A friend in need is a friend indeed: Understanding stress-associated transcriptional networks of plant metabolism using cliques of coordinately expressed genes.
ABSTRACT The response of plants to environmental cues, particularly stresses, involves the coordinated induction or repression of gene expression. In a previous study, we developed a bioinformatics approach to analyze the mutual expression pattern of genes encoding transcription factors and metabolic enzymes upon exposure of Arabidopsis plants to abiotic and biotic stresses. The analysis resulted in three gene clusters, each displaying a unique expression pattern. In the present addendum, we address the composition of each of these three clusters in regard to the functional identity of their encoded proteins as enzymes or transcription factors.
- SourceAvailable from: Jedrzej Szymanski[Show abstract] [Hide abstract]
ABSTRACT: Glycerolipid metabolism of plants responds dynamically to changes in light intensity and temperature, leading to the modification of membrane lipid composition to ensure optimal biochemical and physical properties in the new environment. Although multiple posttranscriptional regulatory mechanisms have been reported to be involved in the process, the contribution of transcriptional regulation remains largely unknown. Here, we present an integrative analysis of transcriptomic and lipidomic data, revealing large-scale coordination between gene expression and changes in glycerolipid levels during the Arabidopsis thaliana response to light and temperature stimuli. Using a multivariate regression technique called O2PLS, we show that the gene expression response is strictly coordinated at the biochemical pathway level and occurs in parallel with changes of specific glycerolipid pools. Five interesting candidate genes were chosen for further analysis from a larger set of candidates identified based on their close association with various groups of glycerolipids. Lipidomic analysis of knockout mutant lines of these five genes showed a significant relationship between the coordination of transcripts and glycerolipid levels in a changing environment and the effects of single gene perturbations.The Plant Cell 03/2014; · 9.58 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Plants need to continuously adjust their transcriptome in response to various stresses that lead to inhibition of photosynthesis and the deprivation of cellular energy. This adjustment is triggered in part by a coordinated re-programming of the energy-associated transcriptome to slow down photosynthesis and activate other energy-promoting gene networks. Therefore, understanding the stress-related transcriptional networks of genes belonging to energy-associated pathways is of major importance for engineering stress tolerance. In a bioinformatics approach developed by our group, termed 'gene coordination', we previously divided genes encoding for enzymes and transcription factors in Arabidopsis thaliana into three clusters, displaying altered coordinated transcriptional behaviors in response to multiple biotic and abiotic stresses (Plant Cell, 23, 2011, 1264). Enrichment analysis indicated further that genes controlling energy-associated metabolism operate as a compound network in response to stress. In the present paper, we describe in detail the network association of genes belonging to six central energy-associated pathways in each of these three clusters described in our previous paper. Our results expose extensive stress-associated intra- and inter-pathway interactions between genes from these pathways, indicating that genes encoding proteins involved in energy-associated metabolism are expressed in a highly coordinated manner. We also provide examples showing that this approach can be further utilized to elucidate candidate genes for stress tolerance and functions of isozymes.The Plant Journal 01/2012; 70(6):954-66. · 6.82 Impact Factor
Plant Signaling & Behavior 6:9 1294-1296; September 2011; ©2011 Landes Bioscience
A friend in need is a friend indeed
Understanding stress-associated transcriptional networks
of plant metabolism using cliques of coordinately
1294 Plant Signaling & Behavior Volume 6 Issue 9
*Correspondence to: Gad Galili; Email: Gad.Galili@weizmann.ac.il
Submitted: 5/23/11; Revised: 7/5/11; Accepted: 7/5/11
Being sessile organisms, plants have to continuously adjust their
gene expression programs to changing environments. Hence,
different external cues may trigger a coordinated expression
of different sets of genes, making it difficult to identify such
expression coordination by classical tools such as Pearson corre-
lation. In a recent report,1 we described a dedicated bioinformat-
ics approach that is able to elucidate specific cues (particularly
abiotic and biotic stresses) in which the change in expression of
specific sets of genes is either positively coordinated (genes are
either co-induced or co-suppressed) or negatively coordinated
(one set is induced and the second set is suppressed). Our study
analyzed the entire collections of Arabidopsis thaliana genes
predicted to encode transcription factors (TFs) and metabolic
enzymes (see details in reference 1), and two steps were used: (1)
assembly of genes into cliques showing co-expression in a num-
ber of cues that is higher than a random background; and (2)
clustering of the cliques. This approach resulted in the forma-
tion of three clusters; the genes in cluster 1 are up-regulated in a
coordinated manner; the genes of cluster 3 are down-regulated
in a coordinated manner. Unexpectedly, the genes of cluster 2
showed a more complex expression pattern, being principally
coordinately down regulated in response to abiotic stresses
and coordinately up-regulated in response to biotic stresses
and UVB stress. Gene overrepresentation enrichment analysis1
had shown that: (1) cluster 1 is enriched in genes belonging to
stress hormone metabolism and to stress-related TFs; (2) clus-
ter 2 is enriched in genes belonging to various pathways, such
as many pathways functioning in energy-associated metabolism
The response of plants to environmental cues, particularly stresses, involves the coordinated induction or repression of
gene expression. In a previous study, we developed a bioinformatics approach to analyze the mutual expression pattern
of genes encoding transcription factors and metabolic enzymes upon exposure of Arabidopsis plants to abiotic and biotic
stresses. The analysis resulted in three gene clusters, each displaying a unique expression pattern. In this article, we
address the composition of each of these three clusters in regard to the functional identity of their encoded proteins as
enzymes or transcription factors.
Tamar Avin-Wittenberg, Vered Tzin, Hadar Less, Ruthie Angelovici and Gad Galili*
Department of Plant Sciences; Weizmann Institute of Science; Rehovot, Israel
Key words: Arabidopsis, stress, metabolism, transcription, bioinformatics, plant
and amino acid metabolism (particularly the Asp-family path-
way that synthesizes the essential amino acids Lys, Thr, Met and
Ile). This enrichment supports previous research illustrating the
important metabolic link of the Asp-family pathway with the
TCA cycle;2-4 and (3) cluster 3 is enriched for genes belonging to
various pathways, such as photosynthesis, tetrapyrrole synthesis,
lipid metabolism and amino acid metabolism.
Since the initial analysis was performed on a group of genes
comprised of enzymes and TFs, in the present addendum article
we first set out to examine the amount of genes encoding enzymes
and TFs in each cluster (Figure 1). Surprisingly, we found that
the ratio between the number of coordinately expressed genes
encoding TFs and enzymes varies between clusters. While in
cluster 1 the amounts of coordinately expressed genes encoding
enzymes and TFs are similar, in clusters 2 and 3 there are many
more coordinately expressed genes encoding enzymes than those
encoding TFs. These results may indicate a different expression
regulation of genes in each cluster. The relatively equal propor-
tions of genes encoding TFs and metabolic enzymes in cluster
1 can be explained by the possibility that coping with stresses
requires activation of multiple regulatory and metabolic genes.
This supposition is strengthened by the enrichment analysis of
cluster 1. The relatively low number of enriched categories in
cluster 1 suggests that the genes in this cluster operate in many
different pathways. The few categories that are enriched in this
cluster are all stress related.
In contrast to cluster 1, clusters 2 and 3 exhibit a much higher
number of coordinately expressed genes encoding enzymes
www.landesbioscience.com Plant Signaling & Behavior 1295
compared to those encoding TFs; with cluster 2 displaying the
most extreme case. In the case of cluster 3, an explanation may
lie in the enrichment analysis. Two greatly enriched pathways in
this cluster are photosynthesis and tetrapyrrole synthesis, path-
ways that are expected to be strongly regulated during stresses
due to their potential negative impact on generating reactive
oxygen species.5-8 Indeed, analysis of genes from these two path-
ways shows very high expression coordination values between
them (data not shown). This may suggest strong co-regulated
synthesis of the enzymes that can be performed by a relatively
small number of TFs, with every TF apparently influencing the
expression of many enzymes simultaneously. The situation in
cluster 2 is more extreme than in cluster 3, namely, the ratio
between TFs and enzymes being about 1:6 compared to about
1:2 in cluster 3. This phenomenon may stem from several
1. The genes encoding enzymes in cluster 2 may be exten-
sively co-regulated by very few TFs.
2. Our analysis used genes that encode TFs according to the
Arabidopsis TF database. It is possible that some genes that func-
tion as TFs regulating the enriched processes in cluster 2 may
have not yet been recognized as possessing this exact function
and therefore, were excluded from our analysis. Future analy-
sis of the whole genome using the gene coordination method
developed in our group may shed light on the functions of some
unknown genes and perhaps add further genes encoding TFs to
3. Our gene coordination approach is based on the recogni-
tion of gene pairs whose expression changes in a similar manner
under a number of conditions above a random background (in
our case, 16 conditions). If a TF is expressed during specific cues
for a very short time, it might affect the production of many
genes encoding enzymes, but will not be coordinated with the
synthesis of these enzymes because the number of coordinated
conditions will not be higher than the random background.
Thus, the enzymes in cluster 2 might be regulated by TFs whose
expression changes for very short periods of time and were,
therefore, excluded from the cluster.
In addition to examining the total number of genes encod-
ing enzymes and TFs in each cluster, we analyzed the number
of genes encoding enzymes and TFs in each of the cliques com-
prising the different clusters. As shown in Figure 2, most of the
cliques contain genes encoding both enzymes and TFs (except
for a few small cliques in cluster 2 that contain genes encoding
only enzymes). In addition, there is a positive linear correlation
Figure 1. number of genes encoding enzymes and transcription factors
(TFs) in each cluster.
Figure 2. number of genes encoding enzymes and TFs per clique in
each cluster. The number of genes encoding enzymes (black diamonds)
and TFs (gray squares) in each clique is plotted as a function of the
clique size. The Linear regression equation and R-square value are
displayed for each plot.
1296 Plant Signaling & Behavior Volume 6 Issue 9
1. Less H, Angelovici R, Tzin V, Galili G. Coordinated
Gene Networks Regulating Arabidopsis Plant
Metabolism in Response to Various Stresses and
Nutritional Cues. Plant Cell 2011; 23:1264-71.
Angelovici R, Fait A, Fernie AR, Galili G. A seed high-
lysine trait is negatively associated with the TCA cycle
and slows down Arabidopsis seed germination. New
Phytol 2011; 189:148-59.
Angelovici R, Fait A, Zhu X, Szymanski J, Feldmesser
E, Fernie AR, Galili G. Deciphering transcriptional and
metabolic networks associated with lysine metabolism
during Arabidopsis seed development. Plant Physiol
Araujo WL, Ishizaki K, Nunes-Nesi A, Larson TR,
Tohge T, Krahnert I, Witt S, Obata T, Schauer N,
Graham IA, Leaver CJ, Fernie AR. Identification of the
2-hydroxyglutarate and isovaleryl-CoA dehydrogenases
as alternative electron donors linking lysine catabolism
to the electron transport chain of Arabidopsis mito-
chondria. Plant Cell 2010; 22:1549-63.
5. Strand A, Asami T, Alonso J, Ecker JR, Chory J.
Chloroplast to nucleus communication triggered by
accumulation of Mg-protoporphyrinIX. Nature 2003;
Saibo NJ, Lourenco T, Oliveira MM. Transcription
factors and regulation of photosynthetic and related
metabolism under environmental stresses. Ann Bot
Christianson JA, Llewellyn DJ, Dennis ES, Wilson IW.
Global gene expression responses to waterlogging in
roots and leaves of cotton (Gossypium hirsutum L.).
Plant Cell Physiol 2010; 51:21-37.
Alboresi A, Dall’osto L, Aprile A, Carillo P, Roncaglia
E, Cattivelli L, Bassi R. Reactive oxygen species and
transcript analysis upon excess light treatment in wild-
type Arabidopsis thaliana vs a photosensitive mutant
lacking zeaxanthin and lutein. BMC Plant Biol 2011;
between the number of genes encoding enzymes or TFs in each
clique and the clique size. In cluster 1, there is a higher correlation
value between the number of genes encoding TFs and clique size
(R2 = 0.90) than between the number of genes encoding enzymes
and clique size (R2 = 0.69). The contribution of genes encoding
TFs to the clique size is greater than those encoding enzymes,
as can be seen from the slope values (+0.67 and +0.34, respec-
tively). In contrast to cluster 1, in clusters 2 and 3, the clusters
in which the total number of genes encoding enzymes is much
higher than those encoding TFs (Figure 1), the genes encoding
enzymes account for most of the clique size. The contribution of
genes encoding TFs to the size of the cliques in cluster 2 is very
low and quite variable, as can be seen by the low coefficient of
determination value (R2 = 0.53) and the low slope value (+0.13)
of the regression line. This is different in cluster 3, where the
contribution of TFs to clique size is less variable (R2 = 0.71) and
more substantial (slope = +0.25).
Looking at specific cliques may also provide the unique oppor-
tunity to examine small, highly coordinated networks of genes
encoding enzymes and TFs and possibly assess their biological
functions and regulatory interplay. As an example of this issue,
we analyzed clique number 70, which contains 24 genes encoding
five TFs and 19 enzymes (see Supplemental Table 2, reference 1).
Pageman over-presentation analysis (mapman.mpimp-golm.
mpg.de/general/ora/ora.shtml) of this clique revealed significant
enrichment of nine enzymes belonging to amino acid metabo-
lism, mainly amino acid degradation. This clique includes genes
encoding catabolic enzymes of the Asp-family pathway, namely,
Lys-Ketoglutarate Reductase/Saccharopine Dehydrogenase (Lys
catabolism)4,9, Branched-chain amino acid Aminotransferase
(likely functioning in Ile catabolism)4,10 and Thr Aldolase (con-
verting Thr into Gly).11 In addition, this clique includes genes
encoding five TFs that possess specific TFs domains, but their
functions have not yet been characterized. Therefore, we pos-
tulate that the genes grouped in clique number 70 may encode
enzymes or TFs serving functions that are related to the catab-
olism of the Asp-family enzyme, and could serve as candidate
genes for further study of the catabolism of amino acids of the
Asp family pathway.
This research was supported by grants from The Israel Science
Foundation (grant No. 764/07) and from the United States–
Israel Binational Agricultural Research and Development Fund
(grant no. IS–3331–02). GG is an incumbent of the Bronfman
Chair of Plant Science at the Weizmann Institute of Science.
9. Tang G, Miron D, Zhu-Shimoni JX, Galili G.
Regulation of lysine catabolism through lysine-keto-
glutarate reductase and saccharopine dehydrogenase in
Arabidopsis. Plant Cell 1997; 9:1305-16.
10. Diebold R, Schuster J, Daschner K, Binder S. The
branched-chain amino acid transaminase gene fam-
ily in Arabidopsis encodes plastid and mitochondrial
proteins. Plant Physiol 2002; 129:540-50.
11. Joshi V, Laubengayer KM, Schauer N, Fernie AR,
Jander G. Two Arabidopsis threonine aldolases are non-
redundant and compete with threonine deaminase for a
common substrate pool. Plant Cell 2006; 18:3564-75.