In Vitro Selection of RNA Aptamers That Bind to Cell Adhesion Receptors of Trypanosoma cruzi and Inhibit Cell Invasion*

Trypanosoma cruzi causing Chagas' disease needs to invade host cells to complete its life cycle. Macromolecules on host cell surfaces such as laminin, thrombospondin, heparan sulfate, and fibronectin are believed to be important in mediating parasite-host cell adhesions and in the invasion process of the host cell by the parasite. The SELEX technique (systematic evolution of ligands byexponential enrichment) was used to evolve nuclease-resistant RNA ligands (aptamer = to fit) that bind with affinities of 40–400 nm to parasite receptors for the host cell matrix molecules laminin, fibronectin, thrombospondin, and heparan sulfate. After eight consecutive rounds of in vitroselection four classes of RNA aptamers based on structural similarities were isolated and sequenced. All members of each class shared a common sequence motif and competed with the respective host cell matrix molecule that was used for displacement during the selection procedure. RNA pools following seven and eight selection rounds as well as individual aptamers sharing consensus motifs were active in inhibiting invasion of LLC-MK2 monkey kidney cells by T. cruzi in vitro.

Trypanosoma cruzi causing Chagas' disease needs to invade host cells to complete its life cycle. Macromolecules on host cell surfaces such as laminin, thrombospondin, heparan sulfate, and fibronectin are believed to be important in mediating parasite-host cell adhesions and in the invasion process of the host cell by the parasite. The SELEX technique (systematic evolution of ligands by exponential enrichment) was used to evolve nuclease-resistant RNA ligands (aptamer ‫؍‬ to fit) that bind with affinities of 40 -400 nM to parasite receptors for the host cell matrix molecules laminin, fibronectin, thrombospondin, and heparan sulfate. After eight consecutive rounds of in vitro selection four classes of RNA aptamers based on structural similarities were isolated and sequenced. All members of each class shared a common sequence motif and competed with the respective host cell matrix molecule that was used for displacement during the selection procedure. RNA pools following seven and eight selection rounds as well as individual aptamers sharing consensus motifs were active in inhibiting invasion of LLC-MK 2

monkey kidney cells by T. cruzi in vitro.
Trypanosoma cruzi, the protozoan causing Chagas' disease, must invade host cells to complete its biological cycle. The parasite exists in three developmental forms. Epimastigotes extracellularly multiply in the gut of insect vectors. The mammalian stage of the parasite life cycle includes amastigotes, an intracellular form that replicates, an epimastigote-like form and trypomastigotes, released to the extracellular milieu to invade new cells or biting insects, thus propagating the infection (1,2).
To contact and invade a mammalian host cell trypomastigotes must cross a network of extracellular matrix proteins on the surface of cells of the blood vessel walls and target organs. Evidence for receptor-ligand interactions in the early events of cell infection by trypanosomes was collected by several laboratories (3,4). Host cell matrix molecules such as fibronectin (5), laminin (6,7), heparan sulfate (8,9), and proteins of the cy-toskeleton exposed at the cell surface (4) bind to the parasite surface and affect host cell invasion.
Recognizing the importance of host cell-parasite interactions, our laboratory has investigated a 85-kDa glycoprotein, termed Tc85-11 (7) and belonging to a large family of 85,000 molecular mass glycoproteins, that is expressed only on cell surfaces of the infective trypomastigote stage and binds to laminin and cytokeratin-18 on host cells (4,7). Tc85-11 has an extensive sequence homology with the gp-85 glycoprotein family that is related to the sialidase-trans-sialidase supergene family (10 -11) and possesses at least two sequence motifs that have been proved to bind to different proteins of the cell matrix (4). Based on collected evidence that strongly supports a crucial function of receptor-ligand interaction between host cell and parasite enabling T. cruzi to invade cells, an in vitro selection technique was herein employed for the isolation of nucleaseresistant RNA aptamers that compete with host cell matrix molecules for their binding sites on T. cruzi cell surfaces (cf. Ref. 12).
The SELEX 1 method (systematic evolution of ligands by exponential enrichment) (13,14) is an oligonucleotide-based combinatorial library selection procedure that has been used extensively to isolate high affinity ligands that bind with picomolar and low nanomolar affinities to a wide variety of proteins and cell surface epitopes, i.e. CD4 and selectins (15)(16)(17), as well as to complex targets such as red blood cell membranes (18), the membrane-bound nicotinic acetylcholine receptor (19 -20), whole virus particles (21), and live African trypanosomes (22). The incorporation of modified nucleotides into the RNA transcripts, resulting in resistance against nuclease attack, has considerably increased the use of aptamers as probes to inhibit protein functions in in vitro and in vivo assays (cf. Ref. 23).
Nuclease-resistant RNA and DNA aptamers selected to block cell adhesion events involved in disease have gained importance in the last years. The biological activity of such aptamers have been shown in in vitro assays where RNA antagonists inhibited selectin-dependent adhesion (17,24). Wang et al. (25) selected RNA aptamers that bind to infectious human cytomegalovirus and inhibit viral infection in vitro, showing the feasibility of the SELEX technique for the evolution of novel compounds that protect cells against infection by pathogens. Furthermore, combinatorially synthesized nuclease-resistant RNA and DNA ligands are promising candidates for use in pharmaceutical and therapeutic applications (cf. Refs. 26 and 27).
The SELEX technique was used to evolve nuclease-resistant RNA aptamers that block in vitro receptor-ligand interactions between T. cruzi trypomastigotes and epithelial monkey kid-ney LLC-MK 2 host cells and partially inhibit cell invasion by these parasites.

Culture of Trypomastigotes and Epimastigotes of T. cruzi (Strain Y)
Trypomastigotes from T. cruzi were obtained by infection of confluent monkey kidney epithelial cells (LLC-MK 2 ) as described previously (28). LLC-MK 2 cells infected with trypanosomes were grown in Dulbecco's modified Eagle's medium containing 2% of fetal bovine serum at 37°C and 5% CO 2 . The trypomastigotes released from the infected cells after 5-6 days are washed three times in Dulbecco's modified Eagle's medium, purified by differential centrifugation through lymphoprep (Nycomed, Oslo, Norway) and resuspended in phosphate-buffered saline (PBS) containing 20 mM glucose at a density of 5 ϫ 10 6 parasites per microliter.

SELEX for Isolation of Aptamers That Displace Matrix Molecules from Cell Surfaces of T. cruzi Trypomastigotes
Laminin (human placenta) and fibronectin (human plasma) were purchased from Invitrogen and thrombospondin (human platelets) from Sigma. The following oligonucleotide, consisting of 108 nucleotides with a 40-nucleotide randomized region flanked on both sides by constant sequences, was synthesized at a 200 nM scale by Biomol, Berlin, Germany: 5Ј-ACCGAGTCCAGAAGCTTGTAGTACT(N 40 )GCCTAGATGG-AGTTGAATTCTCCCTATAGTGAGTCGTATTAC-3Ј (19). The product of the oligonucleotide synthesis was purified on a denaturing PAGE. To create the initial DNA pool for transcription of partial randomized RNA molecules, 10.8 nmol of purified DNA were amplified in a total volume of 12 ml using 11 rounds of error-prone PCR in the presence of 0.5 mM MnCl 2 and 7 mM MgCl 2 . The primers used for amplification were 5Ј-G-TAATACGACTCACTATAGGGAGAATTCAACTGCCATCTA-3Ј and 5Ј-ACCGAGTCCAGAAGCTTGTAGT-3Ј. The PCR product was gel-purified and ethanol-precipitated. For in vitro selection, 20 g of the synthetic DNA pool were transcribed using 1,000 units of T7 RNA polymerase in the presence of 50 Ci of [␣-32 P]ATP and 1 mM 2Ј-OH-(ATP and GTP) and 3 mM 2Ј-F-(dCTP and dUTP), the latter to provide stability of the transcript against RNase activity (29). The RNA pool was diluted to a final concentration of 50 M in the reaction mixture containing 20 mM glucose and heat denatured and renatured to allow for proper secondary structure formation. T. cruzi trypomastigotes, purified by differential centrifugation, were resuspended in 200 l at an approximate concentration of 500 nM possible binding sites for host cell matrix molecules and incubated for 60 min with 100 l of the RNA pool. T. cruzi trypomastigotes and RNA molecules bound to their surfaces were pelleted by centrifugation for 45 s at 13,000 ϫ g, washed once with 500 l of binding buffer, and incubated for 20 min in 50 l of a solution containing 1 M each laminin and fibronectin and 5 M heparan sulfate and 100 nM thrombospondin. The trypomastigotes were pelleted by centrifugation, and the supernatant containing displaced RNA molecules was collected. The collected eluate was phenol-chloroform-and chloroform-extracted using 1:1 supernatant:solvent volume ratios and ethanol-precipitated (80%) in the presence of 10 g of yeast tRNA.
In round seven of SELEX a preselection procedure was employed to remove RNA molecules from the SELEX pool that also bind to noninfective epimastigote surfaces. The RNA pool (3 nmol) that had been obtained by collection and amplification from SELEX round seven was incubated for 60 min with 3 ϫ 10 8 epimastigotes in a total reaction volume of 300 l. The supernatant containing 1.95 nmol of unbound RNA was collected after centrifugation, phenol-chloroform-extracted, ethanol-precipitated, and used for selection of cell matrix moleculedisplaceable RNA molecules bound to trypomastigotes as described.
Following each round of selection, the ethanol-precipitated RNA was reverse-transcribed using avian myeloblastoma reverse transcriptase and amplified. The obtained DNA pools were transcribed for the next round of selection using an input of 10 g of DNA, 1 mM of 2-OHЈ purines and 3 mM of 2-FЈ pyrimidines, and 500 units of T7 RNA polymerase (Invitrogen). Following transcription, the RNA was phenolchloroform-and chloroform-extracted, ethanol-precipitated, and resuspended in diethylpyrocarbonate water.

Individual Aptamer Characterization
After SELEX cycle 8, the cDNAs of the pools of aptamers that were eluted using laminin, fibronectin, heparan sulfate, or thrombospondin as a displacement agent were cut with EcoRI and HindIII at the constant regions and ligated into pGEM-3Z plasmids (Promega) (19). The recombinant vectors were transformed into Escherichia coli strains JM109 or DH5␣ and plated on selective medium. Colonies were picked, and inserts were determined by automatic sequencing using the ABI PRISM big dye terminator kit (PerkinElmer Life Sciences). The se- quences of the previously random inserts were aligned and searched for sequence similarities. RNA structures were predicted by using the MFOLD program (30) at bioinfo.math.rpi.edu/ϳmfold/RNA/form1.cgi.

Binding Analysis
IC 50 Determinations-The IC 50 value is the concentration at which a competitor displaces 50% of a radioactive ligand from its receptor. One microgram of DNA template of SELEX RNA pool-5, -7, or -8 or individual aptamers containing consensus motifs in 20 l of transcription buffer was transcribed in the presence of 0.5 mM GTP, 1.5 mM 2-FЈ pyrimidines, and 50 Ci of [␣-32 P]ATP and purified on a Spin Column 30 (Sigma). 32 P-Labeled RNA (2 ϫ 10 5 cpm) diluted with unlabeled RNA to a final concentration of 20 nM was incubated for 60 min at room temperature with 5 ϫ 10 7 trypomastigotes (130 g of protein) or with 1 ϫ 10 7 epimastigotes (92 g of protein) in the presence of 0.1 g/l t-RNA and increasing concentrations of fibronectin from 0 to 10 M in 45 l of selection buffer. The reaction mixtures were separated by centrifugation (45 s, 12,000 ϫ g). The pellets containing parasites and RNA bound to their surfaces were washed once with 1 ml of selection buffer and scintillation counted.
Saturation Analyses-For saturation binding analysis, constant amounts (2 ϫ 10 7 or 5 ϫ 10 7 trypomastigotes) were incubated with increasing concentrations of 32 P-labeled SELEX pool-7 or individual cloned RNAs falling into structural classes in the presence of 0.1 g/l yeast t-RNA in 60 l of selection buffer for 60 min at room temperature. Non-specific binding was determined in the presence of 8 M unlabeled RNA. The K d was estimated by fitting the data to the equation where B max is the concentration of binding sites for the RNA molecules.

Inhibition of Cell Invasion
Prior to infection, T. cruzi trypomastigotes were preincubated for 15 min with the aptamer solution containing SELEX RNA pool-0, -7, or -8 or individual pooled RNAs based on their consensus sequences in PBS, 20 mM glucose. LLC-MK 2 cells attached to eight-well dishes were infected in a 100:1 parasites/cell ratio with trypomastigotes in the absence or presence of SELEX RNA pools for 2 h at 37°C. The incubation was terminated by removing the medium containing free trypomastigotes and washing twice with PBS. The non-adherent parasites were removed by 5-min incubation with lymphoprep (Nycomed), followed by two washes with PBS. The cells were incubated with Dulbecco's modified Eagle's medium supplemented with 2% fetal bovine serum for 48 h at 37°C, fixed with 100% methanol, and stained with chromomycin A (Molecular Probes), as described previously (4). All experiments were done in duplicates, and 12 photos of each replicate were made, enabling the counting of infected and non-infected cells in samples containing 100 cells each. The data were compared for statistical significance by the unpaired Student's t test. The viability of cells and trypomastigotes in the presence of aptamer solution was confirmed by staining with trypan blue and microscopic analysis.

Isolation of RNA Aptamers That Bind to Host Cell Matrix
Molecule Receptors on T. cruzi-The SELEX technique has been used to evolve nuclease-resistant RNA ligands (aptamers) that bind with high affinity and specificity to receptors of host cell matrix molecules on T. cruzi. A SELEX protocol was established to enrich RNA molecules that bind specifically to the receptors of laminin, thrombospondin, heparan sulfate, and fibronectin on infective trypomastigote stages of T. cruzi. After five SELEX rounds, RNA molecules that bind with high affinity to trypomastigotes were observed. They also bound to some extent to non-infective epimastigote forms as shown by IC 50 studies using fibronectin as a displacement agent (Fig. 1A). The RNA pool was therefore purified by using a preselection procedure where binders to epimastigotes were discarded, followed by selection of RNA molecules that bound only to trypomastigotes and were displaceable by the above-mentioned cell matrix molecules.
Following this preselection, increasing concentrations of fibronectin only displaced 32 P-labeled SELEX pool-7 RNA from trypomastigotes, but not from epimastigotes, indicating that the selected RNAs bind specifically to cell surfaces of infective forms of T. cruzi (Fig. 1B). The fraction of RNA molecules that was adsorbed by epimastigotes was analyzed by sequencing of individual RNA molecules, and none of the consensus motives of fibronectin, thrombospondin, laminin, and heparan-sulfatedisplaceable RNA aptamers were similar to those that bound to trypomastigotes (data not shown). Saturation analyses revealed that the pool of RNA molecules after seven rounds of in vitro selection bound with a binding affinity (K d ) of 172 nM to trypomastigotes ( Fig. 2A). Since further selection rounds did not result in any improvement of the binding affinity to T. cruzi cell surfaces, as shown by IC 50 studies using fibronectin as a displacement agent (Fig. 1C), the post-8 RNA pool that was selected using fibronectin, laminin, thrombospondin, and heparan sulfate as displacement agents was split and eluted with each individual cell matrix molecule.
Binding Affinities and Structures of Selected Aptamers-Aptamers from the four pools obtained after elution with each individual cell matrix molecule were cloned, isolated, and sequenced. Four classes of aptamers were obtained based on structural similarities (classes I-IV). The consensus region found in aptamers that were obtained by displacement with the same cell matrix molecule was specific and was not observed in aptamers that were isolated using another cell matrix molecule (Table I). Twenty-four individual fibronectin-displaceable sequences were cloned and sequenced. Five aptamer sequences contained a consensus motif. Fibronectin clone 4 was represented eight times in the sequenced pool. In 20 individual thrombospondin-displaceable, 28 heparan sulfate-displaceable, and 24 laminin-displaceable RNA aptamers, 8, 13, and 8 sequences were found that contained a respective consensus region (Table I). Based on sequence similarities, the respective individual clones that are displaceable by each matrix molecule employed were pooled and tested for binding affinity and biological activity.
The MFold program (30) was used to predict the secondary structure of individual RNA aptamers from classes I-IV. Interestingly, the structure prediction program integrated the consensus sequence of aptamers from all four classes within stemloop structures (Fig. 3), suggesting that truncated aptamers containing only this stem-loop region may bind to their targets and be biologically active.  individual class containing a consensus motif in the previous random region were pooled and tested for their binding affinity toward infective trypomastigotes. Saturation analyses of the binding of consensus aptamers from classes I-IV (see Table I and Fig. 3) to the corresponding matrix receptors on the parasite surface gave the following binding affinities: 124 nM (fibronectin, Fig. 2B), 400 nM (thrombospondin, Fig. 2C), 40 nM (heparan sulfate, Fig. 2D), and 209 nM (laminin, Fig. 2E). The selected aptamers recognized specifically their targets on trypomastigote surfaces and were not toxic to parasites nor reduced parasite motility. Neither the RNA mixture of SELEX-8 nor individual aptamers falling into structural classes bound specifically to cell membranes of LLC-MK 2 cells and HeLa human fibroblasts (not shown).

Binding Specificity and Affinity of Classes I-IV to Their Receptors on Trypomastigote
Do the Selected Aptamers Inhibit Cell Invasion by T. cruzi?-Invasion of LLC-MK 2 cells was tested in vitro in the presence of increasing amounts of trypomastigotes to determine the optimal ratio of parasites over host cells (Fig. 4). A 100:1 ratio of trypomastigotes over cells and an incubation time of 2 h in the presence and absence of RNA aptamers was a condition that gave a high enough infection to differentiate between infection rates with confidence. The effect of SELEX round 8 RNA pool was dose-dependent as shown in Fig. 5. RNA pools from SELEX rounds 0, 7, and 8 as well as individual aptamers from classes I-IV that were respectively displaceable by each matrix component were tested in this assay. A representative invasion experiment is shown in Fig. 6. The pools following SELEX rounds 7 and 8 as well as fibronectin-, thrombospondin-, heparan sulfate-, and laminin-displaceable aptamers (classes that were pooled based on consensus sequences) inhibited at a concentration of 1 M the invasion of LLC-MK 2 cells by 50 -70% ( Fig. 6; Table II), with thrombospondin-displaceable aptamers being the less effective. Interestingly, the presence of these selected aptamers, which were non-toxic to the cells, also decreased the absolute number of parasites per cell ( Fig. 6; Table  II). A mixture of structural classes such as laminin-and fibronectin-displaceable consensus aptamers did not increase inhibition of cell invasion by T. cruzi. In addition, cell invasion could not be totally aborted by the mixture of RNA aptamers following eight cycles of in vitro selection. These results suggest that the interaction of trypanosomes with their host cells also depends on further receptor-ligand interactions that are not affected by the selected aptamers. Nevertheless, the biological activity of the selected RNA aptamers in inhibiting cell invasion by T. cruzi provides direct evidence that receptor-ligand interactions between the host cell matrix molecules fibronectin, laminin, heparan sulfate, and thrombospondin with their receptors on cell surfaces of trypanosomes are events that concur for the successful invasion of the host cell by the parasite. DISCUSSION The present work demonstrates the importance of the interaction between host cell matrix molecules such as fibronectin, laminin, thrombospondin, and heparan sulfate with their respective host cell receptors on infective trypomastigote forms of T. cruzi. Thrombospondin is a family of secreted proteins widely distributed in the extracellular matrices of several tissues. Whereas fibronectin (5), laminin (6,7) and heparan sulfate (8,9) are known ligands for T. cruzi in the mammalian host, the involvement of thrombospondin in T. cruzi invasion is subjected to investigation. Invasion of Plasmodia is mediated by a thrombospondin-related adhesive protein (TRAP) (31), and thrombospondin is involved somehow with tumor progression (32,33). Notwithstanding, present evidence favors fibronec- tin, laminin, and heparan sulfate as more important molecules in host cell invasion by T. cruzi than thrombospondin.
It is known that trypomastigotes shed molecules bound to their cell surface (34), which may explain quantitative differences between binding affinity and biological effect in vitro. The failure of the selected RNA aptamers, even in higher concentrations, to inhibit completely the invasion of the host cell by the parasite suggests that T. cruzi may use several different mechanisms to invade host cells. This is expected due to the large repertoire of molecules described in the literature as involved in host cell invasion by the parasite (3)(4)(5)(6)(7)(8)(9)(10)(11). In addition to the interaction of the parasite membrane with cell matrix components, parasite proteases (35)(36), and host cell surface proteins such as bradykinin B2 (36), TGF-␤ (37), or mannose (38) receptors have been shown to participate in host cell infection, some of them preparing the host cell for invasion. As T. cruzi invades nearly every cell type except erythrocytes, it may express a variety of receptors and ligands to ensure its interaction with different pattern of molecules expressed on the surface of the respective cell type.
Thus, the SELEX method opens new avenues to block pathogen-host cell interactions. Considering that the 2Ј-hydroxyl group of the ribose is substituted by fluorine, the RNA aptamers are less amenable to RNase attack and, therefore, are ideal for in vivo experiments. Thus, stable oligonucleotides that may inhibit parasite-host cell interaction could be used to develop administrable drugs that would hopefully halt the progress of disease.
Combinatorial library approaches have been successfully employed to block ligand-receptor interactions that are necessary for completion of the life cycle of other parasites such as Plasmodium. Selection in vitro of peptide ligands from a combinatorial phage display library that block adhesion and invasion of insect midgut cells by Plasmodium was described previously (39). Combinatorial libraries may be advantageous for drug development when compared with the rational drug design approach, as the selection of aptamers, for instance, do not require a full understanding of infection by pathogens.
The development of agents that block adhesion events between parasites and host cells have not only become important regarding protection of cells against invasion by pathogens but also to prevent metastatic cancer cells from adhering to endothelial cells of blood vessels and invading the respective tissues. Interestingly the same host cell matrix molecules involved in cell invasion by T. cruzi as shown in this study are used as adhesion ligands by tumor cells (40,41).
Recently, the in vivo use of aptamers is providing more consistent results by improving aptamer pharmacokinetics. Keeping intact aptamers in the blood from hours to days has been accomplished by conjugation with high molecular weight carriers, such as polyethyleneglycol, or by embedding into liposomes (42). As an example, one aptamer to vascular endothelial growth factor (VEGF), patented as NX 1838 is currently been used in clinical trials as a potential therapeutic agent for agerelated macular degeneration (27). Several aptamers that inhibit adhesion events have been shown to be effective in animal models such as those directed against VEGF (43), P-selectin (17) and PDGF (44 -45). Thus, it should be feasible to find active aptamers against in vivo infection by T. cruzi that may open new perspectives to future approaches in the treatment of Chagas' disease.

TABLE II Selected aptamers inhibit the in vitro invasion of LLC-MK 2 cells by T. cruzi
The percentage of infected cells from 2,000 cells as well as the number of parasites per 100 cells and the number of parasites per invaded cells were determined, and results were compared with the control for statistical significance using the unpaired Student's t test. Results are the average of triplicate determinations Ϯ S.D. In all three columns SELEX 0 RNA and SELEX 7 RNA do not differ from each other within a significance level of p Ͻ 0.05. All other conditions were p Ͻ 0.005, with the exception of the earmarked results of the third column. Infection rates in the control experiments made in the absence of RNA aptamers were normalized to 100%. For each condition the concentration of aptamer added was 1 M. Infection and staining were performed as described in the legend of Fig. 6