Plasmodium falciparum Possesses a Cell Cycle-regulated Short Type Replication Protein A Large Subunit Encoded by an Unusual Transcript*

DNA replication in Plasmodium parasites takes place at multiple distinct points during their complex life cycle in the mosquito and vertebrate hosts. Although several parasite proteins involved in DNA replication have been described, the various mechanisms engaged in DNA metabolism of this major pathogen remain largely unexplored. As a step toward understanding this complex network, we describe the identification of Plasmodium falciparum replication protein A large subunit (pfRPA1) through affinity purification and mass spectral analysis of a purified 55-kDa factor. Gel retardation experiments revealed that pfRPA is the major single-stranded DNA binding activity in parasite protein extracts. The activity was expressed in a cell cycle-depend-ent manner with peak activities in late trophozoites and schizonts, thus correlating with the beginning of chromosomal DNA replication. Accordingly, the pfrpa1 message was detected in parasites 20–24 h post-invasion which is in agreement with the expression of other P. falciparum DNA replication genes. Our results show that pfRPA1 is encoded by an unusual 6.5-kb transcript containing a single open reading frame of which only the C-terminal 42% of the deduced protein sequence shows homologies to other reported RPA1s. Like the orthologues of other protozoan parasites, pfRPA1 lacks the N-terminal protein interaction domain and is thus remarkably smaller than the RPA1s of higher eukaryotes.

Plasmodium falciparum causes one of the most life-threatening parasitic diseases in humans being responsible for up to 2 million deaths per year. Malaria pathogenesis is associated with the intracellular erythrocytic stage of the life cycle of the parasite involving repeated rounds of invasion, growth, and schizogony. Parasites that eventually differentiate into gametocytes are taken up by the female anopheline vector where zygote formation and sporogony take place. Sporozoites injected into a human host by the bite of an infective mosquito invade hepatocytes and, after schizogony, release thousands of merozoites capable of invading red blood cells. During each replication event, the timing, rate, and extent of genome multiplication have to be controlled appropriately and coordinated at each developmental stage. Most of the studies on DNA replication in P. falciparum have been conducted during the erythrocytic stages. Several genes involved in eukaryotic chromosomal DNA replication and their encoded proteins have been identified in this parasite, including DNA polymerases ␣ (DNA pol ␣) 1 (1,2) and ␦ (DNA pol ␦) (3,4), proliferating cell nuclear antigen (5), and topoisomerases I (6) and II (7). It has been shown that expression of these genes follows a stagespecific pattern coinciding with the beginning of chromosomal replication which starts 28 -31 h after merozoite invasion and continues through most of schizogony (8).
Due to their complex life cycle and constant immunological pressure exerted by their hosts, the processes of DNA metabolism in Plasmodium parasites must be both very efficient and flexible. The high degree of genetic variability observed in this parasite (9,10) and the fact that DNA replication occurs at five distinct developmental points, namely intrahepatocytic schizogony, intraerythrocytic schizogony, microgametogenesis, premeiotic DNA synthesis, and sporozoite development (11) indicate the operation of highly regulated mechanisms of DNA replication, recombination, and repair. In fact, dynamic processes of DNA metabolism may be one reason for the outstanding success of this parasite because this supports rapid adaptation to environmental challenges such as immune pressure and action of antimalarial drugs. In addition, the unusually high AT content of ϳ80% in the P. falciparum genome (12,13) may also indicate peculiarities in the replication machinery of the parasite. Hence, investigation of the components involved in DNA metabolism of P. falciparum might reveal features unique to this parasite and may consequently lead to the identification of new potential drug targets for malaria therapy.
The eukaryotic single-stranded (ss) DNA-binding protein replication protein A (RPA) plays essential roles in various aspects of DNA metabolism, including replication, recombination, and repair (for review see Ref. 14). The protein has high affinity for ssDNA (15)(16)(17) and binds with much lower affinity to double-stranded DNA (dsDNA) and RNA (18,19). RPA was originally identified as a factor being absolutely required for SV40 replication in vitro (15)(16)(17). In this system RPA interacts with large T-antigen and DNA pol ␣/primase, and these interactions seem to be important in loading DNA pol ␣/primase onto the RPA-coated unwound origin of replication to allow initiation of DNA synthesis to occur (20 -22). Furthermore, RPA is involved in unwinding of dsDNA (23)(24)(25)(26), stimulation of DNA pol ␣/primase activity, and replication factor C-and proliferating cell nuclear antigen-dependent DNA synthesis by DNA pol ␦ (27)(28)(29)(30)(31)(32). RPA exists as a heterotrimeric complex consisting of subunits of ϳ70, 34, and 14 kDa in all eukaryotic organisms examined (14). Among these, genes coding for RPA subunits have been identified and described in two protozoan parasites. In Crithidia fasciculata the large subunit of RPA (RPA1) is only 51 kDa in size (33), and the predicted size of the large subunit of Cryptosporidium parvum RPA is 54 kDa (34). In both RPA homologues the N-terminal protein-interaction domain is lacking, which has been shown to be required for stimulation of DNA pol ␣/primase and important in DNA recombination and repair (35)(36)(37)(38)(39).
During our studies of protein-DNA interactions in P. falciparum promoters, we observed the major ssDNA binding activity in parasite nuclear extracts with binding properties resembling those described for RPAs of other organisms. Affinity purification and mass spectrometric analysis identified a 55-kDa protein as the P. falciparum RPA large subunit homologue (pfRPA1). Interestingly, as in the apicomplexan parasite C. parvum and the trypanosomatid C. fasciculata, this protein lacks the N-terminal protein-interaction domain. Further sequence analysis revealed that the P. falciparum rpa1 transcript is unusually long (ϳ6.5 kb) and encodes a single exon ORF potentially coding for a predicted protein of 1145 amino acids (aa). However, the region sharing homology with other RPA large subunits consists of only the C-terminal 42%. The biological significance of this unusual organization remains unknown. We have also shown that the presence of pfrpa1 message and pfRPA1 activity correlate with timing of chromosomal DNA replication as it has been described for other P. falciparum replication factors.
The identification of a key molecule involved in DNA metabolism in P. falciparum provides a deeper understanding of such essential processes in this parasite and may lead to new pharmacological intervention strategies.

EXPERIMENTAL PROCEDURES
Parasite Cultures-P. falciparum 3D7/K ϩ parasites were cultured in 150-mm Petri dishes at 5% hematocrit as described previously (40) in RPMI medium supplemented with 0.5% albumax (Invitrogen). Growth synchronization was achieved by sorbitol lysis (41) which eliminates all but ring stage parasites.
Probe Labeling and Competitors-Single-stranded oligonucleotides were end-labeled with [␥-32 P]dATP and T4 polynucleotide kinase (Amersham Biosciences) according to the supplier's instructions. The double-stranded oligonucleotide 5B1c was obtained by incubating equimolar amounts of complementary oligonucleotides 5B1f and 5B1rc in 1ϫ React3 (Invitrogen) at 95°C for 5 min followed by slow cooling to room temperature in a heating block. 5B1c was either labeled with [␥-32 P]dATP and T4 polynucleotide kinase (see above) or with Klenow enzyme in a fill-in reaction by incubating 1 pmol of DNA in 1ϫ React2 buffer (Invitrogen) in the presence of 50 M dATP/dGTP/dTTP and 10 Ci (3000 Ci/mmol) of [␣-32 P]dCTP at 30°C for 20 min. Probes were purified using Sephadex G-25 spin columns (Amersham Biosciences). The sequences of probes and competitors are shown in Table I.
Electromobility Shift Assay (EMSA)-EMSA reactions were carried out by incubating 1-2 g of crude nuclear proteins, 5-10 g of crude cytoplasmic proteins, or ϳ0.1-0.5 ng of purified protein (see below) with 5 fmol of radiolabeled probe in EMSA buffer (20 mM Hepes, pH 7.8, 60 mM KCl, 0.5 mM EDTA, 2 mM DTT, 2 mM MgCl 2 , 25 mM ZnCl 2 , 0.1% Triton X-100, 10% glycerol) containing 100 -500 ng of poly(dI-dC) as nonspecific competitor DNA in a 20-l reaction volume for 20 min at room temperature. Binding reactions were analyzed on a 6% polyacrylamide gel in 0.5% TBE. For competition experiments the labeled probe was added 10 min after incubation of protein and competitor DNA. In EMSAs using purified proteins poly(dI-dC) was omitted, and bovine serum albumin (10 g) was added.
UV Cross-linking-To assess the molecular weight of the ssDNAprotein complex ϳ4 or 20 g of crude nuclear or cytosolic protein extracts, respectively, were incubated with 60 fmol of labeled 5B1f oligo in EMSA buffer as described, followed by exposure to UV light for 10 min in a Stratalinker 1800 (Stratagene). Protein-DNA complexes were incubated at 95°C in 1ϫ SDS sample buffer for 5 min and separated on a 12% SDS-polyacrylamide gel. Gels were dried and analyzed by autoradiography. Prestained molecular weight markers were used to estimate the molecular mass of the ssDNA-protein complexes.
Limited Trypsin Digestion-ϳ6 -10 or 20 -40 g of crude nuclear or cytosolic protein extracts, respectively, were incubated with 20 fmol of [␥-32 P]dATP-labeled 5B1f oligonucleotide (in the presence of 1.25 mM PMSF) and UV cross-linked as described above. DNA-protein complexes were incubated with 2 g of porcine pancreas trypsin at room temperature in the same buffer. Aliquots were removed at time points indicated in Fig. 4. Reactions were stopped by the addition of SDS-PAGE sample buffer followed by incubation at 95°C for 5 min. Samples were analyzed SDS-PAGE. Gels were dried and analyzed by autoradiography.
Affinity Purification of the Major ssDNA Binding Activity from Parasite Protein Extracts-As matrix for affinity purification, the biotinylated 90-base oligo 5B1Af (a trimer of 5B1f) was tethered to magnetic streptavidin-coated Dynabeads (Dynal) according to the supplier's instructions. Cytoplasmic extracts from a total of 7.5 liters of asynchronous parasite culture (5-8% parasitaemia) were spun at 3000 ϫ g for 10 min to pellet cellular debris. The supernatant was incubated with 2 mg of Dynabeads (80 pmol oligo/mg beads) in binding buffer BB (20 mM Hepes, pH 7.8, 120 mM KCl, 1 mM EDTA, 1 mM DTT, 1 mM PMSF, 0.1% Triton X-100, 0.65% Nonidet P-40) on a rotating wheel for 1 h at room temperature. Dynabeads were collected by use of a magnetic stand and washed 2 times with 1 ml of wash buffer W1 (20 mM Hepes, pH 7.8, 150 mM KCl, 1 mM EDTA, 1 mM DTT, 1 mM PMSF, 0.01% Triton X-100, 5% glycerol) and 2 times with buffer W2 containing 200 mM KCl. Bound proteins were eluted stepwise twice in 200 l of elution buffer EB1 and twice in EB2 (20 mM Hepes, pH 7.8, 2.5 mM EDTA, 1 mM PMSF, 10% glycerol) containing 0.5 and 1 M KCl, respectively. Purification from nuclear extracts was performed by incubation of 4.5 ml of crude nuclear extract (1-2 g/l) from a total of 2.8 liters of asynchronous parasite culture (5-8% parasitaemia) with 1 mg of Dynabeads (10 pmol of oligo/mg of beads) in binding buffer BB (20 ng poly(dI-dC)/l) as described above. Dynabeads were washed 2 times in 300 l of buffer W1 and once in buffer W3 (250 mM KCl). Bound proteins were eluted stepwise twice in 50 l of EB1 followed by elution in 50 l of EB2. Wash fractions and eluates were analyzed by EMSA and SDS-PAGE and silver staining. Samples were stored at Ϫ80°C.
Mass Spectral Analysis of the ssDNA Binding Activity-A gel piece containing the Coomassie Blue-stained purified cytoplasmic ssDNAbinding factor was washed five times for 1 min each in 30 l of 40% n-propyl alcohol followed by five 1-min washes each in 30 l of 0.2 M NH 4 HCO 3 (50% acetonitrile). The gel piece was dried in a SpeedVac concentrator and then digested with 0.5 g of sequencing grade modified trypsin (Promega) in 10 l of 0.1 M NH 4 HCO 3 for 2 h at 37°C. The gel piece was extracted with 15 l of 0.1% trifluoroacetic acid for 5 min followed by 15 l of acetonitrile for 1 min. Extraction was repeated twice, and the pooled supernatants were dried in a SpeedVac concentrator. Peptides were redissolved in 10 l of 0.1% trifluoroacetic acid, and 5 l were used for mass spectral analysis. Separation of peptides was done on 100-m inner diameter capillary columns packed with POROS R2 material. Mass spectral data were acquired on a TSQ7000 triple quadrupole instrument (Finnigan) with data-controlled switching between precursor ions and daughter ions (43). For precursor ion scan-ning resolution of the first quadrupole was set to 1 Da. For operation in the MS/MS mode, the resolution of the first quadrupole was decreased to transmit a window of 4 Da, and the resolution of Q3 was adjusted to 1.5 Da. The daughter ion spectra were used to identify the protein with SEQUEST program (44).
Isolation of Parasite Total RNA and Northern Analysis-Parasite total RNA was isolated using Trizol (Invitrogen) as described (45), and RNA was stored in formamide at Ϫ80°C. For Northern blot analysis equal amounts of RNA extracted from synchronized parasite cultures was electrophoresed on 1.2% agarose gels (5 mM guanidine isothiocyanate) (45) and vacuum-transferred to a Hybond-XL nylon membrane (Amersham Biosciences). Probes for Northern analysis of pfrpa1 were gel-purified PCR products (see Fig. 6) radiolabeled with [␣-32 P]dCTP using random primers and Klenow polymerase. Hybridization was performed at 42°C in UltraHyb (Ambion).
Nucleotide Sequence Data-Sequence data for pfrpa1 (GenBank TM accession number AL035475) was obtained from the Sanger Center website at www.sanger.ac.uk/Projects/P_falciparum/. Sequencing of P. falciparum chromosome 4 was accomplished as part of the Malaria Genome Project. Preliminary sequence data from the Plasmodium yoellii genome were obtained from the Institute for Genomic Research website (www.tigr.org). This sequencing program is carried on in collaboration with the Naval Medical Research Center.

RESULTS
Major DNA Binding Activity in P. falciparum Nuclear Extracts-In the course of investigations of P. falciparum promoters by gel retardation assays using dsDNA probes, we detected a dominant nonspecific DNA binding activity in parasite nuclear extracts derived from asynchronously growing cultures. This activity was only observed, however, when probes were end-labeled with T4 polynucleotide kinase which also labels ssDNA molecules. In contrast, when we used Klenow enzyme to fill in 4-base 5Ј protrusions at the ends of double-stranded complementary oligonucleotides (ensuring that only doublestranded molecules are labeled), the DNA-protein complex was hardly detected (data not shown). EMSAs using end-labeled single-stranded oligonucleotides showed that this activity was due to the interaction of a nuclear factor with ssDNA. Fig. 1A shows the interaction between this factor and the radiolabeled 30-base single-stranded oligonucleotide probe 5B1f (the sequence of 5B1f corresponds to a conserved motif found in var gene promoters (46)). As shown in Fig. 1B the single-stranded oligonucleotides 5B1f and 5B1rc added at a 20-fold molar excess competed with binding to the labeled probe. However, a double-stranded 155-bp competitor restriction fragment containing the 5B1f sequence (5B1sub5) did not compete for binding. EMSA competition experiments further revealed that the affinity of the nuclear factor was higher for single-stranded polypyrimidine than for polypurine oligomers (Fig. 1B). In contrast, heterogeneous dsDNA was a much weaker competitor, and yeast tRNA did not compete at all even if added at 2000fold weight excess. To further investigate for sequence preferences, we used mutated forms of oligonucleotide 5B1f (5B1fmut1-5, see Table I) in gel retardation competition studies. In these oligos consecutive stretches of six nucleotides each were mutated, where A was replaced by G, G replaced by T, and T replaced by C. Whereas oligos 5B1fmut1, 5B1mut4, and 5B1fmut5 competed equally well or even better compared with the wild type sequence 5B1f, the ssDNA-binding factor had a clearly reduced affinity for oligonucleotides 5B1fmut2 and 5B1fmut3 (Fig. 1C), indicating a certain degree of sequence preference independent of the pyrimidine/purine content.
To investigate whether the ssDNA binding activity was present throughout the intra-erythrocytic life cycle, we performed gel retardation experiments using nuclear extracts prepared from synchronously growing cultures. The ssDNA binding activity was absent in mid-ring stage parasites (8 -16 h postinvasion (hpi)) (Fig. 2). The activity faintly appeared in young trophozoites (16 -24 hpi) and increased to maximal levels in parasites older than 34 hpi. Parasite nuclear extracts derived from the very early ring stage (0 -8 hpi) also contained the ssDNA binding activity.
Comparison of Nuclear and Cytoplasmic ssDNA Binding Activities-A major ssDNA binding activity was also observed in cytoplasmic parasite extracts, but the complexed probe migrated at a slightly different position than the complex formed with the nuclear factor (Fig. 3A). However, when various protease inhibitors were used during protein isolation and EMSAs, and when using fresh cytoplasmic extracts in gel retardation  (70), and sheared salmon sperm DNA at 100-fold weight excess. C, ssDNA-binding affinity analysis using a 20-fold molar excess of wild type oligo 5B1f and mutated oligos 5B1fmut1-5 (see Table I) as competitors. n.e., nuclear extract; pur, 30-base oligonucleotide of random purine sequence; pyr, 30-base oligonucleotide of random pyrimidine sequence. assays, an additional signal migrating at the same position as the nuclear complex was observed (Fig. 3B). This suggested identical activities in both subcellular compartments with proteolytic activities in cytosolic extracts acting on the ssDNAbinding factor during protein isolation and gel retardation experiments. In EMSA affinity assays using a variety of different competitor DNAs, the cytoplasmic and nuclear activities behaved identically (Fig. 3C). These observations supported the assumption that both activities were exerted by the same protein. To investigate this possibility in more detail, we compared limited tryptic digests of UV-cross-linked EMSA reactions by SDS-PAGE (Fig. 4). Without trypsin digestion the major crosslinked complexes in nuclear and cytoplasmic extracts migrated at an equal position (ϳ65 kDa). The additional larger signal observed in nuclear extracts was probably due to another abundant ssDNA binding activity that was unstable under electrophoresis conditions without UV cross-linking. Furthermore, in both experiments an identical pattern of labeled tryptic fragments was observed indicating that the ssDNA binding activity was retained in tryptic fragments of equal size. Taken together, these findings clearly suggested that the activities present in nuclear and cytoplasmic extracts were identical.
Affinity Purification of the ssDNA Binding Activity from Parasite Protein Extracts-Both major ssDNA binding activities from nuclear and cytosolic extracts were purified by affinity purification. Straptavidin-coated magnetic beads with tethered biotinylated 90-base oligonucleotide 5B1Af (a 3-mer of 5B1f) were incubated with crude protein extracts in binding buffer.
After washing, bound proteins were eluted with 0.5 and 1 M KCl, and all fractions were analyzed by SDS-PAGE followed by silver staining (data not shown). Testing for ssDNA binding activity using gel retardation revealed that most of the ssDNA binding activity was eluted at 0.5 M salt (data not shown). The electrophoretic mobilities of ssDNA-protein complexes formed with crude extracts and purified fractions were identical and are presented in Fig. 5A. Fig. 5B shows SDS-PAGE analysis of the 0.5 M cytosolic eluate revealing a dominant band at ϳ55 kDa and two additional enriched proteins at 30 and 25 kDa. Similar results were obtained for the nuclear 0.5 M eluate. However, the purified cytosolic 55-kDa protein had a slightly smaller size than the nuclear factor (data not shown) which is in agreement with the observed difference in mobility of the corresponding ssDNA-protein complexes in EMSA experiments. Furthermore, SDS-PAGE analysis of UV cross-linked ssDNA-protein complexes obtained with crude extracts and the purified factors revealed a size of ϳ65 kDa for each complex (Fig. 5C). These results strongly suggested that the 55-kDa factor was responsible for ssDNA binding because the size difference of 10 kDa observed between the purified protein alone and the UV cross-linked ssDNA-protein complexes was accounted for by the covalent attachment of the 30-base oligonucleotide 5B1f to the binding factor in the UV cross-linked samples. Gel retardation competition studies using the purified proteins from both the nuclear and cytosolic extracts again revealed an identical affinity pattern and the expected difference in complex mobility (Fig. 5D). However, the overall affinity pattern slightly diverged from the results obtained with the crude protein extracts, which might be due to the different binding conditions used in EMSAs (total amount of ssDNAbinding protein, presence of bovine serum albumin, and absence of poly(dI-dC) in EMSAs using the purified proteins). The characteristics of the purified proteins again indicated that the major ssDNA binding activities detected in nuclear and cytosolic extracts were exerted by the same protein.
Mass Spectral Analysis of the Purified ssDNA Binding Factor-To obtain enough protein for mass spectral analysis, the ssDNA binding activity was affinity-purified stepwise from crude cytosolic extracts obtained from a total of 7.5 liters of culture (5-8% parasitaemia) of asynchronously growing parasites yielding a total of ϳ1-2 g of purified protein. Eluates were pooled and separated on a 10% SDS-polyacrylamide gel. The ssDNA-binding factor (55 kDa) was excised and digested with trypsin, and the fragments were subjected to mass spectral analysis. Data base searches with the fragmentation spectra obtained by MS/MS analysis from four tryptic peptides (NVNLVNEEALSGK, ITDSTDSIR, LNEFFFR, and YNYN-FISIDNIK) unambiguously identified this protein as putative P. falciparum replication protein A large subunit (pfRPA1) located on chromosome 4 (GenBank™ accession number AL035475). Furthermore, peptide masses of 7 additional tryptic fragments perfectly matched the peptide mass fingerprint of pfRPA1. The nucleotide and deduced aa sequences of the putative pfrpa1 ORF are shown in Fig. 6. The predicted 1145-aa sequence (134 kDa) shares identities between 34.4 and 39.2% only in its C-terminal 466 aa to the RPA large subunits of Homo sapiens, Saccharomyces cerevisiae, Schizosaccharomyces pombe, C. fasciculata, C. parvum, Xenopus laevis, Caenorhabditis elegans, Drosophila melanogaster, Oryza sativa, and Arabidopsis thaliana (accession numbers P27694, P22336, Q92372, Q23696, AF132307, Q01588, Q19537, Q24492, AF009179, AL021687, respectively). This region includes a conserved zinc finger motif (CX 2-8 CX 10 -15 CX 2 C) (34). However, the consensus spacing of 10 -15 aa between cysteines 2 and 3 is extended to 26 residues in pfRPA1 (Fig. 6). All tryptic peptides  TTC TCT TTC TAT CTA TAT TAT CTA  identified by mass spectral analysis exclusively map to the C-terminal region of the predicted pfRPA1. In addition, the size of 55 kDa estimated for the purified ssDNA-binding factor by SDS-PAGE analysis corresponds to the predicted mass of the C-terminal 466 aa of pfRPA1 (54.6 kDa). It is important to note that the large subunits of RPA of the apicomplexan C. parvum (34) and trypanosomatid C. fasciculata (33) parasites have been reported to be 54 and 51 kDa in size, respectively. These proteins, together with pfRPA1, lack the N-terminal proteininteraction domain found in all other RPA large subunits that have molecular masses between 67 and 73 kDa (14). Moreover, in the N-terminal part of the deduced 1145-aa sequence of pfRPA1, no significant homologies to any known proteins were found. Instead, several interspersed Asn-rich tracts are present within this N-terminal region, a feature that has also been observed in other P. falciparum proteins such as DNA pol ␣ (2).
pfrpa1 Transcript Analysis-To investigate the pfrpa1 transcription pattern in P. falciparum across the intraerythrocytic cell cycle and to test whether the whole 3435-bp ORF was transcribed into a mRNA message, we performed Northern analysis using total parasite RNA isolated from different intraerythrocytic life stages. After hybridization with the radiolabeled probes rpaC and rpaE (Fig. 6), identical signals of a single pfrpa1 message were detected. The transcript appeared in parasites older than 22 hpi and was present until the end of the cycle, whereas it was absent in ring stage parasites and young trophozoites (Fig. 7). Identical results were obtained using hybridization probes rpaA, -B, and -D (data not shown) proving that the entire predicted ORF (3435 bp) was transcribed. The size of the pfrpa1 transcript was ϳ6.5 kb (by comparison to an RNA size standard) indicating the presence of extensive 5Ј-and 3Ј-untranslated regions. DISCUSSION This paper describes for the first time the purification and identification of a DNA-binding protein from the malaria parasite P. falciparum. pfRPA binds ssDNA with high affinity, has a preference for single-stranded polypyrimidines over polypurines, a much lower affinity for dsDNA, and does not bind yeast tRNA. This affinity pattern is consistent for all eukaryotic RPAs investigated to date (14). In addition, we also detected preferences of pfRPA for oligonucleotides of mixed sequence that cannot solely be attributed to the ratio of pyrimidines and purines. This may reflect preferential binding of pfRPA to certain genomic regions like origins of replication or "hot spots" of recombination, and may therefore hold significance for pfRPA function. In fact, partial sequence dependence of RPA binding has also been reported in yeast and humans (18).
We would like to stress here that this ssDNA binding activity heavily interfered in gel retardation analysis of parasite promoters using dsDNA probes. 2 We conclude that the presence of any labeled ssDNA molecules in binding reactions results in the formation of ssDNA-pfRPA complexes that may easily be FIG. 3. Comparison of the major nuclear and cytoplasmic ssDNA binding activities. A, gel retardation analysis of the major nuclear and cytosolic ssDNA binding activities interacting with radiolabeled 5B1f. B, an additional ssDNA-protein complex, having the same electrophoretic mobility as the nuclear complex, was formed with fresh cytosolic extracts in the presence of various protease inhibitors (PI). C, comparison of ssDNA binding affinities exerted by the nuclear and cytosolic activities. 5 fmol of radiolabeled 5B1f (0.05 ng) were incubated with 2 g of nuclear or 8 g of cytosolic extracts. Sheared salmon sperm DNA and pCAM5/3 plasmid DNA were added at 100-and 200-fold weight excess, respectively. Oligonucleotides 5B1fmut1, 5B1mut2, and 5B1mut5 were added at 20-fold molar excess. n.e., nuclear extract; c.e., cytosolic extract.
FIG. 4. Limited trypsin digestion. Nuclear and cytosolic ssDNA binding activities were UV cross-linked to radiolabeled 5B1f. Fragments of partially digested cross-linked ssDNA-protein complexes taken at the time points indicated were separated by SDS-PAGE, and the dried gel was analyzed by autoradiography. The undigested ssDNAprotein complex migrated at an apparent mass of 65 kDa (arrow). Tryptic fragments of the nuclear and cytoplasmic factors that retained binding activity and were of similar size are marked by asterisks. Trypsin incubation times are indicated above each lane. Free probe is apparent as an intense signal at the bottom of the autoradiograph. mistaken as dsDNA-protein interactions. These labeled ssDNA species might originate through labeling with T4 polynucleotide kinase or dissociation of double-stranded probes during preparation and purification. In addition, RPA preferentially binds to AT-rich regions of dsDNA under low salt conditions (26), and this finding merits special attention in light of the AT richness of P. falciparum DNA. We therefore propose to include a 50-fold molar excess of unlabeled oligonucleotides in EMSA reactions using P. falciparum extracts.
pfRPA activity was present in both nuclear and cytosolic extracts. We exclude the possibility of cross-contamination of the cytosolic fraction with nuclear pfRPA during cell fractionation. By using the same protein preparations described in the present study, we recently identified sequence motifs in P. falciparum var gene promoters specifically interacting with activities present in nuclear fractions but completely absent in cytosolic fractions. 2 Furthermore, in human cells the majority of RPA is present in the cytosolic fraction after gentle lysis of cells (15)(16)(17)47), and purification of RPA from cytosolic fractions has been reported (18,19). Compared with the nuclear activity, cytosolic pfRPA1 had a slightly decreased size, and the ssDNA-pfRPA complex showed a corresponding increase in electrophoretic mobility. This was probably due to proteolysis because the use of various protease inhibitors in fresh cytosolic preparations revealed a second complex similar to the one formed with crude nuclear extracts. Interestingly, Seroussi and Lavi (48) reported two closely comigrating complexes using mammalian whole cell extracts. By performing supershift EM-SAs using antibodies against RPA subunits, they provided convincing evidence that the faster migrating complex was due to proteolysis of hsRPA1.
In other eukaryotes such as Crithidia, yeast, tobacco, rodents, and humans, purification schemes for RPA use prefractionation of cell lysates, ssDNA-cellulose columns for affinity FIG. 5. Affinity purification of the major ssDNA binding activity from nuclear and cytoplasmic extracts. A, gel retardation analysis comparing the ssDNA binding activities present in crude nuclear or cytosolic extracts and in the 0.5 M salt elutions obtained from affinity purification. B, SDS-PAGE separation of the cytosolic 0.5 M eluate with enriched bands at 55, 30, and 25 kDa. Crude nuclear and cytoplasmic extracts and molecular weight marker have been used as references. C, UV cross-linked ssDNA-binding reactions were analyzed by SDS-PAGE and autoradiography. The major ssDNA binding activities in crude nuclear extracts and in the 0.5 M eluate from nuclear affinity purification were of equal size. As size references molecular mass marker and an aliquot of the 0.5 M eluate were separated on the same gel and visualized by silver staining. D, competition EMSA comparing the ssDNA binding affinities of the purified nuclear and cytosolic activities. 5 fmol of radiolabeled 5B1f were incubated with ϳ0.1-0.5 ng of purified nuclear and cytosolic factors in the presence of various competitor DNAs. pur, 30-base oligonucleotide of random purine sequence; pyr, 30-base oligonucleotide of random pyrimidine sequence. Yeast tRNA was added at 2000-fold weight excess; the other competitors were added at 20-fold molar excess. n.e., nuclear extract; c.e., cytosolic extract. chromatography, followed by liquid chromatography to obtain highly pure RPA (20, 23, 48 -51). SDS-PAGE analysis of these preparations revealed three distinct protein bands corresponding to the dissociated subunits of RPA. Due to the P. falciparum in vitro culture conditions, it is almost inconceivable to obtain enough protein material for such purification procedures. We therefore purified pfRPA directly from crude nuclear and cytosolic extracts. This probably led to copurification of other proteins with affinity for ssDNA preventing a clear identification of the middle and small subunits (Fig. 5). Although we failed to identify malarial orthologues to these subunits using BLAST (52) at PlasmoDB (www.plasmodb.org) and NCBI Malaria Genetics and Genomics section (www.ncbi.nlm.nih.gov/Malaria/), we clearly expect that P. falciparum is equipped with the middle and small subunits because the heterotrimeric structure of RPA is conserved from kinetoplastids to humans (14). pfRPA1, with the apparent mass of 55 kDa, is remarkably smaller than its orthologues in most other eukaryotes. Sequence analysis revealed that this size difference is accounted for by the absence of the N-terminal protein-interaction domain. pfRPA1 is encoded by a continuous ORF on chromosome 4 (GenBank™ accession number AL035475) encoding a putative 1145-aa protein (134 kDa). Northern analysis using various pfrpa1 probes detected a single large transcript of ϳ6.5 kb encoding the entire ORF. Furthermore, PCR analysis using genomic DNA from six different P. falciparum strains indicated that this locus is conserved (data not shown). Only the Cterminal 466 aa, however, share homology to other RPA large subunits, including a conserved zinc finger motif close to the C terminus. The predicted molecular mass of this 466-aa region (54.6 kDa) is in perfect accordance to the size predicted from SDS-PAGE analysis. The N-terminal part of the predicted protein shows no homologies to any other proteins. Several Asnrich tracts are interspersed in this region, a feature that has also been reported for other P. falciparum proteins (2,7,(53)(54)(55)(56), and more examples emerge from the P. falciparum sequencing project (57,58). These findings, together with the fact that all tryptic fragments from mass spectral analysis exclusively mapped to the C-terminal region, suggest post-translational proteolytic cleavage of a 1145-aa precursor protein generating the 55-kDa pfRPA1. In fact, such a process has been reported for the 51-kDa pfCDP-diacylglycerol synthase where the N-terminal third of a 78-kDa precursor, containing several Asn-rich stretches, appeared as an independent 28-kDa protein in Western analysis (56). Another explanation would be translation of pfRPA1 from an ATG internal to the predicted 3435-bp ORF. However, the absence of frameshift mutations in the whole ORF argues against the latter possibility. Custom   FIG. 6. Sequence of pfRPA1. The nucleotide sequence (GenBank™ accession number AL035475) and deduced aa sequence of putative pfRPA1 are shown. Numbers indicate aa residues in the predicted protein sequence. Tryptic peptides identified by mass spectrometric analysis of the purified protein are boxed. The Cterminal part of the protein sharing sequence homology to other eukaryotic RPA large subunits is typed in boldface letters, and putative internal translation initiation methionines are indicated by black boxes. The conserved zinc finger motif found in all RPA large subunits is shaded in gray, and conserved residues are marked by asterisks. Asn-rich tracts found in the N-terminal half of the predicted protein sequences are indicated in boldface italics. Oligonucleotides used to generate PCR fragments rpaA-rpaE for Northern analysis are shown by arrows.
BLAST searches (www.ncbi.nlm.nih.gov/Malaria/plasmodiumblcus.html) identified the P. yoelli homologue of pfrpa1 coding for the putative pyRPA1 (1168 aa) (TIGR accession c5 m1231). A pairwise sequence alignment using SIM program (searchlauncher.bcm.tmc.edu) revealed an overall sequence identity of 49% to pfRPA1. However, the C-terminal domains (471-aa overlap) share 71% identity, and 80% identity is observed in their N-terminal parts (145-aa overlap), whereas the regions in between lack sequence homology. We also found a Plasmodium vivax ORF (GenBank™ accession number AZ575343) (59) exhibiting 73% sequence identity to the Nterminal 85 aa of pfRPA1. This indicates a conserved organization of rpa1 genes in closely related Plasmodium species.
Replication in intraerythrocytic parasites begins between 28 and 31 hpi and continues until the end of schizogony (8). Accordingly, P. falciparum replication factors like DNA pol ␣ (60), DNA pol ␦ (61), topoisomerases I (62) and II (63), and proliferating cell nuclear antigen (61) are expressed in trophozoites and schizonts. The pfrpa1 message was only transcribed in parasites older than 20 -24 hpi. pfRPA activity drastically increased in parasites older than 24 hpi and lasted until the end of the intraerythrocytic cycle, whereas it was hardly detectable 8 -24 hpi. pfRPA was also present in the nuclear fraction, but not in the cytosol, of very young parasites (0 -8 hpi) (data not shown). Carry-over of nuclear pfRPA from schizonts to young ring stage parasites would explain this observation.
Hence, it appears that pfRPA1, like other P. falciparum replication proteins, is expressed in a cell cycle-dependent manner shortly before chromosomal DNA replication occurs.
Eukaryotic DNA metabolism requires the interaction and cooperation of multiple factors conserved from protists to humans. Among these, RPA takes a central position physically interacting with many proteins involved in DNA replication, recombination, nucleotide excision repair, and transcription (64 -69). Most of these protein interactions involve the N-terminal protein-interaction domain of RPA1. Hence, the absence of this structure in Plasmodium spp., C. parvum, and C. fasciculata implicates that alternative mechanisms employed in DNA metabolism evolved in protozoan parasites. Yeast twohybrid analysis would be a promising way to identify the factors interacting with pfRPA1 in DNA replication, recombina-tion and repair, as well as the pfRPA1 domains involved. This would greatly enhance our understanding of the DNA metabolism network of this major human pathogen.
In summary, we identified the P. falciparum homologue of eukaryotic RPA1 encoded by an unusually long transcript which, like RPA1 of other protozoans, lacks the N-terminal protein-interaction domain. Expression of pfRPA is stage-specifically regulated, and the protein is present at the onset of S-phase in the erythrocytic parasite stages. Our findings support the assumption that major differences in DNA metabolism between P. falciparum and its host exist and may be an indication for peculiarities in the replication machinery of the parasite that could be exploited in novel antiparasite strategies. FIG. 7. Stage-specific Northern analysis of pfrpa1. pfrpa1 PCR fragments were radiolabeled and hybridized to Northern blots of parasite total RNA isolated from synchronized cultures. Equal RNA loading was confirmed by ethidium bromide staining (data not shown). rpaE was generated with primers rpaEf and rpaEr, and rpaC was obtained with primers rpaCf and rpaCr (see Fig. 6).