sequences (average Q20 over
a 4-base sliding window <20). Paired reads that were <25 bp at either end
were discarded. Reads were aligned to the Chlamydomonas reinhardtii
genome v5 assembly with STAR (Dobin et al., 2013) using standard
presets except for intron size, which was set between 20 and 3000 bp
(–alignIntronMin 20 and –alignIntronMax 3000). Greater than 10 million
reads were mapped for each sample (Supplemental Table 1) with uniquely
mapping reads accounting for ;90% of total mapped reads in each
sample. Uniquely mapping reads were assigned to 17,737 version 5.3.1
primary transcripts using HT-Seq (Anders et al., 2015). One read was added
to each transcript in each sample before normalization to allow computation of expression ratios across all samples. This addition had
no measurable effect on normalized gene expression estimates. Expression estimates were normalized to library size (uniquely mapping
reads per million [RPM]) and pairwise Pearson correlations of normalized
samples were computed to evaluate replicate reproducibility. All replicate samples were highly correlated (R > 0.973) except for the pair of 6-h
dark samples taken at ZT18 (R = 0.938). To determine whether one or both of
these replicates was a possible outlier, each was compared with adjacent
time points by Pearson correlation, ZT17 and ZT19, with the assumption that
high similarity between adjacent samples is an indicator of sample integrity
(see Figure 1D showing correlations for all pairwise sample comparisons). Of
Chlamydomonas Diurnal Transcriptome
the two 6D samples, 6D_1 was well correlated with both replicate samples
at times ZT17 (R > 0.99) and ZT19 (R > 0.976). The 6D_2 sample was less
correlated to the ZT17 and ZT19 samples (R = 0.927) and was discarded as
a technical outlier. Because a minimum of two replicates per sample is
required for DESeq2 differential expression analysis, a replicate of 6D_1 was
mapped and quantified in place of the 6D_2 sample. HTSFilter was used to
establish a minimum expression threshold of 1.061 average RPM for detection of differential expression (Rau et al., 2013). A total of 2412 genes that
did not meet this criterion were removed from consideration.
Differential expression analysis was performed on the remaining 15,325
RPM-normalized genes using DESeq2 (Love et al., 2014) by comparing the
replicate average for each gene in each sample with the average normalized expression for the gene across all samples. A total of 13,118 genes
were classified as differentially expressed based on having a greater than
2-fold expression difference from their mean expression in at least one time
point, with a false discovery rate of <0.05. After differential expression
analysis, expression estimates were normalized by the primary transcript
length for each locus to calculate RPKM values. A total of 526 differentially
expressed genes with a maximum RPKM <1 were removed from further
analyses, leaving 12,592 significantly differentially expressed genes, 2179
nondifferentially expressed genes, and 2966 nonexpressed genes.
Coexpression Clustering
A modified implementation of the KMC K-means algorithm implemented in
MeV was used to cluster the 12,592 significantly differentially expressed
genes (Saeed et al., 2003). First, replicate expression estimates were
averaged and then transformed to standard deviations from the mean
expression for each gene [(expression 2 mean)/SD]. The figure of merit
algorithm was used to estimate an appropriate number of clusters (Yeung
et al., 2001). K-means support using Pearson’s correlation was then used
to separate groups of coregulated genes into six initial clusters. Genes that
did not cocluster with the main clusters were grouped into smaller satellite
clusters. A subsequent round of figure of merit and K-means support was
performed on the six main clusters with correlation thresholds varying
between 0.7 and 0.96 and the number of iterations varying between 25 and
100, to produce 18 final clusters. Centroids were computed for each of the
18 clusters and genes from remaining satellite clusters were assigned to
their closest match among the 18 centroids.
JTK Cycle Analysis
The program JTK cycle (Hughes et al., 2010) was used to identify rhythmic
genes as follows. The two replicate data sets were treated as experimental
duplicates. The data were filtered to include only the 14,771 genes that met
minimum expression criteria (maximum average expression > 1.061 RPM
and > 1 RPKM). The temporal spacing between samples was set to 0.5 h
and the target period was set to 48 h (effectively equal to 24 h). A false
discovery rate cutoff (BH.Q < 1E-10) was chosen to extract a set of 12,534
rhythmically expressed genes that was comparable in size to the differentially expressed gene set identified as described above and used for
clustering.
21 of 27
10,000 data permutations were created by random sampling without replacement and cluster assignment (formally equivalent to generating
a hypergeometic distribution). The mean (TermOccurrenceclusterMean) and SD
(TermOccurrenceclusterSD) for each term were calculated from this background distribution and used for comparison with actual distribution data.
Z-scores (z-scoreterm,cluster), P values (p-valuesterm,cluster), and false discovery rates with correction for multiple testing (q-valueterm,clusters) were
calculated for MapMan terms in each cluster using R functions, defined as
follows:
h
z-score term;cluster ¼ TermOccurrence cluster TermOccurrence clusterMean
i
=TermOccurrence clusterSD
h
h
i
p-value term;cluster ¼ 2 pnorm -absðz-scoreÞ
q-value term;cluster ¼ p:adjustðÆvector of all p-valuesæ; method
i
¼ ”FDR”Þ ½cluster
A MapMan term was considered to be significantly enriched in a cluster
if its maximum z-score was >1.96 (corresponding to a P value < 0.05) and
the corresponding false discovery rate was <0.05. Terms associated with
fewer than six loci were not considered.
Gene Identifier Conversions
Reads were mapped to genome version 5.3.1 (v5.3.1) available at Phytozome (http://phytozome.jgi.doe.gov), and expression estimates were
assigned to v5.3.1 gene loci. Conversion between 5.3.1 IDs, the newest
set of IDs from v5.5, and IDs used in previously reported experiments
were done using a correspondence table available from Phytozome
(ChlamydomonasTranscriptNameConversionBetweenReleases.
Mch12b.txt). Four v5.3.1 models were each split into two new genes in
v5.5. For our analyses, we merged the two v5.5 loci back into a single
synthetic gene model as follows: v5.3.1 locus, Cre03.g167650, and synthetic v5.5 locus, Cre03.g167622_Cre03.g167644; v5.3.1 locus, Cre14.
g613050, and synthetic v5.5 locus, Cre14.g613050_Cre14.g613075;
v5.3.1 locus, Cre16.g681700, and synthetic v5.5 locus, Cre16.
g681600_Cre16.g681700; v5.3.1 locus, Cre17.g743300, and synthetic v5.5
locus, Cre17.g743288_Cre17.g743307.
For flagella proteome genes (those with more than two mapped peptides) (Pazour et al., 2005), Cilia Cut genes (Merchant et al., 2007), deflagellation upregulated genes (Albee et al., 2013), and transcription factor
genes (Pérez-Rodríguez et al., 2010; Jin et al., 2014) (Supplemental Data
Sets 7 and 16) that did not have a single v5.3.1 ID, we did a manual BLASTP
search against the v5.3.1 proteome to identify a corresponding gene model
based on high sequence coverage (>50%) and retention of any signature
domains present in the original gene model. For all other transcriptome
comparisons, only IDs with 1:1 correspondences with v5.3.1 ID were used.
Tests for Cluster Membership Enrichment
Functional Annotation of Expression Clusters
Functional annotations for Chlamydomonas reinhardtii v5.3.1 predicted
proteins were obtained from the MapMan website (http://mapman.gabipd.
org) (Thimm et al., 2004) and converted to v5.5 locus IDs (referred to in this
study) based on a correspondence table downloaded from Phytozome
10.1 (http://phytozome.jgi.doe.gov). Level 1 and 2 MapMan ontology terms
were tabulated for the entire expressed transcriptome of 14,771 genes
(>1.061 RPM and >1 RPKM) and used to generate a background distribution model for clusters c1 to c18 along with the 2179 nondifferentially
expressed genes that met the expression threshold criteria above. A total of
Enrichment tests of cluster membership for annotated cell cycle and flagella genes were performed using a resampling method similar to that used
for MapMan Term enrichment (see above). A background distribution of
cluster membership for sample sizes equal to the number of cell cycle
genes or flagella genes was computed and used to derive expected cluster
membership numbers and associated statistics.
Comparisons of cluster membership differences between gene groups
were performed using Fisher’s exact test. The 532 contingency tables
were constructed using observed measurements in individual clusters
versus a background distribution based on expected values from
22 of 27
The Plant Cell
resampling. The flagella cluster group was comprised of genes in clusters
c11 to c14, with genes from all other clusters plus unclustered expressed
genes combined into a second group for each test.
Comparisons of cluster membership differences between Basal Body,
BUG, POC, and Axoneme gene groups were also performed using Fisher’s
exact test. A 435 contingency table was constructed using observed
measurements in clusters and unclustered expressed genes. Clusters c11
to c14 were treated as individual groups and membership in all other
clusters and unclustered genes were combined into a fifth group.
Enrichment of Cluster c11 to c14 Membership in Data Sets
Containing Predicted Flagella Genes
Candidate flagella genes from the flagella proteome (Pazour et al., 2005),
CiliaCut (Merchant et al., 2007), and deflagellation response (Albee et al.,
2013) were scored for membership in flagella clusters (c11 to c14) versus
other clusters (c1 to 10, c15 to c18, and unclustered) and for overlapping
membership between the three studies. Resampling without replacement
from 10,000 replicates (as described in the preceding section) was used to
derive expected null distributions in each overlap category and compared
with actual data derived by separating c11 to c14 genes in each study from
the remaining genes (Figure 6B; Supplemental Figure 15).
Coexpression of Paralogs Encoding Enzymes of Central
Carbon Metabolism
quarter weight, and those 30 min from the central sample given half
weight. Samples ZT1 and ZT24 were each used for their respected
transformations.
Gene expression graphs in Supplemental Figures 20 and 21 were
smoothed using the FFT method of signal processing with two points in
Origin (OriginLab).
Accession Numbers
Primary data are available at the NCBI Gene Expression Omnibus repository under accession number GSE71469. Chlamydomonas locus ID
numbers for genes described in this study are listed in the appropriate
supplemental data sets and can be accessed from Phytozome (http://
phytozome.jgi.doe.gov).
Supplemental Data
Supplemental Figure 1. Comparison of diurnally cycling genes
identified with JTK cycle and DESeq2.
Supplemental Figure 2. Expression of rhythmic genes that were
identified by DE analysis but not JTK cycle.
Supplemental Figure 3. Expression of rhythmic genes that were
identified by JTK cycle but not by DE analysis.
Supplemental Figure 4. Expression profiles of DNA replication genes I.
Pathway signature genes, defined as genes encoding enzymes whose
activity is unique to a specific metabolic pathway, were identified as follows: glyoxylate cycle, MAS1 and ICL1 (encoding malate synthase and
isocitrate lyase); TCA cycle, SDH3 and SDH4 (encoding succinate dehydrogenase subunits 3 and 4); CBB cycle, RBCS1 and SBP1 (encoding
Rubisco small subunit 1 and sedoheptulose-1,7-bisphosphatase);
fermentation, PFL1 (encoding pyruvate-formate lyase); pentose phosphate pathway, GLD1 and GLD2 (encoding glucose-6-phosphate dehydrogenase); and the committed steps of glycolysis, PFK1 and
PFK2 (encoding phosphofructokinase) and gluconeogenesis PCK1
and FBP1 (encoding phosphoenolpyruvate carboxykinase and fructose1,6-bisphosphatase). The expression estimates of these signature genes
were used to determine expression profiles indicative of each pathway by
averaging scaled expression values for the signature genes in a pathway.
Pearson’s correlation coefficients were then calculated between all
paralogs that potentially participate in each of these pathways and
each pathway expression profile to evaluate coexpression between
pathway-containing genes. Paralogs were assigned to pathways on the
basis of highest correlation to the signature genes, using a correlation
threshold $0.6.
Supplemental Figure 5. Expression profiles of DNA replication genes II.
Curation of a Light Stress Response Cluster
Supplemental Figure 16. Ribosomal protein gene transcript contribution to the transcriptome.
A custom light stress response cluster of 280 genes was generated whose
members came almost entirely from a subset of genes in c1 and c2 that
showed a sharp transient peak of expression at ZT1. Criteria for inclusion
were an expression maximum at ZT1 that was $5 RPKM, expression levels
at ZT3, ZT4, and ZT5 no greater than 80, 60, and 50% of the maximum, and
expression at ZT6-ZT24 no higher than 40% of the maximum.
Gene Expression Graph Smoothing
Gene expression graphs in Figure 3 were smoothed by a custom rolling
average function created within the programming shell R (http://www.Rproject.org/). Absolute expression estimates were first averaged and then
normalized to maximum expression. The estimates were then converted to
the weighted average of samples taken within 1 h of each time point with the
central sample given full weight, those 1 h from the central sample given
Supplemental Figure 6. Expression profiles of SMC genes.
Supplemental Figure 7. Expression profiles of cell cycle regulatory
genes.
Supplemental Figure 8. Expression profiles of RDP genes.
Supplemental Figure 9. Expression profiles of anaphase promoting
complex/cyclosome (APC/C) genes.
Supplemental Figure 10. Expression profiles of chloroplast division
genes.
Supplemental Figure 11. Cell cycle gene resampling statistics.
Supplemental Figure 12. Expression patterns of DIV and GEX genes.
Supplemental Figure 13. Resorption and reformation of flagella
during the cell cycle.
Supplemental Figure 14. Flagella and basal body gene cluster
membership statistics.
Supplemental Figure 15. Resampling statistics for comparisons of
flagellar gene group membership overlap.
Supplemental Figure 17. Expression profiles of nuclear genes
encoding subunits of photosynthetic complexes.
Supplemental Figure 18. Expression profiles of genes encoding
subunits of complexes I-III of the mitochondrial electron transport chain.
Supplemental Figure 19. Expression profiles of genes encoding
complexes IV-V of the mitochondrial electron transport chain.
Supplemental Figure 20. Expression profiles of nuclear genes encoding subunits of chloroplast and mitochondrial ATP synthase complexes.
Supplemental Figure 21. Expression profiles of genes encoding
enzymes of tetrapyrrole metabolism.
Supplemental Figure 22. Expression profiles of genes encoding LHClike proteins.
Chlamydomonas Diurnal Transcriptome
Supplemental Figure 23. Expression profiles of transcription factor
genes.
Supplemental Table 1. Illumina read mapping statistics.
Supplemental Data Set 1. Gene expression estimates, cluster
membership, and additional data.
Supplemental Data Set 2. Genes with constant expression.
Supplemental Data Set 3. MapMan term cluster distributions and
statistical enrichment tests.
Supplemental Data Set 4. Cell cycle gene expression estimates.
Supplemental Data Set 5. DIV and GEX genes.
Supplemental Data Set 6. Flagella gene expression estimates.
Supplemental Data Set 7. Flagella gene group membership.
Supplemental Data Set 8. Ribosomal gene expression estimates.
Supplemental Data Set 9. Photosynthetic complex gene expression.
Supplemental Data Set 10. Respiratory complex gene expression.
Supplemental Data Set 11. Tetrapyrrole metabolism-related gene
expression.
Supplemental Data Set 12. Light stress cluster gene expression.
Supplemental Data Set 13. LHC-like gene expression.
Supplemental Data Set 14. MapMan annotations for light stress
cluster.
Supplemental Data Set 15. Central carbon metabolism enzyme gene
expression and pathway assignments.
Supplemental Data Set 16. Transcription factor gene expression.
Supplemental Data Set 17. Nonexpressed genes.
ACKNOWLEDGMENTS
This work was supported by National Institutes of Health Grant R01
GM069521 to J.G.U. and by the University of California-Los Angeles
Department of Energy Institute of Genomics and Proteomics (DE-FC0302ER63421) and National Institutes of Health (NIH) R24 GM092473 to
S.S.M. J.M.Z. was supported by the National Science Foundation Integrative Graduate Education and Research Traineeship Fellowship (0504645).
I.K.B. was supported by a training grant from the NIH (T32ES015457). The
work conducted by the U.S. Department of Energy Joint Genome Institute,
a DOE Office of Science User Facility, is supported by the Office of Science
of the U.S. Department of Energy under Contract DE-AC02-05CH11231.
We thank the Joint Genome Institute for Illumina library preparation and
sequencing and Patrice Salomé for critical reading of the article. We thank
Stephen Douglas and Matteo Pellegrini for their contribution to analysis of
a preliminary diurnal data set.
AUTHOR CONTRIBUTIONS
J.G.U., S.S.M., and J.M.Z. designed the experiments. J.M.Z. performed
the experiments and the bioinformatics analysis of the RNA-Seq data.
J.G.U. and S.S.M. supervised the experiment and bioinformatics analysis.
J.M.Z. and I.K.B. analyzed the data. J.G.U., S.S.M., I.K.B., and J.M.Z.
prepared and edited the article.
Received June 8, 2015; revised July 27, 2015; accepted September 14,
2015; published October 2, 2015.
23 of 27
REFERENCES
Adams, S., Maple, J., and Møller, S.G. (2008). Functional conservation of the MIN plastid division homologues of Chlamydomonas
reinhardtii. Planta 227: 1199–1211.
Albee, A.J., Kwan, A.L., Lin, H., Granas, D., Stormo, G.D., and
Dutcher, S.K. (2013). Identification of cilia genes that affect cellcycle progression using whole-genome transcriptome analysis in
Chlamydomonas reinhardtti. G3 (Bethesda) 3: 979–991.
Anders, S., Pyl, P.T., and Huber, W. (2015). HTSeq–a Python
framework to work with high-throughput sequencing data. Bioinformatics 31: 166–169.
Anwaruzzaman, M., Chin, B.L., Li, X.-P., Lohr, M., Martinez, D.A.,
and Niyogi, K.K. (2004). Genomic analysis of mutants affecting
xanthophyll biosynthesis and regulation of photosynthetic light
harvesting in Chlamydomonas reinhardtii. Photosynth. Res. 82:
265–276.
Ashworth, J., Coesel, S., Lee, A., Armbrust, E.V., Orellana, M.V.,
and Baliga, N.S. (2013). Genome-wide diel growth state transitions
in the diatom Thalassiosira pseudonana. Proc. Natl. Acad. Sci. USA
110: 7518–7523.
Beligni, M.V., and Mayfield, S.P. (2008). Arabidopsis thaliana mutants
reveal a role for CSP41a and CSP41b, two ribosome-associated
endonucleases, in chloroplast ribosomal RNA metabolism. Plant
Mol. Biol. 67: 389–401.
Bernstein, K.A., Bleichert, F., Bean, J.M., Cross, F.R., and
Baserga, S.J. (2007). Ribosome biogenesis is sensed at the Start
cell cycle checkpoint. Mol. Biol. Cell 18: 953–964.
Bišová, K., and Zachleder, V. (2014). Cell-cycle regulation in green
algae dividing by multiple fission. J. Exp. Bot. 65: 2585–2602.
Bisová, K., Krylov, D.M., and Umen, J.G. (2005). Genome-wide annotation and expression profiling of cell cycle regulatory genes in
Chlamydomonas reinhardtii. Plant Physiol. 137: 475–491.
Blaby, I.K., et al. (2014). The Chlamydomonas genome project:
a decade on. Trends Plant Sci. 19: 672–680.
Bolger, A.M., Lohse, M., and Usadel, B. (2014). Trimmomatic:
a flexible trimmer for Illumina sequence data. Bioinformatics 30:
2114–2120.
Bonente, G., Passarini, F., Cazzaniga, S., Mancone, C., Buia, M.C.,
Tripodi, M., Bassi, R., and Caffarri, S. (2008). The occurrence of
the psbS gene product in Chlamydomonas reinhardtii and in other
photosynthetic organisms and its correlation with energy quenching. Photochem. Photobiol. 84: 1359–1370.
Bonente, G., Pippa, S., Castellano, S., Bassi, R., and Ballottari, M.
(2012). Acclimation of Chlamydomonas reinhardtii to different
growth irradiances. J. Biol. Chem. 287: 5833–5847.
Bower, R., Tritschler, D., Vanderwaal, K., Perrone, C.A., Mueller,
J., Fox, L., Sale, W.S., and Porter, M.E. (2013). The N-DRC forms
a conserved biochemical complex that maintains outer doublet alignment and limits microtubule sliding in motile axonemes. Mol.
Biol. Cell 24: 1134–1152.
Carvalho-Santos, Z., Azimzadeh, J., Pereira-Leal, J.B., and
Bettencourt-Dias, M. (2011). Evolution: Tracing the origins of
centrioles, cilia, and flagella. J. Cell Biol. 194: 165–175.
Castruita, M., Casero, D., Karpowicz, S.J., Kropat, J., Vieler, A.,
Hsieh, S.I., Yan, W., Cokus, S., Loo, J.A., Benning, C., Pellegrini,
M., and Merchant, S.S. (2011). Systems biology approach in
Chlamydomonas reveals connections between copper nutrition and
multiple metabolic steps. Plant Cell 23: 1273–1292.
Chang, L., and Barford, D. (2014). Insights into the anaphase-promoting
complex: a molecular machine that regulates mitosis. Curr. Opin. Struct.
Biol. 29: 1–9.
Conant, G.C., and Wolfe, K.H. (2008). Turning a hobby into a job: how
duplicated genes find new functions. Nat. Rev. Genet. 9: 938–950.
24 of 27
The Plant Cell
Coss, R.A. (1974). Mitosis in Chlamydomonas reinhardtii basal bodies
and the mitotic apparatus. J. Cell Biol. 63: 325–329.
Covington, M.F., Maloof, J.N., Straume, M., Kay, S.A., and Harmer,
S.L. (2008). Global transcriptome analysis reveals circadian regulation of key pathways in plant growth and development. Genome
Biol. 9: R130.
Craige, B., Tsao, C.-C., Diener, D.R., Hou, Y., Lechtreck, K.-F.,
Rosenbaum, J.L., and Witman, G.B. (2010). CEP290 tethers flagellar transition zone microtubules to the membrane and regulates
flagellar protein content. J. Cell Biol. 190: 927–940.
Cross, F.R., and Umen, J.G. (2015). The Chlamydomonas cell cycle.
Plant J. 82: 370–392.
Dal’Molin, C.G. de O., Quek, L.-E., Palfreyman, R.W., and Nielsen,
L.K. (2011). AlgaGEM–a genome-scale metabolic reconstruction of
algae based on the Chlamydomonas reinhardtii genome. BMC Genomics 12 (suppl. 4): S5.
Desvoyes, B., Fernández-Marcos, M., Sequeira-Mendes, J., Otero,
S., Vergara, Z., and Gutierrez, C. (2014). Looking at plant cell cycle
from the chromatin window. Front. Plant Sci. 5: 369.
De Veylder, L., Beeckman, T., and Inzé, D. (2007). The ins and outs
of the plant cell cycle. Nat. Rev. Mol. Cell Biol. 8: 655–665.
Dobin, A., Davis, C.A., Schlesinger, F., Drenkow, J., Zaleski, C.,
Jha, S., Batut, P., Chaisson, M., and Gingeras, T.R. (2013). STAR:
ultrafast universal RNA-seq aligner. Bioinformatics 29: 15–21.
Drop, B., Webber-Birungi, M., Yadav, S.K.N., FilipowiczSzymanska, A., Fusetti, F., Boekema, E.J., and Croce, R. (2014).
Light-harvesting complex II (LHCII) and its supramolecular organization in Chlamydomonas reinhardtii. Biochim. Biophys. Acta 1837:
63–72.
Dutcher, S.K. (2009). Basal bodies and associated structures. In The
Chlamydomonas Sourcebook, G.B. Witman and E.H. Harris, eds
(London: Academic Press), pp. 15–42.
Eberhard, S., Drapier, D., and Wollman, F.-A. (2002). Searching
limiting steps in the expression of chloroplast-encoded proteins:
relations between gene copy number, transcription, transcript
abundance and translation rate in the chloroplast of Chlamydomonas reinhardtii. Plant J. 31: 149–160.
Elrad, D., and Grossman, A.R. (2004). A genome’s-eye view of the
light-harvesting polypeptides of Chlamydomonas reinhardtii. Curr.
Genet. 45: 61–75.
Endo, M., Shimizu, H., Nohales, M.A., Araki, T., and Kay, S.A.
(2014). Tissue-specific clocks in Arabidopsis show asymmetric
coupling. Nature 515: 419–422.
Fang, S.-C., de los Reyes, C., and Umen, J.G. (2006). Cell size
checkpoint control by the retinoblastoma tumor suppressor pathway. PLoS Genet. 2: e167.
Fang, W., Si, Y., Douglass, S., Casero, D., Merchant, S.S.,
Pellegrini, M., Ladunga, I., Liu, P., and Spalding, M.H. (2012).
Transcriptome-wide changes in Chlamydomonas reinhardtii gene
expression regulated by carbon dioxide and the CO2-concentrating
mechanism regulator CIA5/CCM1. Plant Cell 24: 1876–1893.
Gaffal, K.P., Arnold, C.G., Friedrichs, G.J., and Gemple, W. (1995).
Morphodynamical changes of the chloroplast of Chlamydomonas reinhardtii during the 1st round of division. Arch. Protistenkunde 145: 10–23.
Gfeller, R.P., and Gibbs, M. (1984). Fermentative metabolism of
Chlamydomonas reinhardtii: I. Analysis of fermentative products
from starch in dark and light. Plant Physiol. 75: 212–218.
Goldschmidt-Clermont, M., and Rahire, M. (1986). Sequence, evolution and differential expression of the two genes encoding variant
small subunits of ribulose bisphosphate carboxylase/oxygenase in
Chlamydomonas reinhardtii. J. Mol. Biol. 191: 421–432.
González-Ballester, D., Casero, D., Cokus, S., Pellegrini, M.,
Merchant, S.S., and Grossman, A.R. (2010). RNA-seq analysis of
sulfur-deprived Chlamydomonas cells reveals aspects of acclimation critical for cell survival. Plant Cell 22: 2058–2084.
Goodenough, U., et al. (2014). The path to triacylglyceride obesity in
the sta6 strain of Chlamydomonas reinhardtii. Eukaryot. Cell 13:
591–613.
Goodenough, U.W. (1970). Chloroplast division and pyrenoid formation in Chlamydomonas reinhardi. J. Phycol. 6: 1–6.
Gómez-Herreros, F., Rodríguez-Galán, O., Morillo-Huesca, M.,
Maya, D., Arista-Romero, M., de la Cruz, J., Chávez, S., and
Muñoz-Centeno, M.C. (2013). Balanced production of ribosome
components is required for proper G1/S transition in Saccharomyces cerevisiae. J. Biol. Chem. 288: 31689–31700.
Harris, E.H. (1989). The Chlamydomonas Sourcebook. (San Diego,
CA: Academic Press).
Harris, E.H. (2001). Chlamydomonas as a model organism. Annu. Rev.
Plant Physiol. Plant Mol. Biol. 52: 363–406.
Hayashi, Y., Sato, N., Shinozaki, A., and Watanabe, M. (2015). Increase in peroxisome number and the gene expression of putative
glyoxysomal enzymes in Chlamydomonas cells supplemented with
acetate. J. Plant Res. 128: 177–185.
Heddad, M., and Adamska, I. (2002). The evolution of light stress
proteins in photosynthetic organisms. Comp. Funct. Genomics 3:
504–510.
Heinnickel, M.L., and Grossman, A.R. (2013). The GreenCut: reevaluation of physiological role of previously studied proteins and
potential novel protein functions. Photosynth. Res. 116: 427–436.
Herrin, D.L., Michaels, A.S., and Paul, A.L. (1986). Regulation of genes
encoding the large subunit of ribulose-1,5-bisphosphate carboxylase
and the photosystem II polypeptides D-1 and D-2 during the cell cycle of
Chlamydomonas reinhardtii. J. Cell Biol. 103: 1837–1845.
Howell, S.H., and Walker, L.L. (1977). Transcription of the nuclear
and chloroplast genomes during the vegetative cell cycle in Chlamydomonas reinhardi. Dev. Biol. 56: 11–23.
Hu, Y., Chen, Z.-W., Liu, W.-Z., Liu, X.-L., and He, Y.-K. (2008).
Chloroplast division is regulated by the circadian expression of
FTSZ and MIN genes in Chlamydomonas reinhardtii. Eur. J. Phycol.
43: 207–215.
Hughes, M.E., Hogenesch, J.B., and Kornacker, K. (2010).
JTK_CYCLE: an efficient nonparametric algorithm for detecting
rhythmic components in genome-scale data sets. J. Biol. Rhythms
25: 372–380.
Hutin, C., Nussaume, L., Moise, N., Moya, I., Kloppstech, K., and
Havaux, M. (2003). Early light-induced proteins protect Arabidopsis
from photooxidative stress. Proc. Natl. Acad. Sci. USA 100: 4921–4926.
Hwang, S., and Herrin, D. (1994). Control of lhc gene transcription by
the circadian clock in Chlamydomonas reinhardtii. Plant Mol. Biol.
26: 557–569.
Hwang, S., Kawazoe, R., and Herrin, D.L. (1996). Transcription of
tufA and other chloroplast-encoded genes is controlled by a circadian clock in Chlamydomonas. Proc. Natl. Acad. Sci. USA 93: 996–
1000.
Idoine, A.D., Boulouis, A., Rupprecht, J., and Bock, R. (2014). The
diurnal logic of the expression of the chloroplast genome in Chlamydomonas reinhardtii. PLoS One 9: e108760.
Jin, J., Zhang, H., Kong, L., Gao, G., and Luo, J. (2014). PlantTFDB
3.0: a portal for the functional and evolutionary study of plant
transcription factors. Nucleic Acids Res. 42: D1182–D1187.
Jinkerson, R.E., and Jonikas, M.C. (2015). Molecular techniques to
interrogate and edit the Chlamydomonas nuclear genome. Plant J.
82: 393–412.
Johnson, C.H., Stewart, P.L., and Egli, M. (2011). The cyanobacterial
circadian system: from biophysics to bioevolution. Annu. Rev. Biophys. 40: 143–167.
Chlamydomonas Diurnal Transcriptome
Johnson, U.G., and Porter, K.R. (1968). Fine structure of cell division
in Chlamydomonas reinhardi. Basal bodies and microtubules. J. Cell
Biol. 38: 403–425.
Johnson, X., and Alric, J. (2013). Central carbon metabolism and
electron transport in Chlamydomonas reinhardtii: metabolic constraints for carbon partitioning between oil and starch. Eukaryot.
Cell 12: 776–793.
Jorgensen, P., Rupes, I., Sharom, J.R., Schneper, L., Broach, J.R.,
and Tyers, M. (2004). A dynamic transcriptional network communicates growth potential to ribosome synthesis and critical cell size.
Genes Dev. 18: 2491–2505.
Kanesaki, Y., Imamura, S., Minoda, A., and Tanaka, K. (2012). External light conditions and internal cell cycle phases coordinate
accumulation of chloroplast and mitochondrial transcripts in the red
alga Cyanidioschyzon merolae. DNA Res. 19: 289–303.
Karpowicz, S.J., Prochnik, S.E., Grossman, A.R., and Merchant,
S.S. (2011). The GreenCut2 resource, a phylogenomically derived
inventory of proteins specific to the plant lineage. J. Biol. Chem.
286: 21427–21439.
Keller, L.C., Romijn, E.P., Zamora, I., Yates III, J.R., and Marshall,
W.F. (2005). Proteomic analysis of isolated chlamydomonas centrioles reveals orthologs of ciliary-disease genes. Curr. Biol. 15:
1090–1098.
Kinmonth-Schultz, H.A., Golembeski, G.S., and Imaizumi, T. (2013).
Circadian clock-regulated physiological outputs: dynamic responses in
nature. Semin. Cell Dev. Biol. 24: 407–413.
Klein, U. (1987). Intracellular Carbon Partitioning in Chlamydomonas
reinhardtii. Plant Physiol. 85: 892–897.
Klein, U. (2008). Chloroplast transcription. In The Chlamydomonas
Sourcebook: Organellar and Metabolic Processes, E.H. Harris and
D.B. Stern, eds (London: Academic Press), pp. 893–914.
Kobayashi, T., and Dynlacht, B.D. (2011). Regulating the transition
from centriole to basal body. J. Cell Biol. 193: 435–444.
Kubo, T., Kaida, S., Abe, J., Saito, T., Fukuzawa, H., and Matsuda,
Y. (2009). The Chlamydomonas hatching enzyme, sporangin, is
expressed in specific phases of the cell cycle and is localized to the
flagella of daughter cells within the sporangial cell wall. Plant Cell
Physiol. 50: 572–583.
Kucho, K., Okamoto, K., Tabata, S., Fukuzawa, H., and Ishiura, M.
(2005). Identification of novel clock-controlled genes by cDNA
macroarray analysis in Chlamydomonas reinhardtii. Plant Mol. Biol.
57: 889–906.
Larson, D.E., Zahradka, P., and Sells, B.H. (1991). Control points in
eucaryotic ribosome biogenesis. Biochem. Cell Biol. 69: 5–22.
Lechtreck, K.-F., Johnson, E.C., Sakai, T., Cochran, D., Ballif, B.A.,
Rush, J., Pazour, G.J., Ikebe, M., and Witman, G.B. (2009). The
Chlamydomonas reinhardtii BBSome is an IFT cargo required for
export of specific signaling proteins from flagella. J. Cell Biol. 187:
1117–1132.
Lee, J., and Herrin, D.L. (2002). Assessing the relative importance of
light and the circadian clock in controlling chloroplast translation in
Chlamydomonas reinhardtii. Photosynth. Res. 72: 295–306.
Lien, T., and Knutsen, G. (1979). Synchronous growth of Chlamydomonas reinhardtii (Chlorophyceae): a review of optimal conditions. J. Phycol. 15: 191–200.
Lindström, M.S. (2009). Emerging functions of ribosomal proteins in
gene-specific transcription and translation. Biochem. Biophys. Res.
Commun. 379: 167–170.
Love, M.I., Huber, W., and Anders, S. (2014). Moderated estimation
of fold change and dispersion for RNA-seq data with DESeq2.
Genome Biol. 15: 550.
Mager, W.H. (1988). Control of ribosomal protein gene expression.
Biochim. Biophys. Acta 949: 1–15.
25 of 27
Malasarn, D., Kropat, J., Hsieh, S.I., Finazzi, G., Casero, D., Loo,
J.A., Pellegrini, M., Wollman, F.-A., and Merchant, S.S. (2013).
Zinc deficiency impacts CO2 assimilation and disrupts copper homeostasis in Chlamydomonas reinhardtii. J. Biol. Chem. 288:
10672–10683.
Matsuo, T., and Ishiura, M. (2010). New insights into the circadian
clock in Chlamydomonas. Int. Rev. Cell Mol. Biol. 280: 281–314.
Matsuo, T., Okamoto, K., Onai, K., Niwa, Y., Shimogawara, K., and
Ishiura, M. (2008). A systematic forward genetic analysis identified
components of the Chlamydomonas circadian system. Genes Dev.
22: 918–930.
McIntosh, K.B., and Bonham-Smith, P.C. (2006). Ribosomal protein
gene regulation: what about plants? Can. J. Bot. 84: 342–362.
Merchant, S.S., et al. (2007). The Chlamydomonas genome reveals
the evolution of key animal and plant functions. Science 318: 245–
250.
Michael, T.P., et al. (2008). Network discovery pipeline elucidates
conserved time-of-day-specific cis-regulatory modules. PLoS
Genet. 4: e14.
Mitchell, M.C., Meyer, M.T., and Griffiths, H. (2014). Dynamics of
carbon-concentrating mechanism induction and protein relocalization during the dark-to-light transition in synchronized Chlamydomonas reinhardtii. Plant Physiol. 166: 1073–1082.
Mittag, M., Kiaulehn, S., and Johnson, C.H. (2005). The circadian
clock in Chlamydomonas reinhardtii. What is it for? What is it similar
to? Plant Physiol. 137: 399–409.
Miyagishima, S.Y., Suzuki, K., Okazaki, K., and Kabeya, Y. (2012).
Expression of the nucleus-encoded chloroplast division genes and
proteins regulated by the algal cell cycle. Mol. Biol. Evol. 29: 2957–
2970.
Monnier, A., Liverani, S., Bouvet, R., Jesson, B., Smith, J.Q., Mosser,
J., Corellou, F., and Bouget, F.-Y. (2010). Orchestrated transcription of
biological processes in the marine picoeukaryote Ostreococcus exposed to light/dark cycles. BMC Genomics 11: 192.
Moulager, M., Corellou, F., Vergé, V., Escande, M.-L., and Bouget,
F.-Y. (2010). Integration of light signals by the retinoblastoma
pathway in the control of S phase entry in the picophytoplanktonic
cell Ostreococcus. PLoS Genet. 6: e1000957.
Morgan, D.O. (2007). The Cell Cycle, E. Lawrence, ed (London: New
Science Press).
Natali, A., and Croce, R. (2015). Characterization of the major lightharvesting complexes (LHCBM) of the green alga Chlamydomonas
reinhardtii. PLoS One 10: e0119211.
Noordally, Z.B., and Millar, A.J. (2015). Clocks in algae. Biochemistry
54: 171–183.
Oey, M., Ross, I.L., Stephens, E., Steinbeck, J., Wolf, J., Radzun,
K.A., Kügler, J., Ringsmuth, A.K., Kruse, O., and Hankamer, B.
(2013). RNAi knock-down of LHCBM1, 2 and 3 increases photosynthetic H2 production efficiency of the green alga Chlamydomonas reinhardtii. PLoS One 8: e61375.
Ohno, S. (1970). Evolution by Gene Duplication. (Berlin: Springer).
Oldenhof, H., Bisová, K., van den Ende, H., and Zachleder, V.
(2004). Effect of red and blue light on the timing of cyclin-dependent
kinase activity and the timing of cell division in Chlamydomonas
reinhardtii. Plant Physiol. Biochem. 42: 341–348.
Olson, B.J.S.C., Oberholzer, M., Li, Y., Zones, J.M., Kohli, H.S.,
Bisová, K., Fang, S.-C., Meisenhelder, J., Hunter, T., and Umen,
J.G. (2010). Regulation of the Chlamydomonas cell cycle by a stable, chromatin-associated retinoblastoma tumor suppressor complex. Plant Cell 22: 3331–3347.
Ostrowski, L.E., Dutcher, S.K., and Lo, C.W. (2011). Cilia and
models for studying structure and function. Proc. Am. Thorac. Soc.
8: 423–429.
26 of 27
The Plant Cell
Panchy, N., Wu, G., Newton, L., Tsai, C.-H., Chen, J., Benning, C.,
Farré, E.M., and Shiu, S.-H. (2014). Prevalence, evolution, and cisregulation of diel transcription in Chlamydomonas reinhardtii. G3
(Bethesda) 4: 2461–2471.
Parker, J.D., Hilton, L.K., Diener, D.R., Rasi, M.Q., Mahjoub, M.R.,
Rosenbaum, J.L., and Quarmby, L.M. (2010). Centrioles are freed
from cilia by severing prior to mitosis. Cytoskeleton (Hoboken) 67:
425–430.
Pazour, G.J., Agrin, N., Leszyk, J., and Witman, G.B. (2005). Proteomic analysis of a eukaryotic cilium. J. Cell Biol. 170: 103–113.
Peers, G., Truong, T.B., Ostendorf, E., Busch, A., Elrad, D.,
Grossman, A.R., Hippler, M., and Niyogi, K.K. (2009). An ancient
light-harvesting protein is critical for the regulation of algal photosynthesis. Nature 462: 518–521.
Pérez-Rodríguez, P., Riaño-Pachón, D.M., Corrêa, L.G.G.,
Rensing, S.A., Kersten, B., and Mueller-Roeber, B. (2010).
PlnTFDB: updated content and new features of the plant transcription factor database. Nucleic Acids Res. 38: D822–D827.
Perry, R.P. (2007). Balanced production of ribosomal proteins. Gene
401: 1–3.
Plotnikova, O.V., Pugacheva, E.N., and Golemis, E.A. (2009). Primary cilia and the cell cycle. Methods Cell Biol. 94: 137–160.
Pracharoenwattana, I., Cornah, J.E., and Smith, S.M. (2005).
Arabidopsis peroxisomal citrate synthase is required for fatty acid
respiration and seed germination. Plant Cell 17: 2037–2048.
Prochnik, S.E., et al. (2010). Genomic analysis of organismal complexity in
the multicellular green alga Volvox carteri. Science 329: 223–226.
Qin, H., Wang, Z., Diener, D., and Rosenbaum, J. (2007). Intraflagellar Transport Protein 27 is a small G protein involved in cellcycle control. Curr. Biol. 17: 193–202.
Ral, J.-P., Colleoni, C., Wattebled, F., Dauvillée, D., Nempont, C.,
Deschamps, P., Li, Z., Morell, M.K., Chibbar, R., Purton, S.,
d’Hulst, C., and Ball, S.G. (2006). Circadian clock regulation of
starch metabolism establishes GBSSI as a major contributor to
amylopectin synthesis in Chlamydomonas reinhardtii. Plant Physiol.
142: 305–317.
Rau, A., Gallopin, M., Celeux, G., and Jaffrézic, F. (2013). Databased filtering for replicated high-throughput transcriptome sequencing experiments. Bioinformatics 29: 2146–2152.
Recuenco-Muñoz, L., Offre, P., Valledor, L., Lyon, D., Weckwerth,
W., and Wienkoop, S. (2015). Targeted quantitative analysis of
a diurnal RuBisCO subunit expression and translation profile in
Chlamydomonas reinhardtii introducing a novel Mass Western approach. J. Proteomics 113: 143–153.
Reddy, A.B., and Rey, G. (2014). Metabolic and nontranscriptional
circadian clocks: eukaryotes. Annu. Rev. Biochem. 83: 165–189.
Robbens, S., Khadaroo, B., Camasses, A., Derelle, E., Ferraz, C.,
Inzé, D., Van de Peer, Y., and Moreau, H. (2005). Genome-wide
analysis of core cell cycle genes in the unicellular green alga Ostreococcus tauri. Mol. Biol. Evol. 22: 589–597.
Saeed, A.I., et al. (2003). TM4: a free, open-source system for microarray
data management and analysis. Biotechniques 34: 374–378.
Salvador, M.L., Klein, U., and Bogorad, L. (1993). Light-regulated
and endogenous fluctuations of chloroplast transcript levels in
Chlamydomonas. Regulation by transcription and RNA degradation. Plant J. 3: 213–219.
Samartzidou, H., and Widger, W.R. (1998). Transcriptional and
posttranscriptional control of mRNA from lrtA, a light-repressed
transcript in Synechococcus sp. PCC 7002. Plant Physiol. 117:
225–234.
Scaife, M.A., Nguyen, G.T.D.T., Rico, J., Lambert, D., Helliwell,
K.E., and Smith, A.G. (2015). Establishing Chlamydomonas reinhardtii as an industrial biotechnology host. Plant J. 82: 532–546.
Schmollinger, S., et al. (2014). Nitrogen-sparing mechanisms in
Chlamydomonas affect the transcriptome, the proteome, and
photosynthetic metabolism. Plant Cell 26: 1410–1435.
Schroda, M., and Vallon, O. (2009). Chaperones and proteases. In
The Chlamydomonas Sourcebook, E.H. Harris and D.B. Stern, eds
(London: Academic Press), pp. 671–729.
Serrano, G., Herrera-Palau, R., Romero, J.M., Serrano, A.,
Coupland, G., and Valverde, F. (2009). Chlamydomonas CONSTANS
and the evolution of plant photoperiodic signaling. Curr. Biol. 19: 359–
368.
Shiratsuchi, G., Kamiya, R., and Hirono, M. (2011). Scaffolding
function of the Chlamydomonas procentriole protein CRC70,
a member of the conserved Cep70 family. J. Cell Sci. 124: 2964–
2975.
Soukas, A., Cohen, P., Socci, N.D., and Friedman, J.M. (2000).
Leptin-specific patterns of gene expression in white adipose tissue.
Genes Dev. 14: 963–980.
Specht, E., Miyake-Stoner, S., and Mayfield, S. (2010). Micro-algae
come of age as a platform for recombinant protein production.
Biotechnol. Lett. 32: 1373–1383.
Swirsky Whitney, L.A., Novi, G., Perata, P., and Loreti, E. (2012).
Distinct mechanisms regulating gene expression coexist within
the fermentative pathways in Chlamydomonas reinhardtii. ScientificWorldJournal 2012: 565047.
Tan, X., Varughese, M., and Widger, W.R. (1994). A light-repressed
transcript found in Synechococcus PCC 7002 is similar to a chloroplast-specific small subunit ribosomal protein and to a transcription modulator protein associated with sigma 54. J. Biol. Chem.
269: 20905–20912.
Tardif, M., Atteia, A., Specht, M., Cogne, G., Rolland, N., Brugière, S.,
Hippler, M., Ferro, M., Bruley, C., Peltier, G., Vallon, O., and Cournac,
L. (2012). PredAlgo: a new subcellular localization prediction tool dedicated to green algae. Mol. Biol. Evol. 29: 3625–3639.
Teramoto, H., Itoh, T., and Ono, T.-A. (2004). High-intensity-lightdependent and transient expression of new genes encoding distant
relatives of light-harvesting chlorophyll-a/b proteins in Chlamydomonas reinhardtii. Plant Cell Physiol. 45: 1221–1232.
Thimm, O., Bläsing, O., Gibon, Y., Nagel, A., Meyer, S., Krüger, P.,
Selbig, J., Müller, L.A., Rhee, S.Y., and Stitt, M. (2004). MAPMAN:
a user-driven tool to display genomics data sets onto diagrams of
metabolic pathways and other biological processes. Plant J. 37:
914–939.
Thines, B., and Harmon, F.G. (2011). Four easy pieces: mechanisms
underlying circadian regulation of growth and development. Curr.
Opin. Plant Biol. 14: 31–37.
Thyssen, C., Schlichting, R., and Giersch, C. (2001). The CO2concentrating mechanism in the physiological context: lowering the
CO2 supply diminishes culture growth and economises starch utilisation in Chlamydomonas reinhardtii. Planta 213: 629–639.
Tirumani, S., Kokkanti, M., Chaudhari, V., Shukla, M., and Rao, B.J.
(2014). Regulation of CCM genes in Chlamydomonas reinhardtii
during conditions of light-dark cycles in synchronous cultures. Plant
Mol. Biol. 85: 277–286.
Tulin, F., and Cross, F.R. (2014). A microbial avenue to cell cycle
control in the plant superkingdom. Plant Cell 26: 4019–4038.
Umen, J.G., and Goodenough, U.W. (2001). Control of cell division
by a retinoblastoma protein homolog in Chlamydomonas. Genes
Dev. 15: 1652–1661.
Umen, J.G., and Olson, B.J.S.C. (2012). Genomics of volvocine algae. In Genomic Insights into the Biology of Algae, G. Piganeau, ed
(Elsevier), pp. 185–243.
Urzica, E.I., Adler, L.N., Page, M.D., Linster, C.L., Arbing, M.A.,
Casero, D., Pellegrini, M., Merchant, S.S., and Clarke, S.G.
Chlamydomonas Diurnal Transcriptome
(2012a). Impact of oxidative stress on ascorbate biosynthesis in
Chlamydomonas via regulation of the VTC2 gene encoding a GDPL-galactose phosphorylase. J. Biol. Chem. 287: 14234–14245.
Urzica, E.I., Casero, D., Yamasaki, H., Hsieh, S.I., Adler, L.N.,
Karpowicz, S.J., Blaby-Haas, C.E., Clarke, S.G., Loo, J.A.,
Pellegrini, M., and Merchant, S.S. (2012b). Systems and transsystem level analysis identifies conserved iron deficiency responses
in the plant lineage. Plant Cell 24: 3921–3948.
Wang, D., Kong, D., Wang, Y., Hu, Y., He, Y., and Sun, J. (2003).
Isolation of two plastid division ftsZ genes from Chlamydomonas
reinhardtii and its evolutionary implication for the role of FtsZ in
plastid division. J. Exp. Bot. 54: 1115–1116.
Wei, H., Persson, S., Mehta, T., Srinivasasainagendra, V., Chen, L.,
Page, G.P., Somerville, C., and Loraine, A. (2006). Transcriptional
coordination of the metabolic network in Arabidopsis. Plant Physiol.
142: 762–774.
Willamme, R., Alsafra, Z., Arumugam, R., Eppe, G., Remacle, F.,
Levine, R.D., and Remacle, C. (2015). Metabolomic analysis of the
green microalga Chlamydomonas reinhardtii cultivated under day/
night conditions. J. Biotechnol. pii: S0168-1656(15)00189-3.
Wilson, R., and Chiang, K.S. (1977). Temporal programming of
chloroplast and cytoplasmic ribosomal RNA transcription in the
synchronous cell cycle of Chlamydomonas reinhardtii. J. Cell Biol.
72: 470–481.
27 of 27
Wittenberg, C., and Reed, S.I. (2005). Cell cycle-dependent transcription in yeast: promoters, transcription factors, and transcriptomes. Oncogene 24: 2746–2755.
Wood, C.R., Wang, Z., Diener, D., Zones, J.M., Rosenbaum, J., and
Umen, J.G. (2012). IFT proteins accumulate during cell division and localize to the cleavage furrow in Chlamydomonas. PLoS One 7: e30729.
Xie, Z., Culler, D., Dreyfuss, B.W., Kuras, R., Wollman, F.A., GirardBascou, J., and Merchant, S. (1998). Genetic analysis of chloroplast c-type cytochrome assembly in Chlamydomonas reinhardtii:
One chloroplast locus and at least four nuclear loci are required for
heme attachment. Genetics 148: 681–692.
Yamaguchi, K., Beligni, M.V., Prieto, S., Haynes, P.A., McDonald,
W.H., Yates III, J.R., and Mayfield, S.P. (2003). Proteomic characterization of the Chlamydomonas reinhardtii chloroplast ribosome. Identification of proteins unique to th e70 S ribosome. J. Biol.
Chem. 278: 33774–33785.
Yeung, K.Y., Haynor, D.R., and Ruzzo, W.L. (2001). Validating
clustering for gene expression data. Bioinformatics 17: 309–318.
Zachariae, W., and Nasmyth, K. (1999). Whose end is destruction: cell division and the anaphase-promoting complex. Genes Dev. 13: 2039–2058.
Zerges, W., and Hauser, C. (2009). Protein synthesis in the chloroplast. In The Chlamydomonas Sourcebook: Organellar and Metabolic Processes, E.H. Harris and D.B. Stern, eds (London: Academic
Press), pp. 967–1025.
High-Resolution Profiling of a Synchronized Diurnal Transcriptome from Chlamydomonas
reinhardtii Reveals Continuous Cell and Metabolic Differentiation
James Matt Zones, Ian K. Blaby, Sabeeha S. Merchant and James G. Umen
Plant Cell; originally published online October 2, 2015;
DOI 10.1105/tpc.15.00498
This information is current as of July 31, 2017
Supplemental Data
/content/suppl/2015/09/16/tpc.15.00498.DC1.html
Permissions
https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298X
eTOCs
Sign up for eTOCs at:
http://www.plantcell.org/cgi/alerts/ctmain
CiteTrack Alerts
Sign up for CiteTrack Alerts at:
http://www.plantcell.org/cgi/alerts/ctmain
Subscription Information
Subscription Information for The Plant Cell and Plant Physiology is available at:
http://www.aspb.org/publications/subscriptions.cfm
© American Society of Plant Biologists
ADVANCING THE SCIENCE OF PLANT BIOLOGY