(Received for publication, April 9, 1997, and in revised form, May 22, 1997)
From the Centre de Génétique
Moléculaire du CNRS, 91198 Gif-sur-Yvette, France, the
§ Centre Orsan de Recherches en Biotechnologies, 91953 Les Ulis, France, and the ¶ Service de Radiophysiologie
Végétale du CEA,
13108 St. Paul-Lès-Durance, France
Two NADPH-cytochrome P450 reductase-encoding cDNAs were isolated from an Arabidopsis cDNA library by metabolic interference in a Saccharomyces cerevisiae mutant disrupted for its endogenous cpr1 gene. ATR1 encodes a protein of 692 amino acids, while ATR2 encodes either a 712-residue protein (ATR2-1), or a 702-residue protein (ATR2-2) depending on the choice of the initiation codon. Comparative analysis of ATR1 and ATR2-1 indicates 64% amino acid sequence identity and the absence of conservation in the third base of conserved amino acid codons. The two Arabidopsis reductases are encoded by distinct genes whose divergence is expected an early event in angiosperms evolution. A poly(Ser/Thr) stretch reminiscent of a plant chloroplastic targeting signal is present at the ATR2-1 N-terminal end but absent in ATR1. The cDNA open reading frames were expressed in yeast. The recombinant polypeptides were found present in the yeast endoplasmic reticulum membrane and exhibited a high specific NADPH-cytochrome c reductase activity. To gain more insight into the respective functions of the two reductases, the Arabidopsis cDNA encoding cinnamate 4-hydroxylase (CYP73A5) was cloned and co-expressed with ATR1 or ATR2 in yeast. Biochemical characterization of the Arabidopsis ATR1/CYP73A5 and ATR2-1/CYP73A5 systems demonstrates that the two distantly related Arabidopsis reductases similarly support the first oxidative step of the phenylpropanoid general pathway.
P450 monooxygenases are involved in various biosynthesis pathways and in degradation of a large range of exogenous compounds (1, 2). In plants, besides their roles in herbicide and pesticide detoxication (3, 4) and in promutagen oxidative activations (5), P450s are widely involved in several secondary metabolisms such as the phenylpropanoid pathway (6, 7). This pathway, which leads to important molecules such as lignins, pigments, coumarins, flavonoids, and phytoalexins (6), involves a common oxidative step: the para-hydroxylation of cinnamate catalyzed by a P450 of the CYP73 family.
In animals, the microsomal P450 system associates in the same organism multiple P450 isoenzymes but a single ubiquitous NADPH-P450 reductase (CPR).1 Conclusive evidence for the presence of a single CPR-encoding gene was provided with Chinese hamster cells, mouse spleen cells (8), and rat hepatocytes (9). Unlike animals, higher plants express multiple forms of CPRs as inferred from Western blot analysis of purified material (10, 11). Durst et al. (12) have recently isolated two distinct partial length cDNAs (HTR1 and HTR3) from Helianthus tuberosus, corresponding to distinct (2.4- and 2.6-kb) mRNA species.
This observation questioned the physiological significance of the multiplicity of plant CPRs. Among hypotheses, one could consider the requirement of different CPRs to support the activities of specific plant P450s or the differential CPR expressions in plant tissues or at various development stages. The possibility of multiple subcellular locations for CPRs in plant-specific organelles in addition to the regular endoplasmic reticulum (ER) location is also of particular interest. To address these questions, full-length Arabidopsis CPR-encoding cDNAs were cloned by a function-based approach involving complementation in yeast of an endogenous CPR defect. The selection relies on the reduced growth rate and the strong ketoconazole hypersensitivity of a yeast strain, which has been disrupted for the CPR1 gene, which encodes yeast microsomal CPR. The addition of ketoconazole, a potent inhibitor of yeast P450 lanosterol demethylase, is required for selection due to the presence of alternate electron donors in yeast, making CPR1 deletion non-lethal. In the cpr1 strain, the CPR defect and ketoconazole inhibition contribute together to cause a full depletion of ergosterol biosynthesis, leading to cell growth arrest (13-15). Screening of an Arabidopsis cDNA library in a multicopy plasmid pFL61 under the transcriptional control of the phosphoglycerate kinase (PGK) promoter (16) was carried out in the W(R) strain. This strain is engineered to overexpress the yeast CPR in galactose- but not in glucose-containing medium (17). The PGK promoter is active on both carbon sources, allowing complementation by heterologous CPRs of the conditional mutant phenotype induced by culture of transformed W(R) cells on glucose. The isolated Arabidopsis CPRs were further characterized for a physiologic coupling with Arabidopsis cinnamate 4-hydroxylase P450 (CYP73A5) by coexpression in yeast. For that purpose, the Arabidopsis cDNA encoding CYP73A5 was cloned from the same library using a hybridization probe derived from the corresponding H. tuberosus CYP73A1 gene (18).
Restriction and DNA modification enzymes were obtained from New England Biolabs. Thermostable Taq and Pfu DNA polymerases were from Boehringer Mannheim and Stratagene, respectively. The SequenaseTM (version 2.0) DNA sequencing kit was purchased from U.S. Biochemical Corp. The pCRScriptTM cloning kit was from Stratagene. NADPH, horse heart cytochrome c, and cinnamic and coumaric acids were from Sigma, and 2-naphthoic acid was from Aldrich.
Yeast Strains and VectorsThe Saccharomyces
cerevisiae strain W303-1B (MATa;
ade2-1; his3-11,-15; leu2-3,-112;
ura3-1; trp1-1) is designated as W(N). Yeast
strain W(R) was constructed by substitution of the natural promoter of
the CPR1 gene (encoding microsomal CPR) by the
galactose-inducible and glucose-repressed promoter
GAL10-CYC1. WR derives from W(N) by disruption of the
CPR1 gene with a TRP1 selection marker (17).
The pFL61 vector has been previously described (16). The yeast expression vector pYeDP60 (V60) contains both URA3 and ADE2 as selection markers, and an expression cassette constituted by GAL10-CYC1 promoter and PGK terminator sequences surrounding a cDNA insertion polylinker (19). The pGP1 vector contains a 5500-bp-long fragment that encompasses the full CPR1 gene at the HindIII site of pUC19. The pYeDP51 yeast expression vector is identical to pYeDP1/8-2 (20), except that the PvuII site close to the yeast origin of replication is replaced by a BglII site.
Using ORS31 and ORS32 as primers (Table I) and pGP1 as a
template, the 428-bp-long CPR1 gene terminator sequence was
PCR-amplified and cloned into the SmaI site of pUC19. The
420-bp-long EcoRI-PvuII fragment of the resulting
vector was ligated to the EcoRI-PvuII fragment of
pYeDP51 encompassing URA3. The 6900-bp-long resulting vector, named pYeDP100, places the inserted cDNA under the
transcriptional control of GAL10-CYC1-inducible promoter and
CPR1 terminator. By PCR amplification using ORS33 and ORS34
primers and pGP1 as a template, a 631-bp-long
HindIII-BglII fragment of the upstream part of
the CPR1 promoter has been obtained and cloned in pUC19. The
HindIII-BglII fragment of the CPR1
promoter was ligated with the larger
HindIII-BglII fragment of pYeDP100, which
encompasses URA3, yielding the 5210-bp-long pYeDP110
integrative vector. The pYeDP110 vector bears GAL10-CYC1
promoter and CPR1 terminator sandwiching a cDNA
insertion polylinker, and the CPR1 5-noncoding region
placed upstream of the URA3 selection marker. The entire integrative fragment can be removed by a NotI digestion. The
respective orientation of the two CPR1 noncoding regions
allows targeted integration of the expression cassette at the
CPR1 locus by homologous recombination.
|
Transformation of yeast strains was performed by a modified lithium acetate procedure as described (21), except for the transformation of W(R) cells with the pFL61 library (see below). Transformed cells were selected onto glucose-containing SGI plates. Culture media, cell cultures, and galactose induction procedures of individual clones in both SLI and YPGE were as described previously (17, 22).
Screening of the Arabidopsis cDNA Expression Library in YeastA sized (inserts >2 kb) cDNA library from Arabidopsis thaliana seedlings including roots (two-leaf stage) in pFL61 yeast expression vector was kindly provided by Dr. M. Minet. In pFL61, cDNAs were inserted as NotI cassettes between the yeast PGK promoter and terminator (16). Yeast W(R) cells were grown on galactose, thus overexpressing yeast CPR, up to a cell density of 2 × 107 cells/ml. Cells are harvested and resuspended at a cell density of 106 cells/ml in a glucose-containing YPGA medium, thus turning off CPR1 gene expression. W(R) cells were further grown on glucose for 5 h until the microsomal CPR activity decreases to the value found in wild-type W(N) cells. About 108 cells were lithium-treated and used for transformation by 10 µg of the Arabidopsis cDNA library, 0.2 mg of heat-denatured herring sperm DNA in 2 ml of 50 mM Tris-HCl buffer, pH 7.4, containing 1 mM EDTA and 0.1 M LiCl. After a 10-min incubation at room temperature, 5 ml of the same buffer containing 50% of poly(ethylene)glycol-4000 was added and gently mixed to the previous 2 ml, and the incubation was continued for 20 min at 37 °C and then 5 min at 42 °C. The cells were sprayed in 0.1-ml aliquots onto 20 plates of SGAI synthetic medium for uracil prototrophy selection.
Two pools of Ura+ transformants were constituted, and cells from each pool were sprayed at 104 cells/plate onto a series of SGAI plates containing increasing concentrations of ketoconazole (0, 5, 10, 20, and 50 µg/ml ketoconazole) and incubated for 36 h at 28 °C. The 52 clones found to be resistant to 10 and 20 µg/ml ketoconazole were striped on SGAI medium containing 10 µg/ml ketoconazole. The plasmidic DNA of 24 of the resistant clones was rescued in Escherichia coli. The BamHI and HindIII digestion patterns revealed five different restriction profiles (profiles A-E).
Isolation of Arabidopsis CYP73A5-encoding cDNAThe Arabidopsis cDNA library in pFL61 was transformed in E. coli by electroporation. About 105 independent clones were recovered and transferred on nitrocellulose. The blots were screened by hybridization in SSC × 6 buffer containing 0.1% SDS and 0.05% milk overnight at 55 °C with the 32P-labeled 1800-bp-long cDNA encoding Helianthus CYP73A1 (18). Of 12 independent clones selected, two strongly hybridize with the same probe after subcloning. The plasmidic DNA of one of the two clones, named pCYP73A5, was isolated and the NotI fragment encompassing the full-length cDNA sequenced.
DNA Sequence DeterminationThe NotI cDNA inserts encoding ATR1 and ATR2 were extracted from pFL61 and subcloned into the unique NotI site of a derivative of pUC9. For ATR1 and ATR2 cDNAs, sequencing was carried out with M13 universal oligonucleotides and a series of specific oligonucleotides belonging to the sequences. For CYP73A5 cDNA, the sequencing using M13 universal oligonucleotides was performed by using nested deletions with exonuclease III of pCYP73A5 digested by KpnI and HindIII. Nucleotide sequences were determined on both strands by the dideoxynucleotide chain termination procedure.
Southern Blot AnalysisThe Arabidopsis genomic DNA (kindly provided by Dr. Clint S. Chapple) was digested with BamHI, EcoRI, or HindIII. The digested aliquots (10 µg) were separated by 0.8% agarose electrophoresis and then transferred to a Nylon membrane (Schleicher and Schüll). The blotted membrane was probed in stringent conditions with 32P-labeled ATR1 or ATR2 open reading frame.
Tailoring of the Selected cDNAsOligonucleotides were synthesized and purified as described previously (19). The 2079-bp-long ATR1 coding sequence was PCR-amplified using ORS24 and ORS25 primers with 10 ng of pFL61/ATR1 as template. These primers introduce a BglII site immediately upstream of the ATG codon and a EcoRI site downstream of the stop codon of the PCR-amplified fragment. The 2139-bp-long ATR2-1 coding sequence was PCR-amplified using the combination of ORS21 and ORS22 primers, and the 2109-bp-long ATR2-2 coding sequence was PCR-amplified using ORS21 and ORS23 primers and 10 ng of pFL61/ATR2 as template in both cases. The three ATR2 primers introduce a BamHI site immediately upstream of the ATG initiation codon and downstream of the stop codon of the PCR-amplified fragments. The expected PCR-amplified products were cloned at the SmaI site of pUC19, giving, respectively, pUC19/ATR1, pUC19/ATR2-1, and pUC19/ATR2-2. The ATR1 coding sequence was then excised by BglII-EcoRI digestion and inserted into pYeDP60 linearized at BamHI and EcoRI unique sites, resulting in plasmid pATR1/V60. The ATR2-1 and ATR2-2 coding sequences were excised by BamHI digestion and inserted into pYeDP60 linearized at BamHI site, a clone with the proper orientation was chosen for each insert giving, respectively, pATR2-1/V60 and pATR2-2/V60. Amplification of the 1518-bp-long CYP73A5 coding sequence was performed by PCR using ATC4H-5 and ATC4H-3 primers, and the product was cloned at the SfrI site of pCRScript vector, giving pCRScript/ATC4H. The primers introduce a BamHI site immediately upstream of the ATG codon and a BglII site downstream of the stop codon. The coding sequence was excised by BamHI-BglII digestion and inserted at the BamHI site of pYeDP60, and a clone with the correct orientation was selected and named pATC4H/V60.
Engineering of WAT11, WAT21, and Derived Yeast StrainsThe
ATR1 and ATR2-1 open reading frames were isolated from pUC19/ATR1 and
pUC19/ATR2-1 as BamHI-EcoRI and
BamHI-BamHI fragment, respectively. The
ATR1 fragment was inserted at the
BamHI-EcoRI sites of pYeDP110, yielding
pATR1/DP110. Similarly, the ATR2-1 fragment was inserted at the unique
BamHI site of pYeDP110, and a clone with the proper
orientation was selected and named pATR2-1/DP110. The NotI
fragment of each plasmid encompassing URA3 and the ATR coding sequence was isolated and used to transform WR cells. Since
in WR
the CPR1 locus is disrupted with the
TRP1 gene, integration events result in a phenotype shift
from Ura
Trp+ to Ura+
Trp
. This procedure yielded two new yeast strains, WAT11
(for ATR1) and WAT21 (for ATR2-1). They express the
Arabidopsis CPR instead of the yeast enzyme when grown on
galactose. On glucose, WAT cells express no CPR activity. The
URA3 selection marker in both WAT strains was mutated to
ura3 by selecting 5-fluoroorotate-resistant clones (23),
giving WAT11U and WAT21U strains.
W(R) and
WAT cells were harvested after a 12-h galactose induction and
microsomal fractions prepared as described in Ref. 17. P450 content in
yeast microsomes was calculated from the reduced carbon monoxide
difference using a differential absorption coefficient (450 versus 490 nm) of 91 mM1·cm
1 (24). Protein
concentrations were determined by the Pierce BCA (bicinchoninic acid)
protein microassay using bovine serum albumin as a standard.
Cinnamate 4-hydroxylase activity was assayed by high pressure liquid chromatography-monitoring 4-hydroxy-cinnamate production as described previously (22). For naphthoate hydroxylation reactions, activity was determined using a 1-ml assay mixture containing 120 µM NADPH, various concentrations of 2-naphthoate, and the microsomal aliquot in 50 mM Tris/HCl buffer, 1 mM EDTA, pH 7.4. The time dependence of the fluorescence emission was monitored at 440 nm, the excitation wavelength being set at 290 nm. The rate of the 2-naphthoate 6-hydroxylation reaction was calculated from the slope of the linear graph obtained. Transformed yeast was grown and assayed for cinnamate bioconversion essentially as described in Refs. 17 and 22.
CPR activity was measured on microsomal fractions as described
previously (22). The rate of cytochrome c reduction was
calculated using a differential absorption coefficient of 21 mM1·cm
1 at 550 nm. The rates
of dichlorophenolindophenol and ferricyanide reductions were monitored,
respectively, at 600 nm (22 mM
1·cm
1) and 420 nm (1 mM
1·cm
1).
Yeast cells W(R) that carry the CPR1 gene under the transcriptional control of the galactose-inducible GAL10-CYC1 promoter were transformed with the Arabidopsis cDNA expression library in pFL61 (16). Extinction of the modified CPR1 gene, when glucose is used as a carbon source for culture, is not lethal (14, 17). However, yeast CPR deficiency caused a dramatic lowering in transformation efficiency due to changes in the cell wall structure. This led us to the design of an "on the fly" transformation strategy (see "Experimental Procedures"). The ketoconazole resistance criterion was used for the selection of clones expressing constitutively an alternate CPR in the pFL61-based cDNA expression library.
Approximately 105 Ura+ colonies were selected,
pooled, and sprayed on a series of SGAI plates containing increasing
ketoconazole concentrations. Fifty-two colonies were sorted out as
resistant to 10 and 20 µg/ml ketoconazole. Twenty-four of them,
selected at random, were further analyzed by restriction digestion.
Five types of restriction profile were evidenced: A (7 clones), B (6 clones), C (5 clones), D (1 clone), and E (5 clones). The plasmidic inheritance of the ketoconazole resistance associated with each type of
sequence was tested by transforming WR, a yeast strain carrying a
permanent CPR1 disruption. Four (A, B, C, and E) of the five
plasmid classes were shown to confer ketoconazole resistance at 40 µg/ml to WR
. Partial cDNA sequencing was used for further identification and showed that classes A, C, and E (~2300-bp-long cDNA NotI fragment) contain related inserts differing
only by their orientation and their 5
- and 3
-noncoding extremities
when class B (2199-bp-long NotI insert) appeared containing
a different open reading frame. The relative amounts of both
ATR mRNAs in the plant were estimated by transforming
E. coli with the Arabidopsis cDNA library in
pFL61. Out of 104 independent clones, two hybridize
ATR1 probe and seven hybridize ATR2 probe (not
shown). This result demonstrates that in the cDNA library used,
ATR1 cDNA is 3 times less represented than
ATR2 cDNA.
The cDNA inserts of
class B and C plasmids were sequenced on both strands. The
ATR1 cDNA is 2199 bp long and encompasses 69 bp of the
5-noncoding region followed by a 2079-bp-long open reading frame and
51 bp of the 3
-noncoding region (Fig. 1). A termination
TGA codon (from
13 to
15) immediately precedes in frame the first
putative initiation codon. The first methionine codon of
ATR1 cDNA closely matches the dicot plant initiator
codon consensus sequence (25). The ATR1 open reading frame
encodes a protein of 692 amino acid residues with an estimated
molecular mass of 76,720 Da, similar to that of rat liver CPR (77 kDa)
(26).
The 2290-bp-long ATR2 cDNA, also obtained as a
full-length clone, consists of a 2139-bp-long open reading frame
sandwiched by a 5-noncoding region of 50 bp and a 3
-flanking region
of 101 bp (Fig. 2). The ATR2 open reading
frame encodes a protein of 712 amino acid residues, which is 64%
identical to ATR1, with an estimated molecular mass of 79,077 Da, more
similar to the Mr observed for higher plant CPRs
(11, 27, 28). The assignment of the starting codon was not
straightforward, since the flanking sequence preceding the first ATG of
ATR2 cDNA does not contain any stop codon in frame and
since two other ATG codons were found in the 15 first codons.
Comparison with the consensus sequences around the initiation codon in
dicot plant mRNAs (25) indicates that both the first and the second
ATG of ATR2 cDNA could be a suitable translation
initiation site; it is therefore difficult to predict which of these is
the true starting site. The ATR2 protein encoded by the first ATG was
designated ATR2-1 and is 712 residues long; the shorter protein,
starting at the second ATG, was named ATR2-2 and is 702 residues long
(Fig. 3).
Fig. 4 shows, based on sequence comparison with
flavodoxin (26, 29) and NADP+-ferredoxin reductase (30,
31), that each ATR presents the conserved segments typical of all known
CPRs for the binding of FMN, FAD, and NADPH. In this respect, ATR1 and
ATR2 are typical NADPH-P450 reductases. In contrast, no clear sequence
similarity is found between the 85 first residues of ATR1 and the 106 first residues of ATR2. However, despite the absence of similarity in their N-terminal part, both ATR1 and ATR2 present in this part an
hydrophobic segment that could act as a membrane anchor, as observed in
other microsomal CPRs. Another distinct difference between the two ATRs
is the net charge of the protein; ATR1 is predicted to be more acidic
than ATR2 at neutral pH, exhibiting a calculated charge
versus pH profile very different from those of ATR2 and
human and yeast CPRs (Fig. 5). The differences in charge
density appear to be mainly localized in the C-terminal moieties of the
ATRs. More unusual is the presence at the N terminus of ATR2 of a
poly(Ser/Thr) stretch that is not found in any other fungal, plant, or
animal CPR. This particular segment could account for the specific
glycosylations found in certain higher plant CPRs. The protein ATR2-2,
encoded from the second ATG, corresponds to ATR2-1 depleted of this
stretch of hydroxylated amino acid residues. Moreover, a second
hydrophobic stretch that could serve as a membrane anchor is found
at the N-terminal side of the canonic ER membrane-anchoring element in
both ATR2-1 and ATR2-2.
Cloning and Sequencing of Arabidopsis CYP73A5
Using as a
probe the coding sequence of Helianthus CYP73A1 (18), one
pFL61 clone of the Arabidopsis library was selected by
hybridization. The 1735-bp-long full-length cDNA contains an open
reading frame of 1518 bp. The translation initiation site was assigned
to the first ATG, since it is preceded in frame by a stop codon (from
37 to
39-positions). The 5
- and 3
-noncoding regions are 48 and
169 bp long, respectively (Fig. 6). This
Arabidopsis cDNA encodes a protein of 505 amino acid
residues with a calculated molecular mass of 57,751 Da, a value typical
of eukaryotic P450s. The high sequence identity found between
Arabidopsis and Helianthus CYP73s (82%) confirms
that cinnamate 4-hydroxylase enzymes share highly conserved
sequence.
Characterizations and Evolution of ATR Genes
A Southern blot
of Arabidopsis genomic DNA digested with different
restriction enzymes was probed with either ATR1 or
ATR2 coding sequence (Fig. 7). With
ATR1 probe, a single band of 12.5 kb is observed with
EcoRI digestion, a site absent in the cDNA, while two
bands are observed with BamHI digestion and three distinct bands with HindIII. On the other hand, with ATR2
probe, two bands are observed with EcoRI and
HindIII and only one with BamHI digestion (absent
from the cDNA). No cross-hybridization is observed. Therefore, the
fact that a digestion gives in each case a single band strongly suggests that each ATR cDNA sequence originates from a
single copy gene.
PCR amplification starting from Arabidopsis genomic DNA and using the primers situated at both extremities of the open reading frames led to the amplification of a 3.9-4.0-kb fragment for ATR1 and of a 3.5-3.6-kb fragment for ATR2. This result confirms the presence of a total of 1.8-1.9 kb and of 1.4-1.5 kb of intronic sequences in the ATR1 and ATR2 genes, respectively. The PCR amplification of CYP73A5 DNA from Arabidopsis genomic DNA as a template using the two primers situated at both extremities of the CYP73A5 coding sequence resulted in amplification of a 1.7-1.8-kb fragment, indicating the presence of a total of about 200 bp of intron sequence, which is a rather small amount.
Dating of the ATR1/ATR2 gene divergence event was attempted by calculation of the level of identity (or ETBI value) between the third base of codons encoding conserved amino acid residues. A reference value (or RTBI value) was also calculated similarly by comparing all possible nucleotide sequences encoding ATR1/ATR2-1 conserved amino acid residues using the observed average codon usage as a ponderation factor. For ATR1/ATR2, ATR1/HSR1, and HSR1/SCR1 couples, Table II shows that ETBI and RTBI values are similar, as expected for distantly related genes (no memory of the third base usage). In contrast, ETBI and RTBI values still significantly differ for H1A1/H1A2 human P450s taken as a control. Such a difference is also detected for the comparison of Helianthus CPR cDNA (HTR1) (12) with ATR2 but not with ATR1. The lower amino acid sequence identity observed between ATR1 and ATR2 (64%) than between ATR2 and HTR1 (73%) and the higher ETBI value for the HTR1/ATR2 couple therefore indicate that the ATR2 gene is probably more distant from the ATR1 gene, although it belongs to the same species, than from the Helianthus HTR1-encoding gene (interspecies comparison).
|
To confirm these findings, a phylogenetic analysis was performed on CPR protein sequences using the CLUSTER algorithm and the BLOSUM62 sequence comparison matrix (32). The phylogenetic tree (not shown) revealed that plant CPRs clearly partition into two divergent subfamilies, one including ATR1 and Mung bean CPR and the second clustering ATR2 and all other plant CPR sequences available today. Interestingly, the two known Helianthus CPR sequences, the full-length HTR1 and the partial length HTR3 (which are 85% similar), cluster in the same subfamily together with ATR2. Thus, Arabidopsis ATR1 and ATR2 sequences constitute the first characterized presence of two distantly related CPR subfamilies in a single organism. These data highly suggest that the corresponding gene divergence probably occurred early during angiosperms evolution.
High Level Expression of Arabidopsis CPRs in YeastTo
optimize expression, the cDNAs encoding ATR1 and ATR2 were
reformatted by deletion of their flanking regions and cloned in
pYeDP60, placing the coding sequences under the transcriptional control
of a GAL10-CYC1 promoter. WR cells were transformed by the resulting vectors, namely pATR1/V60, pATR2-1/V60, and pATR2-2/V60. Upon expression in strain WR
following galactose induction, the three Arabidopsis enzymes confer a resistance to
ketoconazole at least equal to that observed with the original pFL61
clones. This result demonstrates that each ATR is fully competent to
substitute for endogenous yeast CPR and particularly that both ATR2-1
and ATR2-2 are functional in coupling with yeast P450s. Thus, deletion of the poly(Ser/Thr) stretch in ATR2-2 has no detectable consequence on
the subcellular location or function in yeast.
Fig. 8 shows that microsomes prepared from
ATR-expressing yeasts present a dramatically reduced endogenous P450
content as compared with the WR control. The high level of P450s
found in the ER of WR
cells was shown previously to be induced by
CPR deficiency and decreased upon yeast cytochrome
b5 overexpression (15). The expression of ATR1
or ATR2-1 in WR
thus reestablishes a wild type transcriptional
regulation of endogenous P450s in CPR-deficient cells, indicating that
both Arabidopsis CPRs efficiently substitute physiologically
for the endogenous yeast enzyme.
Microsomal fractions prepared from transformed WR cells were
prepared and assayed for several NADH- and NADPH-dependent
acceptor reduction reactions typical of the CPR enzymes (Table
III). As a control, microsomes prepared from WR
cells
transformed by a cDNA-free pYeDP60 exhibited very low
NADPH-cytochrome c reductase activity. NADH does not support
any ATR-dependent reductase activity. In contrast, a strong
microsomal NADPH-cytochrome c reductase activity is observed
for the three ATRs, with ATR2-1 exhibiting an unusually high activity
(2840 nmol/min/mg) as compared with ATR1 and ATR2-2 (100 and 280 nmol/min/mg, respectively). These values are to be compared with a
microsomal specific activity of 100 nmol/min/mg for human CPR
overexpressed in WR
cells using the same vector system. This result
indicates that the N-terminal poly(Ser/Thr) stretch of ATR2 is not
essential for its CPR activity or ER targeting in yeast but
significantly enhances the expressed ATR2-1 reductase activity as
compared with ATR2-2. The Km value of ATR1 and
ATR2-1 are, respectively, 12 ± 2 and 8 ± 1 µM for NADPH and 3 ± 1 and 3 ± 1 µM for
cytochrome c. These values compare well with those reported
for other plant CPRs (10, 27, 33, 34).
|
Analysis of the coupling between any ATR and a plant P450 requires coexpression in yeast. Genomic integration of the ATR1 and ATR2-1 expression cassettes at the CPR1 locus was considered. For this purpose, ATR1 and ATR2-1 coding sequences were cloned into pYeDP110 integration vector, yielding pATR1/V110 and pATR2-1/V110. The pYeDP110 vector places the heterologous coding sequence under the transcriptional control of GAL10-CYC1 promoter and CPR1 terminator and provides sequences from both flanking regions of the CPR1 gene that direct the integration at this locus, resulting in the deletion of most of the CPR1 gene. The ATR1 and ATR2-1 coding sequences were integrated at the CPR1 locus, resulting in two new yeast strains, WAT11 and WAT21, respectively. PCR analysis and Southern blotting confirmed that the full ATR expression cassette has been integrated within the CPR1 locus of each selected clone (not shown).
For each WAT strain, a uracil auxotrophic mutant was selected on
5-fluoroorotate (23) and named WAT11U and WAT21U. When cultivated on
glucose, WAT cells exhibit no microsomal CPR activity and are
phenotypically equivalent to WR cells. But when grown on
galactose, WAT11U and WAT21U cells express, respectively, microsomal ATR1 and ATR2-1 from the integrated expression cassette. Whether on
glucose or in galactose, WAT cells express no yeast CPR. The generation
time on galactose of the engineered strains was compared with that of
the W(R) strain, which overexpresses yeast CPR, and no difference was
observed. On the other hand, WAT11U and W(R) cells grown on glucose
have identical generation times, significantly longer than upon growth
on galactose.
WAT11U and WAT21U cells were grown on galactose to induce ATR expressions. The microsomal NADPH-cytochrome c reductase activity was measured (Table IV). A single integrated copy of the ATR1 expression cassette in WAT11 allows a CPR expression level similar to the value observed with a multicopy plasmid. In contrast, the reductase activity is reduced about 10-fold in WAT21, as compared with the plasmid-based expression, reaching a figure similar to WAT11. This suggests that the observed difference between ATR2-1 and ATR2-2 expression levels on multicopy plasmid might be related to different plasmid copy numbers and not to intrinsic properties of these two CPRs. All of these strains are fully isogenic, with the exception of the modified CPR1 locus. They can thus be used for comparing the effect of the CPR nature on any P450 catalytic properties.
|
The CYP73A5 coding sequence was PCR-amplified and cloned in pYeDP60. The resulting expression vector, pCYP73A5/V60, was used to transform W(R), WAT11U, and WAT21U. Coexpression of cinnamate 4-hydroxylase and Arabidopsis or yeast CPRs was initiated by galactose induction, and bioconversion was assayed by adding cinnamic acid directly into the culture medium. Rapid accumulation of p-coumarate was observed (see Table IV). In contrast, no detectable product formation was observed with the same strains transformed by a cDNA-free pYeDP60 vector. The rate of bioconversion was found to be roughly identical in the three tested strains, suggesting that the three CPRs similarly support CYP73A5 in vivo activity in yeast.
Microsomal fractions of pCYP73A5/V60-transformed W(R), WAT11U, and WAT21U cells were collected, and carbon monoxide-induced difference spectra were recorded on dithionite-reduced yeast microsomes suspensions. Spectrally detectable P450 content was found to be approximately 200 pmol of CYP73A5/mg of microsomal protein in W(R), 100 pmol/mg for WAT11 and WAT21 strains, and undetectable (<10 pmol/mg) in control. These values compare well with those usually observed in microsomes of yeast expressing mammalian P450s. The addition of cinnamate to microsomes from CYP73A5-expressing W(R), WAT11, and WAT21 cells causes a spectral shift with a differential absorption peak centered at 389 nm and a trough at 423 nm (not shown), typical of a ligand-induced low spin to high spin transition of the P450 heme ferric iron.
Microsomes from CYP73A5-expressing yeast cells were assayed for
NADPH-dependent CYP73A5-catalyzed cinnamate and naphthoate hydroxylations. These reactions were found to be strictly
NADPH-dependent and were not supported by NADH. Table IV
lists the apparent kinetic parameters at steady state for the two
substrates of the recombinant CYP73A5 in each of the three microsomal
redox environments. Whatever the nature of the CPR, yeast or
Arabidopsis, the turnover of the recombinant CYP73A5 was
very high, reaching 470 min1 in W(R). Although the level
of NADPH cytochrome c reductase activity was 15-fold lower
in WAT11 as compared with W(R), the cinnamate hydroxylase turnover
number was only reduced 2.3-fold. Intermediate results between these
two cases were obtained with WAT21. Km values of the
P450 for the two substrates tested (cinnamate and naphthoate) were
found to be identical whatever the strain (W(R), WAT11, or WAT21).
These results suggest that ATR2-1 and ATR1 are similarly efficient in
supporting Arabidopsis CYP73A5 activity when coexpressed in
yeast.
In this work, two Arabidopsis cDNAs, whose products increase the resistance of CPR-deficient yeast cells to the antifungal drug ketoconazole and reestablish a physiological expression level of yeast endogenous P450s, have been isolated. These cDNAs encode two distantly related (64% amino acid residue identity) NADPH-P450 reductases, ATR1 and ATR2, which are similarly found in the ER membrane when expressed in yeast. The two ATRs efficiently support the activity of P450 cinnamate 4-hydroxylase from the same plant when coexpressed in yeast. The use of a cloning procedure by metabolic interference allowed us to overcome the strict requirement of sequence similarity, which is critical for cDNA cloning by standard hybridization strategies. Southern blot analysis and sequence comparison in flanking regions of the cDNAs demonstrated that ATR1 and ATR2 correspond to two distinct genes, each existing as a single copy in the Arabidopsis genome. These CPRs are not close allelic variants but are encoded by distantly related genes as shown by the amino acid sequence divergence and the third base randomization in conserved codons. Similar cases have been found in several other Arabidopsis gene families and in particular with phenylalanine ammonia lyases, phytochromes, and chalcone synthases (35-39).
The presence of multiple CPR genes in plants was also suggested by the
multiplicity of bands detected in Western blots of CPR purified to
homogeneity from Helianthus (11), sweet potato (10), and mung bean (27)
and by the multiplicity of CPR peaks eluting from the
2,5
-ADP-Sepharose column when purifying CPRs from periwinkle (28),
tulip bulbs (33), and petunia flowers (34) with, in the latter case, an
absence of antigenic cross-reaction between the proteins eluted in two
peaks. In Helianthus, two partial length cDNAs probably
encoding distinct CPRs have recently been cloned (12), but the
phylogenetic comparison carried out in this work demonstrates that they
cluster in the same plant CPR subfamily together with ATR2. In
contrast, ATR1 belongs to the other plant CPR subfamily, clustering
with the unique mung bean CPR sequence characterized to date. The fact
that ATR2 appears to be more similar to Helianthus HTR1 and
HTR3 than to ATR1 suggests that a parent common to both ATR
genes would have existed before the separation of
Arabidopsis (Brassicaceae) and
Helianthus (Asteraceae). Arabidopsis
is thus the first organism in which the cloning and function of CPR
genes of two distinct subfamilies are achieved. In contrast, a single
CPR-encoding gene has been systematically reported in lower eukaryotes,
like fungi, and in mammals. This raises the question of the
physiological role for two distant CPRs in higher plants. A hypothesis
would be a differential expression of each ATR either during various
stages of plant development or in different plant tissues. There are
several known cases of multiple gene families that encode protein
isoforms with different tissue localizations in Arabidopsis
such as potassium channels isoforms (40), zinc finger proteins (41),
and the myosin gene family (42). There are also examples of plant
protein isoforms encoded by multiple gene families, which present
differential expressions during development of the plant seedling such
as maize manganese superoxide dismutases (43) and, of course, the
Arabidopsis floral homeotic MADS-box regulatory gene family
(44). In other cases, the members of a gene family exhibit both
differential temporal and tissue-specific regulations, such as the
Arabidopsis gibberellin aldehyde-induced transcripts
(45).
Alignments reveal the presence in the two ATRs of all expected conserved FMN-, FAD- and NADPH-binding domains and thus are new members of the ferredoxin-NADP+ reductase family of proteins (30, 31, 46). A very efficient cinnamate bioconversion is observed by co-expression in yeast of ATR1 or ATR2 and Arabidopsis CYP73A5. ATR1, ATR2, and yeast CPR each support rather similar CYP73A5 turnover numbers for cinnamate. Based on these data, the two ATRs would have virtually exchangeable roles. Nevertheless, the relative molar ratio between CPR and P450 expressed in yeast is 10-fold lower in the case of ATR expressions than in the case of yeast CPR in W(R) cells. The apparent equivalence between ATR1 and ATR2 to support cinnamate hydroxylase activity in yeast microsomes might thus be related to a saturation of CYP73A5 activity by high level of CPR. The W(R) or WAT engineered strains permit us to obtain turnover numbers for Arabidopsis and Helianthus CYP73s about 1000-fold higher than the value reported for Medicago CYP73 expressed in wild-type yeast (47). Similarly, expression in E. coli of a functional fusion protein consisting of the Catharanthus CYP73 fused at the N terminus of Catharanthus CPR, was recently reported (48). However, the CYP73 specific activity in this case was found 40-fold lower than in the present work. Functional analysis of the large number of plant P450-encoding cDNAs reported to date is thus expected to take advantage of the engineered WAT strains.
Almost no amino acid sequence similarity is found between the 100 first residues of ATR1 and ATR2-1, except for the presence of the hydrophobic stretch that is critical for the binding of CPRs to microsomal membrane (49). In ATR2-1, this stretch is preceded by a poly(Ser/Thr) N-terminal extension and an unusually long amphipatic sequence segment. This motif, which has no equivalent in any other fungal, plant, or animal CPRs, is highly reminiscent of a chloroplastic targeting sequence (50, 51), thus questioning the authentic subcellular location of ATR2-1 in plants. The observed addressing in yeast could thus be artifactual due to the absence of the plant-specific organites. The presence of P450s in chloroplast still remains debatable, but very recently it has been shown that Linum allene oxide synthase, CYP74, presents a typical N-terminal chloroplastic transit peptide (52). Moreover, it is highly suggested that the gibberellin biosynthesis pathway, which produces phytohormones, is localized at least partially in the chloroplast (Ref. 53 and references therein). ent-kaurene synthesis, from which all gibberellins derive, has been clearly established to take place in the chloroplast (54, 55). Nevertheless, the subcellular location in plants of the P450-catalyzed steps of ent-kaurene and gibberellin oxidations is still debated (53, 56, 57). This leaves open the hypothesis that in Arabidopsis chloroplasts could house some P450-catalyzed reactions with ATR2-1 as a chloroplastic CPR. In this hypothesis, ATR1 would be the Arabidopsis microsomal counterpart. A similar, but clearly established situation, has been found with Arabidopsis signal recognition particles. Distinct genes have been shown to encode two SRP54p isoforms, one cytosolic (58) and the other chloroplastic (59). This raises the question of the occurrence in plants of chloroplast-specific CPR and P450s whose physiological roles remain to be fully investigated.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) X66016 (for ATR1), X66017 (for ATR2), and U37235 (for CYP73A5).
We thank Philippe Fusier and Cécile Le Dû for excellent technical assistance. We especially thank Drs. Danièle Werck-Reichhart and Francis Durst for fruitful discussions.