Abstract
The apocarotenoid zaxinone promotes growth and suppresses strigolactone
biosynthesis in rice. To shed light on the mechanisms underlying its
growth-promoting effect, we employed a combined omics approach
integrating transcriptomics and metabolomics analysis of rice seedlings
treated with zaxinone, and determined the resulting changes at the
cellular and hormonal levels. Metabolites as well as transcripts
analysis demonstrate that zaxinone application increased sugar content
and triggered glycolysis, the tricarboxylic acid cycle and other
sugar-related metabolic processes in rice roots. In addition, zaxinone
treatment led to an increased root starch content and induced
glycosylation of cytokinins. The transcriptomic, metabolic and hormonal
changes were accompanied by striking alterations of roots at cellular
level, which showed an increase in apex length, diameter, and the
number of cells and cortex cell layers. Remarkably, zaxinone did not
affect the metabolism of roots in a strigolactone deficient mutant,
suggesting an essential role of strigolactone in the zaxinone
growth-promoting activity. Taken together, our results unravel zaxinone
as a global regulator of the transcriptome and metabolome, as well as
of hormonal and cellular composition of rice roots. Moreover, they
suggest that zaxinone promotes rice growth most likely by increasing
sugar uptake and metabolism, and reinforce the potential of this
compound in increasing rice performance.
Subject terms: Plant hormones, Plant development, Plant physiology
__________________________________________________________________
Wang et al. report zaxinone as a global regulator of the transcriptome
and metabolome, as well as of hormonal and cellular composition of rice
roots. This study shows that zaxinone promotes rice growth by enhancing
root sugar uptake and metabolism and modulation of cytokinin content,
indicating the potential application of this compound in increasing
rice performance.
Introduction
Carotenoids are widespread pigments fulfilling vital functions in
plants, by protecting the photosynthetic apparatus from photo-oxidation
and harnessing light energy^[54]1. In addition, they are the precursor
of a structurally diverse set of metabolites, generally called
apocarotenoids, which include volatiles, colorants,
signaling/regulatory molecules, and hormones. Apocarotenoids arise
through oxidative break down of carotenoids, which is initiated by
reactive oxygen species (ROS) attack or catalyzed by Carotenoid
Cleavage Dioxygenases (CCDs), an evolutionarily conserved family of
non-heme Fe^2+-dependent enzymes^[55]2–[56]4. The primary cleavage
products are frequently modified by different enzymes, before acquiring
their biological function. For instance, the apocarotenoid plant
hormone strigolactone (SL) is formed by the sequential action of CCD7,
CCD8, more axillary growth1 (MAX1, a cytochrome P450 monooxygenase),
and other enzymes^[57]4–[58]8. SLs and abscisic acid (ABA), a further
carotenoid-derived plant hormone, are key metabolites in establishing
plant’s response to abiotic and biotic stress, and major determinants
of plant development^[59]4,[60]9. Besides, SLs modulate plant’s
architecture in response to nutrients availability, particularly
phosphorus, and mediate, when released in the rhizosphere, the
communication with arbuscular mycorrhizal fungi that supply plants with
water and minerals^[61]10,[62]11. However, SLs are also perceived by
seeds of parasitic plants, such as Striga, which use them as
germination stimulus ensuring the availability of a host required for
the survival of these obligate parasitic weeds^[63]5. To fulfill their
role in plant growth and development, SLs are embedded in a complex
hormonal network in which they affect and are influenced by the
activity of other plant hormones. Indeed, auxin and gibberellins were
reported to interact with SL biosynthesis or signaling in rice and
Arabidopsis^[64]12. Vice versa, it was shown that SLs enhance
cytokinins (CKs) catabolism by modulating CYTOKININ
OXIDASE/DEHYDROGENASE 9 (CKX9) expression to inhibit rice
tillering^[65]13. Similarly, CKs and SLs exert opposite effects on rice
mesocotyl elongation^[66]14,[67]15.
Metabolism is a central process required for the uptake and utilization
of energy and nutrients to ensure the survival, reproduction, growth,
and development of living organisms^[68]16. Thus, primary metabolites
such as sugars, amino acids, nucleotides, organic acids, and fatty
acids are essential for maintaining cellular homeostasis and for
organismal life^[69]17. In fact, metabolites are direct physiological
signatures that are highly correlated with end-phenomes in
plants^[70]17,[71]18. In addition, some primary metabolites act as
signaling molecules regulating plant growth and development. For
instance sugars, such as sucrose, interact with different plant
hormones and regulate bud development and shoot branching by modulating
the signaling of auxin and SLs^[72]12,[73]19,[74]20.
Besides the established plant hormones ABA and SLs, the apocarotenoid
family includes growth regulators, such as anchorene that specifically
promotes the growth of anchor roots in Arabidopsis^[75]21, and
signaling molecules, such as cyclocitral that mediates the response of
plants to high-light and drought stress, and regulates roots
growth^[76]4,[77]22. Recently, we have identified zaxinone, an
apocarotenoid hormone candidate, as a metabolite required for proper
rice growth and development, and characterized a rice CCD, called
ZAXINONE SYNTHASE (ZAS), involved in its biosynthesis^[78]23. A rice
loss-of-function zas mutant showed shoot and root growth impairment, a
lower root zaxinone level, and higher SL content in roots and root
exudates. Exogenous application of zaxinone rescued several zas
phenotypes and resulted in a decrease in SL content and release, and in
promoted root growth^[79]23,[80]24. Treatment of WT seedlings with
zaxinone led also to an obvious increase in root growth and a
suppression of SL formation^[81]23,[82]24. Transcript analysis showed
that zaxinone is a negative regulator of SL biosynthesis at the
transcript level. However, zaxinone application did not enhance root
growth in SL biosynthesis and perception rice mutants, which indicates
an interaction between zaxinone and SLs and suggests the requirement of
a functional SL pathway for zaxinone’s growth-promoting
activity^[83]23.
In the current study, we set out to understand how zaxinone promotes
rice growth. For this purpose, we characterized its effect on rice
primary metabolism and transcriptome in WT and the SL-deficient d17
mutant rice plants, and determined zaxinone’s impact on hormone content
and root anatomy. Our results unraveled enhanced root sugar metabolism
as a major reason for zaxinone growth-promoting activity and point to
modulation of cytokinin content as a likely reason for increased root
cell division activity and enhanced number of cortex cell layers, which
we observed in roots.
Results
Zaxinone treatment increases sugar content and metabolism
To determine the effects of zaxinone on rice at a metabolomic and
transcriptomic level and to get an insight into the dynamics of
triggered changes, we grew rice seedlings hydroponically, applied the
compound at a 5 µM concentration into the growth medium, and collected
root and shoot samples 2, 6 and 24 h after application. A scheme of the
experimental design is shown in Supplementary Fig. [84]1. Gas
chromatography–mass spectrometry (GC–MS) analysis of primary
metabolites in treated roots revealed an up to a 1.5-fold increase in
the level of many sugars, glycolysis- and tricarboxylic acid
(TCA)-cycle intermediates, such as glucose, citric acid, and
2-oxoglutarate (Fig. [85]1a). The position of these metabolites in
cellular sugar catabolism pathways is shown in Fig. [86]1b. At the same
time point, i.e., 6 h after zaxinone application, we also observed an
increase in the content of most of the free amino acids and many of
other analyzed primary metabolites. However, the content of some amino
acids, e.g., leucine, dropped at the 24 h time point to below control
level (Fig. [87]1a). Principal component analysis (PCA) revealed that
the zaxinone effect on root metabolome was more pronounced at 6 and
24 h, compared with the early 2 h time point, with a peak of primary
metabolites accumulation at 6 h (Fig. [88]1a and Supplementary
Figs. [89]2a, Supplementary Fig. [90]3). In more detail, we observed a
significant increase in the levels of the major sugars sucrose,
glucose, and fructose at 2 and 6 h after application of zaxinone, which
was followed by a sharp decrease at 24 h. In contrast, trehalose showed
the highest increase after 24 h, similar to maltose,
glucose-1-phosphate, and glucose-6-phosphate, which indicates a
biphasic response to zaxinone application. In shoot tissues, we also
detected a quick enhancement in the level of sucrose and trehalose 2 h
after zaxinone application. However, many other primary metabolites,
including several sugars, TCA intermediates, and free amino acids,
showed an increase only at the late, 24 h, time point (Fig. [91]1d and
Supplementary Figs. [92]2b, [93]4). These results suggest that
zaxinone application to roots in hydroponically grown seedlings causes
a rapid global change in the primary metabolism of roots and, with a
slight delay, in shoots. In particular, it affects sugar content and
catabolism, which are essential for adenosine triphosphate (ATP)
generation and supply of C skeletons for cellular building blocks.
Fig. 1. The dynamic primary metabolite profiles upon zaxinone treatment.
[94]Fig. 1
[95]Open in a new tab
Primary metabolites extracted from roots for GC–MS, which were
annotated and listed in Supplementary Data 6. Three independent
harvests were concurrently subjected to GC–MS analysis. n ≥ 4
biological replicates. a Heatmap of root tissue showing relative
accumulation of each metabolite as compared to those in control plants.
For each metabolite, the value of the corresponding wild type was set
to 1. Asterisks indicate statistically significant differences as
compared to control by t-test (*p < 0.05, **p < 0.01). b The scheme of
major metabolic changes in the sugar-related metabolites after zaxinone
treatment, which adapted from a. Blue and red color depict a decrease
and increase in metabolic levels compared to the non-treated root
samples, respectively. The data are presented as log2 (fold changes)
from left to right as follows: 2, 6, and 24 h. c Longitudinal-sections
of WT and WT treated with 5 µM zaxinone roots tips after
resin-embedding and staining by the Periodic Acid-Schiff (PAS) reaction
for the visualization of amyloplasts. At higher magnification (15 µm)
the statoliths (st) in the root cap, as well as some tiny amyloplasts
(arrows), are present in the cytoplasm of meristematic cells. The
starch level was quantified in the root tissues. Bar presented as
mean ± SD, n = 8 biological replicates. d Heatmap of shoot tissue of
relative accumulated metabolites in comparison with control. For each
metabolite, the value of the corresponding wild type was set to 1.
Asterisks indicate statistically significant differences as compared to
wild type by t-test (*p < 0.05, **p < 0.01). CTL control, Zax zaxinone.
Assuming that excess sugars may be stored as starch, we measured the
starch level in roots of hydroponically grown seedlings after two weeks
of treatment with zaxinone (5 µM). Indeed, we detected around two-fold
higher starch content in treated roots (Fig. [96]1c), compared to the
mock condition. Taking into consideration that plants produce sugars
through photosynthesis, we investigated the effect of zaxinone on this
process. For this purpose, we performed a time-course measurement of
chlorophyll content and stomatal conductance, two parameters of
photosynthetic activity, in a 3-weeks-old hydroponic grown rice plant
treated with 5 µM zaxinone under greenhouse condition. As shown in
Supplementary Fig. [97]5a, b, we observed an enhancement of both
parameters in leaves of treated rice plants, which indicated an
increased photosynthetic activity and may explain the elevated sugar
levels.
Zaxinone application increases transcript level of genes involved in root
sugar metabolism
We also analyzed the impact of zaxinone treatment on rice
transcriptome, using RNA sequencing (RNA-Seq). A heatmap visualization
of mean-centered, normalized log-expression values for correlated
highly variable genes (HVGs) of the RNA-Seq data confirmed the high
quality of each replicate, which was supported by the PCA plots of HVGs
(Supplementary Fig. [98]6). A Volcano plot of differentially expressed
genes (DEGs), following Deseq2 analysis, revealed the gene expression
pattern at different cutting points with log[2]FoldChange, adjusted
Q-value (False discovery rate, FDR), or a combination of both
(Supplementary Fig. [99]7). In order to have a better picture of the
transcriptome, we decided to use the FDR < 0.05 as a criterion for
further analysis (Supplementary Data [100]1). Zaxinone application led
to significant changes in the transcriptome over time, by increasing
the transcript level of 324, 551, and 350 genes after 2, 6, and 24 h,
respectively, including 38 genes that showed an induction at the three
time points. The application also decreased the transcript level of 136
(2 h), 501 (6 h), and 71 (24 h) genes, ten of which were downregulated
at the three time points (Fig. [101]2a, b). To validate the RNA-Seq
data, we determined the transcript level of 15 selected genes that
showed low to high response to zaxinone treatment, by qRT-PCR. The
resulting correlation analysis (R^2 = 0.87–0.93) indicated that the
RNA-Seq dataset was highly reliable and thus appropriate for pathway
enrichment analysis (Supplementary Fig. [102]8). Gene Ontology (GO
term) analysis of molecular function and biological process showed that
most of the genes regulated by zaxinone were related to metabolic
processes [upregulation: 45 (2 h), 108 (6 h), and 50 (24 h) genes;
downregulation: 16 (2 h), 70 (6 h), and 8 (24 h) genes] and catalytic
activity [upregulation: 72 (2 h), 148 (6 h), and 90 (24 h) genes;
downregulation: 36 (2 h), 124 (6 h), and 22 (24 h) genes]
(Supplementary Data [103]2). Further enrichments with Kyoto
Encyclopedia of Genes and Genomes (KEGG) and PlantCyc pathway unraveled
the induction of genes mediating several annotated sugar metabolism
pathways, including pyruvate metabolism, citric acid cycle, sucrose
degradation, glycolysis, and gluconeogenesis, particularly 6 h after
zaxinone application (Fig. [104]2c), which is in line with the change
in the profile of primary metabolites (Fig. [105]1a). We also confirmed
the annotated pathways (the plant glycolytic pathway and the TCA cycle)
by MapMan software (Supplementary Fig. [106]9). To validate these
changes, we chose ten genes from the OyzaCyc 6.0 database, which are
involved in root glycolysis (Supplementary Data [107]3), and validated
their expression pattern by performing qRT-PCR analysis of the same
samples used for the RNA-seq experiment (Supplementary Fig. [108]10).
Results obtained correlated with and explained the increase in sugar
metabolites in root tissues. In contrast to roots, we detected a less
significant impact on the shoot transcriptome presented in the PCA
plots and DEGs analysis (Supplementary Fig. [109]11 and Supplementary
Data [110]4), which may be anticipated given that the effect of
zaxinone was mainly visible in root tissues, when using the hydroponic
system. To further confirm the results, we performed a parallel
transcript analysis with roots of WT and the zas mutant that contains
less zaxinone and displays retarded growth (Supplementary
Fig. [111]12a). We observed an upregulated transcript pattern of five
glycolytic genes following the zaxinone treatment (Supplementary
Fig. [112]12b), which could explain the capability of zaxinone in
rescuing zas phenotype. Taken together, the transcriptome analysis
supported the hypothesis that the growth-promoting effect is strongly
linked with an increase of sugar metabolism in rice roots.
Fig. 2. Analysis of differentially expressed genes (DEGs) in response to
zaxinone at different time points.
[113]Fig. 2
[114]Open in a new tab
a Numbers of the significantly expressed genes upon zaxinone treatment
(FDR < 0.05). The numbers on the vertical axis represent the three time
points while the horizontal axis reflects the numbers of up- and
down-regulated genes. Up- and down-regulated genes are shown in red and
blue bars, respectively. b Venn diagrams showing the numbers of down
(Dn) and upregulated (Up) genes that overlap between different time
points. c Kyoto Encyclopedia of Genes and Genomes (KEGG) and PlantCyc
pathway enrichment analysis for up- and down-regulated genes, which
were analyzed by The Plant GeneSet Enrichment Analysis Toolkit
(PlantGSEA)
([115]http://structuralbiology.cau.edu.cn/PlantGSEA/index.php).
Zaxinone application does not induce sugar metabolism in the absence of
strigolactones
In our previous study, we showed that zaxinone application did not
promote root growth in rice SL biosynthesis and perception mutants,
indicating that zaxinone’s growth-promoting effect depends on
functional SL biosynthesis and perception^[116]23. This opened the
question of whether the changes in sugar metabolism caused by zaxinone
are also linked to SLs. To answer this question, we applied zaxinone to
d17 and zas mutants and the corresponding WT varieties for 6 h,
following the experimental design shown in Supplementary Fig [117]1,
and analyzed the metabolome of collected root tissues. Results of
metabolome analysis confirmed the accumulation of sugars and TCA cycle
metabolites upon zaxinone treatment in both WT and zas mutants, while
this response was largely absent in the d17 mutant (Fig. [118]3),
demonstrating that the sensitivity of sugar metabolism towards zaxinone
application depends on the presence of a functional SL biosynthetic
pathway.
Fig. 3. The profile of primary metabolites in roots of WT, zas, and d17
mutant.
[119]Fig. 3
[120]Open in a new tab
Primary metabolites extracted from roots for GC-MS, which were
annotated and listed in Supplementary Data 7. n = 4 biological
replicates. a Heatmap of root tissue showing relative accumulation of
each metabolite as compared to those in control plants. For each
metabolite, the value of the corresponding control was set to 1.
Asterisks indicate statistically significant differences as compared to
control by t-test (*p < 0.05, **p < 0.01). b Principal component
analysis (PCA) of root metabolites was performed using Past3 software.
c The scheme of major metabolic changes in central metabolism after
zaxinone treatment, which adapted from a. Orange arrows indicate
sugar-related metabolites that mainly accumulated upon zaxinone
treatment in WT and/or the zas mutant, but not in d17. CTL control, Zax
zaxinone.
Lipids are further important metabolites required for plant’s growth.
To assess the effect of zaxinone on lipid metabolism, we analyzed the
lipid profile of treated root samples 6 h after zaxinone application.
However, we did not detect a positive effect of zaxinone on lipid
metabolism (Supplementary Fig. [121]13).
Zaxinone application promotes cell division in the root apical meristem and
increases the number of cortex layers
It can be assumed that the growth-promoting effect of zaxinone in rice
roots is caused by an increase in cell number and/or size. To determine
changes at a cellular level, we applied zaxinone (at 5 µM
concentration) to hydroponically grown rice seedlings for 2 weeks and
investigated the roots using cotton blue staining. As shown in
Supplementary Fig. [122]14, the treatment with zaxinone enhanced the
length of the root apex, suggesting an increase of cell division or
cell elongation activity. To test this hypothesis, we used
5-Ethynyl-2′-deoxyuridine (EdU) staining that visualizes proliferating
cells and can be monitored by a fluorescent dye. This experiment
revealed that the meristem length and diameter, as well as the number
of cell layers (counted from epidermis to vascular tissue), increased
upon zaxinone application in both primary and the longest crown roots
(Fig. [123]4a, b). To confirm the increase in the number of cell
layers, we performed cross sections of primary and the longest crown
roots of treated WT and zas mutant seedlings, by staining the cell wall
with SCRI Renaissance 2200. In the main root cortex as well as in the
longest crown roots, zaxinone application led to a remarkable increase
in the number of cortex layers from around three to around five, and of
the number of cells in the circumference by ten, which caused an around
50 µm enlargement of root diameters (Fig. [124]4c, d). Moreover, the
enhancement in the number of cortex layers, root diameters, and cell
numbers upon zaxinone application was much more conspicuous in the
crown roots of zas mutant (Supplementary Fig. [125]15).
Fig. 4. Characterization of root development at cellular level upon zaxinone
treatment.
[126]Fig. 4
[127]Open in a new tab
a, b Ethynyl deoxyuridine (EdU) staining for cell proliferation
analysis. Confocal images of rice root showing dividing cells as
captured by EdU staining in Zeiss LSM 710 inverted confocal microscope.
Root meristem length, root diameter, and cell layers (counted from
epidermis to vascular tissue) in both primary roots and crown roots
after 5 µM zaxinone treatment (twice per week) in Nipponbare WT rice
seedlings. Dividing EdU-stained nuclei are shown in green; nuclei
counterstained with Hoechst 33258 are shown in magenta. Images were
acquired using the tile scan function in the Zen software with
automatized stitching. Regions of interest were divided into multiple
tiles and imaged individually. The tiles were then combined via
automatic stitching to create a large overview image. Images are
representative of the total number (n = 10) of seedlings that were
studied. Scale bar: 50 µm. c, d Cross section of the mock and
zaxinone-treated roots stained with SCRI Renaissance 2200. Magenta
indicates the cell wall staining; green shows the auto-fluorescence
marking lignin and suberin deposition (n = 14 biological replicates).
Examples of cell layer and circumference cell number count are
indicated in the cross section d, scale bar: 50 µm. e Quantification of
cytokinins in root tissues of WT and zas mutant after 2-week zaxinone
treatment. Box plot presented as min to max, n = 4 biological
replicates. Asterisks indicate statistically significant differences as
compared to control by t-test (*p < 0.05, **p < 0.01; ***p < 0.001;
****p < 0.0001). CTL control, Zax zaxinone. Ep epidermis, Ex exodermis,
Sc sclerenchyma, Co cortex, En endodermis.
Zaxinone enhances cytokinin glycosylation in rice roots
The changes in root morphology at the cellular level indicate that
zaxinone may affect the hormonal composition in roots, in addition to
its role in determining SL biosynthesis and sugar metabolism. Analysis
of the RNA-Seq data indicated that zaxinone might affect several genes
related to hormone metabolism, including genes involved in jasmonic
acid and auxin biosynthesis and in the glycosylation of CKs
(Fig. [128]2c). Therefore, we determined the changes in the hormone
profile and content of abscisic acid (ABA), gibberellin (GA), auxin
(IAA), salicylic acid (SA), jasmonic acid (JA), and cytokinins (CKs:
trans-zeatin, isopentyladenine, isopentyladenosine, and benzyladenine)
in rice roots 2, 6, and 24 h after zaxinone application. We did not
detect significant changes in the levels of GA, IAA, or SA, compared to
the control, but observed an increase in ABA and JA levels at 2 and
6 h, respectively (Supplementary Fig. [129]16). Notably, the
application of zaxinone led to a significant decrease in the content of
isopentenyladenosine in all treated samples, a reduction of
isopentenyladenine and benyzladenine in the 6, and 6 and 24 h samples,
respectively, and an increase in levels of glycosylated, inactive
trans-zeatin forms. We also observed an increase in trans-zeatin
content 2 h after application (Supplementary Fig. [130]17). These data
indicate that zaxinone may regulate the abundance and pattern of CKs.
To gain insights into the long-term effect of zaxinone on CKs level, we
applied the compound for two weeks and quantified the hormone. In this
experiment, we also included the zas mutant that showed at the cellular
level a stronger response than WT (Supplementary Fig. [131]15). As
shown in Fig. [132]4e, prolonged treatment with zaxinone led to
significant accumulation in the glycosylated, deactivated trans-zeatin.
Finally, we chose four genes annotated by Kyoto Encyclopedia of Genes
and Genomes analysis as cytokinin glycosyltransferases (Supplementary
Data [133]5), and validated their expression levels in the samples used
for the RNA-seq experiment, using qRT-PCR. As anticipated, these genes
were highly induced following 24 h of zaxinone treatment (Fig. [134]2c
and Supplementary Fig. [135]18). We obtained similar results in zas
mutant seedlings exposed to zaxinone treatment (Supplementary
Fig. [136]19). The induction of these genes may explain the
accumulation of the glycosylated forms of CKs. To further validate that
reduction of cytokinin signaling via CK glycosylation is a part of the
downstream effect of zaxinone, we applied 2.5 μM zaxinone to CK
biosynthetic (Os03g49050, Os05g51390, and Os10g33900) and regulatory
(Os06g08440) rice T-DNA insertion mutants (Supplementary
Fig. [137]20a). Interestingly, none of these mutants showed an
increased root growth in response to zaxinone treatment, in contrast to
the corresponding WT controls (Supplementary Fig. [138]20b). This
result supported the hypothesis that the root growth-promoting effect
of zaxinone is dependent on CKs.
Discussion
As the main product of photosynthetic carbon assimilation, sucrose
plays an energy source and an essential role in plant growth and
development. In addition, this sugar acts as a signaling molecule
interacting with hormonal networks and regulating metabolic
pathways^[139]25–[140]27. In this paper, we show that zaxinone
application promotes sugar metabolism in growing plants, leading to the
accumulation of soluble sugars in different tissues, and might enhance
the photosynthetic activity in rice seedlings (Fig. [141]5). We
observed this effect also in zas mutant plants (Supplementary
Fig. [142]5c), which contain less zaxinone in their roots^[143]23.
Compared to WT, these plants also showed a lower chlorophyll content
under control conditions, suggesting that endogenous zaxinone level
might affect the photosynthetic capacity (Supplementary Fig. [144]5c).
The alterations in shoot metabolism were of considerably lower
magnitude and slower in response, compared to those of root metabolism.
The alterations in root sugar metabolism appear to be biphasic with an
initial accumulation of the major sugars sucrose, glucose, and fructose
whose levels subsequently decreased, whilst those of downstream
metabolites including trehalose, glucose-1-phosphate,
glucose-6-phosphate alongside TCA cycle intermediates increased 24 h
following zaxinone treatment. We observed similar metabolic changes to
the initial effect also in zas mutant plants (Fig. [145]3a). Yet, we
cannot rule out that zaxinone might not directly act as signaling
molecules as it may need further modification or trigger its response
via unknown component(s), which might explain the low response at the
2 h time point and the delay of metabolic changes the root and shoot
tissues. Supporting the metabolomics data, the transcriptomic results
indicated that zaxinone induces several sugar-related metabolic
pathways, such as glycolysis that catabolizes hexose units to produce
energy and building blocks for different cellular components
(Fig. [146]5). We also showed that prolonged treatment with zaxinone
led to an increased starch accumulation in roots, which is synthesized
from mobilized sucrose that is produced by photosynthesis in
leaves^[147]26. In rice, disruption in sucrose synthesis or transport
mutations of rice causes growth retardation^[148]28–[149]30. It might
be a possibility that sucrose provides the energy and C-atoms for the
plant to grow and develop, depending on zaxinone, as we saw that
zaxinone rescued Oszas mutant phenotype^[150]23. This trend correlates
with the observed increase in transcript levels of sucrose metabolic
genes. Moreover, we observed some of the sugar metabolites were highly
accumulated in zas mutant compared to WT at control condition, which
may argue that the endogenous zaxinone level is involved in the sugar
metabolism. However, zas also shows high SL content^[151]22 that
indicates the endogenous relative amount of zaxinone and SLs might
contribute to the observed sugar metabolites as well as phenotypic
changes.
Fig. 5. Model of the mechanisms underlying the growth-promoting effect of
zaxinone in rice.
[152]Fig. 5
[153]Open in a new tab
Application of zaxinone might enhance photosynthesis activity (Calvin
cycle) that produces sucrose that is translocated from the shoot
(source) to developing root tissues (sink). Sucrose then can either be
hydrolyzed (glycolysis) into hexose that enters glycolysis and citric
acid cycle (to produce ATP and C-building blocks), or stores as starch
formed by gluconeogenesis. The glucose can also be used for cytokinin
glycosylation that regulates the bioactivity of CKs in rice root
tissues. A combination of these effects results in root phenotypical
changes and cellular events, such as a larger meristem size. In
addition, zaxinone suppresses SL biosynthesis and release, while it
cannot rescue the SL biosynthesis and perception mutants^[154]22 and
does not affect the central metabolism in the SL biosynthetic d17
mutant. Created with “Biorender”.
Roots and young leaves are the major sinks during early developmental
stages, whereas fruit and seeds are the ones accumulating starch during
the reproductive stages^[155]31. Consequently, zaxinone treatment,
which led to an increased starch accumulation in roots, improved the
sink capacity of this organ. However, we cannot currently demonstrate
whether the increase in chlorophyll abundance and stomatal conductance
is due to an enhanced sink-strength of the zaxinone-treated roots, or a
direct effect of zaxinone/derivative thereof. Regardless, our results
demonstrate zaxinone remarkably increases sugar metabolism and might
modulate photosynthesis in rice plants.
The root system is essential for plants to absorb nutrients and water,
which determine plant growth and performance. Application of zaxinone
remarkably increased crown root numbers^[156]23, root apex length as
well as cortex layers and cell numbers in both rice WT and zas mutant,
which indicates a possible interplay with Auxin or CKs. Indeed, CKs
orchestrate root growth and development; and previous studies
documented that this hormone inhibits root elongation by decreasing
root meristem size^[157]32–[158]34. For example, disruption of CROWN
ROOTLESS5 (CRL5) encoding an ERF transcription factor led to a loss of
rice crown root initiation through repression of two negative
regulators of CK signaling (OsRR1 and OsRR2)^[159]35. Indeed, we
detected significant upregulation of the transcript level of OsRR6,
(Os04g57720) and OsRR10, (Os02g35180), two homologs of OsRR1 and OsRR2,
at 2 h after zaxinone application (Supplementary Data [160]1).
Similarly, overexpressing CK OXIDASE/DEHYDROGENASE 4 (OsCKX4,
Os01g71310) led to lower amounts of CKs, which was accompanied by
longer roots and a larger root apical meristem with more cellular
layers^[161]34,[162]36. In this study, we observed an upregulation of
OsCKX4 expression (2 and 24 h; Supplementary Data [163]1) upon zaxinone
treatment, which may contribute to the increased activity of root
meristems and alterations in root architecture. Besides, the plant
hormones profile and transcript analysis showed that zaxinone reduced
the free-form CKs and enhanced glycosylated-CKs biosynthesis, which can
be considered as fine-tuning of their synthesis, metabolism, and
function. Indeed, glycosylation was shown to significantly reduce the
activity of CKs and to affect their transport, signal transduction, and
impact on growth and development^[164]37. The changes in root
architecture caused by CK glycosylation resemble those observed upon
zaxinone treatment to CK biosynthesis and signaling mutants. Therefore,
CKs-glycosylation is of great significance for understanding the
effects of zaxinone and its impact on rice root development
(Fig. [165]5). Previously, we showed that zaxinone did not enhance the
root growth in SL-deficient rice mutants, indicating the requirement
for intact SL biosynthesis^[166]23. Here, we further found that
zaxinone did not increase the sugar metabolism in the SL-deficient rice
d17 mutant, which indicates that the effect of zaxinone on sugar
metabolism is mediated by SLs. Taken together, zaxinone not only
modulates the SL biosynthesis and release^[167]23, but also acts on the
CK signaling pathway by modulating CK activity through glycosylation.
The latter may be the result of increasing sugar content in the root
tissues.
In summary, we provide experimental evidence at the metabolite,
transcript, and cellular level, which demonstrates the role of zaxinone
in regulating central metabolism, determining hormone profile, and
promoting cell division in rice roots (Fig. [168]5). The results
presented explain zaxinone’s growth-promoting effect in rice plants and
may help to develop new strategies to increase the performance of this
and other crops towards sustainable agriculture.
Methods
Plant material and growth conditions
Nipponbare background zas^[169]23, d17^[170]38, and WT rice plants were
grown under controlled conditions (a 12 h photoperiod, 200-µmol photons
m^−2 s^−1 and day/night temperature of 27/25 °C). Rice seeds were
surface-sterilized in a 50% sodium hypochlorite solution with 0.01 %
Tween-20 for 15 min. The seeds were rinsed with sterile water and
germinated in the dark overnight. The pre-germinated seeds were
transferred to Petri dishes containing half-strength liquid Murashige
and Skoog (MS) medium and incubated in a growth chamber for 7 days.
Thereafter, the seedlings were transferred into black falcon tubes
filled with half-strength modified Hoagland nutrient solution with
adjusted pH to 5.8. The nutrient solution consisted of 5.6 mM
NH[4]NO[3], 0.8 mM MgSO[4]·7H[2]O, 0.8 mM K[2]SO[4], 0.18 mM
FeSO[4]·7H[2]O, 0.18 mM Na[2]EDTA·2H[2]O, 1.6 mM CaCl[2]·2H[2]O, 0.8 mM
KNO[3], 0.023 mM H[3]BO[3], 0.0045 mM MnCl[2]·4H[2]O, 0.0003 mM
CuSO[4]·5H[2]O, 0.0015 mM ZnCl[2], 0.0001 mM Na[2]MoO[4]·2H[2]O and
0.4 mM K[2]HPO[4]·2H[2]O.
For metabolomic and transcriptomic analysis, 3-weeks-old seedlings were
grown hydroponically in half-strength modified Hoagland nutrient
solution. Seedlings were further treated with 5 µM zaxinone for 2, 6,
or 24 h, and tissues were collected.
For phenomic experiments, one-week-old seedlings were grown
hydroponically in half-strength modified Hoagland nutrient solution
with or without 5 µM zaxinone for two weeks. Thereafter, plant tissues
were collected for analysis.
For zaxinone application to CK mutants, 1-week-old seedling (TN67
background) were grown hydroponically in ½ strength Hoagland nutrient
solution containing 2.5 µM zaxinone for 2 weeks. The solution was
changed two times per week.
Synthetic zaxinone was purchased (custom synthesis) from Buchem B.V.
(Apeldoorn, The Netherlands).
Analysis of primary metabolites using GC-MS
Frozen ground material, spiked with 60 μg phenyl-β-glucopyranosides,
was homogenized in 750 μL of methanol at 70 °C for 15 min and then
375 μL of chloroform followed by adding 750 μL of water. The polar
fraction was dried under vacuum, and the residue was derivatized for
40 min at 37 °C (in 50 µL of 20 mg mL^−1 methoxyamine hydrochloride in
pyridine) followed by a 30 min treatment at 37 °C with 70 µL of MSTFA.
The GC-MS system used was a gas chromatograph coupled to a
time-of-flight mass spectrometer (Leco Pegasus HT TOF-MS). A Gerstel
Multi Purpose autosampler system injected the samples. Helium was used
as carrier gas at a constant flow rate of 2 mL s^−1 and gas
chromatography was performed on a 30 m DB-35 column. The injection
temperature was 230 °C and the transfer line and ion source were set to
250 °C. The initial temperature of the oven (85 °C) increased at a rate
of 15 °C min^−1 up to a final temperature of 360 °C. After a solvent
delay of 180 s mass spectra were recorded at 20 scans s^−1 with m/z
70–600 scanning range. Chromatograms and mass spectra were evaluated by
using Chroma TOF 4.5 (Leco) and TagFinder 4.2 software^[171]39,[172]40.
Lipid profile by LC–MS
Lipids were extracted based on the protocol published in^[173]41. In
brief, 5 mg of freeze-dry materiel was homogenized and extracted with
1 mL of pre-cooled (−20 °C) extraction buffer (homogenous
methanol/methyl-tert-butyl-ether [1:3] mixture). After 10 min
incubation at 4 °C and sonication for 10 min in a sonic bath, 500 μL of
water/methanol mixture was added. Samples were then centrifuged (5 min,
14,000×g), which led to the formation of two phases: a lipophilic phase
and a polar phase. Five hundred microliters of the lipophilic phase
were collected and dried under vacuum and resuspended in 200 μL of
ASN/isopropanol and used for lipid analysis. Samples were processed
using UPLC-FT-MS on a C[8] reverse-phase column (100 × 2.1 mm × 1.7 μm
particle size, Waters) at 60 °C. The mobile phases consisted of 1% 1 M
NH[4]OAc and 0.1 % acetic acid in water (buffer A) and
acetonitrile/isopropanol (7:3, UPLC grade BioSolve) supplemented with
1 M NH[4]Ac and 0.1% acetic acid (buffer B). The dried lipid extracts
were resuspended in 500 μL of buffer B. The following gradient profile
was applied: 1 min 45% A, 3 min linear gradient from 45% A to 35 % A,
8 min linear gradient from 25 to 11% A, 3 min linear gradient from 11
to 1% A. Finally, after washing the column for 3 min with 1% A, the
buffer was set back to 45% A, and the column was re-equilibrated for
4 min, leading to a total run time of 22 min. The flow rate of the
mobile phase was 400 μL min^−1. The mass spectra were acquired using an
Exactive mass spectrometer (ThermoFisher,
[174]http://www.thermofisher.com) equipped with an ESI interface. All
the spectra were recorded using altering full-scan and all-ion
fragmentation scan mode, covering a mass range from 100–1500 m/z at a
capillary voltage of 3.0 kV, with a sheath gas flow value of 60 and an
auxiliary gas flow of 35. The resolution was set to 10,000 with 10
scans per second, restricting the Orbitrap loading time to a maximum of
100 ms with a target value of 1E6 ions. The capillary temperature was
set to 150 °C, while the drying gas in the heated electrospray source
was set to 350 °C. The skimmer voltage was held at 25 V, while the tube
lens was set to a value of 130 V. The spectra were recorded from minute
1 to 20 of the UPLC gradients. Processing of chromatograms, peak
detection, and integration was performed using REFINER MS 10.0
(GeneData, [175]http://www.genedata.com) or Xcalibur (Version 3.1,
ThermoFisher, Bremen, Germany). In the first approach, the molecular
masses, retention time, and associated peak intensities for the three
replicates of each sample were extracted from the raw files, which
contained the full-scan MS and the all-ion fragmentation MS Data
Processing of MS data included the removal of the fragmentation
information, isotopic peaks, and chemical noise. Further peak filtering
on the manually extracted spectra or the aligned data matrices was
performed. Obtained features (m/z at a certain retention time) were
queried against an in-house lipid database for further annotation
RNA library preparation and transcriptomic analysis
Total rice root RNA was extracted with TRIzol (Invitrogen,
[176]https://www.thermofisher.com/de/de/home.htmL) using a Direct-zol
RNA Miniprep Plus Kit following the manufacturer’s instructions (ZYMO
RESEARCH; USA). RNA quality was checked with a Agilent 2100
Bioanalyzer, and RNA concentration was measured using a Qubit 3.0
Fluorometer. The cDNA libraries were constructed following standard
Illumina protocols and paired‐end sequenced on an Illumina HiSeq 4000
machine by the Bioscience core lab of KAUST. RNA‐Seq reads were aligned
to the O. sativa genome v7.0 downloaded from Phytozome v12.1
([177]http://phytozome.jgi.doe.gov/). Data processing and analysis were
performed using the LSTrAP workflow^[178]42, which included all steps
described below. Adapter sequences were removed from fastq files by
Trimmomatic^[179]43, and aligned to the genome using HISAT2^[180]44.
Read counts aligned to each annotated gene were computed with
HTSeq^[181]45. The results were passed through LSTrAP quality control
and TPM normalized. The mean data were used to cluster and resistance
level was visualized as a heatmap using a hierarchical clustering R
script. Principal component analysis (PCA), a multivariate statistical
technique, was further conducted to examine links between samples. All
analyses were performed using the R statistical package. For
differential gene expression, read counts from HTSeq were analyzed
using the R package DESeq2^[182]46. Genes were considered
differentially expressed based on a P‐value adjusted by the
Benjamini–Hochberg procedure^[183]47 below 0.05. Gene Ontology (GO)
enrichment analysis of all the differentially were selected at
FDR < 0.05 and analyzed by Panther-Gene list analysis^[184]48
([185]http://pantherdb.org/). Kyoto Encyclopedia of Genes and Genomes
and PlantCyc pathway enrichment analysis for up- and downregulated
genes, were then analyzed by The Plant GeneSet Enrichment Analysis
Toolkit (PlantGSEA)^[186]49
([187]http://structuralbiology.cau.edu.cn/PlantGSEA/index.php).
Visualization of DEGs in MapMan followed the instructions as
described^[188]50.
Gene expression analysis
Roots of rice seedlings were ground and homogenized in liquid nitrogen,
and total RNA was isolated using a Direct-zol RNA Miniprep Plus Kit
following the manufacturer’s instructions (ZYMO RESEARCH; USA). cDNA
was synthesized from 1 µg of total RNA using iScript cDNA Synthesis Kit
(BIO-RAD Laboratories, Inc, 2000 Alfred Nobel Drive, Hercules, CA;
USA), according to the instructions in the user manual. Transcript
levels were detected by real-time quantitative RT-PCR (qRT-PCR) which
was performed using SYBR Green Master Mix (Applied Biosystems;
[189]www.lifetechnologies.com) in a CFX384 Touch™ Real-Time PCR
Detection System (BIO-RAD Laboratories, Inc, 2000 Alfred Nobel Drive,
Hercules, CA; USA). Primers used for qRT-PCR analysis are listed in
Supplementary Data [190]8. The gene expression level was calculated by
normalization to the rice housekeeping gene Ubiquitin (OsUBQ)
(Supplementary Data [191]8). The relative gene expression level was
calculated according to 2^−ΔΔCT method^[192]51.
Cotton blue staining for root apex
Apex length and width were assessed in WT rice roots with or without
5 μM zaxinone (applied twice a week). Plants were grown hydroponically
in Hoagland solution (400 µM Pi), and data were collected 3 weeks
post-germination. The primary crown root apex was stained with Cotton
Blue 0.1 %. The apex length was calculated considering the segment
between the root tip and the first root hair.
Ethynyl deoxyuridine (EdU) staining for cell proliferation analysis
Cell proliferation in rice seedlings was evaluated using the Click-iT
EdU Alexa Fluor 488 imaging kit ([193]C10637, Invitrogen) followed the
procedure^[194]52. Plants were incubated in Murashige and Skoog medium
with EdU for 12 h. For each plant, the primary root and the longest
crown root (prior to the formation of lateral roots) were harvested and
fixed in 3.7% formaldehyde for 1 h under vacuum. Then samples were
permeabilized with PBS containing 0.5 % Triton X-100 for 1 h and
incubated for 1 h in the dark with a click-it-reaction cocktail that
was prepared according to the manual, followed by DNA counterstaining
using Hoechst 33342 in PBS under vacuum in the dark for 1 h. Samples
were mounted in clearing solution and incubated in the dark for 2 weeks
at 4 °C as described in the protocol published in ref. ^[195]53. Images
were captured by a Zeiss LSM 880 inverted confocal microscope and
automatically stitched to generate the overview image of the root tip
in ZEN 2.0 while imaging. Dividing EdU-stained nuclei are shown in
green; nuclei counterstained with Hoechst 33258 are shown in blue.
Images are representative of the total number (n = 10) of seedlings
that were studied.
Root cross section, staining, and microscopy
Fresh root segments starting from the root hair emergence zone to the
direction of the shoot (an upward direction, about 0.5 cm from in
differentiation zone) were embedded in 10% low melting agarose and
sectioned using a Leica VT1000S vibratome. The SCRI Renaissance 2200
(SR2200) stain was used to visualize cell walls while the berberine
hemisulfate stain was used to visualize suberin^[196]54. Images were
captured using a Zeiss LSM 880 inverted confocal microscope with
excitation of 405 nm for SCRI or 488 nm for berberine.
Quantification of starch
For starch extraction, excised root systems were rapidly blot-dried on
filter paper and weighed. Samples were then frozen in liquid nitrogen,
transferred to 2-mL Eppendorf tubes (Eppendorf, Hamburg, Germany), and
thoroughly homogenized using a pestle in liquid nitrogen. The samples
were further homogenized in 0.5 mL of absolute ethanol. After the
addition of 0.5 mL of 80% ethanol, the tubes were incubated at 70 °C
for 90 min and then centrifuged for 10 min at 11,337 × g and the pellet
was resuspended in 1 mL of 80% ethanol. Two more washings were
performed with 1 mL of 80% ethanol (and 10 min of centrifugation). The
pellets were finally resuspended in 400 μL of 0.2 m KOH and incubated
at 95 °C for 60 min. After neutralization with 70 μL of acetic acid,
the samples were centrifuged for 10 min and the supernatant was used
for starch quantification (Starch Test-Combination enzymatic analysis
kit, cat. no. 207748; Boehringer, Mannheim, Germany), according to the
manufacturer’s instructions. At least three independent experiments,
including at least three plants each, were performed to obtain all
results of enzymatic starch quantification.
Quantification of plant hormones
For the quantification of endogenous hormone levels, 20 mg of
freeze-dried ground tissues were spiked with internal standards
D[6]-ABA (3.2 ng), D[2]-GA1 (0.08 ng), D[2]-IAA (5.4 ng), D[4]-SA
(0.05 ng), D[2]-JA (0.1 ng), D[5]-trans-zeatin (1.5 ng),
D[5]-trans-zeatin-O-glucoside (2 ng), D[5]-trans-zeatin
riboside-O-glucoside (2 ng), D[6]-N^6-Isopentenyladenine (2 ng),
N[15]-N^6-isopentenyladenosine (2 ng), and D[7]-N^6-Benzyladenine
(2 ng) along with 1.5 mL of methanol as described procedure^[197]51.
The mixture was sonicated for 15 min in an ultrasonic bath (Branson
3510 ultrasonic bath), followed by centrifugation for 10 min at 14,000
× g at 4 °C. The supernatant was collected, and the pellet was
re-extracted with 1.5 mL of the same solvent. Then, the two
supernatants were combined and dried under a vacuum. The sample was
re-dissolved in 150 μL of acetonitrile:water (25:75, v-v) and filtered
through a 0.22 μm filter for LC–MS analysis. Plant hormones were
analyzed using HPLC-Q-Trap-MS/MS with Multiple Reaction Monitoring
(MRM) mode. Chromatographic separation was achieved on a ZORBAX Eclipse
plus C[18] column (150 × 2.1 mm; 3.5 μm; Agilent). Mobile phases
consisted of water:acetonitrile (95:5, v-v) and acetonitrile, both
containing 0.1% formic acid. A linear gradient was optimized as follows
(flow rate, 0.4 mL min^−1): 0 − 17 min, 10–100% B, followed by washing
with 100% B and equilibration with 10% B. The injection volume was 5 μL
and the column temperature was maintained at 40 °C for each run. Mass
spectrometry was conducted in electrospray and MRM mode, in positive
ion mode for cytokinins, and in negative ion mode for the other
hormones. Relevant instrumental parameters were set as follows: ion
source of turbo spray, ion spray voltage of (±) 4500 V, curtain gas of
25 psi, collision gas of medium, gas 1 of 45 psi, gas 2 of 30 psi,
turbo gas temperature of 500 °C, entrance potential of −10 V. The
characteristic MRM transitions (precursor ion → product ion) were
263.2 → 153.1 for ABA; 347.1 → 261.1 for GA1; 174.0 → 129.6 for IAA;
136.6 → 92.8 for SA; 209.0 → 59.0 for JA; 269.2 → 159.1 for D[6]-ABA;
349.1 → 261.1 for D[2]-GA1; 176.0 → 131.6 for D[2]-IAA; 141.0 → 97.0
for D[4]-SA ; 211.0 → 61.0 for D[2]-JA. 225.2 → 136.7 for
D[5]-trans-zeatin; 387.2 → 225.4 for D[5]-trans-zeatin- O -glucoside;
519.2 → 225.2 for D[5]-trans-zeatin riboside-O-glucoside; 210.2 → 137.0
for D[6]-N^6-Isopentenyladenine; 337.0 → 205.0 for
N[15]-N^6-isopentenyladenosine; 233.1 → 98.0 for
D[7]-N^6-Benzyladenine; 220.2 → 136.0 for trans-zeatin; 382.2 → 220.2
for trans-zeatin-O-glucoside; 514.2 → 220.2 for trans-zeatin
riboside-O-glucoside; 204.2 → 136.1 N^6-Isopentenyladenine;
336.0 → 204.0 for N^6-isopentenyladenosine; 226.1 → 91.0 for
N^6-Benzyladenine.
Measurement of photosynthetic parameters
Three-week-old seedlings were grown hydroponically in a
half-strength-modified Hoagland nutrient solution. Seedlings were
further treated with 5 µM zaxinone for 2, 6, or 24 h. Leaf chlorophyll
content was measured by CCM-200 plus chlorophyll content meter
(Opti-Sciences, Hudson, USA), and leaf stomatal conductance was
measured by AP4 Porometer (Delta-T, Cambridge, UK).
Chlorophyll quantification
Chlorophyll was extracted from the leaf segment by following the
procedure according to ref. ^[198]55 with a slight modification.
Briefly, an equal amount of frozen leaf tissue was measured in a 2 mL
Eppendorf tube and ground into a fine powder with 2 mm metal beads. One
mL of 80% acetone was added to each tube and the mixture was vortex for
30 s. The extracted mixture was incubated at room temperature for
10 min. Each sample was subjected to centrifugation at 4282 × g, 4, at
room temperature for 90 s. Then 200 µL of supernatant was collected
from the top of each tube and added to 96 well plates. The plate was
run in the microplate reader (Tecan Infinite M1000 Pro). The absorbance
of Chlorophyll-a (Chla) and Chlorophyll-b (Chlb) was determined by
UV-spectrophotometry at 645 and 663 nm wavelength. Chlorophyll-a,
Chlorophyll-b, and total chlorophyll content were calculated from each
extract by using the following equations:
[MATH: Chla(mgg−1)=12.7(A663)−2.6
9(A645)×V/1000×W
:MATH]
1
[MATH: Chlb(mgg−1)=22.9(A645)−4.6
8(A663)×V/1000×W
:MATH]
2
[MATH: TotalChl(mgg−1)=20.2(A645)+8.0
2(A663)×V/1000×W
:MATH]
3
where
A = absorbance at specific wavelengths
V = final volume of chlorophyll extract
W = fresh weight of tissue extracted
Statistics and reproducibility
A minimum of three independent biological replicates to ensure
reproducibility in all the experiments. Exact biological samples (n)
and mean with error bars are indicated in individual figure captions
and methods. Statistical tests were carried out through one-way
analysis of variance (one-way ANOVA) and Tukey’s post hoc test or
two-tailed Student’s t-tests, using a probability level of P < 0.05,
which was considered to be statistically significant.
Reporting summary
Further information on research design is available in the [199]Nature
Research Reporting Summary linked to this article.
Supplementary information
[200]Supplemental Information^ (2.2MB, pdf)
[201]42003_2021_2740_MOESM2_ESM.pdf^ (215.6KB, pdf)
Description of Additional Supplementary Files
[202]Supplementary Data 1^ (194.9KB, xlsx)
[203]Supplementary Data 2^ (12.1KB, xlsx)
[204]Supplementary Data 3^ (12.5KB, xlsx)
[205]Supplementary Data 4^ (20.4KB, xlsx)
[206]Supplementary Data 5^ (10.5KB, xlsx)
[207]Supplementary Data 6^ (502.2KB, xlsx)
[208]Supplementary Data 7^ (17.4KB, xlsx)
[209]Supplementary Data 8^ (10.8KB, xlsx)
[210]Supplementary Data 9^ (32KB, xls)
[211]Supplementary Data 10^ (169.3KB, xlsx)
[212]Supplementary Data 11^ (248.3KB, xlsx)
[213]Reporting Summary^ (263.6KB, pdf)
Acknowledgements