Contrasting carbohydrate metabolism in sequenced genotypes from a Miscanthus mapping family
A total of 102 genotypes from a paired cross between diploid M. sinensis (“M. sinen 102”) and a diploid M. sacchariflorus (“M. sacch 297”) were established in field conditions and phenotyped. Non-structural carbohydrates were sampled in July 2014, during the summer growing season, and annual yield was obtained at harvest after the following winter. The distribution of carbohydrate concentrations and biomass yield for 98 hybrids were previously reported [17]. After including additional information about number of tillers for the population (Fig. 1 and Suppl. Table S1), we observed significant correlations between number of tillers and starch (r = -0.45, p < 0.001), fructose (r = 0.31, p < 0.005), and total NSC (r = -0.40, p < 0.0001) for the whole family (Suppl. Table S1). We also observed significant correlations between number of tillers and the ratio of sucrose/starch (r = 0.37, p < 0.001), fructose/starch (r = -0.45, p < 0.001), glucose/starch (r = -0.38, p < 0.001) and sucrose/fructose (r = -0.32, p < 0.01). We observed a significant positive correlation between biomass yield and number of tillers (r = 0.62 ± 0.03 for three seasons, p < 0.001).
Four M. sinensis X M. sacchariflorus hybrids from this family (Triangles in Fig. 1) were selected for RNA sequencing in their fourth growing season (2013), based on a higher or lower than the average number of tillers. The two parents of the family were also sequenced (Diamonds in Fig. 1). When the four sequenced hybrids were divided into two groups (genotypes 112 and 90 against genotypes 18 and 120), we observed significant differences between these groups in the number of tillers (p < 0.05), biomass yield quantified as dry weight per plant (p < 0.05), and the final canopy heights (p < 0.05). We also observed a significant difference between these two groups in the concentrations of starch (p < 0.005) and sucrose (p < 0.05), but we did not observe significant differences between groups in the concentrations of fructose or glucose. The most significant difference (p < 0.001) was observed in the total concentration of non-structural carbohydrates (NSC), which was calculated as the sum of the glucose, fructose, sucrose and starch concentrations. We observed significant differences also in the fructose/starch (p < 0.05) and glucose/fructose ratios (p < 0.01). However, any other ratio between concentrations was not significantly different between the groups (Suppl. Table S2).
There was a significant difference between the M. sacchariflorus and M. sinensis parents in NSC (p < 0.05) and sucrose concentrations (p < 0.01). However, there was no significant difference between the parents in the starch, fructose or glucose concentrations (Suppl. Table S2). It is likely an example of heterosis (transgressive segregation) that significant differences in starch, fructose or glucose concentrations were observed in the hybrid progeny but not the parents.
Differential expression (DE) analysis between hybrids and species
We performed RNA-seq from the leaf, stem and root tissue samples extracted from four M. sacchariflorus X M. sinensis interspecific hybrids, and their two parents (Table 1). When the normalised counts obtained from DESeq2 [28] were used to cluster the samples (Fig. 2), the samples firstly grouped by tissue (PC1) and secondly by species (PC2). As a result, the downstream analysis was performed for each tissue separately. Stem and root samples clustered together, and the clustering of these separately from the leaf tissue explained 64% of the variation. Species explains 17% of the variation, with the hybrids falling between the two parents, which were furthest apart from each other.
Table 1: RNA-seq libraries used in this study.
Genotype
|
tissue
|
Library
|
Group
|
112 (hybrid)
|
root
|
LIB2338
|
High NSC / Low yield / Fewer tillers
|
stem
|
LIB2339
|
leaf
|
LIB2340
|
90 (hybrid)
|
root
|
LIB2341
|
stem
|
LIB2342
|
leaf
|
LIB2343
|
120 (hybrid)
|
root
|
LIB2344
|
Low NSC / High yield / Many tillers
|
stem
|
LIB2345
|
leaf
|
LIB2346
|
18 (hybrid)
|
root
|
LIB2347
|
stem
|
LIB2348
|
leaf
|
LIB2349
|
M. sacch 297
|
stem
|
LIB2352
|
Progenitors
|
leaf
|
LIB2353
|
stem
|
SAM1158
|
root
|
SAM1159
|
leaf
|
SAM1160
|
root*
|
SAM1161
|
M. sinensis 102
|
stem
|
LIB2350
|
leaf
|
LIB2351
|
stem
|
SAM1162
|
root
|
SAM1163
|
leaf
|
SAM1164
|
M. sacch = M. sacchariflorus; *root tip
We obtained 1,386 differentially expressed genes (DEG; Suppl. Table S3) in total between the hybrids identified as “High NSC” and “Low NSC” (Fig. 1) at FDR < 0.05 (Fig. 3A). There were 892 DEGs in stems (598 up-regulated and 294 down-regulated), 741 DEGs in leaves (410 up-regulated and 331 down-regulated), and only 253 DEGs in roots (116 up-regulated and 137 down-regulated). 64% of the DEGs in roots were DE in both of the other tissues too, but most DEGs in stem or leaves were exclusively DE in either stem or leaves.
We also compared the expression between the hybrids against each parent and considered a gene as DE if it was DE in both comparisons at FDR < 0.05 (Suppl. Table S4). Under these criteria, there were 2,870 DEGs in roots, 1,464 DEGs in leaves, and 729 DEGs in stems (Fig. 3B). Only 64 among these DEG were also DE between “High NSC” and “Low NSC” hybrids. There were 16,311 DEGs between the hybrids and M. sinensis alone (Suppl. figure S1), and 15,616 DEGs between the hybrids and M. sacchariflorus alone (Suppl. figure S2), this is over a third of the total transcriptome.
Enriched Gene Ontology (GO) terms in DEGs
Enrichment analysis of GO terms over-represented among DE genes allowed us to identify the biological processes (BP) and molecular functions (MF) that are differentially regulated in each tissue. After annotating the reference transcriptome with the homologous proteins and full set of GO terms and (Suppl. Table S5), we simplified the results to the more general “GO slim” terms.
All the significant enrichment “GO slim” terms among DEGs between the “High NSC” and “Low NSC” hybrids were associated with metabolic processes, with the single exception of “response to stress” in stems (Fig. 4; Suppl. Table S6). Among the GO terms in the “biological process” category, the most significantly enriched ones (p < 0.001) were “Carbohydrate metabolism” and “Secondary metabolism” in stem and leaves, and “Generation of precursor metabolites and energy” and “response to stress” in stems. Among the “molecular process” category, “hydrolysis on glycosyl bonds” and “redox activities” were the most significantly enriched (p < 0.0001) in both stems and leaves (Suppl. Table S6).
Thirty-six enzymatic reactions were annotated among DEG in the stem (Table 2). Only six were down-regulated in “High NSC”; four involved in the generation of precursor metabolites and energy, namely 6-phosphofructokinase (EC 2.7.1.11) and Triose-phosphate isomerase (EC 5.3.1.1) in the glycolysis pathway; Malate dehydrogenase NADP(+) (EC 1.1.1.82) in the pyruvate metabolism; and 2-carboxy-D-arabinitol-1-phosphatase (EC 3.1.3.63); and one each in the other GO categories, namely Beta-N-acetylhexosaminidase (EC 3.2.1.52) and carboxypeptidase (EC 3.4.16.-).
Table 2: Carbohydrate and secondary metabolic enzymatic reactions differentially expressed between “high NSC” and “low NSC” Miscanthus hybrids.
Leaf FC or Stem FC = Log2 fold-change expression between “high NSC” / “Low NSC” hybrids in either leaf or stem tissues.
ENZYME NAME
|
ENZIME CODE
|
GENE
|
LEAF FC
|
STEM FC
|
Malate dehydrogenase (NADP(+))
|
1.1.1.82
|
Misin07G271500
|
-3.12
|
-0.85
|
Laccase
|
1.10.3.2
|
Misin06G334400
|
-2.40
|
|
Indole-2-monooxygenase
|
1.14.14.153
1.14.13.138
|
Misin19G207900
|
1.52
|
|
MisinT014900
|
11.40
|
5.99
|
Indolin-2-one monooxygenase
|
1.14.13.138
1.14.14.157
|
MisinT219600
|
2.22
|
|
Misin07G204200
|
5.73
|
|
Misin01G349900
|
-4.68
|
|
Misin09G192700
|
3.53
|
4.73
|
3-hydroxyindolin-2-one monooxygenase
|
1.14.13.139
1.14.14.109
|
MisinT014600
|
10.87
|
8.05
|
Ent-isokaurene C2/C3-hydroxylase
|
1.14.13.143
1.14.14.76
|
Misin01G158200
|
|
1.83
|
Ent-cassa-12,15-diene 11-hydroxylase
|
1.14.13.145
1.14.14.112
|
Misin15G165600
|
5.04
|
5.67
|
Trans-cinnamate 4-monooxygenase
|
1.14.14.91
|
Misin05G312600
|
-2.11
|
|
Camalexin synthase
|
1.14.19.52
|
Misin04G105800
|
1.58
|
|
Aldehyde dehydrogenase (NAD(+))
|
1.2.1.3
|
Misin04G200300
|
0.81
|
|
L-glutamate gamma-semialdehyde dehydrogenase
|
1.2.1.88
|
Misin17G216100
|
4.19
|
4.17
|
Pyruvate dehydrogenase
|
1.2.4.1
|
Misin18G109800
|
|
4.21
|
Delta(24)-sterol reductase
|
1.3.1.72
|
MisinT029700
|
2.52
|
1.50
|
Nicotinamide N-methyltransferase
|
2.2.1.1
|
Misin11G031500
|
|
2.33
|
Sinapoylglucose--malate O-sinapoyltransferase
|
2.3.1.92
|
Misin12G095900
|
2.55
|
3.49
|
Malate synthase
|
2.3.3.9
|
Misin11G121200
|
5.07
|
5.71
|
Glycogen phosphorylase
|
2.4.1.1
|
Misin05G335800
|
|
1.07
|
Sucrose synthase
|
2.4.1.13
|
Misin01G358800
|
|
1.98
|
Sucrose-phosphate synthase
|
2.4.1.14
|
Misin10G070300
|
|
1.10
|
1,4-alpha-glucan branching enzyme
|
2.4.1.18
|
Misin07G352300
|
|
1.92
|
Misin18G276400
|
|
1.77
|
Starch synthase
|
2.4.1.21
|
MisinT393000
|
|
1.63
|
Misin19G100900
|
|
2.44
|
Dimethylallyltranstransferase
|
2.5.1.1
|
Misin04G333300
|
-1.40
|
|
Glutathione transferase
|
2.5.1.18
|
Misin02G293100
|
|
4.15
|
Misin02G286600
|
|
2.01
|
MisinT258000
|
2.57
|
1.55
|
MisinT404400
|
2.51
|
3.37
|
6-phosphofructokinase
|
2.7.1.11
|
Misin12G113600
|
-6.94
|
-6.65
|
Pyruvate kinase
|
2.7.1.40
|
Misin06G200500
|
0.64
|
|
Mitogen-activated protein kinase
|
2.7.11.24
|
Misin01G390800
|
|
1.77
|
Glucose-1-phosphate adenylyltransferase
|
2.7.7.27
|
Misin17G255500
|
|
1.72
|
Fucose-1-phosphate guanylyltransferase
|
2.7.7.30
|
Misin02G490600
|
1.84
|
2.90
|
6-phosphogluconolactonase
|
3.1.1.31
|
Misin07G251100
|
-2.48
|
|
Sugar-phosphatase
|
3.1.3.23
|
Misin10G086500
|
-1.67
|
|
2-carboxy-D-arabinitol-1-phosphatase
|
3.1.3.63
|
Misin10G020200
|
-3.42
|
-2.49
|
Glycerophosphodiester phosphodiesterase
|
3.1.4.46
|
Misin08G144100
|
0.82
|
|
Alpha-amylase
|
3.2.1.1
|
Misin04G207500
|
1.67
|
|
Beta-amylase
|
3.2.1.2
|
Misin02G205400
|
2.36
|
2.37
|
MisinT552400
|
|
3.03
|
Misin04G312400
|
|
5.04
|
Misin15G034600
|
|
1.09
|
Beta-glucosidase
|
3.2.1.21
|
Misin11G141900
|
-1.58
|
|
Misin11G111200
|
-1.26
|
|
Misin06G358300
|
1.27
|
|
Misin12G147300
|
|
1.12
|
Misin11G142000
|
|
1.21
|
Alpha-galactosidase
|
3.2.1.22
|
MisinT167900
|
|
0.81
|
Beta-galactosidase
|
3.2.1.23
|
Misin03G233500
|
-1.03
|
|
Beta-fructofuranosidase
|
3.2.1.26
|
Misin11G067200
|
|
3.63
|
Glucan endo-1,3-beta-D-glucosidase
|
3.2.1.39
|
Misin17G123500
|
-2.79
|
|
Misin16G118700
|
-3.26
|
|
Misin02G326400
|
2.53
|
|
Misin01G337100
|
2.41
|
|
MisinT226600
|
2.45
|
4.22
|
Misin01G145100
|
|
2.22
|
Misin02G115300
|
|
1.99
|
Beta-N-acetylhexosaminidase
|
3.2.1.52
|
Misin03G316100
|
3.91
|
|
Misin17G142700
|
|
-0.57
|
Non-reducing end alpha-L-arabinofuranosidase
|
3.2.1.55
|
Misin10G067800
|
-1.95
|
|
Isoamylase
|
3.2.1.68
|
Misin07G322000
|
-1.02
|
|
Misin17G131000
|
|
1.04
|
Misin03G195400
|
|
1.36
|
Misin04G215400
|
|
1.07
|
Endo-1,4-beta-xylanase
|
3.2.1.8
|
Misin05G078900
|
-2.78
|
|
Glucose-6-phosphate 1-epimerase
|
5.1.3.15
|
Misin06G202700
|
-1.67
|
|
Triose-phosphate isomerase
|
5.3.1.1
|
Misin03G235900
|
|
-0.59
|
Phosphoglycerate mutase
|
5.4.2.11/.12
|
Misin02G341300
|
2.45
|
4.35
|
Ent-copalyl diphosphate synthase
|
5.5.1.13
|
Misin01G047600
|
2.51
|
|
A similar analysis on the enriched GO slim terms among DEGs between hybrids and parents (Suppl. Figure S3; Suppl. Table S7) revealed that the most significantly enriched GO terms (p < 0.01) were in the root and associated with RNA/DNA binding and translation (including ribosome biogenesis and equivalent terms), and several biosynthetic processes. Remarkably, there were no enriched GO terms in the stem between hybrids and parents.
DEG associated with the starch and sucrose metabolism
There were 88 DEGs associated with the enriched “Carbohydrate metabolism” GO term (Suppl. Table S8), specifically 57 DEGs in stems (42 up-regulated and 15 down-regulated) and 44 DEGs in leaves (20 were up-regulated and 24 down-regulated). Thirteen DEGs were common to both tissues and showed close fold-change values in both tissues. All but two of these 88 DEGs could be functionally annotated, 52 and 56 of them had a homologous protein in A. thaliana or rice, respectively.
Twenty-nine DEGs were involved in enzymatic reactions that were part of the starch and sucrose metabolic pathways (KEGG pathway ath00500; Suppl. Figure S4). Among these, all 20 DEGs in stems were up-regulated in “High NSC”, but half of the DEGs in leaves (which were beta-glucosidases) were down-regulated in “High NSC”. Enzymatic proteins in the starch degradation pathway were DE in root and leaves (e.g. AMY3, ISA3, BAM1). At the same time, sucrose metabolism genes in the cytosol were only DE in stems (SUS3, SPS5). Similarly, reactions involving ADP-glucose were only DE in stems (e.g. AGP, SS2, SS3, SBE2).
Twenty-nine genes were annotated as involved in the “generation of precursor metabolites and energy” (Suppl. Table S8), 17 of which could be annotated with an enzymatic code (KEGG pathway ath00010; Suppl. Figure S5). Six genes were involved in starch metabolism (ISA3, DBE1, PFK2, SBE2, PHS2). The phosphofructokinase 2 (PFK2) is the only one clearly down-regulated in “High NSC”. Among the others, a malate synthase (MLS) and an aldehyde dehydrogenase 12A1 involved in siRNAs generation, and an Fts protease (FTSH6) in the chloroplast were all highly up-regulated (FC > 5) in “High NSC”. On the other hand, triosephosphate isomerase (TIM) was down-regulated in “High NSC”.
The relation between 32 DEGs involved in the twelve DE enzymatic reactions in starch and sucrose metabolism, plus three of the glycolysis reactions are summarised in Fig. 5 and Table 3.
Table 3: Thirty-nine differentially expressed genes were involved in twelve reactions in the starch and sucrose metabolism and three of the glycolysis reactions were highlighted in our analysis.
Leaf/stem = Log2 fold-change expression “high NSC” / “Low NSC” hybrids in either lead or stem tissues; Ath/Rice = Homologous protein in Arabidopsis thaliana and rice (The prefix “LOC_” is not included in the name).
GENE
|
LEAF
|
STEM
|
EC
|
Protein name
|
Ath
|
Rice
|
Misin01G145100
|
|
2.22
|
3.2.1.39
|
BG8
|
AT1G64760
|
Os03g45390
|
Misin01G337100
|
2.41
|
|
3.2.1.39
|
Beta-1,3-glucanase
|
|
Os03g25790
|
Misin01G358800
|
|
1.98
|
2.4.1.13
|
SUS3
|
AT4G02280
|
Os03g22120
|
Misin02G115300
|
|
1.99
|
3.2.1.39
|
T11I18
|
AT3G04010
|
Os03g45390
|
Misin02G205400
|
2.36
|
2.37
|
3.2.1.2
|
BAM1
|
AT3G23920
|
Os10g32810
|
Misin02G326400
|
2.53
|
|
3.2.1.39
|
Beta-1,3-glucanase
|
|
Os03g25790
|
Misin02G341300
|
2.45
|
4.35
|
5.4.2.11
|
Phosphoglycerate mutase
|
|
Os03g21260
|
Misin03G195400
|
|
1.36
|
3.2.1.68
|
ISA3
|
AT4G09020
|
Os09g29404
|
Misin03G235900
|
|
-0.59
|
5.3.1.1
|
TIM
|
AT2G21170
|
Os09g36450
|
Misin03G316100
|
3.91
|
|
3.2.1.52
|
HEXO2
|
AT1G05590
|
Os07g38790
|
Misin04G207500
|
1.67
|
|
3.2.1.1
|
AMY1
|
AT4G25000
|
|
Misin04G215400
|
|
1.07
|
3.2.1.68
|
ISA3
|
AT4G09020
|
Os09g29404
|
Misin04G312400
|
|
5.04
|
3.2.1.2
|
Beta-amylase
|
|
|
Misin05G335800
|
|
1.07
|
2.4.1.1
|
PHS2
|
AT3G46970
|
Os01g63270
|
Misin06G202700
|
-1.67
|
|
5.1.3.15
|
F15G16.1/SF10
|
AT3G61610
|
Os01g46950
|
Misin06G358300
|
1.27
|
|
3.2.1.21
|
BGLU42/4
|
AT5G36890
|
Os01g67220
|
Misin07G322000
|
-1.02
|
|
3.2.1.68
|
LSF1/SEX4
|
AT3G01510
|
Os08g29160
|
Misin07G352300
|
|
1.92
|
2.4.1.18
|
SBE2.2
|
AT5G03650
|
Os02g32660
|
Misin10G070300
|
|
1.10
|
2.4.1.14
|
SPS5
|
|
Os11g12810
|
Misin11G067200
|
|
3.63
|
3.2.1.26
|
cwINV4/OsCIN2
|
AT2G36190
|
Os04g33740
|
Misin11G111200
|
-1.26
|
|
3.2.1.21
|
BGLU14
|
AT2G25630
|
|
Misin11G121200
|
|
5.71
|
2.3.3.9
|
MLS
|
AT5G03860
|
Os04g40990
|
Misin11G141900
|
-1.58
|
|
3.2.1.21
|
BGLU45/18
|
AT1G61810
|
Os04g43410
|
Misin11G142000
|
|
1.21
|
3.2.1.21
|
BGLU18
|
|
Os04g43410
|
Misin12G113600
|
|
-6.65
|
2.7.1.11
|
PFK2
|
AT5G47810
|
Os09g30240
|
Misin12G147300
|
|
1.12
|
3.2.1.21
|
BGLU46
|
AT1G61820
|
Os04g43390
|
Misin15G034600
|
|
1.09
|
3.2.1.2
|
Beta-amylase
|
|
|
Misin16G118700
|
-3.26
|
|
3.2.1.39
|
BG1
|
AT3G57270
|
|
Misin17G123500
|
-2.79
|
|
3.2.1.39
|
BG3
|
AT3G57240
|
|
Misin17G131000
|
|
1.04
|
3.2.1.68
|
DBE1
|
AT1G03310
|
Os05g32710
|
Misin17G142700
|
|
-0.57
|
3.2.1.52
|
HEXO3
|
AT1G65590
|
Os05g34320
|
Misin17G216100
|
|
4.17
|
1.2.1.88
|
ALDH12A1
|
AT5G62530
|
|
Misin17G255500
|
|
1.72
|
2.7.7.27
|
AGPL3/APL3
|
|
Os05g50380
|
Misin18G276400
|
|
1.77
|
2.4.1.18
|
Glycogen branching
|
|
|
Misin19G100700
|
|
5.08
|
3.4.24.-
|
FTSH6
|
AT5G15250
|
Os06g12370
|
Misin19G100900
|
|
2.44
|
2.4.1.21
|
SS2
|
|
Os06g12450
|
MisinT226600
|
2.45
|
4.22
|
3.2.1.39
|
BGL2
|
AT3G57260
|
|
MisinT393000
|
|
1.63
|
2.4.1.21
|
SS3
|
AT1G11720
|
|
MisinT552400
|
|
3.03
|
3.2.1.2
|
BAM1
|
AT3G23920
|
Os10g32810
|
DEG associated with other enriched GO terms
The 72 genes annotated as "Response to stress" were involved in a broad range of responses (Suppl. Table S10). On the other hand, the most significantly enriched GO terms in the "Molecular functions" category were associated with metabolic-related enzymatic reactions, namely “oxidoreductase activities” and “hydrolase activities”. The former included 38 cytochrome P450 proteins.
“Secondary metabolism” was enriched in both stems and leaves. 17 of the 19 DEGs in stems were up-regulated, but half of the DEGs in leaves were down-regulated. 16 of the 31 genes involved in the “secondary metabolism” were cytochrome P450 proteins (Suppl. Table S11). Six were included in benzoxazinoids biosynthesis, which is associated with defence in grasses. Another six were involved in terpenoids and phenylpropanoid biosynthesis (KEGG ath00900 and ath00940).
Many of the identified DEG in enriched functions showed no homologies in model organisms and consequently remain uncharacterised. This is the case in 36 DE genes involved in the carbohydrate metabolism (over 88 total), whose function was evidenced by the presence of a protein domain, but with an unclear role. A similar case is noted in two genes involved in the "generation of precursor metabolites", twelve genes involved in the “secondary metabolism”, and 17 genes involved in “response to stress”.