Tài liệu Báo cáo khoa học: Complex transcriptional and translational regulation of iPLA2c resulting in multiple gene products containing dual competing sites for mitochondrial or peroxisomal localization docx

Complex transcriptional and translational regulation of iPLA
2
c
resulting in multiple gene products containing dual competing sites
for mitochondrial or peroxisomal localization
David J. Mancuso
1,2
, Christopher M. Jenkins
1,2
, Harold F. Sims
1,2
, Joshua M. Cohen
1,2
, Jingyue Yang
1,2
and Richard W. Gross
1,2,3,4
1
Division of Bioorganic Chemistry and Molecular Pharmacology, and Departments of
2
Medicine,
3
Chemistry and
4
Molecular Biology
and Pharmacology, Washington University School of Medicine, St. Louis, MO, USA
Membrane-associated calcium-independent phospholipase
A
2
c (iP LA
2

c) contains four potential in-frame methionine
start s ites (Mancuso, D.J. Je nkins, C .M. & Gross, R.W.
(2000) J. Biol. Chem. 275, 9937–9945), but the mechanisms
regulating the types, amount and subcellular localization of
iPLA
2
c in cells are i ncompletely understood. We now:
(a) demonstrate the dramatic transcriptional repression of
mRNA synthesis encoding iPLA
2
c by a nucleotide sequence
nested in the coding sequence itself; (b) localize the site of
transcriptional repression to the most 5¢ sequence encoding
the iPLA
2
c holoprotein; (c) identify the presence of nuclear
protein c onstituents w hich bind to the repressor region by gel
shift analysis; (d) demonstrate the translational regulation of
distinct iPLA
2
c isoforms; (e) identify multiple novel exons,
promoters, and alternative splice variants o f human iPLA
2
c;
(f) document the presence of dual-competing subcellular
localization s ignals in discrete isoforms of iPLA
2
c;and
(g) demonstrate t he functional integrity of an N-terminal
mitochondrial localization signal by fluorescen ce imagi ng

and the presence of iPLA
2
c in the mitochondrial compart-
ment of rat myocardium. The intricacy of the r egulatory
mechanisms of iPLA
2
c biosynthesis in rat myocardium is
underscored by the identification of seven distinct protein
products that utilize multiple mechanisms (transcription,
translation and proteolysis) to produce discrete iPLA
2
c
polypeptides containing either single or dual subcellular
localization s ignals. T his unanticipated complex i nterplay
between peroxisomes and mitochondria mediated by com-
petition for uptake of the nascent iPLA
2
c polypeptide
identifies a new level of phospholipase-mediated m etabolic
regulation. Because uncoupling protein function is regulated
by free fatty acids in mitochondria, these results suggest that
iPLA
2
c processing contributes to integrating respiration a nd
thermogenesis in mitochondria.
Keywords: phospholipase; mitochondria; p eroxisomes; tran-
scription; translation.
Phospholipases A
2
(PLA

2
s) play critical roles in cellular
growth, lipid homeostasis and lipid second messenger
generation by catalyzing the esterolytic cleavage of the
sn-2 fatty acid o f glycerophospholipids [1–5]. The resultant
fatty acids and lysolipids are potent lipid mediators of signal
transduction and a lter the biophysical properties o f the
membrane bilayer, collectively contributing t o the critic al
roles that phospholipases play in cellular adaptation,
proliferation and signaling. PLA
2
s constitute a d iverse
family of enzymes, which include the intracellular phos-
pholipase families, cytosolic PLA
2
s(cPLA
2
) and calcium-
independent PLA
2
s(iPLA
2
) as well as the secretory PLA
2
s
(sPLA
2
).
More than a decade ago, we identified multiple types of
kinetically distinguishable iPLA

2
activities in the cytosolic,
microsomal and mitochondrial fractions from multiple
species of mammalian m yocardium [6–10]. Utilizing the
synergistic power of HPLC in conjunction with MS of
intact phospholipids, initial insights into b oth the canine
and human mitochondrial lipidomes were made [8,11]. Both
human and canine cardiac mitochondria possess a high
plasmalogen content, and plasmalogens are readily hydo-
lyzed by heart mitochondrial phospholipases [7,8]. Both
cytosolic and membrane-associated iPLA
2
activities are
inhibited by the nucleophilic serine-reactive mechanism-
based inhibitor (E)-6-(bromome thylene)-3-(1 -naphthale-
nyl)-2H-tetrahydropyran-2-one (BEL) [12–14]. Recent
studies have shown that BEL has potent effects on
mitochondrial bioenergetics [15] and that fatty acids are a
Correspondence t o R. W. Gross, Washington University School of
Medicine, D ivision of Bioorganic Chemistry and Molecu lar Phar-
macology, 660 South Euclid Avenue, Campus Box 8020, St. Louis,
MO 63110, USA. Fax: +1 314 362 1402; Tel: +1 314 362 2690;
E-mail: rgros
Abbreviations: BEL, (E)-6-(bromomethylene)-3-(1-naphthalenyl)-
2H-tetrahydropyran-2-one; cPLA
2
, cytosolic phospholipase A
2
;ECL,
enhanced chemoluminescence; EMSA, electrophoretic mobility shift

analyses; EST, expressed sequ ence tag; GAPD H, glyceraldehye-
3-phosphate dehydrogenase; iPLA
2
, calcium-independent phosphol-
ipase A
2
; iPLA
2
c, membrane a ssociated calcium-independent phos-
pholipase A
2
(AF263613); MOI, multiplicity o f infection; PLA
2
,
phospholipase A
2
; Sf9, Spodoptera frugiperda cells; sPLA
2
, secretory
phospholipase A
2
; T AMRA, 6-carboxytetramethylrhodamine;
UCP, uncoupling protein.
(Received 25 August 2004, revised 10 October 2004,
accepted 13 October 2004)
Eur. J. Biochem. 271, 4709–4724 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04435.x
rate-determining factor in uncoupling protein (UCP) activ-
ity [16]. Thus, t he role of mitochondrial i PLA
2
activities in

regulating mitochondrial function is just now beginning to
be understood. Mor eover, both fatty acids and lys olipids
alter the physical properties of cell membranes, interact with
specific receptors, and modulate the electrophysiologic
function of m any transmembrane ion channels including
K
+
and Ca
2+
channels in many cells and subcellular
contexts [17–20].
In early studies, we purified canine myocardial cytosolic
iPLA
2
activity (iPLA
2
b) t o homogeneity [21] identifying a
high specific activity, proteolytically activated form of the
gene whose identity w as substantiated by i ts covalent
radiolabeling w ith (E)-6-(
3
H)(bromomethylene)-3-(1-napht-
halenyl)-2H-tetrahydropyran-2-one (radiolabeled BEL)
[12]. However, despite our intense efforts at solubilization
and purification, the membrane-associated iPLA
2
activities
we identified in multip le membrane c ompartments were
resistant to our attempts at their purification. In the
postgenome e ra it became apparent that multiple different

gene products contributed to the many kinetically diverse
activities of membrane-associated iPLA
2
sinmyocardium
possessing distinct molecular masses and substrate selecti-
vities that resided in multiple discrete s ubcellular loci [22–27].
Recently, w e charac terized the genomic organization and
mRNA sequence of a novel iPLA
2
(now termed iPLA
2
c,
GenBank accession number AF263613) located on the long
arm of c hromosome 7 at 118 c
M
[26]. Like other me mbers
of the i PLA
2
family–iPLA
2
a (patatin, found in potato
tubers) [28] and iPLA
2
b [23] – iPLA
2
c contains a consensus
site for nucleotide binding and a lipase consensus motif in its
C-terminal half [26]. Although the intracellular localization
and activity of iPLA
2

b is complex and dynamically
regulated by multiple d ifferent cellular perturbations inclu-
ding ATP concentration [7], calcium-activated calmodulin
[29,30], a nd proteolysis [ 31,32], the biochemical mechanisms
regulating iPLA
2
c in intact tissues are not known with
certainty. For examp le, iPLA
2
c is not activated, stabilized
or bound to ATP under any conditions we have examined,
nor does it associate with calmodulin or possess a discern-
able calmodulin-binding consensus sequence [26]. Like
iPLA
2
b,iPLA
2
c is completely inhibited by low micromolar
concentrations (1–5 l
M
) of the mechanism-based inhibitor
BEL [26].
Previously, we demonstrated th at iPLA
2
c is synthesized
from a 3 .5 kb mRNA containing a putative 2.4 kb co ding
region which was most prominent in heart tissue. The
5¢-region of the 2.4 k b coding sequence of iPLA
2
c contains

four in-frame ATG start sites which can potentially encode
88, 77, 74 and 63 k Da polypeptides [ 26]. However, in i nitial
studies in baculoviral and in vitro rabbit reticulocyte l ysate
systems, we unexpectedly observed that constructs contain-
ing the full-length 2.4 kb sequence encoding the predicted
88 kDa polypeptide resulted instead in the expression of
only t wo protein bands of 77 and 63 kDa [26]. M oreover,
the initial characterization of iPLA
2
c in nonrecombinant
cells demonstrated that hepatic iPLA
2
c was most highly
enriched in the peroxisomal compartment as a 63 kDa
polypeptide [27]. These results raised the intriguing
possibility that iPLA
2
c biosynthesis was transcriptionally
and/or translationally regulated b y as y et unidentified
mechanisms.
To begin to iden tify t he potential modes o f t he regulation
of iPLA
2
c synthesis at the transcriptional and post-
transcriptional levels, an d to i dentify specific mechanisms
modulating iPLA
2
c expression and processing in different
cell types, we examined multiple iPLA
2

c constructs in
different cellular contexts and in intact rat myocardium.
Herein, we demonstrate that iPLA
2
c synthesis is transcrip-
tionally regulated by a transcriptional repressor domain
nested in the 5¢-coding region and translationally regulated
through the differential usage o f downstream A UG start
sites. Moreover, this study identifies an N-terminal mito-
chondrial localization signal a nd demonstrates its functional
integrity by fluorescence colocalization assays. Importantly,
the presence of multiple high molecular m ass iPLA
2
c
isoforms in mitochondria from wild-type rat myocardium
was d emonstrated. T his c omplex interplay o f t ranscrip-
tional and translational, as well as proteolytic, sculpting o f
iPLA
2
c results in a diverse repertoire of biologic products ,
which likely provides the chemical foundations necessary
for iPLA
2
c to fulfill i ts multiple d istinct functional r oles in
mammalian tissues.
Experimental procedures
Materials
[
32
P]dCTP[aP] (6000 C iÆmmol

)1
) and enhanced chemolu-
minescence (ECL) detection reagents were purchased from
Amersham Pharmacia Biotech (Piscataway, NJ, USA). A
human heart cDNA library was purchased from Stratagene
(La Jolla, CA, USA). For PCR, a Perkin-Elmer Thermo-
cycler was employed, and all PCR reagents were purchased
from Applied Biosystems (Foster City, CA, USA). The
Luciferase Assay system and TnT Quick coupled Tran-
scription/Translation system were obtained f rom P romega
(Madison, WI, USA). CV1 cells were generously provided
by D. Kelly (Washington University Medical School).
Vectors pcDNA1.1, pEF1/myc-His and pcDNA 3.1/myc-
His/lacZ were purchased from Invitrogen (Carlsbad, CA,
USA). Vectors pEGFP-N3 and pDsRed-mito were pur-
chased from BD-Biosciences (Palo Alto, CA, USA). Culture
media, CellFECTIN and LipofectAMINE reagents for
transfection, baculovirus vectors and competent DH110Bac
Escherichia coli were purchased from Invitrogen and used
according to the manufacturer’s protocol. QIAfilter plasmid
kits and QIAquick Gel Extraction kits were obtained from
Qiagen (Valencia, C A, USA). Keyhole limpet hemocyanin
was obtained from Pierce (Rockford, IL, USA). BEL was
obtained from Calbiochem (San Diego, CA, USA). Most
other r eagents w ere obtained from Sigma (St. Louis, MO,
USA).
Expression of truncated iPLA
2
c
Constructs encoding the 74- and 63 kDa polypeptides were

prepared as previously described for construction of the f ull-
length iPLA
2
c construct encoding the 88 kDa polypeptide
used for baculoviral expression. In brief, the 74 kDa sense
primer M533 (5¢-TCAAGTCGACATGATTTCACGTTT
AGC-3¢) and the 63 kDa sense primer M530 (5¢-GT
AAGTCGACAATGTCTCAACAAAA GG-3¢)wereeach
paired with reverse primer M458 (5¢-GCATAGCATGCT
4710 D. J. Mancuso et al. (Eur. J. Biochem. 271) Ó FEBS 2004
CACAATTTTGAAAAGAATGGAAGTCC-3¢)forPCR
of  2.0 and 1.7 kb products, r espectively, from the full-
length iPLA
2
c pFASTBac1 construct for cloning via SalI/
SphI sites in to vector pFASTBac1 (Invitrogen). Subsequent
preparation of bacmids, CellFECTIN-mediated transfec-
tion of Spodoptera frugiperda (Sf9) cells to produce virus,
and the Neutral Red agar overlay method for viral plaque
titering were performed utilizing the Bac-to-Bac Baculovirus
Expression System (Invitrogen) according t o the manufac-
turer’s instructions. Sf9 cells were grown and infected for
preparation of recombinant protein extracts as previously
described [26]. In brief, Sf9 cells were cultured in 100-mL
flasks equipped with a magnetic spinner containing supple-
mented Grace’s m edia [26]. S f9 cells at a concentration o f
1 · 10
6
cellsÆmL
)1

were prepared in 50 mL of growth med ia
and incubated at 27 °C for 1 h prior to infection with either
wild-type v irus or recombinant virus containing human
iPLA
2
c cDNA. After 48 h, cells were pelleted by centrifu-
gation, resuspended in ice-cold NaCl/P
i
and repelleted. The
supernatant was decanted and the cell pellet was resus-
pended in 5 mL of ice-cold homogenization buffer (25 m
M
imidazole, pH 8.0, 1 m
M
EGTA, 1 m
M
dithiothreitol,
0.34
M
sucrose, 20 l
M
trans-epoxysuccinyl-
L
-leucylamido-
(4-guanidino) butane and 2 lgÆmL
)1
leupeptin). Cells were
lysed by s onication (20 · 1 s bursts utilizing a Vibra-cell
sonicator at 30% output) and centrifuged at 100 000 g for
1hat4°C. The supernatant was saved (cytosol) and the

membrane pellet was washed by resuspending with a Teflon
homogenizer in 5 mL of homogenization buffer followed by
a brief sonication step (10 · 1 s bursts) before recentrifu-
gation at 100 000 g for 1 h at 4 °C. A fter removal of the
supernatant, the m embrane pellet was resuspended in 1 mL
of homogenization buffer u sing a Teflon homogenizer and
then sonicated (5 · 1 s bursts) to prepare a membrane
fraction.
PLA
2
enzymatic assay and immunoblot analysis
Calcium-independent PLA
2
activity was measured by
quantitating the release of r adiolabeled f atty ac id from
various radiolabeled phospholipid substrates in the presence
of membrane fractions from Sf9 c ells infected with wild-type
or recombinant human iPLA
2
c baculovirus a s previously
described [26]. Protein from baculoviral or reticulocyte
lysate samples was separated by SDS/PAGE [33], trans-
ferred to Immobilon-P membranes by electroelution,
probed with anti-iPLA
2
c Ig and visualized using ECL as
described previously [26].
Northern blot analysis
Total RNA from Sf9 cells was isolated according to the
protocol for RNeasy (Qiagen). In brief, sample was placed

in tissue lysis buffer containing guanine isothiocarbonate
and disrupted by 20–40 s of pulse homogenation with a
rotor stator homogenizer. Total RNA was then recovered
from a c leared lysate after several washes on an RNeasy
mini spin co lumn and elution with RNase-free water.
Recovery of RNA was determined spectrophotometrically
at 260 n m. RNA (2 lg) was fractionated on a 1.25%
agarose Latitude RNA midi gel (BioWhittaker, Walkers-
ville, ME, USA), blotted onto a nylon membrane,
cross-linked b y e xposure to a UV light source for 1 .5 min
and then baked at 85 °C f or 60 min. After prehybridization
in ExpressHyb hybridizatio n buffer (BD Biosciences) for
30 min, the b lot was hybridized 1 h at 68 °C with radio-
labeled iPLA
2
c probe prepared as previously described [26]
in hybridization buffer and then washed with 2· NaCl/Cit
containing 0.1% (w/v) SDS twice for 30 min each, followed
by two washes w ith 0.1· NaCl/Cit containing 0.1% (w/v)
SDSfor40mineachat50°C, as described in the
manufacturer’s instructions. H ybridized sequences were
identified by autoradiography for 16 h.
RNA stability assay
Spinner flasks (100 mL) were infected with equivalent
volumes of each truncated viral iPLA
2
c construct [multi-
plicity of infection (MOI) ¼ 1] and 48 h later, actino-
mycin D was added to a concentration of 10 lgÆmL
)1

.At0,
15, 30, 60, 120 and 2 40 min following actinomycin D
addition, 2-mL aliquots were removed, centrifuged to
collect pellets and quick-frozen in liquid N
2
.RNAwas
then prepared following the RNeasy (Qiagen) protocol.
RNA samples (2 lg) were fractionated on a latitude RNA
midi-gel for northern analysis as described above.
Quantitative PCR
RNA was prepared from Sf9 cell pellets following the
RNeasy protocol supplemented with on-column RNase-
free DNase treatment to remove baculoviral DNA as
described b y t he manufacturer. Completeness of removal of
baculoviral DNA was monitored by including control
samples spiked with p lasmid DNA (either cell pellets from
uninfected Sf9 cells or water blanks). Quantitative PCR of
DNase-treated control samples routinely did not generate
detectable signal. For analysis of actinomycin D -treated test
samples,  0.2–1 lg of the tot al RNA was r everse tran-
scribed using MultiScribe reverse transcriptase in a TaqMan
Gold RT-PCR kit (Applied Biosystems) by incubation for
10 min at 25 °C followed by 30 m in at 48 °C and a final
step of 5 m in at 95 °C a nd 20 ng of cDNA was used per
reaction in quantitative P CR. Specific iPLA
2
c primer pairs
and probe were designed using
PRIMER EXPRESS
software

from PE Bi osystems. Forward and reverse primers, respect-
ively (5¢-AGCTCTTTGATTACATTTGTGGTGTAA-3¢
and 5¢-CACATTCATCCAAGGGCATATG-3¢)wereused
for amplification of an  100 nu cleotide product fl anking
the boundary between exons 5 and 6 o f the iPLA
2
c gene. A
30-mer hybridization probe (5¢-CCCAACATGAAAGC
TAATATGGCACCTGTG-3¢) was designed to anneal
between the PCR primers, at the exon 5/6 boundary,
5¢-labeled with reporter d ye 6-FAM and 3 ¢-labeled
with quenching dye, 6-carboxytetramethylrhodamine
(TAMRA). PCRs were carried out using TaqMan PCR
reagents (Applied Biosystems) as recommended by the
manufacturer. Each PCR amplification was performed in
triplicate, using t he following conditions: 2 min at 50 °C
and 10 m in at 95 °C, followed by a total of 40 two-
temperature cycles (15 s at 95 °C and 1 min at 60 °C). For
the generation o f s tandard curves, s erial dilutions of a
cDNA sample were used and mRNA levels were compared
for various time points after correction using concurrent
Ó FEBS 2004 Regulation of iPLA
2
c biosynthesis (Eur. J. Biochem. 271) 4711
glyceraldehye-3-phosphate dehydrogenase (GAPDH) mes-
sage amplification with GAPDH primers and probe as an
internal standard. Results were plotted as relative mRNA
level vs. time (hours) and the slopes of exponential
trendlines for each construct were compared.
Luciferase assay

PCR primers in Table 1 were used to amplify segments
containing 124 nucleotides of sequence upstream of the
iPLA
2
c 63 kDa start site. All 3 ¢ PCR primers in Table 1
were designed to generate identical Kozak (GCCACC)
sequences [34,35] upstream of the ATG start. In each case,
the sequence around the ATG start is ÔGCCAX
CATGÕ (where ÔXÕ is a ÔCÕ nucleotide in all constructs
except 83 which contains an ÔAÕ nucleotide). In each case,
PCR products were cloned into HindIII/NcoIrestriction
sites within the polylinker region of pGL3-Promoter vector
(pGL3P). Also, because o f the presence of a naturally
occurring NcoI site within the 83 construct, an AflIII
restriction site was utilized at the 3¢-end of this construct
(instead of NcoI) to generate a compatible cohesive end for
cloning into the NcoI restriction site of pGL3-Promoter
vector (pGL3P). Transient t ransfection of CV1 cells with
each of the inhibitory constructs was performed using
LipofectAMINE Plus (Invitrogen). For each transfection,
1–2 lg of luciferase r eporter plasmid was cotransfected with
100 n g of pcDNA 3.1/myc-His/lacZ vector and b-galac-
tosidase activity was measured utilizing the b-galactosidase
enzyme assay system (Promega) for normalization of
results. Background measurements were unifo rmly low
and cell survival was indistinguishable in all transfections
performed. The cells were harvested 24 h later and luciferase
activity was assayed using the luciferase assay syst em
(Promega) following the manufacturer’s protocol. Relative
luminescence values were measured in a Beckman Scintil-

lation counter with a wide-open window.
Subcellular fractionation of rat heart
Subcellular fractionation of rat heart by differential centri-
fugation w as performed essentially as described previously
for r at liver [27]. In brief, rat heart was minced on ice and
then homogenized in 3 vol. (w/v) of ice-cold homogeniza-
tion buffer [0.25
M
sucrose, 5 m
M
Mops,pH7.4,1m
M
EDTA and 0.1% (v/v) ethanol, 0.2 m
M
dithiothreitol
containing protease inhibitors (0.2 m
M
phenylmethylsulfo-
nyl fluoride, 1 lgÆmL
)1
leupeptin, 1 lgÆmL
)1
aprotinin and
15 lgÆmL
)1
phosphoramidon)] using a Potter-Elvehjem
homogenizer at 1000 r.p.m. with 8–10 strokes. The homo-
genate was first centrifuged at 100 g for 1 0 min to remove
cellular debris and then at 1000 g to obtain a nuclear pellet
(nuclear fraction) and a supernatant fraction. The 1000 g

supernatant fraction was further centrifuged at 3000 g for
20 min to collect a heavy mito chondrial pellet ( heavy
mitochondrial fraction). The 3000 g supernatant was then
centrifuged at 23 500 g for 20 m in to collect the light
mitochondrial fraction pellet 2 3 5 00 g (light mitochondrial
fraction). The 23 500 g supernatant was then centrifuged at
70 000 g for 20 m in to collect a second light mitochondrial
pellet (70 000 g light mitochondrial fraction). Utilizing the
above subcellular fractionation technique, the majority of
mitochondria were pr esent in t he 3000 an d 23 500 g pellets,
whereas t he large majority of peroxiso mal marker PMP70
was present in the supernatant.
Promoter analysis
iPLA
2
c seque nces were examined for the presence of
putative promoter e lements utilizing t he inte rnet-based
program
TFSEARCH
(http://150.82.196.184/research/db/
TFSEARCH.html). Promoter activity of iPLA
2
c sequences
was analyzed by cloning sequences upstream of the
luciferase reporter gene i n promoterless vector p GL3-
Enhancer (Promega). The following primers were u tilized
to amplify PCR products containing iPLA
2
c sequence:
P1, 5¢-TCAAGGTACCATGATTTCCTGAAGG-3¢;P2,

5¢-CTGAAGATCTAGCCTTTACTTTCA-3¢;P3,5¢-GC
TAGGTACCAATACAGTAATATATG-3¢;P4,5¢-TGC
TAGATCTCCACCCACTCA-3¢;P5,5¢-TTATGGTACC
TGAAAGGGAATAGCGGC-3¢;P6,5¢-GGCTGGTAC
CCTTGCGCTCCGTC-3¢;P7,5¢-GGAGAGATCTGCG
GGAAGCCGCGACAGA-3¢;p8,5¢-TTCCAGAT CTG
CAGAGATAAGCCTCCC-3¢;p9,5¢-GCGTGAGATCT
CTGGTTGGTTGC-3¢;P10,5¢-ACCAGGTACCGCA
CAGCACGCCCC-3¢; and P11, 5¢-GTCCGGTACCGG
AAGGCAAAACGA-3¢. Primers P1 and P2 were utilized
to amplify a 584-nucleotide product containing sequence
Table 1. PCR primer pairs for lo calization of transcriptional regulatory elements in the 5¢-coding region o f iPLA
2
c. Underlined residues indicate the
locations of HindIII (AAGCTT), NcoI (CCATGG), o r Af l III (ACATGT) restriction sites utilized for cloning PCR p roducts.
Construct PCR primer pairs 5¢-to3¢-sequence
88 88F
GTTGAAGCTTGTGTCTATTAATCTGACTGTA
88R TAGACCATGGTGGCTTATCCTCCAGTAATGC
87 87F GTGTAAGCTTGAAGCAGAGAAGCAAGCAACTG
87R ACTGCCATGGTGGCCTTCACTTTTGGTCCATTTAC
85 85F TGGAAAGCTTGCCACATCAGTCTACAAAG
85R TGCTCCATGGTGGCATCCCAATATGTAAACCA
83 83F GAACCAAGCTTGAAGCACATTCTTGCAGTAAGCA
83R CAAAACATGTTGGCTACGGGACATACAAATGTTCA
80 80F GTTGAAGCTTTTTGAAACTTAGCACTTCTGC
80R ATTCCATGGTGGCTGAAATCATTTCATTTTGATTGCC
74 74F TCAAAAGCTTATGATTTCACGTTTAGCTC
74R CTTTCCATGGTGGCTGTCACTATATTTTTTCA
4712 D. J. Mancuso et al. (Eur. J. Biochem. 271) Ó FEBS 2004

upstream from iPLA
2
c exon 1. For construct I, primers P3
and P4 were utilized to amplify a 584 nucleotide product
containing sequence upstream from iPLA
2
c exon 2. PCR
products for constructs II–IX were prepared as follows:
primers P5 and P4 were paired to amplify a 390-nucleotide
product for construct I I; primers P6 a nd P4 were utilized to
amplify a 197-nucleotide product for construct III; primers
P5 and P8 were employed to amplify a 215-nucleotide
product for construct IV; primers P3 and P8 were utilize d to
amplify a 2 16-nucleotide p roduct for construct V; primers
P3 and P7 were paired to amplify a 409-nucleotide product
for c onstruct VI; primers P5 and P9 were utilized to amplify
a 131-nucleotide p roduct for construct VII; primers P10 and
P9 were paired to amplify a 106-nucleotide product for
construct VIII; and primers P11 and P7 were employed to
amplify a 1 55-nucleotide p roduct for construct IX. PCR
products were subsequently cloned via KpnI/BglII restric-
tion sites into the promoterless vector pGL3-Enhancer
(Promega) and then utilized for LipofectAMINE Plus-
mediated transien t transfection of CV1 ce lls followed 24 h
later by analysis of luciferase activity utilizing the Lucife rase
Assay System o f Promega. Empty p GL3-Enhancer vector
and the SV40-containing promoter vector pGL3-Promoter
were used as controls. MyoD vector used for c otransfection
of CV1 cells with the pre-exon 1 iPLA
2

c construct was
obtained from M. Chin (Harvard Medical School) [36].
Results were normalized to b-gal resulting fro m cotransfec-
tion with a LacZ vector.
5¢-Rapid amplification of cDNA ends (RACE)
5¢-RACE was performed as p reviously described employin g
human heart M arathon-Ready cDNA (BD Bioscien ces)
and primers AP1 and M460 [26]. PCR products were gel
purified with a QIAquick gel extraction kit, subcloned into
pGEM-T vector (Promega), sequenced and analyzed by
alignment with iPLA
2
c sequences.
Electrophoretic mobility shift analyses
Electrophoretic mobility shift analyses (EMSA) were per-
formed with the Promega gel shift assay s ystem according to
the manufacturer’s specifications by using 2 lg of nuclear
protein for each gel shift reaction. For analysis o f t he
5¢-transcription inhibitory region of iPLA
2
c, double-stran-
ded oligonucleotides containing 5 ¢-iPLA
2
c, sequence were
end-labeled with [
32
P]ATP using T
4
polynucleotide k inase,
as instructed by the manufacturer (Promega). Competition

studies were performed by adding a 100-fold molar excess of
unlabeled oligonucleotide or nonspecific control oligo-
nucleotide to the reaction m ixture p rior to the addition of
radiolabeled probe. Reaction mixtures were analyzed on
Novex 6% DNA retar dation polyacrylamide gels in 0.5·
TBE (89 m
M
Tris/HCl, pH 8.0; 89 m
M
boric acid; 2 m
M
EDTA) as the running buffer. Electrophoresis was per-
formed at 298 V for 20 m in, at 4 °C followed by drying of
the gel at 80 °C under vacuum and visualization of DNA–
protein complexes by autoradiography for 12–18 h. Sense
and reverse complement oligonucleotide sequences corres-
ponding to the following sequences were synthesized and
annealed: g50 (5¢-TATTAATCTGACTGTAGATATAT
ATATATTACCTCCTTAGTAATGC-3¢) and random-
ized control g50c (5¢-TTGATAGTTATCTATTACAG
TCTTCTTAGATTGAAACAA-3¢), g177 (5¢-CATACAA
ACATAATAAGATGTAAATGG-3¢) and control g177c
(5¢-TCATCTAAGTACAATAGATAGAAGAAA-3¢),
g230 (5¢-TGTTACTCTCCAAGCAAC CA-3¢) and control
g230c (5¢-GACACTTGTCATCACACTCA-3 ¢). For a na-
lysis of the pre-exon 1 region, myo2 double-stranded DNA
having the sequence 5 ¢-GAAGTACAGGTGTGGCTGG-
3¢ was u tilized along w ith control myo2ctl (5¢-GATCG
TTGTGAAGAGGGCG-3¢). For analysis of the pre-
exon 2 promoter region, Inr double-stranded DNA having

the sequence 5¢-GCGTCACTTCCGCTGGGGGCGG-3¢
was utilized along with randomized control Inrc (5¢-GTG
GCCGGGTGGTCCACCTCGG-3¢).
Mitochondrial target prediction, iPLA
2
c–GFP constructs
and confocal microscopy
The internet-based
MITOPROT
computer program (http://
www.mips.gsf.de/cgi-bin/proj/medgen/mitofilter) [37] was
utilized for prediction of mitochondrial targeting sequences
in iPLA
2
c. To prepare the 74-GFP construct, complement-
ary 5¢-phosphorylated primers (5¢-TCGAGCCAC CAT
GATTTCACGTTTAGCTCAATTTAAGCCAAGTTCC
CAAATTTTAAGAAAAGTAG-3¢ and 5¢-TCGACTACT
TTTCTTAAAATTTGGGAACTTGGCTTAAATAAA
CGTGAAATCATGGTGGC-3¢) were annealed by heat-
ing a 4-l
M
mixture of primers to 95 °C for 3 min followed
by cooling to 22 °C prior to cloning into the Xho1/Sal1
sites of vector pEGFP-N3. Integrity and orientation of the
N-terminal fusion products were verified by sequencing.
Vector pDsRed2-Mito (BD Biosciences), which encodes
a mitochondrial-targeting sequence of human cyto-
chrome c oxidase fused to red fluorescent protein, was
utilized as a mitochondrial marker. HeLa cells were grown

on two-well Laboratory Tek chamber slides to 60–80%
confluency prior to LipofectAMINE Plus (Invitrogen)
mediated single or cotransfection according to the manu-
facturer’s suggested protocol. After 48 h, cells were
washed in NaCl/P
i
, fixed with 4% (v/v) paraformaldehyde,
coverslipped and fluorescence was analyzed utilizing a
Zeiss Axiovert 200 (Carl Zeiss Inc., Thornwood, NY,
USA) equipped with Zeiss LSM-510 confocal system with
a63· oil immersion objective and excitation wavelengths
of 488 and 633 nm. Single transfections with either
pDsRed2-Mito or 74-GFP construct were utilized to
optimize immunofluorescence conditions and eliminate
bleed-through. Filters were optimized for double-label
experiments to minimize bleed-through and fluorescence
images were collected by utilizing Zeiss
LSM
software.
Results
Identification of transcriptional regulatory elements
nested in the 5¢-coding region of iPLA
2
c
In previous work, we demonstrated that expression of a
baculoviral construct encoding the full-length 88 kDa
coding sequence of iPLA
2
c in Sf9 cells resulted instead
in the p roduction of downstream polypeptides of 77 a nd

63 kDa in nearly equal amounts [26]. This was remarkable
because translation initiation almost always occurs at the
Ó FEBS 2004 Regulation of iPLA
2
c biosynthesis (Eur. J. Biochem. 271) 4713
AUG most proximal to the polyhedrin baculoviral promo-
ter [ 38,39]. Accordingly , the virtual absence of the 88 kDa
protein product was unanticipated. To begin identifying
the reasons underlying the differential expression of iPLA
2
c
polypeptides, we prepared pFASTBac1 vectors with the
baculovirus p romotor proximal to each o f the individual
AUG putative translation initiation codons. Analysis of t he
membrane fractions from Sf9 cells infected at identical
MOIs with vector harbori ng the construct containing the
polyhedrin promoter proximal to sequence encoding the
full-length iPLA
2
c 88 kDa polypeptide revealed two bands
of  77 and 63 kDa as previously reported [26] with the
63 kDa being the predominant product (Fig. 1, lane 1). An
uncharacterized band of  50 kDa was also present in a ll
fractions, including the uninfected control ( lanes 4 and 8),
which may represent either endogenous Sf9 cell iPLA
2
c
protein or alternatively nonspecific antibody binding. Ana-
lysis of t he membrane fraction from Sf9 cells infected with
vector harboring the truncation mutant encoding the

putative 74 kDa polypeptide revealed modest bands cor-
responding to the 74- and 63 kDa protein products (Fig. 1,
lane 2). The chemical identity of the minor protein product
of molecular mass > 74 kDa (Fig. 1, lane 2) is unclear and
may be due to secondary processing of the 74 kDa product
which could migrate anomalously. Alternatively, w e cannot
rule out the possibility that a minor amount of 3¢ read-
through from the expression co nstruct occurred. Remark-
ably, expression of the construct containing the polyhedrin
promoter proximal to sequence encoding the predicted
63 kDa product was over 75-fold higher than constructs
encoding either the 74- or 88 kDa protein products (Fig. 1,
lanes 3 and 7).
Lysates f rom v iral infections of the construct producing
the r ecombinant 6 3 kDa product pos sessed r obust P LA
2
activity (as assessed by release of oleic a cid from plasmenyl-
PC) t hat was markedly higher than that manifest in either
the 88- or 74 kDa transfected cells (data not s hown). The
rate of hydrolysis using p lasmenylcholine was similar to t hat
using phosphatidylcholine (each radiolabelled a t the sn -2
position with 9 ,10-[
3
H]oleic acid). These results demonstrate
that the i PLA
2
c enzyme can attack the sn-2 carbonyl a nd
suggest that hydrolysis of these substrates by the 63 kDa
iPLA
2

c occurs predominantly at the sn-2 position.
Measurement of mRNA content and kinetics of mRNA
species encoding individual iPLA
2
c isoforms
Alterations in the amount of iPLA
2
c isoform expression
could be due to changes in mRNA synthesis, differences in
mRNA half-lives, or translational mechanisms for each of
the sequentially truncated coding constructs. Accordingly,
we first examined the amount and stability of mRNA
resulting from each of the constructs in the baculoviral
expression system. Northern analysis revealed only modest
amounts of mRNA mass corresponding to the constructs
encoding the 88 k Da protein and virtually none encoding
the message for the 74 kDa protein (Fig. 2A). Remarkably,
a dramatic increase in the mRNA content in cells
transfected with vector encoding the 63 kDa protein
product w as present (Fig. 2A). These experiments were all
performed at identical MOIs and reproduced on multiple
occasions. After actinomycin D treatment, the half-life of
each mRNA species was compared by two independent
techniques. First, compar isons of iPLA
2
c mRNA mass
expressed from each of the constructs over a 4 h interval
following actinomycin D treatment did not reveal any
discernable differences in mRNA stability by northern
analysis (t

1/2
 1–2 h; Fig. 2B). Second, quantitative PCR
analysis after actinomycin D treatment indicated that
mRNA levels expressed following viral infection with the
63 kDa construct w ere substantially higher than those o f
either the 88- or 74 kDa constructs (t
1/2
 2–4 h; Fig. 2C).
Collectively, these results demonstrated that transcriptional
regulation was a major mechanism underlying the experi-
mentally observed dramatic increase in the 63 k Da protein
mass but did not rule out contributions from translational
mechanisms as well (vide infra).
Localization of the regulatory domain mediating
transcriptional repression of the iPLA
2
c constructs
The observed differences in baculoviral expression patterns
of the sequentially truncated iPLA
2
c message suggested that
a transcriptional i nhibitory element w as present comprised
of nucleic acid sequence encoding the N-terminus of iPLA
2
c
located between the 88- and 63 kDa potential translational
initiation sites. To localize the regulatory domain upstream
ofthe63kDastartsiteofiPLA
2
c responsible for the

observed transcriptional repression, PCR products contain-
ing 124-nucleotide blocks of sequence upstream of the
63 kDa start site were amplified from iPLA
2
c template
and inserted between the SV40 promoter and a luciferase
reporter gene in a pGL3-promoter vector for transient
expression in monkey kidney (CV1) cells. Through this
approach, w e sought to determine which elements in the
5¢-coding sequence a cted as transcriptional repressors in a
mammalian cell line. Constructs corresponding to each of
the first four 124-nucleotide sequences encoding truncated
sequences from the 5¢ of nucleo tide 315 greatly inhibited
luciferase expession (on average  80%), whereas segments
further 3¢ were not inhibitory in comparison with control
Fig. 1. Baculoviral e xpression of truncated iPLA
2
c polypeptides initi-
ating translation at downstream in-frame initiator methionine residues.
Sf9 cells were infected with iPLA
2
c con structs initiating at the 88, 74 or
63 kDa start sites. At 48 h postinfection, cells we re collected and
membrane (lanes 1–4) and cytosolic (lanes 5–8) fractions were p re-
pared a s described in Experimental proced ures. Fractions (100 lg
protein per lane) w ere loade d onto a 10% polyacrylamide gel, resolved
by S DS/PAGE, transferred to a p oly(vinylidene difluoride) me mbrane,
incubated with immunoaffinity-purified antibody directed against
iPLA
2

c and visualization of immunoreactive bands by ECL.
Expressed recombinant polypeptides are de sign ated according to their
expected masses. Lanes 1 and 5, 88 kDa iPLA
2
c; lanes 2 and 6, 74 kDa
iPLA
2
c; lanes 3 and 7, 63 kDa iPLA
2
c; lanes 4 and 8, wi ld-type control
baculovirus. Molecular mass standards are indicated on the left.
4714 D. J. Mancuso et al. (Eur. J. Biochem. 271) Ó FEBS 2004
vector (P<0.001) (Fig. 3 A). Moreover, EMSA of the
5¢-coding region utilized for the above s tudy (nucleotides
1–230 of iPLA
2
c) revealed three separate regions producin g
gel shifts, a ll localized within the identified region of tran-
scriptional repression. Oligonucleotide g 50 was predicted to
contain s ites with a high match to chicken homeobox CdxA
binding sites. Oligonucleotide g177 shares homology with
the Oct1 binding site, whereas oligonucleotide g230 did not
contain a predicted site for binding of nuclear proteins.
Utilizing radiolabeled oligonucleotide dimer g50 (corres-
ponding to residues 6–50 starting from the 88 kDa AUG
codon) a single shifted protein-radiolabeled DNA complex
utilizing HeLa nuclear extract was identified which was
competed out with a 100-fold molar excess of unlabeled
oligonucleotide dimer g50 but not with nonspecific control
g50c oligonucelotide dimer (Fig. 3B, column 1, arrow).

Similarly, oligonucleotide dimers g177 and g230 also
produced shifted bands that were specifically competed
out with 100-fold molar excess unlabeled oligonucleotide
dimer but not with nonspecific control oligonucleotide
dimer (Fig. 3B, columns 2 and 3).
Translational regulation of iPLA
2
c in myocardium
Owing to the obvious complexity of the regulation of
iPLA
2
c resulting from the combined presence of transcrip-
tional and trans lational regulation, we recognized that
current hypotheses relegating the role of iPLA
2
c exclusively
to peroxisomal lipid metabolism were likely limited in
appropriate scope. In prior work, we identified robust
amounts of iPLA
2
activity in the mitochondrial compart-
ment of both canine and human hearts [7,8]. Moreover, we
recognized early on that t he iPLA
2
family of proteins had
the potential for providing substrate f or mitochondrial fatty
acid oxidation by lipid hydrolysis [7], for g enerating lipid
second messengers ( eicosanoids and l ysolipids), for modu-
lating ion channel kinetics [19,40] and for providing fatty
acids for univalent transmembrane ion transport [41].

Accordingly, we c onsidered the possibility t hat myocardial
iPLA
2
c may be present in mitochondria. Western analysi s
demonstrated that iPLA
2
c cosedimented with mitochondria
(in the light mitochondrial fraction) (Fig. 4). Remarkably,
multiple high m olecular mass (63–88 kDa) immunoreactive
proteins were detected in rat mitochondria after differential
centrifugation o f r at hear t homogenates, c onsistent with the
utilization of translation initiation sites producing 88, 77, 74
and 63 kDa protein p roducts. Products corresponding to
the 77 a nd 74 kDa products were t he major i mmuno-
reactive bands. A dditional lower molecular m ass immuno-
reactive bands were also detected (< 60 kDa). Collectively,
these results identified the presence of multiple iPLA
2
c
protein products in mitochondria resulting from usage of
different translation initiation codons in rat myocardium.
Alternative splicing of iPLA
2
c in mammalian tissues
In the years since our first report of the g enomic organiza-
tion of the iPLA
2
c gene, i ncreasing evidence o f extensive
Fig. 2. Analysis of iPLA
2

c mRNA in the baculoviral system. (A) Sf9 cells were infected w ith iPLA
2
c constructs en coding full-len gth (88 kDa) or
truncated 74 an d 63 kDa pro ducts. At 48 h postinfectio n, cells were recovered and total R NA was extracted, fractionated on a latitude RNA gel,
transferredtonylonmembraneandhybridizedwith[
32
P]iPLA
2
c probe followed by autoradiography as described in Experimental procedures.
Lane 88 kDa, RNA f rom 8 8 kDa full-length expression; lane 7 4 kDa, RNA from 74 kDa ex pression; l ane 63 kDa, RNA from 63 kDa expressio n.
The relative positions of RNA size markers in kb are indicated on the left. (B) N orthern analysis of total RNA e xtracted from Sf9 ce lls infecte d for
48 h with re combinant full-length or truncated iPLA
2
c baculovira l c onstructs and then treated with actinomycin D for 0, 0.25, 0 .5, 1, 2 or 4 h prior
to RN A e xtraction. Lane 88 kDa, RNA from 88 kDa fu ll-length expression; lane 7 4 kDa, R NA from 74 kDa expre ssion; lane 63 kDa, RNA f rom
63 kDa expression. The relative positions of RNA size mar kers a re shown in k b on the left. (C) Quantitative P CR analys is of iPLA
2
c mRNA levels.
RNA isolated and DNase treated from 48 h infected Sf9 cells was revers e transcr ibed usin g MultiS cribe reverse transcriptase and the resultant
cDNA (20 ngÆreaction
)1
)utilizedinquantitativePCRasdescribedinÔExperimental proced ures.Õ Log of the relative mRN A level i s p lotted v s. time
(in hours) after actinomycin D addition for RNase-free DN ase-treated R NA extracts of baculoviral extracts e xpressing the 63 kDa ( m), 74 kDa
(j)and88kDa(r)iPLA
2
c p olypeptides.
Ó FEBS 2004 Regulation of iPLA
2
c biosynthesis (Eur. J. Biochem. 271) 4715
alternative splicing in the 5¢-region of t he iPLA

2
c gene has
accumulated along with evidence for the existence of two
previously undescribed exonic sequences within some of the
alternatively spliced iPLA
2
c variants in GenBank
TM
data-
bases. Although previously only present as raw sequence in
the EST database, we now specifically identify two novel
sequences as iPLA
2
c exons. The first exon comprised of 296
nucleotides was located at the 5¢-end of EST sequences
containing iPLA
2
c sequence and is the 5¢-most exonic
sequence located thus far for the iPLA
2
c gene. F or this
reason, t his exon has been designated exon 1 (Fig. 5 ). Based
on its location relative to other iPLA
2
c exons, we have
designated the second new exon as exon 4 . Exon 4 is
comprised of 112 nucleotides and, remarkably, has a high
degree of homology with the human mammalian transpo-
son-like element MaLR repeat sequence. The s ignificance of
this sequence homology in the context o f exons within the

iPLA
2
c gene remains unknown. Thus, the second draft of
the iPLA
2
c genomic map contains 13 exons, the first four of
which contain noncoding sequence (Fig. 5). The first
potential in-frame AUG start site occurs in exon 5 , while
the nucleotide binding and lipase consensus sites occur in
exons 7 and 8, respectively, and the peroxisomal localiza-
tion signal occurs in exon 13 (Fig. 5).
Fig. 4. Immunoblot analysis of iPLA
2
c in subcellular fractionations of
rat heart. E quivalent subcellular fractions (100 lg protein) of rat heart
prepared as described in Experimental procedures were loaded on a
10% gel, resolved by SDS/PAGE , transferred to a poly(vinylidine
difluoride) membrane, incubated with immunoaffinity-purified anti-
iPLA
2
c, a nd im munoreac tive ba nds were visualized by ECL. Lane 1,
rat heart homogenate; lane 2, crude pellet; lane 3, heavy mitochondrial
fraction; lane 4, 23 500 g light mitochondrial f raction; lane 5, 70 000 g
light mitochondrial fraction; lane 6, nuclear fraction. Molecular mass
markers are indicated on the righ t.
Fig. 3. Identification of a regulatory domain within t he coding region of
iPLA
2
c using a luciferase reporter assay system. Th e i nhib itory e ffect of
iPLA

2
c sequences on luc iferase expression were examined by prepar-
ing a series iPLA
2
c-pGL3-Prom oter constructs consisting of
124-nucleotide segments of iPLA
2
c sequence (fro m the region
upstream from the 63 kDa i PLA
2
c start site) cloned immedia tely
upstream fro m the luciferase reporter gene in vector p GL3-Promoter.
CV1 cells were transiently t ransfected with the iPLA
2
c-pGL3-Pro-
moter constructs (100 ng) and harvested 24 h later to assay luciferase
activity as described in Experimental procedures. (A) The regions of
the iPLA
2
c coding sequence included in iPLA
2
c-pGL3-Promoter
constructs 88, 87, 85, 83 and 80 as well as regions correspond ing to
oligonucleotide g50, g 177 a nd g230 used for EMSA are schematically
represented. A portion of t he 5¢ iPLA
2
c coding sequence (iPLA
2
)is
represented in the center of the diagram as a heavy solid bar w ith the

scale in nucleotides ( nt) sh own below. (B) The bar graph i ndicates the
relative luminescent value of iPLA
2
c-pGL3-Prom oter constructs 88,
87, 85, 83, 80 and 74 compared with unmodified pGL3-Promoter
control vec tor u sed i n t he lu cife rase as say system. Results represent the
average o f three sets of data (± SE). Co mp arison of the RFV o f the 80
construct with 88, 87, 85 and 83 constructs (P < 0.001) is indicated
(*). (C) E MSA o f t he iPLA
2
c regulatory domain. EMSA was per-
formed utilizing double-stranded radiolabeled oligonucleotides g50 ( 1),
g177 (2), and g230 (3) as d escribed in Experimental procedures. Lane a ,
negative control minus HeLa nu clear extract; lane b, positive co ntrol
containing HeLa nuclear extract; lane c, competitive assay containing
100-fold molar excess o f unlabeled oligonucleotide; lane d, noncom-
petitive assay containing 100-fold molar excess unlabeled nonspecific
control oligonucleotide. Results are representative of three s eparate
EMSA. Arrows: specific DNA–nuclear protein complex.
4716 D. J. Mancuso et al. (Eur. J. Biochem. 271) Ó FEBS 2004
In addition to transcriptional regulation of mRNA levels,
alternative splicing represents an additional mechanism for
the r egulation of iPLA
2
c biosynthesis. Examination of the
EST database a nd 5¢-RACE analyses revealed a total of ten
different splice variants from eight different tissues which
begin with e ither t he exon 1 or e xon 2 sequence (but do not
contain both) (Fig. 6). Multiple iPLA
2

c splice variants were
identified in a wide range of t issues, including human heart,
smooth muscle, endothelial cell, hippocampus, t estis, pitu-
itary, placenta and pancreas. The predominant splice
variant isolated by 5¢-RACE, and the one most often
present in the EST d atabase, was splice v ariant VI followed
by splice variants V and IV. Multiple splice variants from
different tissues that differ w ith regard to their 5 ¢-terminus
were present. Seven begin with exon 2, whereas three begin
with the exon 1 sequence. Splice variants I and II do not
contain the exon 5 sequence and thus do not contain
sequence for the four alternative AUG start sites initiating
biosynthesis of the 88, 77, 74 and 6 3 k Da iPLA
2
c isoforms.
Instead, based on current information about iPLA
2
c and i ts
splicing, the fi rst in -frame AUG site is downstream o f the
nucleotide binding and lipase consensus domains and thus
encodes a putative potential 33 kDa polypeptide which does
not contain the serine active site. The reasons underlying the
presence of this product are unknown, but it could be
involved in regulatory events similar to s plice variants o f
iPLA
2
b previously identified that do not contain the active
site serine [42–44]. Splice variants III, IV, VIII and IX have
an alternative AG/GT splice site within exon 5 resulting in a
truncated exon 5 that is missing the 88 kDa iPLA

2
c start
site. Intere stingly, the alternative splicing that g enerates
variant IV results in a new 5 ¢ in-frame AUG start site, which
Fig. 6. Splice variants o f iP LA
2
c be ginning with e ither exon 1 or ex on 2. A diagrammatic representation of iPLA
2
c e xons is indicated at the top with
the relative locations of th e 88, 77, 74 a nd 63 kDa ATG start sites indicated by triangles. Vertical arrows indicate th e locations of t he nucleotide
(ATP) and lipase consensus s ites. Representatio ns of ten splice variants a re sho wn b elow with lines indicating splicing across exons. O pen boxes
represent 5¢-untranslated sequence while shaded boxes represent the open reading frame . In addition to the f our upstre am A TG start sites encoding
88, 7 7, 74 and 63 kDa products, all potential in-frame downstream ATG start sites are also indicated with triangles. Slashes ind icate the e xtent of
known seque nce f or e ach ES T. Asterisks designate splice v ariants ide ntified by 5¢-RACE in this study.
Fig. 5. G enomic m ap of iPLA
2
c. The intron–exon boundaries of the iPLA
2
c gene are sho wn in scale (k b). The 13 exons o f the iPLA
2
c gene a re
indicated as boxes. Spaces between the exons represent t he relative sizes of the 12 introns contained within the iPLA
2
c gene. Re gions of the gene t hat
correspond to the nucleotide binding, lipase, and peroxisomal localization consensus sequences are indicated in exons 7, 8 and 13, respectively.
Open boxes at the bottom indicate the n ucleot ide numbers (corresponding to the original BAC genomic clone report, GenBank accession number
AC005058) with the sizes of each exon in nucleotides (nt) and in amino acids (aa) shown within. The asterisk is i nserted to note that different 5¢
extents of exon 2 have been reported in GenBank [26, 45] a s well as in the EST database.
Ó FEBS 2004 Regulation of iPLA
2

c biosynthesis (Eur. J. Biochem. 271) 4717
can potentially encode a polypeptide of 91.6 kDa. Trans-
lation from this upstream A UG site thus results i n an
additional potential N-terminal 43 amino acids from
sequence previously regarded as 5 ¢-untranslated sequence.
Because of the truncation of exon 5 , there is also a loss of 15
amino acids including the 8 8 k Da start site. The c omplete
sequence of two splice v ariants (V and VI) h as been
published [ 26,45]. 5 ¢-RACE w as utilized to clone sequence
corresponding to splice variants III, IV, V, VI, VII, IX and
X in this study. Sequence for splice sequence IX, isolated in
this study by 5¢-RACE of human myocardial cDNA, has
not been previously reported in the EST database. Collec-
tively, these results underscore the complexity in the genetic
and molecular biologic m echanisms regulating the tran-
scriptional processing of iPLA
2
c into moieties suitable fo r
translation o f s pecific polypeptides potentially tailored t o
fulfill specific biologic roles in different tissues.
Identification of alternative promotors present in iPLA
2
c
and demonstration of three MyoD regulatory elements
Alternative promoter usage represents yet another potential
mechanism for the regulation of t he biosynthesis of iPLA
2
c.
Because iPLA
2

c splice variants began with either exon 1
or exon 2, sequences upstream of these exons were next
examined for promoter a ctivity. Accordingly, we prepared
constructs in which 584 nucleotides of upstream iPLA
2
c
sequence from each e xon were utilized to drive luciferase
reporter gene expression in CV1 cells. Sequences upstream
of exon 2 had high promoter activity, whereas the pre-
exon 1 sequence had negligible prom oter activity in CV1
cells (Fig. 7A). Truncation of the 5 ¢ 200 nucleotides of the
pre-exon 2 sequence (Fig. 7B, construct II) resulted in an
 15-fold increase in promoter activity suggesting the
presence of repressor elements in the region 400–600
nucleotides upstream of exon 2. Removal of an additional
200 nucleotides from construct II resulted in the loss of the
majority of activity (construct III) indicating that the region
200–400 nucleotides upstream o f exon 2 contains a signifi-
cant proportion of pre-exon 2 promoter activity. This
conclusion was supported by use of a construct containing
sequence 200–400 nucleotides upstream of e xon 2 (con-
struct IV) which resulted in a fivefold increase in promoter
activity compared with the original construct (I), whereas a
construct containing sequence 400–584 nucleotides up-
stream of exon 2 (V) had only slight promoter activity.
Construct VI, including sequence 200–584 nucleotides
upstream of exon 2 , had promoter activity similar to that
of construct IV. Construct VII (sequence 300–400 nucleo-
tides upstream of exon 2) had no detectable promoter
activity, whereas constructs VIII (sequence 200–300 nucleo-

tides upstream of exon 2) and IX (200–350 nucleot ides
upstream of exon 2) had similar promoter activity com-
pared with t he original construct. Promoter activity of genes
are typically regulated by a complex interplay of multiple
promoter elements and t his is reflected in the data prese nted
in Fig. 7B. These results s uggested that a region 200–400
nucleotides upstream of exon 2 contains a major proportion
of the promoter activity of the pre-exon 2 sequence.
However, this activity is clearly modulated by sequences
upstream a nd downst ream of this region. The region 200–
400 nucleotides upstream of exon 2 includes predicted
A
B
C
Fig. 7. Promoter analysis of the 5¢ flanking region of iPLA
2
c exon 2.
(A) An iPLA
2
c promoter co nstruct containing 584-nu cleotide iPLA
2
c
sequence up stream of exons 1 or 2 inserted upstream via HindIII/NcoI
sites i nto the pr omoterless vector pG L3-Enhancer from Pro mega.
Empty pGL3-Enhancer vector and t he SV40 containing promoter
vector pGL3-Promoter were used as controls. Luciferase activity
measured as relative luminescencevalueisshownforvectorpGL3-
Enhancer constructs utilizing 584 nucleotide of iPLA
2
c sequen ce as an

upstream promoter. Lanes indicate construct s containing as p romo ters
pre-exon 1 s equence (pre-exon 1), p re-exon 2 sequence (p re-exon 2),
and the promoterless vector pGL3-Enhancer (pGL3E). (B) Constructs
I–IX containing sequence upstream from exon 2 were prepared by
PCR a mplificatio n of intronic sequence u pstream from iPLA
2
c exon 2,
cloning the PCR products into promoterless vector pGLE, followed by
transfection of CV1 cells as described in Experimental procedures.
Relative size s and n ucleotid e regions include d in each construct are
indicated as blocks t o the left. Luciferase activity, expressed as relative
luminescence value, for each construct is indicated on the right.
(C) Competitive gel retardation analysis of the pre-exon 2 iPLA
2
c
region utilizing In r dime r. Lane 1 , negative c ontrol min us He La nuc -
lear extract; lane 2, positive control containing HeLa nuclear extract;
lane 3, competitive assay containing 100-fold molar excess Inr dimer;
lane 4, noncompetitive assay c ontaining 100-fold molar excess non-
specific control dimer. Results a re representative of three s eparate
EMSA. Arrow: specific DNA–nuclear protein complex.
4718 D. J. Mancuso et al. (Eur. J. Biochem. 271) Ó FEBS 2004
transcription binding sites for known enhancers including
p300 and c-Myb transcription factor s ites, as well as a
CACG repeat sequence. EMSA with sequences including
the p300, c-Myb and CACG sites did not result in a gel shift
(data not shown). Alignment with homologous rat and
mouse sequence reveals several regions of high homology
upstream from exon 2 including most notably, a 22-nuc-
leotide sequence (5¢-GCGTCACTTCCGCTGGGGG

CGG-3¢) which contains a potential initiato r (Inr) sequence
with the c onsensus sequence P y-Py-A-N-T/A-Py-Py. Inr
sequences typically occur at t he transcription s tart site in
genes [46]. Moreover, EMSA provided evidence that a
specific transcriptional regulatory protein binds to the Inr
region in the pre-exon 2 sequence (Fig. 7C).
Finally, we noted the presence of a se quence correspond-
ing to three E-boxes with the s equence 5¢-CAGGTG-3¢
within the 350 nucleotides upstream of e xon 1 , suggesting
that promoter activity may be activated by a myogenic
factor such as MyoD [47–49]. The E-boxes were located at
)27, )58 and )354 nucleotides upstream of the predicted
start for exon 1. To substantiate the functional importance
of this assignment, additional experiments were performed.
Cotransfection of CV1 cells with MyoD and p re-
exon 1-containing vectors resulted in a fivefold stimulation
of promoter activity compared with MyoD + pGLE
vector cotransfection ( Fig. 8A) . EMSA demonstrated that
sequence corresponding to the middle E-box sequence
produced a gel shift utilizing HeLa cell nuclear extract
(Fig. 8 B). This gel shift was competitively blocked during
incubation with 100-fold molar excess of unlabeled double-
stranded nucleotide sequence containing the E-box but was
not attenuated during incubation with a nonspecific control
(i.e. randomized) sequence. These results suggest the
importance o f iPLA
2
c in muscle tissues as initially predicted
from northern blot analyses [26].
Identification of mitochondrial import sequences

in iPLA
2
c and demonstration of their functional integrity
Recent increases in the understand ing of mitochondrial
import machinery have led t o the generation of computer
algorithms that can accurately assess the potential for
mitochondrial localization of a peptide with great accuracy
and predictive probabilities. Analysis of the sequence of
iPLA
2
c corresponding to the 74 kDa encoded pep tide
identified a putative mitochondrial localization s equence
with a predictive value of > 96% and a predicted cleavage
at amino acid K138. Differential centrifugation analysis
demonstrated the mitochond rial localization f or this poly-
peptide (see above). To confirm the mitochondrial localiza-
tion of the iPLA
2
c polypeptide experimentally by an
independent, functional method not susceptible to intrapre-
parative translocation, the widely accepted technique of
fluorescence dual image analysis overlay was employed. A
commercially available marker for mitochondrial localiza-
tion (pDsRed2-Mito, BD Biosciences) was employed to
identify the mitochondrial compartment. Dual transfection
of HeLa cells with a construct containing the leader
sequence of the 74 kDa polypeptide proximal to GFP
(residues 1–32 of the 7 4 kDa product) and pDsRed2-Mito
containing an authentic mitochond rial localization sequence
(human cytochrome c oxidase) resulted in green and red

punctate patterns, respectively ( Fig. 9, panels 1 and 2).
Merging the two images yielded an intense yellow punctate
pattern indicative of mitochondrial localization mediated
by the N-terminal 74 kDa iPLA
2
c mitochondrial leader
sequence (Fig. 9, panel 3). The absence of bleed-over is
evident f rom t he green fluorescence of the c ell at the bottom
of Fig. 9, panel 1 with no bleed-over into red (indicative of
single transfection with 74-GFP). Similarly, cells in the
upper r ight fl uoresce red ( in panel 2) indicative of single
transfection with pDsRed2-Mito, with no bleed-over into
green (in panel 1). Constructs comprised of the 88 and
77 kDa N-terminal leader sequences proximal to GFP w ere
prepared and showed a similar colocalization (data not
shown). Control experiments with GFP alone demonstrated
a diffuse cytosolic localization (data not shown).
Three independent techniques (computer-predicted local-
ization, differential centrifugation and fluorescence dual
wavelength overlay m icroscopy) collectively demonstrate
Fig. 8. MyoD stimu lation o f p romoter a ctivity i n p re-exon 1 sequence of
iPLA
2
c. (A) Luciferase activity m easured as relative luminescence
value is shown for vector pGL3-Enhanc er construct (pre-exon 1)
containing 584 nucleotides of pre-exon 1 iPLA
2
c sequence as an
upstream promoter. Lanes indicate transfection with pGLE or pre-
exon 1 with (+) or without (–) cotransfection with vector expressing

MyoD (MyoD). *P<0.01. (B) Competitive gel retardation analysis
of th e pre-exon 1 iPLA
2
c region utilizing m yo2 dimer. Lane 1 , negative
control m inus HeLa nuclear e xtract; l ane 2, positive control containing
HeLa nuclear extract; lane 3, competitive ass ay containing 100-fold
excess myo2 dimer; lane 4, noncompetitive assay containing 100-fold
molar excess nonspecific control dimer. Results are representative of
three separate EMSA. Arrow: spe cific D NA–nuclear prote in com plex.
Ó FEBS 2004 Regulation of iPLA
2
c biosynthesis (Eur. J. Biochem. 271) 4719
that iPLA
2
c is localized, in large part, to the mitochondrial
compartment in myocardium. Collectively, these results
demonstrate that use of the three proximal in-frame
translation i nitiation codons sites results in the p roduction
of a polypeptide that is targeted to the mitochondrial
compartment b y i ts N-terminal sequence and, at the same
time, t argeted to the perox isomes by virtue o f its C-terminal
SKL peroxisomal localization sequences. T hus, the presence
of dual localization s equences in a single protein identifies a
competition for nascent iPLA
2
c between the mitochondrial
compartment (from its N-terminal mitochondrial l ocaliza-
tion sequence) and the peroxisomal compartment (from its
C-terminal SKL peroxisomal localization sequences) in
isoforms that carry both signals. It is important to note t hat

synthesis of the 63 kDa isoform obligatorily is relegated to
the peroxisomal compartment by virtue of the absence of
any N-terminal mitochondrial leader sequence in this
isoform.
Discussion
After t he initial discovery of a novel intrace llular c alcium-
independent PLA
2
activity in myocardium [6], subsequent
detailed biochemical characteri zation revealed an unsus-
pected complexity of catalytically active isoforms which
could be d iscriminated based upon their m olecular mass
differences, subcellular localizations, k inetic characteristics,
substrate selectivities, chromatographic profiles and protein
chemical techniques including radiolabeling with [
3
H]BEL
[7–12]. Study of human heart PLA
2
underscored the
complexity of multiple distinct isoforms of iPLA
2
in the
cytosolic, microsomal and mitochondrial compartments in
human heart [7,8]. Among the m embrane-associated activ-
ities i dentified was a subset that localized to the mito-
chondrial c ompartment and which was functionally
discriminated from their cytosolic and microsomal coun-
terparts by multiple criteria [7,8]. Purification of the human
heart cytosolic phospholipase activity revealed a proteoly-

tically activated, high specific activity 40 kDa isoform and
the parent 85 kDa PLA
2
holoenzyme [8]. Activity of i PLA
2
is regulated by ATP, c almodulin and proteolysis as well as
potentially many other factors [23,29]. Recent work has
demonstrated the pleiotropic r oles that iPLA
2
b serves in
mediating multiple diverse biochemical and physiologic
functions including arachidonic acid r elease [50], i nsulin
release [ 51,52], modulation of vascular tone [53], ischemia-
induced arr hythmogenesis [40] and cellular electrophysio-
logic function [18–20].
In contrast to the cytosolic iPLA
2
activity, the mem-
brane-associated calcium-independent activities from nat-
urally occurring sources remained resistant to sustained
efforts a t purification. Recently, we identified a novel
member of the iPLA
2
family from the Human Genome
Project based upon the p resence of its dual s ignature
nucleotide binding and lipase site consensus sequence
motifs which was remarkable for the presence of a
C-terminal SKL peroxisomal localization sequence [ 26].
Last year, the first evidence of the protein product of t his
gene in naturally occurring tissue (i.e. nontransgenic) was

obtained which confirmed i ts predicted peroxisomal local-
ization and demonstrated the presence of a predominantly
63 kDa protein product in liver peroxisomes [27]. However ,
prior northern blot analyses demonstrated that myocar-
dium was the most abundant source of mRNA encoding
iPLA
2
c. Our results confirm the abundance of iPLA
2
c in
myocardium and demonstrate the diverse repertoire of
iPLA
2
c protein products generated from a single gene
under complex regulatory control i ncluding transcriptional,
translational and proteolytic mechanisms. Recently, the
functional importance of BEL-inhibitable calcium-inde-
pendent PLA
2
activities in mitochondrial b ioenergetics has
been recognized [15]. The localization of multiple iPLA
2
c
isoforms to the mitochondrial compartment in myocar-
dium in conjunction with prior work identifying the role of
BEL-inhibitable iPLA
2
enzymes i n mitochondrial bioener-
getics identifies a portion of the diversity of bioch emical
mechanisms which contribute to cellular energy regulation

through m odulating mitochondrial nonesterified fatty acid
content. Multiple mechanisms of regulation of iPLA
2
c
activity are apparent from our results.
First, transcriptional regulation by t he 5¢-coding sequence
was d emonstrated in multiple different contexts. Although
transcriptional repression is usually mediated by noncoding
regions of the gene, many prior precedents for important
Fig. 9. Import of GFP into mitochondria mediated by an N-terminal iPLA
2
c mitochondrial import signal. Fusion construct 74-G FP was prep ared
containing iPLA
2
c sequence from the 74 kDa AUG start codon through nucleotide 414 and cloned in-frame at the 5¢-end of the GFP coding
sequence of vector pEGFP-N3. The construct was cotransfected with vector DSRed2 into HeLa cells. After 48 h, cells were visualized utilizing
confocal microscopy with excitation wavelengths of 488 and 633 nm. Visualization was performed with the following wavelengths: (1) 488 nm,
(2) 633 nm, (3) 488 and 633 nm. The colocalization of iPLA
2
c-GFP and DsRED2-mito was revealed by the nearly identical staining patterns
(comparing 1 a nd 2) and by the ye llow co lor re sulting f rom c ombination of green and red flu orescence in 3.
4720 D. J. Mancuso et al. (Eur. J. Biochem. 271) Ó FEBS 2004
transcriptional regulatory elements within the coding
sequence of many genes exist [ 54–58]. In this study, evidence
for transcriptional regulation of iPLA
2
c by elements nested
within the coding region was based on a dramatic and
concomitant increase i n both mRNA encoding iPLA
2

c and
protein synthesis upon sequential truncation of the 5¢-end of
the i PLA
2
c coding sequence in multiple different cellular
contexts. Moreover, specific 5¢-segments of the iPLA
2
c
sequence also repressed luciferase expression in CV1 cells
when placed upstream of the luciferase reporter gene in a
promoter-driven luciferase expression system. Finally,
EMSA within the identified repressor region supported a
conclusion that binding by multiple, as yet unidentified,
nuclear proteins occurs which could r esult in transcriptional
modulation of iPLA
2
c. These results strongly support the
conclusion that iPLA
2
c biosynthesis is transcriptionally
regulated by a r epressor domain within the first 315
nucleotides of the h uman iPLA
2
c coding sequence. We
recognize that multiple AUG codons have previously been
demonstrated to modulate translational repression in other
systems [ 59] (and also modulate i t in different intracellular
contexts), but these translational effects alone could not be
responsible for t he massive tran scriptional changes present
with iPLA

2
c in the multiple systems under study. Accord-
ingly, the majority of alterations in 63 kDa protein mass in
multiple cellular contexts identified here in are mediated
predominantly by transcriptional r egulation. We specifically
point out that this study identifies only one region of
transcriptional repression in iPLA
2
c and that many other
regions (repression or activation) may be operative in
multiple other nutritional, environmental, or cellular con-
texts found in nature.
The second mechanism regulating iPLA
2
c isoform bio-
synthesis uncovered in this study is translational regulation
mediated, i n part, by the differential usage of downstream
AUG start sites. Previously, we demonstrated that the
63 kDa isoform of iPLA
2
c was predominantly expressed in
rat liver peroxisomes [27]. This study demonstrated that
multiple higher molecular mass immunoreactive poly-
peptides were expressed in rat heart. The molecular masses
of the observed polypeptides were in accordance with the
predicted molecular masses of the iPLA
2
c isoforms origin-
ating from translation initiation at each of the proximal four
in-frame N-terminal methionines. In total, seven p eptides

(some obvious ly representing proteolysis products) were
easily identified by western analysis. Many additional
isoforms are likely present within each band (i.e. post-
translation modifications which can not be discriminated b y
SDS/PAGE), which could modify activity, protein–protein
interactions, cellular half lives, or s ubcellular localization. In
addition, lower abundance peptides are almost certainly
present but cannot visualized by wester n techniques at the
sensitivities u tilized in this study. Thus, differential expres-
sion of specific iPLA
2
c isoforms by programmed usage of
alternative AUG start sites in tissues under different
nutrient conditions may be an important mechanism
allowing iPLA
2
c to fulfill its multiple roles in physiologic
function. The ability to produce distinct protein products
from a single gene suggests that mechanisms for diversifi-
cation and specialization h ave e volved to allow i PLA
2
c to
participate in regulating lipid metabolism in multiple
compartments thereby contributing to orchestrating energy
storage, u tilization and signal transduction pathways in
multiple compartments through utilization of a repertoire of
discrete chemically distinct isoforms by multiple different
mechanisms.
It should be recognized that usage of alternative AUG
start sites provides a mechanism for targeting discrete

iPLA
2
c protein i soforms encoded by a single mRNA
molecule to specific subcellular loci. By omitting the
N-terminal mitochondrial localizaion sequences through
use of the fourth downstream in-frame AUG start site, a
63 kDa protein pro duct is synthesized which is targeted to
the peroxisomal compartment by virtue of a C-terminal
SKL sequence. Thus, the subcellular location and functional
properties of t he synthesized iPLA
2
c polypeptides can be
precisely tailored to adapt to acute changes in cellular
nutritional states or environment.
Alternative splicing represents a third mechanism regu-
lating iPLA
2
c biosynthesis. Alternative splicing results in
the production of at least ten iPLA
2
c var iants. Moreover,
alternative splicing is a mec hanism that c ontributes to the
AUG start site whic h is utilized during translation in the
iPLA
2
c gene by introduction of open reading frames
upstream of the start site as well as by splicing out in-frame
AUG start sites. Evidence demonstrating this mechanism
was obtained from iPLA
2

c splice variants cloned through
5¢-RACE and from analyses of EST sequences present in
the GenBank database. Alternative splicing i s a well-
recognized mechanism o f modulation o f protein fun ction
in other genes where it can influence many cellular processes
including signal transduction, transcriptional regulation,
cellular transformation and subcellular l ocalization in other
genes [60–63]. Accordingly, alternative splicing o f iPLA
2
c
can modulate both the site and the amount of enzymic
activity (e .g. thro ugh remov al of t he N-terminal mitoch-
ondrial localization signal o r removal of the lipase active site
sequence). We have already noted that several splice
variants utilize an a lternative splice junction 44 nucleotide
downstream from the start o f exon 5 resulting in loss of the
AUG initiating translation of the 88 kDa isoform. We
specifically point out that differential utilization of exon 1
vs. exon 2 in splice variants introduces differing upstream
ORFs which could potentially regulate translation.
The identification of ten iPLA
2
c splice variants through
5¢-RACE and analyses of the E ST database demonstrate
that two basic types of splice v ariants of iPLA
2
c exist:
(a) those initiating transcription from exon 1 ; or ( b) those
initiating transcription from exon 2. Thus far, no splice
variants have been identified containing both e xons 1 and 2

either through analysis of s equences within GenBank, by
RT-PCR analysis or by using paired sense and antisense
primers within these exons (D. J. Mancuso & R. W. Gross
unpublished observations). Promoter analyses utilizing
sequence upstream from exons 1 and 2 demonstrated that
the pre-exon 2 sequence strongly drives luciferase expres-
sion when placed upstream of the luciferase gene in a
transient expression system. The pre-exon 2 region analyzed
contains features of a t ypical core promoter witho ut a
TATA box including predicted GC boxes and a region of
homology to a n Inr recognition site. The authenticity of the
predicted Inr sequence as a site for binding of nuclear
proteins was supported by EMSA. It should be noted that
the 5¢-transcription start for messages beginning with exon 2
Ó FEBS 2004 Regulation of iPLA
2
c biosynthesis (Eur. J. Biochem. 271) 4721
may be variable based on differences in the extent o f exon 2
sequence from our 5¢-RACE analyses and from reported
ESTs in the database, which suggest that the 5¢-transcrip-
tional start site of iPLA
2
c may not be precisely regulated, a
feature that is c ommon in genes driven by TATA-less
promoters. Three lines of evidence suggested the p resence of
a second promoter upstream o f exon 1 including: ( a) the
identification of three E-boxes upstream of exon 1; (b) the
MyoD-dependent stimulation of promoter activity from
pre-exon 1 sequences; and (c) EMSA supporting the
conclusion that nuclear proteins transcriptionally modulate

iPLA
2
c. The role of MyoD in muscle differentiation is well
known and it is intriguing to speculate that Myo D utilizes
iPLA
2
c to effect alterations in mitochondrial bioenergetics
during muscle differentiation.
Peroxisomes originated as evolutionarily distant subcel-
lular organelles, whic h were initially de signed to orche strate
the metabolism of fatty acids i nto chemical energy o r lipid
storage in response to cellular metabolic energy require-
ments and each cell’s nutritional and metabolic history [64].
Although yeast contains both peroxisomes and mitochon-
dria, th e enzymes for fatty acid b-oxidation are present only
in the peroxisomal compartment in yeast [65]. Over time, the
enzymes for fatty a cid b-oxidation evolved to f acilitate more
efficient energy generation through integrating NADH
production with oxidative phosphorylation in m itochondria
[66]. In mitochondria, NADH is typically closely coupled to
ATP production maximizing ATP synthesis. In contrast, in
peroxisomes, fatty acid oxidation is an obligatorily uncou-
pled process by virtue of the fact that the first committed
step involves the transfer of electrons to molecular O
2
by
fatty acyl C oA oxidase (b-oxidation) or to O
2
by oxoglutryl
phytanoyl-CoA a-hydrox ylase ( a-oxidatio n). Our results

identify one mechanism which regulates the flow of the
Gibbs free energy inherent in C–C bonds in fatty acids into
the production of either chemical energy or heat by directing
the flow of iPLA
2
c into eith er the peroxisomal compartme nt
(expression of the 63 kDa isoform which does not possess a
mitochondrial localization signal) or into the mitochondrial
compartment (by production of isoforms containing a
mitochondrial localization sequence). W e cannot rule out
the possibility that other compartments also contain one or
more specialized forms of iPLA
2
c in lower amounts.
Recently, we have implicated a role for iPLA
2
c in obesity
by demonstrating that iPLA
2
c (aswellasiPLA
2
b)is
essential in hormone-induced 3T3-L1 cell differentiation
into adipocytes [67]. In these studies, iPLA
2
c message,
protein mass and activity increased during adipo genesis and
siRNA knockdown o f iPLA
2
c-inhibited adipogenesis [67].

Similarly, we demonstrated the dramatic up-regulation of
the iPLA
2
c 63 kDa isoform in adipose tissue of Zucker
obese rats [67]. These results suggest that iPLA
2
c biosyn-
thesis is highly regulated in regard to specific isoform
expression with important physiologic sequelae likely by
one or more mechanisms described here.
The relative amounts of c hemical e nergy g enerated by
mitochondrial metabolism o f fatty acids or energy d issipa-
ted by heat is modulated by the a ctivities of mitochondrial
uncoupling p roteins [ 15]. R ecent studies have also demon-
strated that the rate-determining step in UCP activity is the
interaction of fatty acids with UCP [16]. Moreover,
mitochondria bioenergetics is modulated by BEL-inhibita-
ble calcium-independent p hospholipases A
2
[15].Thus,it
seems clear that phospholipases have the capacity to both
generate fatty acid substrate for mitochondrial energy
production (through liberating fatty acids f rom membrane
phospholipids) and regulate their distinct metabolic outputs
(e.g. determine the relative amounts ATP generation vs.
heat production). The results from these studies provide a
novel framework to begin to understand the genetic,
biochemical and cell biological mechanisms involved i n
the coordinated regulation of peroxisomal and mitochond-
rial lipid metabolism, chemical energy production and

thermal energy generation in m ammalian tissue w hich will
further the understanding of the role of iPLA
2
sinenergy
storage and utilization disorders of lipid metabolism such as
diabetes, obesity and atherosclerosis.
Acknowledgements
This research was supported jointly by grants from the National
Institutes of He alth 5 PO1HL 57278-O7 and 5RO1HL41250-12.
References
1. Leslie, C.C. (1997) Properties and regulation of cytosolic phosp-
holipase A
2
. J. Biol. C hem. 272 , 16709–16712.
2. Balsinde, J. & Dennis, E.A. (1997) Function and inhibition of
intracellular calcium-independe nt phospholipase A
2
. J. Biol.
Chem. 272, 16069–16072.
3. Farooqui,A.A.,Yang,H C.,Rosenberger,T.A.&Horrocks,
L.A. (1997) Phospholipase A
2
and its role in brain t issue.
J. Neurochem. 69, 889–901.
4. Gross, R.W. (1998) Activation of calcium-independent phos-
pholipase A
2
by de pletio n o f internal calcium stores. Biochem. S oc.
Trans. 26, 345–349.
5. Ma, Z . & Turk, J. (2001) The molecular biology of the group VIA

Ca
2+
-independent phospholipase A
2
. Prog. N ucleic Acids R es.
Mol. Biol. 67, 1–33.
6. Wolf, R.A. & Gross, R.W. (1985) Identification of neutral active
phospholipases C which hydrolyzes choline glycerophospholipids
and plasmalogen se lective phospholipase A
2
in canine myo-
cardium. J. Biol. Chem. 260, 7 295–7303.
7. Hazen, S.L. & Gross, R.W. (1992) Identification and character-
ization of human myocardial ph ospholipase A
2
from transplant
recipients suffering from end s tage ischemic heart disease. Circ.
Res. 70, 486–495.
8. Hazen, S.L., Hall, C.R., Ford, D.A. & Gross, R.W. (1993)
Isolation of a human myocardial cytosolic phospholipase A
2
isoform: fast atom bombardment mass spectroscopic a nd re verse
phase HPLC identification of choline and ethanolamine
glycerophospholip id s ubstrates. J. Clin. Invest. 91, 2513–2522.
9. Hazen, S.L., F ord, D.A. & Gross, R.W. (1991) Activation of a
membrane-associated phospholipase A
2
during rabbit m yocardial
ischemia which is highly selective for plasmalo gen substrate.
J. Biol. Chem. 266, 5629–5633.

10. Ford, D.A., Hazen, S.L., Saffitz, J.E. & Gross, R.W. (1991) The
rapid and reversible ac tivation of a c alcium-inde pendent p lasma-
logen-selective phospholipase A
2
during myocardial ischemia.
J. Clin. Invest. 88, 331–335.
11. Gross, R.W. (1985) Identification of plasmalogen as the major
phospholipid constituent of cardiac sarcoplasmic reticulum. Bio-
chemistry 24, 1662–1668.
12. Hazen,S.L.,Zupan,L.A.,Weiss,R.H.,Getman,D.P.&Gross,
R.W. (1991) Suicide inhibition of canine myocardial cytosolic
calcium-independent phospholipase A
2
. J. Biol. Chem. 266, 7227–
7232.
4722 D. J. Mancuso et al. (Eur. J. Biochem. 271) Ó FEBS 2004
13. Zupan, L.A., W ei ss, R.H., Haze n, S.L., Parnas, B.L., Aston,
K.W., Lennon, R.J., Getman, D.P. & Gross, R.W. (1993) Struc-
tural determinants of haloenol lacton e-mediated su icide inhibition
of canine myocardial calcium-independent phospholipase A
2
.
J. Med. Chem. 36, 95–100.
14. Jenkins, C.M., Han, X., Mancuso, D.J. & Gross, R.W. (2002)
Identification of calcium-independent phospholipase A
2
(iPLA
2
)
beta, and not iPLA

2
gamma, as the mediator of argentine vaso-
pressin-induced arachidonic acid release in A-10 s mooth muscle
cells. Enantioselective mechanism-based discrimination o f mam-
malian iPLA
2
s. J. Biol. Chem. 277, 32807–32814.
15. Broekemeier, K.M., Iben, J.R., Levan, E.G., Crouser, E.D. &
Pfeiffer, D.R. (2002) Pore formation and uncoupling initiate a
Ca
2+
-independent d egradation o f m itochond rial pho spholipids.
Biochemistry 41, 7771–7780.
16. Urba
´
nkova
´
, E., Voltchenko, A., Pohl, P., Jezek, P. & Pohl, E.E.
(2003) Transport kinetics of uncoupling proteins. Analysis of
UCP1 reconstituted i n planar lipid bilayers. J. Biol. Chem. 27 8,
32497–32500.
17. Fink, K.L. & Gross, R.W. (1984) Modulation of canine myo-
cardial sarcolemmal membrane fluidity by a mphiphilic com-
pounds. Circ. Res. 55 , 585–594.
18. Gubitosi-Klug, R.A. & G ross, R.W. ( 1996) Fatty ac id ethyl esters,
non-oxidative metabolites of ethanol, accelerate the kinetics of
activation of the human brain delayed rectifier K
+
channel,
Kv1.1. J. Biol. Chem. 271, 32519–32522.

19. Gubitosi-Klug, R.A., Yu, S.P., C hoi, D.W. & Gross, R.W. (1995)
Concomitant acceleration of the activation and inactivation
kinetics of the human delayed rectifier K
+
channel ( Kv 1.1) by
Ca
2+
-independent phospholipase A
2
. J. Biol. Chem. 270, 2885–
2888.
20. Smani, T., Zakharov, S.I., C sutora, P., Le no, E., Trepakova, E.S.
& Bolotina, V.M. (2004) A novel mechanism for the store-oper-
ated calcium i nflux pathway. Nat. Cell Biol. 6, 113–120.
21. Hazen, S.L., Stuppy, R.J. & Gross, R.W. (1990) Purification a nd
characterization of canine myocardial cytosolic phospholipase A
2
:
a calcium inde penden t phospholipase with a bsolu te sn-2 regio
specificity for diradyl glycerophospholipids. J. Biol. Chem. 265,
10622–10630.
22. Larsson Forsell, P .K., Kennedy , B.P. & C laesson, H .E. (1999) The
human calcium-independent phospholipase A
2
gene. Multiple
enzymes with distinct properties from a single gene. Eur. J. Bio-
chem. 262, 575–585.
23. Wolf, M.J. & Gross, R.W. (1996) Expression, purifi cation, and
kinetic c haracterizatio n of a re combinant 80-kDa intracellular
calcium-independent phospholipase A

2
. J. Biol. Chem. 271,
30879–30895.
24. Tang, J., Kriz, R.W., Wolfman, N., Shaffer, M., Seehra, J. &
Jones, S.S. (1997) A novel c ytosolic calc ium-in dependent phos-
pholipase A
2
contains eight ankyrin motifs. J. Biol. C he m. 272,
8567–8575.
25. Stewart, A., Ghosh, M., Spencer, D.M. & Leslie, C.C. ( 2002)
Enzymatic properties of human cytosolic phospholipase A(2)
gamma. J. Biol. Chem. 277, 29526–29536.
26. Mancuso, D.J., Jenkins, C .M. & Gross, R.W. (2000) The genom ic
organization, c om plete mRNA sequence , cloning, and expression
of a novel human intracellular membrane-associated calcium-
independent pho spho lipase A (2). J. Biol. Chem. 275, 9937–
9945.
27. Yang, J., Han, X. & Gross, R.W. (2003) Identification of hepatic
peroxisomal phospholipase A(2) a nd characterization of arachi-
donic acid-containing choline glycerophospholipids in hepatic
peroxiso mes. FEBS Lett. 546, 247–250.
28. Andrews, D.L ., Bemes, B., Summers, M.D. & Park, W.D. (1 988)
Characterization of the lipid acyl hydrolase activity of the major
potato (Solanum tuberosum) tuber protein, patatin, b y cloning
andabundantexpressioninabaculovirusvector.Biochem. J. 252,
199–206.
29. Wolf, M.J. & Gross, R.W. (1996) The c alcium-dependent asso-
ciation and functional coupling of calmodulin with myocardial
phospholipase A
2

. Implications for cardiac cycle-dependent
alterations in phospholipolysis. J. Biol. Chem. 271, 20989–20992.
30. Jenkins, C.M., Wolf, M.J., Mancuso, D .J. & Gross, R.W. (2001)
Identification of the calmodulin-binding domain of recombinant
calcium-independent phospholipase A
2
beta. implications for
structure a nd fu nction. J. Biol. Chem. 276, 7129–7135.
31. Atsumi, G., Murakami, M., Kojima, K ., H ad an o, A., Tajima, M.
& Kudo, I. (2000) Distinct roles of two intracellular phospholipase
A
2
s in fatty acid release in the cell death pathway. Proteolytic
fragment of type IVA cytosolic phospholipase A
2
alpha inhibits
stimulus-induced arachidonate release, whereas that of type VI
Ca
2+
-independent phospholipase A
2
augments spontaneous fatty
acid release. J. Biol. Chem. 275, 18248–18258.
32. Atsumi, G., Tajima, M., Hadano, A., Nakatani, Y., Murakami,
M. & Ikudo, I. (1998) Fas-induced arachidonic acid release is
mediated by Ca
2+
-independent phospholipase A
2
but n ot cyto-

solic pho spholipase A
2
, w hich undergoes p roteolytic inactivation.
J. Biol. Chem. 273, 13870–13877.
33. Laemmli, U.K. (1970) Cleavage of structural proteins during the
assembly of th e he ad of b ac teriophage T4. Nature 227, 680–685.
34. Kozak, M. (1996) Inte rpreting cDNA sequences: some i nsights
from studies on t ranslation. Mamm. Genome 7, 563–574.
35. Kozak, M. (1987) An analysis of 5¢-noncoding sequences from
699 vertebrate messenger RNAs. Nucleic Acids Res. 15, 8125–
8148.
36. Chin, M.T., Maemura, K., Fukumoto, S., Jain, M.K., Layne,
M.D., Watanabe, M ., Hsieh, C M. & Lee, M E. (2000) Cardio-
vascular basic helix loop h elix factor 1, a nove l transcrip tional
repressor expressed preferentially in the developing and adult
cardiovascula r s ystem. J. Biol. Chem. 275, 6381–6387.
37. Claros, M.G. & Vince ns, P. (1996) Computational method t o
predict mitochondrially imported proteins and their targeting
sequences. E ur. J. Biochem. 241, 779–786.
38. Luckow, V.A. & Summers, M.D. (1988) Signals i mportant for
high-level expression of foreign genes in Autographa californica
nuclear polyhedrosis virus expression vectors. Viro logy 167 , 56–7 1.
39. Luckow, V.A. & Summers, M.D. (1989) High level e xpression of
nonfused foreign genes with Autographa californica nuclear poly-
hedrosis virus expression vectors. Virology 17 0 , 31–39.
40. Mancuso, D.J., Abendschein, D.R., Jenkins, C.M., Han, X.,
Saffitz, J.E., Schuessler, R.B. & Gross, R.W. (2003) Cardiac
ischemia activates calcium-independent phospholipase A
2
beta,

precipitating ventricular tachyarrhythmias in t ransgenic mice:
rescue of the lethal electrophysiologic phenotype by mechanism-
based inhibition. J. Biol. Chem. 278, 22231–22236.
41. Zeng,Y.,Han,X.,Schlesinger,P.&Gross,R.W.(1998)Non-
esterified fatty acids induce transmembrane monovalent cation
flux: host–guest interactions as determinants of fatty acid-induced
ion transport. Bioc h emi stry 37, 9497–9508.
42. Larsson Forsell, P .K., Kennedy , B.P. & C laesson, H.E. (1999) The
human calcium-independ ent phosp holipase A
2
gene multiple
enzymes with distinct properties from a single gene. Eur. J.
Biochem. 262, 575–585.
43. Ma,Z.,Wang,X.,Nowatzke,W.,Ramanadham,S.&Turk,J.
(1999) Human pancreatic islets express mRNA species encoding
two distinct catalytically active isoforms of group VI phospho-
lipase A
2
(iPLA
2
) that arise from an exon-skipping m echanism of
alternative splicin g o f the transcript from the iPLA
2
gene on
chromosome 22q13.1. J. Biol. Chem. 27 4, 9607–9616.
44. Larsson, P.K., Claesson, H.E. & Kennedy, B.P. (1998) Multiple
splice variants of the human calcium-indepen dent phospholipase
A
2
and t heir effect on e nzyme activity. J. Biol. Chem. 273, 2 07–214.

Ó FEBS 2004 Regulation of iPLA
2
c biosynthesis (Eur. J. Biochem. 271) 4723
45.Tanaka,H.,Takeya,R.&Sumimoto,H.(2000)Anovel
intracellular membrane-bound c alciu m-indepen dent phospho-
lipase A(2). Biochem. Biophys. Res. Commun. 272, 320–326.
46. Verrijzer, C.P., Chen, J L., Yokomori, K. & Tjian, R. (1995)
Binding of TAFs to core e lements directs promoter selectivity b y
RNA polymerase I I. Cell 81 , 1115–1125.
47. Arnold, H.H. & Winter, B . (1998) Muscle differentiation: more
complexity to the network of myogenic regulators. Curr. Opin.
Genet. Dev. 8, 539–544.
48. Mckinsey, T.A., Zha ng, C.L. & Olson, E.N. (2001) Control o f
muscle development by dueling HATs and HDACs. Curr. Opin.
Genet. Dev. 11, 497–504.
49. Puri, P.L. & Sartorelli, V. (2000) R egulation of muscle regulatory
factors by DNA-binding, interacting proteins, and post-tran-
scriptional modifi cations. J. Cell Physiol. 185, 155–173.
50. Lehman, J.J., Brown , K .A., Raman adham, S., Turk, J. & Gross,
R.W. (1993) Arachidonic a cid release from aortic smooth muscle
cells induced by [Arg8]vasopressin is largely mediated by calcium-
independent pho spho lipase A
2
. J. Biol. Chem. 268, 20713–20716.
51. Ramanadham, S., Gross, R.W., Han, X. & Turk, J. (1993)
Inhibition of arachidonate release by secretagogue-stimulated
pancreatic islets suppresses both insulin secretion and the rise in
beta-cell cytosolic calcium i on concentration. Biochemistry 32 ,
337–346.
52. Ma, Z., Ramanadham, S., W ohltmann, M., Bohrer, A., Hsu, F.F.

& Turk, J. (2001) Studies of insulin secretory responses and of
arachidonic acid incorporation into phospholipids of stably
transfected insulinoma cells that overexpress group VIA phos-
pholipase A
2
(iPLA2beta) indicate a signaling rather than a
housekeeping role for iPLA2beta. J. Biol. Chem. 276, 13198–
13208.
53. Seegers, H.C., Gross, R.W. & Boyle, W.A. (2002) Calcium-
independent phospholipase A(2)-derived arachidonic a cid is
essential for endothelium-dependent r elax ation by acetylcholine.
Exp. Therap. 302, 918–923.
54. Ogbourne, S. & Antalis, T.M. (1998) Transcriptional control a nd
the role of silencers in transcriptional regulation in eukaryotes.
Biochem. J. 331, 1–14.
55. Morel, N. & Massoulie
´
, J. ( 2000) Comparative expression of
homologous proteins. A novel mo de of transc riptional regulation
by the coding sequence folding compatibility of chimeras. J. Biol.
Chem. 275, 7304–7312.
56. Paste
´
, M., Huez, G. & Kruys, V. (2003) Deadenylation of inter-
feron-beta mRNA is mediated by both t he AU-rich e lemen t in th e
3¢-untranslated region and an inst ability sequence in the coding
region. Eur. J. Biochem. 270, 1590–1597.
57. Fallaux, F.J., Hoeben, R .C., Cramer, S.J., van den Wollenberg,
D.J., Briet, E., van Ormondt, H. & van der Eb, A.J. (1996) The
human c lottin g factor VIII cDNA c ontains an auto nomously

replicating sequence consensus- and matrix attach ment region-lik e
sequence that bin ds a nuclear factor, represses h eterologo us gen e
expression, a n d mediates the transcriptional effects of sodium
butyrate. Mol. Cell. Biol. 16, 4264–4272.
58. Xiang,S.,Parsons,H.K.&Murray, M. (1998) Identific ation of a
novel transcriptional silencer in the protein-coding region of the
human CYP2C9 gene. Gene. 209, 123–129.
59. Willis, A .E. (1999) Translational control of growth fa ctor
and proto-oncogene expression. Int. J. Biochem. Cell. Biol. 31,
73–86.
60. Leff, S.E. & R osenfeld, M .G. ( 1986) Complex transcriptional
units: divers ity in gene expression by alternative RNA processing.
Annu. Rev . B iochem . 55, 1091–1117.
61. Breitbart,R.E.,Andreadis,A.&Nadal-Ginard,B.(1987)Alter-
native splicing: a ubiquitous mechanism for the g enerat ion of
multiple protein i soforms f rom s ingle g enes. Annu. Rev. Biochem.
56, 467–495.
62. Proudfoot, N.J., Furger, A. & Dye, M.J. (2002) Integrating
mRNA processing with transcription. Cell 108, 501–512.
63. Maniatis, T. & Tasic, B. (2002) Alternative pre-mRNA splicing
and proteome expansion in metazoans. Nature 418, 236–243.
64. Latruffe, N. & Vamecq, J. (2000) Evolutionary aspects of per-
oxisomes as cell organelles, and of gene s en co ding p eroxisomal
proteins. Biol. Ce ll 92, 389–395.
65. van d er Klei, I.J. & Veenhuis, M. (1997) Yea st peroxisomes:
function and biogenesis of a versatile cell organelle. Trends
Microbiol. 5, 5 02–509.
66. Mannaerts, G.P. & van Veldhoven, P.P. (1996) Functions and
organization of peroxisomal beta-oxidation. Ann. NY Acad. Sci.
804, 99–115.

67. Su, X., Mancuso, D.J., Bickel, P.E., J enkins, C .M. & Gro ss, R.W.
(2004) Small interfering RNA knockdown of calcium-indepen-
dent phospholipase A
2
or inhibits the hormo ne-indu ced
differentiation of 3T3-L1 preadipocytes. J. Bi ol. Chem. 279,
21740–21748.
4724 D. J. Mancuso et al. (Eur. J. Biochem. 271) Ó FEBS 2004