Biochem. J. (1997) 324
(65–73) (Printed in Great Britain)
Gene expression of mitochondrial 3-hydroxy-3-methylglutaryl-CoA synthase in a poorly ketogenic mammal: effect of starvation during the neonatal period of the piglet
Sean H. ADAMS*, Clarice S. ALHO, Guillermina ASINS, Fausto G. HEGARDT† and Pedro F. MARRERO
Unit of Biochemistry, School of Pharmacy, University of Barcelona, Avda. Diagonal, 643, 08028 Barcelona, Spain
The low ketogenic capacity of pigs correlates with a low activity
of mitochondrial 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA) synthase. To identify the molecular mechanism controlling such
activity, we isolated the pig cDNA encoding this enzyme and
analysed changes in mRNA levels and mitochondrial specific
activity induced during development and starvation. Pig mitochondrial synthase showed a tissue-specific expression pattern.
As with rat and human, the gene is expressed in liver and large
intestine; however, the pig differs in that mRNA was not detected
in testis, kidney or small intestine. During development, pig
mitochondrial HMG-CoA synthase gene expression showed
interesting differences from that in the rat: (1) there was a 2–3
week lag in the postnatal induction; (2) the mRNA levels
remained relatively abundant through the suckling–weaning
transition and at maturity, in contrast with the fall observed in
rats at similar stages of development; and (3) the gene expression
was highly induced by fasting during the suckling, whereas no
such change in mitochondrial HMG-CoA synthase mRNA levels
has been observed in rat. The enzyme activity of mitochondrial
HMG-CoA synthase increased 27-fold during starvation in
piglets, but remained one order of magnitude lower than rats.
These results indicate that post-transcriptional mechanism(s) and/or intrinsic differences in the encoded enzyme are responsible
for the low activity of pig HMG-CoA synthase observed throughout development or after fasting.
Abbreviations used: CPT-I, carnitine palmitoyl transferase I; HMG-CoA, 3-hydroxy-3-methylglutaryl-CoA; NEFA, non-esterified
fatty acids; PCS, pig-specific cytosolic HMG-CoA synthase probe; PEPCK, phosphoenolpyruvate carboxykinase; PMS, pig-specific mitochondrial HMG-CoA synthase
probe; PPAR, peroxisome proliferator-activated receptor; RACE, rapid amplification of cDNA ends.
* Present address: Department of Internal Medicine, University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Boulevard, Dallas,
TX 75235-9135, U.S.A.
† To whom correspondence should be addressed.
The cDNA sequence data for pig mitochondrial HMG-CoA synthase will appear in the GenBank Database under the accession number U90884.
INTRODUCTION
Mitochondrial 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA) synthase is a potential regulatory site in the pathway converting
acetyl-CoA to ketone bodies [1–3]. In rats, hepatic ketogenic
capacity increases rapidly during postnatal development [4] or
fasting [5], when the liver mRNA, protein and activity of
mitochondrial HMG-CoA synthase increase [6–11]. Ketogenesis
is also associated with increased substrate availability and
changes in the activities of enzymes central to b-oxidation.
Perinatal and dietary responses of rat hepatic mitochondrial
HMG-CoA synthase coincide with changes in the expression and
activity of liver carnitine palmitoyl transferase I (CPT-I) [6,12–15], which suggests common regulatory mechanisms for
these genes.
Factors responsible for the postnatal induction and starvation-associated changes in mitochondrial HMG-CoA synthase gene
expression remain to be characterized, but probably involve
changes in the hormonal and fatty acid milieux. Administration
of dexamethasone or dibutyryl-cAMP to adult rats elicits
increased liver mRNA, protein and enzymic activity [7,8,16],
whereas opposite results are observed with insulin [7]. In addition,
the rat mitochondrial HMG-CoA synthase gene contains
elements that mediate its multihormonal regulation and tissue
specificity [17]. A response element for peroxisomal proliferator-activated receptor (PPAR) has been characterized in the 5´
promoter region of the rat mitochondrial HMG-CoA synthase
gene, and is a mediator of the responsiveness of the gene to fatty
acid [18].
An unusually low hepatic ketogenic capacity [19–23] and a
lack of hyperketonaemia [24,25] in piglets contrast with what is
observed in newborns of other species. A major degree of control
is probably exerted by mitochondrial HMG-CoA synthase in
pigs, because (1) low ketogenesis in vitro [19,21–23] is evident
despite a postnatal rise in total CPT activity [26], (2) experimental use of C8:0 (which bypasses the CPT-I system in liver) as a ketogenic substrate fails to induce ketosis [20], and (3) mitochondrial HMG-CoA synthase specific activity and immunodectable protein in 48 h-old unsuckled pigs are negligible [21].
There is evidence that ketogenesis improves during development
in pigs [19,20,27], which might be linked to a change in
mitochondrial HMG-CoA synthase activity.
The impact of transcriptional and post-transcriptional modulation of mitochondrial HMG-CoA synthase activity on ketogenesis in rats is not fully understood. Companion studies
examining this enzyme in non-ketotic species promise to uncover
novel or subtle control mechanisms that might be overlooked in
work with rats alone. To this end, we isolated the cDNA of pig
mitochondrial HMG-CoA synthase, and examined its mRNA
and enzyme activity during development and the suckling–fasting
transition. Expression of the gene was detected only in liver and
large intestine; increasing mRNA levels were associated with age
and fasting. During fasting, piglet mRNA rose (similar to starved
adult rat mRNA), despite low enzyme activity. Our results
suggest that the transcriptional control of mitochondrial HMG-CoA synthase is critical during the neonatal period in pigs, and
that post-transcriptional mechanism(s) and/or intrinsic
differences in the encoded enzyme are also responsible for the
low HMG-CoA synthase activity observed in this species.
EXPERIMENTAL
Reagents
The following reagents were purchased from Sigma (St. Louis,
MO, U.S.A.): acetyl-CoA, acetoacetyl-CoA, dextran sulphate,
EDTA, guanidine thiocyanate, Mops, Tris and Triton X-100.
Other chemicals, including agarose, CsCl, dithiothreitol, formaldehyde, formamide, KH2PO4, 2,5-diphenyloxazole and sucrose
were from Boehringer Mannheim (Mannheim, Germany), Carlo
Erba (Milan, Italy) or Merck (Darmstadt, Germany). Radiochemicals ([1-14C]acetyl-CoA and [a]32P]CTP) were from ICN
(Irvine, CA, U.S.A.).
Animals
Sprague–Dawley rats (120 g) fed ad libitum were used in this
study. To induce starvation, food was withdrawn 24 h before the
beginning of the experiment. Rats were decapitated and their
livers were quickly removed and processed for either isolation of
mitochondria or RNA extraction in the same way as pig liver (see
below).
Two litters of commercial crossbred pigs (Sus scrofa) were
used in studies examining developmental changes in mitochondrial HMG-CoA synthase gene expression and enzyme
activity. Typical factory production techniques involve a weaning
process initiated much earlier (i.e. at 2–3 weeks) and more
acutely than observed in nature [28]. To characterize naturally
occurring developmental and/or diet-related changes in mitochondrial HMG-CoA synthase, and to avoid possible artifacts
introduced by the stress of early weaning, piglets were allowed
access to the sow until 8 weeks of age, with adult feed available
after the fifth week. Tissues were obtained from one piglet per
litter at 24 h after birth and at each week from weeks 1–8, as well
as from two mature (60 kg) pigs. Animals seemed to be in the fed
state when sampled, but time of last suckling/feeding was not
determined. For fasting experiments, tissue or blood was obtained
from additional 14-day-old piglets previously starved for 48 h or
allowed to suckle. Fasting at the age of 2 weeks was chosen
because of previous reports of improved ketogenesis [19,27] with
starvation (see the Discussion section).
Tissues were excised from pentobarbitol-anaesthetized (20–60 mg/kg, intracutaneously) pigs, and immediately immersed in
liquid nitrogen; the animals were then killed by exsanguination.
Water-rinsed regions of small intestine corresponding to duodenum, jejunum and ileum, and large intestine samples of
proximal (caecum) and distal (colon) regions were obtained.
Portions of liver used for isolation of mitochondria (see below) were placed in ice-cold homogenization buffer containing
250 mM sucrose, 0.1 mM EDTA and 5 mM Tris/HCl, pH 7.4.
For analysis of plasma hormones and metabolites in fed and
fasted piglets, blood was obtained by jugular venipuncture,
combined with potassium-EDTA (1.5 mg/ml) and placed on ice
until centrifugation at 5000 g to retrieve plasma. Plasma for
glucagon analysis was treated with 500 k-i.u./ml aprotinin
(Boehringer-Mannheim). All plasma samples were stored at
-80 °C.
Mitochondrial HMG-CoA Synthase Activity
All procedures were performed at 4 °C. Liver tissue was homogenized in 5 vol. of cold homogenization buffer by using three
strokes of a motorized loose-fitting apparatus with Teflon pestle.
The supernatant from a 350 g (10 min) centrifugation of homogenate was centrifuged at 15000 g (15 min) to pellet the mitochondria. After washing and resuspension with homogenizing
buffer, the preparation was centrifuged again at 15000 g, then
the pellet was washed and resuspended in buffer containing
0.4 mM dithiothreitol, 100 mM Tris/HCl and 1.5% (v/v) Triton
X-100, pH 8, at a ratio of 2:1 (v/w of homogenized tissue).
Mitochondria were dialysed overnight in 1000 vol. of buffer
containing 20 mM KH2PO4 and 12 mM EDTA, pH 7 [29], and
samples were stored at -20 °C. Mitochondrial HMG-CoA
synthase activity was measured as the incorporation of [1-14C]acetyl-CoA into HMG-CoA at 30 °C for 5 min [29]. The
reaction was initiated by adding protein (500–2000 µg of piglet
or 20–150 µg of rat) preparation to a reaction mixture (final
volume 200 µl) composed of 100 mM Tris/HCl, 1 mM EDTA,
20 µM acetoacetyl-CoA, 200 µM acetyl-CoA (specific radioactivity between 2000 and 4000 c.p.m./nmol). Protein concentrations for activity assays were determined by the biuret method.
Radioactivity was counted in a cocktail containing 67% toluene,
33% Triton X-100 and 0.56% 2,5-diphenyloxazole (v/v/w).
Plasma hormones and non-esterified fatty acids (NEFA)
Plasma insulin and glucagon were measured by standard radioimmunoassay procedures employing insulin (Cys Bio International, Gif-Sur-Yvette, France) or pancreatic glucagon [30],
human antibodies and standards for human insulin or pig
glucagon respectively. Fatty acid concentration was determined
enzymically with an automated system (Bio Merieux). Raw
insulin and glucagon values (µ-i.u./ml and pg/ml) are reported
as molar equivalents by using an insulin conversion factor of 6
(American Diabetes Foundation) and a factor of 3485 pg/pmol
for glucagon.
Isolation of DNA fragments encoding part of the catalytic site
from mitochondrial and cytosolic pig HMG-CoA synthase
All enzymic manipulations were performed following the manufacturers' (Boehringer-Mannheim, Promega) instructions. The
sequence of rat mitochondrial HMG-CoA synthase cDNA
between positions 197 and 216 and positions 542 and 561 [31] is
highly similar to human mitochondrial HMG-CoA synthase
[32], cytosolic HMG-CoA synthase from human [33], rat [34],
hamster [35] and chicken [36]; and cockroach HMG-CoA
synthase 1 [37] and 2 [38]. PIG1 (
) and PIG2 (
) degenerate primers were thus
designed to represent all possible combinations of the different
sequences in the similar regions (bold in the PIG1 and PIG2
primer sequences). Pig genomic DNA was obtained, of which
1 µg was used as a template in a PCR amplification (95 °C,
1 min; 55 °C, 45 s; 72 °C, 1 min; 35 cycles) with PIG1 and PIG2
primers. The 380 bp PCR product was cloned into Bluescript
(SK+) plasmid using ClaI and XbaI restriction sites introduced at
the 5´ end of PIG1 (ClaI) or PIG2 (XbaI) primers (underlined in
the PIG1 and PIG2 sequences). Seven recombinant plasmids
were obtained, of which six were cut by NcoI, as in rat and
human mitochondrial HMG-CoA synthases, which contain such
a site [31,32], and one remained undigested (human [33] and rat
[34] cytosolic HMG-CoA synthase lack an NcoI restriction site;
not shown). All inserts were digested by RsaI and SacI (not
shown), which act on sites present in both mitochondrial and
cytosolic HMG-CoA synthase cDNA from rat and human
[31–34]. A representative clone of each family was sequenced by
the dideoxynucleotide method [39] with the use of an automated
fluorescence-based system (Applied Biosystems), and plasmids
pPMS (region of pig mitochondrial HMG-CoA synthase cDNA,
positions 196–560) and pPCS (region of pig cytosolic HMG-CoA synthase cDNA) were characterized (see the Results section).
Isolation of pig mitochondrial HMG-CoA synthase full-length cDNA
Specific primers for PCR amplification of the 5´ and 3´ ends of
pig cDNA were designed on the basis of the sequence of pPMS,
thus avoiding the amplification of cytosolic HMG-CoA synthase
cDNA. In the 5´ end amplification [40] a rapid amplification of
cDNA ends (RACE) kit from Clontech (Palo Alto, CA, U.S.A.) was used following the manufacturer's instructions. Briefly, 2 µg
of poly(A)+ mRNA from the liver of a 30 kg pig was the template
in the first-strand cDNA synthesis, using RTP5 (5´-TCTCAGTGCCCACTTCCAGC-3´) pig-specific reverse primer. After the
RNA degradation, an anchor sequence (3´-GGAGACTTCCAAGGTCTTAGCTATCACTTAAGCAC-3´) was ligated to the
3´ end of the cDNA first strand with T4 RNA ligase. The 5´ end
of pig mitochondrial cDNA was subsequently amplified by PCR
with an anchor primer (5´-CTGGTTCGGCCCACCTCTGAAGGTTCCAGAATCGATAG-3´) and an AMP5 primer (
), a pig-specific primer located upstream of RTP5. The PCR product was
cloned by taking advantage of an EcoRI restriction site introduced in the anchor sequence and AMP5 primers (bold in the
respective sequences). The insert of the resulting plasmid
(pSIM 3) was sequenced and used as a probe in Northern blot
analyses, and in the isolation of the pig mitochondrial HMG-CoA synthase full-length cDNA.
In the 3´ cDNA amplification, a modification of the method
reported by Frohman et al. [41] was used. A sample (109 plaque-forming units) of a pig cDNA library (lgt10, from Clontech) was
boiled and used as a template in a PCR reaction performed with
PIP3 (5´-AAGCTTCCTTCATGCCGCCTCTCCTG-3´) pig-specific primer and the left arm of lgt10 reverse primer
(5´-TTATGAGTATTTCTTCCAGGG-3´). After the first
amplification a new PCR was performed with AMP3
(
) pig-specific primer and REVT (5´-GACTCGAGTCGACATCGATTTTTTTTTTTTTTTTT-3´) primers. This nested PCR
yielded a single amplification product of approx. 1.4 kb that
hybridized in Southern blot analysis with pPMS (result not
shown). EcoRI (bold in the AMP3 primer sequence) and SalI
(bold in the REVT primer sequence) restriction sites introduced
at the 5´ end of the primers allowed ligation of the PCR product
into Bluescript, generating the recombinant plasmid pSIM4.
The cloned 5´ (pSIM3) and 3´ (pSIM4) ends were joined by
using the single restriction site NcoI (underlined in AMP5 and
AMP3 pig-specific primers). The insert from pSIM3 was liberated
by using EcoRI–NcoI enzyme digestion and the isolated fragment
was cloned into pSIM 4 to generate pSIM5, which carried the
full-length cDNA.
Sequencing strategy of pig mitochondrial HMG-CoA synthase
Figure 1 shows the sequence strategy used. pPMS and pSIM3 were sequenced by using T3 and T7 Bluescript-specific primers;
pSIM 4 was sequenced by using AMP3 pig-specific primer; and
pSIM 5 was sequenced by using AMP5 and PIP3 (5´-CCTTCATGCCGCCTCTCCTG-3´) pig-specific primers. Two new
plasmids, pSIM 4.1 and pSIM 4.2, were generated from pSIM 4,
by using an ApaI single restriction site present in the pig cDNA,
and sequenced with T3 and T7 Bluescript-specific primers. Using
as a target pSIM4 or pSIM5, we observed that specific human
mitochondrial HMG-CoA synthase primers were able to PCR-amplify DNA of the expected size when used with a pig-specific
or Bluescript-specific primer (results not shown); therefore we
were able to sequence this DNA-product with the human-specific
primers. The human reverse primers I5R (5´-CCATTGTGAGTGGAGAGG-3´) and I8R (5´-ATGCTGCTCGTCCACTCG-3´) were used with the pig AMP3 primer; the human forward
primers I3 (5´-GGAGCTGTGGCTATGCTG-3´), I4 (5´-CATCCAGTGCTACTTGCG-3´), I5 (5´-ACCCTTCACCCTTGACGA-3´), F4 (5´-GGCCTCTCAGGACATGTTCG-3´) and F5
(5´-TTTCGAGTATCCCAGGATGC-3´) were used with a
Bluescript T7 primer. The sequence obtained with this approach
was used to design five new pig-specific primers: F1 (5´-TGCTCATGGAGCTCTTCCAGG-3´), F2 (5´-TCACTCTCGACGATTTACAG-3´), F3 (5´-AAGCTGGAAGACACCTACGC-3´),
R1 (5´-GCTTGCTTCCACTGCTTCTCG-3´) and R2 (5´-GTGTGAAATTCACCTTGTGGTAG-3´); such primers were used
to confirm the pig mitochondrial HMG-CoA synthase cDNA
sequence.
Isolation of a pig HMG-CoA lyase cDNA probe
After studying the HMG-CoA lyase cDNA alignments [42] we
followed a reverse transcriptase–PCR approach to isolate a pig
HMG-CoA lyase probe with human primers HL-F2 (5´-CTGCCTCAGAGCTCTTCACC-3´), HL-F3 (5´-AGATCTCCCTGGGGACACC-3´) and HL-R4 (5´-TGGCCAAGTTTCCTGATGCC-3´), surrounding the catalytic Cys-266. Poly(A)+ RNA of
pig intestine was used as a template for the first cDNA synthesis,
using oligo(dT) as a primer, then an amplification was performed
with HL-F2 and HL-R4 primers (94 °C for 1 min, 94 °C for
1 min, 50 °C for 45 s, 72 °C for 45 s; 35 cycles; 72 °C for 20 min),
yielding a smear of the expected size (447 bp). A second semi-nested PCR performed with HL-F3 and HL-R4 yielded a single
band of the expected size (237 bp). The isolated pig-specific
HMG-CoA lyase fragment was 87.8% similar to human HMG-CoA lyase cDNA (between positions 616 and 813 [42]) and
recognized a constitutive single transcript of approx. 1.5 kb.
Northern blot hybridizations
Total RNA was extracted from frozen tissues by the guanidinium
thiocyanate method [43] and stored at -80 °C. The concentration
of RNA was measured from the A260. Northern blot analysis of
total RNA (10 µg) was performed after 1% (w/v) agarose-gel
electrophoresis in 2.2 M formaldehyde, as described [9]. cDNA
probes were labelled with [a-32P]dCTP to a specific radioactivity
of approx. 9×109 c.p.m./µg. Reported absorbance (A) units
derived from Northern blot analyses have been corrected for b-actin mRNA abundance.
cDNA probes
In tissue distribution experiments, purified inserts of pPMS or
pPCS were used. In developmental and starvation-induction
studies the SIM 5 (419 bp 5´ RACE product) was used as a pig
mitochondrial HMG-CoA synthase-specific probe. In some
experiments a pig-specific HMG-CoA lyase (237 bp reverse
transcriptase–PCR product) and human b-actin [9] probes were
also used. For the quantification of mRNA, specific 365 bp
probes corresponding to the human (human probe), pig (pig
probe) or rat (rat probe) mitochondrial HMG-CoA synthase
(positions +148 to +512 in each case, taking the translation
start codon as +1) were generated individually by PCR, with
PIG1 and PIG2 primers and human cDNA [32], pig cDNA (this
study) or rat cDNA [31] as templates.
RESULTS
Isolation of mitochondrial HMG-CoA synthase full-length cDNA
from pig
The isolation of DNA fragments encoding part of the pig
catalytic domain from mitochondrial and cytosolic HMG-CoA
synthases was approached by PCR from genomic DNA. PIG1
and PIG2 degenerate primers, which represent sequences of all
HMG-CoA synthases isolated so far [31–38], were used in the
amplification. Recombinant plasmids derived from the resulting
PCR products showed two different DNA inserts identified as
mitochondrial or cytosolic HMG-CoA synthase by DNA sequence alignment with published HMG-CoA synthase cDNA
species. PMS (380 bp insert; NcoI-sensitive) showed greater
DNA similarity to rat and human mitochondrial HMG-CoA
synthases (89.1% and 90.8% respectively) than to rat or human
cytosolic HMG-CoA synthases (67.1% and 67.5% respectively),
and was thus identified as pig mitochondrial HMG-CoA synthase
DNA. Conversely, PCS (380 bp insert; NcoI-insensitive) showed
a greater similarity to rat and human cytosolic HMG-CoA
synthases (86.6% and 91.6% respectively) than to rat or human
mitochondrial HMG-CoA synthases (65.0% and 65.9% respectively), and was therefore identified as pig cytosolic HMG-CoA synthase DNA.
To obtain the nucleotide sequence of a complete pig mitochondrial HMG-CoA synthase cDNA, a RACE approach was
used. Plasmid pSIM5, carrying the full-length cDNA of pig
mitochondrial HMG-CoA synthase, was created by pSIM 3 (5´
end) and pSIM 4 (3´ end) cDNA fusion by using a single NcoI
restriction site present either in the reverse primer used in the 5´
RACE or in the forward primer used in the 3´ RACE (see the
Experimental section for details). Figure 2 shows the nucleotide sequence of this cDNA (SIM 5) encoding pig mitochondrial
HMG-CoA synthase, together with the predicted amino acid
sequence of the protein. The pig mitochondrial HMG-CoA
synthase full-length cDNA spans 1744 nucleotides (see Figure 2) and encodes a protein of 508 amino acid residues with a predicted
molecular mass of 56715 Da. Amino acid sequence alignment of
pig cDNA with rat and human mitochondrial synthase shows an
identity of 82.1% and 83.2% respectively. The N-terminal end
of the primary translation product shows a sequence of 37
residues that is in agreement with the general composition given
for leader peptides that translocate cytosolic synthesized proteins
into mitochondria [44]. A putative polyadenylation signal was
detected in the 3´ untranslated region, at position +1613
(AATAAA).
Tissue specificity of pig mitochondrial HMG-CoA synthase
A Northern blot experiment with PMS or PCS as probes
concurred with the identity of mitochondrial and cytosolic clones
previously assigned by homology analysis. The PMS probe
hybridized with an approx. 1.8 kb transcript (Figure 3). This transcript size agrees with the full-length pig mitochondrial
HMG-CoA synthase cDNA cloned by RACE (see Figure 2) and is similar to the rat and human mitochondrial HMG-CoA
synthase transcripts [31,32]. When PCS was used as a probe,
Northern blot analysis of pig mRNA showed a constitutive
expression of an approx. 3.6 kb transcript (Figure 3), similar in size to the rat cytosolic HMG-CoA synthase [31]. Figure 3 also shows that a significant expression of mitochondrial HMG-CoA
synthase was detected only in liver, caecum and colon of pigs.
Thus subsequent analyses of developmental or fasting-induced
changes in expression focused on these tissues. This tissue
specificity of expression did not seem to be related to age because
an identical pattern was observed in piglets aged 1, 2 or 3 weeks
(results not shown).
mRNA and activity levels of mitochondrial HMG-CoA synthase
during development and starvation
Figure 4(A) shows Northern blot analyses of pig liver RNA during development. When RNA samples were processed with
SIM3 as a probe, large variations in mitochondrial HMG-CoA
synthase mRNA levels were observed. In contrast, when the
same samples were analysed with a pig-specific probe for HMG-CoA lyase, the mRNA level was essentially unchanged during
suckling. Despite this variability in mitochondrial HMG-CoA
synthase transcript it was clear that (1) mRNA levels remained
minimal for at least the first week of life, and (2) transcript
amounts began to rise by the second and the third weeks and
remained high until maturity, despite initiation of weaning. An
increasing trend in enzyme specific activity did not become
apparent until the fifth week (Figure 4B), after which weekly mean activities (0.03–0.19 m-units/mg of protein) were 3–19-fold
that observed in the first month of life (approx. 0.01 m-units/mg
of protein). The adult pig liver HMG-CoA synthase activity
remained more than one order of magnitude smaller than adult
rat liver activity (approx. 6 m-units/mg of protein, measured in
the same assay; see also [21]). Expression of mitochondrial
HMG-CoA synthase was also investigated in large intestine: in
caecum, mRNA was detected throughout the suckling period but
decreased in mature animals; in colon, mRNA was present at all
stages of development (results not shown).
A 48 h fast in 2-week-old piglets elicited a marked rise in
hepatic mitochondrial HMG-CoA synthase mRNA levels without affecting the b-actin mRNA levels (Figure 5). When the same samples were analysed by using the pig cytosolic HMG-CoA
synthase-specific probe PCS no changes were detected in the
mRNA levels (results not shown). The low mRNA content in the
suckled fed pig hinders the quantification of the degree of
induction, but it seemed to correlate with a 27-fold rise in liver-specific activity of the mitochondrial HMG-CoA synthase after
fasting (0.463±0.102 m-units/mg from fasted animals, compared with 0.017±0.001 m-units/mg from suckling animals).
The activity detected in the starved piglets (approx. 0.5 m-units/mg) was higher than in adult pigs (see Figure 4B) but still one order of magnitude lower than activity detected in fed
(approx. 6 m-units/mg) or starved (approx. 25 m-units/mg; see
also [11,45]) rats. Fasting resulted in marked differences in
transcript levels between the proximal and distal regions of the
large intestine. Mitochondrial HMG-CoA synthase mRNA
(Figure 5) increased 4.8-fold in the caecum (P = 0.07) but did not change in the colon (P < 0.1).
In suckling rats no fasting-induced changes in mRNA levels
have been observed [46]; thus the significant fasting-induced rise
in mitochondrial HMG-CoA synthase mRNA in the liver of
piglets (Figure 5) resembled that observed in fasted adult rats [7,8]. In an attempt to compare transcript amount in the two
models, RNA from liver of fasted animals was probed with pig,
rat and human mitochondrial HMG-CoA synthase probe (human probe; PCR-produced with PIG1 and PIG2 primers by
using the human cDNA clone of mitochondrial HMG-CoA
synthase as a target [32]; this 365 bp human probe was 89.9% or
91.0% similar to rat or pig cDNA). The use of homologous pig
or rat probes shows, as expected, that the detected level of
mitochondrial HMG-CoA synthase mRNA in starved piglets or
rats depends on the probe used (see Figure 6). Despite the differences in enzyme activity discussed previously (approx. 50-fold), mRNA levels quantified with the heterologous human
probe were only slightly (35%) higher in fasted rat than in fasted
pig (Figure 6, top panel; 0.393±0.0204 and 0.291±0.0126 absorbance units respectively; P < 0.05).
Circulating hormones and fatty acids could affect the expression of mitochondrial HMG-CoA synthase [7,8,17,18]. To
gain an insight into their roles in the fasting-induced rise in
mitochondrial HMG-CoA expression (Figure 5), insulin, glucagon and NEFA concentrations were determined in suckled
(n = 7) and fasted (n = 5) piglets. Fasting did not significantly
affect plasma insulin (38±3 pM compared with 48±7 pM) or
glucagon (131±13 pM compared with 136±19 pM) levels. However, fasting plasma NEFA levels were almost double those
observed in suckling animals (602±75 µM compared with
358±17 µM; see also [27,47]).
DISCUSSION
Isolation of pig mitochondrial HMG-CoA synthase cDNA
The strategy for isolating pig mitochondrial HMG-CoA synthase
was based on the assumption that the evolutionary distance
between the different HMG-CoA synthases would not be high.
We previously observed that the different cytosolic and mitochondrial HMG-CoA synthases contained extensively conserved
sequences (results not shown) that would presumably be
conserved in the pig. Accordingly we chose fully conserved
amino acid sequences from published HMG-CoA synthases
(between positions 197 and 216 and positions 542 and 561 of rat
mitochondrial HMG-CoA synthase cDNA [31]) and the corresponding degenerate oligonucleotides were used in PCR experiments. This approach was successful in isolating two DNA
products (365 bp each) encompassing two different fragments of
pig HMG-CoA synthase. By sequence comparison between these
DNAs and those of the rat and human, we deduced that the
amplified fragments corresponded to pig mitochondrial and
cytosolic HMG-CoA synthase. The sequence of the putative
cDNA for mitochondrial HMG-CoA synthase was used to
design specific primers that allowed the isolation of a full-length
cDNA (Figure 2) by RACE.
The identification of the cDNA as mitochondrial HMG-CoA
synthase was based on several criteria: (1) the amino acid
sequence predicted from this cDNA shares extensive conservation
with mitochondrial HMG-CoA synthases from rat (82.1%) and
human (83.2%); (2) the predicted amino acid sequence from
the cDNA contains a region with a high level of identity to that
of the active site of other HMG-CoA synthases, whether cytosolic
or mitochondrial; (3) the N-terminus contains a putative leader
peptide to target the protein into mitochondria [44]; and (4) in
the avian enzyme the cysteine residue involved in the formation
of the acyl-S-enzyme intermediate has been identified [48]; this is
mapped as Cys-166 in the cDNA sequence of mitochondrial
HMG-CoA synthase from rat [31] and human [32], and it is
conserved in the pig cDNA sequence (Figure 2).
Tissue specificity of pig mitochondrial HMG-CoA synthase
As in rats [9,10,31,49] and humans [32], mitochondrial HMG-CoA synthase expression is tissue-specific in pigs. All three
species show mitochondrial HMG-CoA synthase mRNA expression in the large intestine and liver of adult animals (see Figure 3A); but pig differs from human and rat in that the kidney, ovary
and testis lack detectable expression of the gene (Figure 3A, and results not shown). During suckling the expression pattern was
also different in rats and pigs: no expression was found in small
intestine of piglets (results not shown). Regulation of expression
also seemed to differ across tissues because fasting induced a
significant rise in transcript in liver and caecum but not in colon
(Figure 5).
The relevance of intestinal mitochondrial HMG-CoA synthase
expression/activity for whole-animal ketone body metabolism in
newborns is not well defined. Rat neonatal small intestine
possesses the capacity for ketogenesis, as judged by the presence
of ketogenic enzymes and active ketogenesis in vitro [9,50,51].
Therefore the lack of expression in small intestine and the
abundance of mRNA in large intestine of piglets might affect
metabolism. The relative mass (percentage of body mass) of the
small intestine (4%) is similar to that in the liver (3%) in
suckling pigs, whereas the large intestine (approx. 1%) is smaller
(S. H. Adams, P. F. Marrero and F. G. Hegardt, unpublished
work). Nevertheless the physiological effect of intestinal expression of the gene remains unknown.
mRNA and activity levels of mitochondrial HMG-CoA synthase
during development
Hepatic rates of ketogenesis and b-oxidation rise within hours of
the onset of suckling in rats [6,52–54] but remain low in newborn
pigs [19,21–23,27]. The increased ketogenesis in newborn rats is
due to transcriptional and post-transcriptional factors, including
the induction of the expression of mitochondrial HMG-CoA
synthase and CPT-I gene [9,10,14,15] and the depressed inhibition
of these enzymes by succinyl-CoA and malonyl-CoA respectively
[4].
Studies reporting insignificant ketogenesis [19,21–23], a relatively low rate of b-oxidation [21] and significant acetogenesis
[22,23] in newborn pig liver have made it clear that perinatal
changes in lipid metabolism in pigs differ markedly from those
described in other species. Although ketogenesis is minimal in
newborn pigs, studies performed in vitro [19] and in vivo [20,27]
have shown a significant developmental increase in hepatic
ketogenic capacity. Changes in activity of mitochondrial HMG-CoA synthase might be associated with development of ketogenic
capacity in pigs, as suggested previously [20,21]. This study is the
first to address the molecular phenomena regulating mitochondrial HMG-CoA synthase activity in pigs throughout development or after fasting.
The abundance of liver mRNA encoding mitochondrial HMG-CoA synthase rose with age (Figure 4A), indicating that the synthase gene is developmentally controlled. Substantial mRNA
was observed only after a postnatal lag of 1–2 weeks (Figure 4A), a pattern markedly different from the rapid postnatal rise
reported for suckling rats [9,10]. A major species difference was
also noted during the suckling–weaning transition and maturity:
in rat the gene expression and activity of mitochondrial HMG-CoA synthase fall steadily with consumption of adult food
[6,9,10,55] but mRNA levels for the enzyme were persistently
abundant in the liver of older pigs, even after the introduction of
adult food during the final 3 weeks of suckling (Figure 4A). These results illustrate that major differences in regulation of the
mitochondrial HMG-CoA synthase gene can occur across species
during postnatal development.
mRNA and activity levels of mitochondrial HMG-CoA synthase
during starvation
During suckling a number of physiological conditions mimic the
fasted state of adults. In rats, for example, circulating glucagon
and NEFA levels are raised, the insulin level is low and
gluconeogenesis coupled to b-oxidation/ketogenesis operates
concomitantly with adaptational changes in enzyme activity and
transcription [i.e. high for cytosolic phosphoenolpyruvate
carboxykinase (PEPCK), mitochondrial HMG-CoA synthase
and CPT-I] [4]. Thus the fasting of previously suckled rat pups
does not alter the insulin-to-glucagon ratio [56] or lead to
increased levels of mitochondrial HMG-CoA synthase mRNA
[46]. In contrast, starvation in piglets aged 2 weeks elicited a large
stimulation of expression of the liver mitochondrial HMG-CoA
synthase gene (Figure 5), increased ketogenesis from long-chain fatty acids in vitro [19] and increased circulating ketone bodies
[27]. The latter event was not observed in fasted 3-day-old piglet
[27], suggesting that fasting-induced ketogenesis (perhaps
coupled with enhanced hepatic mitochondrial HMG-CoA
synthase activity) occurs only later in development. These results
clearly emphasize that multiple levels of transcriptional control
impact the pig synthase gene, including developmental and
fasting-induced regulators of expression.
The mechanism(s) by which mitochondrial HMG-CoA
synthase expression is induced during the suckling–fasting transition in piglet is not clear. The promoter of the rat mitochondrial
HMG-CoA synthase gene is activated in vitro by the presence of
long-chain fatty acids [17,18], an event mediated by PPAR [18].
Plasma NEFA nearly doubled with 48 h of fasting [27,47]. It is
plausible that increased blood fatty acids have a role in the
fasting-induced stimulation of expression observed in piglets
(Figure 5), and the hypothesis that NEFA/PPAR trans-activate the pig gene is the subject of current investigation in this
laboratory.
The induction of mitochondrial HMG-CoA synthase mRNA
by starvation (Figure 5) was reflected in a 27-fold increase in specific activity. Despite the large amount of transcript, the
increased enzyme activity in pig remained one order of magnitude
lower than that in adult or suckling rat. This suggests that post-transcriptional mechanisms and/or kinetic differences in the
mitochondrial HMG-CoA synthase enzyme also control activity
in pigs. As a first attempt at assessing this hypothesis we
performed Northern blot experiments of mRNA from rat or pig
liver, using as a probe a fragment of the cDNA from human
mitochondrial HMG-CoA synthase. This probe, whose sequence
shared similar global homology with the cDNA from rat or pig
(90%), detected a large increase in transcript after fasting in both
species, consistent with results from the use of homologous
probes. The slight species differences in fasted mRNA levels
detected with the human probe do not seem to explain the 50-fold difference in enzyme activity in the two models (see the
Results section). If the heterologous human probe hybridizes to
HMG-CoA synthase mRNA with similar affinity for rat or pig
samples, then the results shown in Figure 6 might also indicate that the very low activity of the enzyme in pig is not due to low
mRNA abundance but rather to intrinsic differences in enzyme
kinetics and/or differences in post-translational modifications.
The expression of recombinant proteins in Escherichia coli might
elucidate which mechanism best explains the low enzymic activity.
The unusual rise in insulin level in the first 1–2 days of suckling
[25,47,57] and the fasting-induced increase in ketogenesis [27] by
2 weeks of age raised the possibility that changes in mitochondrial
HMG-CoA synthase expression in fasted piglets might be due to
a decrease in the insulin-to-glucagon ratio. However, the results
illustrate that stimulation of expression of the mitochondrial
HMG-CoA synthase gene in fasted 2-week-old piglets does not
seem to be related to changes in circulating insulin or glucagon
concentrations. No significant changes in plasma concentrations
of hormones were observed, and the liver content of cytosolic
PEPCK mRNA as determined by Northern blot hybridization
(results not shown) was not different (50±15 compared with
52±7 absorbance units for fed and fasting animals respectively).
Abundance of cytosolic PEPCK mRNA is a sensitive marker of
the prevailing insulin and glucagon activities in a number of
systems [4,58].
In summary, the patterns of expression and activity of mitochondrial HMG-CoA synthase in piglets lend support to the idea
that this enzyme could modulate the ketone body profile of
suckling and fasting, and the age-related increase of ketogenic
capacity described previously [19,20,27]. Despite increases in
expression by fasting or suckling, the specific activity of the
enzyme in mitochondria remained relatively low, suggesting that
post-transcriptional mechanism(s) and/or kinetic differences in
the mitochondrial HMG-CoA synthase enzyme also control the
expression of the HMG-CoA synthase gene in pigs.
We thank Dr. Carlos Buesa for the isolation strategy of pig HMG-CoA lyase; Dr. Diego
Haro for discussion and critical examination of the manuscript; Dr. Roser Casamitjana
and Dr. Elena Casals of the Hospital Clínic i Provincial de Barcelona for the hormones
and NEFA determinations; and Robin Rycroft for editorial assistance. This work was
supported by grant PB95-0012 from Dirección General de Investigación Científica y
Técnica, Spain. S.H.A. was the recipient of a National Science Foundation/North
Atlantic Treaty Organization Postdoctoral Fellowship. C.S.A. is the recipient of an
Instituto de Cooperación Iberoamericana Doctoral Fellowship.
REFERENCES
1 Williamson, D. H., Bates, M. W. and Krebs, H. A. (1968) Biochem. J. 108, 353–361
Medline 1st Citation
2 Baird, G. D., Hibbitt, K. G. and Lee, J. (1970) Biochem. J. 117, 703–709
Medline 1st Citation
3 Dashti, N. and Ontko, J. A. (1979) Biochem. Med. 22, 365–374
1st Citation
4 Girard, J., Ferré, P., Pégorier, J. P. and Duée, P. H. (1992) Physiol. Rev. 72,
507–562
Medline 1st Citation 2nd 3rd 4th
5 McGarry, J. D. and Foster, D. W. (1980) Annu. Rev. Biochem. 49, 395–420
Medline 1st Citation
6 Bailey, E. and Lockwood, E. A. (1973) Enzyme 15, 239–253
Medline 1st Citation 2nd 3rd 4th
7 Casals, N., Roca, N., Guerrero, M., Gil-Gómez, G., Ayté, J., Ciudad, C. and Hegardt,
F. G. (1992) Biochem. J. 283, 261–264
Medline 1st Citation 2nd 3rd 4th 5th
8 Serra, D., Casals, N., Asins, G., Royo, T., Ciudad, C. J. and Hegardt, F. G. (1993)
Arch. Biochem. Biophys. 307, 40–45
1st Citation 2nd 3rd 4th
9 Serra, D., Asins, G. and Hegardt, F. G. (1993) Arch. Biochem. Biophys. 301,
445–448
Medline 1st Citation 2nd 3rd 4th 5th 6th 7th 8th
10 Thumelin, S., Forestier, M., Girard, J. and Pégorier, J. P. (1993) Biochem. J. 292,
493–496
Medline 1st Citation 2nd 3rd 4th 5th
11 Quant, P. (1990) Biochem. Soc. Trans. 18, 994–995
Medline 1st Citation 2nd
12 Augenfeld, J. and Fritz, I. B. (1970) Can. J. Biochem. 48, 288–294
Medline 1st Citation
13 Foster, P. C. and Bailey, E. (1976) Biochem. J. 154, 49–56
Medline 1st Citation
14 Thumelin, S., Esser, V., Charvy, D., Kolodziej, M., Zammit, V. A., McGarry, J. D.,
Girard, J. and Pégorier, J. P. (1994) Biochem. J. 300, 583–587
Medline 1st Citation 2nd
15 Asins, G., Serra, D., Arias, G. and Hegardt, F. G. (1995) Biochem. J. 306, 379–384
Medline 1st Citation 2nd
16 Quant, P. A., Tubbs, P. K. and Brand, M. D. (1990) Eur. J. Biochem. 187, 169–174
Medline 1st Citation
17 Gil-Gómez, G., Ayté, J. and Hegardt, F. G. (1993) Eur. J. Biochem. 213, 773–779
1st Citation 2nd 3rd
18 Rodríguez, J. C., Gil-Gómez, G., Hegardt, F. G. and Haro, D. (1994) J. Biol. Chem.
269, 18767–18772
1st Citation 2nd 3rd 4th
19 Pégorier, J. P., Duée, P. H., Girard, J. and Peret, J. (1983) Biochem. J. 212, 93–97
1st Citation 2nd 3rd 4th 5th 6th 7th 8th 9th
20 Adams, S. H. and Odle, J. (1993) Am. J. Physiol. 265, 761–765
1st Citation 2nd 3rd 4th 5th 6th
21 Duée, P. H., Pégorier, J. P., Quant, P. A., Herbin, C., Kohl, C. and Girard, J. (1994)
Biochem. J. 298, 207–212
1st Citation 2nd 3rd 4th 5th 6th 7th 8th
22 Lin, X., Adams, S. H. and Odle, J. (1996) Biochem. J. 318, 235–240
Medline Biochem. J. Biochem. J. 1st Citation 2nd 3rd 4th 5th
23 Odle, J., Lin, X., Van-Kempen, T. A. T. G., Drackley, J. K. and Adams, S. H. (1995)
J. Nutr. 125, 2541–2549
1st Citation 2nd 3rd 4th 5th
24 Bengtsson, G., Gentz, J., Hakkarainen, J., Hellström, R. and Persson, B. (1967)
J. Nutr. 97, 311–315
1st Citation
25 Pégorier, J. P., Duée, P. H., Assan, R., Peret, J. and Girard, J. (1981) J. Dev.
Physiol. 3, 203–217
1st Citation 2nd
26 Bieber, L. L., Markwell, M. A. K., Blair, M. and Helmrath, T. A. (1973) Biochim.
Biophys. Acta 326, 145–154
Medline 1st Citation
27 Gentz, J., Bengtsson, G., Hakkarainen, J., Hellström, R. and Persson, B. (1970)
Am. J. Physiol. 218, 662–668
1st Citation 2nd 3rd 4th 5th 6th 7th 8th 9th 10th
28 Newberry, R. C. and Wood-Gush, D. G. M. (1985) Behaviour 95, 11–25
1st Citation
29 Clinkenbeard, K. D., Reed, W. D., Mooney, R. A. and Lane, M. D. (1975) J. Biol.
Chem. 250, 3108–3116
Medline 1st Citation 2nd
30 Unger, R. H., Eisentrant, A. M., McCalls, M. S. and Madison, L. (1961) J. Clin.
Invest. 40, 1280–1289
1st Citation
31 Ayté, J., Gil-Gómez, G., Haro, D., Marrero, P. F. and Hegardt, F. G. (1990) Proc. Natl.
Acad. Sci. U.S.A. 87, 5682–5684
1st Citation 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th
32 Mascaró, C., Buesa, C., Ortiz, J. A., Haro, D. and Hegardt, F. G. (1995)
Arch. Biochem. Biophys. 317, 385–390
1st Citation 2nd 3rd 4th 5th 6th 7th 8th 9th 10th
33 Russ, A. P., Ruzicka, V., Maerz, W., Appelhans, H. and Gross, W. (1992) Biochim.
Biophys. Acta 1132, 329–331
Medline 1st Citation 2nd 3rd 4th
34 Ayté, J., Gil-Gómez, G. and Hegardt, F. G. (1990) Nucleic Acids Res. 18, 3642
1st Citation 2nd 3rd 4th
35 Gil, G., Goldstein, J. L., Slaughter, C. A. and Brown, M. S. (1986) J. Biol. Chem.
261, 3710–3716
Medline 1st Citation 2nd
36 Kattar-Cooley, P. A., Wang, H.-H. L., Mende-Muller, L. M. and Miziorko, H. M. (1990)
Arch. Biochem. Biophys. 283, 523–529
1st Citation 2nd
37 Martínez-González, J., Buesa, C., Piulachs, M. D., Bellés, X. and Hegardt, F. G. (1993)
Eur. J. Biochem. 217, 691–699
1st Citation 2nd
38 Buesa, C., Martínez-González, J., Casals, N., Haro, D., Piulachs, M. D., Bellés, X. and
Hegardt, F. G. (1994) J. Biol. Chem. 269, 11707–11713
Medline 1st Citation 2nd
39 Sanger, F., Nicklen, S. and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. U.S.A. 74,
5463–5467
Medline 1st Citation
40 Edwards, J. B., Delort, J. and Mallet, J. (1991) Nucleic Acids Res. 19, 5227-5232
Medline 1st Citation
41 Frohman, M. A., Dush, M. K. and Martin, G. R. (1988) Proc. Natl. Acad. Sci. U.S.A.
85, 8998–9002
Medline 1st Citation
42 Mitchell, G. A., Robert, M.-F., Hruz, P. W., Wang, S., Fontaine, G., Behnke, C.,
Mende-Muller, L. M., Schappert, K., Lee, C., Gibson, K. M. and Miziorko, H. M.
(1993) J. Biol. Chem. 268, 4376–4381
Medline 1st Citation 2nd
43 Chirgwin, J. J., Przybyla, A. E., MacDonald, R. J. and Rutter, W. J. (1979)
Biochemistry 18, 5294–5299
1st Citation
44 Von Heijne, G. (1986) EMBO J. 5, 1335–1342
Medline 1st Citation 2nd
45 Guzmán, M., Bijleveld, C. and Geelen, M. J. (1995) Biochem. J. 311, 853–860
Biochem. J. Biochem. J. 1st Citation
46 Arias, G., Matas, R., Asins, G., Hegardt, F. G. and Serra, D. (1995) Biochem. Soc.
Trans. 23, 493S
Medline 1st Citation 2nd
47 Swiatek, K. R., Kipnis, D. M., Mason, G., Chao, K.-L. and Cornblath, M. (1968)
Am. J. Physiol. 214, 400–405
1st Citation 2nd 3rd
48 Misra, I., Narasimhan, C. and Miziorko, H. M. (1993) J. Biol. Chem. 268,
12129–12135
Medline 1st Citation
49 Royo, T., Pedragosa, M. J., Ayté, J., Gil-Gómez, G., Vilaró, S. and Hegardt, F. G.
(1993) J. Lipid Res. 34, 867–874
Medline 1st Citation
50 Békési, A. and Williamson, D. H. (1990) Biol. Neonate 58, 160–165
1st Citation
51 Hahn, P. and Taller, M. (1987) Life Sci. 41, 1525–1528
Medline 1st Citation
52 Ferré, P., Satabin, P., Decaux, J.-F., Escrivá, F. and Girard, J. (1983) Biochem. J.
214, 937–942
1st Citation
53 Ferré, P., Pégorier, J. P., Williamson, D. H. and Girard, J. (1978) Biochem. J. 176,
759–765
1st Citation
54 Escrivá, F., Ferré, P., Robin, D., Robin, P., Decaux, J.-F. and Girard, J. (1986) Eur. J.
Biochem. 156, 603–607
1st Citation
55 Decaux, J.-F., Robin, D., Robin, P., Ferré, P. and Girard, J. (1988) FEBS Lett. 232,
156–158
Medline 1st Citation
56 Beaudry, M.-A., Chiasson, J.-L. and Exton, J. H. (1977) Am. J. Physiol. 233,
E175–E180
Medline 1st Citation
57 Lepine, A. J., Boyd, R. D. and Welch, J. (1989) Domest. Anim. Endocrinol. 6,
231–241
Medline 1st Citation
58 Granner, D. and Pilkis, S. (1990) J. Biol. Chem. 265, 10173–10176
Medline 1st Citation
Received
8 October 1996/29 November 1996; accepted 6 January 1997
The Biochemical Society, London © 1997
Full Editorial Board
RSS feeds
Table of contents by email
My BJ toolbox
