(Received for publication, November 26, 1996, and in revised form, February 6, 1997)
From the Friedrich Miescher-Institut, In eukaryotic cells, pre-tRNAs spliced by a
pathway that produces a 3 Transcripts of many tRNA genes in eukaryotes contain a single
short intron, located in a conserved position in the anticodon loop,
which is excised by a different mechanism to that utilized during
nuclear pre-mRNA processing (Fig. 1; reviewed in Refs. 1 and
2). Splicing of pre-tRNA is initiated by endonucleolytic cleavages that result in removal of the intron and formation of two
tRNA half-molecules, a 5
The 2 The plant CPDase was originally purified from wheat as an enzyme that
hydrolyzes nucleoside 2 In this work we describe the molecular characterization of the plant
cyclic phosphodiesterase. The cDNA encoding this protein in
Arabidopsis has been cloned and the properties of the
bacterially overexpressed and purified enzyme have been compared with
those of its purified wheat counterpart. The Arabidopsis
enzyme has no apparent structural resemblance to other known cyclic
nucleotide phosphodiesterases.
Plantlets of Arabidopsis
thaliana, ecotype Columbia C0, were grown in Petri dishes
containing 0.8% agar, 1% sucrose, and MS salts (19) in a 22 °C
growth chamber under a 12-h light/12-h dark cycle. Three weeks after
sowing, leaves, roots, and floral buds were harvested. For leaf strip
incubation, the leaves were sliced and incubated in a culture medium
described by Nagy and Maliga (20), containing 1 mg/liter of
2,4-dichlorophenoxyacetic acid. Aliquots were harvested at different
times of culture for total RNA extraction.
About 5 µg of wheat germ
CPDase, purified as described previously (18), was applied to
the SDS-PAGE 10% gel. After blotting to the polyvinylidene difluoride
membrane (Bio-Rad) and staining with Ponceau S, the approximately
23-kDa CPDase band was excised and treated with trypsin. Proteolytic
peptides were resolved by HPLC and sequenced by Dr. W. S. Lane (Harvard
MicroChem, Cambridge, MA). Three peptides were sequenced (Fig.
2).
A Primer extension was carried out (21) using a
25-mer-specific 32P-labeled oligonucleotide 1 (ACGGTGACGTGAGGAACGAATCTTG), complementary to positions 203-227 of the
cDNA (Fig. 2), 10 units of avian myoblastosis virus reverse
transcriptase, and 100 µg of RNA isolated from the Arabidopsis leaf strip cultures incubated for either 24 or
48 h. The resulting cDNA was purified on a 6% denaturing
polyacrylamide gel and its 3 The
fragment encompassing 226 base pairs of the promoter and upstream
portion of the transcribed region was cloned by inverse PCR (23). The
Arabidopsis DNA was digested with EcoRI and
ligated using T4 DNA ligase. PCR was performed using oligonucleotide 5 (CGATTCCTCATCTGGTAATGC; complementary to positions 130 to 150 in Fig.
2) and oligonucleotide 6 (GCTAATGGAAGCTTTGAGATCC; positions 168 to 189 in Fig. 2) as primers. The amplified fragments were cloned into the
SmaI site of pBluescript and sequenced.
Total RNA
from Arabidopsis organs and leaf strip cultures was isolated
as described by Hall et al. (24). RNA (10 µg/lane) was
separated on a formaldehyde-agarose gel, blotted onto Hybond-N nylon
membrane (Amersham Corp.) by capillary transfer using 20 × SSPE,
and UV-cross-linked to the membrane. The integrity and the amount of
RNA applied to each lane were verified by control hybridizations using
a tomato 25 S rRNA probe (25). The cDNA fragment extending from
positions 430-741 (Fig. 2) was used as a CPDase probe. The histone H4
probe corresponds to the 196-base pair restriction fragment
AccI/DdeI of the coding region of the gene H4A748
(26). The actin probe corresponds to the 570-base pair PCR-amplified
fragment of the Arabidopsis actin gene AAc1 (27).
The genomic DNA was isolated from lyophilized Arabidopsis plants using a procedure similar to that of Murray and Thompson (28).
The probes were labeled with [ For immunoblot analysis, proteins were fractionated by SDS-PAGE and
electroblotted onto the polyvinylidene difluoride membrane. The
membrane was probed with a 1:1000 dilution of the hen polyclonal antibody. The immunoreactive proteins were detected using
peroxidase-conjugated affinity-purified rabbit anti-chicken IgYs
(Dianova) and the ECL Western blotting analysis system from
Amersham.
A
BamHI site was introduced 3 Two hens were immunized with
the purified recombinant CPDase. For primary immunization, 20 µg of
the protein, in Freund's complete adjuvant, was used. After 4 weeks,
20 µg of the protein with Freund's incomplete adjuvant was injected.
Eggs were collected daily starting 2 weeks after the last immunization.
Antibodies were purified from egg yolk according to the three-step
method of Polson and von Wechmar (32). A 2 M
(NH4)2 SO4 precipitation was used
for the complete removal of polyethylene glycol.
All nucleoside
2 For calculation of specific activities
and kinetic analysis, a quantitative assay based on a measurement of
the phosphatase-sensitive nucleotide product was used. All incubations
(20 µl) contained 50 mM Tris-HCl, pH 7.0, and 0.01%
Triton X-100. Concentrations of CPDase and substrates and incubation
times at 30 °C were as indicated in the figure legends. Reactions
were stopped by boiling for 2 min, and 80 µl of 0.1 M
Tris-HCl, pH 8.0, containing 0.2 unit of CIP was added. After
incubation for 10 min at 37 °C, liberated phosphate was assayed
according to Hess and Derr (35). For determination of the
Km and Vmax values, the
assays contained substrates at concentrations of 1.38-12.5
mM. All velocities were calculated from the initial linear
rates. Values were fitted to the Lineweaver-Burk equation by the linear
regression method assuming proportional errors.
Products of enzymatic digestions, performed as described previously
(8), were analyzed by cellulose TLC in solvent A (saturated (NH4)2SO4/3 M sodium
acetate/isopropyl alcohol (80:6:2)) or by polyethyleneimine-cellulose
TLC in solvent B (0.75 M LiCl). The nucleotide standards
and reaction products were visualized under UV light.
The CPDase coding sequence (the
NcoI-BamHI fragment from pQECPDase; see above)
was cloned into the NcoI and BamHI sites of the
pGGS.5 expression vector (kindly provided by Gordon Simpson of this
laboratory; the vector contains a duplicated cauliflower mosaic virus
promoter and a cauliflower mosaic virus poly(A) signal), resulting in
the plasmid pHATCPDase. The CPDase encoded by pHATCPDase contains the
influenza hemagglutinin nonapeptide epitope tag (flu tag, amino acids
YPYDVPDYA) at the C terminus. A similar plasmid (pNRBP43) expressing
the nuclear RNA-binding protein N-RBP43 of Nicotiana
plumbaginifolia with the flu tag fused at the C terminus, was used
as a control for the nuclear-localized protein. Mesophyll protoplasts
of N. plumbaginifolia were transfected by the polyethylene glycol method (36), using 20 µg of plasmid per transfection. Transfected protoplasts were collected 24 h after transfection, washed twice with 10 ml of W5 solution (36), twice with 10 ml of 0.5%
Mes, pH 5.7, containing 175 mM CaCl2, and
suspended in 100 µl of the same Mes/CaCl2 buffer.
Indirect immunofluorescence analysis was performed according to Cairns
et al. (37) and Neuhaus et al. (38), with
modifications. Aliquots of the protoplast suspension were spread on the
surface of detergent- and acetone-washed glass slides. The slides were dried in an oven at 55 °C and kept overnight under vacuum in a desiccator. Samples were fixed for 30 min at room temperature by
overlaying with 100 µl of the fixation solution (4% formaldehyde, 1% Triton X-100 in 0.1 M potassium phosphate buffer, pH
8.0). They were subsequently washed four times for 30 min with 1 ml of
washing solution A (2.5% NaCl in 0.1 M glycine-KOH buffer, pH 8.5), and overlaid for 1 h in a wet chamber with 100 µl of a
solution of the rabbit antiserum against the flu epitope (antibody HA-11; Berkley Antibody Co.), diluted 1:80 with buffer B (2.5% NaCl
solution in 0.1 M Tris-HCl, pH 7.4). Glass slides were
washed four times for 15 min with 1 ml of solution A and overlaid for 30 min with the fluorescein isothiocyanate-conjugated goat anti-rabbit antibody (AffiniPure F(ab The
previously purified wheat germ CPDase (18) was subjected to tryptic
digestion, and three peptide sequences were obtained. One of these, the
20-amino acid-long pep3, showed 80% identity and 100% similarity to
an open reading frame of the Arabidopsis EST, present in the
GenBankTM/EBI data base (accession number T12916[GenBank]). The EST cDNA was
used as a probe to screen the The coding region of the
Arabidopsis cDNA was subcloned in the pQE-60-inducible
expression vector to yield a fusion protein containing six histidine
residues at the C terminus. The tagged protein was overproduced in
E. coli and purified using the nickel-nitrilotriacetic acid
resin (Fig. 3A). The protein was over 95%
pure as judged by SDS-PAGE. The polyclonal antibodies, raised in
chickens immunized with the overexpressed Arabidopsis
CPDase, detected purified Arabidopsis protein on Western
blots and also cross-reacted with the purified wheat CPDase (Fig.
3B). The antibodies did not detect the CPDase in crude
cellular extracts prepared from the leaves of Arabidopsis, but the protein band likely to correspond to the CPDase could be
detected after partial purification of the enzyme (data not shown).
Hence, consistent with previous observations (18), the CPDase appears
to be a nonabundant protein.
The Arabidopsis CPDase hydrolyzed all four nucleoside
2
The wheat CPDase was previously shown to cleave the 1",2"-cyclic
phosphate linkage in the enzymatically produced 32P-labeled
Appr>p to generate Appr-1"p (17). We have used chemically synthesized
Appr>p (33) to demonstrate that the Arabidopsis enzyme has
similar activity. The availability of larger amounts of Appr>p also
allowed us to determine the kinetic parameters for its hydrolysis by
the wheat and Arabidopsis CPDases (see below). Hydrolysis of
Appr>p by the Arabidopsis or wheat enzyme yielded products
of identical mobility in two different TLC systems (Figs. 5,
A and B, lanes 2 and
3). Following additional treatment with CIP,
in both cases the compound comigrated with ADP-ribose (Fig. 5,
A and B, lanes 4 and 5). Treatment of
Appr>p with RNase T2 yielded the Appr-2"p isomer, which chromatographs
more slowly than Appr-1"p on the polyethyleneimine plate (Fig.
5B, lane 7; Ref. 17). As expected, the RNase T2 hydrolysis
product was also sensitive to CIP, yielding ADP-ribose (Fig. 5B,
lane 8).
Reaction requirements of the Arabidopsis CPDase were
determined using A>p as a substrate (Fig. 6). Triton
X-100 stimulated the enzyme activity, and 0.01% detergent was included
in all reactions. The amount of substrate hydrolyzed was linearly
dependent on enzyme concentration up to 20 ng/20 µl. The rates of
A>p hydrolysis were similar at 20, 30, and 37 °C. The optimal
activity was found at pH 7.0. These reaction requirements are similar
to those of the wheat enzyme (18).
The effects of mono- and divalent cations (chloride salts) and EDTA,
previously tested with the wheat enzyme, were determined. Addition of
NaCl to 0.2 or 0.4 M inhibited A>p hydrolysis by 15 and
31%, respectively. Cu2+ and Zn2+ at 0.5 mM inhibited A>p hydrolysis by 93 and 87%, respectively. At 0.5 mM, Mn2+ slightly stimulated (by 5%)
enzyme activity, whereas Ca2+, Mg2+,
Co2+, Ni2+, and EDTA showed no effect at 0.5 mM and were only weakly inhibitory at 10 mM
(data not shown). All these results are consistent with those obtained
for the wheat CPDase (18).
The Km and Vmax for the
Arabidopsis CPDase were estimated for four nucleoside
2 Table I.
Substrate specificity of the Arabidopsis and wheat CPDases
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
,5
-phosphodiester, 2
-phosphomonoester
linkage contain a 2
-phosphate group adjacent to the tRNA anticodon.
This 2
-phosphate is transferred to NAD to give adenosine diphosphate
(ADP)-ribose 1",2"-cyclic phosphate (Appr>p), which is subsequently
metabolized to ADP-ribose 1"-phosphate (Appr-1"p). The latter reaction
is catalyzed by a cyclic phosphodiesterase (CPDase), previously
identified in yeast and wheat. In the work presented here, we describe
cloning of the Arabidopsis cDNA encoding the 20-kDa
CPDase that hydrolyzes Appr>p to Appr-1"p. Properties of the
bacterially overexpressed and purified Arabidopsis enzyme
are similar to those of wheat CPDase. In addition to their
transformation of Appr>p, both enzymes hydrolyze nucleoside
2
,3
-cyclic phosphates to nucleoside 2
-phosphates. For the
Arabidopsis CPDase, the apparent Km
values for Appr>p, A>p, C>p, G>p, and U>p are 1.35, 1.34, 2.38, 16.86, and 17.67 mM, respectively. Southern analysis
indicated that CPDase in Arabidopsis is encoded by a single
copy gene that is expressed, at different levels, in all
Arabidopsis organs that were analyzed. Indirect
immunofluorescence, performed with transfected protoplasts, showed that
CPDase is localized in the cytoplasm. Based on substrate specificity
and products generated, the plant enzyme differs from other known
cyclic phosphodiesterases. The Arabidopsis CPDase does
not have recognizable structural similarity or motifs in common with
proteins deposited in public data bases.
-half terminating in a 2
,3
-cyclic phosphate
and a 3
-half bearing a 5
-hydroxyl group (3-7). In yeast and plants,
these two tRNA exons are ligated to give an unusual
3
,5
-phosphodiester, 2
-phosphomonoester linkage. This reaction,
catalyzed by the RNA ligase (8-11), is a multistep process resulting
in formation of the mature length tRNA containing a 2
-phosphate at the
splice junction (1, 2) (Fig. 1). The ligation pathway leading to the
formation of the 2
-phosphate-bearing tRNA molecules is also conserved
in vertebrates (12), despite the fact that in these organisms most of
the tRNA splicing appears to involve another RNA ligase, an enzyme
which joins two tRNA halves by the regular 3
,5
-phosphodiester (3, 4,
13).
Fig. 1.
Eukaryotic tRNA splicing pathways and
formation of Appr>p and its hydrolysis product Appr-1"p. For
additional details and references, see text.
[View Larger Version of this Image (24K GIF file)]
-phosphate present in the product of the spliced tRNA is removed
by a specific phosphotransferase, previously identified in yeast and
vertebrates (14, 15). Culver et al. (16) found that this
enzyme transfers the 2
-phosphate to an NAD acceptor molecule, to
produce ADP-ribose 1",2"-cyclic phosphate
(Appr>p).1 However, Appr>p is not the
final product of this complex series of reactions. It has been found
recently that Appr>p is converted into ADP-ribose 1"-phosphate
(Appr-1"p) by the action of the cyclic phosphodiesterase (CPDase),
identified in yeast and wheat (17). Although all partial reactions
leading to the formation of Appr>p and Appr-1"p have, to date, only
been demonstrated in yeast (16, 17), the available evidence suggests
that both compounds are also produced, as a result of the tRNA splicing
reaction, in plants and vertebrates (12, 15-17). It has been suggested
that Appr>p, or its hydrolysis product, may perform some as yet
unspecified regulatory function(s) in the cell (16). Conservation of
the Appr>p-forming pathway in vertebrates (12, 15, 16), despite the
fact that most of the cellular tRNA in these organisms seems to be
processed by another pathway (see above), offers some support for this
hypothesis.
,3
-cyclic phosphates to nucleoside 2
-phosphates (18). The biological significance of this reaction is not
known, but the ability of the enzyme to convert the 2
,3
-cyclic phosphate to the 2
-phosphate in mononucleotides but not in
cyclic-phosphate-terminated oligoribonucleotides, together with its
ability to hydrolyze Appr>p, clearly distinguishes it from other known
enzymes having the 2
,3
-cyclic phosphate 3
-phosphodiesterase activity
(17, 18; see "Discussion"). Although the yeast phosphodiesterase
shares many characteristics with the wheat enzyme, it has a different
substrate specificity, hydrolyzing Appr>p to Appr-1"p, but having no
detectable activity on nucleoside 2
,3
-cyclic phosphates (17).
Fractionation experiments performed with yeast and wheat germ extracts
indicated that phosphodiesterases described above are, most probably,
the only cellular activities converting Appr>p to Appr-1"p (17).
Plant Material
Fig. 2.
Nucleotide and derived amino acid sequences
of the A. thaliana CPDase and its cDNA. The
upstream end of the longest cDNA isolated from the
Arabidopsis ZAP cDNA library is marked by an
arrow. The short open reading frame in the 5
-terminal
leader sequence is underlined. Termination codons in the
5
-leader are overlined. Putative polyadenylation signal
sequences are double-underlined. Amino acid sequence
comparisons with the sequenced peptides (pep 1, pep 2, and
pep 3) of the wheat CPDase are boxed. Identical
residues are indicated by dashes, while conserved residues
are in bold. The N-terminal amino acid of pep3 is either Ser
(S) or Thr (T). The nucleotide sequence shown in
this figure has been submitted to the EMBL/GenBank/DDBJ Nucleotide
Sequence Libraries under accession number Y11650[GenBank].
[View Larger Version of this Image (50K GIF file)]
ZAP cDNA library,
prepared with a mixture of the poly(A)+ RNA isolated from
24, 48, and 72 h leaf strip cultures (a gift from J. Fleck,
Institut de Biologie Moléculaire des Plantes du CNRS, Strasbourg,
France), was screened with the partial cDNA clone (the
Arabidopsis EST; GenBankTM/EBI accession number T12916[GenBank]; kindly provided by the Arabidopsis Biological Resource
Center at Ohio State University, Columbus, OH) as a probe.
Hybridizations were performed overnight at 42 °C in 5 × SSPE
(SSPE: 0.18 M NaCl, 10 mM
NaH2PO4, 1 mM Na2-EDTA,
pH 7.7) containing 50% formamide, 5 × Denhardt's solution
(100 × Denhardt's solution is 2% Ficoll, 2%
polyvinylpyrrolidone, 2% bovine serum albumin), 1% SDS, and 50 µg/ml denatured salmon sperm DNA. Filters were subsequently washed in
2 × SSC (SSC: 0.15 M NaCl, 15 mM
Na2citrate) and 0.1% SDS for 30 min at 42 °C and in
0.2 × SSC and 0.1% SDS for 30 min at 42 °C. Twenty-nine
clones were isolated after screening 800,000 recombinant phages. After
excision of the phagemids, the inserts were analyzed by restriction
mapping and sequencing of the ends. The longest clones were
subsequently sequenced on both strands.
Terminus by
RACE
end was tagged with the 5
-phosphorylated
oligonucleotide 2 (pCATCTCGAGCGGCCGCATCA) using T4 RNA ligase (22). The
CPDase cDNA sequence was PCR-amplified using oligonucleotide 3 (GTGAGGAACGAATCTTGG; complementary to positions 202 to 219; Fig. 2) and
oligonucleotide 4 (TGATGCGGCCGCTCGAGA; complementary to the 3
tag) as
primers. The PCR products were cloned into pBluescript (Stratagene) and sequenced.
-32P]dCTP (3000 Ci/mmol,
Amersham) by the random priming method (29). RNA as well as DNA gel
blots were hybridized overnight at 42 °C in 5 × SSPE, 50%
formamide, 10% dextran sulfate, 1% SDS, and 50 µg/ml denatured
salmon sperm DNA. The blots were subsequently washed in 2 × SSC
and 0.1% SDS for 30 min at 42 °C and in 0.2 × SSC and 0.1%
SDS for 30 min at 42 °C and then at 60 °C.
to the CPDase coding sequence
by site-directed mutagenesis (30). The NcoI/BamHI
fragment (the NcoI site is present at the AUG initiation
codon of the CPDase cDNA) was cloned into the pQE-60 vector
(Qiagen) yielding plasmid pQECPDase. In this construct, 10 additional
amino acids (sequence GSRSHHHHHH) are placed in frame at the C terminus
of the recombinant protein. The protein remained soluble during
expression in the Escherichia coli strain BL21(DE3) and was
purified in the native form, under nondenaturing conditions, using the
nickel-nitrilotriacetic acid resin and following the Qiagen protocol.
The purified CPDase was applied to a 10-ml Sephadex G-25 column
equilibrated and eluted with 20 mM Tris acetate, pH 7.6, 0.5 mM dithiothreitol, 0.1 mM EDTA, 0.5 mM dithiothreitol, 5% (v/v) glycerol, 0.01% Triton X-100, and 10 µM phenylmethylsulfonyl fluoride. The protein
concentration was measured by the method of Bradford (31) using bovine
serum albumin as a standard.
,3
-cyclic and 3
,5
-cyclicphosphates, nucleoside 5
-, 3
- and
2
-phosphates, pG>p and inositol 1,2-cyclic phosphate were obtained
from Sigma and Pharmacia Biotech Inc. A>p, G>p, and C>p were
purified by reverse phase HPLC on Nucleosil C4 RP-300 column using 50 mM triethyl-ammonium acetate-acetonitrile gradient.
Products were collected as a single peak. Appr>p was chemically
synthesized as described elsewhere (33). The 32P-labeled
oligoribonucleotide AAAAUAAAAG>p* (asterisk denotes the position of
the label) was prepared as follows: the synthetic oligoribonucleotide
AAAAUAAAAG (0.7 µg), obtained from MWG-Biotech (Munich, Germany), was
3
-terminally labeled using 5
-[32P]pCp (*pCp) and T4 RNA
ligase. The ligation product was then digested with 5 units of RNase T1
yielding AAAAUAAAAGp*. The latter was quantitatively converted into
AAAUAAAAG>p* by incubation with the RNA 3
-terminal phosphate cyclase
purified from HeLa cells (34). The oligonucleotide was recovered by
phenol extraction and ethanol precipitation.
)2 fragment; Jackson/Milan
Analytica AG). The secondary antibody was diluted 1:100 with buffer B,
containing 10 µg/ml Hoechst 33258 dye. After washing four times with
1 ml of solution A and four times with 1 ml of solution C (0.1 M glycine-KOH, pH 8.5), samples were overlaid with a drop
of the embedding material and covered with a cover glass. Samples were
examined with a Zeiss Axiophot microscope and a Leica TCS 4D confocal
scanning laser microscope, using a 63× objective. Images were recorded
using the Leica software (SCANware 4.2) provided with the system and analyzed with the Imaris software on a Silicon Graphics work
station.
Cloning of the cDNA Encoding Arabidopsis CPDase
ZAP cDNA library made with the
poly(A)+ RNA obtained from the Arabidopsis leaf
strip culture. Twenty-nine positive recombinant phages were isolated,
and their inserts were analyzed by restriction mapping and sequencing
of the ends. The longest cDNA obtained from this screening started
approximately 85 nucleotides downstream from the transcription
initiation site, as determined by primer extension (data not shown). To
determine the 5
-terminal mRNA sequence, RACE experiments were
performed using RNA isolated from leaf strip cultures incubated for
either 24 or 48 h. Both RNA preparations yielded cDNA clones
of identical sequence and extending to the same position (position 1 in
Fig. 2). The sequence of the region identified by RACE was
independently confirmed by cloning, using inverse PCR, and sequencing
the promoter region of the gene. The apparent full-length cDNA is
741 nucleotides long, without counting the poly(A) tail (Fig. 2). The
sequence including the presumed AUG initiation codon
(AUCCAUGGA) is similar to the consensus
(AACCAUGGC) established for plant genes (39). The
5
-terminal leader contains one additional AUG in a much less favorable
context, followed by termination codons (Fig. 2). Conceptual
translation of the cDNA yields a 20.5-kDa protein of 181 amino
acids with a predicted isoelectric point of 4.82. The deduced
Arabidopsis protein contains sequences showing significant
similarities with all sequenced peptides derived from the wheat
protein. The greatest sequence homology is for pep3 (see above).
Peptides pep1 and pep2 show 39 and 38% similarity and 39 and 31%
identity, respectively.
Fig. 3.
Gel electrophoresis of the purified
Arabidopsis histidine-tagged CPDase (A) and
immunological relationship to the wheat enzyme (B).
A, a total of 200 ng (lane 1) and 400 ng
(lane 2) of the affinity-purified CPDase was applied to the
10% gel. The gel was stained with Coomassie Blue. B,
SDS-PAGE immunoblot of purified Arabidopsis and wheat
CPDases, probed with the polyclonal antibody raised against the
Arabidopsis protein. No immunoreactive material was detected
with the preimmune antibody (data not shown). Lanes 1 and
2, 20 and 40 ng, respectively, of the Arabidopsis CPDase; lanes 3 and 4, 10 and 50 ng,
respectively, of the purified wheat CPDase. Molecular mass markers are
indicated.
[View Larger Version of this Image (28K GIF file)]
,3
-cyclic phosphates to the corresponding 2
-phosphomonoesters as analyzed by cellulose TLC (Fig. 4). No
3
-phosphomonoester formation could be detected. Thus, like the wheat
protein, the Arabidopsis enzyme has 2
,3
-cyclic nucleotide
3
-phosphodiesterase activity. Nucleoside 3
,5
-cyclic phosphates
(3
,5
-cAMP and 3
,5
-cGMP), inositol 1,2-cyclic phosphate and
guanosine 5
-phosphate, 2
,3
-cyclic phosphate (pG>p), and also the
cyclic phosphodiester in the 2
,3
-cyclic phosphate-terminated
oligoribonucleotide (AAAAUAAAAG>p*, the asterisk indicates a position
of the 32P label) were not hydrolyzed (data not shown);
these compounds are also not substrates for the wheat enzyme (17,
18).
Fig. 4.
Hydrolysis of the 2,3
-cyclic nucleotides by
the Arabidopsis CPDase. Reaction mixtures of 10 µl
were incubated with (even-numbered lanes) or without
(odd-numbered lanes) 200 ng of CPDase and contained the
following nucleotides (each 20 mM): A>p (lanes
1 and 2), G>p (lanes 3 and 4),
C>p (lanes 5 and 6), U>p (lanes 7 and 8). A total of 4 µl of each sample was subjected to
cellulose TLC in solvent A. Positions of nucleotide markers are
indicated.
[View Larger Version of this Image (63K GIF file)]
Fig. 5.
Hydrolysis of Appr>p by the
Arabidopsis and wheat CPDases and characterization of the
hydrolysis product. Appr>p was treated with either wheat
(lanes 2 and 4) or Arabidopsis
(lanes 3 and 5) CPDase or with RNase T2
(lane 7). The samples in lanes 4, 5, and
8 were additionally treated with calf intestinal phosphatase (CIP). Lane 1, untreated Appr>p; lane
6, ADP-ribose marker. A, cellulose TLC in solvent A. B, polyethyleneimine-cellulose in solvent B. Appr>p was not
sensitive to CIP treatment as analyzed by TLC in both systems (not
shown). Note: use of parentheses indicates that no standard
is available for comparison and the identification of the molecule is
based on other criteria (see text).
[View Larger Version of this Image (39K GIF file)]
Fig. 6.
Hydrolysis of A>p by the Arabidopsis
CPDase. A, dependence on enzyme concentration. Assays
were incubated for 30 min at 30 °C. B, kinetics of
hydrolysis at different temperatures. Assays (20 µl) contained 20 ng
of CPDase. C, pH optimum. Assays (20 µl), containing 20 ng
of CPDase, were incubated for 30 min at 30 °C. The following buffers
were used: Mes-NaOH (), Mops-NaOH (
), and Tris-HCl (
).
[View Larger Version of this Image (16K GIF file)]
,3
-cyclic phosphates and Appr>p and compared with those obtained
with the purified wheat enzyme (Table I).
Km values for individual nucleotide substrates were
comparable for both enzymes (but see legend to Table I). However, the
Vmax values obtained with the
Arabidopsis CPDase were 10-25 times lower than those
measured with the wheat enzyme. It is possible that only a fraction of
the Arabidopsis protein overexpressed in E. coli
is enzymatically active. Alternatively, activity of the overexpressed
enzyme may be lower due to the presence of the histidine tag or the
absence of some essential modification of the protein. Based on
relative Vmax/Km values, the specificity of the Arabidopsis CPDase toward the cyclic
nucleotides is C>p:A>p:Appr>p:U>p:G>p = 100:54:36:6:6, whereas for the wheat CPDase the specificity is
A>p:C>p:Appr>p:U>p:G>p = 100:43:39:9:7. Hence,
both enzymes have similar substrate specificities with a preference for
A>p, C>p, and Appr>p over U>p and G>p.
,3
-cyclic phosphates are comparable.
Parameters in the table are the weighted means, together with their
standard errors, of at least two determinations. In the case of
A>p, G>p, and C>p, the Km and
Vmax values determined with the wheat CPDase differ
from the values obtained previously by Tyc et al. (18)
(their Km and Vmax values were: 13.1 and 2094 (A>p), 9.2 and 276 (G>p), 25.2 and 2140 (C>p)). The reason for the discrepancy is not known. Note that
A>p, G>p, and C>p used in this work were purified by
HPLC (see "Experimental Procedures"). Based on the
Vmax/Km values, determined previously (18) and now, the relative specificities of the wheat enzyme
toward four nucleoside 2
,3
-cyclic phosphates are comparable.
Substrate
Arabidopsis
Wheat
Vmax
Km
Vmax/Km
Vmax
Km
Vmax/Km
µmol/min/mg
mM
×
103
µmol/min/mg
mM
×
103
A>p
25 ± 2
1.34
± 0.24
18.5 ± 3.7
597 ± 29
1.09 ± 0.13
547
± 72
G>p
37 ± 4
16.86 ± 2.57
2.2
± 0.4
917 ± 121
24.44 ± 3.77
37 ± 7
C>p
79 ± 7
2.38 ± 0.39
33.3 ± 6.2
695
± 40
3.01 ± 0.27
231 ± 24
U>p
47
± 3
17.67 ± 1.61
2.6 ± 0.3
762 ± 116
15.16
± 2.85
50 ± 12
Appr>p
16 ± 2
1.35
± 0.19
12.1 ± 2.2
261 ± 34
1.21 ± 0.23
216
± 49
The CPDase gene copy number was
estimated in a Southern blot analysis. Only one hybridizing band was
detected in DNA digests carried out with three different restriction
enzymes, consistent with the existence of a single-copy gene (Fig.
7).
Expression of the CPDase gene in various tissues of
Arabidopsis plants and in germinating seeds and young
plantlets was analyzed by Northern blotting (Fig.
8A). The CPDase mRNA is relatively low in
abundance, consistent with the results of Western analysis (see above).
Roots contained slightly higher levels of mRNA than other tissues
analyzed. Still higher levels of mRNA were found in 3-week-old
Arabidopsis plantlets.
Expression of the CPDase gene was also investigated during re-initiation of mitotic activity in leaf strip cultures. When Arabidopsis leaf strips are incubated in a culture medium containing 1 mg/liter of the auxin analogue 2,4-dichlorophenoxyacetic acid, the cells start to proliferate very rapidly. [3H]Thymidine incorporation and Northern blot hybridization, performed with the histone H4 cDNA as a probe, have shown that 48 h after starting the culture most of the cells are in the S phase (40) (see also Fig. 8B). Using this experimental system, the highest CPDase mRNA level was found at 72 h (Fig. 8B), a time when a high number of cell divisions is observed.2 RNase A/T1 mapping, performed with the antisense RNA probe covering the coding region of the CPDase cDNA, also indicated that after 72-h incubation, the level of the CPDase mRNA is approximately five times higher than at the start of the culture (data not shown). The significance of this mRNA accumulation is not understood at present.
Cellular Localization of the CPDase Protein Studied by Indirect ImmunofluorescenceThe intracellular localization of the
Arabidopsis CPDase was determined by an epitope tagging
approach combined with indirect immunofluorescence. The coding sequence
of the CPDase cDNA was cloned in a plant expression vector with an
influenza hemagglutinin (flu) epitope fused in frame to the C terminus
of the protein. The plasmid expressing the tagged protein was
transfected into mesophyll protoplasts of N. plumbaginifolia. The protoplasts were processed for
immunofluorescence microscopy, using a rabbit anti-flu polyclonal
antibody and fluorescein isothiocyanate-conjugated goat anti-rabbit
antibody. Indirect immunofluorescence showed that the protein is
cytoplasmic (Fig. 9B). No staining was seen in untransfected protoplasts visible in the same field (A
and B) or in mock-transfected protoplasts (data not shown).
Transfected protoplasts (B), in which the nucleus was
localized by staining with Hoechst 33258 (A), were also
examined by confocal microscopy. The expressed tagged protein was
clearly excluded from the nucleus and the chloroplasts (C).
In control experiments, in which the flu tag was fused to the protein
N-RBP43 known to be targeted to the
nucleus,3 immunofluorescence was
predominantly nuclear (D-F). These results indicate that
the CPDase is a cytoplasmic protein.
The Arabidopsis CPDase Does Not Share Significant Sequence Similarity with Other Known Phosphodiesterases
A search of
current sequence data bases did not reveal any proteins having
significant sequence similarity with the Arabidopsis CPDase.
The consensus signature motifs of 3,5
-cyclic nucleotide phosphodiesterases (41, 42) are not present in the
Arabidopsis enzyme. Cyclic phosphodiesterases from brain
(43, 44), and tRNA ligase (45), two proteins having
3
-phosphodiesterase activity, also have no significant
similarity with the plant protein.
Removal of the 2-phosphate group present in the products of tRNA
splicing (see Introduction) is catalyzed by a specific
phosphotransferase, first identified in yeast and vertebrates by
Phizicky and co-workers (14, 15). In this reaction, the 2
-phosphate
from tRNA is transferred to give a cyclic phosphate at the 1"-2"
positions of NAD, yielding an unusual ADP-ribose derivative, Appr>p,
and nicotinamide (16). Culver et al. (46) have recently
cloned the gene encoding this enzyme in yeast and demonstrated that it is essential for viability. We describe here isolation of the Arabidopsis cDNA and properties of the encoded enzyme
that metabolizes Appr>p to Appr-1"p. Phosphodiesterases that hydrolyze
Appr>p to Appr-1"p have been previously characterized biochemically in
both yeast and wheat (17). The purified wheat enzyme also hydrolyzes nucleoside 2
,3
-cyclic phosphates in addition to Appr>p. The enzyme
partially purified from yeast does not accept nucleoside 2
,3
-cyclic
phosphates as substrates, but otherwise, its properties are similar to
that of the wheat protein (17).
Evidence presented here indicates that the protein encoded by the
Arabidopsis cDNA and the protein purified previously
from wheat (17, 18) are equivalent. Both enzymes have similar substrate specificity, showing preference for A>p, C>p, and Appr>p over U>p
and G>p. Significantly, neither of the enzymes ring-opens the
2,3
-cyclic phosphate group of pG>p or at the terminus of an
oligoribonucleotide (Ref. 18; this work), a property that distinguishes
these enzymes from other known cyclic nucleotide 3
-phosphodiesterases
(see below). The Arabidopsis and wheat proteins have similar
molecular weights, show immunological cross-reactivity, and, based on
the limited sequence information available for the wheat protein (Fig.
2), are structurally related. Finally, the reactions catalyzed by both
enzymes have identical pH and temperature dependence and are similarily
affected by various inhibitors.
Southern analysis has indicated that CPDase in Arabidopsis is encoded by a single copy gene (Fig. 7). In previous experiments, only single enzymatic activity for the hydrolysis of Appr>p has been observed upon fractionation of yeast extracts, while chromatography of the wheat germ extract on DEAE-cellulose yielded two pools of CPDase, having similar specificity toward N>p and Appr>p substrates (17, 18). It is not known whether the two active pools represent products of two separate genes or two forms of the same protein.
With respect to substrate specificity, the plant CPDase studied in this
work is clearly different from that of other known proteins possessing
cyclic 3-phosphodiesterase activity. The RNA ligase involved in tRNA
splicing, extensively characterized for yeast and wheat (see
Introduction), efficiently hydrolyzes terminal 2
,3
-cyclic phosphates
in oligoribonucleotides (10, 47) but is only poorly active on
2
,3
-cyclic mononucleotides (18). The vertebrate cyclic nucleotide
phosphodiesterase (CNPase), isolated from brain, hydrolyzes
2
,3
-cyclic phosphates in both mono- and oligoribonucleotides (48, 49)
(for activity of CNPase with oligonucleotide substrates, see also Refs.
47 and 50). Most importantly, however, neither of the two enzymes
mentioned above accepts Appr>p as a substrate (17), in marked contrast to the plant CPDase. Based on substrate specificity and generated products, the plant enzymes also differ from the broad specificity vertebrate cyclic phosphodiesterase described by Helfman and Kuo (51).
Consistant with the differences discussed above, the
Arabidopsis CPDase does not have recognizable structural
similarity or motifs in common with the phosphodiesterase domain of the
tRNA ligase (52), with brain CNPase (43, 44), or with other known cyclic phosphodiesterases or RNases, acting on either nucleoside 3,5
-
or 2
,3
-phosphates (41, 53) or on inositol 1
,2
-phosphate (54).
Despite the fact that plant and yeast CPDase share many characteristics
(17), a search of the yeast fully sequenced genome with the
Arabidopsis protein sequence as a query did not identify a
candidate gene coding for the CPDase in yeast. This is perhaps not
surprising, seeing that even wheat and Arabidopsis sequences
diverge considerably (Fig. 2). A search of the public data bases also
did not reveal proteins having significant sequence similarity with the
Arabidopsis enzyme. Thus, the CPDase characterized in this
work belongs to a new group of cyclic phosphodiesterases.
The finding that the 3,5
-phosphodiester, 2
-phosphomonoester-forming
pathway of RNA ligation is conserved in vertebrates, despite the fact
that most of the tRNA splicing in these organisms proceeds via a
different route (see Introduction), led to speculation that production
of Appr>p or Appr-1"p may be important for the cell (2, 16, 17). Other
ADP-ribose and NAD derivatives such as cyclic adenosine 5
-diphosphate
ribose (cADPR) (55, 56) or nicotinic acid adenine dinucleotide
phosphate (NAADP) (57, 58) are known to act as potent Ca2+
release agents in sea urchin eggs and microsomes and, as demonstrated for cADPR, also in mammalian and plant cells (59, 60). Using purified
CPDase and chemically synthesized Appr>p, we have produced Appr-1"p,
and both compounds were tested for calcium mobilizing activity in sea
urchin egg homogenates. Neither Appr>p nor Appr-1"p were found to
effect calcium release or modify calcium release induced by
cADPR.4 Cloning of the cDNA from
Arabidopsis should allow the use of reverse genetics to
manipulate levels of CPDase in this organism and test whether Appr>p
or Appr-1"p serve any regulatory function. Availability of this
cDNA could also be of help in investigating a potential role of
nucleoside 2
,3
-cyclic phosphates or nucleoside 2
-phosphates in
plants. Schildknecht (61) has reported that G>p and A>p may function
as turgor-regulating leaf movement factors in mimosa plants.
It is equally possible that neither substrates or products of the plant
CPDase are biologically active, and the CPDase is a catabolic enzyme
channeling Appr>p and 2,3
-cyclic nucleotides, the latter being
produced by degradation of cellular RNAs by cyclizing endoribonucleases
(62-65), into respective degradation pathways (17, 18).
We thank the Arabidopsis Biological Resource Center at Ohio State University (Columbus, OH) for providing the Arabidopsis EST clone, S. Burkhardt for assistance with Appr>p synthesis, M. Swianiewicz and H. Ryan for assistance with some experiments, Dr. A. Galione (University of Oxford, Oxford, United Kingdom) for testing Appr>p and Appr-1"p for calcium release activity, Dr. J. Hofsteenge for valuable discussions, and Drs. F. Dragon and J. Hofsteenge for critical reading of the manuscript.