THE JOURNAL OF BIOLOO~CAL C H E M I ~ Y Vol. 269, No. 32, Issue of August 12, pp. 20509-20516,1994 Printed in U.S.A.

A Differentially Expressed Gene Family Encoding "Amastin," a Surface Protein of Dypanosoma cruxi Amastigotes"

(Received for publication, December 28, 1993, and in revised form, May 23, 1994)

Santuza M. R. TeixeiraS, David G. Russell$, Louis V. KirchhofFnll, and John E. Donelson$**$$ From the Departments of **Biochemistry and Wnternal Medicine, University of Zowa, the Wepartment of Veterans Affairs Medical Center, the $Howard Hughes Medical Institute, Iowa City, Iowa 52242, and the $Department of Molecular Microbiology, Washington University, St. Louis, Missouri 63110

A new family of closely related glycoproteins, collec- tively called amastins, has been found on the surface of the amastigote form of !#ypanosoma cruzi. The gene family encoding these amastigote-specific proteins was identified by differentially screening an amastigote cDNA library with reverse transcribed poly(A)+ RNA from amastigote and epimastigote stages of the parasite. Amastins are encoded by eight or more tandem genes, at least five of which are distinguished by nucleotide point changes. The 1.4-kilobase amastin mRNAs are 50 times more abundant in amastigotes than in epimastigotes or trypomastigotes. The amastin genes are transcribed to an equal extent in both amastigotes and epimastigotes, indicating that the stage-specific amastin mRNA levels are determined by a post-transcriptional mechanism. Sequence determination of full-length cDNAs reveals an open reading frame encoding 174 amino acids and a 700- base pair 3'-untranslated region. Nascent amastins con- tain four distinct hydrophobic regions of 20-30 amino acids each, 2 at internal locations and 1 each at the N and C termini.

During its life cycle, !lFypanosoma cruzi, the protozoan para- site that causes Chagas' disease, goes through three develop- mental stages. In the midgut of the reduviid bug vector, the parasites multiply as extracellular epimastigotes that eventu- ally migrate to the hindgut, where they differentiate into non- dividing metacyclic trypomastigotes and are excreted. After passing through mucous membranes and skin cuts of the mam- malian host, the infective trypomastigotes can invade a variety of cell types. Once inside a cell, they differentiate into amastig- otes that multiply in the cytoplasm. Amastigotes transform into trypomastigotes and enter the circulation when the host cell ruptures. These bloodstream trypomastigotes spread the infec- tion to other tissues and to a new host if they are ingested by reduviid bugs during a blood meal. Although amastigotes origi- nally were defined as an obligate intracellular form, they also can be released from ruptured cells. Recently, it has been

Health Research Grants AI24711 (to L. V. K.), AI 26889, and AI 34207 *This research was supported in part by National Institutes of

( to D. G. R.) and by the Department of Veterans Affairs Research Ser- vice (to L. V. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This 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 GenBankTMIEMBL Data Bank with accession number(s) U04337, U04338, U04339, U04340, U04341.

Heart Association. (1 Recipient of an Established Investigatorship from the American

$$ 'b whom correspondence should be addressed: DeDt. of Biochem- istry, University of Iowa, Iowa City, IA 52242. %I.: 319:335-7889; Fax: 319-335-6764.

shown that extracellular amastigotes can infect cells both in vivo and in vitro and thus contribute to sustaining the infection in the host (1-3).

Chagas' disease is a major health problem throughout most of Latin America. Almost all acute cases resolve spontaneously in a few weeks, and infected persons enter the indeterminate or asymptomatic phase of chronic II cruzi infection. During this phase, few, if any, bloodstream trypomastigotes are detected, but it is generally assumed that the intracellular amastigotes are present at all times. It is possible that more effective meth- ods for identifying and treating the disease could be developed if the intracellular form of the parasite and the molecular mechanisms controlling its differentiation were better under- stood. However, most of the biochemical data accumulated to date on !l! cruzi have been obtained by studying the insect form epimastigotes, which can be easily cultivated in vitro, or by analyzing the trypomastigotes' capacity to invade mammalian cells in vitro.

Thus, we set out to identify genes that are preferentially expressed by amastigotes in anticipation that these genes will provide new insights into this intracellular form of the para- site. We constructed a cDNAlibrary from RNAof tissue culture- derived amastigotes and used a differential screening proce- dure to isolate cDNAs encoding amastigote-specific proteins. Using this approach, we identified a family of closely related genes expressed at high levels only in amastigotes that encode novel surface glycoproteins, which we designate "amastins."

MATERIALS AND METHODS Parasites-Epimastigote forms of Tulahuen (25) and Y strains (4)

and Sylvio X-1014 clone ( 5 ) of T cruzi were maintained in logarithmic growth phase at 26 "C in supplemented liver digest neutralized medium as described previously (6). The renal carcinoma cell line RA786 (7) was grown in RPMI medium at 37 "C and passed weekly. Culture-derived trypomastigotes and extracellular amastigotes of the Tulahuh strain were obtained by infecting monolayers of RA786 cells grown to 50% confluency with stationary phase parasites cultivated in liver digest neutralized medium. One to two weeks after infection, trypomastigotes released in the supernatant were collected and purified in a 10-15% discontinuous metrizamide gradient (Sigma) as described (8). Amastig- otes were obtained from supernatants of cultures of RA786 cells 3 weeks after infection and purified by centrifugation in a 15-1%21% metriz- amide gradient. The purity (>95%) as well as the morphology of the parasites in these preparations were analyzed by light and electron microscopy.

Construction and Differential Screening of an Amastigote cDNA Library"A unidirectional cDNA library in bacteriophage AZAPII was constructed from 3 pg of poly(AY RNA purified from T cruzi amastig- otes using procedures provided by the supplier of the kits for cDNA synthesis and in uitro phage packaging (Stratagene). Alibrary of 2 x lo6 independent clones was amplified to a titer of 1 x 101o/ml and plated at low density (-5000 phagdplate) in a total of 10 plates. After incubation for 8 h at 37 "C, two replica filters of each plate were hybridized to 32P-labeled, first strand cDNA prepared from 2 pg of epimastigote or amastigote poly(AY RNA using random primers (9). The filters were incubated with each cDNA probe (2 x lo8 total cpm) at 42 "C for 48 h in

20509

20510 Amastin of T cruzi

50% formamide, 6 x SSC, 5 x Denhardt's solution, 10% dextran sulfate, 0.1% SDS, 0.1% sodium pyrophosphate, and 100 pg salmon sperm DNA/ ml. The phage clones that strongly hybridized to the amastigote cDNA and not to the epimastigote cDNA were plaque-purified after two rounds of hybridization. To confirm the purification, a third round of hybridization was performed with polymerase chain reaction-amplified inserts from all isolated clones.

cDNA Characterization-Recombinant phagcmids were excised by coinfecting Escherichia coli XL-1 Blue cells with the selected AZAPII phage and R408 helper phage as described by the commercial supplier of the phage (Stratagene). Plasmid DNAs were prepared using the Qiagen method (Qiagen, Inc., Chatsworth, CA) and cDNA inserts se- quenced using the Sequenase kit (U. S. Biochemical Corp.). The com- plete cDNAsequence of both strands was obtained using oligonucleotide primers spaced about 250 base pairs apart.

Nucleic Acid Manipulations-Total RNA was isolated from parasite cultures or from RA786 cells using LiCl and urea as described previ- ously (10). Poly(A)+ fractions were obtained by hybridization to oligo(dT) using Dynabeads (Dynal Inc., Great Neck, NY). For Northern blots, 3 pg of total RNA were separated in 1.2% agarose gel containing formalde- hyde (11). Parasite DNA was purified as described (121, digested with restriction endonucleases, and fractionated in 0.6 or 0.8% agarose gels using either conventional or pulse field electrophoresis (CHEF-DR I1 system, Bio-Rad). DNA and RNA were transferred to nitrocellulose membranes (Schleicher & Schuell) and hybridized to 32P-labeled probes using standard procedures (12). DNA probes obtained either by polym- erase chain reaction amplification or by restriction endonuclease diges- tion of plasmids were labeled with [a-"PldCTP to high specific activity (lo9 cpdpg) using a random oligonucleotide primer kit (Boehringer Mannheim).

Antibody ProductionSynthetic peptides corresponding to amino ac- ids 36-47 (peptide P1: ATGNTPCITLWG) and amino acids 63-76 (pep- tide P2: EDFRECPSWRLFR) were synthesized by Chiron Minotopes P/L (Clayton, Victoria, Australia) and were conjugated to diphtheria toxoid. Rabbits were immunized with 500 pg of conjugated peptide reconstituted in phosphate-buffered saline and injected with 1 ml of MPISTDM/CWS Emulsion (RIBI ImmunoChem Research, Inc., Hamil- ton, MT). The rabbits received three injections a t 3-week intervals according to the immunization protocol suggested by the adjuvant sup- plier (RIBI) and were bled 12 days after the last injection.

Immunoelectron Microscopy-Tissue culture amastigotes were fLxed in 1% glutaraldehyde in Hepes saline, pH 7.0, as described previously (13). The cells were pelleted, embedded in gelatin, and infiltrated with 2.3 M sucrose/20% polyvinylpyrrolidone in phosphate-buffered saline. The blocks were trimmed, frozen, and sectioned with an RMC MT7/ CR21 cryoultramicrotome and were probed with antibodies. All labeling experiments were conducted in parallel with controls omitting the pri- mary antibody or substituting "non-immune" rabbit IgG or irrelevant mouse monoclonal antibodies against Leishmania antigens. These con- trols were consistently negative at the concentrations of gold-conju- gated second antibodies (Amersham Corp.) used in these studies.

7kanscription in Isolated Nuclei-Nuclei were isolated from epimas- tigotes and amastigotes following a modification of the Nonidet P-40 lysis protocol. Briefly, 1 x lo9 cells were washed in phosphate-buffered saline and resuspended in 10 ml of hypotonic buffer A (10 mM Tris-HC1, pH 7.6,2 mM MgCl,, 5 mM KCI, 2 mM CaCI,, 0.5 mM dithiothreitol, 1 mM EDTA, 1 mM spermidine, 6% polyethylene glycol) and incubated on ice for 10 min. After the addition of 1.5% Nonidet P-40, the cells were lysed by homogenization with 100 strokes in a tight fitting Dounce homoge- nizer. The cell lysate was diluted with 1 volume of 2 x buffer B (0.64 M

sucrose, 40 mM Tris-HC1, pH 7.6,60 mM KC], 1 mM EDTA, 1 mM EGTA, 0.5 mM dithiothreitol, 1 mM spermidine), and the nuclei were centri- fuged a t 2500 x g for 15 min. The nuclear pellet was washed once with 10 ml of 1 x buffer B and resuspended in 100 pl of buffer C (25% glycerol, 50 mM Tris-HC1, pH 8,60 mM KCI, 1 mM dithiothreitol, 0.1 mM EDTA, 1 mM spermidine). Nuclear aliquots were quick-frozen in dry ice and stored at -70 "C. Run-on transcripts were labeled with [a-"PIUTP a t 30 "C for 30 min as previously described (14). The labeled nascent RNA was extracted with phenol-chloroform, precipitated with ethanol, and purified using a G50 column (Quickspin Column, Boehringer Mannheim). Hybridization to slot blots were carried out for 3 days a t 65 "C using 2-5 x lo6 cpm of labeled RNA in 5 ml of hybridization solution and washed as described (14). The immobilized DNAs (5 pg/ slot) consisted of single-stranded sense and anti-sense strands of M13 mp18 or M13 mp19 clones containing the full-length TcA7 cDNA or a 540-base pair fragment encoding part of the large subunit 24Sa rRNA.

A B

1 2 3 4 1 2 3 4

b

7.4 kb- - 4.4 kb-

2.4 kb-

1.3 kb-

cruzi. Northern blot analyses were performed with total RNA (5 pg/ FIG. 1. Amastin mRNA expression during the life cycle of !l!

lane) from the human adenocarcinoma cell line RA786 (lane I ) , and total RNA (3 pgAane) from I: cruzi amastigotes (lane 2) , trypomastig- otes (lane 3 ) , and epimastigotes (lane 4) . Panel A shows the autoradio- graph of a blot probed with a 32P-labeled amastin cDNA (TcA33). The same blot was reprobed with a polymerase chain reaction-generated fragment containing part of the I: cruzi 24Sa rRNA gene (panel B) . Molecular size markers are shown on the left.

RESULTS

Isolation of cDNA Clones Corresponding to Amastigote-spe- cific Genes-Infection of monolayers of RA786 cells with cul- ture-derived trypomastigotes of the Tulahu6n strain of T cruzi resulted in the release of large numbers of amastigotes into the supernatant (>107/ml). In cultures that had been infected for only 1 to 2 weeks, we observed a mixed population with ap- proximately equal numbers of trypomastigotes and amastig- otes in the supernatant, while, after 3 weeks, the supernatant contained more than 90% amastigotes. The two developmental forms were separated on the basis of their difference in cell density using a metrizamide gradient. Several groups have shown, using morphological and biochemical data, that these extracellularly derived amastigotes are indistinguishable from intracellular amastigotes (3,8, 15). We confirmed these obser- vations by analyzing electron micrographs of amastigotes iso- lated from supernatants of infected RA786 cultures as well as by probing Western blots with the amastigote-specific mono- clonal antibody Ssp-4 (16). Ssp-4 detected a double band of proteins with apparent molecular masses of 75 and 85 kDa only in cell lysates of amastigotes but not in epimastigotes (not shown).

To identify genes preferentially expressed in amastigotes, we differentially screened about 50,000 clones of an amastigote cDNA library with 32P-labeled cDNA probes derived from re- verse transcribed poly(A)+ RNA of amastigotes and epimastig- otes. By comparing the signals on replica filters, we identified seven clones that, during three rounds of screening, consist- ently displayed more intense signals on filters hybridized with the amastigote probe. Restriction analysis of the pBluescript plasmids recovered from these seven phage revealed that their cDNA inserts ranged from 0.7 to 1.5 kb.'

To confirm that these cDNAs correspond to stage-specific mRNA, we probed Northern blots containing total RNA from amastigotes, epimastigotes, and culture-derived trypomastig- otes with the inserts purified from each clone. Four of the seven cDNAs inserts detected a single message of 1.4 kb that was 50

The abbreviations used are: kb, kilobaseb); 3'-UTR, 3'-untranslated region.

Amastin of I: cruzi 20511

-T CTCACGTAOO MOCMTOAO CAhACTpGoc GCGATTCTCT ACUXGCGGT GGCCTPCCTP GCCTTCCTGT TCCGTGCTCGT 100 .................................................................................................... --------- .......... "-""". ...................................................................... --------- --------- ................................................................................ -CG ATJGACCAGT TCCGCACGAA GCCXTWXA ACACGCCGTG CATTACGTTC "X CCAMMlATC 200 .................................................................................................... ......................... G.... A...,..GG. ....................... A.,.... ...... A... .G........ ... C....C. .................................................................................................... .................................................................................................... AAGTAl3%&&?X%%XCTTTTG GGAAGATITT CGTGMTGTC CCTCAGTCTT DADGcTGTTC AGWLTGGCCG CGOCGTTCTC CATTATTTCC A m 300 .......... .A........ ................................................................................

U

A . . . . . . . . . ........................................................ T... ..............................

MRZ

............... G..G. .................. C. .............................. A......... .................... .................................................................................................... .................................................................................................... TGCTTGCTOC MCCATACTT GGCGTGCCCO CACACTTCTA CCTAAAATCC CTCA@%%k@ lTGCGACCCT GCTGTIGGTG GTCAGCAITG TGACGGTTGG 400 .................................................................................................... .......... C...GC.... ................... G ......... T ..................................................

891 11

................................................................... .............................. G.. .................................................................................................... GCTCGTl!TG&%&?ECAn;G CCTACCWA CAMCGTOAC GTGGACAACT GCTTTGGAGA ACCGATAGAA WTGGTCGA AATACGGTTC TGGCTTTCTC 500

-1

................................................................... .............................. c.. .......... A.... . . . . . ..... T.... ... C.A.... ........ T. ........AC .... C.CA.. .CAG.T.... ........GG ....... G..

TcA7 T c A l T a l TcA33 TcMl

TCA7 TCAl T C h z l TCA33 T a l

TCA7 TCAl TCA2l KA33 TcA51

TCA7 TCAl TCA2l TCA33 T a l

TCA7 TCAl TCA21 TCA33 .................................................................................................... T c M l .................................................................................................... TCA7 ATCATlETGA TTGGATGGTG TCXGATATPT GTCGCCGTCG TGCTTCTGCG TGTGCTATAG CGGDPGCATC CACCGTCTGC ATGCGCCGCT GGAGGCCAAA 600 TCAl ............................................................ . . . - e . . . . . .............................. TCA2l .................... C.. . . . . . . . .................... .A........ .............................. ....A....C

TcA51 .................................................................................................... TCA7 A-GCGTATPC TGCTCTACTA TGCCCCGTGC CGTGCAGCTT TGTGGATGAC CTCCCTGCM TATTTGATGA UWTACTTSC T P T G C m T A TTTGTTTTAT 700

- TCA33 ........................................................................................ T - ..........

X A l ............................................................................................ A s . . . . .

XA21 .A........ .................... .A........ ...................... C.....TG .. T. . . . . . . C ............ T...A..

TCMl TCA33 Am.....

TCA7 TTTTTlCTTG TmTGTGAC TGCGTCCMT CXCCTGTGT TACCCMTGC TGTTAGAGGT GGAGCGGGTG CTATAATGAC ATGTGTGCAT CTGCTGTCCC 800 TCAl ....................................................... C.. . . ........................................ TCA2l .. G..-.... .... A...G. ............................................................... C.... . . . . . . . .C... TCA33 ................................................................................................. C . . . TCA551 .................................................................................................... TCA7 TCTGTfXGTG TGOGCGAACG CTTGTTCCCC GOCCCAGGGA GAAASATA GATGGTAAGC AGGCAAGAAG AAXGGAAGG AGGAAGAGAA TTACGTTGTG 900 TCAl .............................................. A.. . .................................................. TCA21 ................... C G......... ................ A... .... A..... ..... G.... .;........ .................... %A33 .............................................. A , . . .................................................. T c M l ................ ;... .......................... A... .................................................. F A 7 TACTCTGACG AGGGTCTCCA CGAGGGACCC GCGTCGCCGT CTGTTTGTGA TGoGACGTTT TlTATTlTTA "ITCTACTA CTTGAA'IGCG CGGTGATCXG 1000 TCAl .................................................................................................... TCAZl .............................................................................. AG G.. . . . . . . . .......... TCA33 .................................................................................................... TcA51 ................................................................................................... TCA7 CGATTTXGTG CGTGTTTGCC CGTCWGCG TTPGTGTTTA CTGCTCGCW CCTPCTTCCC GTCGGTCCAC A A m G c C T C TGTPGGCGTC ACGCGATGAA 1100 TCAl .................................................................................................... TCA2l .................................................................................................... TCA33 .................................................................................................... TcA51 .....................................................................................................

............................................................................................. .....................................................................................................

TCA7 TTTTGATTGT GTC-TTCGGT GCTGCCTTTT TTGTTTTTGT GCTAGCGTTT TTTPTTTTTT TTGTGTGTGT OPCTGTGTGA GCGGCACTGC TGAGAGAAAA 1200 TCAl ........................ T.... . .............................. G ......... C.G....A.C -- .................. TCAZl ............. C . . . . . . ....................................... G ..T------- C.G....A.C -- .................. TCA33 ............. C.. . . . . .......................................... T-...... .............................. TcA51 ............. C - . . . . . .................... . . . . . e . - - . ........................................... A.A....

TCA7 AAAAAhAAAA AAAAAAAAAAA TCAl .............. TCA21 .................. TCA33 ....... TcA51 ...............

1221

FIG. 2. Comparison of the nucleotide sequences of five independently isolated amastin cDNh. Dots and dashes denote nucleotide identities and gaps introduced to maximize alignment, respectively. The box at the upper left encloses 17 nucleotides of the II: cruzi 39-nucleotide spliced leader. Start and stop codons are ouerlined. Recognition sites for EcoRI, BglII, and KpnI are shaded.

times more abundant in amastigotes than in epimastigotes or trypomastigotes, as determined by densitometry. Fig. lA shows the result of a Northern blot probed with the cDNA insert in clone TcA33. As expected, RNA isolated from noninfected M786 cells did not hybridize to this T cruzi cDNA. A 540-base pair fragment derived from the large subunit 24Sa rRNA gene (171, obtained by polymerase chain reaction amplification of T cruzi genomic DNA, was used to probe the same blot to correct the signal intensity of the bands for the amount of RNA loaded (Fig. 1B). Northern blots probed with a fifth cDNA revealed a 0.5-kb RNA present at greater levels in amastigotes. A sixth clone, TcA52, hybridized to an RNA species of 1.4 kb, present

only in amastigotes, and to three other larger species, present in both stages (not shown). The seventh cDNA did not detect an RNA in the Northern blot.

Nucleotide Sequences of Amastigote-specific cDNAs-The se- quences of 200-300 nucleotides at both ends of each of the seven amastigote-specific cDNA were determined. The four cDNA clones that hybridized to a 1.4-kb RNA were found to contain very similar sequences at their 3' ends and to have poly(A) tails of about 20 nucleotides, indicating that these cDNAs are derived from closely related genes. The 5' end of one of these clones, TcA21, contains part of the Z! cruzi 39-nucle- otide splice leader, indicating that it is a full-length cDNA.

20512 Amastin Comparison of these nucleotide sequences with sequences in the GenBankTM database did not reveal any substantive simi- larities.

The partial sequence of another of the seven cDNA clones, TcA52, is 85% identical to a member of the previously charac- terized SA-85-1 gene family that is differentially expressed in l! cruzi (18). These 85-kDa surface antigens occur only during the mammalian stages and are trans-sialidases whose activi- ties have been implicated in cell invasion (19). Sequence anal- ysis of the other two cDNAs did not reveal any similarities to previously described sequences.The four cDNAs with similar sequences were selected for further analysis. The cDNA library was rescreened with the TcA33 cDNA to identify additional cDNAs from the same gene family. This probe detected about one positive clone per 1000 clones of the library, reflecting a relatively high abundance of this RNA family in amastigotes. The complete sequences of the three largest clones isolated during the initial differential screening, and the sequences of two full-length cDNAs isolated in the second screening, are compared in Fig. 2. Each of the five cDNAs represents the product of a distinct but closely related gene. Although cDNA clones TcA7, TcA1, TcA33, and TcA51 share >98% identity, clone TcA21 represents a more divergent gene. Three cDNAs were found to contain part of the 5’ splice leader, and, in one of these clones, TcA1, this sequence was attached to a separate downstream site. Seventeen nucleotides downstream of the spliced leader (or two nucleotides in the case of TcAl) is an ATG codon that initiates an open reading frame of 522 nucleotides. This relatively short coding region comprises only 40% of the 1.4-kb mRNA. Although the nucleotide differences among the five cDNAs are approximately equally distributed in the coding region and the first part of the large 3”untranslated region (3’-UTR), the 247 nucleotides in the 3’-UTR region between positions 867 and 1113 are highly conserved. The shortest clone (0.7 kb) isolated during the differential screening contains only the 3’-UTR and is not shown in Fig. 2.

Properties of the Amastin Amino Acid Sequence-Fig. 3A shows an alignment of the amino acid sequences deduced from the five cDNA sequences. The cDNAs encode similar proteins of 174 amino acids with molecular masses of 19.5 ma. We desig- nate this group of proteins the Yamastin‘’ family. Clones TcA7, TcA33, and TcA51 encode identical amino acid sequences, which are shown in the top lane of Fig. 3A. Clone TcAl encodes a protein with only two amino acid substitutions (positions 59 and 145), whereas clone TcA21, the most divergent member of the family, specifies a protein with 23 amino acid changes, 18 of which are conservative substitutions.

A search for homologous proteins in the databases indicated that amastins constitute an unknown class of T cruzi proteins. The most remarkable feature of these proteins is their high degree of hydrophobicity. The hydropathy plot shown in Fig. 3B demonstrates that amastins possess hydrophobic N and C ter- mini and have two additional internal hydrophobic regions of 23 and 22 amino acids, which is a size consistent with that of transmembrane domains. Other characteristics of the amino acid sequence are the presence of 5 cysteine residues (or 6, in TcMl), and numerous serines and threonines but no conven- tional sites for N-linked glycosylation.

Further studies aimed at the biochemical characterization of the recombinant protein have been impaired by the fact that it is not readily expressed in E. coli. Several attempts to produce fusion proteins using pBluescript or pGEX plasmids failed, and, although we did obtain different deletion fragments of amastin as fusions with glutathione S-transferase (not shown), we were not able to express the full-length protein in E. coli. Because these expression systems are designed for high levels

of T cruzi

A 1 M S K L G A I L Y G A V G F L A F L F V L V G T P I D Q F R 30 2 . . . . . . . . . . v . . . . . . . . . . . . . . . . . . . 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

T K E K K A T G N T P C I T L W G V K E D C Q S M K Y E F T 6 0 A . . . G . . . . . . . . . . . . I . . . . H . T . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L .

Pi

F W E D F R E C P S V L R L F R M A G A F S I I S I L L L L 90 V G . . . . . . . . . . . . . . . . E . . . . . . . . . . .

P2

. . . . . . . . . . . . . . . . . . . . . .

A A T I L G V A A H F Y L K S L K I F A T L L L V V S I V T 120 . . . a . . . . . . . c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V G L V W V P M A Y L Y K R D V D N C F G E P I E R W S K Y 150 . . . . . I . . . . F . N H . . . . . . . T . L K T G . . . . . . . . . . . . . . . . . . . . . . . . . . . Q . . . . .

G S G F L I I V I G W C L I F V A V V L L R V L . G . . V . . . . . . . . . . . . . . . . . M . . . . . . . . . . . . . . . . . . . . . . . . .

1 7 4

B 4 .0 5.00 1

I -4 .00 p‘ P2 - 5 . 0 I

1 101 174

amastin cDNAs and a hydrophobicity analysis. A, TcA7, TcA33, FIG. 3. Comparison of amino acid sequences deduced from

and N 5 1 cDNAs encode the same amino acid sequence, which is shown as sequence 1. In sequences 2 (TcA21) and 3 (TcAl), positions identical to those of sequence 1 are indicated by dots. Sequences corre- sponding to synthetic peptides P1 and P2 are underlined. B, hydropho- bicity plot of sequence 1 generated using Hibio MacDNASIS Pro 2.0 and a window size of 6. Hydrophobic regions are assigned a positive index.

of expression, we suspect that amastins might be toxic for bac- teria cells, and, if so, alternative strategies for expression in E. coEi are necessary. It is also worth noting that none of the clones isolated from the unidirectional AZAPII library contain the amastin coding region in frame with the lacZ gene.

Cellular Localization of Amastins-Because the predicted amino acid sequences of amastins strongly suggest a mem- brane localization, we conducted immunoelectron microscopy on tissue-culture amastigotes. The cells were incubated with polyvalent antisera raised against synthetic peptides whose sequences are derived from two structurally discrete regions of the amastin sequence, as shown in Fig. 3. Cryosections of amastigotes incubated with both anti-peptide sera generated comparable labeling patterns. Both antisera showed strong decoration of the cell surface (Fig. 4, A and B) including the membrane of the vestigial flagellum. The antibodies also la- beled intracellular vesicles (Fig. 4C). The distribution of label observed with both of the anti-peptide sera closely paralleled the labeling pattern observed with the mouse monoclonal an- tibody that recognizes the amastigote-specific antigen Ssp-4 (Fig. 40). Controls conducted with “nonimmune,” or irrelevant primary antisera were all negative.

Genomic Organization of Amastin Genes--To determine the number and organization of amastin genes in the l! cruzi ge- nome, we performed Southern blot analysis of total DNA iso- lated from epimastigotes. The DNA was digested with several enzymes, including EcoRI and BglII, whose cleavage sites are present in the coding region, and the blots were probed with the cDNA insert of clone TcA33. Complete digestions resulted in

Amastin of T cruzi 20513

FIG. 4. Cellular localization of amastins in amastigotes. Cryoimmuno- electron micrographs of amastigotes were probed with rabbit antisera generated against the synthetic peptides PI (a) and P2 ( b ) from amastin, followed by 15 nm of gold anti-rabbit IgG. Both antisera strongly label the amastigote plasma- lemma. Label was also associated with in- tracellular vesicles ( c ) , seen in the en- largement ofpanel b. "he labeling pattern observed with both antisera was compa- rable with that seen with a mouse mono- clonal antibody against Ssp-4 ( d ) , fol- lowed by 6 nm of gold anti-mouse IgG. Scale bars = 0.5 Dm.

intense hybridization to a fragment of 3.0 kb (Fig. 51, suggest- ing the presence of multiple genes arranged in tandem. To test this prediction, DNA was partially digested using decreasing concentrations of the enzymes. Hybridization of the cDNA to genomic DNA partially digested with EcoRI and BglII resulted in a ladder pattern in which the fragment sizes are multiples of 3 kb (Fig. 5, A and B, lanes 24). Similar results were obtained using Hind111 and BamHI (not shown). The largest fragment extends to 24 kb, corresponding to eight genes arranged in tandem. More quantitative methods, such as determination of the intensity of the radioactive signal in each band via densi- tometry, indicated that the amastin cluster contains as many as 23 gene copies. The differences detected in the cDNA sequences were reflected in the complete digests with certain enzymes (e.g. EcoRI) where point differences affecting the restriction sites resulted in fragments that are multiples of 3 kb (Fig. 5, lanes 1).

Fig. 5 0 shows a schematic representation of the amastin gene cluster. The positions of the restriction sites for BamHI andHindII1 in the intergenic region were determined by South- ern blot analysis using different combinations of enzymes in double digestions of the DNA as shown in Fig. 5C. Longer exposures of the blots also reveal fragments that probably cor-

b

respond to the ends of the cluster. They can be detected in the EcoRI digest, where two additional fragments of 1.0 and 5.7 kb have signal intensities that might be expected for the gene at each end of a cluster (Fig. 5, A and C , lanes 1 ). We have con- firmed the overall organization of the amastin cluster and the linkage between genes by characterizing A phages isolated from a genomic DNA library of the TulahuBn strain. In one of the clones, we identified two copies of amastin genes with a 1.8-kb intergenic region and a 5.7 kb EcoRI fragment corresponding to the first gene and 5"flanking sequences (not shown).

We also looked for amastin genes in two other isolates of T cruzi and in other trypanosomatids. Genomic DNAs from T cruzi Tulahuh, Y and Sylvio X-10 isolates, !lYypanosoma brucei rhodesiense and Leishmania donovani chagasi were completely digested with EcoRI and BglII and probed with the cDNA TcA33. As shown in Fig. 6, Y strain displays the same gene organization as that described above for TulahuBn strain. In the Sylvio X-10 clone, instead of the 3- and 6-kb EcoRI frag- ments corresponding to one and two repeats, we detected a 9-kb fragment, which corresponds to three copies of the gene (lane 1). The type of polymorphism observed with EcoRI cleavage sites and with other enzymes (e.g. KpnI) (see Fig. 2) was also present in Tulahuh strain. However, when BglII-digested

Amastin of T cruzi Eco RI B Bgl I1

1 2 3 4 5 1 2 3

20514

A

10.1 9.1 : 8.1 - 7.1 - 6.1 - 5.0 - 4.0 - 30-

48.5-

38.4-

29.9-

19.4-

12.2-

8.S-

""

2 0-

1.6-

D H B E L H B E L

1 0-

C 1 2 3 4 5 6

6.0-

4.0- 5.0-

3.0-

2.0-

1.6-

1.0-

H B E L H B E L

FIG. 5. Organization of amastin genes. T cruzi DNA (5 pg/lane) was digested with serially diluted amounta of EcoRI (A) or BglII (B) and analyzed by Southern blots. DNA in lanes 1-5 were digested with 15, 5, 1.5, 0.4, and 0 units of enzyme, respectively. C, DNA was digested to completion with EcoRI (lane 11, BamHI (lane 2), EcoRI and BamHI (lune 3), BglII and BamHI (lane 41, EcoRI and HindIII (lune 5), BglII and HindIII (lane 6) . After separation in 0.6% (A) or 0.8% (C) agarose gels or a 0.8% CHEF gel ( B ) , the digestion products were hybridized with 32P-labeled TcA33 cDNA. Molecular size markers in kb are shown on the left. D, schematic diagram of a cluster of amastin genes inferred from the results of the Southern analyses. Open boxes represent coding regions, shaded boxes correspond to 3'-untranslated regions, and the horizontal line indicates intergenic segments. Cleavage sites for HindIII ( H ) , BamHI ( B ) , EcoRI (E) , and BglII (L) are indicated.

DNAs were analyzed, the same pattern was found in all three strains (Fig. 6, lanes 2). DNA from L. donovani chagasi and T brucei rhodesiense did not hybridize to the amastin probe, in- dicating that amastin-like genes do not occur in these two trypanosomatids.

Dunscription Analysis of Amastin Genes-To investigate whether the dramatic differences in the steady state levels of amastin mRNA in the different developmental stages are due to changes in transcription rates, we performed run-on assays using nuclei isolated from epimastigotes and amastigotes. These assays are designed to indicate the distribution of RNA polymerases on the DNA template, as measured by the hybrid- ization signals of 32P-labeled nascent RNA binding to the cor- responding immobilized DNA fragments. Nuclei were isolated from amastigotes and epimastigotes and the RNA elongated in the presence or absence of 1 mg of a-amanitidml (which inhib- its RNA polymerase 11). Hybridization of nascent RNA to single strand clones containing the full-length amastin cDNA or part of the 24Sa rRNA gene indicated that there is no substantive difference in the transcription of these genes in amastigotes and epimastigotes (Fig. 7, A and B ) . Because the DNA frag- ments were cloned in M13 vectors in both orientations, we could distinguish between transcription from the sense (+) and anti-sense (-) strands. As shown in Fig. 7, only the anti-sense strands of the two genes are transcribed. When nuclei were incubated in the presence of a-amanitin, the level of nascent

amastin RNA dropped to background levels, indicating that these genes are transcribed by RNApolymerase 11. As expected, synthesis of rRNA, which is transcribed by RNA polymerase I, was unaffected by the presence of a-amanitin. (Fig. 7 0 .

DISCUSSION Using differential hybridization screening of an amastigote

cDNA library, we identified several cDNAs derived from the amastin gene family that is preferentially expressed during this intracellular stage of T cruzi. In principle, differential cDNA library screening can be used to identify any differen- tially regulated gene, but, in practice, usually only those genes encoding abundant transcripts are identified. In our case, the amastin RNAs were found to be 50 times more abundant in amastigotes than in epimastigotes and are represented in the amastigote cDNA library at a frequency of 0.1%. Although we do not have a more precise determination of the relative abun- dance of amastin mRNAs, which requires a probe for an RNA of known abundance, we believe amastin RNA to be a major com- ponent of the mRNA pool in T cruzi amastigotes.

The genomes of three different T cruzi isolates contain tan- dem copies of the amastin genes, and the Tulahuh strain has 8-23 gene copies. This pattern of tandem gene organization is commonly found in trypanosomatids, and it is often difficult to determine the exact number of genes. In T brucei, the primary transcripts from at least some gene clusters are polycistronic

Amastin of II cruzi 20515

T Y S Tb Ld

1 2 1 2 1 2 1 2

6.0- 5.0- 4.0-

3.0-

2.0-

1.6-

1.0-

6.0- 5.0- 4.0-

3.0-

2.0-

1.6-

1.0-

lates. Southern blot analyses were performed with DNAs from T. cruzi FIG. 6. Comparison of amastin genes in different T. cruzi iso-

Tulahuh (T) and Y ( Y ) strains and Sylvio X-10/4 clone (S), T. brucei rhodesiense (2%) and L. donouuni chugusi (Ld), digested with EcoRI (lanes 1 ) and BglII (lanes 2 1. Digestion products were separated in 0.8% agarose gels and probed with 32P-labeled TcA33 cDNA. Molecular size markers in kb are on the left.

A B C

- + - + - +

TcA7 ~

1

!

FIG. 7. Nuclear run-on analyses of nascent RNA from amastin genes. In uitro transcription assays were performed with nuclei iso- lated from log epimastigotes (A and C ) and amastigotes ( B ) in the presence ( C ) or absence (A and B ) of 1 mg/ml of a-amanitin. Single strand DNAs corresponding to the sense (+) and anti-sense (-) strands of the full-length TcA7 cDNA ( a ) and part of the gene for rRNA subunit 24Sa ( r ) were slot blotted on nitrocellulose filters and hybridized with the 32P-labeled nascent RNA.

pre-mRNAs that can be initiated as far as 60 kb upstream of the last gene in the cluster (20). The conversion of the pre- mRNAs to monocistronic mRNAs requires intergenic RNA cleavages and at least two processing events to generate the 5' trans-spliced and 3' polyadenylated mature mRNAs. Recent studies with T brucei and Leishmania indicate that expression of these multigene clusters is not regulated during transcrip- tion initiation but is controlled by post-transcriptional events, including the pre-mRNA processing steps (21, 22).

If gene expression in T cruzi is regulated in a similar fash- ion, several mechanisms could account for the increased levels of amastin mRNAs in amastigotes compared with epimastig- otes. Indeed, our nuclear run-on experiments demonstrate that the amastin genes are transcribed equally in both epimastig- otes and amastigotes, even though amastin RNA is found pre- dominately in amastigotes (Figs. 1 and 7). The differential ac- cumulation of amastin mRNAs could involve pre-mRNA processing, RNA transport or localization, and mRNA stability. The presence of a large 3"untranslated region in the amastin mRNAs with no nucleotide differences between positions 847 and 1113 in four of the five cDNAs leads us to speculate that the 3'-UTR is involved in the post-transcriptional regulation.

The deduced amino acid sequences of five members of the amastin family have characteristics of a small, hydrophobic protein that could be targeted to the cytoplasmic membrane by a signal peptide. This interpretation was confirmed by probing cryosections of amastigotes with antibodies against amastin peptides. Immunoelectron microscopy revealed an abundance of the protein on the surface of the cell and in association with intracellular vesicles. The apparent abundance of label is con- sistent with the relative amount of amastin mRNA in amastig- otes. Although the predicted molecular mass of the 174 residue amastins is 19.5 kDa, preliminary biochemical characteriza- tions, including Western blots, indicate that in extracts of T cruzi amastigotes, they are present as a large molecular weight complex that binds concanavalinA(n0t shown). Thus, amastins likely are glycoproteins that have extensive post-translational modifications. Although no conventional signals for N-linked glycosylation are present, 18 serine and threonine residues occur, some of which could be modified by 0-linked sugar chains. Examples of mucin-like glycoconjugates, which are known to participate in the attachment of the parasite to mam- malian cells, have been recently described in metacyclic T cruzi trypomastigotes (23).

The amastin amino acid sequences are not similar to those of known proteins and provide few clues about amastin function. Several T. cruzi surface antigens have been characterized pre- viously, but the corresponding genes for only a few have been cloned. In addition, several monoclonal antibodies directed against the amastigote stage have been generated (16,24), but no genes for the antigens identified in this manner have been cloned. Andrews et al. (16) have described a major surface gly- coprotein named Ssp-4 whose expression is associated with the transformation of trypomastigotes into amastigotes. Ssp-4 is an acidic, mannose-containing glycoprotein of 70-84 kDa that is present at the surface of recently transformed amastigotes. I t is anchored to the membrane by glycosyl-phosphatidylinositol and is progressively shed from amastigotes when they are cul- tivated in the absence of mammalian cells. Unfortunately, re- peated attempts to use antibodies against Ssp-4 and against the amastin peptides to test whether the two proteins are the same or share epitopes were not successful. In another report, Pan and McMahon-Pratt (24) described the purification of an amastigote-specific protein of 83 kDa from T cruzi membranes that has some properties in common with Ssp-4. However, the 20 amino acid N-terminal sequence obtained from that protein that was affinity-purified using the monoclonal antibody shows no homology to the amastin sequence. Thus, it is not known if amastin has been previously detected on the surface of amastigotes using immunological techniques. Nevertheless, identification of this family of hydrophobic surface proteins provides the foundation for examining interactions of the in- tracellular amastigote stage of T cruzi with mammalian cells and establishes a system for studying stage-specific gene ex- pression in this organism.

20516 Amastin of T cruzi Acknowledgment-We thank Noma Andrews for the kind gift of

monoclonal antibodies to Ssp-4.

1.

2.

3.

5. 4.

6.

7.

8. 9.

10.

REFERENCES Pereira, M. E. A. (1990) in Modern Parasite Biology (Wyler, D. J., ed) pp. 64-78,

McCabe, R. E., Remington, J. S., and Araujo, F. G. (1984) Infect. Immun. 46,

Ley, V., Andrews, N. W., Rohhins, E. S., and Nussenzweig, V. (1988) J. Exp.

Silveira, F. T., Dias, M. G., Pardal, l? l?, de Oliveira Lohao, A,, and de Britto Silva, L. H., and Nussenzweig, V. (1953) Folia Clin. Bid. 20, 197-207

Kirchhoff, L. V., Hieny, S., Shiver, G. M., Snary, D., and Sher, A. (1984) J.

Williams, D. M., Sawyer, S., and Remington, J. S. (1976) J. Infect. Dis. 134,

de Carvalho, T. U., and de Souza, W. (1983) Z. Parasitenkd. 69, 571-575 Kriegler, M. (1990) Gene Dansfir and ExpressiontA Laboratory Manual, W. H.

Wilson, M. E., Hardin, K. K, and Donelson, J. E. (1989) J. Immunol. 143,

W. H. Freeman and Company, New York

373-376

Med. 168,649-659

Melo, G. (1979) Hileia Med. Belem 1, 61-62

Immunol. 133,2731-2735

610-614

Freeman and Company, New York

678-684

11. Fourney, R. M., Miyakoshi, J., Day, R. S., and Paterson, M. C. (1988)Focu.s 10, 5-7

12. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Leboratory Manual, 2nd ed., Cold Spring Harbor Laboratory, Cold Spring

13. Russell, D. G., Xu, S., and Chakraborty, P. (1992) J. Cell Sci. 103, 1193-1210 Harbor, NY

14. Clayton, C. E., Fueri, J. E, Itzhaki, J. E., Bellofatto, V., Sherman, D. R., Wisdom, G. S . , Vijayasarathy, S., and Mowatt, M. R. (1990) Mol. Cell. Bid. 10, 30363047

15. Villalta, F., and Kierszenbaum, F. (1982) J. Protozool. 29, 570-576 16. Andrews, N. W, Robbins, E. S., Ley, V.. Hong, K. S., and Nussenzweig. V. (1988)

17. de Arruda, M. V., Reinach, F. C., Colli, W., and Zingales, B. (1990) Mol. Bio-

18. Kahn, S., Trenton, G. C., Wallace, J. C., Hoagland, N. A., and Eisen, H. (1991)

19. Schenkman, S., and Eichinger, D. (1993) Parasitol. Today 9,218-222 20. Johnson, P. J., Koot r , J. M., and Borst, P. (1987) Cell 51, 273-281 21. LeBowitz, J. H., Smith, H. Q., Rushe, L., and Beverley, S. M. (1993) Genes &

22. Kapotas, N., and Bellofatto, V. (1993) Nucleic Acids Res. 21, 40674072 23. Schenkman, S., Ferguson, M. A. G., Heise, N., de Almeida, M. L. C., Mortara,

24. Pan, A. A., and McMahon-Pratt, D. (1989) J. Immunol. 143, 1001-1008 R. A., and Yoshida, N. (1993) Mol. Biochem. Parasitol. 59,293-304

25. Pizzi, T. (1957) Inmunologia de la enfermedact de Chagas. Ph.D. Thesis, Uni-

J. Exp. Med. 167, 300-314

chem. Parasitol. 40, 35-42

Proc. Natl. Acad. Sci. U. S. A. 88, 44814485

Deu. 7,99&1007

versidad de Chile, Santiago