aldolase gene of Plasmodium falciparum - NCBI

The EMBO Journal vol.9 no.5 pp.1645-1649, 1990

Initiation of translation at a UAG stop codon in the aldolase gene of Plasmodium falciparum

Paola Ghersa, Indresh K.Shrivastava, Luc H.Perrin, Heinz Dobeli1, David J.Becherer2, Hugues Matile1, Bea Meier1 and Ulrich Certa Geneva Blood Centre and Department of Medicine, University Cantonal Hospital, 1211 Geneva 4, 'Central Research Units, F.Hoffmann-La Roche Ltd, 4002 Basel and 2Basel Institute for Immunology, 4002 Basel, Switzerland Communicated by J.Scaife

The gene coding for the key glycolytic enzyme fructose-1,6-diphosphate aldolase of the human malaria parasite Plasmodium falciparum lacks a functional AUG initiation codon for translation. Protein sequences of natural or in vitro translated aldolase include the candidate start methionine residue at internal positions. No additional AUG start codon is found in genomic DNA, cDNA or mRNA sequences. Instead, a UAG chain termination codon is recognized as the start signal of protein synthesis in vivo and in vitro. Key words: chain termination/initiation of translation/start codons

Introduction Ribosome-mediated synthesis of proteins with mRNA as template is a fundamental process in nature and it is therefore not surprising that the molecular mechanisms are highly conserved in all organisms studied (Kazak, 1983). Initiator amino-acyl-tRNA binding, peptide bond formation and ribosome translation are virtually identical in all cell types. An AUG initiation codon on the mature transcript is the most common start signal for protein synthesis although GUG, UUG or ACG start signals are recognized with decreasing efficiency by Escherichia coli (Kozak, 1983; Peabody, 1989). Initiation in yeast is restricted to AUG and a point mutation converting AUG to GUG abolished expression of the iso-I cytochrome c gene (Stewart et al., 1971). An ACG codon of human adeno-associated virus RNA embedded in favourable initiation sequence context is recognized with low efficiency as start signal by a rabbit reticulocyte in vitro translation extract (Beccera et al., 1985). Chain termination at UAA, UGA or UAG codons is due to lack of complementary charged tRNA molecules. In several organisms amino-acyl suppressor tRNAs decoding the stop signals have been detected. They have evolved by point mutations in the anticodon of charged tRNA molecules and they have important functions such as supression of in frame termination codons that appear frequently in the coding regions of ciliate genes (Caron and Meyer, 1985). A 50 nt untranslated region is present within the coding sequence of Escherichia coli phage T4 gene 60, which encodes one subunit of DNA topoisomerase type II (Huang et al., 1988). This insert contains chain termination codons in all reading frames. A mechanism is proposed in which folding of the Oxford University Press

untranslated region brings together codons separated by the interruption so that the elongating ribosome does not terminate translation at one of the stop codons. A truncated organ specific variant of the human apolipoprotein B (apo B48) is generated by the tissue specific introduction of a stop codon in intestinal mRNA. The mRNA in the liver lacks this in-frame chain termination codon and as a result a larger protein (apo B100) is translated (Chen et al., 1987). These mutations with physiological significance could be introduced by post-transcriptional mRNA editing in the absence of template DNA which was first discovered in kinetoplast mRNA of trypanosomes (Benne et al., 1986). These recent findings suggest that the function of stop codons may not be restricted to chain termination. Here we describe a novel function of a UAG chain termination codon. We have recently cloned and analysed the aldolase gene of the human malaria parasite Plasmodium falciparum (Certa et al., 1988). A comparison of the amino acid sequence deduced from cloned DNA with peptide sequences of a number of protease V8 peptides from purified parasite aldolase drew our attention to a peptide that contained the putative methionine start codon internal at position three. This eliminated its codon as the translation initiation codon. No additional in-frame AUG initiation codon is present upstream and instead a UAG chain termination codon is recognized as the start signal for ribosomal protein synthesis in vitro and in vivo.

Results Genomic aldolase DNA, cDNA and mRNA lack a functional AUG initiation codon Figure 1 shows the DNA and deduced protein sequence at the 5' end of the P.falciparum aldolase gene. The first methionine residue in frame is contained at position two (Figure 1, codon 8) of a protease V8 peptide sequence (Figure 1, residues 6-20). The lack of an additional in frame ATG upstream in the genomic aldolase DNA sequence of

1

2

3

4

5

6

7

5'--TTT TTA TAT TTT TTT TAG GCT CAT TGC ACT GAA TAT *

ALA

HIS

CYS THR GLU TYR

9 10 11 12 13 14 15 16 17 18 19 20 ATG AAT GCC CCA AAA AAA TTA CCA GCA GAT GTT GCC GAA 8

MIET ASN ALA PRO LYS LY'S

LEU PRO

ALA

ASP VAL

ALA GLU

GAA TTA GCA ACCACCGCCCAA AAGCTT--3' GLU

LEU

ALA

THR

THR

ALA

GLN

LYS

LEU

Fig. 1. DNA and deduced amino acid sequence at the 5' end of the P.falciparum aldolase gene. The amino acid sequence of a protease V8 peptide derived from parasite aldolase is underlined and was determined as described (Certa et al., 1988). The first methionine residue in-frame is marked with an arrow and codons relevant for this study are numbered on top starting with an in-frame TAG stop codon labelled with an asterisk. The DNA sequence of the KI isolate of P.falciparnm was published earlier (Certa et al., 1988) and the 5' sequence is identical in the M-25 isolate (Figure 2).

1645

Fig. 3. In vitro translation of aldolase mRNA in rabbit reticulocytes. Autoradiography of an 8% SDS-polyacrylamide gel. Lane 1: Total P.falciparum labelled with [35S]methionine. Lane 2: P.falciparum aldolase immunoprecipitated with rabbit serum monospecific for parasite aldolase. Lane 3: Immunoprecipitation with pre-immune serum. Lane 4: Immunoprecipitation of P.falciparum aldolase synthesized in vitro with total cellular RNA as template. Lane 5: mRNA as template. Lane 6: same as lane 2. The in vitro synthesis of a second product with a molecular mass of -31 kd is consistent with initiation of translation at an internal ATG codon downstream in the aldolase gene sequence. Molecular markers are indicated in kilodaltons. Fig. 2. Detection of the TAG codon in two genomic DNA sequences of the aldolase gene of the M25 isolate of P.falciparum (panel A; first two sequence ladders). The beginning of the gene sequence was confirmed by chemical sequencing of clone M25-A (panel A, third ladder). Detection of the TAG condon in sequences of aldolase cDNA and mRNA (panel B). The left side shows sequences determined for four independent cDNA clones of the M25 isolate of P.falciparum. The right part shows the mRNA sequence determined by primer extension sequencing of total parasite mRNA. Note that the sequences upstream of the stop codon are identical to genomic DNA. The codes of the clones sequenced are given below the sequencing tracks. Letters G, A, T and C on top of the lanes indicate base-specific sequencing reactions. The in-frame TAG (UAG) stop codon is marked by an arrow and the sequences read 5' to 3' starting from top.

the Kl (Certa et al., 1988) or M25 isolate of P.falciparum suggested an alternative start site for ribosomal protein synthesis. The presence of a TAG splice acceptor site (Figure 1, codon 1; Simmons et al., 1987) seven codons upstream of the ATG in the genomic sequence suggested the possibility that the AUG start signal could be localized on a second exon separated by an intron (Figure 2, panel A). However, the genomic sequence 500 bp upsteam of the initiation site had no AUG codon in-frame with the aldolase coding sequence downstream which had a splice donor consensus sequence (data not shown; Simmons et al., 1987). The lack of a possible intron was confirmed by the sequence of four independent cDNA clones which was identical to genomic DNA (Figure 2, panel B). Direct primer extension sequencing of aldolase mRNA extracted from the parasite confirmed the cDNA sequences and eliminated the introduction of an AUG start site by post-transcriptional editing of mRNA (Figure 2, panel B). The possibility of dideoxy sequencing artifacts was excluded by re-sequencing the relevant part of cloned genomic DNA by chemical sequence determination (Figure 2, panel A; Maxam and Gilbert, 1980). 1646

UAG is recognized as a start codon in vitro by a rabbit reticulocyte lysate The sequencing studies thus eliminated the first in frame ATG (Figure 1, codon 8) as the start codon. The lack of an additional upstream ATG in-frame of mature P.falciparum aldolase mRNA suggested an anomalous start site for initiation of ribosomal protein synthesis. A subset

of parasite-specific, suppressor-like initiation tRNAs carrying mutations in the anticodon sequence could recognize such a start codon. Stage-specific expression of a single tRNA gene could thus turn on the translation of a set of related genes with a unique start codon. These tRNA species are not expected to be present in a conventional rabbit reticulocyte in vitro translation extract. We therefore compared the in vitro translation activity of total cellular parasite RNA with oligo-dT affinity purified mRNA which is presumably free of parasite derived tRNA. Surprisingly, both templates initiated aldolase synthesis and the products had the same molecular mass as aldolase immunoprecipitated from a 35S-labelled culture of P.falciparum (Figure 3). SDS -PAGE purified in vitro synthesized aldolase was then subjected to protein sequence analysis in order to determine the position of the cryptic start codon relative to the first AUG in frame (Figure 1, codon 8). 35S-labelled methionine was recovered in cycles 1 and 8 after Edman degradation. This assigned the in-frame UAG stop codon (Figure 1, codon 1) to the binding site of charged initiation tRNA in vitro. This result strongly suggests that regular rabbit initiation tRNA can recognize the UAG codon and makes it unlikely that P.falciparum derived tRNAs or other molecules are required for initiation at UAG in vitro. This plausible conclusion was directly proven by translating in vitro synthesized aldolase mRNA in a rabbit

Initiation at

Ml 2

3

I

I

I.

4

a

UAG stop codon

!) z C.)

200 - -ae c

.)

97

-

S^a (3

6 8-

E

43

-#_

200

< -

C

o

26

lE_

100

_ 0

L 0

Fig. 4. Translation of in vitro synthesized aldolase RNA. Aldolase cDNA was cloned in an in vitro transcription vector and transcribed with SP6 RNA polymerase. The coding strand as template induced synthesis of one major protein (lanes 3 and 4 arrow; 1 and 0.1 g template, respectively) which migrated at the same position as aldolase immunoprecipitated from a P.falciparum lysate (lane 1). Lanes 5 and 6 show translations of 1 and 0.1 tg antisense RNA as a control, and lane 2 shows the background translation without input RNA. Lane M contains molecular weight marker proteins and sizes are indicated in kilodaltons. The non-radioactive marker proteins in lanes M and 1 were stained with Coomassie blue and aligned with the autoradiography using needle marks.

reticulocyte lysate. The aldolase cDNA of clone M25-1 (starting at the first thymidine residue shown in Figure 2) was subcloned in the transcription plasmid pSPT19 in both orientations relative to the SP6 promoter. After transcription with SP6 RNA polymerase both templates were translated in vitro under the same conditions as parasite-derived mRNA (see above). As expected, only the coding strand (plasmid TP41; see Figure 4) stimulated protein synthesis and the product had the same mobility as marker aldolase synthesized by P.falciparum (Figure 4, lane 1). Protein sequencing of the isolated protein band confirmed the result obtained with parasite-derived template RNA: 35S label was recovered in cycles 1 and 8 (Figure 1, codons 1 and 8). The result eliminates the need of any parasite derived molecules for this unique initiation reaction. We can further exclude the more formal possibility that the aldolase RNA cloned and sequenced belongs to an abundant pool of unprocessed precursor transcripts. A simple U-A inversion by RNA editing, for example, could theoretically convert the UAG stop signal to a common AUG start codon. The ability of rabbit ribosomes to initiate at UAG encouraged us to believe that P.falciparum ribosomes could recognize the same initiation codon in vivo. Initiation at UAG in vivo Computer analysis of the aldolase polypeptide sequence predicts no glycosylation site and initiation at UAG in vivo could account for the identical mobility of natural and in vitro synthesized aldolase on SDS -polyacrylamide gels. Direct protein sequencing of the amino-terminus failed, probably

25

Retention

50 time

75

(minutes)

Fig. 5. Detection of the amino-terminal peptide of P.falciparum aldolase synthesized in vivo. Purified P.falciparum aldolase labelled in vivo with [3H]alanine was cleaved with CNBr in the presence of carrier recombinant aldolase expressed in E.coli. The amino-terminal CNBr peptide sequence of the recombinant protein deviates from the natural protein (MRGSHHHHHHGSELACQYM versus MAHCTEYM) and the expected peptide was therefore chemically synthesized either acetylated (position D: Ac-AHCTEYMNA) or nonacetylated (position C: AHCTEYMNA). CNBr cleavage after the internal methionine residue results in the natural sequence and a shift in the detection profile (top) indicates cleavage (position B: acetylated, cleaved; position A: non-acetylated, cleaved). A total of 75 fractions were collected at the column outlet for scintillation counting. Black bars show the 3H radioactivity detected in each fraction. The highly labelled fractions (2839 c.p.m. total) eluting after 67 min correspond to internal, alanine-rich (39 residues) CNBr peptides of P.falciparum aldolase.

due to a blocking group bound to the first amino acid of the polypeptide. Methionine specific cleavage of P.falciparum aldolase with CNBr would release the aminoterminal octapeptide if initiation occurs at UAG in vivo (Figure 1, residues 1-8). A unique alanine residue (Figure 1, codon 2) allows specific labelling of this predicted peptide in vivo with [3H]alanine. Aldolase purified from a pulse labelled P.falciparum culture was cleaved with CNBr in the presence of recombinant aldolase expressed in E. coli (H.Dobeli, unpublished results) and two synthetic peptides that had the expected sequence (AHCTEYMNA or AcAHCTQYMNA) after cleavage. We included an acetylated form of the hypothetical amino-terminal peptide because the first methionine residue in eukaryotic proteins is frequently removed followed by acetylation of the penultimate amino acid residue. We next established conditions that allowed separation of the non-cleaved and CNBr cleaved marker peptides, acetylated and nonacetylated, by reversed-phase high performance liquid chromatography (HPLC; Figure 5). We chromatographed the CNBr cleavage products of pulse labelled natural aldolase under the same conditions and measured the radioactivity in each fraction. A single radioactive peak eluted at the same position as the CNBr-cleaved 1647

P.Ghersa et al.

acetylated marker peptide which proves indirectly that initiation of aldolase synthesis in vivo occurs at the UAG stop codon. Removal of the start methionine residue and acetylation of the alanine residue (Figure 1, codon 2) explains the failure to obtain a direct amino-terminal amino acid sequence of parasite derived aldolase. We conclude that UAG termination codon can be recognized as an initiation signal for ribosomal P.falciparum aldolase synthesis in vitro and in vivo.

40~ ~ 4 a%CG

"INITIATION" IA

C

UA

Discussion We have shown by protein, DNA, cDNA and direct mRNA sequencing that the aldolase gene of P.falciparum lacks a conventional AUG codon for initiation of ribosomal protein synthesis. Protein sequencing of in vitro synthesized aldolase with either parasite-derived or in vitro synthesized RNA as template selects a UAG chain termination codon as the binding site for methionine charged initiation tRNA. In addition, we detect an in vivo synthesized amino-terminal CNBr peptide which can only exist when the ribosomes initiate at the UAG start site. A UAG codon can thus serve as an initiation and termination signal in the process of ribosomal protein synthesis in P.falciparum. Initiation at UAG may be restricted to P.falciparum and within the organism an exception. The reading frame in all P.falciparum gene sequences contained in the EMBL sequence data library is terminated by a UGA or UAA codon. As in other organisms, AUG appears to be the sole initiation codon although this was not confirmed by protein sequencing. It is therefore possible that UAG is an ancestral form of an initiation signal which has been maintained in the evolution of malaria parasites for functional reasons. The aldolase gene sequence of the rodent malaria parasite P.berghei is identical to P.falciparum including 15 nt upstream of the UAG codon (Figure 1, codon 1). In addition, both polypeptides have the same mobility on 8 % SDS -PAGE gels which suggests that both organisms use the same initiation site for aldolase synthesis (data not

shown). Aldolase is a central key glycolytic enzyme with four allelic subunits. The number of amino acids per aldolase subunit (368 residues) and the primary sequence is highly conserved in evolution, and only trypanosome aldolase has a unique 15 amino acid insertion of a transport signal sequence (Kelley and Tolan, 1986; Marchand et al., 1988). Consistent with our finding, the first methionine in other vertebrate aldolases is removed after translation. In human, rabbit and mouse aldolases, alanine and histidine (Figure 1, codons 2 and 3) are the first two amino acid residues at the amino-terminus of the mature enzyme (Kelley and Tolan, 1986). Initiation of protein synthesis at AUG (Figure 1, condon 8) would remove this partially conserved aminoterminal heptapeptide and it is therefore reasonable to assume that initiation of protein synthesis at the UAG condon is required for aldolase function. This abnormal start codon is in a favourable sequence context for initiation. A guanine residue following the initiation codon is the most frequent base after efficient AUG start codons (Kozak, 1983). This could help stabilize binding of the anticodon sequence (CCAU) of a mismatched initiation tRNA. The aldolase gene sequence upstream of the initiation signal is extremely AT-rich (90%; Certa et al., 1988) 1648

UA UA UA UA UA UA 5'----UUUAUAUUUUUAUA UUACCAGCAGAUGUU----3

I

c DNA

Fig. 6. Secondary structure prediction of P.falciparum aldolase mRNA at the start site of translation. The start point of translation (UAG, 'initiation') and the first AUG in-frame are marked by arrows pointing in the direction of translation. The model implies that the AUG 'methionine' codon is perfectly paired with the upstream sequence which would impair binding of initiation tRNA to the double-stranded transcript. The beginning of the cDNA sequences shown in Figure 2 is marked

by an arrow.

which is commonly found at start sites in yeast, slime moulds and, especially, P.falciparum (Kozak, 1983; Caron and Meyer, 1985; Tanabe et al., 1987). Computer-assisted secondary structure analysis of aldolase mRNA further predicts a hairpin loop structure at the 5' end of the message including both the UAG and AUG codons of the transcript (Figure 6). This sequence is transcribed and present in the cDNA clones shown in Figure 2 and in the aldolase mRNA sequence. According to the model the AUG is paired with the upstream sequence whilst the UAG site is in a more accessible position. This could provide a molecular basis of preferred binding of regular initiation tRNA to the UAGG sequence. Detection of this bifunctional UAG codon indicates that conventional assignment of the first AUG in-frame of eukaryotic mRNA (cDNA) as the start site requires direct confirmation by protein sequence analysis.

Materials and methods Parasites The P.falciparun isolates KI (Hall et al., 1984) and M25 (Perrin et al., 1985) were cultivated as described (Certa et al., 1987). An asynchronous culture of Pfalciparwm (KI isolate) was labelled in alanine-depleted medium (Gibco) for 8 h with 50 ,uCi/ml [3H]alanine (New England Nuclear) at a parasitaemia of 20%. Labelling with [3S]methionine (New England Nuclear) was performed in methionine-depleted medium under the same conditions.

Isolation of radiolabelled P.falciparum aldolase Infected red blood cells were collected by centrifugation and lysed in the presence of 0.5% Triton X-100 by five freeze -thaw cycles and sonication. After high speed centrifugation the aldolase was immunoprecipitated with rabbit antiserum to p41 (Certa et al., 1988). The precipitate was further purified by SDS-PAGE (Laemmli, 1970) using recombinant aldolase as a size marker. The marker lane was stained with Coomassie blue and the gel slice at the same position in the non-stained gel was isolated. 3H-labelled aldolase was subsequently eluted from the gel and dialysed overnight against 10 mM Tris-HCI, pH 7.4.

Initiation at a UAG stop codon Protein sequence analysis Sequencing of protease V8 peptides of P.falciparum aldolase was performed as described (Certa et al., 1988). Immunoprecipitated in vitro translated aldolase was transferred from a SDS gel to a PVDF membrane (Millipore) and sequenced as described (Becherer and Lambris, 1988).

Peptide analysis SDS-gel-purified, in vivo labelled aldolase was mixed with 600 yg P.falciparum recombinant aldolase (H.Dobeli, unpublished results), 200 Ag peptide C (AHCTEYMNA) and 200 Ag peptide D (Ac-AHCTEYMNA) before cleavage with CNBr. The peptides were redcued with 10 mM dithiothreitol (DTT) at pH 9.0 for 90 min at room temperature, and separated by reversed-phase HPLC (RP- 18 column) using 0.1 % trifluoro acetic acid and an acetonitrile gradient designed to resolve the small peptides. Elution was monitored with UV detector at 230 nm and 1 min fractions were collected at the column outlet for scintillation counting. The amino acid

composition of peaks A and B showed threonine, glutamic acid, alanine, threonine and histidine in equal amounts. Methionine was degraded by CNBr and identified as a mixture of homoserine and homoserine -lacton. Cysteine escaped detection with o-phtalaldehyde. The retention times of the DTTreduced peptides A (AHCTEYM), B (Ac-AHCTEYM), C (AHCTEYMNA) and D (Ac-AHCTEYMNA) on the RP-18 column were determined by injection of the cleaved and the non-cleaved peptides. Construction and screening of DNA libraries A genomic library of the M25 isolate of P.falciparum with partially DraI digested DNA was constructed as described (Certa et al., 1987) and screened with the g41-D aldolase gene probe of the KI isolate (Certa et al., 1988). Construction of the M25 cDNA library was described earlier (Certa et al.,

1988).

Sequence analysis of positive clones Two genomic clones (M25-A and M25-B) and four cDNA clones (M25-1, M25-13, M25-29 and M25-50) were isolated and analysed in this study. Restriction analysis revealed that all covered the entire aldolase coding sequence. Restriction fragments containing the 5' end were subcloned in M13 and sequenced as described (Yanisch-Perron et al., 1986; Sanger et al., 1977). The 3 kb EcoRI insert of clone M25-A was subcloned into M13 mpl8 (Yanisch-Perron et al., 1986). Double stranded phage DNA was HindIII restricted and end-labelled with [32P]ATP using T4 polynucleotide kinase. After DraI digestion, a 1 kb fragment containing the 5' sequencing of the aldolase gene was isolated and subjected to base-specific chemical cleavage (Maxam and Gilbert, 1980). Hybridizations and other DNA manipulations were carried out according to standard procedures (Maniatis et al., 1982).

mRNA preparation and sequencing P.falciparnn mRNA was prepared as described (Certa et al., 1988). Primer extension sequencing was performed in the presence of actinomycin D in order to prevent transcription from DNA contaminants in the RNA preparation (Geliebter, 1987). The sequencing primer had the sequence

5'-ATCTGCTGGTAATTTTTTTGGGG-3'. In vitro transcription The insert of cDNA clone M25-1 was released with EcoRl and cloned in both orientations in the polylinker of the plasmid pSPTl9 (Boehringer Mannheim). The resulting plasmids TP41 and TP41R were either digested with SspI (TP41) or with PstI (TP41R). SspI cuts once in the vector - 500 nt upstream of the SP6 promoter and 108 nt downstream of the TAA termination codon of the aldolase gene. The DNA fragments containing the transcription unit were isolated and transcribed with SP6 RNA polymerase using a commercial kit and the suppliers instructions (Pharmacia). After digestion of the template DNA with DNase I the reaction mixture was dialysed and used directly for in vitro translation (see below).

Acknowledgements We thank Y.Burki and D.Rotmann for excellent technical assistance and Dr W.Bannwarth for the synthesis of the mRNA sequencing primer. We are especially grateful to Dr D.Gillessen and A.Trzeciak for synthesizing the amino-terminal peptides. We thank Dr G.Langsley, Professor J.Scaife and Dr M.Steinmetz for discussions and suggestions. Parts of this work were supported by the Swiss National Research Foundation (grant 3.923.0.87).

References Beccera,S.P., Rose,J.A. and Anderson,C.W. (1985) In Calendar,R. and Gold,L. (eds), Sequence Specificity in Transcription and Translation, UCLA Symposia on Molecular and Cellular Biology (New series). Alan R.Liss, New York, Vol. 30, pp. 383-396. Becherer,J.D. and Lambris,J.D. (1988) J. Biol. C7em., 263, 14586-14592. Benne,R., Van Den Burg,J., Brakenhoff,J.P.J., Sloof,P., Van Boom,J.H. and Tromp,M.C. (1986) Cell, 46, 819-826. Caron,F. and Meyer,E. (1985) Nature, 314, 185-188. Certa,U., Rotmann,D., Matile,H. and Reber-Liske,R. (1987) EMBO J., 6, 4137-4142. Certa,U., Ghersa,P., Dobeli,H., Matile,H., Kocher,H.P., Shrivastava,I., Shaw,L. and Perrin,L.H. (1988) Science, 240, 1036-1038. Chen,S.-H., Habib,G., Yang,C.-Y., Gu,Z.-W., Lee,B.R., Weng,S.-A., Silberman,S.R., Cai,S.-J., Deslypere,J.P., Rosseneu,M., Gotto,A.M., Li,W.-H. and Chan,L. (1987) Science, 238, 363-366. Geliebter,J. (1987) Focus (BRL), 9, 5-8. Hall,F.R., McBridge,J., Morga ,J., Tait,A., Zolg,J.W., Walliker,D. and Scaife,J. (1983) Mol. Biochem. Parasitol., 7, 247-265. Huang,W.M., Ao,S.-Z., Casjens,S., Orlandi,R., Zeikus,R., Weiss,R., Winge,D. and Fang,M. (1988) Science, 239, 1005-1012. Kelley,P.M. and Tolan,D.R. (1986) Plant Physiol,, 82, 1076-1080. Kozak,M. (1983) Microbiol. Rev., 47, 1-45. Laemmli,U.K. (1970) Nature, 227, 680-685. Manaitis,T., Fritsch,E.F. and Sambrook,J. (1982) Molecular Cloning: a Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Marchand,M., Poliszcak,A., Gibson,W.C., Wierenga,R.K., Opperdoes,F.R. and Michels,P. (1988) Mol. Biochem. Parasitol., 29, 65-76.

Maxam,A.M. and Gilbert,W. (1980) Methods Enzymol., 65, 499-565. Peabody,D.S. (1989) J. Biol. Chem., 264, 5031-5035. Perrin,L.H., Merkli,B., Gabra,M.S., Stocker,J.W., Chizzolini,C. and Richle,R. (1985) J. Clin. Invest., 75, 17618-1721. Sanger,F., Nicklen,S. and Coulson,A.R. (1977) Proc. Natl. Acad. Sci. USA, 74, 5463-5467. Simmons,D., Woollett,G., Bergin-Cartwright,M., Kay,D. and Scaife,J. (1987) EMBO J., 6, 485-491. Stewart,J.W., Sherman,F., Shipman,N. and Jackson,M. (1971) J. Biol. Chem., 246, 7429-7445. Tanabe,K., Mackay,M., Goman,M. and Scaife,J.G. (1987) J. Mol. Biol., 1%, 273-287. Yanisch-Perron,C., Vieira,J. and Messing,J. (1985) Gene, 33, 103- 119. Received

on

October 8, 1989; revised

on

February 14, 1990

In vitro translation A commercial in vitro translation kit with [35S]methionine label (New England Nuclear) was used following the manufacturers protocol. Total RNA extraction and oligo-dT fractionation of mRNA was performed as described (Certa et al., 1988) and either 2 ug total RNA or 100 ng mRNA template were used per assay. Amounts of 0.1 and 12 ig of in vitro transcribed RNA were translated. Immunoprecipitation was carried out with monospecific rabbit seurm to P.falciparum aldolase (p41; Certa et al., 1988).

The products were separated on 8% SDS-polyacrylamide gels (Laemmli, 1970) and electroblotted onto PVDF membranes for protein sequence determination (Becherer and Lambris, 1988). Staining and autoradiography were carried out according to standard procedures.

1649