PROTEINS: Structure, Function, and Genetics 39:1– 8 (2000)

RESEARCH ARTICLES

Homologs of the ␣- and ␤-Subunits of Mammalian Brain

Platelet-Activating Factor Acetylhydrolase Ib in the
Drosophila melanogaster Genome
Peter J. Sheffield,1 Sarah Garrard,1 Michal Caspi,2 Junken Aoki,3 Hiroyuki Arai,3 Urszula Derewenda,1
Keizo Inoue,3 Beat Suter,4 Orly Reiner,2 and Zygmunt S. Derewenda1*
1
Department of Molecular Physiology and Biological Physics, University of Virginia, Health Sciences Center,
Charlottesville, Virginia
2
Department of Molecular Genetics, The Weizmann Institute of Science, Rehovot, Israel
3
Department of Health Chemistry, Faculty of Pharmaceutical Sciences, University of Tokyo, Japan
4
Department of Biology, McGill University, Montreal, Canada

ABSTRACT The mammalian intracellular brain
platelet-activating factor acetylhydrolase, impli-
cated in the development of cerebral cortex, is a
member of the phospholipase A2 superfamily. It is
made up of a homodimer of the 45 kDa LIS1 protein
(a product of the causative gene for type I lissen-
cephaly) and a pair of homologous 26-kDa ␣-sub-
units which account for all the catalytic activity.
LIS1 is hypothesized to regulate nuclear movement
in migrating neurons through interactions with the
cytoskeleton, while the ␣-subunits, whose structure
is known, contain a trypsin-like triad within the
framework of a unique tertiary fold. The physiologi-
cal significance of the association of the two types of
subunits is not known. In an effort to better under-
stand the function of the complex we turned to
genomic data mining in search of related proteins
in lower eukaryotes. We found that the Droso-
phila melanogaster genome contains homologs of
both ␣- and ␤-subunits, and we cloned both genes.
The ␣-subunit homolog has been overexpressed,
purified and crystallized. It lacks two of the three
active-site residues and, consequently, is catalyti-
cally inactive against PAF-AH (Ib) substrates. Our
study shows that the ␤-subunit homolog is highly
conserved from Drosophila to mammals and is able
to interact with the mammalian ␣-subunits but is
unable to interact with the Drosophila ␣-subunit.
Proteins 2000;39:1– 8. © 2000 Wiley-Liss, Inc.
Key words: PAF-AH(Ib); LIS1; ␣-subunit
continues until cortical layers of neurons are fully laid out
from the deepest to the outermost. Thus, the younger
generations of neurons must negotiate their way through
the layers of older cells. The mechanisms, which control
and regulate this elaborate and complex phenomenon, are
not well understood. Over 25 different syndromes caused
by defects in neuronal migration have been described in
man.1 Most are genetic in nature and affect the cortex
during its development. Perhaps the most dramatic in
phenotype is lissencephaly, a severe developmental disor-
der which results in a virtually smooth cerebral surface
with conspicuous absence of typical cortical convolutions.2
When the condition is associated with dysmorphic facial
appearance, it is known as the Miller-Dieker syndrome.3
The causal gene for this disorder, LIS1, has been identified
and characterized.4,5 It encodes a 45-kDa polypeptide
chain with homology to various members of the WD-repeat
family, a group of predominantly regulatory proteins
which include, among others, the ␤-subunits of trimeric
G-proteins.6 Close homologs of LIS1 have been found in
organisms as diverse as Aspergillus nidulans and Cae-
norhabitis elegans, 7,8 suggesting an important role in a
conserved regulatory pathway.
Interestingly, the LIS1 gene product co-purifies with a
unique dimeric intracellular isoform of phospholipase A2,
the platelet-activating factor acetylhydrolase.9 The entire
oligomeric complex is known in literature as the PAF-
AH(Ib). The substrate of this enzyme, platelet-activating
factor (PAF, 1-O-alkyl-2-acetyl-sn-glycero-3-phosphocho-
line) is a potent phospholipid messenger found in verte-

INTRODUCTION
The six-layered cerebral cortex is the seat of all cognitive
functions in mammals. It is formed in the developing fetus
by progenitor neurons, which originate from specialized
regions deep in the brain and migrate over large distances,
up to a thousand times the length of the cell. This process
Grant sponsor: National Institutes of Health; Grant number:
NS36267; Grant sponsor: US-Israel Binational Science Foundation;
Grant sponsor: National Cancer Institute of Canada.
*Correspondence to: Zygmunt S. Derewenda, Department of Molecular
Physiology and Biological Physics, University of Virginia, Health Sci-
ences Center, Charlottesville, VA 22908. E-mail: zsd4n@virginia.edu.
Received 8 June 1999; Accepted 2 November 1999

© 2000 WILEY-LISS, INC.

2 P.J. SHEFFIELD ET AL.
brates, as well as among animals belonging to lower
phyla.10 How PAF may be involved in the development of
the cerebral cortex is not understood.
In man, PAF acts primarily as a mediator of inflamma-
tory and allergic reactions. It causes microvascular leak-
age, smooth muscle contraction, endothelial adhesion, and
general cell activation and has been implicated in asthma,
HIV pathogenesis, carcinogenesis, and apoptosis.11–15 It is
known to play an essential role in neurophysiology as a
synapse messenger, inducer of the early-response genes, in
ischemia, seizures and long-term potentiation.16 –19 Consis-
tent with the diversity of its biological functions, PAF is
found both in plasma and in cell cytosol throughout the
human body. Intra- and extracellular levels of the phospho-
lipid are tightly regulated at the point of biosynthesis as
well as hydrolytic degradation. The latter is catalyzed by a
diverse family of plasma and cytosolic enzymes (PAF-
acetylhydrolases), which remove the acetyl group from the
sn-2 position and can therefore be classified as phospho-
lipases A2.20,21 One of these enzymes is the intracellular
PAF-AH(Ib), the physiological function of which is not
understood.9,22,23
As already pointed out, the PAF-AH(Ib) is an oligomeric
complex. As originally purified from bovine brain, it con-
tained a dimer of two homologous catalytic subunits ␣1
and ␣2 (63% amino acid sequence identity), which are not
related to any other known proteins.9,22 Both are 26-kDa,
single polypeptide chain proteins. It is now known (see
below) that ␣1 and ␣2 homodimers may also be found in the
complex. When expressed individually in Escherichia coli,
the ␣-subunits form homodimers. Recently, the crystal
structure of the bovine ␣1 homodimer has been determined
at 1.7 Å resolution.24 The molecular structure underscores
the uniqueness of the protein, which has a tertiary fold
reminiscent of small GTPases and harbors a chymotrypsin-
like Ser-His-Asp triad in its active site. Interestingly, the
expression patterns of ␣1 and ␣2 appear to be regulated
independently during development, so that in rat ␣1 is
expressed exclusively in fetal cells, while ␣2 is present in
both fetal and mature neurons. In man and cow however,
␣1 can also be detected in the adult brain (Manya et al.25;
O. Reiner, unpublished data). These observations suggest,
that both LIS1 (the ␤-subunit) and the ␣1-subunit function
synergistically during cortex development, although the
details are not known.
There is circumstantial evidence that LIS1 may be
involved in the regulation of nuclear movement in the
migrating neurons. The only well characterized non-
mammalian homolog of LIS1 is the fungal NudF gene
product (Aspergillus nidulans), a molecule shown to con-
trol the migration of fungal nuclei into the growing germ
tube after they undergo division.8 Recently, LIS1 has been
shown to associate with a mammalian homolog of NudC,
another protein in the same pathway.26 Furthermore,
LIS1 interacts with tubulin and affects microtubule dynam-
ics.27 This evidence appears to link LIS1 logically to the
cytoskeleton. In contrast, the biological role of the catalytic
subunits in the PAF-AH(Ib) complex remains a complete
enigma. Their putative role in neuron migration and
cortex development has no experimental support, except
for two reports that inhibition of the PAF-AH activity may
affect neuronal migration.28,29
Given the uncertainties surrounding the structure-
function relationships in the PAF-AH(Ib) complex, we
turned to genomic data mining in an effort to find homo-
logu of the ␣- and ␤-subunits, whose sequences might
reveal new information. Further, we thought it would be a
considerable advantage to use a simpler organism to probe
the role of the PAF-AH(Ib) complex in development. Using
the Berkeley Drosophila melanogaster genome database
(http://www.fruitfly.org) we identified two candidate DNA
clones. One of these (EST clone LD20496) contained a
hitherto unrecognized ⬃300-bp DNA fragment that, in one
of the reading frames, showed 44% identity with the
␣-subunit. The other (genomic P1 clone DS07424), ap-
peared to code for a homolog of LIS1 as judged by putative
exon sequences. To obtain complete protein sequences we
cloned both genes from an embryonic D. melanogaster
library. Further more, we studied the developmental
expression profiles of the proteins, and we overexpressed,
purified, and crystallized the ␣-subunit homolog. The
results shed new light on the possible evolutionary origins
of the PAF-AH(Ib) complex.
MATERIALS AND METHODS
Bacterial Strains and Plasmids
pBluescript (Stratagene, La Jolla, CA) was used in all
cDNA cloning and sequencing experiments. E. coli strain
XL1Blue (Stratagene) was used for DNA subcloning and
DNA manipulations. The GST expression construct, pGST-
parallel1,30 was expressed in strain BL21.
Cloning of the D. melanogaster ␣-Subunit
Homolog (Dm␣PAFAH)
An embryonic cDNA library (12–24 hr) was supplied by
Dr. P. Adler (Dept. of Biology, University of Virginia). The
cDNA’s were unidirectionally cloned into pNB40 (created
by N. Brown) and maintained in strain DH5␣. Two prim-
ers specific for EST clone LD20496 were designed: primer
A, [5⬘CCCTGTGTGCTGCCCACCCCATTGCCCG 3⬘] and
primer B, [5⬘GCGAATGCTGAAGTTAAGGCAGTGCAG-
CGGCGC 3⬘]. These primers generated a 201-bp amplicon
identical to clone LD20496. Three primers, two proximal
and one distal to the cDNA insertion point in pNB40
vector, were also designed: proximal primers: NB40-1, [5⬘
GGCTTGTACATATTGTCGTTAGAACGCGGC 3⬘], NB40-2,
[5⬘GGTTCGCTGACCATTTCCGGGTGCGGGACGGCG
3⬘], and the distal primer: NB40-3, [5⬘CGTTGGCCGAT-
TCATTAATGCAGCTGGC 3⬘]. The universal T7 primer
[5⬘AATACGACTCACTATAG 3⬘], which bound distal to
the insertion point, was also used. Using these primers
and the polymerase chain reaction (PCR), amplicons
greater than 690 bp in length (the minimum length
required to encode a 230-residue protein, equivalent to the
mammalian ␣ proteins) were isolated and cloned into
pBluescript. A 950-bp amplicon was consistently produced
from several PCR experiments and this was shown to
contain the entire Dm␣PAFAH gene. Several Dm␣PAFAH
 PAF-ACETYLHYDROLASE SUBUNIT HOMOLOGS IN D. MELANOGASTER
clones were identified and five, isolated from different PCR
reactions, were sequenced (ABI automated DNA se-
quencer). DNA sequences were analyzed and a contiguous
sequence was assembled using the Wisconsin GCG suite of
programs (GenBank accession number AF098069).
Cloning of D. melanogaster LIS1-homolog (DmLIS1)
Translation of the Drosophila genomic P1 clone, DS07424
(GenBank accession AC004320) allowed a putative cDNA/
protein sequence to be assembled. Two oligonucleotides
specific for the 5⬘ and 3⬘ ends of the putative cDNA were
designed [N-terminal primer: 5⬘ CAGTATGAccATGGTGT-
TGTCGCAGCGGCAGCGCG 3⬘ and C-terminal primer: 5⬘
CTTCCTTAATTTAACGACATTCCCAGACCTTTACTG
3⬘]. The N-terminal primer contained a two base pair
mutation to introduce an Nco I site. A 1.2-kb amplicon was
isolated by PCR and subcloned into pBluescript. Several
clones were identified and sequenced. The sequence data
was analyzed and a contiguous sequence was assembled
using the Wisconsin GCG package. The sequence matched
the putative cDNA sequence, with only a few differences
observed at the predicted exon junctions (GenBank acces-
sion number AF098070).
Expression, Purification and Crystallization
of the Dm␣PAFAH
An Nco I site and a Xho I site were introduced into
Dm␣PAFAH by site directed mutagenic PCR, Nco I primer
[5⬘ GGGCTGCAGGCCATGGATCCCTGTGTGC 3⬘] and
Xho I primer [5⬘ GGATAGCTGCTCGAGATTCCTTTTAT-
TCGG 3⬘]. The resultant 717-bp amplicon was subcloned
into pGST-parallel130 as an Nco I/Xho I fragment and
clones verified by DNA sequencing. pGSTpara1::
Dm␣PAFAH was transformed into E. coli strain BL21 and
expression experiments were done in liquid culture at
37°C in 1–2 L of Luria broth (ampicillin, 50 ␮g/ml).
Expression was induced at an O.D.600nm of 0.4 – 0.6 by the
addition of IPTG (final conc. 0.5 mM) and growth contin-
ued for a further 4 h at 37°C. The cells were harvested and
re-suspended in lysis buffer I (50 mM Tris-Cl pH 8.5, 300
mM NaCl, 1ml/gram cell pellet) and lysed by sonication.
Soluble protein extracts were recovered and GST fusion
protein was purified by glutathione sepharose 4B affinity
chromatography. Briefly, protein extracts were bound to
pre-equilibrated glutathione sepharose 4B (Pharmacia,
Gaithersburg, MD) columns for 2– 4 h at room tempera-
ture, drained and washed with 2 liters lysis buffer I at 4°C.
Fusion protein was eluted using 20 ml buffer II (50 mM
Tris-Cl, pH 8.0, 10 mM reduced glutathione). The GST tag
was removed from Dm␣PAFAH by treatment with rTEV
protease (10°C for 24 –36 h, Life Technologies, Bethesda,
MD). Excess glutathione was removed by dialysis against
2 L of buffer I and the GST tag removed by passing
through a second glutathione sepharose column. Pure
Dm␣PAFAH (⬎ 95%) was obtained after size exclusion
chromatography, performed on a BioCad Sprint system
(PerSeptive Biosystems) using a Superdex 75 column
(Pharmacia). A standard Tris buffer (50 mM Tris-Cl pH
8.0, 200 mM NaCl) was used throughout the purification.
3
Crystallization experiments were performed by the hang-
ing-drop vapor-diffusion method at 21°C using Linbro
24-well tissue culture plates. Initial screening was done
using the Hampton Research crystal screen I and 2 with
protein concentration of 8 mg/ml in 25mM Tris-HCl, pH
8.5. Protein and precipitant were mixed in a ratio of 2:1.
PAF Acetylhydrolase Activity Assay
Recombinant bovine ␣1 and ␣2 proteins, as well as the D.
melanogaster homolog, were expressed and purified from
E. coli as described previously28 and above, in fusion with
either a His- or GST- tag. The standard incubation system
for the assaying of PAF acetylhydrolase comprised 50 mM
Tris-HCl (pH 7.4), 5 mM EDTA, 20 nM of [3H-acetyl] PAF
(DuPont-New England Nuclear, Boston, MA), and the
recombinant enzymes in a total volume of 125 mL. After 30
min at 37°C the reaction was stopped by adding 1.25 mL of
chloroform/methanol (4:1) and 125 mL of water. Aliquots
(300 mL) of the upper phase were used for radioactivity
measurement by ␣-counter (Pharmacia BioTech) to deter-
mine the amount of liberated acetate.
Northern Blots
These were prepared as previously described.31
Binding of the DmLIS1 to the Human ␣-Subunits
The yeast two-hybrid system was employed to analyze
putative interactions of the DmLIS1 protein. The LexA
protein (pEG-202 plasmid) was fused by genetic recombina-
tion to two following baits: hLIS1, DmLIS1, Dm␣PAFAH ,
bovine ␣1 and ␣2. The hLIS1, rat ␣-subunits and Dm␣PAFAH
were also fused to the B42 activation domain (pJG4-5
plasmid) in order to test all possible cross interactions.
Yeast transformation was performed according to the
manual (OriGene Technologies, Inc.). Detection of the
different interactions was done using the X-gal assay.
RESULTS AND DISCUSSION
The D. melanogaster Genome Contains a Homolog
of Human LIS1
A search of the Drosophila genome database (Berkeley)
identified a previously reported genomic P1 clone which,
when translated to the three possible frames, appeared to
contain a gene coding for a LIS1 homolog. From this
sequence, allowing for the presence of putative introns, it
was possible to predict a 1.2 kb cDNA encoding a 410
residue protein that had a similarity of 79.6% and an
identity of 70.3% when compared to hLIS1. Our direct
cDNA cloning confirmed this prediction: the inferred amino
acid sequence matched the hypothetical sequence, with
the exception of a few residues located at the predicted
exon junctions (GenBank accession code AF098070). The
high similarity between the protein and gene structures of
LIS1 in human and Drosophila is noteworthy. A previ-
ously reported homolog of LIS1 from Aspergillus nidulans,
the product of the NudF gene, also shows a high level of
amino acid similarity (60% similarity and 43% identity)
with the DmLIS1, suggesting that the protein has been
highly conserved throughout evolution. Using a BLAST
4 P.J. SHEFFIELD ET AL.
Fig. 1. Amino acid sequence alignment for the Drosophila and bovine
␣-subunits. The secondary structure elements, as inferred from the
crystal structure of the bovine ␣1-subunit, are shown above the lineup.
The stars above the lineup show the identities between ␣1 and Dm␣,
whereas those beneath the lineup show the identities between ␣2 and
search against genomic databases, we also identified a
close homolog of LIS1 in a Zebra fish EST clone (95%
similarity and 94% Identity, accession No. AI964975) and
in C. elegans (68% similarity and 56%, accession No.
CAB03282) but no homolog was found in prokaryotes.
NudF has been shown as essential for the migration of
nuclei in Aspergillus nidulans. LIS1 is hypothesized to
exert its physiological function during cortex development
by controlling the migration of nuclei in neurons. It will be
of great interest to see if similar functions can be identified
for the D. melanogaster and C. elegans homologs. If the
function of the LIS1-like protein is distinctly associated
with the migration of nuclei, it is not surprising that we do
not find them among prokaryotic organisms. [Since accep-
tance of this paper the cloning of the DmLIS1homolog has
been reported by others. Liu Z, Xie T, Steward R. Lis1, the
Drosophila homolog of a human lissencephaly disease
gene, is required for germline cell division and oocyte
differentiation. Development 1999;126:4477– 4488.]
A Homolog of the ␣-Subunit of PAF-AH(Ib) Is Also
Found in D. melanogaster Genome
Our genomic data mining also revealed the presence in
the D. melanogaster genome of a homolog of the ␣-subunit.
To date, there are no reports in the literature of homologs
of this protein, other than from mammals such as mouse
Dm␣. The boxes show amino acids involved in the dimerization of the
bovine ␣-subunits, and the arrows show the residues which in the bovine
protein form the catalytic triad. Structural elements H0 ⫺ H5 ⫽ alpha
helices, S1 ⫺ S6 ⫽ beta sheet.
and rat. Based on the Drosophila EST clone LD20496, we
isolated and sequenced a full-length cDNA clone of the
Drosophila melanogaster PAF-AH(Ib) ␣-subunit homolog
(GenBank accession AF098069). The DNA sequence over-
lapped with the Drosophila genomic EST clone LD20496
and confirmed that LD20496 was the 5⬘ end of the coding
region, containing the initiation ATG codon. The predicted
amino acid sequence shared 43% identity with the bovine
PAF-AH(Ib) ␣1 and 47% identity with PAF-AH(Ib) ␣2 (Fig.
1). No other clones showing homology to the ␣-subunits
were isolated. This could suggest that Drosophila, unlike
higher eukaryotes, harbors only one gene for this protein,
although no low-stringency blot has been done to confirm
this. Once the Drosophila genome project has been com-
pleted and made available it will be possible to identify
other homologs using bioinformatics. Until then we cannot
discount the existence of other homologs of the PAF-
AH(Ib) ␣-subunits. However, an extensive BLAST search,
of GenBank and the existing Flybase, identified only 7
PAF-AH(Ib) ␣-subunits homologs, all mammalian. Thus,
D. melanogaster is the first known invertebrate harboring
genes coding for both types of the subunits found in the
mammalian brain PAF-AH(Ib). The obvious questions
that arise from this observation are: is the Dm␣PAFAH a
dimeric hydrolase like its mammalian counterparts, is it
 PAF-ACETYLHYDROLASE SUBUNIT HOMOLOGS IN D. MELANOGASTER 5

Fig. 2. SDS gel of Dm␣PAFAH purification from

pGST-paral::Dm␣PAFAH M ⫽ Markers, 1 ⫽ total cell
protein uninduced, 2 ⫽ total cell protein induced, 3 ⫽
soluble protein fraction, 4 ⫽ flow through, 5 ⫽ GST-
Dm␣PAFAH elution, 6 ⫽ rTEV treated (30 kDa band ⫽
doublet made up of GST and Dm␣PAFAH ) and 7 ⫽ pure
Dm␣PAFAH after size-exclusion chromatography.

co-expressed with DmLIS1, and does it interact with it to

form a complex?

The D. melanogaster ␣-Subunit Homolog Has No

Enzymatic Properties
In spite of the significant sequence similarity between
Dm␣PAFAH and its mammalian homologs, the active site
residues are not conserved. The nucleophilic Ser47 (hu-
man ␣1 numbering) is replaced by a cysteine, and His195
of the catalytic triad is replaced by an asparagine (this
result has been observed in several independent clones).
Moreover, one of the essential residues in the acetyl-
recognition pocket, Leu194 is replaced by a lysine. In
contrast, most of the residues involved in the dimer
interface in the mammalian proteins are conserved in the
D. melanogaster homolog. We have therefore inferred from
the sequence that the protein would not show any catalytic
activity using PAF as a substrate. To test this interpreta- Fig. 3. Activity assay for Dm␣PAFAH. (1) GST-Dm␣PAFAH, (2) His-
Dm␣PAFAH, (3) bovine GST-PAFAH ␣1, (4) bovine His-PAFAH ␣1. Details
tion, the protein was over-expressed in E. coli with high of the assay are given in the text.
yield and purified to homogeneity paving a way for struc-
tural and functional analysis (Fig. 2). This allowed us to
carry out activity assays, which showed that the protein is
indeed inactive against PAF as a substrate (Fig. 3). On the since been grown in 1.6 –1.8 M (NH2)4SO4, 4 –9% dioxane,
other hand, and in contrast to our expectations, gel 4% PEG400, 100 mM MES, pH 6.8 by mixing the protein
filtration suggested a predominantly monomeric character with the precipitant at a ratio of 2:1 respectively. Prelimi-
of the molecule, unlike the mammalian proteins (Fig. 4). nary data were collected from a frozen crystal to 2.5 Å
We therefore conclude, that in Drosophila, the protein resolution at beamline X9B of the National Synchrotron
plays a non-catalytic role and that the mammalian protein Light Source (Brookhaven National Laboratory). Crystals
may also have a primary physiological function that is not are orthorhombic with a F222 space group, with a unit cell
catalytic and that the PAF-acetylhydrolase property ac- of a ⫽ 95.3 Å, b ⫽ 97.4 Å c ⫽ 137.0 Å. The total number of
quired independently during evolution may be a modulat- observations was 26,357, reduced to 10,303 unique reflec-
ing or regulatory function. tions (85.1% completeness) with an overall Rmerge of 2.6%.
The specific volume is 3.06 Å2/Da, assuming one molecule
Dm␣PAFAH Has Been Successfully Crystallized per asymmetric unit. Structure solution is in progress and
for X-Ray Studies will be reported elsewhere.
In order to pave the way for structural characterization
The Two D. melanogaster Gene Products Do Not
of Dm␣PAFAH, we carried out crystallization experiments,
Interact With Each Other
as described in Materials and Methods. Micro-crystals
formed after 4 –5 days in 1.6 M (NH2)4SO4, 10% dioxane, We then set out to determine if DmLIS1 and Dm␣PAFAH
100 mM NaOAc, pH 6.5. Larger crystals (10 ⫻ 70 ␮m) have are co-expressed and interact in Drosophila in a way
6 P.J. SHEFFIELD ET AL.
Fig. 4. Size exclusion profile for Dm␣PAFAH. The experiment was
conducted using a pure Dm␣PAFAH sample isolated as a monomeric
fraction from a previous size exclusion chromatography run.
reminiscent of mammals. We probed developmental RNA
blots with the two cDNAs for both proteins (Fig. 5). The
␣-subunit gene gives rise to a major and a minor transcript
of similar size (1.4 and 1.25 kb; Fig. 5) while DmLIS1
encodes four different transcripts with distinct transcript
accumulation patterns. One transcript of 2.8 kb is ex-
pressed in most or all stages of the life cycle, two smaller
transcripts of 1.6 and 1.8 kb accumulate predominantly
during oogenesis and appear to be deposited into the egg
where they get degraded before or around the blastoderm
stage. Based on Northern blot hybridizations with differ-
ent genomic fragments as probes, it appears that these
three types of transcripts differ primarily in their 3⬘UTR
(data not shown). In addition, a 3.2-kb transcript accumu-
lates predominantly in the adult head fraction and in
males. Comparing the two expression patterns reveals
that the two genes are co-expressed during oogenesis, the
first four hours of embryogenesis and during pupation
(lL3, eP, and possibly lP). However, there are phases and
tissues that predominantly accumulate DmLIS1 mRNA
with little or no mRNA encoding the ␣-subunit. The zygotic
genome is activated after approximately four hours of
embryogenesis. In the subsequent phase from 4 –20 hours,
DmLIS1 mRNA accumulation increases while the ␣-sub-
unit mRNA levels continue to drop. The most striking
difference, however, is the very strong accumulation of
DmLIS1 in adult heads where we do not detect mRNAs
encoding the ␣-subunit.
Given that a significant level of co-expression has been
observed—although our experiments did not address the
issue of co-localization—we employed the two-hybrid yeast
system to test the possibility of DmLIS1 forming a complex
with Dm␣PAFAH. We also included the mammalian pro-
teins in the assays to test for inter-species dimerization.
Perhaps the most exciting outcome of our study is the
demonstration that in the yeast two-hybrid system the D.
melanogaster LIS1 protein interacts with the mammalian
␣1 and ␣2 subunits (Fig. 6A), but not with the D. melano-
gaster homolog (Fig. 6B). It has been observed that LIS1
interacts in vitro with several proteins other than the
␣-subunits of the brain PAF-AH(Ib), including PH do-
mains,32 NudC,26 tubulin,27 and co-purifies with the spleen
protein tyrosine kinase TPK-IIB.33 Which of these interac-
tions is physiologically relevant is not known, but the
observation that those epitopes on LIS1 which are in-
volved in the PAF-AH(Ib) ␣-subunit binding are conserved
in the D. melanogaster protein, suggests that the PAF-
AH(Ib) complex is indeed functionally important. What is
difficult to interpret at this point, is the inability of the
DmLIS1 to associate with the ␣-subunit from the same
species. We can only speculate that the evolution of the
PAF-AH activity was in some way linked to the ability of
the ␣-subunit to associate with the LIS1 protein.
Fig. 5. Autoradiograph of a single poly(A)⫹ RNA blot
that was hybridized with a cDNA probe (a), with the
DmLIS1 probe (DmLIS1) and with a cDNA encoding the
Drosophila ribosomal protein S15A (34) as a loading
control. poly(A)⫹ RNA was from five different embryonic
stages (E). The range of age of these embryos is given in
hours after egg deposition (at 25°C). Also loaded are
samples from first and second instar larvae (L1, L2) and
from early third and late third instar larvae (eL3, lL3), and
from early and late pupae (eP, lP). Adult poly(A)⫹ RNAs
stem from an appendage fraction containing mainly legs,
wings and antennae, a head fraction, a thorax and
abdomen fraction, whole males and whole females and
ovaries.
 PAF-ACETYLHYDROLASE SUBUNIT HOMOLOGS IN D. MELANOGASTER 7

Fig. 6. A. The yeast strain EGY48 was

transformed with different “prey” and “bait” in
order to check specific interactions. Transfor-
mants were plated on X-gal galactose contain-
ing plates. 1. pEG202 (empty bait) ⫹ m␣1, 2.
pEG202 ⫹ m␣2, 3. DmLIS1 ⫹ pEG202 (empty
prey), 4. DmLIS1 ⫹ m␣1, 5. DmLIS1 ⫹ m␣2.
B. The yeast strain EGY48 was transformed
with different “prey” and “bait” in order to
check specific interactions between the Dro-
sophila and mammalian LIS and ␣-proteins.
Transformants were streaked on X-gal, galac-
tose containing plates. Ds, Drosophila, m,
mammalian, pEG202 (empty bait) and pJG-45
(empty prey).

CONCLUSION those amino acids is sufficient to confer catalytic activity

Using D. melanogaster embryonic cDNA library we have
cloned the homologs of the ␣- and ␤-subunits of the
mammalian PAF-AH(Ib). This is the first report of the
presence of both of these genes in any lower eukaryote.
The two genes are co-expressed during oogenesis and
embryogenesis, suggesting that both proteins play a role in
early development, supporting the notion that the mamma-
lian PAF-AH(Ib) complex plays a similar role. The level of
the amino acid conservation of the ␣1 and ␤ subunits of
PAF-AH(Ib) from fruit fly to human is remarkable, but
consistent with the reported strict conservation of these
proteins within mammals. In fact, LIS1 appears to be one
of the most conserved proteins between cow and man (3
substitutions in a 410 amino acid long polypeptide chain),
while the ␣1 subunit is not far behind (5 differences among
231 amino acids). This high level of conservation is sugges-
tive of a critical function that these proteins must play in
all the organisms where they are found.
The ␣-subunit homolog has been overexpressed, purified
and crystallized, paving a way for structural studies. It
lacks the PAF-AH catalytic activity, as two of the three
residues in the putative catalytic triad, as determined by
the sequence alignment, are missing. It will be interesting
to see if site-directed mutagenesis aimed at replacing
on the Drosophila melanogaster protein.
Our results pave the way for a better understanding of
the physiological role of the PAF-AH(Ib) complex in cere-
bral development.
ACKNOWLEDGMENTS
This work was supported in part by the National Insti-
tutes of Health, Grant NS36267 (to Z.S.D.), a grant from
the US-Israel Binational Science Foundation (to O.R. and
Z.S.D.), and a grant from the National Cancer Institute of
Canada (B.S.). We would also like to thank Dr. Jay Hirsch,
Department of Biology, University of Virginia, for his help
at the beginning of this project.
The sequences reported in this article were submitted to
GenBank™ National Center for Biotechnology Informa-
tion, Bethesda, MD 20894, with accession codes AF098069
and AF098070.
REFERENCES
1. Pearlman AL, Faust PL, Hatten ME, Brunstrom JE. New direc-
tions for neuronal migration. Curr Opin Neurobiol 1998;8:45–54.
2. Dobyns WB, Truwit CL. Lissencephaly and other malformations
of cortical development: review. Neuropediatrics 1995;26:132–
147.
3. Jones KL, Gilbert EF, Kaveggia EG, Opitz JM. The Miller-Dieker
syndrome. Pediatrics 1980;66:277–281.
8 P.J. SHEFFIELD ET AL.
4. Reiner O, Carrozzon R, Shen Y, et al. Isolation of a Miller-Dieker
lissencephaly gene containing G protein beta-subunit-like re-
peats. Nature 1993;364:717–721.
5. Reiner O, Sapir T. Abnormal cortical development; towards eluci-
dation of the LIS1 gene product function (review). Int J Mol Med
1998;1:849 – 853.
6. Neer EJ, Schmidt CJ, Nambudripad R, Smith TF. The ancient
regulatory-protein family of WD-repeat proteins. Nature 1994;371:
297–300.
7. Wilson R, Ainscough R, Anderson K, et al. 2.2 Mb of contiguous
nucleotide sequence from chromosome III of C. elegans. Nature
1994;368:32–38.
8. Xiang X, Osmani AH, Osmani SA, Xin M, Morris NR. NudF, a
nuclear migration gene in Aspergillus nidulans, is similar to the
human LIS-1 gene required for neuronal migration. Mol Biol Cell
1995;6:297–310.
9. Hattori M, Adachi H, Tsujimoto M, Arai H, Inoue K. Miller-Dieker
lissencephaly gene encodes a subunit of brain platelet-activating
factor acetylhydrolase. Nature 1994;370:216 –218.
10. Sugiura T, Fukuda T, Miyamoto T, Waku K. Distribution of alkyl
and alkenyl ether-linked phospholipids and platelet-activating
factor-like lipid in various species of invertebrates. Biochim
Biophys Acta 1992;1126:298 –308.
11. Venable ME, Zimmerman GA, McIntyre TM, Prescott SM. Platelet-
activating factor: a phospholipid autacoid with diverse actions. J
Lipid Res 1993;34:691–702.
12. O’Flaherty JT, Wykle RL. PAF and cell activation. In: Barnes PJ,
Page CP, Henson P, editors. Platelet activating factor and human
disease. Blackwell Scientific Publications: Oxford; 1989. p 117–
137.
13. El Azzouzi B, Jurgens P, Benveniste P, Thomas Y. Immunoregula-
tory functions of paf-acether. IX. Modulation of apoptosis in an
immature T cell line. Biochem Biophys Res Commun 1993;190:
320 –324.
14. Shukla D. Platelet-activating factor receptor and signal transduc-
tion mechanisms. FASEB J 1992;6:2296 –2301.
15. Bennett SA, Leite LC, Birnboim HC. Platelet activating factor, an
endogenous mediator of inflammation, induces phenotypic trans-
formation of rat embryo cells. Carcinogenesis 1993;14:1289 –1296.
16. Bazan NG, Rodriguez de Turco EB. Platelet-activating factor is a
synapse messenger and a modulator of gene expression in the
nervous system. Neurochem Intern 1995;26:435– 441.
17. Kumar R, Harvey S, Kester N, Hanahan D, Olson M. Production
and effects of platelet-activating factor in the rat brain. Biochim
Biophys Acta 1988;963:375–383.
18. Panetta T, Marcheselli VL, Braquet P, Spinnewyn B, Bazan NG.
Effects of a platelet activating factor antagonist (BN 52021) on
free fatty acids, diacylglycerols, polyphosphoinositides and blood
flow in the gerbil brain: inhibition of ischemia-reperfusion induced
cerebral injury. Biochem Biophys Res Commun 1987;149:580 –
587.
19. Kato K, Clark GD, Bazan NG, Zorumski CF. Platelet-activating
factor as a potential retrograde messenger in CA1 hippocampal
long-term potentiation. Nature 1994;367:175–179.
20. Stafforini DM, Prescott SM, Zimmerman GA, McIntyre TM.
Mammalian platelet-activating factor acetylhydrolases. Biochim
Biophys Acta 1996;1301:161–173.
21. Stafforini DM, McIntyre TM, Zimmerman GA, Prescott SM.
Platelet-activating factor acetylhydrolases. J Biol Chem 1997;272:
17895–17898.
22. Hattori M, Adachi H, Tsujimoto M, Arai H, Inoue K. The catalytic
subunit of bovine brain platelet-activating factor acetylhydrolase
is a novel type of serine esterase. J Biol Chem 1994;269:23150 –
23155.
23. Hattori M, Adachi H, Aoki J, Tsujimoto M, Arai K, Inoue K.
Cloning and expression of a cDNA encoding the beta-subunit
(30-kDa subunit) of bovine brain platelet-activating factor acetyl-
hydrolase. J Biol Chem 1995;270:31345–31352.
24. Ho YS, Swenson L, Derewenda U, et al. Brain acetylhydrolase
that inactivates platelet-activating factor is a G-protein-like tri-
mer. Nature 1997;384:89 –93.
25. Manya H, Aoki J, Watanabe M, et al. Switching of platelet-
activating factor acetylhydrolase catalytic subunits in developing
rat brain. J Biol Chem 1998;273:18567–18572.
26. Morris SM, Albrecht U, Reiner O, Eichele G, Yu-Lee LY. The
lissencephaly gene product Lis1, a protein involved in neuronal
migration, interacts with a nuclear movement protein, NudC Curr
Biol 1998;8:603– 606.
27. Sapir T, Elbaum M, Reiner O. Reduction of microtubule catastro-
phe events by LIS1, platelet-activating factor acetylhydrolase
subunit. EMBO J 1997;16:6977– 6984.
28. Adachi T, Aoki J, Manya H, Asou H, Arai H, Inoue K. PAF
analogues capable of inhibiting PAF acetylhydrolase activity
suppress migration of isolated rat cerebellar granule cells. Neuro-
sci Lett 1997;235:133–136.
29. Clark GD, McNeil RS, Bix GJ, Swann JW. Platelet-activating
factor produces neuronal growth cone collapse. Neuro report
1995;6:2569 –2575.
30. Sheffield PJ, Garrard S, Derewenda ZS. Overcoming expression
and purification problems of RhoGDI using a family of “parallel”
expression vectors. Protein Expr Purif 1999;15:34 –39.
31. Suter B, Romberg LM, Steward R. Bicaudal-D, a Drosophila gene
involved in developmental asymmetry: localized transcript accu-
mulation in ovaries and sequence similarity to myosin heavy
chain tail domains. Genes Dev 1989;3:1957–1968.
32. Wang DS, Shaw R, Hattori M, Arai H, Inoue K, Shaw G. Binding
of pleckstrin homology domains to WD40/beta-transducin repeat
containing segments of the protein product of the Lis-1 gene.
Biochem Biophys Res Commun 1995;209:622– 629.
33. Brunati AM, James P, Guerra B, Ruzzene M, Donella-Deana A,
Pinna LA. The spleen protein-tyrosine kinase TPK-IIB is highly
similar to the catalytic domain of p72syk. Eur J Biochem 1996;240:
400 – 407.
34. Lavoie C, Tam R, Clark M, Lee H, Sonenberg N, Lasko P.
Suppression of a temperature-sensitive cdc33 mutation of yeast by
a multicopy plasmid expressing a Drosophila ribosomal protein.
J Biol Chem 1994;269:14625–14630.