(Received for publication, June 5, 1997)
From the Laboratory for Plant Hormone Function, Frontier Research Program, The Institute of Physical and Chemical Research (RIKEN), Wako, Saitama 351-01, Japan and the ¶ Department of Bioproduction, Faculty of Agriculture, Yamagata University, Tsuruoka, Yamagata 997, Japan
ent-Kaurene is the first cyclic
diterpene intermediate of gibberellin biosynthesis in both plants and
fungi. In plants, ent-kaurene is synthesized from
geranylgeranyl diphosphate via copalyl diphosphate in a two-step
cyclization catalyzed by copalyl diphosphate synthase and
ent-kaurene synthase. A cell-free system of the fungus
Phaeosphaeria sp. L487 converted labeled geranylgeranyl
diphosphate to ent-kaurene. A cDNA fragment, which
possibly encodes copalyl diphosphate synthase, was isolated by reverse
transcription-polymerase chain reaction using degenerate primers based
on the consensus motifs of plant enzymes. Translation of a full-length
cDNA sequence isolated from the fungal cDNA library revealed an
open reading frame for a 106-kDa polypeptide. The deduced amino acid
sequence shared 24 and 21% identity with maize copalyl diphosphate
synthase and pumpkin ent-kaurene synthase, respectively. A
fusion protein produced by expression of the cDNA in
Escherichia coli catalyzed the two-step cyclization of
geranylgeranyl diphosphate to ent-kaurene. Amo-1618
completely inhibited the copalyl diphosphate synthase activity of the
enzyme at 106 M, whereas it did not inhibit
the ent-kaurene synthase activity even at 10
4
M. These results indicate that the fungus has a
bifunctional diterpene cyclase that can convert geranylgeranyl
diphosphate into ent-kaurene. They may be separate
catalytic sites for the two cyclization reactions.
Gibberellins (GAs)1 are one of an important group of phytohormones regulating many aspects of plant growth and development. Some fungal species produce GAs as secondary metabolites (1). Gibberella fujikuroi is a rice pathogenic fungus producing high amount of GAs. The fermentation and biosynthesis of GA in G. fujikuroi were well characterized, since some GAs are produced industrially using the fungus (2, 3).
GA is unequivocally synthesized from ent-kaurene in both fungi and plants (4, 5). ent-Kaurene is a tetracyclic diterpene hydrocarbon formed from geranylgeranyl diphosphate (GGDP) via copalyl diphosphate (CDP). The pathway of ent-kaurene biosynthesis was first confirmed using cell-free systems from G. fujikuroi (6). The two-step cyclization was thought to involve two different enzymes: copalyl diphosphate synthase (CPS, formerly ent-kaurene synthase A) and ent-kaurene synthase (KS, formerly ent-kaurene synthase B). These enzymes were partially purified from Fusarium moniliforme, an anamorph of G. fujikuroi. (7). It was suggested that CPS and KS in the fungus might be tightly associated, because the two activities could not be separated. On the other hand, the two activities were successfully separated from a plant enzyme preparation from Marah macrocarpus by chromatographic methods (8). The two enzymes are possibly localized in plastids (9). Quite recently, genes encoding both enzymes have been cloned from plants: CPS from Arabidopsis (10), maize (11) and pea (12), and KS from pumpkin (13). Several other GA biosynthetic enzymes have been cloned from plants as well (14). In contrast, no GA biosynthetic enzymes have been cloned from fungi.
GA biosynthesis in another GA-producing fungus,
Phaeosphaeria sp. L487 has been characterized recently (15).
This fungus produces significant amounts of GA1 through a
pathway similar to that in higher plants: namely, 3-hydroxylation of
GA9 and GA20 occurs to form GA4 and
GA1, respectively (15, 16). This contrasts to G. fujikuroi, in which 3
-hydroxylation occurs early in the pathway
of GA12-aldehyde and GA9 is not converted to
GA4. Although ent-kaurene is an intermediate of
GA biosynthesis in both fungi and plants (17), we were unable to
compare the characteristics of the plant CPS and KS with fungal CPS and
KS at molecular levels. Little is known about enzymes and genes
involved in ent-kaurene biosynthesis of
Phaeosphaeria sp. L487. In addition, none of the diterpene
cyclases from fungi have been isolated yet. Therefore, we focused on
fungal-type CPS (FCPS) of Phaeosphaeria sp. L487. In this
paper we report cDNA cloning and characterization of an enzyme
involved in ent-kaurene biosynthesis in the fungus
Phaeosphaeria sp. L487.
[1-3H]GGDP was purchased from Amersham. [3H]CDP was a gift from Dr. Saito of the Institute of Physical and Chemical Research (18). Amo-1618 was obtained from Serva (Heidelberg).
Preparation of Cell-free ExtractsEleven-day-old mycelium of Phaeosphaeria sp. L487 grown as described (15) was harvested by filtration and homogenized in potassium phosphate buffer (50 mM, pH 8.0) containing dithiothreitol (2 mM), pepstatin A (0.1 mM), and (p-amidinophenyl) methanesulfonyl fluoride hydrochloride (10 mM) using Dyno-Mill (Willy A. Bachofen, Basel, Switzerland). The supernatant after centrifugation (2,000 × g at 4 °C for 15 min) was dialyzed three times for 45 min against phosphate buffer (50 mM). The dialysate was centrifuged at 200,000 × g at 4 °C for 90 min to obtain soluble enzyme fraction (S-200). The protein concentration was determined with Bio-Rad standard assay (19) using bovine serum albumin as the standard.
Isolation of Nucleic AcidsPhaeosphaeria sp. L487 was cultured in GOM liquid medium (8% glucose, 1.5% oatmeal, 0.5% KH2PO4, 0.1% MgSO4·7H2O, 0.1% NH4NO3) for 6 days at 26 °C on a reciprocal shaker. The mycelium was filtered and stored in liquid nitrogen. The frozen mycelia (4.8 g) were ground to fine powder in liquid nitrogen, then total RNA was extracted by SDS-phenol methods (20). Poly(A)+ RNA (4.9 µg) was isolated from the total RNA (300 µg) by Dynabeads oligo(dT)25 (Dynal, Oslo, Norway).
Polymerase Chain ReactionTwo degenerate oligonucleotides
were synthesized based on the common amino acid sequences of plant CPSs
(12, 13, 14). The degenerate primers used for the amplification
were 5-GCITA(T/C)GA(T/C)ACIGCITGGGT-3
and
5
-(A/G)AAIGCCATIGCIGT(A/G)TC(A/G)TC-3
. PCR was performed in a
50-µl reaction volume using Expand HF buffer with 1.5 mM MgCl2, 0.25 mM dNTP, 8 µM of each
primer, 1.75 units of the enzyme (Expand HF, Boehringer Mannheim) and 1 µl of double-stranded cDNA solution as described below with the
following program: denaturation at 94 °C for 1 min, annealing at
46 °C for 1 min, extension at 72 °C for 2 min (35 cycles), then
final extension at 72 °C for 7 min. PCR products were directly
cloned into pCRII vector (Invitrogen).
Double-stranded
cDNA was synthesized from 4.9 µg of poly(A)+ RNA by
reverse transcription with Superscript II (Life Technologies, Inc.)
followed by DNA polymerase (Takara) and RNase H treatments. The
cDNA, ligated with EcoRI/NotI adaptor
(Pharmacia Biotech Inc.), was size fractionated using a SizeSep 400 spun column (Pharmacia) and ligated to EcoRI-digested ZAP
II (Stratagene). The phage was packaged in vitro with
Gigapack III Gold (Stratagene) and transfected into E. coli
XL1-Blue. Hybond N+ membrane (Amersham Corp.) replicas were
screened with the PCR product, which had been labeled by random priming
(rediprime DNA labeling system, Amersham). Hybridization was
performed at 68 °C for 16 h in Rapid-hyb buffer (Amersham)
containing salmon sperm DNA (200 µg/ml). The filters were washed with
2 × SSC at 68 °C for 15 min, with 1 × SSC, 0.1% SDS at
68 °C for 30 min then with 0.1 × SSC, 0.1% SDS at 68 °C
for 30 min. The positive plaques were purified by three more rounds of
screening as described above. The cDNA was in vivo
excised into pBluescript II SK(
).
To determine the nucleotide sequence of FCPS cDNA, deletion clones were prepared from the pBluescript-inserted cDNA with ExoIII/mung bean nuclease deletion kit (Stratagene) according to manufacturer's instruction. FCPS cDNA and its deletion clones were sequenced using the dye primer cycle sequencing method (PRISM 377, Applied Biosystems).
To align the FCPS sequence with other sequences, the Genetics Computer Group programs (University of Wisconsin) were used. Homology search was carried out using the BLAST program. To make phylogenetic tree, PHYLIP programs (version 3.5) were used (J. Felsenstein, University of Washington, Seattle).
Isolation of the 5The sequence of
5 end of the FCPS cDNA was determined using Marathon cDNA
amplification kit (CLONTECH). The first strand cDNA was reverse-transcribed from poly(A)+ RNA (0.5 µg) using a gene specific primer 1: 5
-CCAGCTACCGTCCTCGCCTTG-3
(complementary to nucleotides 202-222). Double-stranded cDNA with adaptors was prepared according to the manufacturer's instruction and
subjected to a PCR using adaptor primer 1:
(5
-CCATCCTAATACGACTCACTATAGGGC-3
and a gene-specific primer 2:
5
-GGCCGCCCAGGCTGTGTCGTAGAT-3
(complementary to nucleotides 106-129).
The PCR-amplified DNA fragment (~300 bp) was directly cloned into the
pCR II vector.
Poly(A)+ RNA (0.3 µg), prepared from 6-day-old mycelium of Phaeosphaeria, was size-fractionated with a 0.8% agarose gel containing formaldehyde and blotted on to a Hybond N+ membrane. The membrane was hybridized using 32P-labeled probe from PCR products (873 bp) by random priming at 68 °C and finally washed with 0.1 × SSC containing 0.1% SDS at 68 °C for 30 min.
Expression of Protein and Preparation of Fungal CPSThe
cDNA inserted in pBluescript was digested with NotI. The
digested DNA fragment was ligated to NotI-digested pGEX 4T-3 (Pharmacia). E. coli JM109 cells harboring the ligated
plasmid, pGEX-FCPS, was grown in Luria-Bertani medium containing
ampicillin (50 µg/ml) and 0.2% glucose at 37 °C. When
A600 reached 0.7, expression of the
glutathione S-transferase (GST) fusion protein was induced by addition of isopropyl-1-thio--D-galactoside (IPTG) to
a final concentration of 1 mM. The cells were collected
after 24-h induction at 20 °C and washed with 50 mM
phosphate buffer, pH 8.0. Cells were resuspended in a lysis buffer (50 mM potassium phosphate buffer, pH 8.0, 10% glycerol, 2 mM dithiothreitol, 5 mM MgCl2) and
sonicated on ice. After centrifugation at 15,000 × g
for 25 min, the supernatant was used as a crude soluble fraction. The protein concentration of the solution was estimated with the Bio-Rad protein standard assay as described above. The soluble protein was
analyzed by SDS-polyacrylamide gel electrophoresis with Coomassie Brilliant Blue staining (21). The GST-fusion protein was detected by
Western blot analysis using GST detection module (Pharmacia).
CPS and KS activities of the cell-free system and the recombinant protein were measured as described previously (15, 18). In the case of the cell-free system, each substrate was incubated in 100 µl of the S-200 enzyme solution containing Mg2+ (5 mM) and uniconazole (1 mM), as ent-kaurene oxidase inhibitor (22), at 30 °C for 30 min. A ent-kaurene fraction after incubation was extracted with n-hexane and separated on silica gel TLC with n-hexane. The radioactivity of the ent-kaurene region on the TLC plate (RF 0.6-1.0) was measured by liquid scintillation counting of the silica gel.
To identify metabolites by gas chromatograph-mass spectrometry (GC-MS), GGDP (2 µg) was incubated in 100 µl of enzyme solution containing Mg2+ (5 mM) and uniconazole (1 mM) at 30 °C for 1 h. The incubation was stopped by heating (65 °C) for 5 min. Six units of bacterial alkaline phosphatase (Takara) was added into the reaction mixture and incubated overnight at 37 °C. After incubation, 100 µl of acetone was added to terminate the hydrolysis reaction. The mixture was extracted three times with 100 µl of n-hexane. The n-hexane extracts were concentrated under gentle nitrogen flow and injected into a 15-m fused silica capillary column (DB-1, J & W Scientific, Folsom, CA) coupled to an ion-trap GC-MS (GCQ, Thermo Quest, San Jose, CA).
A crude soluble enzyme fraction (S-200) was prepared from Phaeosphaeria mycelium. CPS and KS activities were measured from the conversion of [3H]GGDP and [3H]CDP to ent-kaurene, respectively (Table I). The denatured enzyme by heating was used as negative controls. The activities were lost by adding 20 mM EDTA. This effect of EDTA indicated that the enzyme requires a divalent metal ion such as Mg2+, for the cyclization, as has been reported for other plant terpene cyclases (23, 24). To confirm the production of ent-kaurene from GGDP by the enzyme, the n-hexane extract from the reaction with unlabeled GGDP was analyzed by full-scan GC-MS. The product was detected at 6:02 after injection: m/z 272 ([M]+; relative intensity, 32%); m/z 257 (relative intensity, 100%), m/z 229 (relative intensity, 97%), and m/z 213 (relative intensity, 93%). This compound was identified as ent-kaurene by comparing with full mass spectrum of the authentic compound. These results demonstrated that the presence of both CPS and KS activities in mycelia of Phaeosphaeria sp. L487. These activities were found exclusively in the S-200 fraction, suggesting that the enzymes for ent-kaurene biosynthesis are soluble. This is also the case for in G. fujikuroi and higher plant enzymes (17).
|
To design degenerate primers for reverse transcription-PCR, plant CPS sequences were compared. Several combinations of primers were designed based on the consensus motifs. Double-stranded cDNA was synthesized from poly(A)+ RNA prepared from 6-day-old mycelium, and PCR was carried out. Only one combination of primers encoding AYDTAW and DDTAMAF motifs yielded a specific DNA fragment of the anticipated size. Sequence analysis of the DNA fragment revealed 873 bp encoding 291 amino acid residues that showed significant similarity to sequences of CPS (10-12), KS (13), and abietadiene synthase (AS) (25). RNA blot analysis was performed using poly(A)+ RNA prepared from 6-day-old mycelium of Phaeosphaeria to detect a 3-kilobase transcript that hybridized with the PCR product (data not shown).
Using this PCR product as a hybridization probe, 2.5 × 105 plaques from a cDNA library constructed in ZAP
II were screened. Four positive clones were isolated and further
purified. One of the clones, pFCPS, was further characterized. Sequence
analysis revealed that the 3,034 bp pFCPS cDNA insert contained an
open reading frame of 2,838 bp encoding 946 amino acid residues. To determine the 5
end of the FCPS cDNA, 5
-rapid amplification of
cDNA ends experiments were performed. The 5
-end of the FCPS transcript was mapped at 24 bp upstream from the 5
end of pFCPS. Since
there was no ATG codon in the extended sequence, we concluded that the
2,838-bp open reading frame encodes FCPS. The complete sequence of FCPS
cDNA is shown in Fig. 1.
The predicted amino acid sequence of FCPS has significant homology with
those of diterpene cyclases from plants; 20% identity with AS (grand
fir, Ref. 25), 20% with casbene synthase (castor bean, Ref. 26), 24%
with maize CPS (11), 22% with pumpkin KS (13), and 21% with taxadiene
synthase (pacific yew, Ref. 27). The FCPS sequence is not closely
related (~19% identity) to any of the microbial sesquiterpene
cyclases (28-30). Sequence alignment with plant diterpene cyclases
revealed that the motifs, YDTAW, QXXDGSW, and DVDDTA, were
conserved in the FCPS sequence (Fig. 2).
The electron-rich aromatic residue of the QXXDGSW motif is
proposed to stabilize intermediate cation during the cyclization process (31). FCPS contains aspartate- and aspartate/glutamate-rich motifs (DDVLD: amino acids 132-136 and DEVIDEVVD: amino acids 683-691, Fig. 2). These functions have been strongly suggested to
mediate substrate binding by chelation of the divalent metal ion (32,
33). The transit peptide in plant diterpene cyclases is rich in serine
and threonine residues, with a net positive charge (34), is found in
the amino-terminal region. However, the amino-terminal region of FCPS
was not rich in serine and threonine. The calculated pI of the first 30 amino acids of the FCPS was 4.8 (complete amino acid sequence:
calculated pI = 5.2). Therefore, the transit peptide region is not
present in FCPS (Fig. 1).
Expression and Functional Analysis of the Recombinant Protein in E. coli
The pFCPS was introduced into pGEX 4T-3 and the GST-FCPS fusion protein was expressed in E. coli after induction with 1 mM IPTG. The IPTG-induced protein was detected in the soluble fraction as a 132-kDa band on SDS-polyacrylamide gel electrophoresis. It was confirmed to be GST-FCPS by Western blot analysis using GST detection module (data not shown).
The soluble enzyme fraction containing GST-FCPS was used for enzyme
assays. Interestingly, when [3H]GGDP was incubated with
the soluble enzyme, radioactivity showing formation of a diterpene
hydrocarbon was detected by TLC analysis. Furthermore, when
[3H]CDP was incubated with the enzyme, radioactivity
associated with the ent-kaurene-like compound was also
detected. To confirm production of ent-kaurene from GGDP by
the FCPS, full-scan GC-MS analysis was carried out. To identify
residual GGDP and CDP as geranylgeraniol and copalol by GC-MS,
respectively, the reaction mixture after incubation of unlabeled GGDP
with the enzyme was treated with alkaline phosphatase. After the
hydrolysis, nonpolar compounds were extracted with n-hexane.
After full-scan GC-MS analysis, ent-kaurene was identified.
At the same time, copalol was also detected in the reaction mixture
(Fig. 3). When pea CPS (PsCPS) fusion
protein was incubated under the same condition, conversion of GGDP to
CDP, but not to ent-kaurene, occurred (Fig. 3). Furthermore,
when pumpkin KS (CmKS) fusion protein was added to PsCPS fusion
protein, ent-kaurene was produced (Fig. 3). The KS activity
of the FCPS was confirmed using [3H]CDP (Table
II). Thus, it is clear that the FCPS
fusion protein catalyzes both the cyclization of GGDP to CDP and that
of CDP to ent-kaurene (Fig.
4). Hence, here we changed the name of
FCPS to fungal-type ent-kaurene synthase or FCPS/KS as a
bifunctional enzyme.
|
Amo-1618 is a quaternary ammonium compound that inhibits CPS activity
(7, 35). When Amo-1618 was added to the reaction mixture at a
concentration of 106 M, conversion of
[3H]GGDP to ent-kaurene by FCPS/KS was
completely inhibited. However, conversion of [3H]CDP to
ent-kaurene was not inhibited, even at 10
4
M (Table II). This suggested that sites for CPS and KS
activities are distinct in FCPS/KS.
We report the isolation and characterization of a cDNA
encoding the fungal ent-kaurene synthase, FCPS/KS, from
Phaeosphaeria sp. L487. The genes for the fungal
sesquiterpene cyclases, trichodiene synthase from Fusarium
sporotrichoides (28) and aristolochene synthase from
Penicillium roqueforti, have been characterized (29).
However, although many fungal species produce interesting bioactive
diterpene compounds, such as the GAs and fusicoccins (36), none of the
genes encoding diterpene cyclases have been isolated from fungi. Thus,
this is the first report of the cloning of a diterpene cyclase of
fungal origin. We initially focused on the cDNA cloning of fungal
CPS, since the information about the sequence of CPS was obtained from
various plants. The predicted amino acid sequence of FCPS/KS isolated
from Phaeosphaeria sp. L487 has homology with the plant CPS
and KS. The phylogenetic tree (Fig. 5)
indicates that FCPS/KS has a significant evolutional relationship with
plant CPS. However, whereas both CPS and KS have a transit peptide in
their amino-terminal regions, this is absent in FCPS/KS. FCPS/KS
contains neither membrane translocation sequences or membrane spanning
regions. Therefore, FCPS/KS is likely to be localized in the cytoplasm,
as are the fungal sesquiterpene cyclases (29). This is also consistent
with the presence of enzyme activity in the soluble fraction of the
cell-free preparation from Phaeosphaeria sp. L487. In
plants, all GGDP-derived carbon skeletons are synthesized within
plastids and sesquiterpene hydrocarbons within the cytoplasm or
endoplasmic reticulum (37).
We revealed that FCPS/KS is a bifunctional cyclase having both activities of CPS and KS in plants. FCPS/KS catalyzes the two-step cyclization of GGDP to ent-kaurene via CDP with a single polypeptide. Amo-1618 inhibits CPS activity in FCPS/KS at low concentrations. In contrast, it does not inhibit the KS activity, even at high concentrations. Similar inhibition characteristics are observed in enzyme preparations of G. fujikuroi and Marah macrocarpus (35). It has been suggested that, in plants, CPS and KS may interact with each other to synthesize ent-kaurene from GGDP (6). This suggests that the catalytic sites for the two cyclizations are separated. Furthermore, the similar modes of action of Amo-1618 on ent-kaurene biosynthesis in fungi and plants indicate that FCPS/KS of Phaeosphaeria sp. L487 and CPS-KS complex of plants have similar structure. At present, the two active sites responsible for CPS and KS activities within the FCPS/KS cannot be determined by sequence comparison. In contrast, AS is also a bifunctional enzyme which catalyzes the cyclization of GGDP to abietadiene via labdadienyl diphosphate (25). The amino-terminal region of AS, containing the DXDDTA motif, resembles CPS, whereas the carboxyl-terminal region, containing the DDXXD motif, resembles KS (Fig. 2). The two active sites within the AS suggested a fusion of elements of CPS-type enzymes and KS-type enzymes (25). Thus, there are some examples where biosynthetic genes are separate in one organism but linked in another (38-40). FCPS/KS is an attractive model for understanding of structure-activity relationship of the enzymes involved in ent-kaurene biosynthesis between fungi and plants. To understand the catalytic sites of FCPS/KS, more detail studies will be required. Comparison of FCPS/KS with diterpene cyclases from G. fujikuroi and other fungi will be necessary for understanding the evolution of GA biosynthesis in fungi.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) AB003395.
We are grateful to Dr. Peter Hedden (IACR-Long Ashton Research Station, University of Bristol) and Prof. Tai-ping Sun (Duke University) for their critical readings of the manuscript. We thank Drs. Shinjiro Yamaguchi and Tahar Ait-Ali (Frontier Research Program, RIKEN) for providing KS and CPS recombinant proteins, respectively, and Yukiji Tachiyama for expert technical assistance on DNA sequencing.