INTRODUCTION

Top Abstract Introduction Materials & Methods Results & Discussion References

The mitochondrial pyruvate dehydrogenase complex (mtPDC)¹ catalyzes the oxidative decarboxylation of pyruvate, yielding acetyl-CoA and NADH to support the Krebs cycle and oxidative phosphorylation (1, 2). In addition to the mtPDC, plants have a second PDC localized in the plastid stroma, where it provides acetyl-CoA for fatty acid biosynthesis (3, 4). The plastid and mitochondrial PDCs have considerably different properties reflecting their unique roles in metabolism (reviewed in Ref. 5). The most striking difference is regulation by reversible phosphorylation. The mtPDC, unlike the plastid counterpart, has an associated PDH kinase and phospho-PDH phosphatase, which catalyze reversible phosphorylation of the  subunit of the PDH component (reviewed in Ref. 6). Pyruvate dehydrogenase kinase (PDK, EC 2.7.1.99) catalyzes inactivation of PDC while the phospho-PDH phosphatase reactivates mtPDC. Consequently, the phosphorylation status of mtPDC is determined by the activities of these opposing enzymes, and physiological effectors of plant PDKs would thus regulate the phosphorylation status of mtPDC in vivo (6).

In most plants, leaf mtPDC is phosphorylated in the light by a photosynthesis-photorespiration-sensitive mechanism (7, 8). For C₃ plants, this is most likely due to photorespiratory glycine metabolism that occurs in the leaf mitochondria during photosynthesis. Glycine oxidation generates large amounts of NADH to support mitochondrial ATP production, as well as NH₄⁺, which stimulates PDH kinase (9). Consequently, mtPDC is negatively regulated as glycine oxidation increases. To further understand reversible phosphorylation of mtPDC in plants and its role in the control of Krebs cycle activity, we have undertaken a molecular analysis of maize PDK.

The molecular cloning of the first PDK cDNA from rat (10) showed the deduced primary amino acid sequence lacked typical eucaryotic Ser/Thr kinase domains, but had the domains diagnostic of procaryotic two-component histidine kinases (reviewed in Ref. 11). Procaryotic histidine kinases autophosphorylate on a His followed by phosphotransfer to Asp (or Glu) of their response regulator protein to transduce cellular signals (reviewed in Ref. 12). Based on similarities to histidine kinases, PDKs could also utilize His for phosphotransfer, but unlike histidine kinases PDKs phosphorylate Ser residues. There is as yet no evidence for His autophosphorylation for the PDKs. This new class of eucaryotic protein kinases also includes the branched chain -keto acid dehydrogenase kinase (BCKDH kinase; Ref. 13), which regulates a related mitochondrial -keto acid complex involved in branched chain amino acid degradation. Although the two kinases are related, they are specific for their respective complex (14).

The domains responsible for substrate recognition, complex association, and catalysis by PDKs have not yet been identified. More information on the structure-function relationships of this unique class of eucaryotic protein kinases might become apparent from comparison of primary amino acid sequence data from divergent organisms. Based on similarities to the rat PDK, putative PDKs have also been identified from humans (15, 16), fruit fly (17), and nematodes (18). Here we report the molecular cloning of two plant PDK homologues, and characterization of the recombinant proteins.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results & Discussion References

Identification of Maize Pyruvate Dehydrogenase Kinase cDNAs-- The maize cDNAs were identified from the Pioneer Hi-Bred International (Johnston, IA) maize expressed sequence tag data base by screening with the entire rat PDK amino acid sequence (10). Multiple cDNAs with varying degrees of homology were identified using the Basic Local Alignment Search Tool (BLAST; Ref. 19). Two unique maize cDNAs that were homologous to the C-terminal portion of the rat PDK and of sufficient length were identified for subsequent sequencing.

Preparation of a Mitochondria Matrix Fraction and Kinase-depleted PDC-- Mitochondria were isolated from etiolated maize (B73, Illinois Seed Foundation, Urbana, IL) shoots according to procedures described previously (20). Isolated mitochondria were resuspended in 30 mM TES-KOH, pH 7.5, 2 mM DTT, then homogenized on ice with a Polytron (Brinkmann, Westbury, NY). The homogenate was centrifuged in a TL100 centrifuge at 100,000 × g using a TL100.3 rotor for 15 min. Supernatants, termed the 100K enzyme, were concentrated with an Amicon ultrafiltration membrane (XM300, 300-kDa cut-off). This protein was layered onto a 10-50% linear glycerol gradient containing 50 mM TES-KOH, pH 7.5, and 2 mM DTT, then centrifuged in an SW28 rotor for 16 h at 25,000 rpm. The PDC activity peak at approximately 30% glycerol had only 5% (± 4%, n = 7) PDK activity and was termed kinase-depleted PDC. PDC activity assays were performed as described previously (20).

Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR) of Maize RNA-- Approximately 60 µg of maize total RNA isolated from various organs was treated with 5 units of RNase-free DNase (Boehringer Mannheim) for 2 h at 37 °C in 10 mM MgCl₂, 1 mM dithiothreitol, 50 units of RNasin RNase inhibitor (Promega, Madison, WI). The RNA was then extracted with phenol, precipitated with ethanol, and resuspended in nuclease-free water. The RNA was quantitated by absorbance at 260 nm and diluted to 10 ng/µl for use in RT-PCR.

Construction of Expression Plasmids-- Primers DDR 193 (5'-gtcacgcccggggaattcaccATGGCGTCGGAGCCG-GTGGCGCGG) and DDR 194 (5'-tcattactcgagctgcagctatcaTTACGGCAAGGGTTCCTCCGA) were used to amplify PDK1 and PDK2 open reading frames (amino acids 1-347, PDK1; 1-364, PDK2) corresponding to the region between 56 and 1098 (PDK1) and 78 and 1172 (PDK2) base pairs. To ensure translation termination, the two remaining nonsense codons were introduced into DDR194 (underlined). Restriction sites were introduced into each primer at the 5' end (lowercase letters) to facilitate subcloning of the PCR fragment in the proper reading frame. The EcoRI (DDR 193) and XhoI (DDR 194) sites were used to subclone the PCR products into pET28a expression vector (Novagen, Madison, WI), which encodes six His residues, followed by an 11-amino acid T7 epitope tag upstream of the multiple cloning site. The EcoRI and PstI (DDR194) sites were used to subclone into pMAL-cRI (New England Biolabs, Beverly, MA), which encodes for a 385-amino acid maltose-binding polypeptide upstream of the multiple cloning site. Thermal cycling reactions (50 µl total volume) contained 10 mM Tris-HCl, pH 7.9, 0.5 mM MgCl₂, 200 µM dNTPs, 5 units of Taq polymerase (Promega, Madison, WI), 2 ng of plasmid cDNA template, 5% dimethyl sulfoxide, and 20 pmol of each primer. Cycling conditions were 94 °C for 5 min, initial denaturation, followed by 30 cycles of 30 s at 94 °C, 30 s at 50 °C, 2 min at 72 °C with 6-s extensions for the last step of each cycle.

Expression and Purification of Recombinant His₆-tagged PDK1 Protein and Preparation of Antibodies-- A single colony of recombinant Escherichia coli BL21(DE3) was inoculated into 2 ml of LB medium supplemented with kanamycin (50 µg/ml), and incubated with shaking at 37 °C overnight. The cells were transferred to 100 ml of LB plus kanamycin in a baffled Erlenmeyer flask and shaken at 37 °C until the A₆₀₀ reached 0.4 (2-4 h). The target gene was induced by adding isopropyl-1-thio--D-galactopyranoside to a final concentration of 0.1 mM and continuing shaking at 37 °C for another 4-8 h. The cells were harvested by centrifugation (5000 × g for 20 min) and recombinant protein purified by nickel-nitrilotriacetic acid chelate chromatography under denaturing conditions according to Qiagen (Chatsworth, CA) protocols. Dialyzed recombinant protein (50 µg) was emulsified with 0.5 ml of Freund's complete adjuvant (Sigma) and injected into New Zealand White rabbits. The rabbits were boosted once with 50 µg of recombinant protein plus incomplete adjuvant. SDS-PAGE and immunoblotting were performed as described by Thelen et al. (20).

Expression and Purification of a MBP-PDK Fusion Protein-- The MBP-PDK chimera was expressed similarly to PDK1 in BL21 E. coli host cells with ampicillin selection. The target gene was induced with 1 mM isopropyl-1-thio--D-galactopyranoside for 6 h at 37 °C. The cells were harvested and resuspended in ice-cold MBP wash buffer (20 mM Tris-HCl, pH 7.4, 0.2 M NaCl, 10 mM 2-mercaptoethanol, 5 mM EDTA) plus 1 mM phenylmethanesulfonyl fluoride, 1 mM benzamidine, and 1% (v/v) Triton X-100. The resuspended cells were disrupted by ultrasonic treatment using three 45-s pulses at 50 watts, while cooling on ice between pulses. The suspension was centrifuged at 10,000 × g for 15 min. The supernatant was applied to an amylose column (New England Biolabs, Beverly, MA) that had been equilibrated with MBP wash buffer. The column was subsequently washed with 50 volumes of MBP wash buffer. Bound MBP fusion protein was eluted with three volumes of MBP wash buffer plus 1% (w/v) maltose. The purified fusion protein was dialyzed for 16 h in 20 mM TES-KOH, pH 7.4, 10% glycerol, 1 mM DTT, 0.1 mM phenylmethanesulfonyl fluoride, and 0.1 mM benzamidine. After dialysis, the fusion protein was concentrated with an Amicon ultrafiltration membrane (PM30, 30-kDa cut-off) and stored at 80 °C in 0.1-ml aliquots.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials & Methods Results & Discussion References

Several conventional approaches were unsuccessful in obtaining cDNA clones encoding plant PDKs. However, two cDNAs encoding putative PDKs were identified from a maize expressed sequence tag data base (Pioneer Hi-Bred International), based on sequence similarity to the rat PDK. The two cDNAs are 1332 and 1602 bp in length with open reading frames starting with ATG codons at bases 55 and 78 and in-frame stop codons at 1096 and 1170 bases, respectively. Although no stop codons upstream of the initiating methionine were found, the translation start codon could be predicted when the deduced amino acid sequences of the two maize PDKs were aligned. For clarity, we will term the 1332- and 1602-bp cDNAs and the proteins they encode as PDK1 and PDK2, respectively. These cDNA clones encode polypeptides of 347 and 364 amino acids with calculated molecular weights of 38,867 and 41,327, respectively, and share 78% overall amino acid identity. A major difference between these isoforms is the 16 fewer amino acids (168-183 of PDK2) within PDK1. The function of this domain and reason for its absence in PDK1 are uncertain, although it may impart unique properties to this isoform.

The first 30 residues of both PDK1 and PDK2 (Fig. 1) are rich in Ala, Arg, and Val residues and can form amphipathic  helices. These are characteristic of mitochondrial targeting sequences (22). Another feature is the low abundance of acidic residues, although approximately 40% of plant mitochondrial targeting sequences contain such residues (23) including the maize PDKs, which contain three in the first 30 amino acids. Processing generally occurs at sites with Arg residues 3 and 10, 11 amino acids upstream (22, 24). Arg⁸ and Arg¹⁶ at the 11 and 3 position may signal processing after Gly¹⁸. Alternatively, Arg²¹ and Arg²⁸ (conserved Lys²⁸ for PDK2) are spaced properly and could also signal processing after Met³⁰.

View larger version (98K): [in this window] [in a new window]   Fig. 1.   Comparison of the deduced amino acid sequences of rat and maize PDKs. The deduced amino acid sequence for the rat PDK (GenBank accession no. L22294; Ref. 10) was aligned with those from maize using the GeneWorks software package from IntelliGenetics, Mountain View, CA. Shaded residues indicate amino acid identity. Histidine kinase domains are underlined, and essential residues within these domains are marked with an asterisk. Conserved Trp and Cys residues are indicated with a closed circle. Other conserved domains mentioned in the text are indicated with dashed lines.

The deduced amino acid sequences of the maize PDKs are approximately 37% identical to mammalian PDKs (for rat, see Refs. 10 and 25; for human, see Refs. 15 and 16). The homologous regions span the entire polypeptide, although the C-terminal half has the highest similarity as illustrated by comparison to the rat PDK (Fig. 1). The relatedness of PDKs and procaryotic histidine kinases is confined to five subdomains (11), defined by essential residues also conserved in the maize PDKs (Fig. 2). Within subdomain I is a conserved His¹¹⁷ (numbering according to PDK2), that in the procaryotic His kinases is the autophosphorylation site, involved in phosphotransfer. In mammalian PDKs, e.g. rat, two motifs on either side of this invariant His, (R¹⁴⁶NR¹⁴⁸)_rat and (P¹⁵⁴TMAQGV¹⁶⁰)_rat, are conserved, but not conserved in the BCKDH kinase and are conservatively substituted to (R¹¹⁴XR¹¹⁶)_maize and (P¹²²(T/A)(M/I)AXGV¹²⁸)_maize in the maize PDKs. In the mammalian PDKs, the motif (K²⁸⁰ NAMRAT²⁸⁶)_rat containing an essential Asn²⁸¹ within subdomain II of the rat PDK is proposed to be the hinge region allowing the ATP binding domains to interact with the phosphotransfer domain (11). In maize, the canonical Asn is present along with the two basic residues (K²⁴²NXXRAX²⁴⁸)_maize. Subdomain III of the rat PDK (S³¹⁵DRGGG³²⁰)_rat, containing the signature DXGXG is also present in both maize PDKs. Subdomain IV contains the essential Tyr²⁷⁹ residue within the conserved motif (F²⁹³XYXYSTA³⁰⁰)_maize. Subdomain V is defined by a glycine-rich motif (A³¹⁷GXGXG³²²)_maize, which like subdomain III has essential Gly residues that may be involved in ATP binding (11).

View larger version (32K): [in this window] [in a new window]   Fig. 2.   Model of maize PDKs. The five subdomains of the procaryotic histidine-like kinases, previously identified by Harris et al. (11) to be found in the mitochondrial protein kinases, are shown by dark shading and Roman numerals I-V. Residues also conserved in the procaryotic histidine-like kinases are denoted with an asterisk. The six additional subdomains found in PDKs and discussed in this report are indicated by light shading and letters A-F. The amino acid consensus sequence for each subdomain is indicated. The rat PDK (10) and rat BCKDH kinase (13) were included for comparison.

In addition to the five conserved subdomains characteristic of the procaryotic two-component histidine kinases, the maize PDKs have six other subdomains that are also well conserved in mammalian PDKs and semi-conserved in the BCKDH kinase (Fig. 2). The N-terminal motif (F⁷⁸LXXELP(V/I)RXA⁸⁸)_rat is conserved in all PDKs and the BCKDH kinase, suggesting that it has a common role in all keto acid dehydrogenase kinases. The most conserved domain throughout the entire family of polypeptides is not one of the five His kinase subdomains but rather a 15-amino acid motif between subdomains I and II with the consensus (F¹⁵¹LDRFYMSRIXIRML¹⁶⁵)_maize. Interestingly, this motif is located immediately upstream of the 16 "missing" amino acids in PDK1. This motif is less conserved in the BCKDH kinase FLDXXXXSRXXIRML, suggesting a PDK-specific function. Another highly conserved motif is immediately downstream of subdomain V with the consensus (P³²⁴ISRLYAXYFXGDL³³⁷)_maize, corresponding to PXSRXYAXYXXGXL in the BCKDH kinase. Four residues downstream from this motif is the consensus (S³⁴¹XEGYGTDA³⁴⁹)_maize for PDKs and SXXGXGTDX for the BCKDH kinase. Overall, these four domains in the maize PDKs are more similar to the mammalian PDKs than to the BCKDH kinase, further suggesting the maize proteins are PDKs.

An interesting feature of the maize PDK primary sequence is the paucity of Trp and Cys residues. Trp⁸³ is conserved in all PDKs but not the BCKDH kinase and is part of the consensus (V⁸⁰XXWYXXS⁸⁷)_maize, possibly involved in PDK-specific function (Fig. 2). Cys²⁰⁶, found in the conserved motif (A²⁰²RXXCXXY²⁰⁹)_maize and conserved in all PDKs and the BCKDH kinase, might be involved in catalysis or inter- but probably not intrathiol disulfide exchange since PDK1 does not contain a second Cys.

Dendrogram analysis with putative and characterized PDKs reveals at least three groups of related PDKs (Fig. 3). The Ascaris suum and hypothetical Caenorhabditis elegans proteins form a group that may also contain the putative Drosophila melanogaster PDK. The characterized mammalian PDKs form a closely linked group more related to other animal than plant PDKs. The BCKDH kinase is divergent from all PDKs, and the procaryotic His kinase PhoM (26) outlies the entire group.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Dendrogram analysis of dehydrogenase complex kinases. GenBank accession numbers are as follows: A. suum, U94519; C. elegans hypothetical protein ZK370.5, M98552; D. melanogaster, D88814; human, L42450; rat, L22294; and maize (AF038585 PDK 1, AF038586 PDK 2). The rat branched chain keto acid dehydrogenase kinase (M93271) and the procaryotic histidine kinase PhoM (M13608) were included for comparison. Clustal analysis was performed with GeneWorks software. The length of horizontal lines indicates inverse degree of relatedness. The percent amino acid identity and identity plus similarity of the various proteins with maize PDK2 are indicated.

Expression Pattern of PDK Isoforms-- The high degree of nucleotide similarity between PDK1 and PDK2 prevented the specific detection of transcripts when using restriction enzyme-digested DNA fragments as probes by Northern analysis. However, 20-base oligonucleotides, designed to regions of low homology, enabled specific detection of transcripts by RT-PCR. The oligonucleotide pair for PDK1 did not amplify PDK2 cDNA, nor did PDK2 primers amplify PDK1 cDNA, under the same conditions RT-PCR was carried out (data not shown). After DNase treatment, the RNA was devoid of DNA, as determined by PCR amplification prior to reverse transcription. Although RT-PCR is, at best, semiquantitative, it enables detection of low abundance transcripts with high specificity, which is why it was used. The overall pattern of expression for PDK1 transcript is clearly different from the transcript for PDK2 (Fig. 4). Whereas PDK1 appears to be somewhat constitutive in its expression pattern, PDK2 is up-regulated in leaves. The higher expression of PDK2 in green leaves might enable acute response to mitochondrial ATP concentration during photosynthesis, a model consistent with the photosynthetic-induced inhibition of mtPDC activity observed in maize and other plants (7, 8).

View larger version (49K): [in this window] [in a new window]   Fig. 4.   RT-PCR analysis of maize PDK isoforms. Isoform-specific oligonucleotides were used to amplify transcript from total RNA isolated from 5-day dark-grown seedlings, roots and leaves from 14-day light-adapted seedlings, and ear husks and shoots derived from adult (~90 day), greenhouse-grown maize. Exactly 5 µl of the PDK1 and PDK2 reactions were resolved on a 1.5% agarose gel. Molecular sizes are indicated.

Heterologous Expression of Maize PDKs-- Recombinant His₆-tagged PDK1 expressed in E. coli was greater than 95% insoluble under all conditions tested; therefore, it was purified under denaturing conditions. Recombinant His₆-tagged PDK1 and PDK2 polypeptides were typically 95% pure after affinity chromatography and migrated at 44 ± 2 (S.D., n = 7) and 48 ± 3 (S.D., n = 3) kDa, respectively (Fig. 5). The recombinant proteins were slightly larger than the predicted mass of 38,867 and 41,327 Da, because the His₆ and T7 epitope tags plus multiple cloning site add 37 amino acids (approximately 3.5 kDa) to the N terminus. However, even when the N-terminal tags are accounted for, the recombinant proteins migrate 3-4 kDa larger upon SDS-PAGE than predicted. This anomaly was also observed with protein translated in a rabbit reticulocyte lysate (data not shown). In this case, no additional amino acids were present at the N terminus, yet the apparent size was 3-4 kDa larger than predicted. Size discrepancies have also been observed with another subunit to this complex, the dihydrolipoamide acetyltransferase, which migrates 10-15 kDa slower during SDS-PAGE. Recombinant PDK1 was used as an antigen to raise rabbit polyclonal antibodies. Antibodies to PDK1 recognized both PDK recombinant proteins by immunoblot analysis (data not shown). Purified IgG immunoprecipitated 75% of PDK activity from a 100K maize mitochondrial extract while the preimmune IgG had no effect (Fig. 6A). The non-precipitated kinase activity might be due to an immunogenically distinct kinase or incomplete precipitation. PDK1 antibodies did not recognize any polypeptides from a total maize mitochondrial fraction (Fig. 6B, lane 1). However, upon enrichment for PDC activity by rate-zonal sedimentation a 43 ± 2 (S.D., n = 5) kDa polypeptide was decorated with PDK antibodies (Fig. 6B, lanes 2 and 3). This polypeptide was not detected after glycerol gradient fractionation of PDC, in agreement with the loss of kinase activity in this fraction (Fig. 6B, lane 4). The smaller size of the mitochondrial matrix protein, compared with recombinant PDKs, is due to the absence of the His₆ and T7 epitope tags and processing of the mitochondrial targeting peptide.

View larger version (11K): [in this window] [in a new window]   Fig. 5.   Purification of recombinant PDK fusion proteins. The PDK proteins were expressed either as His6-tagged proteins or as fusions with MBP. Approximately 1 µg of each purified fusion protein was resolved by SDS-PAGE and stained with Coomassie Blue. Molecular weight markers are indicated.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   A, immunoremoval of PDK activity from a maize mitochondrial matrix fraction. Purified anti-rPDK1 IgGs were incubated with a 100K enzyme fraction and protein A-agarose. The 100K supernatant was incubated with 10 mM KF (to inhibit phospho-PDH phosphatase) and assayed for MgATP-dependent inactivation of PDC with 2 mM MgATP for 50 min at 25 °C. Minus ATP controls for each time point were included. B, immunoblot of PDC purification fractions probed with anti-PDK1-His6 antibodies. Purified mitochondria from etiolated maize shoots were homogenized and centrifuged 100,000 × g for 15 min, supernatant was termed "100K enzyme." The 100K enzyme was centrifuged at 400,000 × g for 6 h and the resuspended pellet termed "400K enzyme." The 400K enzyme was layered onto a linear 10-50% glycerol gradient and centrifuged at 25,000 rpm for 16 h in a SW28 rotor. PDC activity fractions were pooled and termed "glyc gradient." Approximately 25 µg of protein was loaded in each lane of the immunoblot.

View larger version (24K): [in this window] [in a new window]   Fig. 7.   MBP-PDK2-mediated MgATP-dependent inactivation of kinase-depleted maize PDC. Approximately 65 µg of kinase-depleted maize PDC was incubated with 3 µg of recombinant protein and MgATP at the indicated concentration, at 25 °C. Kinase-depleted PDC showed 3% inactivation with 500 µM MgATP after 35 min. Inset shows the incorporation of 32P from [-32P] ATP into the E1 subunit, without and with MBP-PDK2. Incorporation of 32P proceeded for 2 h at 25 °C with 10 µM Mg-[-32P]ATP (specific activity = 10 mCi/mmol), stopped with sample buffer, and resolved on SDS-PAGE.