(Received for publication, December 18, 1996, and in revised form, February 12, 1997)
From the Department of Biochemistry, Gunma University School of Medicine, Maebashi 371, Japan
Phospholipase D (PLD) is implicated in important cellular processes such as signal transduction, membrane trafficking, and mitosis regulation. Recently, cDNA for human PLD1 (hPLD1) was cloned from HeLa cells (Hammond, S. M., Altshuller, Y. M., Sung, T-C., Rudge, S. A., Rose, K., Engebrecht, J., Morris, A. J., and Frohman, M. A. (1995) J. Biol. Chem. 270, 29640-29643). hPLD1 is stimulated by phosphatidylinositol 4,5-bisphosphate and the small GTP-binding protein known as ADP-ribosylation factor 1. Here we report the cloning and characterization of cDNA for a different type of PLD (rat PLD2 (rPLD2)) from rat brain. We synthesized highly degenerate amplimers corresponding to the conserved regions of eukaryote PLDs and performed polymerase chain reaction on a rat brain cDNA library. Using the amplified sequence as the probe, we cloned a rat brain cDNA clone that contained an open reading frame of 933 amino acids with an Mr of 105,992. The deduced amino acid sequence showed significant similarity to hPLD1 with a large deletion in the middle of the sequence. When the sequence was expressed in the fission yeast Schizosaccharomyces pombe, PLD activity was greatly increased. The activity was markedly stimulated by phosphatidylinositol 4,5-bisphosphate, but not by ADP-ribosylation factor 1 and RhoA. Rat brain cytosol known to stimulate small GTP-binding protein-dependent PLD did not stimulate rPLD2 expressed in S. pombe. The transcript was detected at significant levels in brain, lung, heart, kidney, stomach, small intestine, colon, and testis, but at low levels in thymus, liver, and muscle. Only a negligible level was found in spleen and pancreas. Thus rPLD2 is a novel type of PLD dependent on phosphatidylinositol 4,5-bisphosphate, but not on the small GTP-binding proteins ADP-ribosylation factor 1 and RhoA.
Phospholipase D (PLD)1 catalyzes the hydrolysis of phosphatidylcholine (PC) to phosphatidic acid and choline (1). A variety of signal molecules such as hormones, neurotransmitters, and growth factors are known to induce the activation of PLD in a wide range of cell types. Hence PLD is implicated in a broad spectrum of physiological processes and diseases, including metabolic regulation, inflammation, secretion, mitogenesis, oncogenesis, neural and cardiac stimulation, diabetes, and senescence (for reviews, see Ref. 2). Despite its crucial importance in signal transduction, the molecular structure and characteristics of PLD enzyme are only poorly understood.
Multiple PLD isoforms exist in mammalian tissues. Several factors were
reported to stimulate PLD activity in vitro, including unsaturated fatty acid (3), phosphatidylinositol 4,5-bisphosphate (PIP2) (4), monomeric GTP-binding proteins (G proteins)
such as ADP-ribosylation factor 1 (ARF1) (5, 6) and RhoA (7, 8),
protein kinase C (9), and calmodulin (10). Massenburg et al.
(11) showed that two major forms of PLD activity in rat brain membranes
can be separated into ARF-dependent and
oleate-dependent enzymes, clearly indicating that these are
distinct isoforms. Both oleate-dependent and
ARF-dependent types of PLD were recently highly purified
from pig lung and brain, respectively (12, 13). In addition, there may
be multiple forms of small G protein-dependent PLD
including ARF-sensitive, RhoA-sensitive, and ARF-, RhoA-sensitive PLDs.
Siddiqi et al. (14) reported that the cytosolic fraction of
HL-60 cells contained a soluble PLD activated by ARF, but not RhoA.
Malcolm et al. (8) showed rat liver plasma membrane PLD to
be sensitive to RhoA, but not to ARF. PLDs in HL-60 membranes and brain
membranes are thought to contain the ARF-, RhoA-sensitive type since
ARF and RhoA acted on the HL-60 enzyme in a synergistic manner
(14-16). PIP2, another important activator of PLD, was
shown to be generally required for the small G
protein-dependent PLDs (5, 6, 8, 17), but not
required for the oleate-dependent PLD (12). However,
Liscovitch et al. (4) demonstrated that the activation
of PLD of rat brain membranes will occur without the addition of
GTPS. Hence there may be a PIP2-sensitive, small G
protein-insensitive PLD in the brain. To further confirm and characterize these putative PLD isoforms, it is extremely important to
purify the enzymes and clone the encoding cDNA.
The cloning of cDNA for higher eukaryote PLD was initially reported by Wang et al. (18), who isolated the cDNA for castor bean (Ricinus communis) PLD by using oligonucleotide probes based on the amino-terminal amino acid sequence of the purified protein (19). Using sequence similarity to castor bean PLD, the Saccharomyces cerevisiae SPO14 gene (20) was identified as the yeast PLD gene (21, 22). By searching for human expressed sequence tags bearing an amino acid sequence similar to the SPO14 sequence, Hammond et al. (23) obtained human cDNA for PLD and designated it hPLD1. Characterization of recombinant hPLD1 revealed that it was activated by PIP2 and ARF1. The activity was strongly inhibited by oleate in agreement with the data obtained for rat brain ARF-dependent PLD (11). In the present study, we cloned cDNA for a novel type of PLD from rat brain (rPLD2) using highly degenerate PCR primers corresponding to the conserved regions among various eukaryote PLDs. The obtained cDNA showed sequence similarity to hPLD1 and contained a significant deletion in the middle of the sequence. Characterization of the rPLD2 expressed in the fission yeast Schizosaccharomyces pombe revealed that the PLD was stimulated by PIP2, inhibited by oleate, and thus partially resembled hPLD1. In sharp contrast to hPLD1, however, rPLD2 was not affected by ARF1 or RhoA, indicating that rPLD2 is a novel type of PLD.
Total RNA was
extracted from 1 g of male rat brain using TRIZOL
reagent (Life Technologies, Inc.) according to the manufacturer's instructions. Poly(A)+ RNA was obtained from total RNA
using Oligotex-dT30 Super (Japan Roche, Tokyo). Oligo(dT)-primed
cDNA synthesis was carried out using a cDNA synthesis kit (Life
Technologies, Inc.). cDNAs were ligated to the ZAP II arms
(Stratagene, La Jolla, CA), and packaged using GigaPack II Gold
packaging extracts (Stratagene).
To obtain the
probe for cDNA screening, PCR was performed using a rat brain
ExCell cDNA library as the template. Recombinant phages were
precipitated with 20% (w/v) polyethylene glycol 6000 and 2 M NaCl, resuspended in 10 mM Tris-HCl, pH 8.0, containing 1 mM EDTA, and treated at 95 °C for 5 min.
The first round PCR was performed at 93 °C for 1 min, 55 °C for 2 min, and then 72 °C for 1 min for 35 cycles using the phage solution
as the template. The second round PCR was performed under similar
conditions using 1:50 volume of the first PCR amplification mixture as
the template. The 5
-amplimer DBP1 (5
-GCIMGICAYTTYRTICARMGITGG-3
,
where I is inosine, M is A or C, R is A or G, and Y is C or T)
corresponded to the amino acid sequence ARHF(V/I)QRW (hPLD1, amino acid
residues 694-701 (23); S. cerevisiae PLD1, 888-895 (21);
S. pombe open reading frame SPAC2F7.16c, 769-776; Fig. 1).
The 3
-amplimer DBP2 (5
-TCRTTDATRTTIGCISWICCDAT-3
, where D is A, G,
or T, S is C or G, and W is A or T) corresponded to the amino acid
sequence IGSANIN(D/E) (hPLD1, 909-1,016; S. cerevisiae
PLD1, 1,109-1,116; S. pombe open reading frame SPAC2F7.16c,
993-1,000; Fig. 1). The amplified 660-base pair fragment
(DBT1 in Fig. 1) was gel-purified and cloned into the
pGEM-T vector (Promega, Madison, WI).
Plaques were formed using a rat brain ExCell cDNA library, then
transferred to Hybond N+ membranes (Amersham Corp.) and screened with
digoxigenin-labeled DBT1 as the probe by hybridizing at 68 °C overnight in a solution containing 5 × SSC (1 × SSC = 0.15 M NaCl in 0.015 M sodium citrate), 1%
blocking reagent (Boehringer Mannheim), 0.02% SDS, and 0.1% sodium
N-lauroylsarcosine. The filters were washed twice with
2 × SSC in 0.1% SDS at room temperature and twice in 0.1 × SSC in 0.1% SDS at 60 °C. Positive phages were located using a
digoxigenin-labeled nucleic acid detection kit (Boehringer Mannheim)
according to the manufacturer's instructions. Plasmids were obtained
from the isolated phages by in vivo excision. The clone DB1
obtained (Fig. 2) was cleaved with PstI (one site within the
insert; another site in the vector), and the excised 300-base pair
fragment was used as the probe for the second-round screening of a rat
brain
ZAPII cDNA library to obtain the full-length cDNA
(DB3) using similar screening conditions.
DNA Sequence Analysis
Both strands of DNA were sequenced using a DNA sequencing kit (Perkin-Elmer) on the 373A DNA sequencer (Perkin-Elmer) after subcloning into pBluescript II (Stratagene).
Reverse Transcription-Polymerase Chain Reaction (RT-PCR)Total RNA (0.1 µg) was reverse-transcribed and
subjected to PCR amplification using a one-step RT-PCR kit (Toyobo,
Osaka) according to the manufacturer's instructions. The reverse
transcription was carried out at 60 °C for 30 min, and PCR
amplification was performed at 94 °C for 1 min, 56 °C for 1 min,
and then at 60 °C for 1.5 min for 40 cycles using primers
5-TCAAGGCCAGATACAAGATACC-3
(rPLD2 sense primer corresponding to
positions 2,042-2,063, DBP3 in Fig. 3),
5
-CACGTAGACTCGGAAACACTGC-3
(rPLD2 antisense primer corresponding to
positions 2,352-2,373, DBP4 in Fig. 3), or
5
-TCCACCACCCTGTTGCTGTA-3
(glycerol-3-phosphate dehydrogenase sense
primer) and 5
-ACCACAGTCCATGCCATCAC-3
(glycerol-3-phosphate
dehydrogenase antisense primer).
Yeast Strain, Culture, and Transformation
S.
pombe strain TKP1 (h ade6-704
leu1-32 ura4-D18) was used as the host for transformation.
Edinburgh minimal medium and YES medium were used for culture (24).
Adenine, L-leucine, uracil, and thiamine were added at
concentrations of 100, 100, 50, and 5 mg/liter, respectively, where
required. Yeast cells were grown aerobically at 32 °C and then
transformed by the lithium acetate method as described (25). Standard
procedures for S. pombe manipulation were as described
(24).
A 3.4-kilobase pair SalI-XbaI fragment (the SalI site was in the vector, and the XbaI site was blunt-ended) containing the entire open reading frame for rPLD2 was isolated from DB3 and subcloned into the SalI/SmaI site of the pREP3X vector to yield pREP3X-rPLD2. pREP3X was a derivative of the S. pombe expression vector pREP1 (26) containing the S. cerevisiae LEU2 marker and a thiamine-repressible promoter.
Preparation of S. pombe Extract and Cell FractionsYeast cells grown in Edinburgh minimal medium in the presence of thiamine were washed three times with the medium without thiamine, cultured in 100 ml of the same thiamine-free medium for 24 h at 32 °C, and finally harvested by centrifugation at 2,000 × g for 5 min. The cells were washed with water and suspended in 1 ml of the extraction buffer containing 50 mM Hepes/NaOH, pH 7.0, 1 mM EDTA, 1 mM EGTA, 0.1 mM dithiothreitol, 2 µM p-amidinophenylmethanesulfonyl fluoride, and 300 mM sucrose. The cells were disrupted by vigorously vortexing with 1.5 g of glass beads (0.3 mm diameter) four times for 1 min at 4 °C. The beads and the cell debris were removed by centrifugation at 2,000 × g for 10 min. The cell extract thus obtained was further centrifuged at 100,000 × g for 60 min to obtain the cytosolic and membrane fractions. The membrane fraction was suspended in 0.2 ml of the extraction buffer. The samples were kept at 4 °C until use.
PLD AssayThe PLD activity in the cell extracts was
determined essentially as described (5) by measuring the
transphosphatidylation activity in the presence of ethanol. Unless
otherwise stated, the cell extracts were incubated with 50 mM Hepes/NaOH, pH 7.5, 200 mM KCl, 200 mM NaCl, 3 mM MgCl2, 2 mM CaCl2, 3 mM EGTA, 1 mM dithiothreitol, 400 mM ethanol, 140 µM phosphatidylethanolamine (PE), 12 µM
PIP2, and 8.6 µM
1-palmitoyl-2-[1-14C]palmitoyl-PC (130 dpm/pmol)
(Amersham Corp.) in a total volume of 100 µl for 60 min at 37 °C.
PC, PIP2, and PE were added as the mixed micelles (5).
GTPS, ARF1, and RhoA were included as indicated. The reaction was
stopped with 100 µl of 1 M HCl. The lipid product was
extracted with 1 ml of chloroform/methanol (2:1, by volume) after
adding 0.5 ml of 170 mM NaCl. Lipids were separated on a
silica gel 60 thin-layer plate (Merck) with chloroform/methanol/acetic acid (13:3:1, by volume). All experiments were performed at least twice, and the representative data were presented.
Recombinant N-myristoylated ARF1 was prepared from Escherichia coli expressing human ARF1 and yeast myristoyl-CoA:protein N-myristoyltransferase (27) as described by Randazzo et al. (28). Recombinant isoprenylated RhoA was prepared from Sf9 cells expressing human RhoA as described by Mizuno et al. (29).
To isolate a novel PLD
cDNA from rat brain, we synthesized degenerate primers (DBP1 and
DBP2) corresponding to the conserved domain of various eukaryote PLDs
(Fig. 1). An open reading frame of S. pombe
(SPAC2F7.16c on chromosome I; GenBankTM Z50142[GenBank]) was also
included in the comparison because it closely resembled other eukaryote
PLDs and was hence thought to be putative S. pombe PLD. We
used the degenerate primers for PCR with a rat brain ExCell cDNA
library as the template. The 660-base pair PCR product (DBT1
in Fig. 2) was cloned into the pGEM-T vector, and the
obtained clones were subjected to restriction analysis. All of the 48 clones analyzed showed the same restriction pattern, indicating that
all clones were identical. Sequence analysis revealed that the sequence
significantly resembled that of hPLD1 (23). Thus DBT1 was used to
screen a rat brain
ExCell cDNA library as the probe. From the
2 × 105 plaques screened, one positive clone was
obtained and designated DB1 (3.1 kilobase pairs, Fig. 2). The
5
-portion of DB1 was subsequently used as the probe in the second
round of screening to obtain the full-length cDNA. Five clones were
obtained from 2 × 105 plaques of a rat brain
ZAPII
cDNA library and analyzed by restriction analysis and in some cases
by partial sequencing analysis. Clone DB3, which contained the largest
insert (4.6 kilobase pairs, Fig. 2), was selected, and the complete
nucleotide sequence was determined on both strands. DB3 contained a
single large open reading frame capable of encoding 933 amino acids
with a calculated molecular weight of 105,992 (Fig. 3).
The 5
-untranslated region contained an in-frame termination codon
followed by the predicted translation start site conforming to the
Kozak consensus (30). The 3
-untranslated region contained a
poly(A)+ tail preceded by the polyadenylation signal at
nucleotides 4,530 to 4,539 (31). The predicted amino acid sequence was
compared with that of hPLD1 by dot matrix analysis (Fig.
4A). Overall, the sequences of the two
proteins are 50% identical, indicating that the protein is
structurally related to hPLD1. Notably, there was a significant
deletion of about 120 amino acid residues in the middle of the
predicted sequence compared with hPLD1. We designated this protein rat
phospholipase D2 (rPLD2) since we had obtained a rat cDNA clone
encoding a protein 91% identical to
hPLD1.2 rPLD2 also contained the consensus
regions conserved among the eukaryote PLDs (Figs. 1 and 4B).
Thus rPLD2 belongs to the same family of protein as hPLD1, yeast PLDs,
and castor bean PLD. Of particular interest, the sequence IGSANIN
(which was used for designing the PCR primer (Fig. 1)) was perfectly
conserved in all of the five eukaryote PLD sequences compared,
suggesting the importance of this sequence in the activity and/or
structure of eukaryote PLD.
Expression and Characterization of rPLD2
We next sought to
express the isolated cDNA using the fission yeast S. pombe as the expression system. It has been shown that many
mammalian cDNAs (for example, protein kinase C (32) and phosphatidylinositol 3-kinase (33)) can be efficiently expressed in
S. pombe. The major advantages of this system are that
transfectants are easy to handle and are stable. Thus S. pombe cells were transfected with either vector alone or rPLD2
cDNA. Cells were disrupted by vortexing with glass beads,
fractionated into the cytosolic and membrane fractions, and then
assayed for PLD activity using labeled PC embedded in the
PIP2-PE mixed micelles as described by Brown et
al. (5). As the most reliable assay of PLD, we measured the
transphosphatidylation activity of the enzyme (34-36). The enzyme was
incubated with the PC-PIP2-PE micelles in the presence of
ethanol, and the phosphatidylethanol formed was separated by thin-layer
chromatography and counted. Although no obviously increased protein
band was observed on SDS-polyacrylamide gel electrophoresis, the
membrane fraction of the rPLD2 transfectants displayed a 26-fold increase in PLD activity when compared with the vector transfectants, confirming that the cDNA indeed encoded a functional PLD (Fig. 5). The cytosolic fraction from the rPLD2-transfected
cells also displayed a 4.4-fold increase in PLD activity. Thus the
major portion of the expressed PLD activity was localized in the
membrane fraction, although rPLD2 had no extended hydrophobic amino
acid stretch as examined by the method of Kyte and Doolittle (37). When
the extract from rPLD2 transfectants was incubated in the absence of
ethanol, phosphatidic acid was formed instead of phosphatidylethanol (data not shown).
Properties of rPLD2
We next examined the enzymatic properties
of the rPLD2 enzyme using the membrane fraction of the transfectant. A
test mixture was prepared by removing PE and PIP2 from the
standard mixture to examine their effects. As shown in Fig.
6A, PE or PIP2 were stimulatory
to the enzyme. Remarkably, PE and PIP2 synergistically activated the enzyme. This synergistic effect was only obtained with
this combination. When phosphatidylserine was used instead of PE,
stimulation by PIP2 was completely abolished, and
phosphatidylinositol was inhibitory. When increasing concentrations of
PIP2 were used in the presence of PE, stimulation occurred
at low concentrations of PIP2 and leveled off at 10 µM (Fig. 6B). The extent of stimulation (5.8-fold) and the required PIP2 concentration were
comparable with those reported previously for PLD activities in
mammalian tissues (11, 13, 17) and Sf9-expressed hPLD1 (23).
Many reports implicated ARF1 (5, 6) and RhoA (7, 8) as the regulators
of PLD activity. hPLD1 was reported to be activated by ARF1 (23). As
shown in Fig. 7, A and B, however, we found that rPLD2 was not activated by either ARF1 or RhoA in the
presence of GTPS. This was in sharp contrast to the case of hPLD1,
which was sensitive to both ARF1 and RhoA (23). In addition, rat brain
cytosol, which is known to stimulate small G
protein-dependent PLD very effectively (15, 38, 39), had no
stimulatory effect on rPLD2 activity, although it was very effective on
the rat brain membrane PLD that was used as the control (Fig.
7C). Thus the response of rPLD2 to the small G proteins was
completely different from that of hPLD1. As shown in Fig. 7D, 10
4 M concentrations of oleic
acid strongly inhibited the enzyme as in the case of hPLD1 (23).
Tissue Distribution of rPLD2 mRNA as Determined by RT-PCR
We examined the expression of rPLD2 mRNA in various
rat tissues by RT-PCR with a pair of rPLD2-specific primers
(DBP3 and DBP4 in Fig. 3). As shown in Fig.
8, rPLD2 mRNA was expressed in brain, lung, heart,
kidney, stomach, small intestine, colon, and testis with the highest
level occurring in the lung. Thymus, liver, and muscle expressed
relatively low levels of rPLD2 mRNA. The transcript was almost
negligible in spleen and pancreas. Although faint, the presence of an
additional band seen in liver, kidney, and muscle suggested the
occurrence of an alternatively spliced form of rPLD mRNA in these
tissues.
The present cloning of rPLD2 (a second form of PLD) will give us an important tool for elucidating the function and role of the PC signaling in mammalian cells. The presence of multiple PLD isoforms has been predicted on the basis of the multiplicity of enzyme activators (2) and the presence of multiple forms of short DNA fragments resembling yeast and plant PLD in the DNA data base (23, 40). Recently, Yoshimura et al. (41) isolated three DNA fragments (rat PLDa, rat PLDb, and rat PLDc) that ostensibly represent a partial sequence of different PLD cDNAs by RT-PCR on rat C6 glioma cell RNA using degenerate primers for the conserved regions in hPLD1 and S. cerevisiae PLD1. The sequence of rat PLDc was 99% identical to the rPLD2 sequence reported here, indicating that rat PLDc is a partial sequence of rPLD2. Besides rPLD2, we have isolated two additional full-sized PLD cDNAs from rat brain designated rat PLD1a and rat PLD1b.2 Sequence comparison has revealed that rat PLD1a and rat PLD1b probably correspond to rat PLDa and rat PLDb, respectively. Although not yet cloned, rat tissues contain another type of PLD (oleate-dependent enzyme (1, 12)). Thus rat tissues express at least four different PLDs. The existence of such multiple forms of PLD explains why related yet different data have been reported for PLD activities in different cells and tissues. It is of great interest to know the specific roles of individual PLD isoforms in mammalian tissues.
The programs of systematic genome sequencing have supplied numerous new cDNA sequences named expressed sequence tags (42) as well as useful information to search for homologues of the sequence of interest. By searching the human expressed sequence tag library using the BLAST program (43), five GenBankTM clones (D20091[GenBank], R02092[GenBank], R69739[GenBank], R83570[GenBank], and R93485[GenBank]) were found to have a strong sequence similarity to rPLD2. For example, R93485[GenBank] showed an 84% homology to rPLD2, but only a 63% homology to hPLD1 at the nucleotide level. This clone is thought to be a partial sequence of the human version of rPLD2. Recently, Ribbes et al. (40) searched the expressed sequence tag library for hPLD1 homologues and pointed out that R93485[GenBank] may be a new PLD isoform. Further analysis of R93485[GenBank] will clarify the presence of an rPLD2-type enzyme in human tissues.
Comparison of rPLD2, hPLD1, and other eukaryote PLDs exposed many interesting aspects. Sequence comparison revealed that four distinct regions are conserved in all of the five known eukaryote PLD sequences (rPLD2, hPLD1, S. cerevisiae PLD1, S. pombe putative PLD1, and castor bean PLD (Fig. 4B)). These conserved regions probably play specific roles in the activity and/or structure of PLD. The fourth carboxyl-terminal consensus contains the sequence IGSANIN, which is followed by a highly conserved hydrophilic amino acid stretch (Fig. 1). This sequence is perfectly conserved in all of the compared eukaryote PLDs and also in the rice and maize PLDs recently cloned (44). We are interested in elucidating the functional role of this consensus in the PLD enzymes.
rPLD2 and hPLD1 resemble each other considerably and share a sensitivity to PIP2 and oleic acid. However, their responses to small G proteins are clearly distinguishable: hPLD1 is stimulated by ARF1 and RhoA, but rPLD2 is not. The major difference in their sequences is a large deletion in the middle of the rPLD2 sequence (Fig. 4A). Hence it is tempting to speculate that this part of the hPLD1 sequence might be responsible for binding to and/or activation by ARF1 and RhoA. Sequence comparison is also expected to provide useful information as to the PIP2 binding site in the PLD enzyme. Like hPLD1 and rPLD2, S. cerevisiae PLD is sensitive to PIP2 (21). Thus rPLD2, hPLD1, and S. cerevisiae PLD1 are thought to commonly contain the PIP2-binding site. There are six conserved regions shared by these three sequences, at least one of which should contain the PIP2-binding site. This information should be useful for future identification of the PIP2-binding region in rPLD2 as well as in hPLD1. Interestingly, these regions are also present in the S. pombe putative PLD1 sequence, although PIP2 stimulation has not yet been reported for S. pombe PLD.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EBI/DDBJ Data Bank with accession number(s) D88672[GenBank].
We thank Dr. R. Kahn (NCI, National
Institutes of Health) and Dr. Y. Takai (Osaka University) for kindly
providing the E. coli strain expressing human ARF1 and the
recombinant virus expressing RhoA in Sf9 cells, respectively. We also
thank Dr. J. I. Gordon (Washington University) for the yeast
NMT1 gene and Dr. K. Hosaka (Gunma University) for the rat
brain ExCell cDNA library. We are indebted to Dr. K. Miyamato
(Gunma University) for valuable advice in the construction of the
phage cDNA library.