ABSTRACT
The gene encoding the primase small subunit was isolated from genomic DNA of
strain K1 of the human malarial parasite
Plasmodium falciparum
. Isolation of a complete cDNA clone revealed the presence of 15 introns in the genomic
sequence. This is unprecedented for
Plasmodium
genes, which usually contain no or only 1 or 2 introns. The gene is present as
a single copy and the cDNA contains an open reading frame of 1356 nt encoding a
protein of 452 amino acids. A single mRNA of 2.1 kb was identified by Northern
blotting. Comparison of the amino acid sequence with five eukaryotic small primase subunits revealed the presence of eight conserved regions. Sequence alignments allowed the identification of
putative motifs A, B and C that are essential features of the catalytic centre
of DNA polymerases, RNA polymerases and reverse transcriptases. Also,
similarity of a C-terminal region of
~
100 amino acids to a conserved region in herpes virus primases,
[alpha]
-like DNA polymerases and RNA polymerase II was noted. The complete gene
was expressed as a fusion product containing an N-terminal polyhistidine tag using a baculovirus expression vector. The protein was
overproduced in insect cells and purified. Activity assays demonstrated the
ability of the p53 subunit to initiate
de novo
primer formation.
Nuclear chromosomal DNA replication and its regulation involves the highly coordinated interaction of many proteins, substrates and DNA (
1
,
2
). Three DNA polymerases, designated [alpha], [delta] and [epsilon] are absolutely required for DNA synthesis (
3
,
4
). DNA polymerase [alpha] is essential for initiation of DNA synthesis of both the lagging and the
leading strand. In virtually all eukaryotic organisms analysed so far, DNA
polymerase [alpha] has been isolated as a four subunit complex consisting of polypeptides
of 165-182, 68-86, 54-60 and 46-50 kDa. When separated from the complex only the largest subunit displays DNA polymerase activity
(
5
), whereas the two smallest subunits can be dissociated from the complex as a heterodimer
containing primase activity (
6
).
Primase is absolutely essential to chromosomal DNA replication by catalysing
de novo
synthesis of discrete length oligoribonucleotides (7-10 bp) that constitute the primers required for subsequent DNA synthesis by the polymerase [alpha] subunit (
7
-
9
). Recent studies have yielded ambiguous results on the mechanism of action of the primase
heterodimer with respect to the question of whether or not the smallest subunit
alone is capable of
de novo
initiation of primer synthesis (
10
-
13
). Cloning and heterologous expression of the
Plasmodium falciparum
primase small subunit described in this report will help to resolve this
controversy and give us a better understanding of the fundamental mechanisms of
action.
Another major reason for investigating the replication machinery of malaria parasites, which rely completely on their own set of replication
enzymes (
14
), is its potential suitability as a target for new anti-malarial drugs. DNA polymerases have proven to be important target enzymes
for anti-microbial chemotherapy. We anticipate that primases are equally important
targets, as no other RNA polymerase can substitute for its action (
15
). Indeed, it has recently been shown that primase can be a target of inhibition
for several nucleotide analogues (
16
-
19
).
The presence of primase activity in malaria parasites was demonstrated in our efforts to purify and characterize
Plasmodium
DNA polymerases (
20
). Some of the identified DNA polymerases appeared to be
in vitro
targets of inhibition of a new class of nucleotide analogues and this has led
us to extensive
in vitro
and
in vivo
studies on the possibilities of using acyclic phosphonate analogues of
adenosine as anti-malarial drugs (
21
-
23
).
Plasmidium falciparum
strain K1 (
24
) was grown in culture as described (
21
,
25
). Parasites from asynchronous cultures were isolated by saponin lysis (
21
) and DNA and RNA were prepared according standard procedures (
26
). Poly(A)
+
RNA was purified by chromatography on oligo(dT)-cellulose (
26
). Southern and Northern blotting were performed as described (
26
).
The sequences of degenerate 23mer oligonucleotide P1 [5'-GAATT (A/T)GT(A/T)TTTGATAT(A/T)GATAT] and 24mer P2 [5'-CCAACA(G/A)TG(T/A)AC(T/A)CCTCTTCT(T/A)CC] were derived from two highly conserved regions [ELVFDIDM
and GRRG (A/V)HCW] between mouse (
27
) and yeast (
28
) small subunit primase using the preferred
P.falciparum
codon usage (
29
). PCR was performed on 100 ng K1 DNA with 200 ng P1 and P2 (1 min 95oC, 2 min 40oC, 3 min 65oC, 40 cycles, 1.3 mM MgCl
2
). A PCR product of ~300 bp was isolated from a low melting point agarose gel (
26
), cloned into the
Hin
cII site of pUC21 and sequenced. The cloned fragment (clone PriPCR7) was used as
a probe on Southern blots with multiple digests of K1 DNA in order to map
suitable restriction sites (see map in Fig.
1
). Fragments within a selected size range of digests of K1 DNA with
Hin
dIII/
Xba
I (3000-6000 bp) and
Bsp
HI (1500-4000 bp) were isolated from a low melting point agarose gel (
26
) and cloned into the
Hin
dIII/
Xba
I site and
Nco
I sites respectively of pUC21. The resulting libraries were screened with a
32
P-labelled insert of PriPCR7 and clones PriHX321 and PriBB251 were isolated.
Convenient fragments were subcloned into pUC20 and pUC21 and both strands were sequenced (T7 Sequencing Kittm; Pharmacia). Occasionally gaps in the sequence were closed by using specifically designed sequence primers.
Five hundred nanograms of oligonucleotide P4 (5'-AAATTAGTAAAAATGCTGGTACA), mapping 97 bp downstream of the putative
TAA stop codon, were hybridized to 5 [mu]g total K1 RNA (5 min, 65oC in H
2
O) and elongated by 20 U Superscripttm RNase H
-
reverse transcriptase for 2 h at 42oC (50 mM Tris-HCl, pH 8.3, 75 mM KCl, 3 mM MgCl
2
, 2.5 mM dNTP). The cDNA generated was amplified by PCR with 200 ng sense primer
P3 (5'-TTTTTTATTTACATTTCTTTTGA), mapping 32 bp upstream of the putative ATG start codon, and 200 ng primer P4 (30 s 95oC, 45 s 50oC, 1 min 70oC, 1.5 mM MgCl
2
). A PCR fragment of ~1500 bp was cloned in pCRtmII using the TA Cloningr Kit (Invitrogen) (clone PricDNA3). Convenient fragments were subcloned and sequenced. Comparison of the 1356 bp open reading frame (ORF)
with the genomic sequence identified 15 introns. Stop codons in-frame with the coding sequence were present within the cDNA at the 5'- and 3'-ends, indicating that the complete gene was cloned and no additional 5' or 3' protein encoding exons are
present.
Two hundred nanograms of oligonucleotide P5 (5'-GATAATACTTCATGAAAATGG), creating a
Rca
I site at the initiating methionine codon (position 486 according to the map in Fig.
1
), and 200 ng oligonucleotide P6 (5'-CTTAAATTCTGCAGGGTTACT), creating a
Pst
I site 253 bp downstream of the translation start site, were hybridized to 200 pg clone PricDNA3 and amplified by PCR (1 min 95oC, 1 min 48oC, 1 min 70oC). The resulting PCR fragment was digested with
Rca
I and
Pst
I and cloned into the
Nco
I and
Pst
I sites of pAcSG His NT-B (PharMingen). Sequence analysis of the resulting clone (PriPCR11)
confirmed the sequence. Subsequently, the 1235 bp
Bcl
I-
Nsi
I fragment of clone PricDNA3 was cloned into the
Bcl
I/
Pst
I sites of PriPCR11 to give clone BacPri53cDNA, encoding the complete primase fused to an N-terminal hexahistidine tag. Recombinant baculovirus was produced by co-transfection of Sf9 insect cells with BacPri53cDNA and BaculoGoldtm viral DNA (PharMingen) according the recommended procedures (PharMingen). The integrity of
the recombinant virus was checked by Southern blotting.
Sf9 insect cells were grown in TMN-FH medium containing 10% fetal calf serum (PharMingen) or in Grace's
medium (Gibco) supplemented with lactalbumin hydrolysate and Yeastolate (Gibco) containing 10% fetal calf serum according to standard procedures (
30
). Cells were infected with ~10 plaque forming units of recombinant baculovirus per cell and harvested
after 72 h. Cells were washed twice with phosphate-buffered saline. Approximately 2 * 10
7
cells were resuspended in 1 ml lysis buffer (100 mM NaCl, 50 mM Tris-HCl, pH 7.5, 1% NP40) and incubated for 10 min at 0oC. The extract was centrifuged (20 000
g
, 4oC, 15 min) and the supernatant was adjusted to 10% glycerol and stored at -20oC for further analysis (fraction S). The pellet was resuspended
in 1 ml pellet extraction buffer (500 mM NaCl, 100 mM Tris-HCl, pH 8,4, 1% Triton X-100, 15% glycerol), sonicated (2 * 5 s at 0oC) and centrifuged (20 000
g
, 4oC, 15 min). The supernatant was stored at -20oC (fraction P) and the pellet resuspended in 0.5 ml
denaturation buffer (6 M guanidinium-HCl, pH 7.5, 100 mM Tris-HCl, pH 7.5), stirred for 15 min at room temperature and
centrifuged (20 000
g
, 4oC, 15 min). The supernatant was stored at -20oC (fraction D) until purification on Ni-NTA-agarose. Ni-NTA-agarose was equilibrated by
repeated washing in denaturation buffer. Column material (200 [mu]l packed volume) was added to 1 ml fraction D and incubated for 3 h at room
temperature on a rotating platform. The slurry was packed in a column and the
bound denatured protein was allowed to re-fold by applying a linear gradient of 6-1 M urea (5 ml, 1.5 h). Protein was eluted with a linear gradient
of 0-0.5 M imidazole in 50 mM sodium phosphate, pH 6.5, 200 mM NaCl, 10% glycerol (15 ml, 2 h). Fractions containing recombinant primase were pooled, dialysed (50 mM
Tris-HCl, pH 7.5, 100 mM KCl, 20% glycerol) and stored in small aliquots for
use in activity assays (fraction NA).
Maximal primase activity was obtained when 500 ng poly(dT) were incubated with
purified recombinant primase (fraction NA) in a buffer containing 50 mM NaCl, 2
mM MgCl
2
, 200 [mu]M ATP, 0.2 mg/ml BSA, 50 mM MOPS (20 [mu]l), pH 7.6, for 15 min at 30oC. Subsequently the newly formed oligo(A) primers were elongated by
the addition of 0.75 U Klenow and 200 [mu]M [[alpha]-
32
P]dATP (0.4 Ci/mmol) and incubation for 30 min at 30oC. The reaction was terminated by the addition of 250 [mu]l 20 [mu]M EDTA, 0.1 mg/ml BSA and the amount of radioactivity incorporated
into nucleic acid was determined by precipitation with 100 [mu]l 50% trichloroacetic acid. Incorporation of radioactivity showed a linear increase for primase concentrations of 2-100 ng/ 20 [mu]l and for a timespan of 10-30 min. Conditions were optimized within these limits.
A cDNA encoding the full-length primase small subunit of
P.falciparum
(strain K1) was obtained by RT-PCR (Fig.
1
A). Specific sense and antisense primers were derived from sequences flanking the primase small subunit gene (Fig.
1
B), which was cloned from
P.falciparum
genomic DNA libraries making use of oligonucleotides based upon two highly
conserved regions between yeast and mouse 49 kDa primase subunits.
Sequence analysis of the 1507 bp cDNA clone revealed a single ORF of 1356 bp
encoding a protein of 53 kDa. The protein is encoded by a gene interrupted by
15 introns (Fig.
1
B) and exists as a single copy gene, as shown by Southern blots of single and
double digests of K1 DNA with 15 restriction enzymes (blots not shown).
Malarial genes have either a few or no introns at all. The one exception
reported to date concerns a 41 kDa bloodstage antigen gene, interrupted by
eight introns, from which several alternatively spliced transcripts are derived
(
31
). All primase small subunit introns are typical examples of malarial introns in
being short and extremely AT rich. Comparison of the 15 primase donor and acceptor splice junctions (Fig.
1
C) with a recent compilation of 48 malarial 5' and 3' splice sites (
32
) shows a substantial deviation from the deduced consensus sequence at only
three positions. Position -1 in the 5' splice site is almost exclusively a purine (94% in 15 primase
introns versus 71% in 48 other
P.falciparum
introns). At position +5 of the 5' splice site a T is preferred in primase introns (53 versus 17%) and at
position +1 of the 3' splice site a G is preferred (53 versus 29%). Adding the 15 primase sequences to the most recent compilation randomizes the
consensus sequence at two positions (from G/A to N at +5 of the 5' splice site and from A to G/A at +1 of the 3' splice site).
Superimposition of the crystal structures of polymerases from three different
families (
Escherichia coli
DNA polymerase I, bacteriophage T7 RNA polymerase and HIV-I reverse transcriptase) identified a remarkably similar folding of the catalytical centres of the
enzymes (
34
-
39
), suggesting a common evolutionary ancestry (
40
). Sequence similarity in between separate, highly conserved polymerase families (
39
-
45
) is limited to a few functionally important residues (
39
,
46
,
47
). Three catalytically essential acidic residues are present at identical
positions at the bottom of a large cleft (the `palm domain', by virtue of the
anatomical similarity of a polymerase to a right hand). They serve as ligands
coordinating two metal ions involved in the nucleotidyl transfer reaction. The
three acidic residues are located within highly conserved sequence motifs A and
C (Fig.
4
) that show only limited similarity in between separate families (
39
,
46
,
47
). Furthermore, in the three-dimensional models of Klenow and T7 RNA polymerase interactions take place
between residues located on one face of a similarly positioned [alpha]-helix and the dNTP and the template strand. This [alpha]-helix (motif B) is located in the structurally less
well conserved `fingers domain' bordering the cleft. The absence of motif B
from HIV reverse transcriptase may reflect the different template requirement (
39
). The catalytically important residues of motifs A-C have putatively been identified in highly conserved, co-linear arranged motifs of several other polymerase families (
39
,
46
).
Figure
3
shows an alignment of eukaryotic primase small subunit sequences. Conserved
regions I-V have been defined from alignments of mammalian (mouse and human) and
yeast sequences (
12
,
27
). In this compilation additional conserved regions are denoted Ia, VI and VII.
The
P.falciparum
protein has unique insertions on either side of region Ia.
Block IV contains the invariant Asp X Asp motif present in motif C of the [alpha]-like DNA polymerase family (
39
,
46
; see Fig. 4). Mutational studies showed these two residues to be essential for
catalytic activity (
48
). Alignment of eight bacterial primase sequences identified the presence of the
Asp X Asp motif in a highly conserved block showing some additional homology to
the putative motif C of eukaryotic primases (Fig.
4
). Putative motifs A and B can also be identified in conserved blocks of
eukaryotic and prokaryotic primases, with again some additional homology
between the two primase classes. For motif B the interaction of four amino
acids with the [beta] and [gamma] phosphates (R754 and K758) and sugar moiety (F762) of the dNTP and
with the template strand (Y766) has been proposed in a model complex of
E.coli
DNA polymerase I (
35
,
36
). These residues are not absolutely conserved in most of the other polymerase families, including the generally accepted
motif B of [alpha]-like DNA polymerases (see Fig.
4
). In some cases an aromatic residue might be replaced by a bulky aliphatic
residue. Figure
4
includes typical examples of the large subunits of RNA polymerase classes I, II
and III and of two viral RNA polymerases. Three highly conserved regions can be aligned with almost perfectly matched invariant residues at the critical positions.
Spacing of the motifs is very similar to other families. We conclude that the
proposed uniform architecture of the catalytic centres of different polymerase
families may also apply to primases and multisubunit RNA polymerases.
Databank searches (EMBL and SwissProt) did not identify any similarity of
eukaryotic primases to known amino acid sequences. Potentially significant
sequence similarity was only found by eye (Fig.
5
), between C-terminal regions of eukaryotic primases (blocks V-VII), herpes virus primases (block V;
49
), [alpha]-type DNA polymerases (boxes CT-1-3;
44
,
50
) and eukaryotic RNA polymerase II. Similarity between eukaryotic primases and
polymerases of different classes included in Figure
5
ranges from 30% (human primase/
P.falciparum
DNA polymerase [alpha]) to 40% (human primase/
P.falciparum
DNA polymerase [delta]). Random similarity trials (arbitrarily shifting alignments a few amino
acids) yielded values of 14-17%. Mutational studies on HSV-1 DNA polymerase indicated the involvement of this domain in DNA
binding and association with accessory factors (
51
,
52
). The first region of variable length in the alignment functions as a hinge
region in HSV-1 polymerase (
52
) and size variations are tolerated. Mutations in the C-terminal region of several polymerases [yeast primase (
15
), mouse primase (
48
), HSV-1 primase (
49
) and mouse DNA polymerase [alpha] (
53
)] alter kinetic parameters, but a common evolutionary origin and/or functional
equivalence of the aligned domains can only be revealed by detailed structural
or functional studies.
Biochemical studies on wild-type and specifically altered primase can relate the structural features described above to specific enzymatic
properties. Detailed kinetic studies on
P.falciparum
primase require amounts of pure enzyme only obtainable by heterologous
expression. We have cloned the 53 kDa primase coding sequence in a baculovirus
expression vector fused to an N-terminal hexahistidine tag, allowing purification by nickel-agarose affinity chromatography. A 57 kDa protein (53 kDa plus 4
kDa hexahistidine tag) is present in extracts of Sf9 cells infected with the recombinant virus and can be purified on Ni-NTA-agarose (Fig.
6
, lane 7). As no specific antibodies are available for definite identification
of the recombinant product, a control recombinant virus expressing a 67 kDa
part of a
P.falciparum
polymerase was included (lane 12) to demonstrate that affinity purification of
proteins of expected size was linked to infection with the corresponding
recombinant virus. The primase 53 kDa subunit is partially dissolved at high
salt concentrations (fraction P, lane 4), but >75% is only dissolved in 6 M
guanidinium-HCl (fraction D, lane 6). Between 10 and 20% is only soluble in SDS
loading buffer (lane 5). Maximal amounts of primase can be isolated 72 h post-infection, although relatively more primase is soluble (50-75%) in 500 mM NaCl 48 h after infection (not shown). The native
protein of fraction P binds poorly to the nickel column (not shown), whereas
denatured recombinant primase (fraction D) efficiently binds and elutes
(fraction NA, lane 7). Most likely the polyhistidine tag is hidden within the
three-dimensional structure of the native protein. A similar observation was
made on
Drosophila
50 kDa primase produced in
E.coli
(
54
).
We like to thank Drs Albert Cornelissen and Peter van der Vliet for helpful
discussions and Dr Cornelissen for critical reading of the manuscript. This
work was supported by the Netherlands Minister of Development Cooperation.
Responsibility for the contents and for the opinions expressed rest solely with
the authors; publication does not constitute and endorsement by the Netherlands
Minister for Development Cooperation.
REFERENCES
Return