Transcriptomic comparison in the leaves of two aspen genotypes having similar carbon assimilation rates but different partitioning patterns under elevated [CO2]

Authors


  • Dedication: This paper is published in memory of Dr David F. Karnosky, Professor, School of Forest Resources and Environmental Science, Michigan Technological University and director of the Aspen FACE site.

Author for correspondence:
Gopi K. Podila
Tel: +1 256 824 6263
Email: podilag@uah.edu

Summary

  • • This study compared the leaf transcription profiles, physiological characteristics and primary metabolites of two Populus tremuloides genotypes (clones 216 and 271) known to differ in their responses to long-term elevated [CO2] (e[CO2]) at the Aspen free-air CO2 enrichment site near Rhinelander, WI, USA.
  • • The physiological responses of these clones were similar in terms of photosynthesis, stomatal conductance and leaf area index under e[CO2], yet very different in terms of growth enhancement (0–10% in clone 216; 40–50% in clone 271). Although few genes responded to long-term exposure to e[CO2], the transcriptional activity of leaf e[CO2]-responsive genes was distinctly different between the clones, differentially impacting multiple pathways during both early and late growing seasons.
  • • An analysis of transcript abundance and carbon/nitrogen biochemistry suggested that the CO2-responsive clone (271) partitions carbon into pathways associated with active defense/response to stress, carbohydrate/starch biosynthesis and subsequent growth. The CO2-unresponsive clone (216) partitions carbon into pathways associated with passive defense (e.g. lignin, phenylpropanoid) and cell wall thickening.
  • • This study indicates that there is significant variation in expression patterns between different tree genotypes in response to long-term exposure to e[CO2]. Consequently, future efforts to improve productivity or other advantageous traits for carbon sequestration should include an examination of genetic variability in CO2 responsiveness.

Introduction

Atmospheric CO2 concentration ([CO2]) has risen by ~30% in the last 250 yr, and data from monitoring stations, together with historical records extracted from ice cores, show that atmospheric [CO2] is now at a level higher than at any time in the last 650 000 yr (Meehl et al., 2007). Driven by the addition of 6–8 Pg carbon yr−1 from anthropogenic sources, atmospheric [CO2] is predicted to continue to rise by an additional 50% by 2050 (Meehl et al., 2007). The Intergovernmental Panel on Climate Change (IPCC) estimates that between 46 and 56% of terrestrial carbon stocks are found in forest biomes, and that actions to preserve and enhance this carbon sink would probably increase the global terrestrial carbon stock by 60–87 Pg carbon by 2050, thereby offsetting c. 15% of the anthropogenic emissions predicted for the same period (Prentice, 2001; Brown, 2002).

Aggressive afforestation is part of the action required to meet this potential for increasing the terrestrial carbon stock, and managed plantations of highly productive tree species, such as members of the genus Populus, are an attractive system to help achieve this goal (Deckmyn et al., 2004; Perlack et al., 2005). Recent advances in poplar genomic resources include the collections of over 410 000 expressed sequence tags (ESTs) in GenBank (http://www.ncbi.nlm.nih.gov/dbEST/), the development of cDNA and whole-genome microarrays (Déjardin et al., 2004; Broschéet al., 2005; Rinaldi et al., 2007) and the full genome sequence available from P. trichocarpa (http://genome.jgi-psf.org/Poptr1_1/Poptr1_1.home; Tuskan et al., 2006). These advances have promoted the idea that poplar species hold great potential to be bred or engineered for an increased suitability for carbon sequestration (Sims et al., 2006).

Plants, including trees, sense and respond to increasing [CO2] through increased photosynthesis and reduced stomatal conductance (Ainsworth & Rogers, 2007; Hyvönen et al., 2007; Liberloo et al., 2007; Riikonen et al., 2008; Taylor et al., 2008). All other effects of elevated [CO2] (e[CO2]) on plants and ecosystems are derived from these two fundamental responses. However, genetic and environmental bottlenecks can determine both the magnitude of these primary responses to e[CO2] and the capacity to assimilate carbon into increased above-ground biomass (Karnosky et al., 2005; Liberloo et al., 2006; Luo et al., 2006, 2008; Ainsworth & Rogers, 2007). In trees, growth is dependent on the internal balance between carbon and nitrogen, and this balance depends not only on nitrogen resources directly available from soil and internal sources, but also on nitrogen resources from seasonal storage pools in bark, wood and roots (Luo et al., 2008). A few studies have explored the transcriptomics of individual tree species as they respond to e[CO2] (Gupta et al., 2005; Taylor et al., 2005; Druart et al., 2006). Each study identified relatively few transcriptional changes in response to e[CO2], and the results between species were quite variable, making it difficult to draw solid conclusions. Although the responses of photosynthesis, growth and biomass accumulation in trees grown at e[CO2] are well documented (Nowak et al., 2004; Ainsworth & Long, 2005; Norby et al., 2005; Ainsworth & Rogers, 2007), the molecular mechanisms that determine how different tree species achieve a balance between carbon and nitrogen assimilation, storage and eventual growth remain largely unknown. Consequently, there is a need to increase our understanding of metabolic/physiological processes that may limit the response of trees to increasing [CO2].

Free-air CO2 enrichment (FACE) provides a realistic platform on which to investigate the response of trees to e[CO2] (Long et al., 2004) and, with particular reference to this study, avoids artificial restriction of sink capacity, canopy development and nutrient supply. In this study, we have taken advantage of an ongoing FACE experiment located near Rhinelander, WI, USA (Karnosky & Pregitzer, 2006), referred to as the Aspen FACE site (http://aspenface.mtu.edu/). Here, five distinct genotypes of quaking aspen (Populus tremuloides) have been grown at e[CO2] (ambient plus 200 ppb) since 1997.

Poplar trees growing under e[CO2] generally have a larger leaf size (Oksanen et al., 2001; Riikonen et al., 2008), increased stem and branch growth (Isebrands et al., 2001; Karnosky et al., 2005; King et al., 2005; Kubiske et al., 2007) and increased root biomass (King et al., 2001; Lukac et al., 2003). However, at Aspen FACE, CO2 responsiveness was found to be genotype dependent. Clone 271 is highly responsive to e[CO2], whereas the growth of clone 216 is not significantly stimulated by e[CO2], despite the fact that both clones show similar increases in photosynthetic rates under e[CO2] (Noormets et al., 2001b; Riikonen et al., 2008; Taylor et al., 2008). Clone 271 has delayed senescence and develops ~50% more stem biomass than clone 216 in response to e[CO2] (Table 1) (Karnosky et al., 2005; Kubiske et al., 2007). By focusing on the tissues that most directly sense and respond to e[CO2], this paper provides an initial examination of the abundance of leaf gene transcripts and biochemical responses in growth-responsive (clone 271) and growth-unresponsive (clone 216) genotypes of P. tremuloides grown at ambient and e[CO2]. It is hypothesized that such an analysis of clones with markedly different growth responses to e[CO2] will allow us to identify some of the genetic trends associated with a sustained utilization of an increased carbon supply and a superior capacity for above-ground biomass production. To our knowledge, this is the first study to directly compare the leaf transcriptomes of tree genotypes of the same species having similar carbon uptake, but very different growth responses when grown at e[CO2] in the field.

Table 1. Summary of physiological changes in response to long-term elevated [CO2] (e[CO2]) in aspen (Populus tremuloides) clones 216 and 271
  Clone 216 response to e[CO2]Clone 271 response to e[CO2]Reference(s)
  1. ns, no significant response; +, trend of increase (< 10%); −, trend of decrease (< 10%); ++, significant increase (10–30%); −−, significant decrease (10–30%); +++, large and significant increase (> 30%); −−−, large and significant decrease (> 30%).

Leaf criteria
 Photosynthesis++++++ Noormets et al. (2001b); Riikonen et al. (2008); Taylor et al. (2008)
 Chlorophyll content−−−−−− Wustman et al. (2001)
 Rubisco−−ns Wustman et al. (2001)
 Stomatal conductance−−−− Noormets et al. (2001a); Riikonen et al. (2008)
 Stomatal frequencynsns Karnosky et al. (2003); Mankovska et al. (2005)
 Leaf area index (LAI)++++ Karnosky et al. (2005) Riikonen et al. (2008); Taylor et al. (2008)
 Leaf thickness++++ Oksanen et al. (2001, 2003)
 Leaf cell wall thicknessns−− Oksanen et al. (2003)
 Leaf surface waxesns++ Karnosky et al. (2002); Mankovska et al. (2005)
 Leaf isoprene emissionsnsns Calfapietra et al. (2007)
 Leaf cytoplasmic lipids+++ Oksanen et al. (2001)
 Leaf starch content++++ Kopper & Lindroth (2003)
 Leaf total nitrogen−−−− Kopper & Lindroth (2003)
 Rust occurrence+ns Karnosky et al. (2002); Mankovska et al. (2005)
 Tent caterpillar performancens−− Kopper & Lindroth (2003)
 Delayed autumn senescence+++ Riikonen et al. (2008); Taylor et al. (2008)
Stem/wood criteria
 Diameter growth++++ Isebrands et al. (2001); Kubiske et al. (2007)
 Volume growth++++ Karnosky et al. (2005); Kubiske et al. (2007)
 Height growthns+++ Isebrands et al. (2001); Kubiske et al. (2007)
 Cell wall percentage+ns Kaakinen et al. (2004)
 Wood starch contentns+ Kaakinen et al. (2004)
 Wood total nitrogen−− Kaakinen et al. (2004)

Materials and Methods

Experimental site and plant material

The study was performed on leaf tissues collected near Rhinelander, WI, USA (89.5°W, 45.7°N) at the Aspen FACE experimental site, which has been operating since 1997. The design and operation of this experimental site have been described in detail elsewhere (Dickson et al., 2001; Karnosky et al., 2005; Karnosky & Pregitzer, 2006). In short, 12, 30-m diameter, plots are contained within a 32-ha site in a full factorial design. The plots are exposed to ambient air (control) and air enriched with CO2, ozone, or CO2 and ozone in combination. In this study, only plots receiving ambient air ([CO2] = 372 µl l−1) and air enriched with CO2 (target [CO2] = 560 µl l−1) were used. The site is divided into three blocks, each containing a CO2 and a control plot. e[CO2] plots were fumigated from sunrise to sunset for the length of the growing season (May–September), and 1-min averages of [CO2] in the e[CO2] plots were within 20% of the target concentration. The soil at the Aspen FACE site is a sandy loam and is characteristic of a moderate to poor soil for the region. Nitrogen is considered to be the principal limiting nutrient, and no supplemental fertilizer has been added (Ellsworth et al., 2004; Karnosky & Pregitzer, 2006).

As a result of the significant differences in growth response to e[CO2], this study utilized two trembling aspen (Populus tremuloides Michx.) genotypes: clone 216 and clone 271. To avoid variation derived from different leaf emergence dates on indeterminate shoots, leaves were sampled from short shoots. Short shoots have a determinate growth form and all leaves expand rapidly at the same time, at the start of the growing season (Cox, 2005). Likewise, to avoid variation resulting from leaf developmental stage, only mature, fully expanded leaves were used for analysis. As bud-break in this region occurred on 19 May and bud-set occurred in late August (for clone 216) and early September (for clone 271), harvest dates of 15 June and 17 August 2005 were chosen to ensure recently mature early-season leaf collections and mature, but not yet senescing, leaf collections in the late-season time point. Upper canopy sunlit leaves were harvested between 10:00 and 11:00 h on each date. Tissues were frozen in liquid nitrogen within 20 s of harvesting and stored at −80°C until analysis. To avoid potential variation derived from time-of-day effects, all pairs of control vs. e[CO2] samples were collected within 10 min of one another.

Physiological data collection

Stem volume growth comparisons of aspen clones 216 and 271 in response to e[CO2] were collected for 8 yr, as described previously (Karnosky et al., 2003; Kubiske et al., 2007). Maximum instantaneous photosynthesis (Amax) of aspen clones 216 and 271 was compared under ambient and e[CO2] using an LI-6400 portable photosynthesis system (Li-Cor Inc., Lincoln, NE, USA), and the data were analyzed as described previously (Noormets et al., 2001a,b). Data were expressed as the mean of three leaves each from three trees measured during midday and replicated three times (6, 7 and 11 July 2006) from the full-sun canopy.

cDNA microarrays

The PICME (Platform for Integrated Clone Management; http://www.picme.at) Populus microarrays are composed of ~28 000 elements, including 6528 ESTs of Populus interamericana from INRA-Nancy (Rinaldi et al., 2007), 12 202 ESTs of Populus alba × Populus tremula from INRA-Orleans (Déjardin et al., 2004) and 8160 cDNA clones of Populus euphratica from the University of Helsinki (Broschéet al., 2005). This set of leaf, root and xylem cDNAs corresponds to ~10 000 different predicted gene models in the P. trichocarpa genome sequence (Tuskan et al., 2006; Rinaldi et al., 2007). The arrays are well documented for leaf transcript analysis (Rinaldi et al., 2007; Fluch et al., 2008) and are fully described under platform GPL6036 stored in the Gene Expression Omnibus (GEO) at the National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/geo).

We followed the widely accepted Minimum Information About a Microarray Experiment (MIAME) guidelines for microarray analysis and verification (Zimmermann et al., 2006). For each of the two-channel arrays, RNA collected from plants grown in the ambient (control) ring was compared directly with RNA derived from the same genotype grown under e[CO2], following cDNA synthesis and labeling procedures in the Superscript™ Indirect cDNA Labeling System (Invitrogen, Carlsbad, CA, USA). Three independent biological replicates, derived from trees grown in three independent replicate FACE rings, were analyzed using equal amounts (40 pmol) of labeled samples. In addition, each clone and time point included dye swap reciprocal two-color experiments for each biological replicate (Churchill, 2002; Allison et al., 2006). Thus, six data points per cDNA (three biological replicates with two technical replicates each) were used.

Slides were scanned using a VersArray ChipReader™ scanner (BioRad, Hercules, CA, USA) at 5-µm resolution. Cyanine-3 (Cy3) and cyanine-5 (Cy5) images were aligned and spots were flagged using VersArray Analyzer 5.0 (BioRad). After local background subtraction, the signal intensity was log2 transformed and normalized by the LOWESS algorithm with a smoothing parameter of 0.2, using GeneGazer software (Bio-Rad). Normalized intensity values were filtered by a coefficient of variance (CV) cut-off of 0.25, and spots having intensities below 100 in both channels were also excluded from further analysis. Filtered gene lists were subjected to t-tests with a false discovery rate at 0.05 (P < 0.05) (see Fig. S1, Supporting Information, for examples of data linearity). Principal component analysis (PCA) was conducted using standard correlation and hierarchical cluster analysis (HCA) was performed using the Euclidean distance metric and average cluster linkage. The complete expression dataset, sample details, RNA extraction, cDNA preparations, labeling of cDNAs and hybridization procedures are available with the series accession number GSE14881 in GEO at NCBI.

Quantitative real-time reverse transcriptase-polymerase chain reaction (RT-PCR)

DNA-free total RNA was used for cDNA synthesis with oligo-dT primers, as described previously; however, it is important to note that independent RNA preparations were used for quantitative RT-PCR analyses. In this way, the quantitative RT-PCR samples represent additional biological replicates from the same field sampling. EST sequences for genes of interest from the microarray analysis were retrieved from GenBank and used to identify their cognate gene models from the Joint Genome Initiative (JGI) Populus genome portal. Because a large fraction of Populus genes are derived from genome-wide duplication (Tuskan et al., 2006), closely related gene family members were aligned in order to evaluate the likelihood of cross-hybridization on the EST arrays. Depending on sequence homology, gene- or group-specific primers (Table S1, see Supporting Information) flanking 137–290-bp amplicons near the 3′-untranslated regions (3′-UTRs) were designed on the basis of JGI-predicted cDNA sequences and the corresponding GenBank Populus EST sequences.

Real-time PCRs were performed using the Mx3000P Real-Time PCR System (Stratagene, La Jolla, CA, USA) with the ABsolute QPCR SYBR Green Mixes (ABgene Inc., Rochester, NY, USA). The PCR parameters were as follows: 15 min at 95°C, followed by 40 cycles of 15 s at 95°C, 1 min at 58–66°C (Table S1) and 1 min at 72°C. Each reaction was performed in duplicate with three biological replicates (from replicate rings) and no-template controls, using cDNA synthesized from 2.5 ng of total RNA. Because the housekeeping genes were expressed relatively strongly in these tissues, PCR amplification was performed using diluted cDNA (200-fold for cyclophilin and ubiquitin-conjugating enzyme E2, and four-fold for actin) in order to improve relative transcript abundance estimates of weakly expressed genes. The specificity of the amplification was verified using dissociation curve analysis at the end of each run, employing MxPro software (Stratagene). The threshold level was manually adjusted and applied consistently across all amplifications to obtain the threshold cycle (CT) values. Relative target transcript levels were normalized to the geometric mean of three housekeeping genes using the ΔCT method (Tsai et al., 2006).

Biochemical analysis

Frozen leaf tissues, derived from the same tissue collections as described above, were powdered in liquid nitrogen and transferred to microcentrifuge tubes; 375 µl of 80% (v/v) ethanol, 10 mM 2-(N-morpholino)ethanesulphonic acid (MES), pH 6.0 was added to equal quantities of tissue. The tubes were mixed thoroughly and incubated at 80°C for 30 min. Extracts were clarified by centrifugation (4500 g, 10 min), and the supernatant was decanted into a 96-well, deep-well microplate, stored at 4°C; 225 µl of 80% (v/v) ethanol, 10 mM MES pH 6, and 375 µl of 50% (v/v) ethanol, 10 mM MES pH 6, were added in two additional, separate incubations to the pellet and supernatants pooled in the deep-well plate.

Glucose, fructose, sucrose and starch content were determined from the supernatant and pellet resulting from the ethanol extract using the continuous enzymatic substrate assay described by Rogers et al. (2006). The total amino acid content in the ethanol extract was determined using a fluorogenic-based microplate assay, as described previously (Rogers et al., 2006). Protein content was determined using a commercial kit (Pierce, Rockford, IL, USA) as described previously (Ainsworth et al., 2007). Nitrate was determined by conversion to nitrite by nitrate reductase and subsequent colorimetric determination by the Griess reaction (Cross et al., 2006). Biochemical statistical analyses were based on the treatment plot as the experimental unit. Values for three leaves per clone per plot (subsamples) were averaged to provide the sample estimate for each treatment plot (n = 3 blocks). The experimental design was a split-split-plot with [CO2] as the main plot factor. Clone was nested within [CO2] as a split-plot factor and date was the split-split-plot factor. A mixed model ANOVA was used for statistical analysis (SYSTAT, SPSS Inc., Chicago, IL, USA, 2000).

Results and Discussion

Physiological measurements alone do not explain the superior ability of clone 271 to allocate carbon towards growth, whilst clone 216 remains largely unresponsive under e[CO2] conditions

e[CO2] has been shown to increase photosynthesis in plants, including trees (Ainsworth & Long, 2005; Hyvönen et al., 2007), largely through increases in Rubisco-limited photosynthesis in leaf tissues, the tissues that most directly sense and respond to e[CO2] (Wustman et al., 2001; Long et al., 2004; Ainsworth & Rogers, 2007). At the Aspen FACE site (http://aspenface.mtu.edu/), photosynthetic enhancement has ranged from 10 to 80% depending on the tree species and seasonal variation (Noormets et al., 2001b; Riikonen et al., 2008; Taylor et al., 2008). This enhancement translated into an increase in leaf and stem carbohydrate/starch content (Kopper & Lindroth, 2003; Kaakinen et al., 2004), as well as subsequent above-ground and below-ground growth (Isebrands et al., 2001; King et al., 2001, 2005; Karnosky et al., 2003, 2005; Norby et al., 2005; Kubiske et al., 2007).

Physiological data on each of the five different aspen clones grown at the Aspen FACE site have been collected since 1997 (see references in Table 1). One of the most outstanding differences between these clones is the growth enhancement responses resulting from e[CO2] between clone 216 (0–10% enhancement) and 271 (40–50% enhancement) (Fig. 1). Comparisons of published physiological data have shown that both clones have similar photosynthetic rates, a similar CO2-induced stimulation in photosynthesis, similar leaf area index (LAI) and a similar increase in LAI in response to e[CO2], suggesting that the differences in growth stimulation are a result of different strategies in carbon partitioning (Noormets et al., 2001a,b; Karnosky et al., 2005; Riikonen et al., 2008; Taylor et al., 2008). We confirmed that clones 216 and 271 had similar leaf photosynthetic rates at e[CO2]. Indeed, clone 216 showed a greater stimulation at e[CO2] (13.66 ± 2.06 µmol m−2 s−1 in controls vs. 18.76 ± 0.85 µmol m−2 s−1 in e[CO2]) than did clone 271 (16.33 ± 0.61 µmol m−2 s−1 in controls vs. 18.64 ± 0.01 µmol m−2 s−1 in e[CO2]) as a result of the lower rates at current [CO2] (Fig. 2).

Figure 1.

Annual stem volume growth (mean ± SE) comparison of aspen (Populus tremuloides) clones 216 and 271 in response to long-term elevated [CO2] (e[CO2]).

Figure 2.

Maximum light-saturated photosynthesis (Amax) of aspen (Populus tremuloides) clones 216 and 271 compared under ambient and long-term elevated [CO2] (e[CO2]). Data are expressed as the means ± SE of three leaves each from three trees.

Seasonal variation in the response of photosynthesis to e[CO2] has a large influence on the overwintering capacity of forest trees (Eamus & Ceulemans, 2001). Both clones 216 and 271 showed similar photosynthetic enhancement throughout the season (Noormets et al., 2001b; Riikonen et al., 2008); however, clone 271 showed significant delayed senescence in response to e[CO2], which extended the growing season by as much as 2 wk compared with 216 (G. Taylor, pers. comm.; Riikonen et al., 2008). Such extension of the growing season is likely to account for c. 2–3% of the gross primary productivity of clone 271 (Taylor et al., 2008). Other differences between the clones include minor variations in leaf surface waxes, cytoplasmic lipids, wood starch and nitrogen levels, cell wall thickness, as well as a slight improvement in the capacity of clone 271 to tolerate insects and pathogens (Table 1). Although such minor differences do not seem to explain the significant differences between the growth responses of these clones, they do suggest that each clone may use different molecular strategies for carbon accumulation and utilization (Isebrands et al., 2001; Wustman et al., 2001; Karnosky et al., 2003, 2005; Kubiske et al., 2007). As a means to elucidate such molecular differences that may underlie the difference in CO2 responsiveness between clones 216 and 271, we examined their leaf transcript profiles via microarrays.

Transcriptomics reveals that genotypic variation in transcriptional activity is one of the key factors controlling how clones 216 and 271 respond to e[CO2]

Microarray analysis of ∼26 800 cDNA probes (representing ∼10 000 different gene sequences) demonstrated that relatively few leaf transcripts (245) showed a significant response to long-term e[CO2] in either early (June) or late (August) time points. As visualized by PCA, 83% of the observed variation between leaf transcript expression in clones 216 and 271 can be explained by the interplay of developmental/seasonal differences between the time points and genotypic differences between the clones (Fig. 3). Such interactions between environment and genotype have also been observed in stress and defense responses of aspen genotypes (Osier & Lindroth, 2006).

Figure 3.

Principal component analysis (PCA) of normalized transcript intensities with low coefficient of variance (CV < 0.25), showing the principal components explaining the majority of the observed array variation. Principal component #1 correlates with developmental/seasonal differences (early vs. late time points). Principal component #2 correlates with genotypic differences [aspen (Populus tremuloides) clone 216 vs. clone 271].

Comparisons of array data between both clones and time points revealed 184 significantly different (P < 0.05) leaf transcripts that responded to long-term e[CO2] for at least one clone in one of the two time points. Early-season clone 216 showed 52 differentially expressed genes and only 14 genes in the late-season time point. By contrast, early-season clone 271 showed 102 differentially expressed genes and 25 genes in the late-season time point (Tables 2–5). In addition, very few of the leaf e[CO2]-responsive transcripts overlapped between the datasets for each clone or time point (Fig. 4a). These genotypic and developmental/seasonal differences can also be visualized by HCA, which depicts normalized transcript intensity ratios as coded colors (Fig. 4b). As in PCA, the developmental/seasonal effect was a stronger driver for global gene expression than genotypic difference, leading to the clustering of clone 216 and 271 data by time point. HCA also showed that both the number and magnitude of gene expression changes in response to e[CO2] were greater earlier in the growing season than late in the season.

Table 2. Differentially regulated genes in leaves of early-season clone 216 grown under elevated [CO2] (e[CO2])
  Biological functionDefinition of best BlastXPoplar gene modelFold changeNumber of occurrences
  1. Functional categories were based on the biological/molecular functions and standardized KOGG descriptions from the Joint Genome Initiative (JGI) gene ontology within the poplar database. Bold and italic type in the fold change column indicates two-fold or greater up-regulation, and bold type indicates two-fold or greater down-regulation.

Cell rescue defense virulence
  DefenseThaumatin-like protein [Arabidopsis thaliana]estExt_Genewise1_v1.C_LG_IX1261−1.371
  Response to woundingWound-induced protein [Arabidopsis thaliana]eugene3.001910632.251
Energy metabolism
  Electron transportThioredoxin [Glycine max]gw1.XVII.864.11.561
  PhotosynthesisELIP [Brassica rapa ssp. pekinensis]eugene3.1514100011.261
  Electron transportCytochrome c oxidase subunit VIa precursor [Oryza sativa (japonica cultivar-group)]estExt_fgenesh4_pg.C_LG_VII1235−1.451
  PhotosynthesisPSII DI protein [Actinidia arguta]fgenesh4_pg.C_scaffold_201000016−1.481
  Carbon fixationRibulose bisphosphate carboxylase [Gossypium hirsutum]eugene3.01230087−1.511
  PhotosynthesisPhotosystem I reaction center subunit II, chloroplast precursor (PS I subunit 5)eugene3.00100819−1.741
  PhotosynthesisPhotosystem II 32 kDa protein [Riccia huebeneriana]gw1.XIII.2252.1−1.791
General metabolism
  DiphosphoinositolPhosphatidylinositol/phosphatidylcholine transfer protein [Oryza sativa (japonica cultivar-group)]fgenesh4_pm.C_LG_V000599−1.431
  Chlorophyll biosynthesisGlutamate-1-semialdehyde aminotransferase [Arabidopsis thaliana]eugene3.00150799−1.471
  OxidoreductaseOxidoreductase family protein [Arabidopsis thaliana]estExt_Genewise1_v1.C_410478−1.511
  Chlorophyll biosynthesisGeranylgeranyl hydrogenase [Glycine max]eugene3.00120625−1.601
  Hydrolase activityATNUDT4 (Nudix hydrolase homolog 4) At1g18300 [Arabidopsis thaliana]grail3.0031000301−1.611
  Asparagine synthaseAsparagine synthase (glutamine-hydrolyzing)eugene3.01070081 −2.11 1
Protein fate folding modification
  ProteinaseAspartic proteinase [Theobroma cacao]estExt_fgenesh4_pg.C_1660009 2.32 1
  UbiquitinUbiquitin/ribosomal protein CEP52 – wood tobaccoestExt_Genewise1_v1.C_LG_XVI2419−1.391
  Transcriptional regulatorSyringolide-induced protein 1-3-1B [Glycine max]estExt_Genewise1_v1.C_LG_I0681−1.491
  Proteinase inhibitorOSJNBb0116K07.15 [Oryza sativa (japonica cultivar-group)]fgenesh4_pg.C_LG_I000354−1.551
  Detoxification proteinCopper chaperone [Populus alba × Populus tremula var. glandulosa]gw1.X.3496.1−1.591
  UbiquitinPolyubiquitin [Arabidopsis thaliana]eugene3.00060405−1.851
Secondary metabolism
  Cell wallPeroxidase (EC 1.11.1.7) – upland cottongrail3.0111002302 2.55 2
  Nitrogen utilizationPhenylcoumaran benzylic ether reductase [Populus balsamifera ssp. trichocarpa]estExt_fgenesh4_pm.C_LG_II01641.781
  BiosynthesisChalcone synthase (CHS1, clone CHS7)eugene3.00140920−1.231
  OxidoreductaseLaccase (diphenol oxidase), putative [Arabidopsis thaliana]fgenesh4_pg.C_scaffold_107000055−1.391
  OxidoreductaseLaccase [Populus trichocarpa]estExt_fgenesh4_pg.C_LG_VIII0541−1.771
  Protein targetingIntegral membrane protein [Phaseolus vulgaris]estExt_fgenesh4_pg.C_1780025−1.681
Translation
  Protein biosynthesisRibosomal Pr 117 [Triticum aestivum]estExt_Genewise1_v1.C_LG_II4011−1.241
  Protein biosynthesisRibosomal protein L2 [Schizanthus pinnatus]fgenesh4_pg.C_scaffold_2160000002−1.661
Transport
  Phosphate carrierPhosphate transporter, mitochondrialestExt_fgenesh4_pg.C_LG_I2387−1.401
  Ion transportCyclic nucleotide-gated channel family At5g14870eugene3.00410125−1.401
  AquaporinPlasma membrane intrinsic protein [Populus tremula × Populus tremuloides]estExt_Genewise1_v1.C_LG_III0271−1.751
  Xanthine/uracilXanthine/uracil permease family protein At2g26510 [Arabidopsis thaliana]gw1.40.581.1−1.761
  AquaporinAquaporin [Ricinus communis]eugene3.00280238−1.871
Unclassified
  Phosphate carrierPredicted P0524G08.101 gene product [Oryza sativa (japonica cultivar-group)]gw1.XIII.3248.1−1.271
  MitochondrionMitochondrial protein of unknown function, overexpression suppresses an rpo41 mutation affecting mitochondrial RNA polymerasegw1.XIII.1305.1−1.331
  UnknownHypothetical protein Npun02000359 [Nostoc punctiforme PCC 73102]fgenesh4_pg.C_scaffold_14076000001−1.35 2
  Cell wallProjectineugene3.00100655−1.361
  MitochondrionSterility protein 1 [Phaseolus vulgaris]eugene3.02970009−1.481
  Leucine-rich repeatHcrVf2 protein [Malus floribunda]eugene3.00101000−1.511
  Protein phosphorylationProtein kinase, putativefgenesh4_pg.C_LG_XI001073−1.571
  UnknownUnknowngrail3.0068005901−1.581
  UnknownUnknowneugene3.186940001−1.601
  Zinc-fingerUnknowneugene3.02270010−1.60 10
  Conserved protein Mo25AT5g47540/MNJ7_13 [Arabidopsis thaliana]estExt_fgenesh4_pg.C_LG_VI0133−1.611
  UnknownUnknowngw1.XVI.2246.1−1.621
  PSI assemblyycf3 protein [Panax ginseng]estExt_Genewise1_v1.C_14480001−1.681
  UnknownExpressed protein [Arabidopsis thaliana]estExt_fgenesh4_pg.C_LG_XIV0052−1.721
  UnknownUnknown protein [Oryza sativa (japonica cultivar-group)]estExt_fgenesh4_pg.C_LG_IX1163−1.771
  Protein phosphorylationUnknown protein [Arabidopsis thaliana]estExt_fgenesh4_pm.C_LG_XI0040−1.781
  Zinc-bindingZinc-binding protein [Oryza sativa (japonica cultivar-group)]grail3.0031020101−1.991
Table 3. Differentially regulated genes in leaves of early-season clone 271 grown under elevated [CO2] (e[CO2])
  Biological functionDefinition of best BlastXPoplar gene modelFold changeNumber of occurrences
  1. Functional categories were based on the biological/molecular functions and standardized KOGG descriptions from the Joint Genome Initiative (JGI) gene ontology within the poplar database. Bold and italic type in the fold change column indicates two-fold or greater up-regulation, and bold type indicates two-fold or greater down-regulation.

Cell fate
  SenescenceSenescence-associated protein [Pisum sativum]gw1.7267.9.1−1.711
Cell rescue defense virulence
  Defense, membrane permeabilityThaumatin/osmotin [Arabidopsis thaliana]gw1.I.8918.1 6.72 1
  Nutrient reservoirGermin-like protein [Arabidopsis thaliana]fgenesh4_pg.C_scaffold_426000001 5.53 1
  Proteinase inhibitorKunitz trypsin inhibitor 3 [Populus balsamifera ssp. trichocarpa × Populus deltoides]estExt_fgenesh4_pg.C_LG_XIX0984 3.00 5
  Proteinase inhibitorKunitz trypsin inhibitor TI3 [Populus tremula]estExt_Genewise1_v1.C_14340008 2.29 1
  DefenseThaumatin-like protein [Actinidia deliciosa]grail3.00200190021.851
  DefenseThaumatin-like protein [Arabidopsis thaliana]estExt_Genewise1_v1.C_LG_IX1261−1.671
  Glutathione metabolismGlutathione S-transferase GST 23 [Glycine max]estExt_fgenesh4_pg.C_LG_III1026−1.821
  Response to woundingProtease inhibitor 2 [Zinnia elegans]eugene3.10700003 −2.29 1
  TranscriptionDREB-like protein [Cucumis melo]gw1.XVIII.1182.1 −2.69 1
Cellular communication
  Protein transportGTP-binding protein [Oryza sativa (japonica cultivar-group)]estExt_fgenesh4_pm.C_LG_I0698−1.651
  Transcriptional regulationDentin sialophosphoprotein preproprotein [Homo sapiens]gw1.XV.2357.1−1.721
Development systemic
  Meristem developmentASYMMETRIC LEAVES2-like gene 1 protein [Arabidopsis thaliana]gw1.V.3873.11.821
  TranscriptionMADS-box protein PTM5 [Populus tremuloides]gw1.XIV.941.1−1.581
Energy metabolism
  Reductase1,4-Benzoquinone reductase [Oryza sativa (japonica cultivar-group)]estExt_Genewise1_v1.C_LG_XI23371.791
  GlycolysisEnolase, isoform 1 [Hevea brasiliensis]estExt_fgenesh4_pm.C_2801321.441
  PhotosynthesisPhotosystem I reaction center subunit II, chloroplast precursor (PS I subunit 5)eugene3.00081422−1.301
  ChloroplastOxygen-3volving enhancer protein 1estExt_fgenesh4_kg.C_LG_VII0034−1.391
  PhotosynthesisRibulose-1,5-bisphosphate carboxylase/oxygenase activase 1 [Gossypium hirsutum]gw1.VIII.2613.1−1.43 2
  Putative thioredoxinUnknown [Arabidopsis thaliana]estExt_Genewise1_v1.C_LG_XIII1459−1.441
  PhotosynthesisChlorophyll a/b-binding protein precursor [Euphorbia esula]grail3.0002067901−1.44 3
  PhotosynthesisUltraviolet-B-repressible protein [Gossypium hirsutum]estExt_Genewise1_v1.C_280006−1.461
  PhotosynthesisRibulose-1,5-bisphosphate carboxylase/oxygenase large subunit [Trithrinax acanthocoma]gw1.211.2.1−1.641
  Electron transportBlue copper protein [Pisum sativum]estExt_fgenesh4_pg.C_LG_II0928−1.991
  Phosphatidylcholine metabolismPhospholipase D [Ricinus communis]fgenesh4_pm.C_scaffold_440000162.021
General metabolism
  Amino acid transportAmino acid carrier [Ricinus communis]eugene3.06730001 12.29 1
  Nitrogen metabolismAminomethyltransferase precursor [Mesembryanthemum crystallinum]estExt_fgenesh4_pm.C_LG_XI0012−1.351
  Spermidine biosynthesis S-Adenosylmethionine decarboxylase [Prunus persica]gw1.123.91.1−1.511
  Fatty acid biosynthesisAcetyl-CoA carboxylase β-subunit [Ipomoea batatas]estExt_Genewise1_v1.C_1330169−1.531
  UDP-glucosyl transferaseUnknownfgenesh4_pg.C_LG_I000366−1.571
  Tyrosine biosynthesisUnknowngw1.97.33.1−1.681
  Pectin metabolismUnknowneugene3.00290334−1.701
  Hydrolase, putativeHypothetical protein T1K7.26 –Arabidopsis thalianagrail3.0010026002−1.711
  TransferaseGlycosyltransferase [Gossypium raimondii]estExt_Genewise1_v1.C_LG_V4069−1.731
 Electron transportP450 monooxygenase [Gossypium arboreum]eugene3.01270031−1.781
  Phosphatidylserine biosynthesisPhosphatidylserine sythase family proteineugene3.00081172−1.831
  Cell wallPectin methylesterase [Populus tremula × Populus tremuloides]grail3.0150001701−1.871
  Methionine biosynthesisCobalamine-independent methionine synthase [Solenostemon scutellarioides]grail3.0066005802−1.911
  Electron transportRespiratory burst oxidase homolog [Nicotiana benthamiana]estExt_Genewise1_v1.C_LG_III1154−1.991
  Response to pestsChitinase-like protein [Gossypium hirsutum]estExt_Genewise1_v1.C_LG_X05432.111
Protein fate folding modification
  Protein foldingCyclophilin [Populus alba × Populus tremula]estExt_fgenesh4_pg.C_LG_IX0365−1.511
  UbiquitinUbiquitin-conjugating enzyme-like protein [Oryza sativa (japonica cultivar-group)]gw1.107.100.1−1.551
  UbiquitinUbiquitin/ribosomal protein CEP52 – wood tobaccoestExt_Genewise1_v1.C_LG_XVI2419−1.581
  EndopeptidaseAt1g67700/F12A21_30 [Arabidopsis thaliana] contains domain OLIGOPEPTIDASEestExt_Genewise1_v1.C_LG_X4972−1.781
Regulation interaction with cellular environment
  Light-inducibleLight-inducible protein ATLS1 [Arabidopsis thaliana]grail3.0023033701−1.121
  Na+/H+ antiporter, putativeSodium hydrogen antiporter [Arabidopsis thaliana]eugene3.00012381−1.831
Secondary metabolism
  OxidoreductaseLaccase [Arabidopsis thaliana]eugene3.00060812−1.601
  Cell wallPeroxidase [Populus nigra]estExt_fgenesh4_pm.C_LG_I0036−1.671
  Lignin biosynthesisCinnamoyl-CoA reductase [Arabidopsis thaliana]gw1.II.2127.1−1.761
  Transferring hexosyl groupsUDP-glucuronosyltransferase [Arabidopsis thaliana]gw1.II.1756.1−1.871
Subcellular localization
  EndopeptidaseSignal peptide peptidase [Galega orientalis]estExt_fgenesh4_pm.C_LG_IX0436−1.701
  Cell adhesion, cell wallFasciclin-like AGP 4 [Populus alba × Populus tremula]eugene3.00131210−1.821
  Cell adhesion, cell wallFasciclin-like AGP 5 [Populus alba × Populus tremula]grail3.0094006801 −2.05 1
Transcription
    Ankyrin repeat-like protein [Oryza sativa (japonica cultivar-group)]fgenesh4_pg.C_LG_VIII0015831.741
  Protein biosynthesisRemorin family protein [Arabidopsis thaliana]estExt_Genewise1_v1.C_LG_XII1403−1.441
  TranscriptionApetala2/ethylene responsive factor [Populus alba × Populus tremula]gw1.158.150.1−1.461
  Putative polyadenylationFY protein [Oryza sativa (japonica cultivar-group)]gw1.III.2677.1−1.561
  Nucleic acid bindingSpliceosomal protein FBP21estExt_fgenesh4_pg.C_LG_IX0581−1.681
  TranscriptionEthylene-responsive factor-like protein 1 [Capsicum annuum]gw1.XIV.2445.1−1.721
  myb transcription factorTranscription factor [Arabidopsis thaliana] member of the R2R3 factor gene familyfgenesh4_pg.C_LG_VII000767−1.841
Translation
  Protein biosynthesisTranslation initiation factor B04 [Helianthus annuus]estExt_Genewise1_v1.C_LG_X0725−1.271
  Protein biosynthesisRibosomal protein L17 [Oryza sativa (japonica cultivar-group)]estExt_Genewise1_v1.C_LG_XV2068−1.351
  Protein biosynthesisAcidic ribosomal protein P1aeugene3.00021666−1.471
  tRNA synthetasetRNA synthetase class I (C) family protein [Arabidopsis thaliana]estExt_fgenesh4_pm.C_LG_X0474−1.501
  Protein biosynthesis40S ribosomal protein, putativeestExt_fgenesh4_pm.C_LG_VI0643−1.561
  Protein biosynthesis40S ribosomal protein S8estExt_Genewise1_v1.C_LG_IX3421−1.661
Transport
  Nitrate transporterNitrate transporter [Populus tremula × Populus tremuloides]fgenesh4_pm.C_LG_IX000714 2.11 1
  Vesicle coat proteinOSJNBb0017I01.5 [Oryza sativa (japonica cultivar-group)]estExt_Genewise1_v1.C_1240035−1.631
  Vesicle transportSynaptobrevin [Arabidopsis thaliana]gw1.I.7543.1−1.681
  Phosphate carrierPlastidic phosphate translocator-like protein2 [Mesembryanthemum crystallinum]estExt_fgenesh4_pm.C_LG_I0645−1.681
 Protein transportLongevity-assurance (LAG1) family protein At1g13580/F13B4_25 [Arabidopsis thaliana]gw1.VIII.2932.1−1.791
Unclassified
  GTP-aseUnknowneugene3.000908161.871
  UnknownUnknowneugene3.000117741.621
  Protein phosphorylationCholine kinase [Arabidopsis thaliana]grail3.00280242011.591
  MetabolismNo hit foundeugene3.015000331.371
   Arabidopsis protein of unknown function DUF266At5g11730 [Arabidopsis thaliana]estExt_fgenesh4_pm.C_LG_VI06981.341
  UnknownUnknowneugene3.00191070−1.271
  UnknownUnknowneugene3.01350001−1.311
  Glycolipid transferGlycolipid transfer protein-like [Oryza sativa (japonica cultivar-group)]eugene3.00100340−1.361
  UnknownUnknowneugene3.00120229−1.411
  UnknownTO71-3 [Taraxacum officinale]grail3.11474000101−1.491
  Phosphate carrierPREDICTED P0524G08.101 gene product [Oryza sativa (japonica cultivar-group)]gw1.XIII.3248.1−1.521
  UnknownAt3g60850 [Arabidopsis thaliana]eugene3.00140158−1.541
  UnknownHypothetical protein [Arabidopsis thaliana]estExt_Genewise1_v1.C_LG_XIV1980−1.561
  Membrane proteinIntegral membrane family protein-like [Oryza sativa (japonica cultivar-group)]estExt_fgenesh4_pm.C_LG_X0826−1.581
  GTP bindingNTGP4 [Nicotiana tabacum]gw1.41.462.1−1.591
  UnknownUnknown protein [Oryza sativa (japonica cultivar-group)]estExt_Genewise1_v1.C_95310004−1.601
  ProteinaseOTU-like cysteine protease-like [Oryza sativa (japonica cultivar-group)]gw1.XIV.2913.1−1.62 2
  UnknownUnknowneugene3.00102359−1.631
  UnknownNtEIG-E80 [Nicotiana tabacum]gw1.XVI.3414.1−1.641
  UnknownUnknown [Arabidopsis thaliana]eugene3.00040773−1.641
  UnknownOJ991214_12.11 [Oryza sativa (japonica cultivar-group)]grail3.0010069001−1.691
  Protein phosphorylationUnknowngrail3.0140005901−1.701
  Chlorophyll metabolismAminotransferase [Arabidopsis thaliana]fgenesh4_pm.C_LG_V000349−1.731
  UnknownNo hit foundgrail3.1043000101−1.731
  Calcium bindingUnknown [Arabidopsis thaliana]grail3.0242000102−1.731
  Protein dephosphorylationAT4g31860/F11C18_60 [Arabidopsis thaliana]eugene3.00061958−1.751
  Protein phosphorylationNo hit foundgw1.29.518.1−1.951
  Zinc-fingerNEW1 domain containing protein isoform [Oryza sativa (japonica cultivar-group)]gw1.VII.2281.1−1.991
  Protein phosphorylationSecreted glycoprotein 3eugene3.01340020 −2.17 1
  Carbohydrate metabolismNo hit foundeugene3.00090917 −2.55 1
  StructuralAt5g09820 [Arabidopsis thaliana]gw1.I.3658.1 −2.62 1
Table 4. Differentially regulated genes in leaves of late-season clone 216 grown under elevated [CO2] (e[CO2])
  Biological functionDefinition of best BlastXPoplar gene modelFold changeNumber of occurrences
  1. Functional categories were based on the biological/molecular functions and standardized KOGG descriptions from the Joint Genome Initiative (JGI) gene ontology within the poplar database. Bold and italic type in the fold change column indicates two-fold or greater up-regulation, and bold type indicates two-fold or greater down-regulation.

Cellular communication
  Receptor kinaseReceptor kinase-like protein [Oryza longistaminata]grail3.00670118011.201
Development systemic
  Meristem maintenanceNAC domain protein NAC4 [Glycine max]fgenesh4_pg.C_scaffold_107000083 6.91 1
Energy metabolism
  Ion transportATPase β-subunit [Nicotiana sylvestris]estExt_Genewise1_v1.C_LG_VIII1306−1.281
  PhotosynthesisChlorophyll a/b-binding protein precursor – upland cotton chloroplastgrail3.0002067901−1.381
  PhotosynthesisOxygen-evolving enhancer protein 1 precursor [Bruguiera gymnorrhiza]estExt_fgenesh4_kg.C_LG_VII0034−1.411
  Iron metabolismThioredoxin [Arabidopsis thaliana]estExt_Genewise1_v1.C_LG_XIII2097−1.431
General metabolism
  Cell wall biosynthesisUDP-sugar pyrophosphorylase [Pisum sativum]eugene3.00020715 2.39 1
  MetabolismCytochrome P450 monooxygenase – maize (fragment)estExt_Genewise1_v1.C_917800061.181
Protein fate folding modification
  UbiquitinUbiquitin-specific protease 26 [Arabidopsis thaliana]estExt_fgenesh4_pg.C_LG_II07321.701
Storage
  Cell wallAcid phosphatase class B family protein At1g04040/F21M11_2 [Arabidopsis thaliana]estExt_Genewise1_v1.C_LG_II40121.331
Systemic regulation of/interaction with environment
  Response to stressIron transport protein 2 [Ricinus communis]eugene3.00130613 2.46 1
Unclassified
  UnknownUnknown protein [Arabidopsis thaliana]gw1.IX.4902.1 2.70 1
  UnknownUnknowneugene3.00051241−1.281
  Protein phosphorylationMitochondrial protein of unknown function, overexpression suppresses an rpo41 mutation affecting mitochondrial RNA polymerasefgenesh4_pg.C_LG_XII001221−1.671
Table 5. Differentially regulated genes in leaves of late-season clone 271 grown under elevated [CO2] (e[CO2])
  Biological functionDefinition of best BlastXPoplar gene modelFold changeNumber of occurrences
  1. Functional categories were based on the biological/molecular functions and standardized KOGG descriptions from the Joint Genome Initiative (JGI) gene ontology within the poplar database. Bold and italic type in the fold change column indicates two-fold or greater up-regulation, and bold type indicates two-fold or greater down-regulation.

Cell cycling processing
  Chromosome assemblyRegulator of chromosome condensation (RCC1) family protein [Arabidopsis thaliana]eugene3.012300711.14 1
Cell rescue defense virulence
  Ion bindingMetallothionein 2b [Populus trichocarpa × Populus deltoides]eugene3.000913351.39 17
  Ion bindingMetallothionein 1a [Populus balsamifera ssp. trichocarpa × Populus deltoides]estExt_fgenesh4_pg.C_20800031.364
  Ion bindingMetallothionein 1b [Populus balsamifera ssp. trichocarpa × Populus deltoides]estExt_Genewise1_v1.C_14200361.33 14
  Ion bindingMetallothionein 1a [Populus balsamifera ssp. trichocarpa × Populus deltoides]eugene3.012000811.32 10
  Proteinase inhibitorKunitz trypsin inhibitor 3 [Populus balsamifera ssp. trichocarpa × Populus deltoides]estExt_fgenesh4_pg.C_LG_XIX0984 −2.58 2
Cellular organization
  Tubulin bindingTubulin folding cofactor D [Arabidopsis thaliana]gw1.XIV.724.11.45 1
  Actin depolymerizingActin-depolymerizing factor 2 [Petunia × hybrida]estExt_fgenesh4_pg.C_LG_I1779−1.56 1
Protein fate folding modification
  Unknown31.2-kDa small heat shock family protein/hsp20 family protein [Arabidopsis thaliana]eugene3.001408831.40 1
  Proteinase inhibitorCystatin-like protein [Citrus × paradisi]fgenesh4_pg.C_LG_VI000133−1.47 1
Secondary metabolism
  TransferaseAER [Nicotiana tabacum]eugene3.001610691.21 1
  Spermidine biosynthesisArginine decarboxylase 1 [Datura stramonium]estExt_Genewise1_v1.C_LG_IV0804−1.86 1
Subcellular localization
  Chromosome assemblyHistone H3.2 protein [Oryza sativa (japonica cultivar-group)]estExt_Genewise1_v1.C_LG_II09221.30 1
Transcription
  DNA bindingmyb family transcription factor At4g36570 [Arabidopsis thaliana]gw1.VII.3578.11.81 1
  Protein catabolismProtein involved in mRNA stabilitygw1.VIII.145.11.23 1
  Histone modificationTAZ zinc finger family protein/zinc finger (type) family protein [Arabidopsis thaliana]estExt_fgenesh4_pg.C_1350050−1.22 1
  TranscriptionTranscription factor of the Forkhead/HNF3 familyestExt_Genewise1_v1.C_LG_II4045−1.32 1
Translation
  Protein biosynthesisTranslation initiation factor B04 [Helianthus annuus]estExt_Genewise1_v1.C_LG_X07251.24 1
Transport
  AquaporinPlasma membrane intrinsic protein [Populus tremula × Populus tremuloides]grail3.00490303021.35 1
Unclassified
  UnknownUnknowngw1.XVI.2635.1 2.47  1
  UnknownProtein [Arabidopsis thaliana]eugene3.001200981.82 1
  UnknownHypothetical chloroplast RF2 [Eucalyptus globulus ssp. globulus]eugene3.1106700011.41 1
  UnknownUnknowngrail3.00060395011.26 1
  UnknownSimilar to unknown protein [Arabidopsis thaliana]fgenesh4_pg.C_LG_X0012371.16 1
  Membrane proteinUnknowngrail3.0012030401−1.61 1
Figure 4.

(a) Venn diagram comparing the differently expressed transcripts that were statistically significant (P < 0.05) for each aspen (Populus tremuloides) clone and time point. The analysis indicates that there is little overlap between the statistically significant sets of transcripts that are differentially expressed in response to long-term elevated [CO2] (e[CO2]). (b) Hierarchical cluster analysis (HCA) of normalized transcript intensity ratios having statistical significance (P < 0.05) in at least one condition for early and late time points for clones 216 and 271. Yellow indicates up-regulation in response to e[CO2] compared with control. Blue indicates down-regulation compared with control. Gray indicates little change compared with control. Stronger changes occur in response to e[CO2] earlier in the growing season than late in the season. Red branches indicate data that show statistically significant differential expression between the control and e[CO2] samples for early-season clone 271. Green branches indicate data that are not statistically significant for early-season clone 271.

Thus, similar to other studies on poplar tree transcriptomic responses to e[CO2] (Gupta et al., 2005; Taylor et al., 2005; Druart et al., 2006), we observed that field-grown poplar trees exposed to long-term e[CO2] displayed few significant leaf transcriptional changes compared with trees grown under ambient (control) conditions, despite often significant leaf and stem growth enhancements. The fold changes of the responsive genes in each of these studies were also typically small, making statistically significant detection less likely. Similar small numbers of e[CO2]-responsive genes and fold changes have been observed in a variety of other plant species (Li et al., 2007), and the most likely explanation is that these studies, like our own, are observing the acclimated response to a chronic treatment rather than the instantaneous response to an acute treatment (Ainsworth et al., 2007).

Clones 216 and 271 showed distinctly different e[CO2]-responsive sets of leaf transcripts that appeared to differentially impact on multiple biological pathways (Tables 2–5). Many of these gene expression trends showed significant differences in control and experimental expression levels between the clones, often including opposing patterns between the two clones. Such examples are shown in Tables S2–S5 (see Supporting Information), which provide expanded, processed datasets, including spot information, direct comparisons of pair-mean differences between clones and experiments, results of CV analysis and t-tests, and log data of both control and experimental samples. Such clonal and opposing trends were also assessed on a sample (∼10%) of the genes identified by the microarrays by performing quantitative RT-PCR analyses, using independent RNA samples as a means to more stringently test the patterns (Fig. 5). The combined results highlight the fact that innate transcriptional differences between different tree genotypes are an important and significant factor behind how trees respond to e[CO2]. This may help to explain some of the variation seen in studies making use of other cDNA microarrays (including the POP1 and POP2 microarrays used in Taylor et al., 2005 and Druart et al., 2006). If genotypes of the same plant species can vary dramatically in their response to e[CO2], the significant variation observed between different species is not surprising.

Figure 5.

Quantitative reverse transcriptase-polymerase chain reaction (RT-PCR) comparisons of selected genes between aspen (Populus tremuloides) clones 216 and 271 in response to long-term elevated [CO2] (e[CO2]). Included genes are as follows: RCA, ribulose-1,5-bisphosphate carboxylase/oxygenase activase; Lhcb1, chlorophyll a/b-binding protein; PSI, photosystem I reaction center subunit II; MT2b, metallothionein 2b; MT1, metallothionein 1; Thau, thaumatin; PCBER, phenylcoumaran benzylic ether reductase; POX, peroxidase; KTI, kunitz trypsin inhibitor 3; PR1, pathogenesis-related protein; ASP, aspartic proteinase; SAG, senescence-associated protein; UDPGP, UDP-sugar pyrophosphorylase; GT43, glycosyl transferase; DREB, DREB-like transcription factor; chitinase; cystatin; NTR, nitrate transporter.

It should be noted that there are some limitations to the PICME arrays used in this study in that they are not fully comprehensive and contain a moderate level of gene redundancy. It may be argued that, because these arrays contain ∼3600 leaf/shoot ESTs, ∼9000 xylem/cambium ESTs and ∼8600 root ESTs, they cannot be effectively used for leaf transcriptomics. However, leaves also contain xylem tissue, and many of the represented genes/ESTs are expressed in many different tissues. For example, of the leaf expression patterns assessed by quantitative RT-PCR (Fig. 5), only seven of these sequences were originally identified in leaf/shoot EST libraries. The remainder were identified as ESTs in other tissues (six from xylem/cambium and eight from root EST libraries), yet they are clearly expressed in leaves. There are many similar examples throughout our microarray data, as well as the data of leaf transcriptomics studies using the PICME arrays and/or similar POP1 and POP2 cDNA arrays (Broschéet al., 2005; Taylor et al., 2005; Druart et al., 2006; Rinaldi et al., 2007; Fluch et al., 2008). We also view the redundancy of the PICME arrays as an advantage, as it provides additional hybridization signals for each gene. In this respect, the transcriptional patterns of redundant clones on the PICME arrays corresponded well. Examples can be found in Tables 2–5, as well as the tables in the Supporting Information.

Clones 216 and 271 respond to e[CO2] through fundamentally different biological strategies that impact multiple pathways

As a means to highlight the primary (most significant) leaf gene expression trends observed between clones 216 and 271 as they sense and respond to e[CO2], we grouped the e[CO2]-responsive genes into functional categories based on the biological/molecular functions and standardized KOGG descriptions from the JGI gene ontology within the poplar database. Here, we discuss the key functional categories independently.

Cell rescue and defense  Clone 271 appears to have a different approach to defense compared with clone 216. Although early-season clone 216 shows a slight decrease in the expression of some leaf cell rescue and defense genes (Table 2), early-season clone 271 demonstrates the up-regulation of members of this gene class, including the up-regulation of thaumatin (Thau, 6.72-fold), germin-like protein (5.53-fold), pathogenesis-related protein (PR1), chitinase and repeated occurrences of kunitz trypsin inhibitors (KTI3, ~3.00-fold) (Table 3; Fig. 5). Indeed, at both early and late time points, clone 216 has either significantly reduced or opposing trends in the expression of such genes compared with clone 271. In the late-season samples, these trends reverse for clone 271 (showing down-regulation in response to e[CO2]), whereas the expression levels and responses remain approximately the same in clone 216 (Tables 4, 5; Fig. 5). In addition, late-season clone 271 differential gene expression is dominated by a slight up-regulation in metallothionein genes (MT1a and MT2b), which are expressed at much higher levels than by clone 216 (Table 5). Metallothionein genes are also involved in cell defense and stress responses through their ability to bind metal ions. Likewise, there was a significant up-regulation of a dehydration-responsive element-binding (DREB)-like transcription factor in late-season clone 271 compared with clone 216 (Fig. 5). DREB transcription factors have been shown to regulate the expression of other stress-related genes in the biotic stress signaling pathway (Agarwal et al., 2006). In addition, it has been possible to engineer stress tolerance in transgenic plants by manipulating the expression of DREBs (Agarwal et al., 2006), and so this may be one gene that helps clone 271 deal with stress late in the growing season.

Such observations indicate that the strategies for defense and leaf stress responses between clones 216 and 271 are different, and that clone 271 seems to take action early in the season to protect itself against attack and/or other stresses, whereas clone 216 directs this energy elsewhere. This correlates well with the fact that clone 271 shows an improved ability to deal with rust infections in the leaves and tent caterpillar attack compared with clone 216 (Kopper & Lindroth, 2003; Mankovska et al., 2005; Osier & Lindroth, 2006). Likewise, clone 271 appears to adapt to late-season stresses largely through altered transcriptional activity, whereas clone 216 does not.

Photosynthesis  It is interesting to note that, despite the well-characterized e[CO2]-induced increases in photosynthesis (Table 1; Fig. 2), the microarray data indicated a slight (< 1.75-fold) down-regulation of leaf genes related to photosynthesis and energy metabolism (Tables 2, 3). This phenomenon has also been observed in other projects using Populus species to characterize the effects of e[CO2], employing similar EST microarrays as used in our study (Taylor et al., 2005). Quantitative RT-PCR analysis indicated an e[CO2]-induced trend in photosynthesis, including Rubisco activase (RCA), chlorophyll a/b-binding protein (Lhcb1) and photosystem I reaction center subunit II (PSI). However, significant expression differences were only observed as a result of genotypic or seasonal variations, rather than those triggered by e[CO2] (Fig. 5). Such minor differences between microarray and quantitative RT-PCR results are probably caused by some level of cross-hybridization on the cDNA arrays, and are not uncommon when assessing genes having small fold-changes. The primers designed for quantitative RT-PCR are often more gene or group specific, and can thus discriminate desired expression from that of related genes. Overall, the expression trends (relative increases and decreases) for the quantitative RT-PCR results are consistent with the arrays, especially when comparing trends between clones 216 and 271.

Translational-, post-translational- and metabolism-level adjustments are also key to long-term physiological and growth responses in these trees. For example, in terms of photosynthesis, there was a substantial increase in the levels of RCA transcripts in clone 216 at the late time point compared with clone 271 (Fig. 5). RCA controls the overall activity of photosynthesis by post-translationally regulating the activity of the Rubisco enzyme through the removal of ribulose 1,5-bisphosphate (RuBP) from inactive Rubisco. Likewise, changes in the activity of Rubisco are related to alterations in source–sink relationships that may be controlled by limitations in the supply of nitrogen and/or water (Long et al., 2004; Ainsworth & Rogers, 2007; Luo et al., 2008). This correlates with changes in the expression levels of nitrogen utilization genes (see below) and water transport genes, including aquaporin/plasma membrane intrinsic protein (Tables 2, 5). Reduced leaf transcript abundances of aquaporins in response to e[CO2] were also observed for clone 216 by Gupta et al., (2005) and, combined with observed reductions in stomatal conductance (Table 1), may suggest that e[CO2]-treated trees are able to manage water more efficiently (Tjoelker et al., 1998).

Primary sugar metabolism  There was no effect of e[CO2] on the levels of glucose, fructose or sucrose, and only a marginally significant increase in starch (P = 0.094) for each clone (Fig. 6). However, there were significantly higher glucose (P = 0.004), fructose (P = 0.015), sucrose (P = 0.013) and starch (P = 0.022) contents in clone 271 compared with clone 216. The effect of e[CO2] on starch content was more pronounced in clone 271, as indicated by the significant CO2 × clone interaction, and this suggests that clone 271 may partition more carbon into transitory leaf starch at the beginning of the season to provide a greater carbon pool for nocturnal metabolism. The high sucrose levels in clone 271 could be indicative of a high flux of sucrose into the phloem, as well as a high capacity for carbon utilization. Lower levels of sucrose in clone 216 might indicate that more carbon is being funneled into leaf secondary metabolism rather than export.

Figure 6.

Levels of carbon and nitrogen metabolites in the leaves of aspen (Populus tremuloides) clones 216 and 271 grown at current [CO2] and long-term elevated [CO2] (e[CO2]). Leaves were taken from the same upper canopy collection as those used in the array and quantitative reverse transcriptase-polymerase chain reaction (RT-PCR) analyses. Significant (P < 0.05) effects of CO2, clone and date and the interactions (CO2 × clone, clone × date, CO2 × clone × date) are indicated in the panels. Marginally significant (P < 0.1) effects are indicated by parentheses surrounding the treatment.

Changes in gene expression related to leaf carbohydrate metabolism were also observed in the array data. Increases in a cell wall-associated glycosyl transferase (GT43) were seen in both clones. Glycosyltransferases are enzymes that act as catalysts for the transfer of monosaccharide units from an activated sugar phosphate to an acceptor molecule. Likewise, UDP-sugar pyrophosphorylase (UDPGP) is involved in general metabolism by controlling the levels of the internal hexose phosphate pool by providing metabolites for both cell wall synthesis and starch synthesis. The observed increase in expression (2.39-fold) of this sugar pyrophosphorylase in late-season clone 216 suggests that glucose and fructose may be directed towards cell wall thickening in the leaves or possibly the stems of clone 216 (Table 4). We consider this possibility next.

Secondary metabolism  This study reveals that secondary metabolism genes, such as phenylcoumaran benzylic ether reductase (PCBER) and cell wall peroxidase (POX), are more strongly up-regulated in the leaves of clone 216 at the early time point (1.78-fold and 2.55-fold, respectively) in response to e[CO2] (Tables 2, 3; Fig. 5). Opposite trends in PCBER between the clones also suggest opposing activity of the phenylpropanoid pathway between clones 216 and 271 in response to e[CO2]. In addition, clone 271 exhibits the down-regulation of genes involved in leaf secondary metabolism, including the lignin biosynthesis genes cinnamoyl-CoA reductase (−1.76-fold) and laccase (−1.60-fold). There is a corresponding down-regulation of leaf cell wall-related genes in clone 271, including fasciclin-like arabinogalactan protein genes (−2.0-fold) and pectin methylesterase (−1.87-fold). Combined with the up-regulation of UDPGP described above, such differences in expression patterns under e[CO2] suggest that clone 216 directs more of its fixed carbon to storage and/or defense in the form of cell wall compounds derived from the phenylpropanoid pathway.

Previous studies have reported that clone 216 clearly differs in leaf and wood structure compared with clone 271 (Oksanen et al., 2001, 2003; Kaakinen et al., 2004; Table 1). Clone 216 has significantly smaller fiber and vessel lumen diameters compared with clone 271, and clone 216 has the largest wall percentage, as well as the smallest vessel percentage, of any of the clones at the Aspen FACE site (Kaakinen et al., 2004). Interestingly, there was a significant (6.91-fold) up-regulation of a NAC [for NAM (no apical meristem), Arabidopsis NAC domain containing protein (ATAF1/2) and CUC2 (cup-shaped cotyledons 2)] domain transcription factor in the leaves of late-season clone 216 in response to e[CO2] (Table 4). NAC domain transcription factors have been implicated as transcriptional switches that regulate secondary wall synthesis in fibers (Zhong et al., 2006). Ectopic over-expression of such NAC transcription factors in the leaves of Arabidopsis results in the activation of the expression of secondary wall biosynthetic genes, leading to massive deposition of secondary walls in cells that are normally nonsclerenchymatous (Zhong et al., 2006). As secondary walls are important carbon sinks, the up-regulation of this NAC transcription factor may be a factor controlling carbon deposition into secondary cell wall thickening in the leaves of clone 216 compared with those of clone 271.

Protein stability and nitrogen metabolism  Another intriguing difference between clones 216 and 271 was the up-regulation of KTI3 in clone 271 and an opposing down-regulation of KTI3 in clone 216 in response to e[CO2] at the early time point (Table 3; Fig. 5). This trend reverses for clone 271 later in the season (Table 5; Fig. 5). Although KTI3 is typically associated with defense responses, it participates in such functions by impacting on the fate of specific proteins (Major & Constabel, 2008). In conjunction with alterations in KTI levels, we also observed alterations in the expression levels of various proteases, as well as genes involved in the ubiquitin pathways, in both clones. In addition, there was a substantial increase in the activity of an amino acid carrier gene (12.29-fold), as well as an increase in nitrate transporter (2.11-fold) in the leaves of early-season clone 271 (Table 3). This nitrate transporter showed an even stronger response (3.69-fold increase) in late-season clone 271 and only marginal changes in clone 216 (Table S3). Together, these results suggest that e[CO2] treatment has different impacts on protein stability and nitrogen metabolism, especially for clone 271. To test this possibility, we again turned to biochemical assays to assess the level of nitrogen resources.

The free leaf amino acid pool is a good indicator of nitrogen status as it represents the currently available pool of reduced nitrogen for growth and is representative of whole-plant nitrogen status. The response of free amino acids to growth at e[CO2] was unclear. There was a significant interaction between clone and date (P = 0.005), driven by the higher amino acid levels in clone 216 in August (Fig. 6). There was also a significant interaction between CO2 and clone (P = 0.047) and a marginally significant interaction between CO2, clone and date (P = 0.072). Clone 216 showed no response in June, but an increase in August, whereas clone 271 showed a consistently lower amino acid content at e[CO2] (Fig. 6). There was no effect of CO2 or clone on protein content, but there was a significant clone × date interaction (P = 0.009), driven by the higher protein content in clone 216 in August. This trend was also seen in nitrate content, where there was a significantly higher content in clone 216 compared with clone 271 (P = 0.018), which was markedly higher in August (clone × date, P = 0.049). As the sucrose content is indicative of the available pool of carbon for growth, the sucrose to free amino acid ratio can provide an indication of potential shifts in carbon to nitrogen status. There was a significant effect of clone (P = 0.039) and a significant interaction between clone and date (P = 0.023), indicating a shift in the availability of reduced carbon and nitrogen metabolites. This shift was confined to clone 271 in August, and was driven by an increased availability of fixed carbon, which was not matched by a similar increase in reduced nitrogen. This contrasts with the response of clone 216, which was able to match increases in carbon availability occurring later in the season with increases in free amino acid content.

Regulation of transcription  One last gene functional category that showed significant responses to e[CO2], whilst also differing between the clones, was transcription factors. We have already pointed out the importance of DREB and NAC domain transcription factors in the regulation of stress-related genes in clone 271 and possible leaf cell wall thickness in clone 216. However, transcription factors, including those involved in systemic development, appear to play an important role in clone 271 in response to e[CO2], whereas no changes in such factors are observed in clone 216 at either early or late season (Tables 2–5). Some of these clone 271-specific e[CO2]-responsive transcription factors have known functions in stress responses and biomass development, including ethylene response factors (McGrath et al., 2005), Myb family transcription factors involved in the development of cell size in poplar leaves and stems (Karpinska et al., 2004; McGrath et al., 2005) and one MADS-box gene known to participate in cell expansion in the vascular tissues of P. tremuloides (Cseke et al., 2007). Such alterations in regulatory factors correlate well with the known growth responses in clone 271, and their lack of response in clone 216 may play a role in the e[CO2] unresponsiveness of this clone.

Conclusions

Trees allocate assimilated carbon between growth, respiration, storage and chemical defense, and changes in resource availability, such as light, nitrogen or atmospheric [CO2], can affect tree species quite differently. Growth-dominated species tend to allocate ‘extra’ carbon (relative to nitrogen) to growth, whereas differentiation-dominated species tend to allocate it to the production of carbon-based secondary compounds, including phenylpropanoid biosynthesis, and subsequent morphological differentiation, such as cell wall thickening (King et al., 2005). This study has demonstrated that poplar clones 216 and 271 show distinctly different e[CO2]-responsive genes during long-term exposure to e[CO2]. In most cases, expression differences caused by developmental/seasonal and/or genotypic variations were greater than those triggered by e[CO2] treatment, influencing multiple pathways that work synergistically to control the overall response to e[CO2]. Much of this variation may be a result of the activity (or inactivity) of specific transcriptional factors that may be orchestrating overall development. Clone 271 clearly has a responsive, growth-dominated strategy that seems to be based on a ‘less expensive’ defense strategy, allowing it to be better adapted to respond to stress. Clone 216 appears to devote more energy towards carbon-based secondary compounds and possibly relocation of carbon to processes other than growth. Thus, we label clone 216 as having a preparative, differentiation-dominated strategy. Such strategies have been described previously to help explain the variable e[CO2]-induced growth enhancements of different tree species. However, to our knowledge, this study is the first to provide evidence that such differences in fundamental biological strategies also occur within genotypes of the same tree species. Consequently, selection for improved productivity or carbon sequestration needs to take into account the genetically distinct responses, not only of different tree species, but also of different genotypes within a species. Selection of e[CO2]-responsive clones, such as clone 271, may provide the best means to improve carbon sequestration potential in future plantations.

Acknowledgements

This work was sponsored primarily by the US Department of Energy's Office of Biological and Environmental Research, Grant No. DEFG02-04ER63792. Operations of Aspen FACE are principally supported by the US Department of Energy's Office of Biological and Environmental Research, Grant No. DE-FG02-95ER62125. A.R. was supported in part by the US Department of Energy Office of Science Contract No. DE-AC02-98CH10886 to Brookhaven National Laboratory. We would like to thank Dr Ramesh Thakur for his help in collecting Aspen FACE Site tissues used in this study. We would also like to thank Scott Harding and Michelle Jarvie for their assistance with RNA extractions for quantitative RT-PCR analysis. We thank Dr Francis Martin and Dr Silvia Fluch for supplying the PICME annotation files used in this study. The ESTs printed on the PICME poplar arrays were produced by INRA-Nancy (Rinaldi et al., 2007), INRA-Orleans (Déjardin et al., 2004) and the University of Helsinki (Broschéet al., 2005) within the framework of the LIGNOME and ESTABLISH programmes, respectively.

Ancillary