(Received for publication, February 6, 1997, and in revised form, May 2, 1997)
From the Department of Biochemistry and Molecular Biology, Indiana University School of Medicine, Indianapolis, Indiana 46202-5122
Protein X, recently renamed dihydrolipoamide dehydrogenase-binding protein (E3BP), is required for anchoring dihydrolipoamide dehydrogenase (E3) to the dihydrolipoamide transacetylase (E2) core of the pyruvate dehydrogenase complexes of eukaryotes. DNA and deduced protein sequences for E3BP of the human pyruvate dehydrogenase complex are reported here. With the exception of only a single lipoyl domain, the protein has a segmented multi-domain structure analogous to that of the E2 component of the complex. The protein has 46% amino acid sequence identity in its amino-terminal region with the second lipoyl domain of E2, 38% identity in its central region with the putative peripheral subunit-binding domain of E2, and 50% identity in its carboxyl-terminal region with the catalytic inner core domain of E2. The similarity in the latter domain stands in contrast to E3BP of Saccharomyces cerevisiae, which is quite different from its homologous transacetylase in this region. The putative catalytic site histidine residue present in the inner core domains of all dihydrolipoamide acyltransferases is replaced by a serine residue in human E3BP; thus, catalysis of coenzyme A acetylation by this protein is unlikely. Coexpression of cDNAs for E3BP and E2 resulted in the formation of an E2·E3BP subcomplex that spontaneously reconstituted the pyruvate dehydrogenase complex in the presence of native E3 and recombinant pyruvate decarboxylase (E1).
The pyruvate dehydrogenase complex
(PDC)1 catalyzes the
oxidative decarboxylation of pyruvate with the formation of
CO2, acetyl-CoA, and NADH. The eukaryotic complex has 30 copies of a tetrameric (2
2) pyruvate
decarboxylase (E1) component noncovalently bound along the
edges of an icosahedral 60 meric dihydrolipoamide acetyltransferase (E2) core (1). Twelve copies of a homodimeric
dihydrolipoamide dehydrogenase (E3) component are believed
held on the faces of the E2 core by a corresponding number
of monomeric E3-binding proteins (protein X,
E3BP) (2, 3). With the exception of E3BP, the
role of each enzymatic component in the overall reaction catalyzed by
the complex is basically understood. E1 catalyzes a
thiamine diphosphate-dependent oxidative decarboxylation of pyruvate and the reductive acetylation of a lipoyl residue covalently attached to the lipoyl domain of E2. E2 then
catalyzes transfer of the acetyl group to coenzyme A, leaving a reduced
E2 lipoyl group that the E3 component uses as
an electron source for FAD-dependent reduction of
NAD+ to NADH.
Called protein X because its function was not apparent at the time it was originally discovered to be a component of eukaryotic PDCs (4-6), E3BP has generated considerable interest because of sequence similarity with E2 (7-10) and evidence that it has a covalently bound lipoyl moiety (4, 6, 8, 11). Its function was difficult to establish in early studies because of a very tight association with the E2 core of the complex. Very elegant limited proteolysis and immunological studies (12, 13) provided the first evidence that protein X contributes to the binding and function of E3. Cloning of the E3BP gene of Saccharomyces cerevisiae (14) revealed a protein structure that resembles the E2 component of yeast in its amino terminus but not the remainder of the molecule. Subsequent gene disruption studies provided definitive proof that protein X should be considered an E3-binding protein (15). Questions left unsettled included whether mammalian E3BP is completely analogous to yeast E3BP, whether the lipoyl moiety of mammalian E3BP is functionally important for catalytic activity of the complex, and whether the inner core of the mammalian E3BP has transacetylase activity.
The deduced amino acid sequence for the first mammalian (human)
E3BP is presented here. Although similar in their
amino-terminal, lipoyl-bearing domains, the mammalian E3BP
and the yeast E3BP are markedly different in their
carboxyl-terminal regions. Indeed, the human E3BP is more
homologous to mammalian E2 throughout its primary sequence
than it is to yeast E3BP. Previous evidence that the
mammalian E3BP contains only a single lipoyl domain in its amino terminus (8), rather than the two tandemly arranged lipoyl domains characteristic of mammalian E2, is confirmed. The
active site histidine residue characteristic of transacetylases of all -ketoacid dehydrogenase complexes is not conserved in the mammalian E3BP.
Two clones found by searching of the expressed sequence tag data base of the National Center for Biotechnology Information (National Library of Medicine, Bethesda, MD) were analyzed in this study. H58032 was identified by the presence of sequence encoding for the amino-terminal peptide of E3BP, and T77385 was identified by a sequence encoding an internal peptide of E3BP. The cDNA of H58032 in pT7/T3 vector is 2,335 base pairs long; the cDNA of T77385 in Lamfid vector is 1,965 base pairs long.
SequencingSequencing of double-stranded plasmid DNA was done by the Biochemistry Biotechnology Facility (Indiana University) using Taq DyeDeoxy terminator cycle sequencing kit with AmpliTaq DNA polymerase, (Perkin-Elmer) following manufacturer instructions. Both strands were sequenced.
Construction of a Bacterial Expression Vector for the Human E2 Component of the Pyruvate Dehydrogenase ComplexTo
remove the mitochondrial targeting sequence and 3-non-coding region
from the cDNA for the human E2 component (kindly
provided as a generous gift by Dr. Mulchand Patel, SUNY at Buffalo,
NY), NdeI and XhoI sites flanking the coding
region were created by polymerase chain reaction (PCR) with
Pfu DNA polymerase (Stratagene). The sense/antisense PCR
primers correspond to bases 163-190 and 1827-1850 of human
E2 cDNA (CAT ATG AGT CTT CCC CCG CAT CAG
AAG GTT CC(A/C)TC GAG AGT TAC AAC AAC ATA GTG ATA GGT). The
sequence coding for NdeI restriction site was added to sense
primer, whereas the sequence coding for XhoI site was added
to antisense primer (underlined). The respective
NdeI/XhoI fragment was subcloned between
NdeI and XhoI sites of pET-23a expression vector
(Novagen) (plasmid PDHE223a). The fidelity of constructs was
established by nucleotide sequencing.
The expression cassette for human E2 (carrying T7 promoter, ribosome-binding site, cDNA for E2 component, and T7 terminator) was cut from pPDHE223a with BglII and NaeI. Resulting DNA was purified, blunt-ended with T4 DNA polymerase, and ligated into the pACYC vector (New England Biolabs) cut with HindIII and blunt-ended with T4 DNA polymerase (plasmid PDHE2). Plasmids were transformed into competent TG-1 cells. Transformants were selected on YT agar plates containing 35 µg/ml of chloramphenicol. Colonies that expressed chloramphenicol resistance were screened for the presence of inserts by PCR with the above primers.
Construction of a Bacterial Expression Vector for Human E2·E3BP Subcomplex of the Pyruvate Dehydrogenase ComplexNcoI and EcoRI restriction sites flanking the coding region of E3BP cDNA were constructed by PCR with Pfu DNA polymerase. The sense, NcoI containing, primer (TGG CCC ATG GGT GAT CCC ATT AAG ATA CTA ATG) corresponds to bases 168-191 of the E3BP cDNA. Antisense primer (CTT GAA TTC CTA GGC AAG TCG GAT AGG ATT CT) corresponds to bases 1492-1514 of the E3BP cDNA and contains an EcoRI restriction site. Resulting cDNA of approximately 1.3 kilobases was digested with NcoI/EcoRI and ligated between NcoI/EcoRI site of bacterial expression vector pET-28a (Novagen) (plasmid PDHE3BP28a). pPDHE3BP28a was digested with DraIII/SphI to obtain an E3BP expression cassette. This DNA was blunt-ended with T4 DNA polymerase and ligated into pPDHE2 plasmid cut with BamHI and blunt-ended with T4 DNA polymerase (plasmid PDHE2/E3BP). Resulting plasmids were transformed into competent TG-1 cells. Transformants were selected on TY agar containing chloramphenicol. Plasmids carrying both cDNAs were identified by PCR. Construction of the bacterial expression vector for E1 component of PDC (plasmid PDHE1) has been described elsewhere (16).
Expression of Recombinant Polypeptides in Escherichia coliCompetent BL21(DE2) cells obtained from Novagen were
transformed with pPDHE2 or pPDHE2/E3BP plasmids. Transformants were
selected on TY agar plates containing 35 µg/ml chloramphenicol.
Several colonies displaying antibiotic resistance were tested for their ability to express recombinant polypeptides and were used to prepare glycerol stocks. For expression experiments, 10 µl of glycerol stock
(pPDHE1 was transformed in BL21(DE3)pLys cells) was inoculated into 1 liter of M9ZB medium supplemented with 35 µg/ml chloramphenicol or
100 µg/ml ampicillin for pPDHE2/E3BP and pPDHE1 plasmids,
respectively. The cells were allowed to grow at 37 °C until the
A600 reached 0.6-0.7. At this point, the
cultures were shifted to room temperature. After temperature
equilibration, cultures were supplemented with 50 µg/ml thiamin for
pPDHE1 or 0.2 mM lipoic acid for pPDHE2/E3BP and induced
with 0.4 mM
isopropyl--D-thiogalacto-pyranoside. Cells were cultured
for another 20-24 h at room temperature and then harvested.
Harvested
cells were resuspended in 75 ml of buffer A (50 mM
potassium phosphate, pH 7.4, 0.5 mM EDTA, 20 mM
-mercaptoethanol, 100 mM KCl) supplemented with 0.5%
Triton X-100 plus a mixture of protease inhibitors (0.1 mM
phenylmethylsulfonyl fluoride, 10 µg/ml pepstatin A, 20 µg/ml
leupeptin, 0.2 mM benzamidine, 1% aprotinin). Cells were
sonicated five times for 15 s on ice with 1-min intervals.
Extracts were clarified by centrifugation at 40,000 × g for 20 min. Clear extracts containing E1 and
either E2 or E2·E3BP subcomplexes
were combined and recombinant proteins were precipitated with 12%
(w/v) polyethylene glycol (PEG, Sigma)-8000 at 4 °C for 20 min. This
12% PEG pellet was resuspended in 30 ml of buffer A containing 0.1%
Triton X-100 and applied on a Sepharose 4B column (bed volume, 500 ml)
equilibrated with the same buffer. Column fractions (5 ml) were assayed
by PDC activity and Western blot analysis. Pooled fractions containing
human PDC components were centrifuged at 150,000 × g
for 10-12 h. Precipitated PDC was resuspended in a small volume of
buffer A containing 0.5% Triton X-100 and protease inhibitors and
stored at
70 °C.
Activity of PDC was assayed by the reduction of NAD+ to NADH as described previously (17). Five units of porcine dihydrolipoamide dehydrogenase (Sigma) and 0.5 mM pyruvate were used in the assay. SDS-PAGE analysis was performed according to Laemmli (18). Procedure for Northern analysis was as described elsewhere (19). Human heart poly(A)+ RNA was obtained from CLONTECH. Protein was determined according to Lowry et al. (20) with bovine serum albumin as a standard.
A search of the
expressed sequence tag data base of the National Center for
Biotechnology Information revealed four clones that encoded protein
sequences similar to those reported previously (7, 8, 10) for the amino
terminus and an internal region of bovine E3BP. Two of
these clones (H58032 and T77385) were sequenced and found identical to
one another except that T77385 is 370 base pairs shorter at its 5 end.
Clone H58032 has an open reading frame of 1503 base pairs that encodes
a protein of 501 amino acid residues with a calculated molecular weight of 54,085 (Fig. 1). Evidence that this
cDNA encodes E3BP is provided by a comparison of the
deduced protein sequence with the previously published (7, 8, 10)
amino-terminal sequences of bovine heart and kidney E3BP as
well as internal sequences obtained with Arg C proteolytic fragments of
these proteins (Table I). Although no
previous data for the human protein are available in the literature, the almost perfect match (45 of 50 residues) with the bovine protein indicates that the cDNA obtained in this study encodes human
E3BP. Based on the amino-terminal sequence of 22 amino
acids published previously for bovine heart E3BP (Table I),
the mature form of E3BP is assumed to start after a
53-amino acid presequence that has numerous Arg, Ser, and Leu residues
arranged in an order consistent with other mitochondrial targeting
sequences (21). The assignment of where the mature protein begins is
provisional, however, because the mature native human protein has not
been subjected to direct amino acid sequencing and there is
disagreement in the literature as to whether the amino terminus of the
bovine protein is free or blocked (7, 8). If it is assumed that mature
E3BP of human starts at the same residue found for the
bovine heart protein, the human mature protein has 448 amino acids with
a calculated molecular weight of 48,040, in reasonably good agreement
with the molecular weight of 50,000 estimated by SDS-PAGE (4-6). The sequence that corresponds to the previously determined sequences of the
protease Arg C-derived fragments of bovine heart and kidney E3BP follows an Arg-130 in the human protein. Cleavage of
the human protein at this residue would generate polypeptides with calculated molecular weights of 14,033 and 34,025, in good agreement with the estimated molecular weights (15,000 and 35,000) by SDS-PAGE of
the polypeptides produced by protease Arg C cleavage of the bovine
proteins (10, 22).
|
Northern blot analysis of human heart poly(A)+ RNA (CLONTECH) using 32P-labeled cDNA for E3BP as a probe revealed a single hybridizing message of approximately 2.7 kilobases (data not shown), indicating cDNA of clone H58032 is almost complete. The DNA sequence encoding this message is located on chromosome 11 in the region p12-13 (National Center for Biotechnology Information Human Gene Map). It is interesting that the human E2 gene has also been mapped to chromosome 11 but in a region (q21-23) far removed from that of the E3BP gene. Lack of any long stretches of identical nucleotide (not shown) and amino acid sequences (presented and discussed below) in E2 and E3BP also indicates that different genes encode these proteins.
Expression, Reconstitution, and Partial Purification of the Human Pyruvate Dehydrogenase ComplexE3BP is an essential
component of PDC because it positions E3 to accept
electrons from the reduced lipoyl group of the E2 component
(12, 13, 15). Since the only function of E3BP may be to
anchor E3 to the complex and E3BP may not have
an intrinsic catalytic activity of its own, proof that the cDNA
characterized in this study encodes E3BP would best be
provided by reconstitution of the overall reaction catalyzed by PDC.
Without the presence of E3BP as an integral component of
the complex, the enzymatic activity of the complex for its complete
reaction (pyruvate + NAD+ + CoA acetyl-CoA + NADH + CO2) would be expected to be very low for want of high
affinity binding of E3 to the complex. With E3BP present in the complex, E3 should bind and
reconstitute PDC with a high specific activity.
The mammalian PDC consists of three single enzyme subcomplexes capable
of self-assembly: 1) E1 subcomplex (tetramer of two E1 and two E1
subunits); 2)
E2·E3BP subcomplex (consists of 60 copies of
E2 and 12 copies of E3BP); and 3)
E3 subcomplex (dimer of two identical subunits). This
modular organization allows for independent expression of the different
single enzyme subcomplexes, which can then be mixed for reconstitution
of the multienzyme complex. We took advantage of this fact, generating
the catalytically active E1 subcomplex by co-expression of
cDNAs for E1
and E1
subunits and
independently generating the catalytically active dihydrolipoamide
acyltransferase or E2·E3BP subcomplex by
co-expression of cDNAs for E2 and E3BP. For
comparison purposes, the dihydrolipoamide acyltransferase component
lacking E3BP was obtained by expression of the cDNA for
E2 alone. The dihydrolipoamide dehydrogenase subcomplex from porcine heart is commercially available in excellent purity, making its expression for the reconstitution studies unnecessary.
Reconstitution of the PDC was achieved by mixing extracts of E. coli in which the E1 subcomplex had been expressed
with extracts of E. coli in which either the
E2·E3BP complex or E2 alone
subcomplex had been expressed. Reaction mixtures containing expressed
E1 alone plus porcine E3, expressed
E2·E3BP alone plus porcine E3, or
expressed E2 alone plus porcine E3 did not show
PDC activity above that due to the native bacterial enzyme activity
present in these extracts (Fig. 2).
However, combining extracts containing E1 and
E2·E3BP in a reaction cocktail containing
porcine E3 reconstituted the mammalian PDC as evidenced by
a more than 3-fold increase in activity above that of the native
bacterial enzyme activity (Fig. 2). Combining extracts containing
E1 with extracts containing E2 but lacking
E3BP in a reaction mixture containing the same amount of
porcine E3 resulted in no increase in PDC activity, showing
that the cDNA encodes for E3BP and confirming its
requirement for the reaction catalyzed by the complex.
To establish whether E3BP physically associates with other
components of the PDC, extracts of E. coli containing the
different recombinant proteins were combined to reconstitute
subcomplexes with and without E3BP. The
E1·E2·E3BP subcomplex, produced
by combining extracts containing E1 and
E2·E3BP, and the
E1·E2 subcomplex, produced by combining
extracts containing E1 and E2, were then partially purified by PEG precipitation and gel filtration on Sepharose
4B. Both subcomplexes eluted from the gel filtration column as large
aggregates right after the void volume, indicative of assembly of
macromolecular complexes from the components (data not shown). The
isolated E1·E2·E3BP subcomplex
added to a reaction mixture containing porcine E3 exhibited
a specific enzyme activity of 12 units/mg of protein. This value is
lower but still quite respectable relative to previously reported
specific activities (18-23 units/mg of protein) for native bovine
enzymes (23), particularly since the subcomplex was only partially
purified. Incomplete lipoylation of the E2 component could also be a
contributing factor. In contrast to the high activity of the
E1·E2·E3BP subcomplex, the
isolated E1·E2 subcomplex had a much lower
specific activity (<0.8 units/mg of protein) when assayed under the
same conditions. SDS-PAGE analysis of the respective preparations
revealed that the catalytically active subcomplex contains predominant
protein bands corresponding to the molecular weights of
E1, E1
, E2, and
E3BP (Fig. 3). In contrast,
the much less active subcomplex obtained by expression of the
E2 cDNA without E3BP cDNA lacks the
protein corresponding to E3BP. These findings provide
evidence that the E3BP is physically associated with
E2 and confirm that it is required for PDC activity.
Comparison of the Amino Acid Sequences of Human E3BP and Human E2
Analysis of the human E3BP
amino acid sequence indicates a structure quite similar to human
E2, i.e. a segmented domain structure joined by
linker regions rich in alanine and proline residues (Fig.
4A). Overall sequence identity
is 43% for the two proteins. One difference is that E3BP
has one putative lipoyl-bearing domain in its amino-terminal region
rather than the two tandemly repeated lipoyl-bearing domains present in
E2. Residues 4-83 of E3BP exhibit 44%
sequence identity with the first lipoyl-bearing domain and 46%
identity with the second lipoyl-bearing domain of E2 (Fig. 4B). Thus, the alignment can be made with either
lipoyl-bearing domain, but because of the slightly greater degree of
homology, the E3BP sequence has been aligned in Fig. 4 with
the second lipoyl-bearing domain of E2. That
E3BP should have a lipoyllysine residue in its
amino-terminal region is expected from several previous studies demonstrating acetylation of this protein upon incubation of PDC with
either radioactive pyruvate or acetyl-CoA (4, 6, 8, 11). Lipoyllysine
residues are typically found to occur 43-44 residues from the amino
terminus of mammalian E2 proteins (26). A lysine residue is
located at assigned position 44 in the deduced sequence of mature
E3BP, and this lysine occurs in a typical consensus sequence for a lipoyl-attachment site (branched chain
residue-Glu-Ser/Thr-Asp-Lys-Ala-Xaa-branched chain residue) (26) and
is, therefore, highly likely to be the site of lipoylation. The
existence of a glycine at position 55 is of significance because of
recent evidence that a glycine residue situated 11 residues to the
carboxyl-terminal side of a lipoylation site is required for
lipoylation of E2 by lipoyltransferases (27).
Human E3BP also has an internal region that exhibits
considerable sequence identity with the putative
E1/E3-binding domain (28) of human
E2 (Fig. 4B), a relationship pointed out
previously by Sanderson et al. (10) in their analysis of the
first 23 amino-terminal residues of an Arg C proteolytic fragment of
bovine E3BP. Thirty-eight percent sequence identity occurs
in this domain (E2 residues 271 to 303 versus
E3BP residues 131 to 163) where E1 is believed
to bind to E2 and where E3 most likely binds to
E3BP. Further support that E3 binds in this
region to E3BP is provided by considerable sequence
identity with E3-binding domains of the E2
components of human branched chain -ketoacid dehydrogenase complex
(30%) (26, 29) and E. coli
-ketoglutarate dehydrogenase
complex (50%) (30). It is noteworthy that the protease Arg C cleavage site of E3BP (7, 10) is located adjacent to the
amino-terminal side of the E1/E3 binding-domain
(Fig. 4B). This is consistent with the previous suggestion
of Sanderson et al. (10) that exposure of this domain with
disruption of its structure most likely accounts for loss of
E3 binding upon cleavage of E3BP at the Arg C
proteolytic site (7).
The carboxyl-terminal regions of E3BP and E2 are remarkably similar (Fig. 4B). This is the inner core domain of E2 where acetyl transferase activity as well as the self-assembly sequences are located. These proteins share 50% sequence identity in this region, beginning with residues 221 and 334 of E3BP and E2, respectively, and extending to their carboxyl termini. The carboxyl-terminal region of human E3BP lacks the highly conserved histidine residue found in the active site motif of all dihydrolipoamide acyltransferases. This histidine residue is invariably present in the highly conserved sequence of DHRXXDG in proteins with acyltransferase activity (31-34). The corresponding sequence in human E3BP is DSRXXDD (residues 221 to 226; Fig. 4B). Thus, the active site histidine is replaced by a serine residue, and the carboxyl-terminal aspartate of the motif is replaced by a glycine residue, making it unlikely that this protein possesses acetyltransferase activity.
Based on a comparison of the primary sequences E2 of different species, two domains (residues 428 to 440 and 462 to 483) within the carboxyl-terminal region of E2 have been proposed involved in homo- and hetero-subunit interactions (28). Although not established that these domains are interaction sites between E2 subunits, it is interesting to note that 9 of 13 residues in the first domain and 17 of 22 residues in the second are identical in E3BP and E2.The hinge regions separating the three major segments of E3BP and E2 show the least sequence identity (15%) of any region of these proteins (Fig. 4B). The hinge of E2 that connects lipoyl-bearing domain 2 with the peripheral subunit binding domain is rich in both proline and alanine residues. The corresponding region of E3BP is rich in proline but contains no alanine residues. Collagenase specifically cleaves within this hinge of E2 because it contains an amino acid sequence (PAGP) recognized by this protease. Absence of this sequence from the hinge of E3BP prevents cleavage by collagenase, accounting for the specific removal of lipoyl-bearing domains from E2 by collagenase digestion of the PDC (35). The hinge separating the E1/E3-binding domain from the inner core is rich in proline and alanine residues in both proteins, but this domain is 28 residues longer in E3BP and consequently has many more proline and alanine residues in this region than E2.
Lipoyl-bearing proteins have invariably been found to migrate on SDS-PAGE gels with aberrantly high molecular weights (29, 36), a feature attributed to an extended region of helical structure due to domains rich in alanine and proline residues (37). It is not apparent why this disparity between SDS-PAGE-estimated molecular weight (50,000) and amino acid content-calculated molecular weight (48,040) is not observed for E3BP.
Comparison of the Amino Acid Sequences of Human E3BP and Yeast E3BPThe human E3BP shows significant sequence similarity to the yeast E3BP only in amino-terminal, lipoyl-bearing domains of these proteins (49% identity between residues 4-83) (Fig. 4B). In this region, both proteins have the signature motif of a lipoyl-bearing domain as discussed above in the sequence comparison with E2. Beyond residue 83, however, very little sequence similarity is found for these proteins that have been assigned a similar function in their respective complexes. The yeast protein lacks a well defined proline-alanine-rich domain that would readily identify a hinge region. On the other hand, a short sequence exists between residues 166 and 175 that has significant homology to sequences within the E1/E3-binding domains of human E2 and E3BP. Conservation of this sequence in both the human and the yeast forms of E3BP suggests this motif may be particularly important for E3 binding. However, it is known that E3-deficient PDC from beef heart cannot be reconstituted with yeast E3 (10), a finding that presumably reflects poor binding of yeast E3 to the mammalian E3BP as a consequence of marked differences in sequence and therefore structure of the E3-binding domains of yeast and mammalian E3-binding proteins.
The carboxyl-terminal region of yeast E3BP shows no sequence similarity with either human E2 or E3BP (Fig. 4B). The active site motif DHRXXDG characteristic of dihydrolipoamide acyltransferases is not present in yeast E3BP, and no evidence has been presented for acyltransferase activity with either the native or recombinant proteins (38, 39).
The remarkable differences between yeast E3BP and mammalian E3BP coupled with the remarkable similarity between mammalian E2 and mammalian E3BP raise interesting questions, not the least of which is whether anchoring E3 to E2 is the only function of mammalian E3BP. We are particularly interested in a possible role for this protein in mediating regulation of the activity of the various isoenzymes of pyruvate dehydrogenase kinase that have been discovered recently (19, 40). The availability of the recombinant E3BP and the capability for reconstitution of PDC completely from recombinant proteins should facilitate future studies of the function of protein X.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) AF001437.
We thank Dr. Byoung J. Song for help with the expression system for E1 and Patricia A. Jenkins for help in preparation of this manuscript.