The plant energy-dissipating mitochondrial systems: depicting the genomic structure and the expression profiles of the gene families of uncoupling protein and alternative oxidase in monocots and dicots

Abstract

The simultaneous existence of alternative oxidases and uncoupling proteins in plants has raised the question as to why plants need two energy-dissipating systems with apparently similar physiological functions. A probably complete plant uncoupling protein gene family is described and the expression profiles of this family compared with the multigene family of alternative oxidases in Arabidopsis thaliana and sugarcane (Saccharum sp.) employed as dicot and monocot models, respectively. In total, six uncoupling protein genes, AtPUMP1–6, were recognized within the Arabidopsis genome and five (SsPUMP1–5) in a sugarcane EST database. The recombinant AtPUMP5 protein displayed similar biochemical properties as AtPUMP1. Sugarcane possessed four Arabidopsis AOx1-type orthologues (SsAOx1a–1d); no sugarcane orthologue corresponding to Arabidopsis AOx2-type genes was identified. Phylogenetic and expression analyses suggested that AtAOx1d does not belong to the AOx1-type family but forms a new (AOx3-type) family. Tissue-enriched expression profiling revealed that uncoupling protein genes were expressed more ubiquitously than the alternative oxidase genes. Distinct expression patterns among gene family members were observed between monocots and dicots and during chilling stress. These findings suggest that the members of each energy-dissipating system are subject to different cell or tissue/organ transcriptional regulation. As a result, plants may respond more flexibly to adverse biotic and abiotic conditions, in which oxidative stress is involved.

Alternative oxidase, expression profile, oxidative stress, uncoupling proteins

This paper is available online free of all access charges (see http://jxb.oxfordjournals.org/open_access.html for further details)

Introduction

In recent years, several groups have asked why plants possess two energy-dissipating systems, an uncoupling protein (UCP, also termed PUMP in plants) and an alternative oxidase (AOx), in their mitochondria. The proton electrochemical gradient energy-dissipating pathway involving PUMP and the redox energy-dissipating pathway involving AOx lead to the same final effect, i.e. a decrease in the efficiency of oxidative phosphorylation and an increase in heat production. Both systems play a role in protecting plant cells against oxidative stress. PUMPs are probably involved in optimizing the mitochondrial proton electrochemical potential (Δμ_H⁺) across the inner mitochondrial membrane, while AOx appears to function predominantly in thermogenic processes (Meeuse, 1975; Day et al., 1980; Borecký et al., 2001a). Both systems are present simultaneously in the mitochondria of green tomato fruit (Sluse and Jarmuszkiewicz, 2000), thus raising the question of a specific physiological role for each of them and of a possible functional link between them (shared regulation). The systems do not appear to work at their maximal rates simultaneously, since increasing concentrations of free fatty acids (FFA) block the activity of AOx (K_i ∼4 μM) while activating PUMP (K_m ∼10 μM; Sluse and Jarmuszkiewicz, 2000). Thus, FFAs could function as switching agents between the energy-dissipating systems. Studies of AOx and PUMP expression and of their activities during the ripening of tomato (Sluse and Jarmuszkiewicz, 2000; Almeida et al., 2002) or mango fruit (Considine et al., 2001) have shown that AOx and PUMP may work sequentially.

To date, the properties and physiological roles of two isoforms of PUMP and four isoforms of two types of AOx have been studied. Vercesi et al. (1995) discovered a plant counterpart of mammalian uncoupling protein (primarily UCP, now UCP1; Ricquier and Kader, 1976) in potato and named it plant uncoupling mitochondrial protein (PUMP). Its biochemical and physiological properties are similar to those of UCP1 (Jarmuszkiewicz et al., 1998; Ježek et al., 1996, 1997; Ježek and Garlid, 1998; Kowaltowski et al., 1998; Sluse et al., 1998; Vercesi et al., 1997, 1998; Almeida et al., 1999; Costa et al., 1999; Nantes et al., 1999). From 1997 to 1999, four additional animal UCPs and a second PUMP were identified (UCP2, Fleury et al., 1997; UCP3, Boss et al., 1997; UCP4, Mao et al., 1999; BMCP1/UCP5, Sanchis et al., 1998; plant AtUCP2, Watanabe et al., 1999). Recently, a comprehensive review on these members of UCP was reported by Ledesma et al. (2002).

The finding of uncoupling proteins in plants raised the question of their physiological role, originally considered to be solely heat production in non-shivering thermogenesis in hibernating mammals. Recent hypotheses favour a more general role, the regulation of energy metabolism in mitochondria (Ježek and Garlid, 1998; Skulachev, 1998; Ricquier and Bouillaud, 2000) in order to avoid an extremely high Δμ_H⁺, which can lead to the excessive production of reactive oxygen species (ROS; Skulachev, 1998; Brandalise et al., 2003a; Considine et al., 2003).

In chilling-sensitive plants, such as sugarcane (Saccharum sp.; Tai and Lentini, 1998), oxidative stress is a major component in the chilling response (Pinhero et al., 1997). ROS, i.e. hydrogen peroxide, superoxide, and hydroxyl radicals, can react with DNA, lipids, and proteins, to cause severe cellular damage (Sato et al., 2001). Since mitochondria represent one of the major sources of ROS during cold stress, PUMP and AOx may serve to prevent ROS production in this organelle (Kowaltowski et al., 1998; Maxwell et al., 1999).

In plants, two discrete gene families encoding alternative oxidases were described (for a review, see Considine et al., 2002). The genes of the AOx1 type family seem to be induced by different types of stress and are present in both monocot and dicot plants. The genes of the AOx2 type family are usually expressed in a constitutive or developmental manner in dicots, but are absent in all monocots examined to date (Considine et al., 2002). This molecular distinction suggests a divergence in AOx across plant families and may even have implications for its physiological role in different plant species.

New attempts to determine the physiological role of PUMP and AOx isoforms in plants have focused on their gene regulation (Ito et al., 2003; Clifton et al., 2005; Dojcinovic et al., 2005). In Arabidopsis, five genes encoding AOx (Considine et al., 2002) and two genes encoding PUMP (Maia et al., 1998; Watanabe et al., 1999) have been identified. A sequence encoding another putative PUMP was also found in the Arabidopsis genome (Hanák and Ježek, 2001). In this report, a probably complete gene family of uncoupling proteins in plants, which consists of up to six members, PUMP1–6, identified in the genome of the dicotyledonous Arabidopsis thaliana or in a database of 237 954 ESTs of monocotyledonous sugarcane (Vettore et al., 2003) is described and characterized. One of the new Arabidopsis PUMP genes, AtPUMP5, was expressed in E. coli and the recombinant protein was reconstituted into proteoliposomes. Similarly to AtPUMP1, this recombinant AtPUMP5 induced linoleic acid-mediated H⁺ flux that was sensitive to ATP and GTP. Furthermore, four members of the AOx1 gene family in sugarcane were also identified. One of the Arabidopsis AOx genes (AtAOx1d) was found to belong not to the AOx1 gene family but to a third AOx gene family, called the AOx3 family. The divergences in expression profiles of PUMPs and AOxs in specific tissues/organs and during chilling-induced stress suggest that each PUMP and AOx isoform can have a specific physiological role. The regulation of gene expression of PUMPs and AOxs is discussed.

Materials and methods

Plant growth and chilling stress treatment

Sugarcane plantlets (Saccharum sp. cv. SP80-3280), propagated axenically in vitro (Nogueira et al., 2003), were kept at 26 °C on a 16/8 h day/night cycle with a photon flux density of 70 μE m⁻² s⁻¹. Four-week-old plantlets were transferred to 4 °C under the same photoperiod conditions. Control plantlets were maintained at 26 °C. The leaves of control and chilling-treated plantlets were harvested after 0, 6, 12, 24, and 48 h of treatment (six plantlets per time point). The expression pattern of the SsPUMP and SsAOx genes in leaves, roots, and stem, was examined using 1-month-old sugarcane plantlets (Saccharum sp. cv. SP80-3280) cultivated in a greenhouse. Flowers were obtained from sugarcane plants (Saccharum sp. cultivar SP80-87432) grown at the Copersucar experimental station (http://www.ctc.com.br).

Seeds of Arabidopsis thaliana ecotype Columbia (Col-0) were surface-sterilized and transferred to Petri dishes containing MS medium (Murashige and Skoog, 1962). The seedlings were transferred to fresh MS medium every week and maintained in a growth chamber at 22 °C on a 16/8 h day/night cycle with a photon flux density of 70 μE m⁻² s⁻¹ (Maia et al., 1998). Twenty-one-day-old seedlings were incubated either at 4 °C or at room temperature (22 °C) and the above-ground organs were harvested after 0, 3, 6, 12, 24, and 48 h of treatment (60 seedlings per time point). To analyse the expression profiles of AtPUMPs and AtAOxs in different plant organs, seeds were germinated under controlled environment conditions: 70 μE m⁻² s⁻¹, day/night temperature of 22 °C, and a photoperiod of 16/8 h day/night. Leaf, flower, and root samples were obtained from 10-week-old plants.

Identification and in silico analysis of sequences encoding sugarcane and Arabidopsis PUMPs and AOxs

The Arabidopsis thaliana genome database (TIGR, http://www.tigr.org/tdb/ath1/htmls/index.html) was analysed to find sequences homologous to UCP/PUMP or AOx using the tBLASTn algorithm (Altschul et al., 1997) and UCP1–5, AtPUMP1–2, AtAOx1a–1c, and AtAOx2 as drivers. The same approach, including the novel Arabidopsis homologues, was used to identify SAS (sugarcane assembled sequences; http://sucest.lad.ic.unicamp.br/public; Vettore et al., 2003) sharing similarities to PUMPs and AOxs. Sugarcane PUMP and AOx sequences were also identified by keyword searches using a provisional functional assignment generated by automated annotation, and the resulting hits were validated against existing homologues in public databases (NCBI, http://www.ncbi.nlm.nih.gov).

To obtain the full-length cDNA sequences that represented both sugarcane PUMP and AOx genes, the longest EST of each identified SAS was resequenced using an ABI PRISM 3700 DNA sequencer and sequence data were analysed using the PHRED-PHRAP package (http://www.phrap.org/).

The full-length cDNAs representing the AtPUMP and AtAOx genes were obtained by RT-PCR, using clone-specific forward and reverse (in parentheses) primers for each gene: AtPUMP1, 5′-GCTTTAGCCGTAATCGTCG-3′ (5′-GTTGCTCTCATTCCCTCTGC-3′), AtPUMP2, 5′-CAATGGCGGATTTCAAAC-3′ (5′-CTAGGGATCTGAGAATCAATCG-3′), AtPUMP3, 5′-ATGGAGCGGAGCCGAGTG-3′ (5′-CGCGTTAATCAGAAACTGATGC-3′), AtPUMP4, 5′-ATGGGAGTCAAAAGTTTCG-3′ (5′-TCAAAAATCTCGAAGCAGC-3′), AtPUMP5, 5′-CCAGTGAGATCAGCTCCAATTC-3′ (5′-CGCCACCATCATTCATCTTC-3′), AtPUMP6, 5′-TGTCTTCTTCCATTCCAAC-3′ (5′-TGAAGAAATATGGGAATCTC-3′), AtAOx1a, 5′-ATGATGATAACTCGCGGTGGAGC-3′ (5′-GCAACATTCAAAGAAAGCCGAATC-3′), AtAOx1b, 5′-ATGATGATGAGTCGTCGCTATG-3′ (5′-CCCATTAAAGCCCATTTAGG-3′), AtAOx1c, 5′-CAAATCTCCCTTGAATCCG-3′ (5′-GCTCTTCTGATTCAGTGATATCC-3′), AtAOx1d, 5′-CCCAACTGTTGTTACTCATG-3′ (5′-CACAGCTTTGTGACTTTGTC-3′), and AtAOx2, 5′-ATGGGTATGAGTTCTGCATCG-3′ (5′-TTAGTGATAACCAATCGGAGCTG-3′). Reverse transcription was done on 5 μg of total RNA extracted from Arabidopsis seedlings using Superscript II RNase H⁻ reverse transcriptase (Invitrogen, USA) under the recommended conditions. PCR was then done using the following conditions: 94 °C for 3 min and 35 cycles of 94 °C for 1 min, 52–60 °C for 45 s, and 72 °C for 1 min 30 s. The single PCR products were isolated from agarose gels using Concert™ rapid gel extraction system (Invitrogen, USA), then cloned into the pGEM-T EASY vector (Promega, USA) and resequenced.

The amino acid sequences deduced from the cDNAs of the PUMP candidates mined in the databases were initially aligned with the corresponding protein sequences of all known UCPs/PUMPs using the CLUSTALX program (weight matrix Gonnet 250; Thompson et al., 1997). The aligned sequences were grouped and subsequently aligned with 80 sequences representative of known UCPs/PUMPs, together with malate/2-oxoglutarate carriers (M2OMs) and the dicarboxylate carriers (DICs), as the phylogenetically closest members of the mitochondrial anion carrier family (MACF). To validate further the identified gene sequences, phylogenetic trees (1000 bootstraps) were inferred using the MEGA2 program (Kumar et al., 2001). Sequences from the following list (accession numbers in parentheses) were used for the alignment and phylogenetic analyses: ODC1_YEAST (Q03028), ODC2_YEAST (Q99297), UCP2_HUMAN (P55851), UCP2_CANFA (Q9N2J1), UCP2_RAT (P56500), PsUCP2 (AAG33984), UCP2_MOUSE (P70406), UCP2_PIG (O97562), XlUCPput (AAH44682), UCP2_BRARE (Q9W720), UCP2_CYPCA (Q9W725), UCP3_MOUSE (P56501), UCP3_RAT (P56499), PsUCP3 (AAG33985), UCP3_CANFA (Q9N2I9), UCP3_PIG (O97649), UCP3_HUMAN (P55916), UCP3_BOVIN (O77792), UCP-gallus (AAL35325.2), MgUCP (AAL28138), EmUCP (AAK16829), PmUCP2 (AAL92117), UCP1_MOUSE (P12242), UCP1_RAT (P04633), UCP1_MESAU (P04575), PsUCP1 (AAG33983), DgUCP1 (AAM49148), UCP1_RABIT (P14271), UCP1_BOVIN (P10861), UCP1_HUMAN (P25874), GmPUMP1b (AAL68563), GmPUMP1a (AAL68562), MiPUMP-1 (AAK70939), LePUMP (AAL82482), AtPUMP1 (CAA11757), SfUCPa (BAA92172), SfUCPb (BAA92173), HmUCP (BAC06495), ZmPUMP (AAL87666), OsPUMP2 (BAB40658), OsPUMP1 (BAB40657), TaPUMP1a (BAB16385), TaPUMP1b (BAB16384), AtPUMP2 (NP_568894), DIC_MOUSE (Q9QZD8), RnDCput (NP_596909), DIC_HUMAN (Q9UBX3), CeM2OM (NP_509133), NcDCput (XP_327953), M2OM_MOUSE (Q9CR62), M2OM_RAT (P97700), M2OM-pig (AAD01440), OaM2OM (AAF44754), M2OM_HUMAN (Q02978), M2OM_BOVIN (P22292), NtDTC (CAC84545), StM2OM (CAA68164), AtM2OM (NP_197477), PmM2OM2 (S65042), SsM2OM (AY644469), PlaFalM2OM (CAD51134), PlaYoeM2OM (EAA21506), AtPUMP5 (F14M13_10), AtPUMP4 (F22K18_230), AtPUMP6 (T5E8_270), UCP5_MOUSE (Q9Z2B2), MmUCP5 (NP_035528), RnUCP5 (NP_445953), UCP5_HUMAN (O95258), UCP4_HUMAN (O95847), MmUCP4 (BAC66453), RnUCP4a (CAC20898), RnUCP4b (CAC20899), RnUCP4c (CAC20900), CeUCP (NP_505414), AtPUMP3 (F7A19_22), UCPhom-yeast (S25357), and DNC_HUMAN (Q9HC21).

The complete amino acid sequences sharing homology with known UCPs, PUMPs, M2OMs, and DICs were analysed using the pattern prediction program package MEME-MAST, version 3.0 (Bailey and Elkan, 1994; Bailey and Gribskov, 1998; http://www.sdsc.edu/MEME). Finally, the identified sequences of sugarcane and Arabidopsis were checked for (i) the number of energy transfer protein signature (ETPS, P-x-[DE]-x-[LIVAT]-[RK]-x-[LRH]-[LIVMFY], PROSITE accession number PDOC00189) and (ii) the presence of these ETPS in variants specific for uncoupling proteins (Borecký et al., 2001a).

Equivalent in silico analysis was done for AOx candidates. Representative fungal and plant mitochondrial AOx sequences plus two putative plastidic AOxs were retrieved from the NCBI database. Sequences from the following list (accession numbers in parentheses) were used for alignment and phylogenetic analyses: VinAOx0a (AAK61349), VinAOx0b (AF279690), AniAOx (AAN39883), EniAOx (BAA93615), AOX_ASPNG (O74180), AfuAOx (AAL87459), AcaAOx (AAD29681), BfAOx (CAD42731), MfAOx (AAL24516), BgAOx (AAL56983), AOX_NEUCR (Q01355), GspAOx (AAN39884), PaAOx (AAK58849), MgAOx (AAG49588), CaAOx0a (AAF21993), CaAOx0b (AAC98914), AOX_HANAN (Q00912), YlAOx (CAD21442), CneAOx (AAM22475), CrAOx0a (AAG33633), CrAOx0b (T07947), Cw80AOx (BAA23725), PapAOx (CAE11918), AOX_TRYBB (Q26710), MiAOx1b (AAK70936), PtPtAOx1a (CAB64356), AOX1_SOYBN (Q07185), AtAOx1a (Q39219), AtAOx1d (NP_564395), MiAOx1a (AAK70935), AtAOx1b (O23913), AtAOx1c (O22048), AOX1_SAUGU (P22185), StAOx1 (2208475A), MiAOx1c (AAK70937), AOX2_TOBAC (Q40578), AOX1_TOBAC (Q41224), LeAOx1a (AAK58482), CroAOx (BAA23803), LeAOx1c (AAP92756), LeAOx1b (AAK58483), PtPtAOx1b (CAB72441), ZmAOX2 (AAL27796), OsAOx1a (BAA28772), TaAOx1a (BAB88646), TaAOx1b (BAB88645), ZmAOX1 (AAL27795), TaAOx1c (BAB88646), OsAOx1c (BAB71944), ZmAOX3 (AAL27797), OsAOx1b (BAA28771), AOX2_ARATH (O22049), LeAOx2 (AAP92755), AOX2_SOYBN (Q41266), VuAOx2a (CAC42836), CsAOx2a (AAP35170), AOX1_MANIN (Q40294), AOX3_SOYBN (O03376), VuAOx2b (CAD12835), CsAOx2b (EAA32850), OsAOxPut (AAL76179), DdAOx (BAB82989), AtAOxput (NP_567658), and LeAOxPlastid (AAG02286).

Digital mRNA expression profiling

The digital expression profiling of the SsPUMP and SsAOx transcripts among sugarcane tissues was analysed as an estimate of their relative abundance calculated as a count of the number of ESTs in the given library pool normalized by the total number of ESTs in each SUCEST library pool (pooled from 25 different EST libraries prepared from mRNA isolated from different tissues (Vettore et al., 2003). The library pools are listed as the following (total number of ESTs from each pool in parentheses): IL, plants inoculated with endophytic bacteria (24 430); M, meristem (39 116); F, floral organs (52 430); L, leaves (19 676); R, roots (25 302); S, seeds (17 106); and ST, stem (49 785).

The Arabidopsis ESTs were retrieved from libraries described in the Arabidopsis Gene Index database at The Institute for Genomic Research (TIGR; http://www.tigr.org/tdb/tgi/agi). Libraries containing more than 9000 ESTs and completely described for tissues or organs were used to further digital mRNA expression profiling analysis. All subtractive libraries were removed from the analysis and selected non-normalized EST libraries were pooled into five library pools. The library pools are listed (total number of ESTs from each pool in parentheses): A, above-ground organs (12 985); R, roots (17 573); F, floral organs (9120); G, green siliques (12 589); and S, seeds (10 800). The sequences representing AtPUMP and AtAOx transcripts were used as drivers against TIGR Arabidopsis gene index and corresponding tentative consensus (TC) sequences (227 742 TCs as of 24 April, 2003; Quackenbush et al., 2000) were analysed in order to estimate the relative abundance of each gene in the selected library pools.

Nuclear-encoded subunits of ATP-synthase (subunit γ, accession D88374 or CA121698) and two respiratory chain representatives, Complex I (51 kDa subunit, accession NM_120938 or CA093348) and Complex IV (subunit Vb, accession NM_112434 or CA129721) were included in the analysis as mitochondrial nuclear gene controls.

RNA gel-blot analysis

Total RNA was isolated from chilling-treated and untreated sugarcane and Arabidopsis tissues using Trizol reagent (GibcoBRL, USA) according to the manufacturer's instructions. Twenty micrograms of total RNA were electrophoresed in a 1% (w/v) agarose gel containing formaldehyde and transferred to a Hybond-N⁺ filter (Amersham Pharmacia Biotech, USA), as described by Sambrook et al. (1989). The filters were hybridized with full-length cDNA fragments of SsPUMPs, SsAOxs, AtPUMPs, and AtAOxs labelled with α-³²P dCTP. Hybridization was done at 42 °C for 16 h (Sambrook et al., 1989). The blots were then washed at high stringency and exposed to imaging plates. Digitalized images of the RNA-blot hybridization signals were analysed using the Image Gauge software (Fujifilm, Japan).

Semi-quantitative RT-PCR

Semi-quantitative RT-PCR was used to detect the accumulation of AtPUMP4, AtPUMP5, AtAOx1b, and AtAOx1d transcripts. cDNA was synthesized from 500 ng (AtAOx1b–1d and AtPUMP5) and 1 μg (AtPUMP4) of total RNA from leaves, flowers, and roots using Superscript II RNase H⁻ RT (Invitrogen, USA). Primers for the constitutive gene encoding adenine phosphoribosyltransferase (APRT, Moffatt et al., 1994) were used as an internal control to normalize the quantity of total RNA used in each sample. The PCR was done with a 12.5-fold dilution of the first-strand cDNA as a template under the following conditions: 94 °C for 3 min and 27 cycles of 94 °C for 1 min, 52–60 °C for 45 s, and 72 °C for 1 min 30 s. The experiment was repeated twice. Amplified fragments were electrophoresed in 1% (w/v) agarose gels and visualized by ethidium bromide staining.

Expression, isolation and reconstitution of recombinant AtPUMP5, and fluorescent monitoring of H⁺ fluxes in proteoliposomes

AtPUMP5 was expressed in Escherichia coli and isolated according to Borecký et al. (2001b). The recombinant AtPUMP5 (100 μg) was reconstituted in proteoliposomes consisting of 39 mg phosphatidyl choline, 1.66 mg cardiolipin, and 0.66 mg phosphatidic acid at a lipid:protein ratio of 410:1. The proteoliposomes internal medium [IM; 28.8 mM tetraethylammonium N-tris [hydroxymethyl]methyl-2-aminoethanesulphonate (TEA-TES; [TES⁻]_free=9.2 mM), pH 7.2, 84.4 mM TEA₂SO₄ and 0.6 mM TEA-EGTA] contained 2 mM SPQ probe.

H⁺ flux assays were based on the quenching of SPQ fluorescence by TES⁻ (Orosz and Garlid, 1993). Vesicles (30 μl; 1.23 mg lipids) were added to 1.5 ml of external medium [EM; 28.8 mM TEA-TES (9.2 mM TES⁻), pH 7.2, 84.4 mM K₂SO₄, and 0.6 mM TEA-EGTA]. Fatty acids were added after 20 s and the H⁺ efflux was initiated by 1.3 μM valinomycin 10 s later. The fluorescence was calibrated to [H⁺] by the addition of 1 M KOH aliquots in the presence of 1.5 μM nigericin to proteoliposomes suspended in IM. Fluorescent data were converted into ‘H⁺ traces’ by fitting with modified Stern–Volmer equation

\[[\mathrm{H}^{{+}}]{=}(1/K_{\mathrm{q}})(F_{0}{-}F)/(F{-}L)\]

where F is the experimental and F₀ the unquenched fluorescence. Parameters K_q (quenching constant) and L (background, mostly light scattering) were obtained by linear regression of F versus (F₀–F)/[KOH] plot. Initial rates of H⁺ efflux were derived from ‘H⁺ traces’, multiplied by the internal proteoliposome volume (V, estimated from volume distribution of the SPQ; Garlid et al., 1996), and corrected for lipid content to yield final rates in nmol H⁺ s⁻¹ mg⁻¹ lipid.

Results

PUMP gene family

Homology-based searches (tBLASTn algorithm; E-value <10⁻⁴⁰) in the Arabidopsis genome database yielded 21 genomic sequences (BAC, TAC, and PL clones) comprising PUMP-like ORFs. Alignment of the protein sequences deduced from these clones with known UCPs and PUMPs, together with M2OMs and DICs, led to the identification of six Arabidopsis genes highly similar to known UCPs/PUMPs (Fig. 1). Using the same search that included all the six Arabidopsis sequences as drivers, an additional 28 PUMP-like sugarcane sequences representing EST clusters produced by the assembly of 237 954 ESTs, which were referred to SASs (Vettore et al., 2003), were identified. Alignment of all of these SASs resulted in the identification of five non-redundant sequences containing ORFs highly similar to known UCPs/PUMPs (Fig. 1). The phylogenetic analysis of the six Arabidopsis and five sugarcane PUMPs along with the UCPs/PUMPs plus M2OM and DIC representatives revealed a complex group that consisted exclusively of uncoupling proteins of all types (Fig. 2A, groups I–VI), distinct from the branch containing M2OMs and DICs (Fig. 2A, group VII).

Fig. 1.

Alignment of UCP/PUMP homologues. Six Arabidopsis and five sugarcane PUMP amino acid sequences deduced from corresponding cDNAs were aligned with 16 representatives of UCP/PUMP. Nine known M2OM/DIC were included to identify their possible orthologues. Amino acid characters: white on black, conserved residues; black on dark grey, identical residues; and black on light grey, similar residues. ETPS are shown in rectangular boxes. The new monocotyledonous AAAA motif is underlined. The accession numbers for each protein sequence are listed in the Materials and methods.

Fig. 2.

Unrooted phylograms of the amino acid sequences of uncoupling proteins (A) and alternative oxidases (B) of all types. The seven groups distinguished within the uncoupling protein tree are: I, UCP1–3; II, PUMP1–2; III, PUMP4–6; IV, UCP5; V, yeast ODC, VI, UCP4 and PUMP3; and VII, M2OM and DIC. The six groups discriminated within the alternative oxidase tree are: I, monocot AOx1; II, dicot AOx1; III, dicot AOx2; IV, possible dicot AOx3; V, plastidic dicot AOx; and VI, fungal AOx. Phylograms were generated as a consensus of 1000 bootstrap replicates by the Neighbor–Joining method (the bootstrap values are indicated close the branch divisions). The scale bar indicates the relative amount of change along branches. The Arabidopsis and sugarcane PUMP and AOx family members are in bold. The accession numbers are given in the Materials and methods.

The novel putative PUMP genes identified within the Arabidopsis genome are summarized in Table 1. AtPUMP1 and AtPUMP2 (AtUCP2) were characterized previously (Maia et al., 1998; Watanabe et al., 1999). The sequence of AtPUMP3 was previously submitted to GenBank as a ‘putative mitochondrial uncoupling protein’. The sequences representing AtPUMP4–6 were identified for the first time in this work. Mapping of these AtPUMPs revealed that each gene is located on a different chromosome, except for AtPUMP2 and AtPUMP6, which are on the opposite ends of chromosome 5 (Table 1). Figure 3A illustrates the genomic structure of PUMP genes in the Arabidopsis genome. Intron/exon composition of AtPUMP1 and AtPUMP2 were almost identical, with only first exon and first and sixth introns differing in their size (Borecký et al., 2001a). Genomic structures of AtPUMP3 and AtPUMP6 were also very similar, whereas AtPUMP4 was similar to AtPUMP5.

Table 1.

PUMP family members from Arabidopsis and their mapping to chromosomes

Protein	CHR positiona	CHR locusb	No. of ESTs	Highest identity (%)
AtPUMP1	20656336-20658442	At3g54110	16	UCP2 (45%)
AtPUMP2	30436384-30434148	At5g58970	2	UCP3 (41%)
AtPUMP3	5488903-5487246	At1g14140	1	UCP4 (41%)
AtPUMP4	1782788-1783729	At4g24570	20	UCP3 (31%)
AtPUMP5	10640322-10641674	At2g22500	13	UCP2 (31%)
AtPUMP6	2950514-2949242	At5g09470	0	UCP3 (31%)

a

Chromosome position in base pairs.

b

Chromosome locus based on MIPS (http://mips.gsf.de/proj/thal/proj/thal_overview.html).

Fig. 3.

Gene structure of Arabidopsis PUMP (A) and AOx (B) genes. Filled boxes represent exons and lines represent introns and 5′- and 3′-untranslated regions according to Locus data from The Arabidopsis Information Resource (TAIR). Bar corresponds to 500 bp of chromosomal DNA.

The five putative sugarcane PUMP orthologues identified in this work were denominated SsPUMP1–5 (Saccharum sp. PUMP1–5; Table 2). Comparison of the protein sequences of SsPUMPs with PUMPs from other plants revealed a new specific insertion of four alanines in the proximity of the first ETPS (underlined in Fig. 1), which may represent a putative monocotyledonous PUMP-specific motif. This motif is also present in the maize PUMP recently identified in our laboratory (Brandalise et al., 2003b). A similar alanine-rich region was also found in PUMPs from wheat (Murayama and Handa, 2000) and rice (Watanabe and Hirai, 2002). This motif is not present in PUMPs from dicotyledonous plants (Fig. 1). In addition to the PUMP sequences identified in sugarcane, a gene encoding a putative malate/2-oxoglutarate carrier was also detected (named here as SsM2OM), and was, along with other M2OM proteins, phylogenetically much closer to known M2OMs than PUMP4–6 (Fig. 2A).

Table 2.

PUMP and AOx family members from sugarcane

Protein	SAS IDa	Accession	No. of ESTs	Highest identity (%)
SsPUMP1	SCRFFL5034A04.g	AY64460	1	UCP2 (47%)
SsPUMP2	SCEPAM2011H12.g	AY64461	22	UCP3 (45%)
SsPUMP3	SCRLSB1040H12.g	AY64462	1	UCP4 (48%)
SsPUMP4	SCCCRZ2C01E09.g	AY64463	31	UCP3 (36%)
SsPUMP5	SCCCLR1076E12.g	AY64464	37	UCP5 (42%)
SsAOx1a	SCACCL6008G11.g	AY64465	5	TaAOx1a (77%)
SsAOx1b	SCRURT3064E04.b	AY64466	2	OsAOx1b (39%)
SsAOx1c	SCCCCL3080G10.b	AY64467	3	OsAOx1c (48%)
SsAOx1d	SCBGLR1095F09.g	AY64468	1	OsAOx1c (30%)

a

Assembled sequences from the SUCEST database (Vettore et al., 2003).

All members of the MACF possess the ETPS (Borecký et al., 2001a). The known uncoupling proteins have three copies of this signature, while other MACF members possess only one or two copies (Table 3). The three copies of the ETPS signature found in all of the new PUMPs displayed specific variants for the PUMPs already identified (Borecký et al., 2001a), while the M2OM sequences lacked the second signature copy (Table 3). The novel PUMP members were also screened for the presence of four UCP-specific signatures proposed by Ježek and Urbánková (2000). PUMP3 from both Arabidopsis and sugarcane showed all three signatures with the exception of one substitution (a non-polar amino acid for Tyr) in the middle of the second signature (Table 3). PUMP4–6 differed slightly; they lacked the negatively charged amino acid (Asp or Glu) in the middle of the first signature and had Gly at this position, and the terminal hydrophobic amino acid (Phe, Leu, or Ile) was substituted for His (Table 3). In the second signature, there were three substitutions of a non-polar amino acid for Thr (positions 1, 6, and 11). In the third signature, a negatively charged amino acid was substituted for Gln (position 13) and the first Pro was changed to Leu (Table 3). All of the new PUMPs had a complete fourth signature, found in all UCPs/PUMPs but UCP5 (Ježek and Urbánková, 2000). Moreover, the MEME-MAST prediction of additional motifs revealed an M2OM/DIC-specific motif [AG]PM[TV][VM][LM]T[FLW]IFL[EM]Q[LMI][NRQ]K located at the C-end of the molecule. This motif was not detected in any of the already known or new UCPs/PUMPs.

Table 3.

Specific variants of the energy transfer protein signature with an additional tripeptide in known UCPs/PUMPs compared with new PUMP genes

Protein name	1st signature	2nd signature	3rd signature
UCP1a	PLDTAKVRL-QLQ	PTEVVKVR[LM]-QAQ	PVDVVKTRF-[IV]NS
UCP2a	PLDTAKVRL-QLQ	PTDVVKVRF-QAQ	PVDVVKTRY-MNS
UCP3a	PLDTAKVRL-QLQ	PTDVVKVRF-QAQ	PVDVVKTRY-MNS
UCP4a	PLDLTKTRL-QMQ	PTDLVKV••-QMQb	PADVIKSRI-MNQ
BMCP1a	PVDLTKTRL-QVQ	PTDVLKIRM-QAQ	PVDVVRTRM-MNQ
AtPUMP1-2	PLDTAKVRL-QIQ	PTDLVKVRL-Q[AS]E	PVDV[VM]KSRM-MGD
AtPUMP3	PIDLTKTRM-QLH	PADLVKVRMQAD	PADVVKTRM-MNG
AtPUMP4	PLDLIKVRM-QLQ	PADVAMVRM-QAD	PVDVIKTRV-MNM
AtPUMP5	PLDLIKVRM-QLQ	PADVAMVRM-QAD	PVDVIKTRV-MNM
AtPUMP6	PLDLIKVRM-QLH	PADVAMVRM-QAD	PIDVVKTRM-MN•b
SsPUMP1	PLDTAKVRL-QLQ	PTDLVKVRL-QAD	PVDVVKSRM-MGD
SsPUMP2	PLDTAKVRL-QLQ	PTDLVKVRL-QAE	PVDVVKSRM-MGD
SsPUMP3	PLDAVKTRL-QLH	PADLMKVRM-QAD	PADVIKTRM-MNQ
SsPUMP4	PLDLIKVRM-QLQ	PADVAMVRM-QAD	PVDVVKTRM-MNM
SsPUMP5	PLDLIKVRM-QLQ	PADVAMVRM-QAD	PVDVVKTRM-MNM
M2OMa	PIDMVKVKI-QLG	–	PFDYVKTQI-QKM
AtM2OM	PIDMIKVRI-QLG	–	PFDYVKTQI-QKM
SsM2OM	PIDMVKVRI-QLG	–	PFDYVKTQI-QKM

a

Signature variants are representative for proteins of all species.

b

Filled circles represent possible amino acid deletions.

Because the AtPUMP5 gene was highly expressed in all organs/tissues tested, as well as in response to chilling stress (described below), its protein product was biochemically characterized. AtPUMP5 was expressed in E. coli, solubilized from inclusion bodies and reconstituted in proteoliposomes. AtPUMP5 mediated H⁺ flux that exceeded the basal H⁺ flux in liposomes without the recombinant protein and that was activated by linoleic acid (K_m=45 μM, V_max=1.4 mmol H⁺ s⁻¹ mg⁻¹ lipid; Fig. 4A) with similar affinity to linoleic acid as AtPUMP1 (K_m=43 μM, V_max=0.32 mmol H⁺ s⁻¹ mg⁻¹ lipid; Borecký et al., 2001b). The proton flux was unaltered in the presence of 5 mM 2-oxoglutarate. The LA-mediated H⁺ flux was weakly inhibited by ATP (K_i=15 mM, Hill coefficient 0.95) and GDP (K_i=10 mM, Hill coefficient 0.98; Fig. 4B). Taken together, the sequence homology analysis of PUMP1–6 and the functional characterization of recombinant AtPUMP5 provided strong evidence that the novel genes identified in Arabidopsis and sugarcane encode true PUMPs and not any other MACF member.

Fig. 4.

(A) Kinetics of linoleic acid-induced H⁺ fluxes in protein-free liposome (open circles) and proteoliposomes with the reconstituted recombinant AtPUMP5 (solid squares) in the presence of 1.3 μM valinomycin. The derived K_m of AtPUMP5-mediated H⁺ flux is 45.91 μM and V_max is 1.40 nmol H⁺ s⁻¹ mg⁻¹ lipid. (B) Dose–response curves for the inhibition of fatty acid-induced AtPUMP1-mediated H⁺ fluxes by GDP (open squares) and ATP (solid triangles) at pH 7.1. The dose–response curves (solid lines) represent inhibition of the outwardly oriented PN binding sites (∼50%). The derived K_i are 10.07 mM (GDP, two experiments) and 15.05 mM (ATP), with Hill coefficients of 1 for both plots.

Alternative oxidase gene family

The five Arabidopsis AOx protein sequences (Saisho et al., 1997; Thirkettle-Watts et al., 2003) were used as drivers to identify AOx orthologues in the SUCEST database. Eleven SASs found as homologous to AOx were reblasted to the GenBank non-redundant database and six of them showed high identity levels (E-value <10⁻⁴⁰) to known AOxs (SsAOxs for Saccharum sp. AOxs, Table 2). Protein sequences deduced from these SsAOxs and AtAOxs, together with 36 representatives of known AOxs from plants and fungi were aligned (Fig. 5). Phylogenetic analysis (Fig. 2B) showed that four putative SsAOxs (Table 2) together with AtAOx1a–1c were grouped within AOx1 type groups (Fig. 2B, group I and II) but not with AOx2 (Fig. 2B, group III), in agreement with recently published results (Considine et al., 2002) showing that monocots lack AOx2 genes. Interestingly, AtAOx1d together with LeAOx1b and PtPtAOx1b formed a new branch, suggesting the existence of a third AOx type: AOx3 (Fig. 2B, group IV).

Fig. 5.

Alignment of alternative oxidase homologues. Five Arabidopsis and four sugarcane AOx amino acid sequences deduced from corresponding cDNAs were aligned with selected known plant and fungi AOx1 and AOx2 types. Amino acid character: white on black, conserved residues; black on dark grey, identical residues; and black on light grey, similar residues. The accession numbers for each protein sequence are given in the Materials and methods.

Digital mRNA expression profiling, RNA gel-blotting, and semi-quantitative RT-PCR analysis of PUMP and AOx genes

EST datasets have been used recently to extract information on gene expression levels (Mekhedov et al., 2000). The rationale of the so-called ‘digital mRNA expression profiling’ is that the number of EST sequences representing a given mRNA obtained from a cDNA library is proportional to the abundance of this mRNA species in the tissue originally used to make the library (Audic and Claverie, 1997). However, some precautions need to be implemented to prevent potentially misleading gene expression results. First, records of individual EST sequences were individually inspected and clone IDs were retrieved and displayed in a catalogue, such that EST sequences obtained from opposite ends of the same cDNA clone were counted as one. Second, only non-normalized sugarcane and Arabidopsis EST collections were used for this analysis. Therefore, the expression profiling of members of the PUMP and AOx gene families in several tissues and/or organs of sugarcane and Arabidopsis was assessed using the information available in the SUCEST and TIGR Arabidopsis EST databases. ESTs representing three respiratory chain genes (ATP synthase, Complex I, and Complex IV subunits) in both species were found in most cDNA pools analysed, suggesting that these libraries are representative of the mitochondrial–nuclear transcriptome (Fig. 6A, B). The results showed that SsPUMP1 transcripts were detected only in floral organs whereas those of SsPUMP3 were found only in stems (Fig. 6A). SsPUMP2 was expressed in most tissues, but preferentially in the meristem and roots. SsPUMP4 and SsPUMP5 were more abundant in stems, leaves, and plantlet inoculated with endophytic bacteria pools (Fig. 6A; Vettore et al., 2003).

Fig. 6.

Relative abundance of PUMP- and AOx-representing ESTs in sugarcane (A) and Arabidopsis (B) tissue-specific library pools and transcript accumulation of SsPUMPs and SsAOxs (C) and of AtPUMPs and AtAOxs (D) in different plant organs. (A, B) The top panels show PUMP ESTs; the middle panels show AOx ESTs; and the bottom panels show ESTs of the nuclear-coded subunits of respiratory chain complexes I (CI) and IV (CIV), and of the ATP-synthase of both plant species. Sugarcane EST library pools: IL, plants inoculated with endophytic bacteria; M, meristem; F, flowers; L, leaves; R, roots; S, seeds; ST, stem. Arabidopsis EST library pools: A, above-ground organs; R, roots; F, flowers; G, green siliques; S, seeds. (C) Total RNA isolated from sugarcane organs was subjected to RNA-blot analysis using probes prepared from full-length cDNAs for SsPUMP2, SsPUMP4, SsAOx1a, and SsAOx1d. L, leaves; ST, stem; R, roots; F, flowers. (D) Total RNA isolated from Arabidopsis organs was subjected to semi-quantitative RT-PCR analysis using specific primers for AtPUMP4, AtPUMP5, AtAOx1b, and AtAOx1d. L, leaves; F, flowers; and R, roots.

In sugarcane, the expression of AOx genes attained much lower levels than the expression of PUMP genes. In general, AOx gene expression was observed preferentially in non-green tissues, in contrast to SsPUMPs (Fig. 6A). ESTs of SsAOx1a and SsAOx1b were found in roots, while SsAOx1c was present in the stem and meristem. SsAOx1d was expressed only in leaves (Fig. 6A).

Interestingly, transcripts of all Arabidopsis PUMPs were detected in roots, but were undetectable in floral organs; no AtPUMP6 expression was detected in any tissue/organ (Fig. 6B), suggesting that it could be a pseudogene or expressing in very low levels. AtPUMP1 was expressed in several tissues, in agreement with the results of Maia et al. (1998). AtPUMP2 was detected in roots and green siliques, whereas AtPUMP3 expression was found only in roots (Fig. 6B). AtPUMP4 showed high levels of expression in roots and seeds, whereas AtPUMP5 expressed predominantly in above-ground organs and roots, and less in green siliques. AtAOx1a expression was observed in above-ground organs and AtAOx1c was detected only in floral organs (Fig. 6B). As with sugarcane, the Arabidopsis AOx genes showed extremely low overall levels of expression when compared with AtPUMPs.

Although the digital prediction of tissue-specific expression can provide clues that may improve our understanding of the function of these proteins, it should be noted that the procedures used to construct the cDNA libraries in large-scale EST projects could undermine subsequent expression profile analysis (Koo and Ohlrogge, 2002). Thus, digital expression profile analysis should be validated experimentally, using methods such as RNA gel-blots and/or reverse transcription-PCR. Thus, the expression of two SsPUMPs (SsPUMP2 and 4), and two SsAOxs (SsAOx1a and 1d) was analysed using RNA gel-blots. Figure 6C shows that SsPUMP2 was expressed preferentially in roots, while only traces of its transcript were found in flowers, leaves, and stems. RNA gel-blots of SsPUMP4 indicated a high level of expression in non-reproductive tissues, especially in leaves and roots. SsAOx1a was expressed at very low levels in all tissues except flowers, whereas SsAOx1d was undetectable in all of the tissues analysed (Fig. 6C). These results indicated that, generally, the RNA gel-blots agreed with the results obtained by digital mRNA expression profiling analysis.

A semi-quantitative RT-PCR was employed to analyse the expression of two novel AtPUMPs (AtPUMP4 and AtPUMP5) and two already known AtAOxs (AtAOx1b and AtAOx1d) because of the low EST frequency of AtAOxs in the library pools analysed (Fig. 6B). AtAOx genes were included in the analysis for comparison of expression patterns of AtPUMPs and AtAOx under the same experimental conditions. The results of RT-PCR demonstrated that AtPUMP4–5 were ubiquitously expressed in leaves, floral organs, and roots at considerably high levels. AtAOx1b transcripts were detected only in floral organs at very low levels whereas AtAOx1d was expressed in all three organs, although the expression in roots was low (Fig. 6D). The discrepancies observed between ‘digital mRNA expression profiling’ and semi-quantitative RT-PCR might be the result of different development stages of the tissues/organs used in both analyses, since these genes were reported to be temporally modulated (Finnegan et al., 1997). Despite of these differences, the Arabidopsis AOx genes showed extremely low overall levels of expression when compared with AtPUMPs, analogously to sugarcane PUMP and AOx genes.

PUMP and AOx expression profiles in response to chilling stress

RNA gel-blots revealed that SsPUMP2 transcript accumulation was unaltered when sugarcane plantlets were submitted to chilling stress (4 °C) for up to 48 h (Fig. 7A). SsPUMP1 and SsPUMP3 were undetectable during chilling treatment suggesting that these genes are not induced by low temperature (data not shown). By contrast, SsPUMP4 and SsPUMP5 were strongly induced, with the transcript levels of both reaching a maximum level after 48 h at 4 °C (Fig. 7A). The level of SsAOx1a transcripts was unaffected by chilling stress, whereas SsAOx1c was slightly induced after 6 h at 4 °C, and returned to basal levels after 12 h (Fig. 7A). By contrast, SsAOx1b and –1d transcripts were undetectable in plantlets exposed to cold (data not shown).

Fig. 7.

Chilling-regulated transcript accumulation of PUMPs and AOxs in sugarcane (A) and Arabidopsis (B). Total RNA from the leaves of chilling-treated and untreated plantlets colleted after 0, 3 (Arabidopsis only), 6, 12, 24, and 48 h at 4 °C was subjected to RNA-blot analysis using probes prepared from full-length cDNAs for SsPUMPs and SsAOxs (A), or AtPUMPs and AtAOxs (B).

AtPUMP1 has already been described as being chilling-inducible (Maia et al., 1998). Therefore the accumulation of transcripts of the other AtPUMPs identified in this work was analysed. As shown in Fig. 7B, AtPUMP2 was not induced by chilling treatment, in agreement with the findings of Watanabe et al. (1999). AtPUMP3 also appeared not to be regulated by low temperature (Fig. 7B). By contrast, AtPUMP4 and AtPUMP5 were strongly induced after 3 h of cold treatment, and returned to basal levels after a 12 h exposure to low temperature (Fig. 7B). Although their expression profiles were similar, the levels of transcript accumulation for both AtPUMPs were different (Fig. 7B). Finally, AtPUMP6 transcripts were undetectable throughout the entire exposure to cold (data not shown).

Figure 7B also illustrates the AtAOxs gene expression in response to low temperature. Under chilling stress, the mRNA levels of AtOx1a increased with time to reach a maximum level after 48 h at 4 °C (Fig. 7B). However, the accumulation of AtAOx1b and –1c transcripts was undetectable throughout the exposure to cold (data not shown). Surprisingly, AtAOx1d, a recently described member of the Arabidopsis AOx gene family (Considine et al. 2002), was clearly down-regulated after exposure to low temperature (Fig. 7B), indicating a possibly different role for this AOx gene in the response to cold stress. Finally, AtAOx2 showed very low levels of transcript accumulation, and was not responsive to chilling stress, in agreement with the results reported by Considine et al. (2002).

Discussion

Identification and tissue-enriched expression profiles of PUMP and AOx gene families in sugarcane and Arabidopsis

The biochemical and physiological roles of PUMP and AOx energy-dissipating systems in plants have been extensively discussed (Sluse et al., 1998; Almeida et al., 1999; Jarmuszkiewicz et al., 2000; Sluse and Jarmuszkiewicz, 2000). Both systems can play a similar role in mitochondrial energy-linked processes, either through tissue-specific thermogenesis and/or the maintenance of tissue temperature, or by protecting plant cells against oxidative stress and/or tuning the mitochondrial

\({\Delta}{\mu}_{\mathrm{H}^{{+}}}\)

across inner mitochondrial membrane (Meeuse, 1975; Day et al., 1980; Borecký et al., 2001a).

This report describes probably the whole family of PUMP-encoding genes present in dicot Arabidopsis and in monocot sugarcane. In Arabidopsis, besides AtPUMP1 (Maia et al., 1998), AtPUMP2 (Watanabe et al., 1999), and a putative plant uncoupling protein gene (AtUCP4) reported by Hanák and Ježek (2001) and classified here as AtPUMP3, three new members, AtPUMP4, AtPUMP5, and AtPUMP6, were identified (Table 1). The location of AtPUMP1, AtPUMP2, AtPUMP4, and AtPUMP5 within or close to duplicated regions of the Arabidopsis genome (Blanc et al., 2000) and the gene structure similarities among pairs of the family members suggest these genes originated through gene duplication events (Fig. 3A). These observations are in keeping with the phylogenetic subfamilies defined in Fig. 2A, corroborating the existence of three distinct groups of PUMPs.

The five AtPUMP orthologues were also identified in sugarcane and named as SsPUMP1, SsPUMP2, SsPUMP3, SsPUMP4, and SsPUMP5 (Table 2). The amino acid sequences deduced from all of the identified genes exhibited the structural features of uncoupling proteins (Table 3). The major feature of PUMPs, the presence of three ETPS (Borecký et al. 2001a), was conserved in all novel PUMP isoforms. In addition, PUMPs were clearly distinguished in specific branches of the phylogenetic trees when analysed together with other members of the MACF family (Fig. 2A). In general, type 1 and 2 PUMPs from both Arabidopsis and sugarcane are homologous to the animal UCPs belonging to types 1, 2, and 3 while type 3 PUMPs are similar to UCP4, and PUMPs of type 4, 5, and 6 formed a distinct PUMP group, analogous to the group formed by the UCP5 members (Fig. 2A).

Millar and Heazlewood (2003) and Picault et al. (2004) classified as putative DICs the two Arabidopsis DNA clones, At4g24570 and At2g22500, that have been identified here as AtPUMP4 and AtPUMP5, respectively (Table 1). The results presented here showed clearly that these genes had the conserved specific motifs characteristic of PUMPs (Fig. 1; Borecký et al., 2001a). In addition, these proteins conserved the sequence of UCP-specific motifs proposed previously (Ježek and Urbánková, 2000) in a high degree. The two substitutions of the negatively charged amino acids for Gly or Gln do not contradict the proposed involvement of these signatures in fatty acid anion binding and translocation (Ježek and Urbánková, 2000). In addition, the M2OM/DIC-specific motif found in all M2OM including AtM2OM and SsM2OM was absent in all new PUMPs. This feature provides strong support that PUMP3–6 are true uncoupling proteins. The functional analysis of reconstituted recombinant AtPUMP5 revealed that this novel PUMP mediated linoleic acid-dependent H⁺ flux (Fig. 4A). Kinetic analysis of linoleic-acid activation of AtPUMP5-mediated H⁺ flux showed that the affinity of AtPUMP5 to this acid was similar to AtPUMP1 (Borecký et al., 2001b). The AtPUMP5-mediated proton flux was unaltered in the presence of 5 mM 2-oxoglutarate strengthening the suggestion that this protein is a true PUMP and not an M2OM. The LA-mediated H⁺ flux was weakly sensitive to ATP and GDP (Fig. 4B) with inhibitory constants in the proposed physiological range (2–15 mM for ATP; Jarmuszkiewicz et al., 2004). However, the weak affinity of AtPUMP5 to purine nucleotides together with the strong induction of the AtPUMP5 gene under stress conditions (Fig. 7A) suggest that the activity of this protein is more probably regulated at the transcriptional level.

The AtAOx1 and 2 family members have been described (Considine et al., 2002; Thirkettle-Watts et al., 2003). Data mining of the sugarcane EST database for genes encoding alternative oxidase revealed four AOx1 family members (SsAOx1a–1d). The phylogenetic analysis showed that SsAOx1a–1d and three Arabidopsis (AtAOx1a–1c) actually belonged to the AOx1 type. An orthologue for Arabidopsis AOx2 type was not found in sugarcane (Fig. 2B). Nonetheless, to our surprise, AtAOx1d, initially classified into the AOx1 type gene family (Considine et al., 2002), clustered with tomato AOx1b (LeAOx1b) and Populus tremula×Populus tremuloides AOx1b (PtPtAOx1b) into a discrete group (group IV, Fig. 2B). This new group was separate from the AOx1 (Fig. 2B, groups I and II) and AOx2 groups (group III, Fig. 2B). Thus, the existence of a new AOx family is proposed, AOx3-type that is probably present exclusively in dicotyledonous plants.

Several reports have described the complex regulation of AOx and PUMP activities on the protein level (for review see Borecký and Vercesi, 2005). The activity of both proteins is regulated by redox state of Coenzyme Q, ROS, and FA (in the opposite direction). AOx activity also depends on concentrations of O₂ and intra-mitochondrial pyruvate and on the redox state of the mitochondrial matrix. PUMP activity is inhibited by extra-mitochondrial purine nucleotides. However, little is known about the factors that affect the transcriptional regulation of PUMPs as well as the differential regulation of the multigene family members in organ/cell-type. The existence of multiple members of the PUMP protein family suggests that these genes may be under the control of factors in a cell-, tissue-, or organ-specific manner. In addition, AtAOx expression profiles were only recently investigated (Thirkettle-Watts et al., 2003). In this report, the expression analyses of PUMPs and AOxs were performed in parallel, to evaluate their expression profiles under the same conditions. The results showed that PUMP members were generally expressed at higher levels in both plants when compared with AOx members (Fig. 6B, D). When seeking for specific gene expression in individual tissues/organs, PUMP and AOx genes displayed very different tissue-enriched expression patterns in both plant species (Fig. 6). In silico and experimental results suggested that AtPUMP4 and AtPUMP5 and their sugarcane counterparts are the most expressed PUMP genes, while most of the AOx family members were expressed at extremely low levels (Fig. 6). The low abundance of AOx transcripts may reflect the low AOx activity observed in tissues of non-thermogenic plants (Finnegan et al., 1997). However, precautions should be taken when interpreting physiological and metabolic responses based only on transcriptional profiles.

Influence of chilling stress on PUMP and AOx expression

Among the SsAOx genes, only SsAOx1c responded to cold treatment, and increased the transcript level after 6 h of low temperature exposure. Chilling treatment did not induce SsAOx1a, whereas SsAOx1b and –1d were not detected at any of the time points analysed (Fig. 7A). Although no expression was detected for some AOxs (and PUMPs as well) in both plants during chilling treatment (data not shown), the possibility exists that they may be down-regulated by colder temperatures. Arabidopsis AOx1a was up-regulated during chilling stress after a 12 h exposure (Fig. 7B). This gene is induced by antimycin A, an inhibitor of Complex III in the mitochondrial respiratory chain (Saisho et al., 1997). Moreover, the overexpression of AOx1a in tobacco transgenic cells reduced the formation of mitochondrial ROS, suggesting a possible role for AOx in protecting plant cells against oxidative stress caused by biotic and/or abiotic stress (Maxwell et al., 1999).

Considine et al. (2002) proposed that AOxs have a housekeeping function in respiratory metabolism and a protective function during stress. Nevertheless, the unexpected down-regulation of AtAOx1d by low temperature (Fig. 7B) might indicate an additional role for AOx proteins. AtAOx1d fell within a novel subgroup of AOx (Fig. 2B, group IV) and its very different response to chilling stress corroborate with a third type of dicot AOx (AOx3), as mentioned above in the phylogenetic analysis. In Arabidopsis, a cold-tolerant plant species, acclimation to low temperatures leads to the modulation of gene expression, including up- and down-regulation of many different genes (Fowler and Thomashow, 2002). These authors suggested that transferring Arabidopsis from a warm to a low temperature produces ‘waves’ of changes in its transcriptome. It is possible that these ‘waves’ can differentially regulate members of a gene subfamily, such as the AOx families.

Analysis of the expression profile of the SsPUMP genes during chilling stress revealed that SsPUMP2 was not chilling-regulated (Fig. 7A). By contrast, a strong response was observed for SsPUMP4 and SsPUMP5 after a 12 h exposition to 4 °C (Fig. 7A). Interestingly, the EST frequency of SsPUMP4 and SsPUMP5 was elevated in EST libraries prepared from sugarcane plantlets inoculated with the nitrogen-fixing bacteria Acetobacter diazotroficans or Herbaspirillum rubrisubalbicans (Vettore et al., 2003), suggesting that these genes may respond to biotic and abiotic stress signalling. The Arabidopsis counterparts of SsPUMP4 and SsPUMP5 were also up-regulated by exposure to low temperature, but displayed a very different transcript accumulation pattern (Fig. 7B). Analogously to other UCPs/PUMPs (Laloi et al., 1997; Maia et al., 1998), these proteins may be involved in reducing ROS generation during chilling stress in monocot and dicot plants.

As shown in Fig. 7, the expression profiles of PUMP compared with the AOx genes were quite different in sugarcane and Arabidopsis, suggesting that multiple regulatory pathways were likely to be involved in PUMP and AOx gene expression. Interestingly, computational inspection of the promoter region (within a 1.0 kb region upstream of the transcription initiation site) of AtPUMP4, AtPUMP5, and AtAOx1a using PLACE (Higo et al., 1999) and PlantCARE (Lescot et al., 2002) indicated the presence of several copies of a TCTCC core sequence. This sequence is recognized by the ADR1 transcriptional factor that is involved in oxidative processes (Simon et al., 1991) and activates peroxisomal proteins (Simon et al., 1991), which also possess ETPS (Jank et al., 1993).

Another interesting finding was that SsPUMP4 and –5 were long-term up-regulated by chilling stress, while SsAOx1c was up-regulated transiently (Fig. 7A). In Arabidopsis, the opposite was observed (Fig. 7B), i.e. AtPUMP4 and AtPUMP5 were induced rapidly and transiently and AtAOx1a was long-term up-regulated after exposure to low temperature. Sugarcane, a chilling-sensitive plant species, appears not to have a complete cold-regulated signal pathway compared with Arabidopsis and other temperate-climate species (Nogueira et al., 2003). This finding suggests that, in addition to the temporal regulation of PUMPs and AOxs within a plant species, there could also be species-dependent regulatory pathways in the response to chilling stress.

Although recent reports have already described the intricate transcriptional regulation of AOxs (Thirkettle-Watts et al., 2003; Borecký and Vercesi, 2005; Dojcinovic et al., 2005), the results of this work suggest that the members of PUMP and AOx energy-dissipating system are subject to different cell/tissue/organ or even stress-specific transcriptional regulation. As a result, plants may respond more flexibly to adverse biotic and abiotic conditions, in which oxidative stress is involved. Furthermore, the present report is aimed at enhancing knowledge of PUMP and AOx regulation with characteristics of their regulation on the transcriptional level, with emphasis on the existence of several isoforms of both PUMP and AOx. As the papers describing biochemical/physiological aspects of these proteins usually refer to ‘PUMP’ or ‘AOx’ as singular proteins, it should be noted that this result suggests that each PUMP as well as each AOx isoform may have a different role. Thus the observations of the ‘PUMP’ (or ‘AOx’) overall activity can be a result of co-operative activities of various members of the corresponding family that should be taken into account. Further genetic and corresponding functional studies on different uncoupling protein and alternative oxidase family members in various plant species are necessary to establish the true physiological roles of the isoforms of these proteins.

*

These authors contributed equally to this work.

†

Present address: Instituto de Biociências, Departamento de Genética, Universidade Estadual Paulista (UNESP), 18618–000, Botucatu, SP, Brazil.

The authors thank AS Zanca, S Meire, GS Ferraz, CR Brambile, and AL Beraldo for excellent technical assistance and L Neubauer for providing the AtPTR specific primers. This work was supported by FAPESP grants to PA (no. 98/12250-0) and to AV (no. 03/08514-1). FTSN was the recipient of fellowships from FAPESP and JB, KAPO, and IGM were recipients of fellowships from CNPq.