Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Transcriptome Analysis of Potato Leaves Expressing the Trehalose-6-Phosphate Synthase 1 Gene of Yeast

Abstract

Transgenic lines of the potato cultivar White Lady expressing the trehalose-6-phosphate synthase (TPS1) gene of yeast exhibit improved drought tolerance, but grow slower and have a lower carbon fixation rate and stomatal density than the wild-type. To understand the molecular basis of this phenomenon, we have compared the transcriptomes of wild-type and TPS1-transgenic plants using the POCI microarray containing 42,034 potato unigene probes. We show that 74 and 25 genes were up-, and down-regulated, respectively, in the mature source leaves of TPS1-transgenic plants when compared with the wild-type. The differentially regulated genes were assigned into 16 functional groups. All of the seven genes, which were assigned into carbon fixation and metabolism group, were up-regulated, while about 42% of the assigned genes are involved in transcriptional and post-transcriptional regulation. Expression of genes encoding a 14-3-3 regulatory protein, and four transcription factors were down-regulated in the TPS1-transgenic leaves. To verify the microarray results, we used RNA gel blot analysis to examine the expression of eight genes and found that the RNA gel blot and microarray data correlated in each case. Using the putative Arabidopsis orthologs of the assigned potato sequences we have identified putative transcription binding sites in the promoter region of the differentially regulated genes, and putative protein-protein interactions involving some of the up- and down-regulated genes. We have also demonstrated that starch content is lower, while malate, inositol and maltose contents are higher in the TPS1-transgenic than in the wild-type leaves. Our results suggest that a complex regulatory network, involving transcription factors and other regulatory proteins, underpins the phenotypic alterations we have observed previously in potato when expressing the TPS1 gene of yeast.

Introduction

One major approach to improve drought tolerance in crop species is to express genes encoding either metabolic enzymes or transcription factors, which exert their effects through various mechanisms of action [1]. Genes of different origins involved in trehalose metabolism have been used in a number of plant species to improve their drought tolerance [2]. Trehalose, a non-reducing disaccharide consisting of two glucose molecules is a very abundant sugar in nature. In bacteria, yeast and desiccation-tolerant plants it accumulates under osmotic/dehydration stress [3], and helps cells to survive by protecting membranes and proteins [4]. In other plants, however, trehalose is synthesised at an almost undetectable level. In Escherichia coli, yeast and plants, trehalose is synthesised in a two-step process. First, trehalose-6-phosphate (T6P) is synthesised from glucose-6-phosphate (G6P) and UDP-glucose (UDPG) by trehalose phosphate synthase (TPS) and then T6P is converted into trehalose by trehalose phosphatase (TPP). In the yeast Saccharomyces cerevisiae, an enzyme complex, consisting of trehalose-6-phosphate synthase (TPS) and trehalose-6-phosphate phosphatase (TPP) encoded by the genes TPS1 and TPS2, respectively, is involved in the synthesis of trehalose, while in Escherichia coli the corresponding genes are otsA and otsB. The trehalose biosynthetic genes in plants have been mainly studied in Arabidopsis, and three classes of proteins have been distinguished based on domain structure, similarity to the yeast TPS1 and TPS2 genes, and the absence/presence of phosphatase boxes in the TPP domain. Ectopic expression of TPS1 and otsA in different plant species and overexpression of AtTPS1 in Arabidopsis improved drought tolerance, but had diverse effects on plant development and resulted in other phenotypic changes in certain species [2], [3].

To improve drought tolerance of potato (Solanum tuberosum), we have previously introduced the TPS1 gene of yeast into the cultivar White Lady, under the control of a drought-inducible potato promoter, StDS2 [5]. Although the transgenic plants became drought tolerant, it was determined that the transgene was expressed at a very low level even under optimal growth conditions and the transgenic plants displayed certain morphological and physiological changes when compared with the wild-type. For example, they grew slower, had a lower CO2 fixation rate and stomatal density was reduced by about 35% [6].

Our observations, and the results of others, highlight the importance of analysis of transgenic plants in order to understand how and why the inserted genes can have such pleiotropic effects. To study this, we have analysed the transcriptome of the wild-type and TPS1-transgenic potato plants under unstressed conditions using a microarray, which contains 42K potato unigene probe sequences [7]. Statistical analysis revealed that 99 genes are expressed differentially, and functional annotation revealed that a number of genes are associated with carbohydrate metabolism, while a large proportion (about 42%) of the genes with known function are involved in transcriptional and translational regulation of gene expression. Our results suggest that complex regulation operating at different levels might underpin the observed phenotypical and biochemical changes of the TPS1-transgenic potato plants.

Results

Physiological changes in potato plants expressing the TPS1 gene of yeast

Control (S. tuberosum cv. White Lady) and two TPS1-transgenic lines [6] were grown under well-watered conditions as described in the Materials and methods section. Five physical and biochemical parameters of the lines were then measured, and the results are shown in Table 1. Water and protein content of the transgenic plants did not change compared to the wild-type. Chlorophyll content of the transgenic leaves was slightly, but not significantly, higher than in the wild-type leaves. In contrast, shoot mass and leaf area of the TPS1-transgenic lines were, on average, about 35 and 24% lower, respectively, than in the wild-type.

thumbnail
Table 1. Measured parameters, as indicated, of potato plants grown under optimal conditions.

https://doi.org/10.1371/journal.pone.0023466.t001

Transcriptome analysis of the TPS1-transgenic plants

Previously published observations [6] and the morphological and physiological changes detected in TPS1-transgenic plants grown under well-watered conditions described above prompted us to investigate this phenomenon further. To do this, we performed a transcriptome analysis using a potato microarray [8], with which we could monitor expression of a large number of genes simultaneously. Total RNA was isolated from fully expanded leaves of six-week-old plants with the characteristics shown in Table 1. Total RNA was transcribed into fluorescently-labelled cDNA, which was then hybridised to the microarrays in three technical repeats per biological replicates. Images of the hybridised microarrays were analysed by ArrayPro software, and within-array. Loess-normalised data were collected from nine parallel microarrays. After removing bad-spot data, datasets were quantile-normalised between arrays, which adjusts variations in microarray data arising from technical rather than any biological differences (Figure S1). Thus data from the technical and biological replicates become more comparable, a prerequisite for statistical analysis. Between-array-normalised data were then transformed into log2 values and exported into the web-tool ArrayMiner for statistical analysis. Using the empirical Bayesian option of ArrayMiner, 99 genes with a q-value lower than 0.05 were returned, and we consider this to represent significant differences in gene expression between TPS1-transgenic and wild-type leaves across all nine arrays. Of these genes, 74 were up- and 25 were down-regulated in the TPS1-transgenic plants.

Annotation of the differentially expressed genes

The 99 differentially expressed genes, which we have identified in the microarray experiments, were exported into the MapMan software for functional annotation. Of these 99 genes, 53 were assigned into different functional groups (bins), while the bin of “not assigned” genes contains 46 genes (Table S1) of which 36 encode unknown, hypothetical or putative proteins. To confirm the annotations of the differentially expressed genes, we performed a BLAST analysis of the potato unigene sequences from which the microarray oligonucleotide probes were designed [7] against the recently completed genome sequence of the doubled monoploid Solanum phureja DM1-3 516R44 [Potato Genome Sequencing Consortium, http://www.potatogenome.net/index.php/Main_Page], which is phylogenetically the same species as S. tuberosum [8]. Homologies were displayed in the potato genome browser [http://www.potatogenome.net/index.php/Main_Page] and both UniProt [www.uniprot.org] Solanaceae entries and corresponding Arabidopsis genes of the genomic loci were recorded (Table 2 and Table S2). After this second approach, seven of the MapMan-assigned 53 genes were discarded because of discrepancies between the MapMan and genome data. Thus in total, 46 assigned genes have been obtained (Table 2), and were used in further analyses.

thumbnail
Table 2. Functional annotation of differentially expressed genes.

https://doi.org/10.1371/journal.pone.0023466.t002

TPS1-transgenic potato plants display reduced growth and CO2 fixation rate [6], which may be linked to carbohydrate metabolism at the molecular level. Therefore, the seven genes in Table 2, which are associated with carbon fixation and metabolism, and are up-regulated in the TPS1-transgenic plants, may be significant. Four of these genes are associated with photosynthesis, two with major carbohydrate metabolism, and one with the tricarboxylic acid cycle. Of the photosynthesis-associated genes, pbsY encodes a photosystem II thylakoid membrane protein [9] and RbcS is a nuclear gene family member encoding small subunits of the Rubisco complex localised in the chloroplast stroma [10]. Phosphoglycerate kinase (PGK1) and fructose bisphosphate aldolase (ALDP1) are also stromal proteins, both having a function in the Calvin cycle, while ALDP1 is also involved in glycolysis [11]. One of the major carbohydrate-metabolism-associated genes encodes a sucrose synthase, which catalyzes the conversion of sucrose into UDP-glucose and fructose. The particular gene (SUS3), which is up-regulated in TPS1-transgenic plants, was shown to be expressed at the highest levels in stems and roots of non-transgenic plants [12]. The other gene in this functional group is an adenine nucleotide carrier protein [13].

A large proportion (about 42%) of the assigned genes belongs to functional groups of RNA, DNA and protein-associated genes. One of these is an Alfin1-like PHD-finger transcription factor, a second is an AG-motif binding protein 4 (AGP4), similar to the GATA zinc-finger transcription factor GATA5 and two others are similar to the MADS-box TFs AGL8 and AGL24. All four TF genes were down-regulated in TPS1-transgenic plants. In general, PHD-finger proteins are thought to be chromatin mediated transcriptional regulators, but one of them, Alfin1 is a promoter-binding TF [14]. GATA factors are zinc finger domain-containing DNA binding TFs, which are involved in diverse developmental and environmental pathways, including responses to light. It is worth to note that the corresponding Arabidopsis protein of AGP4, GATA5, is expressed in all mature plant tissues at an almost constant level, and is up-regulated in light-grown plants [15]. MADS-box proteins are DNA-binding TFs involved in plant developmental processes, including floral development and transition between vegetative and reproductive phases [16]. Based on the UniProt database [http://www.uniprot.org], there are two corresponding proteins of AGL8 in potato, POTM1 and SCM1, which are 96% identical to each other. Suppression lines of POTM1, which belongs to the same MADS-clade as AGL8 [17], produce truncated shoot clusters from stem buds and exhibit enhanced axillary bud growth instead of producing a tuber [18]. The other MADS-box protein, AGL24, is homologous to StMADS11 and belongs to the StMADS11 clade of MIKC-type MADS-box proteins [19]. STMADS11 was isolated from S. tuberosum, and is expressed in all vegetative tissues [17]. In the protein-associated functional group, there are two genes, which are up-regulated in TPS1-transgenic plants and encode calmodulin-binding proteins. Calmodulins are Ca2+-binding proteins, which interact with a large number of structurally and functionally diverse proteins [20]. Five genes, four of which are up- and one down-regulated, encode proteins involved in the ubiquitin-proteasome pathway. Two of these proteins are a ubiquitine-conjugating (E2) and a ubiquitine-ligase (E3) protein working in a cascade to ubiquitinate target proteins, which then are transferred into the 20S proteolytic unit of the 26S proteasome for degradation [21]. One protein is a Kelch repeat-containing F-box family protein, which is a subunit of the E3 ubiquitine-ligase complex [22]. Two other genes encode 20S proteasome components. E2 and E3 proteins are encoded by large gene-families [23] and different combinations of these proteins provide for very selective ubiquitination and consequently degradation of cellular proteins in proteasomes [21]. A CLP protease (CLPP) was also up-regulated in TPS1-transgenic plants. CLPP is one of those chloroplast proteases, which is located in the stroma, and forms a proteolytic complex with other proteases and is assumed to be a housekeeping protease [24].

Another gene in Table 2 encodes for a protein that is homologous to GRF3 of Arabidopsis, which is a (ψ)-type 14-3-3 protein (TFT4) expressed in stems, leaves and flowers. 14-3-3 proteins, which are ubiquitous in animals and plants, bind commonly, but not exclusively to phosphorylated target proteins and are considered of great significance because they act as central regulators of metabolism and signalling in plants [25].

Verification of microarray results

Because the two TPS1-transgenic lines did not show any significant differences in terms of the measured physical and biochemical parameters (Table 1), only one of the lines, T2, was used for the microarray experiments to compare its transcriptome with the non-transformed control. However, to check the reliability of our microarray results, both TPS1-transgenic lines were analysed in RNA gel blot analysis. For this, total RNA, isolated from the leaves of the wild-type and TPS1-transgenic lines, was separated on agarose gels, transferred to membranes and probed with gene-specific radioactive probes for eight of the differentially regulated genes. Thus the expression of about 17% of the assigned genes was assessed by RNA blot analysis. After scanning the autoradiographs, the ratio of the signal between the TPS1-transgenic and the wild-type lines was calculated. Line T1, whose transcriptome was not examined by microarray, gave similar results to line T2 (Figure 1A) indicating that, very likely, gene expression in both lines follows the same pattern. Although the expression ratios for all tested genes were slightly different in the microarray and the RNA gel blot experiments, they had a strong positive correlation with an r-value of 0.9369 (Figure 1B). Thus our microarray results can be assumed to be correct and reliable.

thumbnail
Figure 1. Validation of microarray results.

A) RNA gel blot analysis of selected genes. Phosphorimage analysis was used to quantify the intensity of hybridisation. The expression ratios between T1, T2 and wild-type plants are shown below the lanes as log2 values. Ethidium bromide-stained rRNA bands are shown as loading controls. WT, wild-type; T1, T2, two independent TPS1-transgenic lines. B) Correlation between microarray and northern results in T2/wild-type comparison. A statistically significant correlation (r = 0.9369, p = 2.376e−06) was obtained for all genes tested.

https://doi.org/10.1371/journal.pone.0023466.g001

Leaf carbohydrate and starch content

The microarray results showed that a sucrose synthase gene (SUS3) and six other genes associated with photosynthesis and carbon metabolism are up-regulated in TPS1-transgenic leaves. We investigated, therefore, the relative levels of major carbohydrates and starch in the TPS1-transgenic leaves as compared with the wild-type levels.

Carbohydrates and starch were extracted from the same pool of leaves used for microarray analysis. GC-MS analysis revealed that the amounts of D-fructose, D-galactose, D-glucose, sorbitol, and sucrose are largely similar in each line (data not shown), while the amounts of inositol and maltose, and the organic acid, malate, are increased in the TPS1-transgenic lines (Figure 2A). In terms of starch content, the levels in the wild-type showed a large variation between the biological repeats (4.4, 1.7, 0.5 µmol hexose equivalent/g FW), but were always proportionally higher than in TPS1-transgenic leaves (Figure 2B). Photosynthetic partitioning into starch is finely regulated, and the amount of carbohydrate stored is dependent upon the environmental conditions experienced by the plant, particularly day length [26]. Thus small differences in day length, light intensity and temperature in the greenhouse during the three consecutive plant tests may explain the variation in starch content. When compared with starch, much less variation was detected in malate (17.9±1.37 µmol/g FW), inositol (0.91±0.03 µmol/g FW) and maltose (0.52±0.01 µmol/g FW) content. Although TPS1 mRNA was detectable even under unstressed conditions in the transgenic plants [6], no trehalose (<0.006 µmol hexose equivalent/g FW) was detected either in wild-type or TPS1-transgenic leaves. Under water-deficit-stress, the level of TPS1 mRNA slightly increased compared to well-watered conditions [6], but trehalose was still undetectable (data not shown). This may be due to the high trehalase activity detected in dicotyledonous plant species [2].

thumbnail
Figure 2. Relative amounts of sugars (A) and starch (B) in wild-type (WT) and TPS1-transgenic (T1, T2) leaves.

Bars and error bars represent the mean ± SE derived from three independent experiments. Filled and open circles denote differences significant at P = 0.01 and P = 0.05 (t probe) levels, respectively, when compared with the wild-type. The carbohydrate concentrations in wild-type leaves in the three independent experiments were as follows: malate 19.4, 16.7, 17.6 µmol/g FW; inositol 0.89, 0.9, 0.94 µmol/g FW; maltose 0.51, 0.53, 0.52 µmol/g FW; starch 4.4, 1.7, 0.5 µmol hexose equivalent/g FW. These are regarded as 100% values for comparison with the equivalent samples from the transgenic leaves.

https://doi.org/10.1371/journal.pone.0023466.g002

Interaction analysis

The large proportion of genes associated with transcriptional and translational regulation identified in the microarray experiments prompted us to investigate some elements of the putative regulatory network that may underpin the observed differential gene expression. Unfortunately, in potato, such interaction information is very limited. For example, protein orthologs from solanaceous species are not listed in the databases of OMA (http://www.cbrg.ethz.ch/research/orthologous/index), Roundup (http://roundup.hms.harvard.edu/site/index.php), and BioGRID (http://thebiogrid.org) for protein-protein interactions. In another interaction database, IntAct (http://www.ebi.ac.uk/intact/main.xhtml), there are only 27 interactions out of 266,855 which have S. tuberosum proteins as the interacting partner, but only a fraction of these (seven interactions) are real binary interactions; the others are co-complexes. Thus acknowledging the problems associated with gene orthology in general, and in particular in the case of gene families [27], our analyses had to be based on putative Arabidopsis orthologs of the assigned potato sequences (Table 2). We assumed that orthologs have the same function in different species [28], and so for the protein-protein interaction analyses, we chose only those genes (28 in all) which have corresponding genes in both Arabidopsis and Solanaceae (Table S2), while putative TF-binding site analyses were performed for all genes in Table 2.

Firstly, we have searched the AthaMap database for transcription factors, which bind to the −500 to +50 region of the assigned genes. In particular, we were interested to see whether the four transcription factors (Alfin1, C2C2 GATA, AGL8, AGL24) that have been detected as differentially expressed genes in our microarray experiments bind to the promoter region of putative orthologs of the assigned potato genes. The results of the search are shown in Table S3. Alfin1 is predicted to bind to the promoter region of all but seven genes. Bound promoters include AGL8, AGL24 and C2C2 GATA, but not the Alfin1 gene itself, suggesting that Alfin1 does not regulate its own synthesis at the transcription level. At the time of writing this manuscript, there was no entry in the AthaMap database for the two MADS-box proteins, AGL8 and AGL24, the orthologs of which are down-regulated in the TPS1-transgenic plants. However, the MADS-box protein, AGL15, does bind to the promoter region of all four TFs, while another, AGL2 binds to the promoter of the C2C2 GATA gene (Table S3).

We also performed an additional search of the AtcisDB database. While AthaMap returns information about TFs which bind to the promoter region of Arabidopsis genes, AtcisDB contains information about the DNA motifs to which the TFs bind. This search revealed the presence of CArG MADS protein-binding boxes in the promoter of AGL8 and Alfin1. GATA binding sites were identified in the promoter region of all four TFs. In addition to these specific findings, a number of additional putative TF-binding sites were identified in the promoter region of all assigned genes, revealing a very complex matrix of the assigned genes, the TFs and their binding sites (Table S4).

As the part of our analysis, we have investigated whether any protein-protein interactions occur between the proteins encoded by the up- and down-regulated genes. For this, the BioGRID and IntAct databases were interrogated using the Locus ID of the corresponding Arabidopsis genes of 28 potato genes (Table 2). We have found that 13 proteins have proven binary interactions in the databases. All of these interactions are shown in Figure 3 and in Table S5. We have identified two interactions in which both partners are proteins encoded by orthologs of differentially expressed potato genes. Namely, these are the NDPK1/CAT2, and the AGL8/AGL24 interactions. The biggest network of interacting proteins is associated with the MADS-box proteins, AGL8 and AGL24, which also interact with each other. They have twelve and 14 interacting partners, respectively, with seven of these proteins common to both. Their interacting partners are mostly Agamous, Agamous-like or other MADS-box proteins, but AGL8 also interacts with three calmodulins. MADS-box proteins form homo- or hetero-dimers and are considered as combinatorial transcription factors [29], which explains the interaction of AGL8 and AGL24 with a number of other MADS-box proteins. The second largest protein network consists of eleven proteins involved in carbohydrate and nucleotide metabolism, redox and protein processes. Some proteins of this network are connected by some common interacting partners, such as a protein kinase and a ubiquitin-protein ligase. The 14-3-3 protein has three interacting partners, a nitrate reductase, another general regulatory factor (GRF2) and a transcription initiation factor. Some minor interactions, involving 2 or 3 interacting proteins were also identified.

thumbnail
Figure 3. Networks of interacting proteins.

Proteins encoded by differentially expressed genes from the microarray experiments and their interacting partners are labelled with black and grey circles, respectively. Proteins which are present on both the microarray and the partner lists (Table S5) are labelled by red circles. Proteins with similar functions are boxed: 1, kinases; 2, ubiquination; 3, calmodulins/calmodulin binding. AGL16, a guard cell-specific transcription factor interacting with AGL24, an Agamous-like transcription factor, is in boldface (see text for explanation). AGL24 interacts with itself too, indicated with a circled line. For protein descriptions and available Solanaceae protein names see Table S5.

https://doi.org/10.1371/journal.pone.0023466.g003

Discussion

In a previous paper we have reported that introducing the TPS1 gene of yeast into S. tuberosum cv. White Lady resulted in drought-tolerance accompanied by certain pleiotropic effects, which could be observed even under well-watered conditions [6]. In the Solanaceae, similar studies have been undertaken with tobacco and tomato by introducing either the same TPS1 gene or its E. coli ortholog, otsA, under the control of a constitutive (CaMV35S), tissue specific (Rubisco, patatin, 16SrRNA) or drought-inducible (AtRAB18) promoter [2], [3]. Almost all of the transformed plants displayed drought-tolerance, but the strong constitutive promoter combined with either TPS1 or otsA caused phenotypical changes in all three species, similar to those observed by us. While these previous studies focused on the biochemical aspects of drought-tolerance in transgenic solanaceous plants, in the current study we have applied genomic and bioinformatic approaches to investigate the differences between wild-type and transgenic potato plants and in doing so have identified 99 genes, which are either up- or down-regulated in the leaves of TPS1-transgenic potato plants when compared with the wild-type under well-watered conditions.

Previously, it has been shown that T6P inhibits SnRK1 activity in extracts of Arabidopsis seedling and leaf tissues of different ages, with the exception of mature leaves [30]. This study also found that seedlings expressing otsA show opposite regulation of the SnRK1 target genes. By comparing this study with our microarray results we have found 22 assigned potato genes out of 46 whose expression was changed both in otsA-transgenic Arabidopsis seedlings and in mature leaves of TPS1-transgenic potatoes. However, 10 of these displayed an inverse regulation (Table 2) suggesting that further work is necessary to unequivocally establish the effect of T6P on SnRK1 in mature leaves of potato.

The expression of two ortholog genes, GRF3 and TFT4, which encode a member of the 14-3-3 protein family, is reversed in the otsA and TPS1-transgenic plants, respectively. TFT4 is down-regulated in TPS1-transgenic plants which also displayed reduced leaf area, while GRF3 influences leaf growth in Arabidopsis [31]. Among the proteins with which binding of 14-3-3 proteins has been demonstrated is a cauliflower TPS [32] while TPS5, 6, and 7 of Arabidopsis also bind to 14-3-3 proteins if the Ser22 and Thr49 residues are phosphorylated [33]. Since the TPS1 protein of yeast is 40% identical with TPS5, and TPS1 contains Ser and Thr residues at the same locations as the Arabidopsis TPS isoforms, we postulate that phosphorylation of yeast TPS1 and its interaction with 14-3-3s may also exist in potato. Binding of yeast TPS1 to 14-3-3 can influence the activities of house-keeping proteins such as nitrate reductase whose interaction with GRF3 has been demonstrated in Arabidopsis (Figure 3). The down-regulation of TFT4 might result in reduced 14-3-3 availability leading to an imbalance in ion homeostasis and hormone signalling in which 14-3-3s have well understood functions [25].

Four transcription factors, similar to two MADS-box proteins (AGL8 and AGL24), a GATA factor and an Alfin1 TF, are down-regulated in the leaves of TPS1-transgenic plants. These four TFs may bind to the promoter regions of other differentially expressed genes in a very complex pattern (Table S3 and S4). In addition, the two MADS-box proteins interact with a number of other MADS factors (Table S5), which is in good agreement with the proposed combinatorial regulation of vegetative development by MADS factors [19].

Two genes, SUI1B and EF1a, that encode proteins, which bind Ca2+-binding calmodulin proteins, are up-regulated in TPS1-transgenic potato plants. In addition, AGL8 also interacts with three calmodulins (Figure 3 and Table S5) and together this indicates that Ca-signalling might have an important role in the TPS1-transgenic plants. We also propose that protein phosphorylation via calmodulins, which has an important role not only in regulatory cascades but also in protein degradation, may be affected. We have also found that several genes, encoding proteins involved in the ubiquitin-proteasome pathway, are differentially expressed, mostly up-regulated, in the TPS1-transgenic plants when compared with the wild-type (Table 2). Furthermore, an ubiquitin-ligase (UBC2) and the Rubisco small subunit (RBCS-3B), the genes for which are up-regulated in TPS1-transgenic plants, interact with HUB1 and CHIP ubiquitin-ligases, respectively. Altogether there are five proteins in the protein networks which are involved in ubiquitination. In addition to their numerous roles in regulation, 14-3-3 proteins can either inhibit or promote degradation of phosphorylated protein to which they bind [34], and so the down-regulation of a gene encoding such a protein can influence the turnover of a number of proteins. Together, these findings indicate that posttranslational regulation might also have a role in the development of the phenotypes observed in the TPS1-transgenic plants.

In terms of regulatory cascades, it is now well established that plants respond to environmental stresses via mechanisms involving sugar signalling and hormonal factors [35]. It is thus not surprising that we have found hormone metabolism and stress-related genes that are differentially regulated in the TPS1-transgenic plants even under well-watered conditions (Table 2). In addition, a number of genes assigned into other functional groups are known to be stress-responsive (data not shown). A recent example of such interlocking regulatory cascades is the observation that mutant plants with impaired nitrate reductase are also dehydration resistant [36].

It appears to be inconsistent that a Rubisco small subunit gene is up-regulated (Table 2), while CO2 assimilation is reduced in the TPS1-transgenic plants [6]. However, it has been shown that the protein abundance and activity of Rubisco is not always correlated with changes in the amount of rbcS transcript. Moreover, the CO2 fixing reaction catalysed by Rubisco is reversible and molecular oxygen can also be a substrate of the Rubisco complex [37]. It is possible therefore, that the up-regulation of rbcS has an effect on these reactions and together with the inherently complex regulation of Rubisco results in the net reduction in CO2 assimilation, which may also be the primary reason of the decreased starch content of leaves. Another explanation for the observed reduced CO2 assimilation might be the lower stomata density observed in TPS1-transgenic plants [6]. It has been shown that the density and development of stomatal complexes on the epidermis of Arabidopsis thaliana leaves depend, in part, on the microRNA-mediated regulation of AGL16 [38], which is a member of the MADS-box protein family and expressed in guard cells [39]. In this respect it is intriguing therefore that AGL16 is among the interacting partners of AGL24 whose corresponding gene (StMADS11) is down-regulated in TPS1-transgenic plants. Assuming that these proteins exist in potato, an altered interaction between them, due to the down-regulation of StMADS11, might be an alternative explanation for the lower stomata density and reduced CO2 assimilation rate of TPS1-transgenic leaves.

Although, the level of pyruvate dehydrogenase, phosphoglycerate kinase and fructose-bisphosphate aldolase mRNAs are increased in TPS1-transgenic potato leaves no significant changes in levels of major carbohydrates, glucose, fructose, and sucrose were detected while the amounts of malate, inositol and maltose were increased. Pathways interconnecting these enzymes and metabolites are shown in Figure S2. It is worth noting that carbohydrate metabolism in drought tolerant transgenic plants harbouring different trehalose biosynthetic enzymes can be very different. For example, introducing the otsA and otsB genes into rice resulted in slightly elevated levels of glucose, fructose and sucrose under both well-watered and drought-stress conditions [40], while in our potato lines transformed with the TPS1 gene the levels of these sugars remained quite constant under all conditions. This very likely reflects certain differences in the carbohydrate metabolism between dicots and monocots.

So, putting all information together we propose that molecular interactions and complex regulatory mechanisms at transcriptional, translational and post-translational levels underpin the pleiotropic effects in drought-tolerant potato, harbouring the TPS1 gene of yeast.

Materials and Methods

Ethics statement

This work did not raise ethical issues.

Plant material and growth conditions

Solanum tuberosum cv. White Lady wild-type and TPS1-transgenic plants T1 and T2 [6] were vegetatively propagated from single-node stem segments in tissue culture and maintained at 24°C under a 16 h light / 8 h dark regime on RM medium [41]. Six-week-old plants were transferred to pots containing A260 sterile soil (Stender, Germany) and were grown in a greenhouse in summer, under natural light, at 20–28°C, and at a soil water content of 70%. After six weeks in the greenhouse, mature source leaves of vegetative growth-phase plants were sampled for further analysis, four hours after sunrise.

Physical and biochemical measurements

To determine total shoot mass, the entire aerial part of plants were harvested by excising the stem one cm above the soil and the collected material weighted. The moisture content of leaves was calculated from the fresh and dry weight and expressed as the percentage of the fresh weight. To measure leaf area, freshly collected leaves were scanned and their area was determined using PhotoShop software. To measure chlorophyll and protein contents, leaves were powdered in liquid nitrogen and 500 mg of the powder was vortexed in 5 ml of ice-cold acetone, followed by centrifugation at 13000 rpm for 10 min at 4°C to remove cell debris. The supernatant was neutralised with an equal volume of 1 M Tris-HCl, pH 8.0, and then the absorbance was measured at 645 and 663 nm. Chlorophyll content was calculated from the absorbance data and expressed as mg per g fresh weight. For protein measurement, 100 mg of the powdered leaf material was vigorously mixed with 400 µl of 0.1 M Na-phosphate buffer, pH 7.8, followed by centrifugation at 13000 rpm for 30 min at 4°C. The concentration of the total soluble proteins in the supernatant was measured using a dye-binding method [42].

Extraction, derivatisation and analysis of potato leaf carbohydrates were carried out as described by [43] using a quadrupole-type GC-MS system (Finnigan Trace/DSQ, Thermo Electron Corp.). The chromatograms and mass spectra were evaluated using the XCALIBUR software (Thermo Electron Corp.) and the NIST 2.0 library.

Starch was isolated from 150 mg of leaf tissue powdered in liquid nitrogen by incubation in 1 ml of a solvent containing 80% (v/v) ethanol and 5% (v/v) formic acid at 80°C for 10 min. After centrifugation at 13000 rpm for 10 min the supernatant was removed and the pellet resuspended in 1 ml of 80% (v/v) ethanol and incubated at 80°C for 5 min. The pellet was harvested again by centrifugation and after washing twice with 80% (v/v) ethanol, the starch was solubilised with 400 µl of 0.2 N KOH at 95°C for 1 h. The solution was neutralised by 70 µl of 1N acetic acid and cleared by centrifugation after which 100 µl of the supernatant was mixed with 10 µl of Lugol solution (2 g KI, 1 g I2 dissolved in 150 ml distilled water) in a 96-well microplate, and the developed colour measured at 595 nm in a Multiskan EX (Labsystems) microplate reader. A calibration curve was prepared, using solubilised corn starch (Sigma) as control, to determine starch content in the leaf samples.

RNA isolation and cDNA synthesis

For the microarray experiments, plants were grown in three biological replicates, each containing six plants. All fully expanded leaves of all plants of each replicate were pooled and then total RNA was extracted from the pools as described [44]. Fluorescently labelled cDNA was synthesised from 20 µg total RNA using a SuperScript Plus Direct labelling kit (Invitrogen), according to the manufacturer's instructions. Alexa Fluor 647- and 555-labelled dUTPs were used for the transgenic and wild-type samples, respectively, and anchored oligo (dT) was used as the primer. The labelled cDNA was purified using a MinElute PCR purification kit (Qiagen), and was quantified using a NanoDrop spectrophotometer. The cDNA yield varied between 1.06 and 2.31 µg, while specific dye incorporation was between 25 and 113 pmol dye/µg cDNA.

Microarray processing

POCI potato microarrays (4×44 K; [7]) were purchased from Agilent. One microgram of each of the Alexa Fluor 647- and 555-labelled cDNAs were combined, dried in a SpeedVac, and dissolved in 20 µl of water. The hybridisation mixture was set up using a Gene Expression Hybridisation Kit (Agilent) and applied to the microarray, which was then hybridised at 65°C for 17 h, followed by two subsequent washes in the appropriate Agilent wash solutions for one minute each at RT and 37°C, respectively. The dried microarrays were then scanned with a Genetix microarray scanner at 100% laser power and 50–70% gain settings.

Data analysis and mining

Three technical replicate microarray hybridisations per biological replicates were performed, so altogether we obtained data from nine arrays. Scan images were analysed using ArrayPro software, and the raw signals were within-array-normalised using the local regression (Loess) function of the software. Bad quality (empty or dirty) spots were manually removed and data between arrays were quantile-normalised in Excel [45]. Box and whisker plots [46] were created using an Excel template (http://www.vertex42.com/ExcelTemplates/box-whisker-plot.html). Quantile-normalised data were log2 transformed and statistical analysis was performed by the empirical Bayesian option of the web-tool ArrayMiner [47], which applies the method of Benjamini and Hochberg [48] to adjust the significance threshold to prevent false positive discoveries. A q-value, an adjusted p-value for multiple testing, as a significance score for each gene was returned, and we considered genes having a q-value smaller than 0.05 as significant discoveries. Genes displaying significant changes in expression were annotated into functional categories using the MapMan software [49], [50]. Potato microarray unigene sequences were analysed by BLAST against the S. phureja genome (http://www.potatogenome.net/index.php/Main_Page) for further annotation. Data for Solanaceae proteins and Arabidopsis genes and proteins were collected from the UniProt (http://www.uniprot.org) and the TAIR (http://www.arabidopsis.org/) databases, respectively. Identification of transcription factors and their binding sites in the promoter regions of the putative Arabidopsis orthologs of the assigned genes was performed by searching the AtcisDB [51] and AthaMap [52] databases. Protein-protein interactions were identified using the BioGrid (http://www.thebiogrid.org) and IntAct (http://www.ebi.ac.uk/intact/main.xhtml) databases, and were visualised with software Osprey [53]. Microarray data were submitted to ArrayExpress under accession number E-MEXP-3221.

RNA gel blot analysis

RNAs quantified using a NanoDrop spectrophotometer were loaded in equal amounts (20 µg) into agarose gels, separated and blotted onto Hybond-N membranes as described [54]. To generate hybridisation probes, PCR amplifications were carried out using a S. tuberosum leaf cDNA library [55] as template and gene-specific primers (Table S6). PCR products were isolated from agarose gels using the GFX PCR DNA and Gel Band Purification Kit (GE Healthcare), and radioactively labelled by random priming [54]. Hybridisation was carried out in Church buffer [56] at 65°C for overnight. The filter was washed for 20 min at 65°C twice in 2xSSC [54] containing 0.1% (v/v) SDS and once in 0.2xSSC with 0.1% (v/v) SDS.

Supporting Information

Figure S1.

Between array normalisation of microarray data. Data for TPS1-transgenic (A) and wild-type plants (B) from nine microarrays were quantile normalised.

https://doi.org/10.1371/journal.pone.0023466.s001

(PDF)

Figure S2.

Pathway map of carbohydrate metabolism. The map is based on the KEGG database (http://www.genome.jp/kegg/pathway.html). Red, green and black letters and dots represent increased, decreased and unchanged gene expression and carbohydrate levels, respectively, in TPS1-transgenic potato leaves when compared with the wild-type. I, inositol; F, fructose; G, galactose; Gl, glucose; M, maltose; Mt, malate; S, sorbitol; Su, sucrose; FBA, fructose-bisphosphate aldolase; PDH, pyruvate dehydrogenase; PGK, phosphoglycerate kinase; RBC, RuBisCo; SUS, sucrose synthase.

https://doi.org/10.1371/journal.pone.0023466.s002

(PDF)

Table S1.

List of differentially expressed genes not assigned to functional categories using MapMan software.

https://doi.org/10.1371/journal.pone.0023466.s003

(DOC)

Table S2.

Functional annotation of differentially expressed potato genes. Three approaches were used, functional grouping using MapMan software, BLAST searching of the S. phureja genome and searching the UniProt database for Solanaceae orthologs.

https://doi.org/10.1371/journal.pone.0023466.s004

(XLS)

Table S3.

Transcription factors binding to the promoter region of putative Arabidopsis orthologs of differentially expressed potato genes.

https://doi.org/10.1371/journal.pone.0023466.s005

(XLS)

Table S4.

Transcription factor binding sites in the promoter region of putative Arabidopsis orthologs of differentially expressed potato genes.

https://doi.org/10.1371/journal.pone.0023466.s006

(XLS)

Table S5.

Protein-protein interactions involving putative Arabidopsis orthologs of potato proteins encoded by assigned genes in microarray experiments.

https://doi.org/10.1371/journal.pone.0023466.s007

(XLS)

Table S6.

PCR primers used to amplify cDNA fragments to generate hybridisation probes for verification of the microarray results.

https://doi.org/10.1371/journal.pone.0023466.s008

(DOC)

Acknowledgments

We thank Dr. Ibolya Molnár-Perl (Institute of Chemistry, Department of Analytical Chemistry, L. Eötvös University, Budapest) for help in GC-MS analysis, Monika Kiss for technical assistance and Dr. Zoltán Havelda for discussion. The authors are very grateful to Professor Michael McManus (Massey University, Palmerston North, New Zealand) and Dr. Iain Manfield (University of Leeds, Leeds, United Kingdom) for their critical reading of the manuscript.

Author Contributions

Conceived and designed the experiments: ZB MK FM. Performed the experiments: MK FM KB JZ. Analyzed the data: MK FM BK ZJ. Wrote the paper: ZB FM.

References

  1. 1. Cattivelli L, Rizza F, Badeck F-W, Mazzucotelli E, Mastrangelo AM, et al. (2008) Drought tolerance improvement in crop plants: An integrated view from breeding to genomics. Field Crops Res 105: 1–14.
  2. 2. Iturriaga G, Suárez R, Nova-Franco B (2009) Trehalose metabolism: from osmoprotection to signaling. Int J Mol Sci 10: 3793–3810.
  3. 3. Iordachescu M, Imai R (2008) Trehalose biosynthesis in response to abiotic stresses. J Integr Plant Biol 50: 1223–1229.
  4. 4. Jain NK, Roy I (2009) Effect of trehalose on protein structure. Protein Sci 18: 24–36.
  5. 5. Dóczi R, Csanaki C, Bánfalvi Z (2002) Expression and promoter activity of the desiccation-specific Solanum tuberosum gene, StDS2. Plant Cell Environ 25: 1197–1203.
  6. 6. Stiller I, Dulai S, Kondrák M, Tarnai R, Szabó L, et al. (2008) Effects of drought on water content and photosynthetic parameters in potato plants expressing the trehalose-6-phosphate synthase gene of Saccharomyces cerevisiae. Planta 227: 299–308.
  7. 7. Kloosterman B, De Koeyer D, Griffiths R, Flinn B, Steuernagel B, et al. (2008) Genes driving potato tuber initiation and growth: identification based on transcriptional changes using the POCI array. Funct Integr Genomics 8: 329–340.
  8. 8. Spooner DM, Núñez J, Trujillo G, Herrera del Rosario, Guzmán F, et al. (2007) Extensive simple sequence repeat genotyping of potato landraces supports a major reevaluation of their gene pool structure and classification. PNAS 104: 19398–19403.
  9. 9. Friso G, Giacomelli L, Ytterberg AJ, Peltier J-B, Rudella A, et al. (2004) In-depth analysis of the thylakoid membrane proteome of Arabidopsis thaliana chloroplasts: new proteins, new functions, and a plastid proteome database. Plant Cell 16: 478–499.
  10. 10. Sawchuk MG, Donner TJ, Head P, Scarpella E (2008) Unique and overlapping expression patterns among members of photosynthesis-associated nuclear gene families in Arabidopsis. Plant Physiol 148: 1908–1924.
  11. 11. Rutschow H, Ytterberg AJ, Friso G, Nilsson R, van Wijk KJ (2008) Quantitative proteomics of a chloroplast SRP54 sorting mutant and its genetic interactions with CLPC1 in Arabidopsis. Plant Physiol 148: 156–175.
  12. 12. Fu H, Park WD (1995) Sink- and vascular-associated sucrose synthase functions are encoded by different gene classes in potato. Plant Cell 7: 1369–1385.
  13. 13. Millar AH, Heazlewood JL (2003) Genomic and proteomic analysis of mitochondrial carrier proteins in Arabidopsis. Plant Physiol 131: 443–453.
  14. 14. Wei W, Huang J, Hao Y-J, Zou H-F, Wang H-W, et al. (2009) Soybean GmPHD-type transcription regulators improve stress tolerance in transgenic Arabidopsis plants. PLoS ONE 4: 7209.
  15. 15. Manfield IW, Devlin PF, Jen CH, Westhead DR, Gilmartin PM (2007) Conservation, convergence, and divergence of light-responsive, circadian-regulated, and tissue-specific expression patterns during evolution of the Arabidopsis GATA gene family. Plant Physiol 143: 941–958.
  16. 16. Messenguy F, Dubois E (2003) Role of MADS box proteins and their cofactors in combinatorial control of gene expression and cell development Gene 316: 3–21.
  17. 17. Carmona MJ, Ortega N, Garcia-Maroto F (1998) Isolation and molecular characterization of a new vegetative MAD-box gene from Solanum tuberosum L. Planta 207: 181–188.
  18. 18. Rosin FM, Hart JK, Van Onckelen H, Hannapel DJ (2003) Suppression of a vegetative MADS box gene of potato activates axillary meristem development. Plant Physiol 131: 1613–1622.
  19. 19. de Folter S, Immink RGH, Kieffer M, Pařenicová L, Henz SR, et al. (2005) Comprehensive interaction map of the Arabidopsis MADS box transcription factors. Plant Cell 17: 1424–1433.
  20. 20. Kim MC, Chung WS, Yun D-J, Cho MJ (2009) Calcium and calmodulin-mediated regulation of gene expression in plants. Mol Plant 2: 13–21.
  21. 21. Myung J, Kim KB, Crews CM (2001) The ubiquitin-proteasome pathway and proteasome inhibitors. Med Res Rev 21: 245–273.
  22. 22. Sun Y, Zhou X, Ma H (2007) Genome-wide analysis of Kelch repeat-containing F-box family. J Integr Plant Biol 49: 940–952.
  23. 23. Kraft E, Stone SL, Ma L, Su N, Gao Y, et al. (2005) Genome analysis and functional characterization of the E2 and RING-Type E3 ligase ubiquitination enzymes of Arabidopsis. Plant Physiol 139: 1597–1611.
  24. 24. Sjögren LLE, Stanne TM, Zheng B, Sutinen S, Clarke AK (2006) Structural and functional insights into the chloroplast ATP-dependent Clp protease in Arabidopsis. Plant Cell 18: 2635–2649.
  25. 25. Oecking C, Jaspert N (2009) Plant 14-3-3 proteins catch up with their mammalian orthologs. Curr Opin Plant Biol 12: 760–765.
  26. 26. Zeeman SC, Smith SM, Smith AM (2007) The diurnal metabolism of leaf starch. Biochem J 401: 13–28.
  27. 27. Fang G, Bhardwaj N, Robilotto R, Gerstein MB (2010) Getting started in gene orthology and functional analysis. PLoS Comput Biol 6(3): e1000703.
  28. 28. Tatusov RL, Koonin EV, Lipman DJ (1997) A genomic perspective on protein families. Science 278: 631–637.
  29. 29. Immink RG, Kaufmann K, Angenent GC (2010) The ‘ABC’ of MADS domain protein behaviour and interactions. Semin Cell Dev Biol 21: 87–93.
  30. 30. Zhang Y, Primavesi LF, Jhurreea D, Andralojc PJ, Mitchell RA, et al. (2009) Inhibition of SNF1-related protein kinase1 activity and regulation of metabolic pathways by trehalose-6-phosphate. Plant Physiol 149: 1860–1871.
  31. 31. Kim JH, Choi D, Kende H (2003) The AtGRF family of putative transcription factors is involved in leaf and cotyledon growth in Arabidopsis. Plant J 36: 94–104.
  32. 32. Moorhead G, Douglas P, Cotelle V, Harthill J, Morrice N, et al. (1999) Phosphorylation-dependent interactions between enzymes of plant metabolism and 14-3-3 proteins. Plant J 18: 1–12.
  33. 33. Harthill JE, Meek SE, Morrice N, Peggie MW, Borch J, et al. (2006) Phosphorylation and 14-3-3 binding of Arabidopsis trehalose-phosphate synthase 5 in response to 2-deoxyglucose. Plant J 47: 211–223.
  34. 34. Huber SC, MacKintosh C, Kaiser WM (2002) Metabolic enzymes as targets for 14-3-3 proteins. Plant Mol Biol 50: 1056–1063.
  35. 35. Rosa M, Prado C, Podazza G, Interdonato R, González JA, et al. (2009) Soluble sugars: Metabolism, sensing and abiotic stress: A complex network in the life of plants. Plant Signal Behav 4: 388–393.
  36. 36. Lozano-Juste J, León J (2010) Enhanced abscisic acid-mediated responses in nia1nia2noa1-2 triple mutant impaired in NIA/NR- and AtNOA1-dependent nitric oxide biosynthesis in Arabidopsis. Plant Physiol 152: 891–903.
  37. 37. Saibo NJ, Lourenço T, Oliveira MM (2008) Transcription factors and regulation of photosynthetic and related metabolism under environmental stresses. Ann Bot 103: 609–623.
  38. 38. Kutter C, Schöb H, Stadler M, Meins F , Si-Ammour A (2007) MicroRNA-mediated regulation of stomatal development in Arabidopsis. Plant Cell 19: 2417–2429.
  39. 39. Alvarez-Buylla ER, Liljegren SJ, Pelaz S, Gold SE, Burgeff C, et al. (2000) MADS-box gene evolution beyond flowers: Expression in pollen, endosperm, guard cells, roots and trichomes. Plant J 24: 457–466.
  40. 40. Garg AK, Kim JK, Owens TG, Ranwala AP, Choi YD, et al. (2002) Trehalose accumulation in rice plants confers high tolerance levels to different abiotic stresses. PNAS 99: 15898–15903.
  41. 41. Murashige T, Skoog F (1962) A revised medium for rapid growth and bioassays with tobacco tissue culture. Physiol Plantarum 15: 473–497.
  42. 42. Bradford M (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72: 248–254.
  43. 43. Schauer N, Zamir D, Fernie AR (2004) Metabolic profiling of leaves and fruit of wild species tomato: a survey of the Solanum lycopersicum complex. J Exp Bot 56: 297–307.
  44. 44. Stiekema WJ, Heidekamp F, Dirkse WG, van Beckum J, de Haan P, et al. (1988) Molecular cloning and analysis of four potato tuber mRNAs. Plant Mol Biol 11: 255–269.
  45. 45. Smyth GK, Speed TP (2003) Normalization of cDNA microarray data. Methods 31: 265–273.
  46. 46. Mitra A (1998) Fundamentals of Quality Control and Improvement, 2nd edn. Prentice Hall, New Jersey.
  47. 47. Glaab E, Garibaldi J, Krasnogor N (2009) ArrayMining: a modular web-application for microarray analysis combining ensemble and consensus methods with cross-study normalization. BMC Bioinformatics 10: 358.
  48. 48. Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate – a practical and powerful approach to multiple testing. J Roy Stat Soc B Met 57: 289–300.
  49. 49. Thimm O, Bläsing O, Gibon Y, Nagel A, Meyer , S , et al. (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.
  50. 50. Rotter A, Usadel B, Baebler S, Stitt M, Gruden K (2007) Adaptation of the MapMan ontology to biotic stress responses: application in solanaceous species. Plant Methods 3: 10.
  51. 51. Palaniswamy SK, James S, Sun H, Lamb RS, Davuluri RV, et al. (2006) AGRIS and AtRegNet. A platform to link cis-regulatory elements and transcription factors into regulatory networks. Plant Physiol 140: 818–829.
  52. 52. Bülow L, Engelmann S, Schindler M, Hehl R (2009) AthaMap, integrating transcriptional and post-transcriptional data. Nucleic Acids Res 37: D983–D986.
  53. 53. Breitkreutz BJ, Stark C, Tyers M (2003) Osprey: a network visualization system. Genome Biol 4: R22.
  54. 54. Sambrook J, Fritsch EF, Maniatis T (1989) Molecular Cloning: A Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor.
  55. 55. Bánfalvi Z, Kostyál Z, Barta E (1994) Solanum brevidens possesses a non-sucrose-inducible patatin gene. Mol Gen Genet 245: 517–522.
  56. 56. Church GM, Gilbert W (1984) Genomic sequencing. Proc Natl Acad Sci USA 81: 1991–1995.