J Biol Chem, Vol. 275, Issue 14, 10218-10227, April 7, 2000

AU-rich Elements in the 3'-Untranslated Region of a New Mucin-type Gene Family of Trypanosoma cruzi Confers mRNA Instability and Modulates Translation Efficiency^*

Javier M. Di Noia^§, Iván D'Orso^§, Daniel O. Sánchez^¶, and Alberto C. C. Frasch^¶

From the Instituto de Investigaciones Biotecnológicas, Universidad Nacional de General San Martín, C.C. 30, 1650 San Martín, Pcia. de Buenos Aires, Argentina

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Trypanosoma cruzi has a complex mucin gene family of 500 members with hypervariable regions expressed preferentially in vertebrate associated stages of the parasite. In this work, a novel mucin-type gene family is reported, composed of two groups of genes organized in independent tandems and having very short open reading frames. The structures of deduced proteins share the N and C termini but differ in central regions. One group has repeats with the consensus Lys-Asn-Thr⁷-Ser-Thr³-Ser(Ser/Lys)-Ala-Pro and the other a Thr-rich sequence of the type Asp-Gln-Thr^17-20-Asn-Ala-Pro-Ala-Lys-Asp-Thr^5-7-Asn-Ala-Pro-Ala-Lys. In both cases, expected mature core proteins are around 7 kDa. Both groups, named L and S, respectively, differ in the structure of genomic loci and mRNA, with differential blocks in the 3'-untranslated region. The highest mRNA level for S and L groups are in the epimastigote stage but they show distinct developmentally regulated patterns. Transcripts are short lived and their steady-state abundance is regulated post-transcriptionally with increased mRNA stability in insect stage epimastigote. AU-rich sequences, similar to ARE motives known to cause mRNA instability in higher eukaryotes, are present in the 3'-untranslated region of the transcripts. In transfection experiments this sequence is shown to be functional for the L group destabilizing its mRNA in a stage-specific manner. Furthermore, an effect of this AU-rich region on translation efficiency is shown. To our knowledge, this is the first time that a functional ARE sequence-dependent post-transcriptional regulation mechanism is reported in a lower eukaryote.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Mucins are heavily O-glycosylated molecules, with carbohydrates representing up to 80% of the molecular mass, whose protein cores are rich in Thr/Ser and Pro residues. In the human pathogen Trypanosoma cruzi, the ethiological agent of Chagas' disease, there are mucin-type glycoproteins that have been increasingly studied because of their role as the main sialic acid acceptors (1, 2) in the reaction catalyzed by trans-sialidase (3, 4). This unique enzymatic activity seems to have been selected during evolution to circumvent the lack of de novo synthesis of this sugar in this protozoan parasite (5). The composition and diversity of heavily O-glycosylated mucin-type molecules have shown particular features, like O-glycosidically linked N-acetylglucosamine instead of N-acetylgalactosamine (6, 7), and a large mucin gene family of about 500 members having hypervariable regions (8). Because of their size and structure, T. cruzi mucins are comparable to endothelial mucin-type adhesion molecules, better than to high molecular weight epithelial mucins. Additionally, T. cruzi being a digenetic parasite, there are at least two different groups of mucin molecules among life cycle stages: one present in the insect vector forms of the parasite and the other in the mammalian stage (2). T. cruzi infection is established in the mammal by the insect-derived stage metacyclic trypomastigote. It rapidly invades host cells, differentiates to the intracellular replicative stage amastigote, that in turn differentiates to the infective mammal form trypomastigote. This form is released from cells and circulates in blood infecting other cells and being eventually ingested by the insect with its blood meal. Ingested trypomastigotes differentiate to epimastigotes, that are the replicative insect stage. This form migrates along the digestive tract until it differentiates to metacyclic trypomastigotes that are eliminated with the faces closing the circle.

In cell-derived trypomastigotes, mucins were described as major surface glycoproteins anchored by glycophosphatidylinositol (GPI)¹ (1, 9). These mucins migrate in denaturing gels as a smear between 60 and 200 kDa and bear a sialylated epitope involved in the invasion of nonphagocytic cells (10, 11). Additionally, as sialic acid acceptors they play a role in resistance of trypomastigotes to lysis mediated by the alternative pathway of complement (12). A different set of mucin-type glycoproteins exist in the insect dwelling stages, originally described as a periodic acid-Schiff reagent positive double band migrating in the 35/50-kDa range (13). Both, epimastigotes (the replicative insect stage) and metacyclic trypomastigotes (the infective insect stage differentiated from the epimastigote), have mucins that only differ in the nature of the lipid contained in the GPI that anchor them to the plasmatic membrane (14). Despite this, only metacyclic trypomastigote mucins, but not epimastigote ones, were reported to mediate adhesion of the parasite to vertebrate cells (16); even when recent reports indicated an inverse correlation between 35/50 mucins content and strain infective capacity (17, 18). Additionally, the binding of these kind of T. cruzi mucins to L-selectin was reported (19) supporting their role as an adhesion molecule. The molecular mass of the underlying apomucins should be around 5 kDa as suggested by compositional (1) and mass spectrometry analysis (15) but no mucin-like genes are known with these characteristics.

The first mucin-type gene in T. cruzi was reported by our group (20) and resulted in belonging to a very large and complex mucin-type gene family having highly conserved sequences at the corresponding N and C termini of the deduced protein, flanking variable central regions (21). The central domain of the so-called TcMUC family can encode a variable number of tandem arrayed repeats with the consensus sequence Thr⁸-Lys-Pro² or nonrepeated unique central domains that can vary widely within and among strains (8, 15). The mucin nature of the former group of TcMUC products is now clear (15).² The TcMUC family has about 500 members per haploid genome and many, if not all of them, are transcribed simultaneously in a given parasite population (8). The molecular masses of the products deduced from these genes, that are around 18-20 kDa, along with the information from transfection experiments,² point to repeat-containing members of this family as those encoding apomucins present in the mammalian trypomastigote stage. In agreement with this, TcMUC mRNA level is higher in cell-derived trypomastigote stage (21), even when it is expressed in all stages (22) and some individual members mRNA level can be high in the epimastigote stage (8).

In Trypanosomatids, transcription is polycistronic and the major regulation point is exerted at the post-transcriptional level (23). Evidence so far implicate elements present in the 3'-untranslated regions (UTR) of mRNA in determining stage-specific steady state level (24-27). Despite the general validity of this fact, only one mechanism of post-transcriptional regulation is known in some detail in Trypanosomatids. It involves a secondary structure at the 3'-UTR of procyclin mRNA of T. brucei, which modulates transcript stability and translation efficiency in specific developmental stages of the parasite (28).

Two major findings are reported here. First, the characterization of a new mucin-type gene family in T. cruzi, with two gene groups, both of which encode very short deduced polypeptides having all the features of apomucins present in insect derived stages. We consequently named it TcSMUG, for T. cruzi small mucin-like gene family. Due to the similar structure with some TcMUC family products, TcSMUG is likely to encode mucin-type glycoproteins. Second, a post-transcriptional regulation mechanism, mediated by an AU-rich element, is reported for T. cruzi. AU-rich elements (ARE) present in the 3'-UTR of certain mRNAs, are cis-acting motives responsible for the post-transcriptional regulation of inflammatory cytokines, growth, and transcription factors and oncoproteins in mammals mediating mRNA decay and affecting its translation positively and negatively (29-32). This finding has an additional evolutionary interest. So far, ARE have been found to be functional in higher eukaryotes, being Echinodermata the lower branch where they were described (33). The TcSMUG family presents regulated expression throughout the life cycle of the parasite. Differences in steady state mRNA levels among the developmental stages are due to post-transcriptional regulation mechanisms affecting the half-life of these mRNAs. Three mechanisms acting on TcSMUG transcripts are inferred from the results herein. 1) A stabilizing mechanism acting in the epimastigote stage. 2) A destabilizing mechanism mediated by an AU-rich region present in 3'-UTR. 3) An effect of AU-rich element improving translation efficiency.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Parasites-- T. cruzi CL-Brener cloned stock (34) and RA strain (35) were used. Different forms of the parasites were obtained as described elsewhere (15). Purity of the different Trypanosoma stages was determined by conventional microscopy and was at least 95%.

Gene Cloning-- Clones MLe-1 and MLt-1 were obtained from a T. cruzi CL-Brener stock cDNA library constructed by reverse transcriptase-polymerase chain reaction and used as described elsewhere (8). Clones 5.3, 5.4, 7.2, 8.3, 8.4 and 13.4 were obtained from screening a high density arrayed genomic library of the same T. cruzi clone (36), with a probe encompassing the 5' end coding region of the MLe-1 clone (probe N, see Fig. 1). Several positive cosmids were obtained and 18 were randomly chosen, digested with restriction enzymes, and analyzed by Southern blot with the same probe. Positive bands of different sizes were cloned from cosmids 5, 7, 8, and 13 in pBlueScript II KS+ (Stratagene, La Jolla, CA) and one or two clones (when positive bands with different size existed in the Southern blot as in cosmids 5 and 8) were sequenced. Several cDNA sequences were obtained from GeneBank dbEST, identified under the T. cruzi genome project. Thirteen cDNA clones were obtained from our T. cruzi genome project facilities and fully sequenced. GeneBank EST data base accession numbers of ESTs analyzed are: TENS1086, TENS1287, TENS1277, TENS1275, TENS1286, TENS1706, TENS1109, TENS1032, TENS1934, TENS1879, TENS1084, TENS1620, TENS0733, TENS1731, TENS0939, TENS1223, TENS0947, TENS1107, TENS0908, TENS0224, TENS1141, TENS0181, TENS0177, TENS0164, TENS1785, TENG0272, TENF0276, TENF0308, and TENF3023.

Transcription in Isolated Nuclei (Run-on Assay)-- Nuclei were isolated following a modification of the Nonidet P-40 lysis protocol (37). Briefly, 1 × 10⁹ cells were washed in phosphate-buffered saline and resuspended in 10 ml of hypotonic buffer A (10 mM Tris-HCl, pH 7.6, 2 mM MgCl₂, 5 mM KCl, 2 mM CaCl₂, 0.5 mM dithiothreitol, 1 mM EDTA, 1 mM spermidine, 6% PEG 3500) and incubated on ice for 10 min. After the addition of 1% Nonidet P-40, cells were lysed by homogenization with a tissue grinder. The cell lysate was diluted with 1 volume of 2× buffer B (0.64 M sucrose, 40 mM Tris-HCl, pH 7.6, 60 mM KCl, 1 mM EDTA, 1 mM EGTA, 0.5 mM dithiothreitol, 1 mM spermidine), and the nuclei were centrifuged 15 min at 5000 rpm. The nuclear pellet was washed once with 10 ml of 1× buffer B and resuspended in 100 µl of buffer C (25% glycerol, 50 mM Tris-HCl, pH 8, 60 mM KCl, 1 mM dithiothreitol, 0.1 mM EDTA, 1 mM spermidine). The transcripts synthesized were labeled with [-³²P]UTP at 30 °C for 30 min as described previously (38). The labeled nascent RNA was extracted with phenol:chloroform:isoamylic alcohol, 25:24:1, precipitated with ethanol and sodium acetate, and quantified in a -counter. The probe activity was of 10⁷ cpm for epimastigotes and 10⁶ for trypomastigotes. Dot blot hybridizations were carried out on immobilized selected clones for 3 days at 65 °C using the labeled RNA in 10 ml of hybridization solution (0.5 M NaH₂PO₄, 7% SDS, 1 mM EDTA, 1% bovine serum albumin, 100 µg/ml tRNA). Washes were carried out at high stringency conditions (20 mM NaH₂PO₄, 0.1% SDS at 65 °C). The immobilized DNA (3 µg/dot) consisted of the plasmid pT7T3d18 (39) containing different TcSMUG cDNAs; full-length T. cruzi 24S rRNA (kindly given by Dr. E. Bontempi) and a 1-kilobase cDNA fragment of the -tubulin as positive controls. pT7T3d18 was used as background control.

Drug Treatments-- Epimastigote cultures were taken in logarithmic growth phase at a cell density of 3 × 10⁷/ml and treated with actinomycin D (ActD) (Sigma), at a final concentration of 10 µg/ml, known to inhibit transcription in Trypanosomatids (27, 28). Aliquots were taken at different times after addition of the inhibitor. Trypomastigotes were harvested from infected cell culture medium and resuspended in minimal essential medium, 5% fetal bovine serum (Life Technologies, Gaithersburg, MD) at the same cell density and treated equally. Cycloheximide (CHX) (Sigma) was used at a final concentration of 50 µg/ml (40) in the same conditions. In experiments with both drugs, the parasites were preincubated with CHX for a period of 4 h and then ActD was added without CHX withdrawal. This point was considered as time 0 of the treatment. It was confirmed by microscopy that the parasites were viable at all time points used for the experiments. All the aliquots were harvested by centrifugation, washed with phosphate-buffered saline, and frozen at 70 °C until RNA extraction.

Constructions and Transfection-- The Chloramphenicol acetyltransferase (CAT) gene was amplified from vector pBCSK+ (Stratagene) using oligonucleotides CAT/se, gggATGGAGAAAAAAATCACTGGATATA, and CAT/as, cccaagcttTTACGCCCCGCCCTGCCA. The complete TcSMUG intergenic region (IR) was amplified by PCR from the genomic cosmid clone 8 having group L members. The whole IR was fused downstream from CAT and the IR fragment from BamHI site to 5'as/SmaI was fused upstream. The PCR primers contained restriction enzymes sites to facilitate subsequent cloning steps as indicated in Fig. 7A. Oligonucleotides used were: 3'se/HindIII, cccaagcttGAGGACGGGGCGGGGCGCGT; 3'as/XhoI, ggcctcgagCATTGTCACTTGGGCCTTTGT; and 5'as/SmaI, gggCATTGTCACTTGGGCCTTTGTG. AU-rich element deletion was generated by PCR, using primers flanking the AU-rich region with an EcoRI clamp (oligonucleotides ARE-1, cggaattcTTGTGAAAGGGATGTTCGC and ARE-2, cggaattcGGCGCACCGACGCTTCCCCC). Each fragment was cloned in the pTEX vector (41), kindly provided by Dr. J. M. Kelly and/or its derived pRIBOTEX kindly given by Dr. R. Hernández (42), within the cloning sites indicated in Fig. 7. Final constructions were Lc, having cat flanked by the original 5' and 3' IR, and LcAU, essentially identical but with the 50-bp AU-rich region deleted from 3'-IR (exactly the region indicated in Fig. 3C). Transfections were carried out with a BTX 600 electroporator in a 2-mm gap cuvette. 300 × 10⁶ parasites were harvested and washed with BHT medium (43), resuspended in 0.4 ml of BHT with 100 µg of supercoiled plasmid DNA. The electroporation setting was: 1500 microfarads, 335 V, and 24 . Parasites were recovered in 5 ml of BHT supplemented with 10% fetal calf serum (Life Technologies, Gaithersburg, MD) and 48 h later geneticin (Sigma) was added at a final concentration of 500 µg/ml. The neo resistance gene, also present in both plasmids, was used for selection and as an internal control of transfection levels since it is transcribed polycistronically from the same promoter (41, 42). The polyadenylation site of the CAT mRNA derived from Lc and LcAU constructions both in pTEX- and pRIBOTEX-transfected parasites was determined by RT-PCR using the oligonucleotide anchord(T) 5'-GCGAGCTCCGCGGCCGCG(T)₁₈-3' as primer for reverse transcription using the Superscript II enzyme (Life Technologies, Gaithersburg, MD) as indicated by the manufacturer, on total RNA from the parasites. PCR was performed on first strand product using CAT/se and an oligonucleotide with the anchor sequence of anchord(T). The products were cloned in pGEMT-Easy (Promega, Madison, WI) and sequenced.

CAT Assay-- An equal number of parasites from each transfected population was harvested, washed once with 0.25 M Tris-HCl, pH 8, and cellular extracts were prepared by four freeze-thaw cycles and heat inactivation. Cell lysates were assayed for CAT activity as described previously (42). Reactions were conducted for 1 h at 37 °C with cellular extracts prepared from 10⁸ parasites. This time was previously adjusted to fit within the linear range of the assay. Conversion of [¹⁴C]chloramphenicol to acetylated forms was analyzed by thin layer chromatography and quantified by densitometry.

General Methods-- DNA was purified using the conventional proteinase K-phenol extraction method (44). Southern blot analysis for genomic and cosmid DNA was performed following standard protocols (44). Partial digestion of cosmids was done as described in the laboratory manual (45). All the probes were radioactively labeled with [-³²P]dCTP (NEN Life Science Products Inc.) by polymerase chain reaction as in Ref. 46. Probe N was obtained using oligonucleotides ATGATGCTGCGCCGCGTCCT and GTCGCCCCCAGCAGCCTCA on clone TcSMUG-MLe-1, probe 3'L using oligonucleotides GGGGGAAGCGTCGGTGCGCC and TGTCGCTATCGGCTCCCTGG on clone TcSMUG-7.2, and probe 3'S using oligonucleotides ATGAGGACGGGGCGGGGCG and TCAGCGCCGAAAACGTAC on clone TcSMUG-13n22.

Sequencing was performed either manually using Sequenase 2.0 (Amersham Life Science, Cleveland, OH) or by dye terminator cycle sequencing chemistry in an ABI PRISM 377 DNA Sequencer (Perkin-Elmer). Computer analysis of sequences was done on Lasergene package (DNASTAR Inc, Madison, WI) Some of the alignments were done using DIALIGN 2.0 software (47). Sequence similarities in GeneBank data bases were analyzed using BLAST algorithm at the National Center for Biotechnology Information Internet site.

RNA was purified using TRIzol reagent (Life Technologies) following the manufacturer's instructions. Northern blots were carried out as described in Ref. 48. Zeta-Probe nylon membranes (Bio-Rad) were used for all blottings.

Densitometry was done using software ImageQuant v3.2.2 (Molecular Dynamics). Autoradiographies were optically scanned at 300 dots per inch resolution, and the signal intensity quantified. A calibration of the software was done to assure linearity of the determinations under our working conditions.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The TcSMUG Family Is Composed by Two Groups of Genes having Very Small Mucin-type Open Reading Frames-- While screening for mucin-type genes in T. cruzi, two clones were identified and sequenced that were clearly different from members of the TcMUC gene family previously described (21). These partial clones contained the spliced leader (49), 5'-UTR and an incomplete open reading frame (ORF) encoding a stretch of Thr. Both clones were named MLe-1 and MLt-1 because they were mucin-like and obtained from epimastigote and trypomastigote stages, respectively. They were almost identical except for the size of the Thr-rich region. A probe corresponding to the N terminus of the deduced amino acid sequence was generated (probe N, Fig. 1), and used to screen a genomic cosmid library of T. cruzi. Six clones containing the complete ORF were sequenced. Further sequences were obtained from the currently ongoing genome project for T. cruzi from an expressed sequence tags (EST) data base, searching for cDNAs homologous to the different regions of the genomic clones. Twenty-eight cDNAs found in the EST data base were analyzed to have an idea of the diversity of the family, since most of them contained all the ORF. Eleven of them, plus two new cDNA clones were fully sequenced to determine the features of the mRNA. Sequence analysis of all the ORFs showed two kinds of deduced amino acid sequence (Fig. 1). This dichotomy was later correlated to differences in the corresponding genes and mRNA structure. These groups were named L and S for long and short because of the transcript length (see below). Twenty-one out of the 28 cDNAs analyzed belonged to the S group and 7 to the L group.

View larger version (42K): [in this window] [in a new window]   Fig. 1.   Structure of the proteins deduced from TcSMUG genes. A, alignment between a member of the L group (clone 7.2) and another from the S group (clone 13f5) compared with a repetitive member of the previously described TcMUC family (EMUCe-3) (21). Identities among two or more sequences are shaded, dashes indicate gaps introduced for best alignment. A representative repeat is boxed for clone 7.2 and EMUCe-3. Letters a, b, c, and d are to indicate the different regions schematized in the graphic. The asterisk indicates a conserved N-glycosylation site. The putative GPI addition site (the  cleavage site and the following positions +1 and +2, the spacer and hydrophobic tail) as predicted using current knowledge (51) and comparison with several known or predicted Trypanosomatids' signals are indicated. B, schematic in scale representation of both kinds of deduced TcSMUG proteins indicating the percentage of identities between conserved regions and the consensus sequence of central Thr-rich regions. N indicates a conserved N-glycosylation site. The region encompassed by probe N is indicated.

Regarding the coding region, both kinds of deduced sequences have very similar N and C termini regions with an average 80% of identity and higher similarity values because of conservative changes (see Fig. 1A for an alignment between L and S group members). Within each group both regions were almost identical in all sequences analyzed. A minor subgroup was found within S group with some amino acid changes at the boundary between the N terminus and central region (not shown). The N terminus has a putative signal sequence extending to amino acids 17-20 as predicted by von Heijne's method (50). The last 30 amino acids of the C terminus are compatible with a GPI anchor signal (51). Were these two regions processed, the average predicted molecular mass for the mature proteins would be around 7 kDa with Thr representing as much as 50% of the residues. Eight amino acids, mostly polar ones, account for over 85% of the mature proteins. The most represented amino acids were, 34% Thr, 7.2% Ser, 7.2% Pro, 8.4% Lys, 4.8% Gly, 6.4% Glu, 4.8% Asp, and 14.4% Ala for the L group (average of 4 members) and 43.1% Thr, 5.7% Gln, 5.5% Pro, 4.2% Asn, 5.7% Gly, 5.7% Glu, 5.6% Asp, and 13.4% Ala for the S group (average of 10 members). The main difference between groups L and S resides in the central region. While the L group has 1 to 3 tandem arrayed repeats with the consensus Lys-Asn-Thr⁷-Ser-Thr³-Ser(Ser/Lys)-Ala-Pro, the S one has the sequence Asp-Gln-Thr^17-20-Asn-Ala-Pro-Ala-Lys-Asp-Thr^5-7-Asn-Ala-Pro-Ala-Lys (Fig. 1B). Consistently, the first Thr run is longer than the second that, with very low frequency, could be either duplicated or absent. So, differences among members arise from different number of repeats in the L group, and eventually from different numbers of Thr runs in the S group. An interesting feature in the latter group is the difference in the number of Thr of each run resulting from insertion/deletion of complete codons. Some clones were identical in nucleotide sequence but for the presence of one single Thr codon. Due to the different origins of sequences analyzed (i.e. genomic cloning, cDNA library, RT-PCR) it is unlikely that this could be due to a cloning artifact.

A polyclonal antibody against a fusion protein containing the complete ORF of TcSMUG-13f5 expressed in pMALc2X vector (New England Biolabs, Beverly, MA) was raised. This serum failed to recognize any band in Western blots of total lysates of the four developmental stages of T. cruzi. This would be the expected result if the protein were shielded from antibodies by post-translational modifications. This possibility is supported by the overall similarities of TcSMUG products with members of the repeat-containing group of TcMUC family (21). An alignment is shown of representative members of the three kinds of sequences in Fig. 1A. Similarities in the signal peptide, GPI anchor sequence, and indicated N-glycosylation site are high enough to assume that they will be functional, as was demonstrated for TcMUC and will be reported elsewhere.² The mucin nature is more hypothetical, but as Thr⁸-Lys-Pro² repeats from TcMUC genes are good substrates of the T. cruzi O-glycosylation initiating enzyme (7), the high structural and compositional similarity of repeats and Thr runs from L and S groups with TcMUC repeats suggest that they could be O-glycosylated, as is the case for TcMUC.²

The TcSMUG Genes Are Arranged in Tandem in the T. cruzi Genome with 74 Genes Estimated per Haploid Genome-- Cosmids containing TcSMUG genes were analyzed by Southern blot with the N probe. The resulting band pattern suggested a tandem array of these genes, because several different restriction enzymes, later found to be present only once in the genomic clones sequenced, gave the same pattern (not shown). Fig. 2A (right panel), shows the PstI restriction pattern for cosmid 7, that is compatible with the presence of at least 7 genes in tandem, with a repetitive unit of 1300 bp. Cosmids 8 and 13 gave the same unit as cosmid 7, all of them belonging to the L group (Fig. 2A, left panel). An additional slightly larger band in cosmids 8 and 13 is due to different number of repeats in the ORF of individual tandem units, as determined by DNA sequencing. The same applies to cosmid 5, except that the repetitive units are 900 and 950 bp. Hybridization of cosmids with probes of different intergenic regions suggested that both, S and L genes, are arranged in tandems but are segregated from each other (not shown). In Fig. 2B, a genomic DNA Southern blot is shown where HindIII only partially digested the genomic DNA of T. cruzi, showing the same kind of pattern typical of tandemly arranged genes, at least for some of the L group genes. Sequence data indicates that this enzyme does not cut within the S group as expected from the observed pattern. Total digestions of genomic DNA hybridized with probe N showed a simple pattern with bands of 1300 and 900 bp, such as in the case of Sau3AI for enzymes cutting only once in each tandem unit. Enzymes that do not cut within the tandems usually gave very large bands as in XbaI digestions. The EcoRI site seems to be polymorphic, being present only in some tandem units of S group genes but not in L group ones, and this was confirmed by cosmid digestion and sequencing (not shown). Finally, several HaeIII sites are present within the ORF of the sequenced genes, explaining the lack of signal in the Southern blot.

View larger version (51K): [in this window] [in a new window]   Fig. 2.   Genomic structure of TcSMUG family. A, Southern blot of cosmids 7, 8, 5, and 13, containing TcSMUG genes (left panel) scaled to a Southern blot of a partial digestion of cosmid 7 (right panel). In all cases DNA was digested with PstI and separated in agarose gels, transferred to nylon membranes, and hybridized with probe N under stringent conditions (0.1 × SSC 65 °C). Size in kilobase pairs (kbp) is indicated for the units of the tandem repeats as calculated using HindIII digested  DNA (Life Technologies, Gaithersburg, MD) and HaeIII X174 (New England Biolabs, Beverly, MA) as markers. B, total T. cruzi DNA was digested with the enzymes indicated and treated as described above. HindIII digestion was partial.

A rough estimation of the gene number was made as follows. A high density arrayed cosmid library, screened with N probe in high stringency conditions, yielded 242 positive clones. Normalizing to the library complexity (36) this represents 10.6 positive cosmids per haploid genome. From total and partial digestions of cosmids and genomic DNA we know that all genes are arranged in tandem, and that each tandem has at least seven genes. So, the estimation of gene number for the CL-Brener cloned stock of T. cruzi is 74 genes per haploid genome, without discriminating between both TcSMUG groups.

Genes from S and L Group Are Transcribed as Developmentally Regulated, Different Sized mRNA, Distinguished by their 3'-UTR-- In order to analyze the expression of the TcSMUG family, total RNA was extracted from the different stages of development of T. cruzi and hybridized in Northern blots with probe N, conserved in all the members. In Fig. 3A the general pattern of expression throughout the life cycle is shown, with two bands of 1400 and 1000 bases. As will be detailed below, groups L and S of the family have the larger differences in their 3'-UTR sequences. To find out if the upper band corresponded to transcripts of group L and the lower band to group S, two probes named 3'S and 3'L that recognize differential regions in their 3'-UTR were used (see Fig. 3C). Hybridization of epimastigote stage RNA with these probes indicated that this was indeed the case (Fig. 3B).

View larger version (45K): [in this window] [in a new window]   Fig. 3.   mRNA expression pattern of TcSMUG. A, Northern blot of total RNA from the main T. cruzi developmental stages epimastigotes (E), cell-derived trypomastigotes (T), metacyclic trypomastigotes (MT), and amastigotes (A) was performed. Total RNA was extracted and fractionated in denaturing gels, transferred to nylon membranes, and hybridized with probe N. Size in bases as calculated using RNA markers (Life Technologies, Gaithersburg, MD) are indicated. B, one filter containing epimastigotes RNA treated as described above was sequentially hybridized with the indicated probes. C, in scale schematic comparison of the mRNA structure of TcSMUG family L and S groups. Dashed lines are spacers for best alignment. Percentages of homology between conserved regions (gray boxes) are indicated. Regions present in only one mRNA are drawn as white boxes. The probes used in B are indicated. AU-rich elements are indicated as black boxes and their sequences shown. D, the genomic sequences of the conserved trans-splicing acceptor sites and polyadenylation site used in over 95% of cDNA clones analyzed are indicated.

Both, L and S groups were differentially regulated among parasite developmental stages. We initially focused on epimastigote and trypomastigote as representative insect and mammalian dwelling stages, respectively; since the greater differences in steady state mRNA levels were observed between them. In epimastigote, both transcripts were present at the highest level when compared with the other stages (Fig. 3A). In contrast, only the L group was detectable at this time in trypomastigotes. Longer exposures allowed us to detect both bands and perform densitometry. The comparative steady state mRNA levels were quantified after normalization performed by hybridization with 24S rRNA. The values obtained for the L group were 5.9-fold higher in epimastigotes than in trypomastigotes, a ratio that rose to 10:1 for the S group. Developmental regulation exist in the other stages too, with group L being present in all the stages, but with higher levels in the replicative epimastigote and amastigote stages. On the other hand, group S mRNA was mainly present in insect-associated stages being barely detected in vertebrate associated ones (Fig. 3A).

Comparisons of the nucleotide sequence of intergenic and untranslated regions of many genomic and cDNA clones from both groups showed 95% sequence homology in the 130 bp upstream of the ORF and indicated that the same region is used for trans-splicing (49). Any of three conserved AG are used as 3' acceptor sites, leading to the presence of almost identical 5'-UTR of about 55 bases in length in all the cDNAs analyzed (see Fig. 3D). Sequences downstream from the ORF showed the larger differences between both groups, but members were still very homogeneous within each group. This fact is in turn reflected in different 3'-UTR of mRNAs that explain the different sizes of the transcripts as shown in Fig. 3B. This region is composed of highly conserved boxes, shared by both groups mRNA, and other boxes exclusive of each group. A schematic comparison between UTRs of L and S group is shown in Fig. 3C. After the stop codon the next 30 bases are identical in both groups, then there are 40 bases present only in the L group, after which a 82% homologous 90-base stretch is present. Then there is a region of 80 bases present only in the S group and again a highly conserved 100-base region. Following, we found an AU-rich region in both transcripts. In the L group there are two AU₃AU₃A motifs within this region. These sequences are similar to ARE that were described in the 3'-UTR of mRNA of higher eukaryotes, being sites where regulatory factors can act, affecting messenger stability (29). After this point an insertion occurs in the L group, a previously described short interspersed repeat element (SIRE) (52) of 400 bases is present, that is the major cause for the different length of both transcripts. Finally there is a 40-base region with 67% homology where polyadenylation occurs. The poly(A) addition site is conserved for both groups (Fig. 3D). A minority of the L cDNA clones used a polyadenylation site within the SIRE (not shown).

Differences in the Steady State of TcSMUG mRNA Levels Observed between Developmental Stages Are Due to Post-transcriptional Events-- A difference in the steady state mRNA levels among parasite stages was observed for both L and S TcSMUG groups (Fig. 3A). As gene expression regulation in Trypanosomatids is largely attributed to post-transcriptional mechanisms (23) we wondered if this kind of regulation could account for the observed differences. We focused first on the two extreme cases (epimastigote and trypomastigote) to analyze the half-life of TcSMUG mRNA, by incubating them in the presence of ActD, a drug that inhibits transcription by RNA polymerase II. Again, a clear difference was obtained between epimastigote and trypomastigote stages. In epimastigotes, both L and S group mRNAs were affected by ActD treatment, and the half-lives were higher than 6 h in both cases (Fig. 4). When the same experiment was done on trypomastigotes, the half-life of the L group mRNA was found to be less than 0.5 h (Fig. 5). The signal for the S group mRNA could be hardly detected under our conditions but after long exposures the half-life gave similar values than for the L group (not shown). These results indicate that, at least in part, the differences in the steady state mRNA level between epimastigote and trypomastigote stages are due to differential stability of TcSMUG mRNA.

View larger version (65K): [in this window] [in a new window]   Fig. 4.   Stability of TcSMUG transcripts in the epimastigote stage of T. cruzi. A, Northern blot of total RNA from epimastigotes treated or not with ActD and/or CHX for the times indicated above each lane. Hybridizations were sequentially performed with probe N and a 24S rRNA probe (shown under each panel), under stringent conditions (0.1× SSC, 65 °C). B, quantification of TcSMUG L group mRNA levels of the bands from the Northern blots show in A. Autoradiographies were scanned, the signals quantified using ImageQuant v3.2.2 (Molecular Dynamics), and the ratio of TcSMUG L group mRNA to 24S rRNA plotted. C, same for TcSMUG S group. Closed circles, control without treatment; crosses, actinomycin D; open circles, cycloheximide; closed squares, cycloheximide plus actinomycin D.

View larger version (68K): [in this window] [in a new window]   Fig. 5.   Stability of TcSMUG transcripts in the trypomastigote stage of T. cruzi. A, Northern blot of total RNA from cell-derived trypomastigotes treated or not with ActD and/or CHX for the times indicated above each lane. Hybridizations were sequentially performed with the probe N and a 24S rRNA probe (shown under each panel), under stringent conditions (0.1× SSC, 65 °C). B, quantification of TcSMUG L group mRNA levels of the bands from the Northern blots show in A. Autoradiographies were scanned, the signals quantified using ImageQuant v3.2.2 (Molecular Dynamics), and the ratio of TcSMUG L group mRNA to 24S rRNA plotted. C, same for TcSMUG S group. Closed circles, control without treatment; crosses, actinomycin D; open circles, cycloheximide; closed squares, cycloheximide plus actinomycin D.

As an effect of transcription initiation could not be discarded, run-on assays were conducted in epimastigotes and trypomastigotes (Fig. 6). Relative values were determined by densitometric quantification resulting from hybridization of dots, containing clones of a member of the L group and a member of the S group, with total RNA synthesized in isolated nuclei in the presence of [-³²P]UTP. Values were substracted from background, taken as the signal obtained from the plasmid without insert. Values for one S group member (clone 13f5) were similar between both stages, as well as for one L group member (clone Mle-1), being the epimastigote to trypomastigote ratios of 1.06 and 1.73 for clones 13f5 and Mle-1, respectively. Thus, the presence of the S group mRNA in epimastigote in contrast with its virtual absence in trypomastigote (see Fig. 3A), should be regulated primarily by post-transcriptional events since it is transcribed equally in both stages. The same applies for the L group, because the observed differences in transcription initiation cannot account for differences in steady state mRNA levels. Transcription activity is lower in trypomastigotes than in epimastigotes (Ref. 53 and this work). If -tubulin could be considered as an indicator of the transcriptional activity of the stage, and the values normalized to it, then the transcription initiation rate would be even in both stages for the L group and half in epimastigotes than in trypomastigotes for the S group, increasing the importance of post-transcriptional events in mRNA abundance.

View larger version (26K): [in this window] [in a new window]   Fig. 6.   Comparative values of transcription initiation for L and S group members between epimastigote and trypomastigote stages. Run-on experiments on epimastigote and cell-derived trypomastigote stages were conducted by labeling nascent RNA in isolated nuclei in the presence of [-32P]UTP (NEN Life Science Products Inc.). Signals were scanned and quantified using ImageQuant v3.2.2 software (Molecular Dynamics) within the linear range. Values are in arbitrary units and corrected to internal background controls in each case. 13f5 is a member of the S group, Mle-1 is a member of the L group, -tubulin and 24S rRNA ribosomal RNA were included as constitutive positive controls. Differences in the signal between 13f5 and Mle-1 are due to different sizes of immobilized clones and do not indicate relative values of transcription between them. The corresponding dot blot hybridizations are shown under each column.

Effects of Protein Synthesis Inhibitors on the Stability of TcSMUG mRNA Family-- One possible explanation for the differences in TcSMUG mRNA stability between epimastigote and trypomastigote stages could be the presence of protein factors affecting its turnover, either positive or negatively. To test this hypothesis, half-life determinations were repeated in the presence of protein synthesis inhibitors for both stages. Parasites were pretreated during 4 h with CHX, and ActD was then added, without withdrawal of CHX, to assure a complete effect on protein synthesis inhibition (40). This double treatment in epimastigotes caused the TcSMUG mRNA half-life to diminish to about 2 h for L and 2.5 h for S group (Fig. 4, B and C, respectively), at least a 3-fold decrease as compared with parasites without the CHX treatment. On the other hand, ActD plus CHX-treated trypomastigotes showed no difference with those treated only with ActD for the L group mRNA level (Fig. 5), at least for the time points tested. Again, CHX treatment by itself showed no effect on the level of this mRNA (Fig. 5). Cycloheximide by itself do not seem to exert any significant effect on mRNA levels since after treatment with ActD the mRNA levels of L and S transcripts in both stages remained constant (Figs. 4 and 5). So, the reduction in mRNA levels is not because cycloheximide could be indirectly reducing the transcription rate for the TcSMUG genes but rather because the mRNA is being more rapidly degraded. Two simple possibilities to explain these results are discussed below.

AU-rich Elements in the 3'-UTR of L Group Affects mRNA Stability and Modulates Translation Efficiency-- AU-rich sequences known as ARE were described in higher eukaryotes as a destabilizing cis-acting element causing short lived mRNAs (29). We tested if the AU-rich sequences found in the 3'-UTR of TcSMUG transcripts were working alike in the ancient protozoan T. cruzi. Two constructs were generated in the pTEX Trypanosomatid vector (41), which has an ActD-sensitive promoter, to measure their half-lives. Both constructs were almost identical having the CAT gene flanked by the intergenic regions from a group L member, either intact (named clone Lc) or with a 50-bp deletion encompassing the AU-rich region shown in Fig. 3C (clone LcAU). This region of L group is more similar to the AREs described up to date (29) than the one in the S group. From the experiments described above we thought that 3'-UTR AU-rich regions might be responsible of the observed differences, functioning as a destabilizing element in trypomastigotes, but being protected in epimastigotes. However, no differences were found between the half-lives of the Lc and LcAU mRNA when compared in epimastigote stage under the same conditions than in Fig. 4 (not shown). So, AU-rich regions do not seem to be involved in the CHX induced destabilization of TcSMUG mRNA in epimastigotes.

We wanted to find out if this AU-rich region was acting at another developmental stage of the parasite. The optimum would be to assess this in cell-derived trypomastigotes. However, it is not possible to obtain enough transfected cell-derived trypomastigotes to perform Northern blots by directly transfecting them, due to the fact that they do not replicate, and to the low transfection efficiency achievable. On the other hand, to obtain cell-derived trypomastigotes from transfected epimastigotes would require undergoing the complete life cycle, extending times too much, and increasing the probabilities of selecting unexpected populations. So, we chose to compare the mRNA level of Lc and LcAU between epimastigotes and metacyclic trypomastigotes. The steady state mRNA level of the L group was about 6 times higher in epimastigotes than in metacyclic trypomastigotes (see Fig. 3A), similarly to the differences between epimastigotes and cell-derived trypomastigotes. In addition, metacyclic trypomastigotes can be rapidly differentiated in axenic culture from transfected epimastigotes, keeping the time between transfection and measures to a minimum. The same constructions, Lc and LcAU, were expressed in pRIBOTEX vector, a pTEX derivative with a rRNA promoter that allows higher expression levels and shorter times of selection (42). Constructs were transfected in epimastigotes from which metacyclic trypomastigotes were derived. The steady state level of CAT mRNA was measured by Northern blot and densitometry for Lc and LcAU in both stages (Fig. 7B). This value was only normalized to the transfection level using the ratio of CAT mRNA to neo mRNA to allow comparison among different populations (Fig. 7C). While CAT mRNA levels coming from both LcAU and Lc constructions were similar in epimastigotes (1.4 ratio), in metacyclic trypomastigotes the level of CAT mRNA coming from LcAU was 5.8-fold higher than that coming from Lc. Thus, the presence of the AU-rich sequence in the 3'-UTR of the CAT mRNA seems to be responsible for the 10-fold decrease in mRNA level observed at the steady state in metacyclic trypomastigotes with respect to the same construction in epimastigotes (Fig. 7C). Since both constructions used the same ribosomal promoter and have identical sequences, except for the presence of the 50-bp AU-rich region, this effect must occur at a post-transcriptional level and be mediated by this sequence. Thus, these results indicate a stage-specific destabilizing effect for this AU-rich sequence. The same experiment was done using the pTEX vector, in which the reporter is under an RNA polII promoter, and the same qualitative stage-specific effect was obtained. Again, LcAU and Lc showed fairly the same mRNA levels in the epimastigote stage (1.3 ratio) but the mRNA level of LcAU was augmented in metacyclic trypomastigotes increasing the ratio of LcAU to Lc mRNAs to 17 for this stage, even higher than that of 5.8 obtained with pRIBOTEX (results not shown). For all the parasite populations, the polyadenylation site used by the CAT mRNA was determined by sequencing the cDNA of the construction obtained by RT-PCR. In all cases, the same site was found to be used. So an effect in mRNA processing due to the AU-rich region deletion can be discarded. This further supports a direct effect of this region on mRNA stability.

View larger version (29K): [in this window] [in a new window]   Fig. 7.   AU-rich elements in 3'-UTR of TcSMUG affect mRNA stability and translation efficiency. A, a schematic representation of the constructions used is shown. PCR primers used for the constructions are indicated. The restriction sites used were SmaI (S), HindIII (H), BamHI (B), XhoI (X), and BamHI (B). UTR found in the original transcripts are indicated, along with the trans-splicing site (ag), polyadenylation site (pA), and polypyrimidin tract (pPy). B, Northern blots of transfected parasite populations are shown. The same filter was sequentially hybridized with CAT, neo, and 24S rRNA probes. C, CAT mRNA levels in epimastigote (E) and metacyclic trypomastigote (MT) for both constructions were quantified by densitometry using ImageQuant v3.2.2 software (Molecular Dynamics) and normalized to transfection level (signal from probe neo). Two independent experiments gave similar results. D, thin layer chromatography of the CAT activity assay of the same parasite populations used in the experiment shown in B (upper panel). The arrow indicates the origin. In the lower panel, a bar representation of the ratio of CAT activity (normalized to parasites used in the measures) versus CAT mRNA level (normalized to 24S rRNA) is shown. E, epimastigotes; MT, metacyclic trypomastigotes. Two independent experiments gave similar results.

An effect on translation efficiency was also demonstrated for ARE in the case of tumor necrosis factor  (31). We assessed if ARE was also having an effect on translation in T. cruzi. CAT activity was measured in each transfected population (Fig. 7D, upper panel) and normalized to the parasite number used. The ratio of CAT activity to CAT mRNA, normalized to the Northern blot loading by 24S rRNA was obtained. This ratio is indicative of the translation efficiency of each construction and thus can be compared among the different parasite stages (Fig. 7D, lower panel). The construction Lc showed a higher ratio of activity versus mRNA than LcAU both in epimastigotes and metacyclic trypomastigotes, indicating that the presence of the AU-rich sequence improves the translation efficiency. This effect, unlike that on mRNA stability, is not stage specific.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We describe here a novel family of mucin-type genes present in T. cruzi, named TcSMUG, which are distinct from the previously described TcMUC gene family (21). Even though TcMUC and TcSMUG family members share a general mucin-like structure, members of the novel TcSMUG family are much more conserved and homogeneous, with only two groups of very similar genes. Conversely, the TcMUC family is much larger (about 500 members), heterogeneous, and complex. Some sequence similarities exist at the DNA and protein level between both families with around 61 and 50% similarity values, depending on the members compared.

We have recently obtained evidence indicating that repeats containing members of TcMUC encode apomucins of molecules present in cell-derived trypomastigote stage.² We propose here that TcSMUG genes might encode the protein cores of mucin molecules present in epimastigote and metacyclic trypomastigote, known as 35/50-kDa antigens (54). Albeit indirect, there is evidence supporting our hypothesis. First, the structure of the products encoded by TcSMUG genes is mucin type, with Thr-rich central regions very similar to TcMUC repeats that are O-glycosylated in vivo.² Second, there is a concordance between the molecular mass of the TcSMUG mature product and that calculated for the apomucins of the epimastigote stage (1, 15), both around 7 kDa. Third, there is a correlation between the mRNA levels of TcSMUG transcripts and the stages where the 35/50-kDa mucins are expressed, at least for the S group. Amino acid composition of the predicted mature products of TcSMUG is similar to that reported for mucins purified from epimastigotes, although not identical (15). This is not surprising since a high diversity and heterogeneity was described for T. cruzi mucins (15) and experimental amino acid determinations were always done on heterogeneous material. Mucin-type molecules could be important in the insect vector stages, as interactions of the parasite with lectins from the insect have been described (55) and implicated in the differentiation process (56).

The family described here is composed of two groups, named L and S. Even if their N and C termini are almost identical, their central regions are similar in composition but clearly distinctly organized. While the L group had perfect repeats, the S group presented Thr runs with typically two NAPAK motives. The two groups of genes seem to have diverged from a common ancestor because of the high sequence homology, but then they have remained separated. Even their loci structure are different, the insertion of a SIRE sequence in the intergenic region of the L group is the major difference. Furthermore, they have different mRNA structures. While the 5'-UTRs are almost identical, the 3'-UTRs have three differential regions between L and S group (white boxes in Fig. 3C). The presence of SIRE in the 3'-UTR of the L group is interesting since some expression regulatory functions were reported for this element, positioned in the IR (57). The sequence present in the TcSMUG L group genes is somewhat degenerated with respect to SIRE, so further work must be done to know if this sequence has indeed an effect on the observed mRNA level differences between S and L groups. Differences in the 3'-UTR between L and S group could be responsible for the differential expression throughout the life cycle of the parasite. Several mechanisms have been proposed, involving 3'-UTRs in developmental regulation of several genes in Trypanosomatids, including RNA stability and translation modulation (24-26, 28). However, to our knowledge, only one mechanism was described, in T. brucei, involving a conserved stem-loop forming sequence present in 3'-UTR of procyclin transcripts and regulating its stability and translation (28).

It is shown in this work that regulation of mRNA abundance of TcSMUG genes occurs mainly at the post-transcriptional level in three of the four major developmental stages of T. cruzi. We first compared the mechanisms affecting steady state mRNA levels between epimastigotes and cell-derived trypomastigotes. While transcription initiation rates are comparable between these two stages, as estimated from run-on assays, the mRNA abundance is clearly higher in epimastigote, for both L and S mRNAs. TcSMUG mRNAs half-life is reduced when it is determined in the presence of the protein synthesis inhibitor cycloheximide. These mRNAs were not visibly affected by CHX treatment in the trypomastigote stage at the time points tested, maintaining a short half-life of less than 0.5 h as determined in the absence of CHX. In other words the instability observed in trypomastigotes is not dependent on translation. There are two simple hypotheses for explaining this behavior. One is that these mRNA were intrinsically short-lived ones and a mechanism to protect them from degradation, mediated by some labile factor(s), could be acting in epimastigotes but not in cell-derived trypomastigotes. Alternatively, the mRNA could be protected while actively translated and then becoming short lived in the absence of translation. Neither of these two hypotheses would involve the AU-rich regions (see below).

A clue for a functional destabilizing mechanism came from comparing steady state levels of reporter constructions, bearing or not the AU-rich sequence present in the 3'-UTR of the L group transcripts. The L group IR was chosen because its AU-rich sequence has features more similar to those described in AREs present in mammalian cells (30). The comparison was done between epimastigote and metacyclic trypomastigote stages for both constructions. It was demonstrated that the AU-rich region was a mRNA destabilizing element since both constructions, transcribed from the same promoter and differing only in the presence of this 50-bp region, clearly differed in mRNA abundance in the metacyclic trypomastigote stage (Fig. 7C). On the other hand, transfection experiments showed no effect in epimastigote stage for this sequence. CAT mRNA with or without the AU-rich region showed the same half-life. This suggests that this AU-rich region is functioning in a stage-specific manner to down-regulate the L group mRNA levels when the parasite differentiates from epimastigote to metacyclic trypomastigote. Even when the transcription rate was not measured in metacyclic trypomastigote, the effect seen on the reporter transcript stability is strong enough to account for the differences observed in the abundance of the L groups mRNA between these two stages.

It was shown that ARE-directed mRNA decay in mammalian cells is linked to cell transformation, cell growth and differentiation, cell adhesion, and in the immune response regulation (29). In the 3'-UTR of the L group, the motif UUAUUUAU(U/A) is repeated twice, such as the cases of ARE decay-regulated interleukin 3 (58) and interferon  mRNAs (59). It was reported that one copy of the 9-nucleotide sequence UUAUUUAUU is weakly destabilizing and two copies (as is the case in the L group UTR) provides a strong destabilizing effect (60, 61). The smaller AU₃A motif is repeated 4 times in the AU-rich region of the L group 3'-UTR and some groups reported that this pentanucleotide rather than the nonamer is the minimal functional motif of AREs (62). Transcripts of the S group carry two AU₄A motives within the AU-rich region in their 3'-UTR that is different from the typical AREs. However, non-AU₃A containing ARE can also mediate mRNA decay (29). Furthermore, it was recently reported that the (AUUUU)₃A motif bound to an elav-like protein (63). The elav-like protein family is formed by RNA-binding proteins recognizing AU-rich elements in 3'-UTR and stabilizing mRNAs (64). It is important to mention that, even when much work have been done on AREs in mammalian cells, little is known about its conservation during the evolution. The lower organism with functional AREs characterized was an echinoderm (33). In the same work, ARE-like sequences were reported in Plasmodium falciparum and Dyctiostelium dyscoideum but simply by homology searching in data bases. We report here for the first time, AU-rich motives that are functional in regulating mRNA stability, associated with 3'-UTR of transcripts that are mainly regulated by destabilizing mechanisms, in a protozoan organism, presenting this finding an evolutionary interest. This mechanism could be the ancestor of the one in metazoans, and raises the possibility that it could also be functional in Plasmodium and Dyctiostelium. We therefore propose that these regions are indeed ARE since they are clearly functionally and structurally related to the ones described in higher eukaryotes.

To have an idea of the possible importance of this mechanism in T. cruzi, a GeneBank nucleotide data base search was done using the L group's AU-rich region as query, specifically on T. cruzi existing sequences. This search yielded several dozens hits. In many of them, where the complete mRNA is known, 1 to 5 class I and II ARE motives (29) were found in 3'-UTR of mRNAs. Selected examples of characterized molecules are shown in Table I. When the T. cruzi EST data base was searched, several hundred hits were obtained but as these are not characterized entries it is difficult to assign them to 3'-UTR regions. However, it is clear that these kind of sequences are thoroughly present in T. cruzi mRNAs, often in the 3'-UTR. Post-transcriptional mechanisms, like ARE-dependent ones, with a very fast and irreversible action mode, are well suited for quick responses triggered by sudden environmental changes like those that suffer a parasite when changing host or in the case of T. cruzi when it invades or is released from cells. Given that ARE-like sequences are functional in T. cruzi, as suggested by our findings, it should be possible to find proteins implicated in this mechanism similar to those described in metazoans. We searched the T. cruzi EST data base and found an EST with high probabilities to be an elav-like protein. Several other RNA-binding proteins exist in T. cruzi, where post-transcriptional regulation seems to be the rule for genetic expression. The fact that both cis- and trans-acting elements of ARE-mediated mRNA decay were present in T. cruzi open an interesting research field for future work.

	ACKNOWLEDGEMENTS

We thank Dr. J. M. Kelly for providing pTEX vector, Dr. R. Hernández for pRIBOTEX vector, and Dr. E. Bontempi for the 24S rRNA clone. We thank Liliana Sferco and Berta Franke de Cazzulo for parasite cultures, Ramiro Verdún and the sequencing laboratory technicians for automated sequence help, and Carolina Levy for help with some of the subcloning from cosmids. We also thank Dr. J. J. Cazzulo for critical reading of the manuscript.

	FOOTNOTES

^* This work was supported by grants from the UNDP/World Bank/WHO Special Program for Research and Training in Tropical Diseases, the Swedish Agency for Research Cooperation with Developing Countries, the Consejo Nacional de Investigaciones Cientificas y Técnicas (CONICET), Argentina, the Agencia Nacional de Promoción Científica y Tecnología, Argentina, and Fundación Antorchas, Argentina.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EMBL Data Bank with accession number(s) AF203085 to AF203105.

Contributed equally to the results of this work.

^§ Fellow of the Consejo Nacional de Investigaciones Cientificas y Técnicas, Argentina.

^¶ Researcher for the Consejo Nacional de Investigaciones Cientificas y Técnicas, Argentina.

Supported in part by an International Research Scholar Grant from the Howard Hughes Medical Institute. To whom correspondence should be addressed: Instituto de Investigaciones Biotecnológicas-Universidad Nacional de General San Martín, Av. Gral. Paz s/n, INTI, Edificio 24, 1650-San Martín, Pcia. de Buenos Aires, Argentina. Tel.: 54-11-4752-9639; Fax: 54-11-4752-0021; E-mail: cfrasch@iib.unsam.edu.ar.

² G. D. Pollevick, J. M. Di Noia, M. L. Salto, C. Lima, S. Leguizamón, R. M. de Lederleremer, and A. C. C. Frasch, manuscript in preparation.

	ABBREVIATIONS

The abbreviations used are: GPI, glycophosphatidylinositol; UTR, untranslated region; EST, expressed sequence tag; ORF, open reading frame; CHX, cycloheximide; ActD, actinomycin D; CAT, chloramphenicol acetyltransferase; IR, intergenic regions: neo, neomycin phosphotransferase; ARE, AU-rich elements; SIRE, short interspersed repeat element; RT-PCR, reverse transcriptase-polymerase chain reaction; bp, base pair(s).

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES