Regulation of genes encoding the major surface protease of Leishmania chagasi via mRNA stability

Molecular & Biochemical Parasitology 142 (2005) 88–97

Regulation of genes encoding the major surface protease of Leishmania chagasi via mRNA stability Jay E. Purdy a,∗ , John E. Donelson b , Mary E. Wilson a,c,d a

Department of Internal Medicine, University of Iowa, Iowa City, IA 52242, USA b Department of Biochemistry, University of Iowa, Iowa City, IA 52242, USA c Department of Microbiology, University of Iowa, Iowa City, IA 52242, USA d The Veterans’ Affairs Medical Center, Iowa City, IA 52242, USA

Received 4 October 2004; received in revised form 16 March 2005; accepted 23 March 2005 Available online 9 April 2005

Abstract The intercoding regions between many Leishmania sp. genes regulate their mRNA expression. The MSPL mRNA, encoding a subclass of the major surface protease (MSP) of Leishmania chagasi, increases in abundance, when protein synthesis is arrested, while ␣-tubulin (α-TUB) mRNA and most other mRNAs do not. We found that the intercoding region between MSPL-coding regions, when cloned downstream of the ␤-galactosidase reporter gene (β-GAL), caused β-GAL mRNA to increase 8- to 10-fold after inhibiting protein synthesis with cycloheximide. Stable L. chagasi transfectants containing hybrid MSPL/α-TUB intercoding regions cloned downstream of β-GAL were made. The α-TUB intercoding region induced high-level baseline β-GAL mRNA that increased only 1.3-fold after incubation with cycloheximide. In contrast, the MSPL intercoding region, as well as constructs containing nucleotides 303–505 from the MSPL 3 UTR, caused steady-state β-GAL mRNA levels in the absence of cycloheximide that were approximately 10% of α-TUB constructs. These levels increased between 4.4- and 13.2-fold after cycloheximide was added. Constructs containing half of this region (303–394 or 395–505) produced intermediate levels of β-GAL mRNA and intermediate levels of cycloheximide induction. The kinetics of cycloheximide induction of β-GAL mRNA was similar with region 303–505 constructs as with constructs bearing the entire endogenous MSPL intercoding region. Furthermore, region 303–505 increased reporter mRNA abundance after cycloheximide by increasing mRNA half-life. Hence, we have identified a 202-nucleotide region within the MSPL 3 UTR that is in part responsible for cycloheximide induction. We hypothesize that this region may interact with labile regulatory protein factor(s). © 2005 Elsevier B.V. All rights reserved. Keywords: Leishmania; Gene regulation; Major surface protease; mRNA stability; Cycloheximide; Intercoding region; GP63

1. Introduction Regulation of eukaryotic genes at the level of mRNA stability was first suggested in 1976, when Malt and co-worker reported that the half-life of certain murine mRNAs were prolonged, when compared to the majority of murine mRNAs [1]. Since then, many important eukaryotic genes regulated Abbreviations: MSPL, major surface protease logarithmic subclass; βGAL, ␤-galactosidase gene; α-TUB, ␣-tubulin; PFR2, paraflagellar rod gene 2 ∗ Corresponding author. Tel.: +1 319 353 7229; fax: +1 319 384 7208. E-mail address: [email protected] (J.E. Purdy). 0166-6851/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.molbiopara.2005.03.010

at the level of mRNA stability have been identified including, but not limited to, genes encoding GM–CSF [2], c-fos [3], interleukin-8 [4], and interleukin-3 [5]. Among the pathogenic Trypanosomatid protozoan parasites, Leishmania sp. cause visceral, cutaneous, and other forms of leishmaniasis, whereas Trypanosoma sp. cause African sleeping sickness or Latin American Chagas’ disease. These evolutionarily related organisms have two- or three-stage life cycles involving an insect vector and mammalian host. Within their insect vector, the promastigote stage of Leishmania sp. develops as an extracellular, long, flagellated, and motile parasite. During a blood meal, the insect vector transfers infectious promastigotes into a mammalian

J.E. Purdy et al. / Molecular & Biochemical Parasitology 142 (2005) 88–97

host, where they are rapidly internalized by macrophages. Over the next 24–48 h, the parasite loses its flagellum, becomes spherical, and transforms to the obligate intracellular amastigote stage. Accompanying this change in morphology, we postulate that some genes required for survival within the macrophage are up-regulated, while others necessary for survival within the insect vector are down-regulated. The majority of genes in Trypanosomatids are transcribed as part of large polycystronic pre-mRNA units [6]. Few or no promoters for genes transcribed by RNA polymerase II have been found in members of this family of protozoa, and RNA abundance is often not regulated by the initiation and/or rate of transcription [7]. Hence, gene regulation largely occurs post-transcriptionally. Documented mechanisms through which Trypanosomatids regulate gene expression include changes in mRNA stability [7–14], the rate of precursor RNA processing [15], and the rate of translation [10,11,16–18]. Trypanosomatid genes whose mRNAs are regulated at the level of stability include genes for HSP83 [12], several glucose transporters [13], procyclin [11,19], mucin [10], surface antigen FL-160 [9], paraflagellar rod PFR2 [14], and GP63 [also called major surface protease (MSP)] [7,8]. However, the molecular mechanisms employed by Leishmania sp. to alter mRNA half-life are less well understood. Mishra et al. identified a sequence within the 3 UTR of PFR2C (AUGUAuAGUu) of Trypanosoma brucei that is responsible for mRNA stability [14]. Functional AU- elements have also been identified in Trypanosoma cruzi 3 UTRs [10] but it is unclear if such regulation occurs in Leishmania sp. Labile protein factors may play a role in regulation of some trypanosomatid genes, since procyclin/PARP mRNA of Trypanosoma brucei [19] and MSPL mRNA of Leishmania chagasi [20] increase with the arrest of protein synthesis. The increase in MSPL mRNA is not due to polysome stabilization, since chemicals that block different steps of protein synthesis (cycloheximide, pactamycin, puromycin) had a similar effect [20]. While the abundance of procyclin/PARP mRNA increases due to changes in both RNA processing and mRNA half-life [11,15,21], MSPL mRNA abundance is not associated with a change in the rate of transcription [20]. Rather, cycloheximide-induced changes in MSPL mRNA abundance is secondary to changes in mRNA half-life alone [20]. The mechanism responsible for the half-life regulation remains unknown. L. chagasi causes visceral leishmaniasis in the New World. The MSP genes of L. chagasi are arranged in a head-to-tail tandem array [22]. Three classes of differentially expressed MSP genes (MSPL, MSPS, MSPC) differ primarily by unique sequences in their 3 UTRs [20]. Although all are transcribed at a constant rate, the mRNA of the MSPL subclass is only expressed in vitro during logarithmic phase growth of L. chagasi promastigotes [23]. Furthermore, inhibition of protein synthesis increases the abundance of MSPL mRNA but not mRNAs encoding MSPS, MSPC, ␣-tubulin (␣-TUB), or ATPase [7,20,24]. The decrease in MSPL mRNA during in

89

vitro growth from logarithmic to stationary phase and the increase in mRNA abundance after protein synthesis inhibition are due to a corresponding changes in MSPL mRNA half-life [7,20]. The 5 UTRs and coding regions of MSPL and MSPS mRNAs are very similar, suggesting that their half-lives are regulated by unique sequences within their 3 UTRs [23]. The 1258 bp intercoding region between the tandemly repeated L. chagasi MSPL-coding regions is comprised of a 3 UTR of 750-nucleotides, an intergenic region of 357-nucleotides, and a 151 bp 5 UTR of the downstream MSPL gene. During this study, we used sequence replacements to identify a 202-nucleotide region in the 3 UTR of the MSPL genes that destabilizes MSPL mRNA at baseline. This destabilization decreases with the inhibition of protein synthesis resulting in an increase in MSPL mRNA (and β-GAL reporter gene) abundance. We hypothesize that this region interacts with labile, sequence-specific regulatory protein factor(s).

2. Materials and methods 2.1. L. chagasi culture L. chagasi (strain MHOM/BR/00/1669) parasites were originally isolated from a Brazilian patient with visceral leishmaniasis. To prevent attenuation, cultures are serially passaged through golden hamsters [25]. Promastigotes were cultured at 26◦ in hemoflagellate-modified minimal essential medium (HOMEM) containing heat inactivated fetal calf serum [25]. L. chagasi promastigotes were used in logarithmic stage growth, defined as a parasite concentration between 1 × 107 and 3 × 107 /ml, for electroporation, RNA isolations with and without cycloheximide, and DNA isolations [26]. Transfected L. chagasi were maintained in 100 ␮g G418/ml. Protein synthesis was arrested with 5 ␮g cycloheximide/ml (Sigma, St. Louis, MO) for 4 h [20]. RNA synthesis was inhibited by 10 ␮g actinomycin D/ml [7]. 2.2. Total DNA isolation and Southern blot analysis Genomic DNA was harvested from 18 ml of late logarithmic phase L. chagasi by incubating in 0.1% sarkosyl with 10 ␮g RNase/ml, and 0.1 mg proteinase K/ml for 30 min at 37 ◦ C. DNA was extracted with phenol/chloroform/isoamyl alcohol and precipitated with ethanol. DNA was digested to completion with ScaI, which cleaves each of the reporter plasmids only once. Southern blotting was performed using standard techniques [27]. 2.3. Creation and naming of β-GAL-containing reporter constructs bearing different intercoding regions Constructs (Fig. 1) were created via splice overlap extension (SOE) PCR [28] using primers shown in Table 1.

90

J.E. Purdy et al. / Molecular & Biochemical Parasitology 142 (2005) 88–97

Fig. 1. Increase in mRNA abundance after protein synthesis inhibition. (A) Schematic representation of reporter constructs permanently transfected in L. chagasi and maintained as episomes: gray box ( ), β-GAL; dotted line (· · ·), α-TUB intercoding sequence; black box (
), segments of the MSPL intercoding region conserved with MSPS intercoding region; crossed-hatched boxes ( ), segments of the 3 UTR in the MSPL intercoding region divergent from MSPS intercoding region; open box (), portion of the MSPL intercoding region between the most 3 -poly(A) site and the start codon of the downstream gene; upward arrows, sites of endogenous α-TUB poly(A) addition; downward arrows, sites of endogenous MSPL poly(A) addition. The vertical rectangle depicts region 303–505 described in the text. (B) Total RNA was harvested from transfected L. chagasi that had been incubated for 4 h without or with cycloheximide. Northern blots were sequentially hybridized with the β-GAL-coding sequence to determine the abundance of reporter mRNA, and with the α-TUB-coding sequence for a loading control. The intensity of autoradiograph bands was determined via densitometry and calculated as a ratio of β-GAL to α-TUB to correct for loading variation. The average intensity of reporter mRNA without cycloheximide is shown in gray and with cycloheximide is shown in black. (C) The ratio of intensity ¯ is shown along with its standard error (S.E.) and the number of independent Northern blots (N) used to obtain with cycloheximide/without cycloheximide (X) these data. Each N represents an independent RNA preparation from two to four independent clones.

Constructs ␣/(673)/␣, ␣/(583)/␣, and ␣/(235)/␣ required two rounds of SOE PCR to join three fragments. Template DNAs for PCR amplification were plasmids containing the α-TUB intercoding region [24] and the MSPL intercoding region [23]. For the purposes of this work, we define “intercoding region” as the sequence between the stop codon of the upstream gene and the methionine start codon of the downstream gene in a tandem cluster. After the final round of PCR amplification, the total reaction mix was resolved on an agarose gel and the appropriately sized fragment excised and purified. Flanking primers (P1, P2, and P3; Table 1) incorporated an XbaI site to allow digestion and insertion into the pX-β-GAL shuttle vector [29] downstream of βGAL (vector kindly provided by Steve Beverley). The entire cloned region was sequenced at least twice at the University of Iowa DNA Facility using an Applied Biosystems Model 373A stretch fluorescent automated sequencer to ensure the proper sequence prior to being used for transfection. The name of each plasmid reflects its composition. Plasmids containing a hybrid intercoding region whose 5 -most end comes from the MSPL intercoding region begin with the letter “L” (for MSPL). Plasmids containing a hybrid intercoding region whose 5 -most end or 3 -most end is α-TUB in origin begin or end with “␣.” The number in parenthesis indicates the number of base pairs (bp) of the MSPL intercoding region included in the construct. For example, in plasmid L(394)/␣ (see Fig. 1), β-GAL is followed by a fusion product between the first 394 bp after the endogenous MSPL stop codon and the α-TUB intercoding region. The

one exception to this number definition in the plasmid name is plasmid ␣(958) in which 958 represents the entire 958 bp of the endogenous α-TUB intercoding region. The length of α-TUB intercoding region included in other constructs was: L(394)/␣, 660 bp; L(505)/␣, 582 bp; L(625)/␣, 490 bp; L(745)/␣, 389 bp; L(870)/␣, 302 bp; L(979)/␣, 212 bp; ␣/(673)/␣, 298 bp 5 of the MSPL sequence and 212 bp 3 ; ␣/(583)/␣, 298 bp 5 and 212 bp 3 ; ␣/(235)/␣, 552 bp 5 and 212 bp 3 . L. chagasi promastigotes were electroporated with the plasmid reporter constructs (Fig. 1), and individual colonies were selected on agar plates containing medium 199 and 40 ␮g G418/ml and expanded in liquid culture with 100 ␮g G418/ml [30]. 2.4. Northern blots and RNA half-life analysis Total RNA was purified [31] from L. chagasi cultured under conditions described below. RNA was analyzed by Northern blotting using standard techniques [27]. Radiolabeled fragments used as probes included: (i) a 1.4 kb EcoRI fragment including the α-TUB-coding region isolated by PCR amplification of L. chagasi genomic DNA [26]; (ii) a 3.2 kb BamHI fragment of pX-β-GAL bearing β-GAL [29]; (iii) a 0.4 kb EcoRI/PstI fragment of an L. chagasi 18S rRNA gene [32]; (iv) a 0.3 kb PCR fragment of a unique region in the MSPL 3 UTR [22]. Bands detected via autoradiography were quantified using an Alpha Innotech Co. (San Leandro, CA) AlphaImager 3300.

J.E. Purdy et al. / Molecular & Biochemical Parasitology 142 (2005) 88–97 Table 1 Oligonucleotide primers used for SOE PCR construction of reporter plasmids shown in Fig. 1 (sequence 5 –3 ) ␣(958) P1 GC TCTAGA GGTACACTCGTGCCGCGC P2 GC TCTAGA GGCTGAAAAAGAAGAAAG L(1258) P3 GC TCTAGA CGGTGGATAGGACGGGTG L(979)/␣ [P2 and P3] P4 GGCACATGGCGGCAGCAG TGTTGGCGTGCTCACGGGT P5 ACCCGTGAGCACGCCAACA CTGCTGCCGCCATGTGCC L(870)/␣ [P2 and P3] P6 CGTCTCTTCAGGAGCTGC GGGTCGCTCGGGGAGGCG P7 CGCCTCCCCGAGCGACCC GCAGCTCCTGAAGAGACG L(745)/␣ [P2 and P3] P8 TTTTCACGTCGGAGTGGT GTCGTGCCGACGACGTGC P9 GCACGTCGTCGGCACGAC ACCACTCCGACGTGAAAA L(625)/␣ [P2 and P3] P10 CAGGCTCGCAGGTGCGTG CGTCCGCGCGCAGAGGCC P11 GGCCTCTGCGCGCGGACG CACGCACCTGCGAGCCTG L(505)/␣ [P2 and P3] P12 CTTTCTCTTCATCTATCA TGACCGCGCCGGAGCGCC P13 GGCGCTCCGGCGCGGTCA TGATAGATGAAGAGAAAG L(394)/␣ [P2 and P3] P14 CGCGCGTGTGCATGCGCG CGCACGCACCCGCGGATG P15 CATCCGCGGGTGCGTGCG CGCGCATGCACACGCGCG ␣/(235)/␣ [P1, P2, P4 and P5] P16 GGGGGAGGTACCTCGGTG GCGCGCCGACATCACCCG P17 GCACACGCGCGCGCGCTG CACCGAGGTACCTCCCCC ␣/(583)/␣ [P1, P2, P4 and P5] P18 GAGAAGGCAGACACACCG GCAGACGCACGCACGAGG P19 CCTCGTGCGTGCGTCTGC CGGTGTGTCTGCCTTCTC ␣/(673)/␣ [P1, P2, P4, P5 and P18] P20 CCTCGTGCGTGCGTCTGC TTCACTTCCTTATTTGTC Primers used to make each reporter plasmid are listed below the named plasmid with primers used, but previously described, listed in brackets to the right of plasmid name. Bold type emphasizes restriction sites and underlined sequences indicate the portion of each primer complementary to or homologous to MSPL sequences (those portions not underlined are complementary to or homologous to α-TUB sequences).

3. Results 3.1. Cycloheximide induction of β-GAL mRNA regulated by MSPL/α-TUB hybrid intercoding regions L. chagasi were electroporated with the reporter constructs shown in Fig. 1A. The parent plasmid, pX-B-GAL, allowed constructs to be maintained as stable episomes as long as selective antibiotic pressure is maintained. Total RNA was harvested from transfected logarithmic phase L. chagasi promastigotes without cycloheximide or with cycloheximide for 4 h. We previously showed that the increase in MSPL mRNA

91

abundance after the addition of cycloheximide is detected 1 h after protein synthesis inhibition, increases to 78% of maximum at 4 h, and peaks at 20 h [20]. Thus, the majority of induction was expected at the 4 h time point. Fig. 1B shows the average ± S.E. abundance of β-GAL mRNA from each transfectant incubated without (gray bars) or with (black bars) cycloheximide after controlling for loading and differences in length of exposure. Fig. 1C reports the mean ± S.E. ratio of β-GAL mRNA in transfectants after cycloheximide/without cycloheximide. Thus, a ratio of 2.0 indicates that mRNA becomes two-fold more abundant after protein synthesis inhibition. Plasmid constructs ␣(958) and ␣/(235)/␣, which contained wholly or primarily α-TUB intergenic sequences downstream of the β-GAL-coding region, resulted in a very high baseline level of β-GAL mRNA (gray bars) that increased to only a small extent after protein synthesis inhibition (black bars, 1.2- and 1.4-fold, respectively). Plasmid ␣(958) contained only α-TUB intercoding sequences and caused β-GAL mRNA to behave much like endogenous α-TUB mRNA [24]. We showed previously that the abundance, half-life, and hence rate of transcription of β-GAL mRNA transcribed from the plasmid ␣(958) is similar to endogenous α-TUB mRNA [24]. The steady-state abundance of these mRNAs likely represents the level of mRNA resulting from transcription of the genomic α-TUB cluster and without cycloheximide-inhibitable degradation. Hence, no significant induction was noted after cycloheximide addition, since no protein synthesis-dependent regulation of mRNA abundance occurred under these conditions. In contrast, plasmids L(505)/␣, L(625)/␣, L(745)/␣, L(870)/␣, L(979)/␣, L(1258), and ␣/(673)/␣, all produced low levels of β-GAL mRNA without cycloheximide, which was induced 4.4- to 13.2-fold after protein synthesis inhibition. Plasmid L(1258) contains the entire 1258 bp MSPL intercoding region downstream of β-GAL. This plasmid produced a low level of β-GAL mRNA at baseline (i.e. without cycloheximide), which was induced 8.3-fold by cycloheximide. However, even with induction, the amount of β-GAL mRNA resulting from L(1258) remained below the baseline level of β-GAL mRNA observed with ␣(958) and ␣/(235)/␣. The difference in reporter mRNA abundance from plasmids ␣(958) and L(1258) appears to be due to specific destabilization of β-GAL mRNA from L(1258), an effect that is reversed, at least in part, by protein synthesis inhibition. The baseline destabilization and cycloheximide-induced increase of β-GAL mRNA was retained with plasmids lacking: (i) the 279 bp intergenic region downstream of the 3 UTR; (ii) the 3 end of the 3 UTR; (iii) the 5 end of the 3 UTR including most of the MSPL/MSPS common region (indicated by black boxes in Fig. 1A). Thus, β-GAL mRNAs from plasmids containing a region within the MSPL 3 UTR between 303 and 505 nucleotides downstream of the MSPL stop codon (indicated by the tall rectangle in Fig. 1A) behave similarly to endogenous MSPL mRNA including low baseline mRNA levels and induction with protein synthesis inhibition.

92

J.E. Purdy et al. / Molecular & Biochemical Parasitology 142 (2005) 88–97

It must be noted that while baseline β-GAL mRNA levels from plasmid L(505)/␣ was low, cycloheximide increased mRNA abundance only 4.4-fold. Thus, we conclude that the element responsible for baseline mRNA destabilization lies between nucleotides 303 and 505, and that an element in this same region contributes partially to cycloheximide induction. Comparison of constructs L/(505)␣ and L(745)/␣ suggests that an additional element(s) contributing to the cycloheximide-dependent mRNA induction lies between nucleotides 505 and 745. Two plasmids, L(394)/␣ and ␣/(583)/␣, resulted in intermediate basal levels of β-GAL mRNA and intermediate levels of induction after cycloheximide (2.2- and 3.4-fold, respectively). Each of these constructs contains a different half of the apparent critical region 303–505 nucleotides downstream of the MSPL stop codon required for full basal repression and partial cycloheximide induction. The critical region was retained, and full gene repression with partial cycloheximide induction occurred in constructs L(505)/␣ and ␣/(673)/␣, which had flanking sequences replaced by α-TUB sequences on their 5 and 3 ends, respectively. It should be noted that there was no correlation between basal mRNA levels or cycloheximide-induced levels of mRNA and the presence or absence of endogenous poly(A) sites. For example, L(745)␣ contains no endogenous MSPL or α-TUB poly(A) sites but was induced by cycloheximide 11.8-fold. On the other hand, L(870)/␣ contains two poly(A) sites including the primary and most proximal MSPL poly(A) site, and was induced 10.4-fold. 3.2. Lack of effect of cycloheximide on plasmid abundance While we postulate that the cycloheximide induced changes in mRNA abundance resulting from region 303 to 505 was due to changes in mRNA stability, we could not rule out the possibility that cycloheximide differentially affected the abundance of the reporter plasmid constructs. To rule out this possibility, total DNA was harvested from L. chagasi promastigotes containing the ␣(958), ␣/(583)/␣, ␣/(673)/␣, and L(1258) reporter constructs. DNA was digested with ScaI, for which each of the reporter plasmids has a single restriction site. Southern blots were hybridized sequentially with a β-GAL-specific probe to assay for plasmid abundance and with an α-TUB-coding region probe hybridizing to endogenous α-TUB genes to assay for loading control (Fig. 2). The intensities of the bands were measured via autoradiography and the ratio of plasmid β-GAL to genomic α-TUB DNA calculated to correct for loading. The ratio of plasmid abundance in samples with versus without cycloheximide is indicated in the row labeled “change” in Fig. 2. No significant changes in plasmid abundance were noted after the addition of cycloheximide to any of the transfectants (range 0.95- to 1.08-fold increase after cycloheximide, Fig. 2). Thus, the changes in reporter β-GAL mRNA abundance after the protein synthesis inhibition were not due to changes in plasmid abundance. In

addition, the sizes of the ScaI fragment detected on β-GAL blots confirmed that the plasmids were maintained by the cells as an episome, rather than integrating into the genome. It was also possible that the hybrid intercoding regions could influence the baseline plasmid copy number. We, therefore, confirmed that the average copy number of each plasmid was similar. For each DNA preparation, the ratio of β-GAL plasmid DNA to endogenous α-TUB DNA is indicated in the row labeled “Abund” in Fig. 2. This value represents the plasmid copy number per unit of chromosomal DNA. The intensity of the ␣(958) band was arbitrarily assigned the value of 1.0 with the other plasmids ranging from 0.75- to 1.1-fold greater. Thus, little difference was noted in the amount of plasmid per unit of total DNA, suggesting that the copy number per cell of ␣(958), ␣/(583)/␣, ␣/(673)/␣, and L(1258) was similar. 3.3. Kinetics of cycloheximide-induced changes in β-GAL mRNA We have shown that β-GAL mRNA from ␣(958) bearing the α-TUB intercoding region behaves nearly identically to endogenous α-TUB mRNA in its response to cycloheximide and in different stages of growth [24]. Similarly, the change in β-GAL mRNA levels in cells bearing plasmids ␣/(673)/␣ and L(1258) 4 h after addition of cycloheximide is similar to the change in MSPL abundance after protein synthesis inhibition. Therefore, we postulate that these plasmids contain elements responsible for regulation of MSPL mRNA abundance. If the β-GAL mRNAs from plasmids ␣/(673)/␣ and L(1258) are regulated by a mechanism common to endogenous MSPL, we would predict that the kinetics of cycloheximide induction of β-GAL mRNA in transfectants containing these constructs would be similar to that of endogenous MSPL mRNA. Furthermore, β-GAL mRNA from plasmid ␣/(583)/␣ has a baseline abundance and induction levels between those of L(1258) and ␣(958) (plasmids containing the MSPL intercoding and α-TUB intercoding region, respectively). Therefore, we would expect the kinetics of the cycloheximide induction effect to be intermediate as well. Total RNA was harvested from L. chagasi promastigotes transfected with ␣(958), ␣/(583)/␣, ␣/(673)/␣, and L(1258) between 1 and 24 h after the addition of cycloheximide. Northern blots were hybridized sequentially with a [32 P]labeled β-GAL DNA probe to assay for mRNA abundance and [32 P]-labeled α-TUB DNA as a loading control (Fig. 3). Band intensity was measured by autoradiography, and expressed as the mean fold-change from time zero amounts. As predicted, the kinetics of cycloheximide induction of β-GAL mRNA in transfectants containing the L(1258) plasmid, which contains the entire MSPL intercoding region, were similar to that of endogenous MSPL mRNA (Fig. 3). There was no induction of β-GAL mRNA in transfectants containing the α-TUB intercoding region construct ␣(958) similar to endogenous α-TUB mRNA [24]. However, mRNA from plasmid ␣/(673)/␣ changed with kinetics similar to MSPL

J.E. Purdy et al. / Molecular & Biochemical Parasitology 142 (2005) 88–97

93

Fig. 2. The relative plasmid copy number before and after cycloheximide. Total DNA from L. chagasi permanently transfected with ␣(958), ␣/(583)/␣, ␣/(673)/␣, or L(1258) either without cycloheximide (−), or 4 h after the addition of cycloheximide (+). DNA was digested to completion with ScaI prior to resolution. Southern blots were hybridized sequentially with the β-GAL-coding sequence to determine the abundance of reporter mRNA, and the α-TUB-coding sequence as a loading control. Autoradiograph bands were quantified via densitometry and adjusted for loading. The ratio of the abundance with cycloheximide divided by the abundance without cycloheximide (change) ± S.D. is shown in the bottom panel. The relative abundance of plasmid levels compared to the other plasmids ± S.D. is shown (Abund) along with n, the number of independent DNA preparations compared.

and L(1258), and had similar levels of baseline mRNA suppression (Figs. 1 and 3). These data suggest that all three constructs contain the elements required for baseline suppression and cycloheximide induced de-repression. In contrast, the cycloheximide-induced rate of change of β-GAL mRNA in the ␣/(583)/␣ transfectant was mid-way between that of β-GAL mRNA in L(1258) and ␣(958). One should recall that the only difference between ␣/(583)/␣ and ␣/(673)/␣ is an additional 90 bp of the MSPL intercoding region (the first half of the region between 303 and 505) in the latter construct. 3.4. Half-life of β-GAL mRNA We previously showed data suggesting that the changes in abundance of endogenous MSPL mRNA are caused by changes in mRNA half-life [7,20,24]. Hence, we examined whether changes in β-GAL mRNA abundance from ␣/(673)/␣ and to a lesser extent from ␣/(583)/␣ were also due to changes in mRNA half-life.

To investigate mRNA half-life, transfectants containing ␣/(673)/␣ or ␣/(583)/␣ were incubated in either growth medium alone, or medium containing cycloheximide. Four hours later, actinomycin D was added. RNA was harvested from transfectants between 1 and 12 h after nascent RNA synthesis was arrested with actinomycin D. Northern blots of total RNA were hybridized sequentially with a β-GAL probe to assess mRNA abundance, followed by α-TUB as a loading control (Fig. 4). The half-life of β-GAL mRNA in transfectants containing ␣/(673)/␣ was 1.8 h and increased to 11.1 h after cycloheximide addition (Table 2 and Fig. 4). These data are comparable with the half-life of endogenous MSPL mRNA (1.6 and 10.7 h, respectively (data not shown), which explains the majority of the increase in mRNA abundance noted in Fig. 1. Similarly, the half-life of β-GAL mRNA from plasmid ␣/(583)/␣ was 4.6 h without cycloheximide (Fig. 4). This value is intermediate between the 1.8 h half-life of β-GAL mRNA in ␣/(673)/␣ and L(1258) transfectants and the

94

J.E. Purdy et al. / Molecular & Biochemical Parasitology 142 (2005) 88–97 Table 2 Summary of the half-life of mRNA either with or without cycloheximide Endogenous ␣-TUB β-GAL from ␣(958) β-GAL from ␣/(583)/␣ β-GAL from ␣/(673)/␣ Endogenous MSPL

Without Chx (h)

With Chx (h)

Fold-change

7.3 7.7 4.6 1.8 1.6

12.0 13.5a 14.4a 11.1 10.7a

1.6 1.8 [24] 3.1 6.2 6.7 [7]

The estimated fold-increase in mRNA half-life is reported on the right. a Extrapolated value.

7.7 h half-life in ␣(958) transfectants [24]. Thus, it is not surprising that the baseline abundance of β-GAL mRNA from plasmid ␣/(583)/␣ was intermediate compared to transfectants containing L(1258) and those containing ␣(958) (Table 2). Furthermore, when the ␣/(583)/␣ data in Fig. 4 are extrapolated, the half-life of reporter mRNA from ␣/(583)/␣ appears to increase to about 14.4 h after inhibition of protein synthesis, similar to that of endogenous α-TUB mRNA(Table 2). This 3.2-fold cycloheximide-induced halflife increase is greater than the 1.7-fold half-life increase noted from transfectants containing only α-TUB sequences [␣(958)] [24]. Furthermore, this increase in half-life explains why ␤-GAL mRNA from ␣/(583)/␣ increased 3.4-fold in abundance (Fig. 1) with cycloheximide, while mRNA from ␣(958) increased only 1.2-fold. 3.5. Sequence analysis of the 303–505 region of the MSPL intercoding region

Fig. 3. The kinetics of change of mRNA abundance of β-GAL mRNA after inhibition of protein synthesis. Total RNA was harvested from L. chagasi stably transfected with ␣(958), ␣/(583)/␣, ␣/(673)/␣, or L(1258) 0–24 h after the addition of cycloheximide. Northern blots were hybridized sequentially with the β-GAL-coding sequence to determine the abundance of reporter mRNA and the α-TUB-coding sequence for a loading control. Bands were quantified using densitometry. Data shown are the average of at least three Northern blots prepared from independent RNA preparations, and graphed as a percent of maximum intensities with (S.E.) bars. Similar data for endogenous MSPL mRNA is included in the graph for comparison.

The sequence between 303 and 505 within the MSPL intercoding region appears responsible for suppression of MSPL mRNA abundance in the absence of cycloheximide, and partly responsible for the de-repression of MSPL mRNA abundance with cycloheximide. This region of 202 nucleotides can be divided into four segments. Two are highly similar to corresponding sequences within the MSPS intercoding region (black boxes, Fig. 5) and two segments are unique to MSPL (crossed-hatched boxes, Fig. 5). Our data suggest that the full cycloheximide-responsive effect requires elements in both the 5 and 3 halves of this 202-nucleotide region. Constructs containing only the first 91 bases [L(394)] or the remaining 111 bases [␣/(583)/␣] have only partial mRNA destabilization at baseline and partial induction with cycloheximide (Fig. 1). In contrast, transfectants expressing plasmids containing neither of these regions express high levels of β-GAL mRNA, whereas transfectants with constructs containing both halves demonstrate fully repressed β-GAL mRNA at baseline (and dramatically increased mRNA levels after protein synthesis inhibition).

4. Discussion There are two well-documented trypanosomatid genes whose mRNAs increase in abundance in response to protein

J.E. Purdy et al. / Molecular & Biochemical Parasitology 142 (2005) 88–97

95

Fig. 4. The half-life of β-GAL mRNA from plasmids ␣/(673)/␣ or ␣/(583)/␣ either with or without cycloheximide. Total RNA was harvested from L. chagasi stably transfected with either ␣/(673)/␣ or ␣/(583)/␣ for 0–12 h after the arrest of RNA synthesis. The addition of actinomycin D was preceded by 4 h of protein synthesis inhibition for RNA displayed under “+cycloheximide.” Northern blots were hybridized sequentially with the β-GAL-coding sequence to determine the abundance of reporter mRNA, and α-TUB-coding sequence for a loading control. Bands were quantified using densitometry. Graphs shown on the right summarize the average of at least three Northern blots of independent RNA preparations and depict the mean percent of maximum intensities with S.E. bars.

synthesis inhibitors: the EP-procyclin gene of T. brucei [19] and MSPL of L. chagasi [20]. A similar increase in MSPL mRNA was observed using inhibitors that block protein synthesis at different steps of translation (cycloheximide, puromycin, pactamycin) [20]. Hence, the increase in MSPL mRNA is due to a cycloheximide-dependent increase in half-life, rather than changes in the rate of transcription or polysome stabilization [20]. The mechanism responsible for this increase, however, is unknown. Herein, we showed that the 3 UTR region located 303–505 nucleotides downstream of the MSPL stop codon is responsible for the suppression of mRNA in logarithmic promastigotes, and partly responsible for the cycloheximide-induced increase in the steady-state abundance of endogenous MSPL mRNA. Parasites transfected with plasmid constructs

containing the entire 303–505 region cloned downstream of β-GAL produced very low levels of baseline β-GAL mRNA, similar to endogenous MSPL mRNA. Inhibition of protein synthesis resulted in an increase in β-GAL mRNA abundance, suggesting there is a sequence-dependent destabilization of this mRNA. This conclusion is supported by the fact that the construct lacking any of this sequence [␣/(235)/␣] expressed β-GAL mRNA at a high baseline level, whereas constructs containing only half the sequence [L(394)/␣, L(583)/␣] had baseline levels of β-GAL mRNA that were lower than ␣/(235)/␣ but still much higher than constructs containing all of region 303–505 [␣(673)/␣]. Constructs containing the 202 bp region exhibited a partial (4.4-fold) cycloheximide effect, whereas constructs containing additional 240 bp downstream (e.g. L(745)␣) underwent a full

Fig. 5. Analysis of the MSPL intercoding region between nucleotides 303 and 505. (A) The amount of identity between MSPL and MSPS in the four segments of this 303–505 region is shown above, and the % AU content is shown below the schematic representation. Areas with a high degree of homology are shown in black boxes (
), while divergent areas are shown in crossed-hatched boxes ( ). (B) The location and sequence of two conserved, or nearly conserved, sequences (A and B) identified in both unique regions are shown.

96

J.E. Purdy et al. / Molecular & Biochemical Parasitology 142 (2005) 88–97

cycloheximide effect (11.8-fold induction). Therefore, it seems a likely possibility that additional sequences may play a role in full induction of β-GAL mRNA after protein synthesis inhibition. If cycloheximide induction of mRNA regulated by region 303–505 was due to the removal of mRNA suppression, one would expect the half-life of these mRNA to be low at baseline and to increase with protein synthesis inhibition. In fact, the half-life of 303–505 containing mRNA [␣/(673)/␣] was 1.8 h compared to the half-life of 7.7 h for constructs lacking this entire region [␣(958)] or 7.3 h for endogenous α-TUB (Table 2). Transfectants expressing constructs containing only half of the 303–505 region had intermediate half-life values [T1/2 of ␣/(583)/␣ 4.6 h]. In the presence of cycloheximide, the β-GAL mRNA half-life in transfectants containing the 303–505 region increased up to 11.1 h, a level that is comparable to the β-GAL mRNA half-life in transfectants lacking this region [13.5 h for ␣(958) and 12.0 h for endogenous α-TUB]. The similar level of cycloheximide induction and mRNA half-life between endogenous MSPL mRNA and reporter gene transfectants containing the 303–505 region of the MSPL 3 UTR suggests that regulation of these mRNAs occurs through a similar mechanism. Likewise, the kinetics of β-GAL mRNA change in transfectants containing the 303–505 region were similar to the kinetics of cycloheximide induction of endogenous MSPL mRNA. Furthermore, we confirmed that changes in mRNA abundance were not due to changes in plasmid copy number. Prior studies of trypanosomatids have provided insight into gene regulation at the level of mRNA stability. AU- and G-rich sequences have been identified in T. cruzi that destabilize their mRNAs, similar to higher eukaryotic AREs [10,33]. A protein factor that binds to AU-rich sequences appears to target these mRNAs for degradation [34]. Neither AU- nor G-rich elements have been identified in Leishmania sp. to date. Beside MSPL, another known trypanosomatid protein whose mRNA increases with protein synthesis inhibition is EP-procyclin in T. brucei. A cycloheximide responsive element in the EP-procyclin 3 UTR has not been mapped [15]. The data described here and our prior work suggest that MSPL mRNA in L. chagasi is destabilized by a labile protein factor(s) that binds to region 303–505, targeting this mRNA for degradation. Region 303–505 from MSPL contains no sequence homology to the PFR2 element of T. brucei [14], nor does it contain sequences consistent with AU- or G-rich elements identified in other Trypanoasomatidae. In fact, none of the four segments of region 303–505 are AU-rich except the first 22 nucleotides that are conserved with MSPS (Fig. 5). Furthermore, no significant secondary structure was identified in this region using the Wisconsin GCG Package (Genetics Computer Group Inc.). However, sequence comparison of the two unique segments within region 303–505 found two conserved sequences present in both segments. A single copy of sequence A (TTCCCTG) occurs in both unique segments

(Fig. 5B). One copy of sequence B (CCCGGCGC) occurs in the first unique segment and three copies in the second unique region. Work is ongoing to determine what role, if any, these sequences play in MSPL mRNA regulation and to identify the protein(s) factor that binds to and destabilizes mRNA containing region 303–505. The element affecting Paraflagellar Rod Gene PFR2 mRNA abundance characterized by Mishra et al. [14] was identified in Leishmania mexicana rather than Trypanosoma cruzi as reported above.

Acknowledgements The authors thank to Melissa Miller for excellent technical support. This work was supported by a Merit Review grant from the Department of Veteran’s Affairs (MEW), and National Institutes of Health grants R01 AI45540 (MEW), RO1 AI48822 (MEW), R01 AI32135 (MEW and JED), and K08 AI055804-01 (JP). Work was performed during funding by NIH Training Grant AI07343.

References [1] Ouellett AJ, Malt RA. Accumulation and decay of messenger ribonucleic acid in mouse kidney. Biochemistry 1976;15: 3358–61. [2] Bagby GC, Shaw G, Heinrich MC, et al. Interleukin-1 stimulation stabilizes GM–CSF mRNA in human vascular endothelial cells: preliminary studies on the role of the 3 AU rich motif. Prog Clin Biol Res 1990;352:233–9. [3] Greenberg ME, Shyu AB, Belasco JG. Deadenylylation: a mechanism controlling c-fos mRNA decay. Enzyme 1990;44: 181–92. [4] Sprenger H, Lloyd AR, Lautens LL, Bonner TI, Kelvin DJ. Structure, genomic organization, and expression of the human interleukin-8 receptor B gene. J Biol Chem 1994;269:11065–72. [5] Stoecklin G, Hahn S, Moroni C. Functional hierarchy of AUUUA motifs in mediating rapid interleukin-3 mRNA decay. J Biol Chem 1994;269:28591–7. [6] Myler PJ, Audleman L, deVos T, et al. Leishmania major Friedlin chromosome 1 has an unusual distribution of protein-coding genes. Proc Natl Acad Sci USA 1999;96:2902–6. [7] Brittingham A, Miller MA, Donelson JE, Wilson ME. Regulation of GP63 mRNA stability in promastigotes of virulent and attenuated Leishmania chagasi. Mol Biochem Parasitol 2001;112:51–9. [8] Kelly BL, Nelson TN, McMaster WR. Stage-specific expression in Leishmania conferred by 3 untranslated regions of L. major leishmanolysin genes (GP63). Mol Biochem Parasitol 2001;116:101–4. [9] Weston D, La Flamme AC, Van Voorhis WC. Expression of Trypanosoma cruzi surface antigen FL-160 is controlled by elements in the 3 untranslated, the 3 intergenic, and the coding regions. Mol Biochem Parasitol 1999;102:53–66. [10] Di Noia JM, D’Orso I, Sanchez DO, Frasch AC. 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. J Biol Chem 2000;275:10218–27. [11] Furger A, Schurch N, Kurath U, Roditi I. Elements in the 3 untranslated region of procyclin mRNA regulate expression in insect forms of Trypanosoma brucei by modulating RNA stability and translation. Mol Cell Biol 1997;17:4372–80.

J.E. Purdy et al. / Molecular & Biochemical Parasitology 142 (2005) 88–97 [12] Aly R, Argaman M, Halman S, Shapira M. A regulatory role for the 5 and 3 untranslated regions in differential expression of hsp83 in Leishmania. Nucleic Acids Res 1994;22:2922–9. [13] Burchmore RJ, Landfear SM. Differential regulation of multiple glucose transporter genes in Leishmania mexicana. J Biol Chem 1998;273:29118–26. [14] Mishra KK, Holzer TR, Moore LL, LeBowitz JH. A negative regulatory element controls mRNA abundance of the Leishmania mexicana paraflagellar rod gene PFR2. Eukaryot Cell 2003;2:1009–17. [15] Fluck C, Salomone JY, Kurath U, Roditi I. Cycloheximide-mediated accumulation of transcripts from a procyclin expression site depends on the intergenic region. Mol Biochem Parasitol 2003;127:93–7. [16] Nozaki T, Cross GA. Effects of 3 untranslated and intergenic regions on gene expression in Trypanosoma cruzi. Mol Biochem Parasitol 1995;75:55–67. [17] Zeiner GM, Sturm NR, Campbell DA. The Leishmania tarentolae spliced leader contains determinants for association with polysomes. J. Biol. Chem. 2004;278:38269–75. [18] Boucher N, Wu Y, Dumas C, et al. A common mechanism of stageregulated gene expression in Leishmania mediated by a conserved 3 -untranslated region element. J Biol Chem 2002;277:19511–20. [19] Graham SV, Barry JD. Polysomal, procyclin mRNAs accumulate in bloodstream forms of monomorphic and pleomorphic trypanosomes treated with protein synthesis inhibitors. Mol Biochem Parasitol 1996;80:179–91. [20] Wilson ME, Paetz KE, Ramamoorthy R, Donelson JE. The effect of ongoing protein synthesis on the steady-state levels of Gp63 RNAs in Leishmania chagasi. J Biol Chem 1993;268:15731–6. [21] Hotz HR, Biebinger S, Flaspohler J, Clayton C. PARP gene expression: control at many levels. Mol Biochem Parasitol 1998;91:131–43. [22] Roberts SC, Swihart KG, Agey MW, et al. Sequence diversity and organization of the msp gene family encoding gp63 of Leishmania chagasi. Mol Biochem Parasitol 1993;62:157–71. [23] Ramamoorthy R, Donelson JE, Paetz KE, et al. Three distinct RNAs for the surface protease gp63 are differentially expressed during development of Leishmania donovani chagasi promastigotes to an infectious form. J Biol Chem 1992;267:1888–95.

97

[24] Purdy JE, Donelson JE, Wilson ME. Leishmania chagasi: the ␣tubulin intercoding region results in constant levels of mRNA abundance despite protein synthesis inhibition and growth state. Exp Parasitol, in press. [25] Pearson RD, Steigbigel RT. Mechanism of lethal effect of human serum upon Leishmania donovani. J Immunol 1980;125:2195–201. [26] Beetham JK, Myung KS, McCoy JJ, Wilson ME, Donelson JE. Glycoprotein 46 mRNA abundance is post-transcriptionally regulated during development of Leishmania chagasi promastigotes to an infectious form. J Biol Chem 1997;272:17360–6. [27] Sambrook J, Russel DW. Molecular cloning: a laboratory manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 2001. [28] Horton RM, Cai ZL, Ho SN, Pease LR. Gene splicing by overlap extension: tailor-made genes using the polymerase chain reaction. Biotechniques 1990;8:528–35. [29] Beverley SM, Ismach RB, Pratt DM. Evolution of the genus Leishmania as revealed by comparisons of nuclear DNA restriction fragment patterns. Proc Natl Acad Sci USA 1987;84:484–8. [30] Kapler GM, Coburn CM, Beverley SM. Stable transfection of the human parasite Leishmania major delineates a 30-kilobase region sufficient for extrachromosomal replication and expression. Mol Cell Biol 1990;10:1084–94. [31] Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate–phenol–chloroform extraction. Anal Biochem 1987;162:156–9. [32] Anderson KA, Britigan BE, Wilson ME. Regulation of inducible heat shock protein 70 genes in Leishmania chagasi. Am J Trop Med Hyg 1996;54:471–4. [33] D’Orso I, Frasch AC. Functionally different AU- and G-rich ciselements confer developmentally regulated mRNA stability in Trypanosoma cruzi by interaction with specific RNA-binding proteins. J Biol Chem 2001;276:15783–93. [34] D’Orso I, Frasch AC. TcUBP-1, a developmentally regulated U-rich RNA-binding protein involved in selective mRNA destabilization in trypanosomes. J Biol Chem 2001;276:34801–9.