2006–2014 Nucleic Acids Research, 1999, Vol. 27, No.
9  1999 Oxford University Press
Combinatorial selection of high affinity RNA ligands to
live African trypanosomes
Matthias Homann and H. Ulrich Göringer*
Laboratorium für Molekulare Biologie, Genzentrum der Ludwig-Maximilians-Universität München, Am Klopferspitz 18a,
82152 Martinsried, Germany
Received January 14, 1999; Revised and Accepted March 3, 1999
ABSTRACT
African trypanosomiasis is a parasitic disease caused
by a specific class of protozoan organisms. The best-
studied representative of that group is Trypanosoma
brucei which is transmitted by tsetse flies and multi-
plies in the blood of many mammals. Trypanosomes
evade the immune system by altering their surface
structure which is dominated by a layer of a variant
surface glycoprotein (VSG). Although invariant surface
proteins exist, they are inaccessible to the humoral
immune response. Using a combinatorial selection
method in conjunction with live trypanosomes as the
binding target, we show that short RNA ligands
(aptamers) for constant surface components can be
isolated. We describe the selection of three classes of
RNA aptamers that crosslink to a single 42 kDa protein
located within the flagellar pocket of the parasite. The
RNAs associate rapidly and with high affinity. They do
not discriminate between two different trypanosome
VSG variant strains and, furthermore, are able to bind
to other trypanosome strains not used in the selection
protocol. Thus, the aptamers have the potential to
function as markers on the surface of the extracellular
parasite and as such they might be modified to function
as novel drugs against African trypanosomiasis.
INTRODUCTION
African trypanosomes are unicellular, uniflagellated protozoan
parasites. They cause African trypanosomiasis or sleeping
sickness, a chronic disease in humans as well as in wild and
domestic animals. An estimated 50 000 000 people world wide
are at risk of such an infection, and Smith et al. (1) recently
estimated a number of about 300 000 newly infected cases/year.
Moreover, the disease in domestic animals considerably limits
agricultural development in a substantial part of Africa, thus
indirectly affecting an even larger number of people (2,3). None
of the available therapeutic measures against the various diseases
is very effective and therefore novel approaches are needed in the
development of drugs against trypanosome infections (4).
African trypanosomes are transmitted by tsetse flies, and as
extracellular parasites they multiply within the peripheral blood,
*To whom correspondence should be addressed. Tel: +49 89 8578 2475; Fax: +49 89 8578 3810; Email: goeringe@biochem.mpg.de
the capillary beds and within the tissue fluids of the infected hosts.
During this bloodstream life cycle stage, trypanosomes are
covered with a layer of approximately 10 000 000 molecules of
a glycoprotein species known as variant surface glycoprotein
(VSG). VSG molecules have a molecular mass of ∼60 kDa, they
homodimerise and are glycosylphosphatidylinositol-anchored
within the plasma membrane. The VSG surface induces a strong
T-cell-independent IgM response as well as a T-cell-dependent
B-cell response which elicits VSG-specific IgG (5; reviewed in
6). The parasites, however, evade the host immune system by
temporarily expressing different VSG variants (7–9). This
phenomenon has been termed antigenic variation and has its
molecular basis in the surface presentation of structurally
polymorphic N-terminal domains of the different VSG variants.
Though the C-termini of the various VSG molecules are
structurally conserved, they are buried within the VSG layer and
thus not accessible to the humoral immune response (10–12). The
trypanosome genome encodes a repertoire of about 1000 different
vsg genes, but only one VSG is expressed at a given time. The
switching frequency from one variant to the next has been
estimated to range from 10–2 to 10–7/cell generation and seems
to be a stochastic event (13,14). Thus, the VSG surface acts as an
exclusion barrier for larger molecules, such as antibodies, while
its variable characteristics cause the inability of the infected host
to clear the infection. Besides, the surface changes complicate any
vaccination approach.
Although the entire surface of bloodstream stage African
trypanosomes is covered by the VSG protein, the coat nevertheless
shows a highly dynamic structure. VSG molecules show rapid
lateral movement within the cell membrane (15). In addition,
other surface components have been identified, including invariant
surface glycoproteins (ISGs), receptor complexes and transporter
molecules (16–20). Some of these molecules are embedded
within the VSG layer and distributed over the entire surface of the
trypanosome cell body including the flagellum. Other surface
components are localised to the flagellar pocket, an invagination
of the plasma membrane around the base of the flagellum which
functions as an exo- and endocytosis site (reviewed in 21).
Moreover, the trypanosome surface is not an impenetrable casing.
Molecules of lower molecular mass, such as the protease trypsin
(23 kDa), have been shown to cleave VSG molecules at a site
deep within the protein layer (17,22). This suggests the existence

of transient molecular cavities within the layer of the VSG
homodimers.
In this study, we address the potential of tagging the surface of
live bloodstream stage trypanosomes with molecules that bind with
high affinity and specificity. If such molecules were small enough
to reach and interact with invariant elements of the cell surface,
binding would be independent of the expressed VSG and could be
used to direct the immune response to the trypanosome surface. In
vitro evolution methods in conjunction with combinatorial nucleic
acid libraries offer effective strategies for selecting nucleic acid
ligands to biological targets. The SELEX (systematic evolution
of ligands by exponential enrichment) protocol (23,24) has been
widely used to isolate high affinity binders, so-called aptamers, to a
variety of biological targets (reviewed in 25) including complex
targets such as the membranes of human red blood cells (26).
Using this methodology, we report the isolation of three classes
of RNA aptamers that bind to live African trypanosomes. The
binding is specific for the infective bloodstream life cycle stage
of the parasites, because insect stage trypanosomes are not
recognised by the selected RNAs. Individual aptamers are able to
bind rapidly to the parasite surface with high affinity and
specificity for a single protein located within the flagellar pocket.
Although two separate selections with two cloned trypanosome
strain variants were performed, the selected aptamers show no
variant-specific features and furthermore are capable of binding
other, unrelated trypanosome strains.
MATERIALS AND METHODS
Trypanosomes
The bloodstream life cycle stage of Trypanosoma brucei was
grown in HMI-9 medium (27) supplemented with 10% (v/v)
heat-inactivated bovine foetal calf serum (FCS). Incubation was
at 37
C in 5% CO2/95% air. The following strains were used:
T.brucei 427-MITat serodeme, variant clones MITat 1.2 and MITat
1.4 (28); AnTat 1.1 (29); ILTAT 1.1 (30). Insect stage trypanosomes
were grown at 27
C in SDM-79 medium supplemented with 10%
(v/v) FCS (31).
Oligodeoxynucleotide and RNA pool synthesis
DNA oligonucleotides were synthesised by automated solid support
chemistry using O-cyanoethyl-N,N-diisopropyl-phosphoramidites.
The starting DNA library was a 79mer of sequence GAAT-
TCAGTCGGACAGCG(N)40GATGGACGAATATCGTCTCCC
which contained a central sequence stretch of 40 randomised
nucleotides. The DNA was purified by reverse phase liquid
chromatography and 1.1 mg were amplified during seven PCR
cycles in a total volume of 60 ml using primers A (GC-
GAAGCTTTAATACGACTCACTATAGGGAGACGATATTC-
GTCCATC) and B (GGTCGAGAATTCAGTCGGACAGCG).
Of the 7.5 mg of double-stranded PCR product, 1.5 mg were
transcribed into 2.5 mg RNA (reaction volume 7.5 ml) from
which 1.15 mg were used in the first round of each selection. All
subsequent RNA pools were transcribed from 50–150 µg of the
PCR-generated double-stranded DNA templates in a final
volume of 1–1.5 ml. Reactions were performed in the presence of
5 µCi [α-32P]UTP (3000 Ci/mmol) in 20 mM NaxHyPO4, pH 7.9,
8 mM MgCl2, 20 mM DTT, 4 mM spermidine, containing 1 mM
each NTP and 500 U T7 RNA polymerase. The synthesised
RNAs were extracted with phenol, precipitated with ethanol and
2007
Nucleic Acids
Nucleic Acids Research,
Research,1994,
1999,Vol.
Vol.22,
27,No.
No.19 2007
redissolved in a buffer containing 20 mM NaxHyPO4, pH 7.4,
2 mM MgCl2, 130 mM NaCl, 5 mM KCl, 20 mM glucose and
0.2 mM β-mercaptoethanol. A final purification was achieved by
size exclusion chromatography on Sephadex G50 columns using
the same buffer. Before each round of selection, all RNAs were
pre-incubated at 37
C for 30 min.
In vitro selection
Trypanosomes were harvested at a cell density of 1–2 × 106 cells/ml
and extensively washed in binding buffer (20 mM NaxHyPO4,
pH 7.4, 2 mM MgCl2, 130 mM NaCl, 5 mM KCl, 20 mM glucose,
0.2 mM β-mercaptoethanol). RNA binding was performed in a final
volume of 0.5–1.5 ml with 32P-labelled pool RNA (10–20 µM)
and either 5 × 108 (rounds 1–7) or 2 × 108 cells/ml (rounds 8–13).
After an incubation for 60 min at 30
C, unbound and weakly
associated RNAs were washed off by four consecutive washes
with 2 ml binding buffer. Trypanosomes and bound RNAs were
collected by centrifugation and the bound RNAs recovered by
two phenol extractions followed by precipitation with ethanol.
Isolated RNA was reverse transcribed (150 µl reaction volume,
10 µg primer B, 100 U M-MuLV RT), amplified by PCR and the
resulting DNA templates were transcribed into RNA which was
used for the next round of selection. DNA templates from pool 12
were digested with HindIII and EcoRI and cloned into pUC118.
Sequences of 53 clones of each selection were determined by
dideoxy terminator sequencing.
Determination of binding parameters
Uniformly 32P-labelled pool RNAs (5 nM, ∼20 000 c.p.m.) or
RNA aptamers (3–5 nM, 10 000–50 000 c.p.m.) were incubated
with T.brucei cells in 100 µl binding buffer at cell densities of
1–3 × 108/ml for bloodstream stage cells and 5 × 109/ml for insect
stage trypanosomes. After a 30 min incubation at 37
C, the cells
were washed (four times with 2 ml binding buffer), collected by
centrifugation and the percentage of bound RNA was determined
by scintillation counting. Average relative binding efficiencies
were calculated from three independent experiments and were
normalised to the binding efficiency of pool 1 RNA. Equilibrium
dissociation constants were derived from experiments determining
the percentage of bound RNA at varying trypanosome cell
numbers. Kd values were calculated based on the Scatchard
equation r/[A] = n/Kd – r/Kd with the slope of the plot
representing –1/Kd; where A is the free RNA, r = B/[cells] and B
is bound RNA. The number of binding sites on the trypanosome
surface (n) was derived from the x-intercept of the Scatchard plot.
Association rates were determined with uniformly 32P-labelled
RNA aptamer 2-16 (0.75 nM, ∼50 000 c.p.m.) and varying
amounts of T.brucei cells in 50 µl binding buffer at 37
C. Parasite
cell densities varied between 2.5 × 107 and 2 × 108/ml
corresponding to calculated target concentrations of 3–24 nM
(assuming 72 000 targets/cell). Under the conditions of a large
excess of one binding partner, a pseudo-first-order rate constant
kobs can be determined from kobs = ln 2/t
. RNA aptamers and
cells were pre-incubated at 37
C and the binding reaction was
started by adding cells to the RNA. Aliquots were taken at time
points between 0.5 and 30 min, diluted 1:500 in binding buffer
and centrifuged immediately. The percentage of bound and
unbound RNA was determined for each time point by scintillation
counting and plotted against time.
2008 Nucleic Acids Research, 1999, Vol. 27, No. 9
Run-off transcription, 5′-end-labelling and biotinylation
RNA aptamers were transcribed from linearised plasmid DNA
templates using T7 RNA polymerase following standard pro-
cedures. DNA templates (1 µg) were transcribed in a buffer
containing 20 mM NaxHyPO4, pH 7.9, 8 mM MgCl2, 20 mM
DTT, 4 mM spermidine and 1 mM each NTP. Radioactive
labelling of the RNAs was achieved by using [α-32P]UTP
(10–50 µCi) and 5 µM UTP in the transcription reaction. The
DNA was digested with DNase I and RNA transcripts were
phenol extracted and purified from non-incorporated NTPs by
size exclusion chromatography. Aptamer RNAs were 5′-end-
labelled using alkaline phosphatase and T4 polynucleotide kinase
and [γ-32P]ATP according to standard procedures.
Fluorescence labelling of aptamer 2-16 was achieved by
tagging the BODIPY TMR-C5 fluorophore (Molecular Probes)
to the 5′-end of the RNA. The reaction was performed according
to the manufacturer’s specification and fluorophore-conjugated
2-16 RNA was dissolved in a PBS-based buffer at 0.4 µg/µl.
Biotinylation of aptamer 2-16 was performed by in vitro
transcription in a reaction mixture containing 1 mM each GTP,
CTP and ATP, 0.05 mM UTP and 5–50 µM biotin-16-UTP
(Boehringer Mannheim). RNA transcripts were separated from
non-incorporated nucleotides by size exclusion chromatography,
precipitated with ethanol and dissolved in a PBS-based binding
buffer at 20 ng/µl.
RNA structure determination
Theoretical secondary structures were calculated using the
MFold subroutine of the GCG software package based on a free
energy minimisation algorithm (Wisconsin Package v.9.1; Genetics
Computer Group, Madison, WI). Calculations were performed
for 37
C, which is the optimal growth temperature for blood-
stream stage trypanosomes. Structure-specific enzymatic probing
experiments were performed at limiting enzyme concentrations
with cobra venom RNase V1 (0.25 U/ml), RNase T1 (250 U/ml)
and RNase A (10 mU/ml). Cleavage reactions were done with
0.3 µg aptamer RNA at 25
C in 20 mM NaxHyPO4, pH 7.4,
2 mM MgCl2, 130 mM NaCl, 5 mM KCl. Samples were analysed
on denaturing 12% (w/v) polyacrylamide gels. CD spectra and
UV melting curves were recorded as described (32).
Photo-crosslinking of RNA aptamers to living cells
Uniformly 32P-labelled RNA aptamers (5–10 nM) were incubated
with T.brucei cells (2–8 × 108/ml) in binding buffer at 37
C for
20 min. Following incubation, the cells were put on ice and
irradiated with UV light (254 nm) for 10 min at an energy dose
of 0.12 J/cm2. The samples were heat denatured (95
C, 5 min),
digested with a cocktail of RNase A, RNase T1 and DNase I
(30 min, 37
C) and finally analysed by discontinuous SDS–
PAGE. Gels were stained with Coomassie brilliant blue, dried and
the 32P-labelled proteins were visualised by autoradiography.
Fluorescence in situ hybridisation
Bloodstream stage trypanosomes were harvested at a cell density
of 2 × 106/ml and washed in binding buffer (four times). Samples
of 2 × 107 parasites were incubated at 37
C with either
biotinylated (2 ng/µl) or 5′-BODIPY TMR-C5-labelled aptamer
RNA (10 ng/µl) in 40 µl binding buffer. After 20 min the cells
were washed (twice with 2 ml binding buffer) to remove unbound
RNA and put on ice. Cells were smeared onto micoscope slides,
briefly air dried and mounted in ProLong Antifade (Molecular
Probes). In the case of biotinylated aptamer preparations, the
parasites were further incubated with a 1:100 dilution of a
monoclonal Cy3-conjugated anti-biotin antibody (Dianova).
Nuclear and kinetoplast DNAs were stained with Hoechst
H-33342 (Molecular Probes). Slides were examined with a digital
fluorescense microscope. Images were captured and merged
using IP Lab Spectrum 3.1.2c (Scanalytics Inc.). Assembly into
figures and false colouring were performed using Adobe
Photoshop 3.0.3 (Adobe Systems Inc.).
RESULTS
Selection of RNAs that bind to live trypanosomes
For the selection we used a starting pool of 1016 RNA molecules
with an estimated complexity of 2 × 1015 unique sequences. The
library contained 40 randomised nucleotides flanked by primer
binding sites 21 and 24 nt in length, which together restricted the
final molecular mass of the selected aptamers to a value around
25–27 kDa. In the case of the protease trypsin, this has been
shown to be a molecular size still able to penetrate the VSG layer
(22). Two independent selections were performed in parallel
using two different T.brucei variant antigen clones as target cells:
MITat 1.2, a cell line which stably expresses VSG 221, and MITat
1.4, which express VSG 117 on their surface (28). These two
variant strains were further chosen because of steric differences
in the VSG packing on the parasite surface: they respond
differently when incubated with an anti-ISG75 antibody or with
the protease trypsin (17). Lastly, the crystal structure for the
N-terminal domain of the MITat 1.2 VSG is known to a resolution
of 2.9 Å (33).
Both parasite variants were grown as bloodstream stage
trypanosomes and RNA binding was accomplished over a 60 min
incubation period. Despite some RNA degradation, a significant
amount of full-length RNAs remained intact after the incubation
(Fig. 1A) which was processed further. In both experiments,
13 rounds of selective binding and subsequent amplification were
performed. Maximal enrichment of trypanosome-interacting
RNAs was achieved in both selections in round 12 at which time
∼18% of the input RNAs was associated with the parasite cells
(Fig. 1B). Confirmation of a reduction in the pool complexities
with increasing rounds of selection was derived from analytical
RNase T1 digests of the different 32P-labelled RNA pools.
Starting at rounds 8 and 9, distinct RNase cleavage products could
be resolved electrophoretically, indicating the selective enrichment
of defined RNA sequences (Fig. 1C).
DNA templates from RNA pools of round 12 were cloned and
the sequences of 53 clones from each selection experiment were
determined. A majority of these sequences can be grouped into
two families (I and II) based on common sequence motifs
(Fig. 2). Family I members are characterised by an invariant CGT
and a conserved NGC base triplet which are flanked by sequences
that show a high degree of co-variation, suggesting base pairing.
Two of the family I sequences were found in both selections. The
first one is represented by the identical aptamers 2-11 and 4-2,
accounting for ∼50% of all clones in both selections. The second
sequence (aptamers 2-16 and 4-10) appeared at a frequency of
24% in the selection using MITat 1.2 trypanosomes and at a
frequency of 7% in the selection with MITat 1.4 parasites. Thus,
 2009

Nucleic Acids
Nucleic Acids Research,
Research,1994,
1999,Vol.
Vol.22,
27,No.
No.19 2009

Figure 1. Selective enrichment of RNA ligands that interact with live trypanosomes. (A) Degradation kinetics of 32P-labelled pool 1 RNA (85 nt) upon incubation
with either MITat 1.2 or MITat 1.4 trypanosomes. (B) Graphical representation of the percentage of trypanosome-bound RNAs in each round of selection using MITat
1.2 or MITat 1.4 cells. Binding was performed at 30
C using 5 nM RNA and a cell density of 3 × 108 cells/ml. (C) Analytical RNase T1 digestion of different RNA
pools from both selections. Uniformly 32P-labelled aptamer pools were digested with RNase T1 at limiting conditions and the resulting hydrolysis products were
separated in denaturing polyacrylamide gels and analysed by autoradiography. The appearance of distinct cleavage products (arrowheads) indicate a reduction in the
pool complexity and thus an enrichment of a few selected sequences. The very similar digestion pattern in both selection experiments indicates the presence of similar
RNA sequences.

almost 70% of the sequences were identical in both selections.
The remaining family I sequences are unique to each selection
and were either found in several identical isolates or as single
sequences. Family II (Fig. 2) contains only five members; all are
characterised by four repeats of a GGGN sequence motif.
Aptamer 4-30 of that group was found in three identical clones,
while all other clones were unique. Six sequences could not be
assigned to either of the two families and were grouped into a
separate set (family III) (Fig. 2).
Binding efficiencies of the RNA aptamers
Relative binding efficiencies for all aptamers were determined
using both parasite variant strains as targets (Fig. 2). Each
aptamer RNA shows an increased relative binding capacity
compared to the starting randomised RNA pool. The values range
between 2 and 40%, which corresponds to a 12- to 280-fold
increase in the relative binding efficiency. Thus, weak as well as
strong binders were selected in the two experiments. A more
thorough inspection shows that the family I aptamers can be
divided into a high affinity group which binds 150- to 280-fold
stronger to either MITat 1.2 or MITat 1.4 cells and a group of
RNAs with only moderate affinity (14- to 50-fold stronger
binding). The family III RNAs also contain weak and strong
binders, while the aptamers of family II show intermediate
binding. The relative binding efficiencies range between a 50- to
90-fold increase to MITat 1.2 cells and a 50- to 180-fold stronger
binding to MITat 1.4 cells. None of the RNAs is capable of
discriminating between MITat 1.2 and MITat 1.4 cells, regardless
of which parasite variant was used in the selection protocol. This
indicates either that the RNAs interact with the structurally
constant domains of the two different VSGs or that identical
surface targets were selected in both SELEX experiments. In
contrast, all RNAs are incapable of recognising the surface of
insect stage trypanosomes (binding efficiencies ≤0.1%), indicating
a bloodstream stage specificity of the aptamer/trypanosome
interaction.
RNA structure determination
To gain insight into the folding details of the three aptamer
families, we computed energy-minimised secondary structures
for all RNA sequences. In addition, selected aptamers were
experimentally probed with single- and double-strand-specific
RNases and analysed by CD spectroscopy and UV melting. The
combination of all data resulted in a secondary structure model for
the family I RNAs consisting of a hairpin loop as shown in
Figure 3A (left). On one side, the main helical element of the
hairpin is interrupted by the invariant CGU triplet, which forms
a bulge that protrudes from the helical geometry. The conserved
NGC sequence is located within the apical loop sequence of the
hairpin. Additionally, a pseudoknotted structure can be formed
involving a sequence stretch of the 3′ primer binding sequence
(PBS) (Fig. 3A, right). In the case of aptamer 2-16, the
2010 Nucleic Acids Research, 1999, Vol. 27, No. 9

Figure 2. DNA sequences and relative binding efficiencies of all aptamers from both selection experiments. The name of each aptamer is followed by its clone
abundance (in parentheses) and the primary sequence of the randomised region. Base complementary sequences are underlined and co-varying nucleotide positions
are indicated as capital letters in bold. Relative binding efficiencies were measured for MITat 1.2 and MITat 1.4 cells and were normalised to the binding efficiency
of the starting RNA pool, which was set to a value of 1. Standard deviations were estimated from three independent binding experiments. Members of family I (92 out
of 106 clones) contain an invariant CGT and a conserved NGC triplet (shaded boxes). Family II members (seven out of 106 clones) contain four repeats of a GGGN
box (shaded in grey) located at the 3′-end of the randomised region. Family III members (seven out of 106 clones) show no similarity to either of the two families;
three RNAs, however, contain stretches with similar sequences (shaded boxes). Capital letters indicate nucleotides from the 5′ and 3′ PBS involved in base pairing.
aThe sequence includes five clones with one or two base changes. bThe sequence includes one clone with a single base change.

Figure 3. Structure models for family I (A) and family II aptamers (B) derived from sequence comparison and enzymatic probing data of 5′-32P-labelled 2-16 RNA
(family I) or 5′-32P-labelled 4-29 aptamer (family II). Sequences derived from the constant 5′ and 3′ primer binding sites (PBS) are represented as shaded lines.
Nucleotide accessibilities for the various enzymes used for the structure probing are given in the legend below the drawing. For the family I aptamers a pseudoknotted
structure can be formed (A, right) involving a sequence stretch of the 3′ PBS which is supported by several of the RNase V1 hydrolysis sites.

pseudoknotted structure is supported by prominent RNase V1
sites at positions A70, G69, U64, G44 and G43 and the ability to
form the pseudoknot seems to distinguish strong family I
aptamers from weak binders. Lastly, multiple transitions in UV
melting experiments suggest additional higher order foldings
(data not shown). CD measurements verified the typical A-form
characteristics for the helical elements: a strong positive elipticity
around 270 nm and negative elipticities at 240 and 210 nm (data
not shown).
The repetitive GGGN sequence motifs of the family II
aptamers are likely to fold into a G-quartet structure (34,35) as
shown in Figure 3B. The four GGGN elements show different

Figure 4. Photo-crosslinking and quantitative characterisation of the binding reaction of aptamer 2-16. (A) Gel separation of UV-crosslinking products of uniformly
32P-labelled aptamer 2-16 with MITat 1.2 cells in the absence (– comp.) or presence (+ comp.) of a 100-fold molar excess of unlabelled 2-16 RNA or tRNA (+ tRNA).
(B) (Left) UV-crosslinking of aptamer 2-16 to bloodstream form (BF) MITat 1.2 and MITat 1.4 cells or to insect stage cells (procyclic cell form, PCF). (Right)
Coomassie-stained SDS–PAGE gel showing the separation of a trypanosome whole cell lysate and purified VSG protein. Molecular masses are given in kDa.
sensitivities to RNase T1 with the third GGGN motif being
insensitive to the nuclease. As in the case of the family I RNAs,
nucleotides of the 3′ PBS were identified as contributing to the
folding of the RNAs by base pairing with a sequence stretch at the
5′-end of the randomised region. The formed helix extends the
structural arrangement of the G-quartet. No consensus structure
could be derived for the aptamers grouped into family III,
although three of the RNAs contain short stretches with identical
sequence (Fig. 2). Lastly, in contrast to the aptamers of family I,
no structural differences between the strong and weak binders of
families II and III could be identified.
Identification and localisation of the binding target(s)
To identify the binding target(s) on the parasite surface, we
performed zero distance photo-crosslinking experiments. A
representative set of results is shown in Figure 4A using the
strongest binding RNA, aptamer 2-16, as the ligand. Two
crosslinking products with apparent molecular masses of 42 and
29 kDa were identified. The absence of any UV irradiation or
digestion with proteinase K (data not shown) after the photo-
crosslinking reaction completely abolished the formation of the
two products, thus identifying them as RNA–protein complexes.
No radioactive signal with the apparent molecular size of VSG
molecules was detected (∼60 kDa), which indicated that the VSG
protein is not the target for the binding of the aptamer RNAs.
The specificity of the two crosslinks was tested in competition
experiments with homologous and heterologous RNA competitors
(Fig. 4A). Crosslinking to the 42 kDa protein was abolished in the
presence of a 100-fold molar excess of unlabelled 2-16 RNA,
whereas a 100-fold molar excess of tRNA had no effect. This
indicates a specific aptamer–protein interaction for the 42 kDa
protein. In contrast, the 29 kDa band shows several unusual
features. First, it cannot be competed away with either competitor
RNA (Fig. 4A). Second, it was only found at lower parasite cell
densities ≤108 cells/ml (compare Fig. 4A with B). Third, no
crosslinking was achieved at incubation intervals ≥20 min (data
not shown). Since the SELEX cycles involved 60 min incubations
2011
Nucleic Acids
Nucleic Acids Research,
Research,1994,
1999,Vol.
Vol.22,
27,No.
No.19 2011
Figure 5. Trypanosome-bound aptamer 2-16 localises to the flagellar
pocket. In situ fluorescence microscopy of bloodstream stage trypanosomes
(MITat 1.4) incubated with biotinylated 2-16 RNA and immunodecorated with
a Cy3-conjugated anti-biotin antibody (yellow). Counter-staining with
Hoechst H-33342 (pink) indicates the nucleus and kinetoplast. Phase contrast
images of two trypanosome cells in different orientations are shown, merged
with the fluorescence images in false colour. Arrowheads indicate the punctated
staining due to binding of the aptamer. The area coincides with the flagellar
pocket of the cells as indicated in the line drawing to the right.
of the RNA pools with the parasite cells, the 29 kDa protein
cannot be the specific aptamer binding target.
As expected from the binding data, aptamer 2-16 could be
crosslinked to the same 42 kDa target in both MITat 1.2 and MITat
1.4 cells (Fig. 4B). No crosslinking was achieved with insect
stage parasites, again demonstrating the specificity of the RNA
binding reaction for the bloodstream life cycle stage. Weak
binding aptamers could be crosslinked to the same 42 kDa
polypeptide, although with lower efficiency (data not shown).
To localize the 42 kDa binding target on the trypanosome
surface we monitored the binding of aptamer 2-16 by in situ
fluorescence microscopy. Direct staining with a fluorophore-
conjugated 2-16 RNA preparation as well as indirect staining
with a fluorophore-coupled antibody against biotinylated 2-16
RNA gave identical results. The bound RNA localised to a
punctated area proximal to the kinetoplast (Fig. 5). This likely
represents the flagellar pocket, the main endo- and exocytosis site
of the parasite.
2012 Nucleic Acids Research, 1999, Vol. 27, No. 9

Quantitative characterisation of the binding reaction of the parasite. Three RNA ligand families were isolated. A major
group containing almost 90% of all selected RNAs was
As above, we used aptamer 2-16 for a quantitative analysis of the characterised by an invariant CGU sequence in addition to a
binding reaction. The dissociation constant for the interaction of highly conserved NGC base triplet. The molecules were shown
2-16 RNA with either MITat 1.2 or MITat 1.4 cells was to adopt an irregular hairpin structure with the invariant CGU
determined in a Scatchard analysis. Varying concentrations of the sequence in a bulged out, single-stranded conformation and the
radioactively labelled RNA (0.1–250 nM) were incubated with a conserved NGC sequence located within the loop of the hairpin.
constant number of parasite cells (5 × 108 cells/ml). The resulting This likely indicates sequence-specific contacts to the binding
binding curve (Fig. 6A) shows saturation characteristics demon- target, especially in the case of the invariant CGU (see below).
strating the presence of a defined number of binding sites which Family II aptamers share the potential of folding into G-quartet
can be titrated with the aptamer RNA. From the Scatchard structures. Since several other selection/amplification experiments
analysis of eight independent measurements with different target have resulted in G-quartet nucleic acid ligands (36–40), it is not
cell concentrations we derived a dissociation constant (Kd) of clear whether this reflects a concealed selection pressure towards
60 ± 17 nM indicating a high affinity interaction. The target very stable RNA structures inherent to the SELEX protocol.
concentration at saturation corresponds to a number of 72 000 ± However, a comparison of the RNase stability of representative
18 000 binding sites/parasite cell. RNAs of all three aptamer families in comparison to the stability
The association rate of aptamer 2-16 to the surface of the of the initial RNA pool showed no significant differences (data
parasites was determined under pseudo-first-order conditions, not shown). Therefore, RNA stability was not the main selection
equivalent to a large excess of cell surface target (25 nM) over criterion in the two experiments.
aptamer RNA (0.75 nM). Figure 6B shows a typical example of
a binding kinetic which reveals an initial association phase
Table 1. Relative binding efficiencies of RNA aptamers to ILTAT 1.1 and
(0–20 min) with defined first-order characteristics and a pseudo-
AnTat 1.1 trypanosomes
first-order association rate (kobs) of 0.06 min–1, equivalent to a
half maximal association time (t
) of 12 min (t
= ln 2/kobs).
Aptamer family RNA AnTat 1.1 ILTAT 1.1
After 25–30 min, however, RNA binding slows down and a slight
decrease in the association was observed at later time points Pool 1 1 1
(50–60 min). Hence, initial pseudo-first-order association rates Pool 12 3.1 ± 0.2 3.7 ± 0.6
(kobs) were determined at varying target concentrations (4- to
Family I 2-11 5.3 ± 2.6 8.3 ± 3.0
30-fold molar excess over aptamer RNA). As expected for a
bimolecular association reaction, a linear dependence of kobs on 2-16 2.4 ± 0.5 3.5 ± 1.5
the target concentration was identified (Fig. 4D). From the slope 4-25 6.5 ± 2.2 9.7 ± 1.1
of the curve (kass = kobs/[target]) an association rate constant (kass) 4-13 6.0 ± 0.3 7.0 ± 0.4
of 4.3 × 104 M–1 s–1 was derived.
Family II 2-5 11.4 ± 3.7 14.4 ± 3.7

Binding to other trypanosome strains 2-10 17.1 ± 4.0 21.2 ± 12.0

All aptamers were unable to distinguish between the two MITat
strain variants that were used in the two selections. This raised the
question whether other, non-MITat trypanosomes could also be
recognised. Thus, we determined the binding efficiency of the
selected RNAs to two additional T.brucei strains, the well-
characterised pleomorphic strains ILTAT 1.1 (30) and AnTat 1.1
(29). Binding efficiencies of aptamers from all three families
were determined and normalised to the binding of pool 1 RNA
(Table 1). Although both strains showed a 5- to 10-fold higher
non-specific affinity for RNA (compared to the two MITat
variants), pool 12 RNA still bound 3–4 times stronger to these two
strains than pool 1 RNA. This clearly indicated that RNA species
were selected that also have an affinity for the ILTAT and AnTat
trypanosomes. Individual members of family I and family III
bound 3–9 times stronger than pool 1 RNA and the G-quartet
structures of family II showed a 10- to 25-fold stronger binding.
These binding strengths result in 10–30% of the RNAs bound to
the parasite surface, which are comparable values to the binding
efficiencies to the MITat 1.2 and MITat 1.4 parasites.
DISCUSSION
We demonstrate the isolation of high affinity ligands from a
combinatorial RNA library to the surface of live African
trypanosomes. The selection was performed with bloodstream
stage trypanosomes which represent the infective life cycle stage
4-29 17.7 ± 3.0 24.7 ± 4.8
Family III 2-18 3.9 ± 1.3 6.6 ± 1.6
2-29 4.0 ± 0.9 5.1 ± 1.4
4-19 6.0 ± 2.1 8.2 ± 0.4
All aptamers can be subgrouped into high and low affinity
binders. Differences between the two groups could only be
manifested in the case of the family I RNAs. Strong binders are
likely to adopt a stable pseudoknot structure which involves in
part the conserved NGC base triplet of the hairpin loop sequence.
However, further experiments are required to substantiate this
hypothesis. In the case of the other two families, the binding
strength differences must be due to other structural criteria.
Among several possibilities, the length of the individual stem
structures, a possible coaxial stacking of the helices and/or the
spacing between conserved elements might account for this
effect. Nevertheless, all tested RNAs specifically crosslink to the
same 42 kDa protein which is present on the surface of both the
MITat 1.2 and MITat 1.4 parasites. This is surprising for two
reasons. First, given the complexity of the cell surface, one might
have expected the selection of aptamers to different surface
targets similar to what has been described by Morris et al. (26),
using human erythrocyte ghosts. Second, none of the aptamers
interacts with the most abundant polypeptide on the trypanosome
surface, the VSG protein. Although we cannot exclude the

Figure 6. (A) Determination of the dissociation constant (Kd) for the binding reaction between aptamer 2-16 and MITat 1.2 cells. (Insert) Binding curve using varying
concentrations of 2-16 RNA at a constant cell density of 5 × 108 cells /ml. (Main) Scatchard analysis of the binding data. The slope of the curve represents –1/Kd, while
from the x-intercept the number/concentration of target molecules can be derived. (B) Determination of the association rate of aptamer 2-16 (0.8 nM) with MITat
1.2 cells (2 × 108 cells/ml) under pseudo-first-order conditions. Values within the 0–20 min time frame are fitted according to a first-order equation, giving a
pseudo-first-order rate constant (kobs) of 0.06 min–1. (Insert) Determination of kobs at different target cell densities. The slope of the curve represents kass. Assuming
a target number of 72 000 molecules/cell, kass was calculated to be 4.3 × 104 M–1 s–1.
possibility that the 42 kDa protein displays an epitope dominance as
described by others (41–43), we attribute both phenomena to the
unique situation of the trypanosome cell surface. The VSG layer is
not only acting as a physical barrier, it also shows dynamic properties
which enhance its protective function. VSG-bound antibodies, for
instance, have been shown to be eliminated from the trypanosome
surface by a rapid endocytotic turnover of the entire VSG layer
through the flagellar pocket (44,45). Thus, any VSG-bound RNA
was probably removed during the selection via this recycling
reaction and, consequently, stable surface binding might have been
restricted to the less abundant invariable surface elements.
Since the binding of the aptamers is restricted to the flagellar
pocket of the parasites it is tempting to speculate that the 42 kDa
protein might be ESAG 7, a subunit of the transferrin receptor.
ESAG 7 has a molecular mass of 42 kDa, is located within the
flagellar pocket and is exclusively expressed during the blood-
stream life cycle stage of the parasite (46,47). In contrast, we
estimated a copy number of about 70 000 molecules/trypanosome
cell for the 42 kDa protein which is above the reported value of
2–3 × 103 copies for the transferrin receptor (reviewed in 48).
Thus, further experiments must be conducted to confirm this
hypothesis, in addition to experiments that address the fate of the
RNA after binding to the flagellar pocket.
Lastly, we cannot exclude the possibility that next to the
interaction with the 42 kDa protein additional contact points between
the aptamer and the trypanosome surface exist. Yang et al. (49)
recently reported the selection of G-quartet DNA ligands with
cellobiose. This opens up the possibility that the G-quartet RNAs of
family II interact, at least in part, with a carbohydrate moiety. The
trypanosome surface is abundantly glycosylated (50) and thus sugar
residues could contribute to the RNA binding domain.
Three additional features of the aptamers are worth discussing.
First, none of the RNAs bind to insect stage trypanosomes. This
is likely to indicate that the surface target is bloodstream stage
specific. Alternatively, the binding target might be accessible
only in bloodstream stage cells because of steric and/or surface
charge differences of the two life cycle stages. Second, none of the
aptamers are capable of discriminating between the two parasite
strain variants used in the selection scheme. This demonstrates
2013
Nucleic Acids
Nucleic Acids Research,
Research,1994,
1999,Vol.
Vol.22,
27,No.
No.19 2013
that the aptamer/trypanosome interaction is independent of the
particular VSG variant expressed on the trypanosome surface.
Third, two unrelated trypanosome strains are also recognised by
the aptamers. This may suggest that the selected RNAs can be
applied to a greater range of trypanosome strains, provided the
target protein is sterically accessible.
In summary, the selection of high affinity nucleic acid ligands
presented in this study extends the range of applications of the
SELEX methodology to living cells as binding targets. The
identified RNA aptamers show high affinity binding and
specificity for a single protein and as such they have the potential
to be used as diagnostic as well as therapeutic tools. Their ability
to bind to an invariant element on the trypanosome surface opens
up the possibility of side-stepping the antigenic variation of the
VSG coat. Binding to the flagellar pocket, the main site for endo-
and exocytosis events, might further allow for a specific targeting
of toxic components to intracellular compartments of the parasite.
ACKNOWLEDGEMENTS
We thank U. F. Müller for discussion and input in the quantification
of the binding data. E. Weyher-Stingl is thanked for her help
recording the CD spectra, M. Engstler and M. Boshart for
providing T.brucei strains and F. Melchior for providing access to
her digital imaging facility. P. Overath, G. A. M. Cross and A. S.
Paul are thanked for critically reviewing the manuscript. We
acknowledge the contributions of H. H. Shu during the early
phase of the project, the secretarial help of A. Kohlhuber and the
purification of VSG by M. Worbs. Metabion is thanked for expert
oligodeoxynucleotide synthesis. The work was supported by a
grant from the German Research Foundation (DFG) to H.U.G.
REFERENCES
1 Smith,D.H., Pepin,J. and Stich,A.H.R. (1998) Br. Med. Bull., 54, 341–355.
2 Hursey,B.S. and Slingenbergh,J. (1995) World Anim. Rev., 3–4, 67–73.
3 Taylor,K.A. (1998) Int. J. Parasitol., 28, 219–240.
4 Van Gompel,A. and Vervoort,T. (1997) Curr. Opin. Infect. Dis., 10, 469–474.
5 Takayanagi,T. and Enriquez,G.L. (1973) J. Parasitol., 59, 644–647.
6 Sternberg,J.M. (1998) Chem. Immunol., 70, 186–199.
2014 Nucleic Acids Research, 1999, Vol. 27, No. 9

7 Borst,P. and Rudenko,G. (1994) Science, 264, 1872–1873. 31 Brun,R. and Schönenberger,M. (1979) Acta Trop., 36, 289–292.
8 Cross,G.A.M. (1996) Bioessays, 18, 283–291. 32 Schmid,B., Riley,G., Stuart,K. and Göringer,H.U. (1995) Nucleic Acids Res.,
9 Rudenko,G., Cross,M. and Borst,P. (1998) Trends Microbiol., 6, 113–117. 23, 3093–3102.
10 Carrington,M., Miller,N., Blum M., Roditi,I., Wiley,D. and Turner,M. 33 Freymann,D., Down,J., Carrington,M., Roditi,I., Turner,M. and Wiley D.
(1991) J. Mol. Biol., 221, 823–835. (1990) J. Mol. Biol., 216, 141–160.
11 Blum,M.L., Down,J.A., Gurnett,A.M., Carrington,M., Turner,M.J. and 34 Phillips,K., Dauter,Z., Murchie,A.I.H., Lilley,D.M.J. and Luisi,B. (1997)
Wiley,D.C. (1993) Nature, 362, 603–609. J. Mol. Biol., 273, 171–182.
12 Carrington,M. and Boothroyd,J. (1996) Mol. Biochem. Parasitol., 81, 35 Cheng,A.J. and van Dyke,M.W. (1997) Gene, 197, 253–260.
119–126. 36 Macaya,R.F., Schultze,P., Smith,F.W., Roe,J.A. and Feigon,J. (1993)
13 Lamont,G.S., Tucker,R.S. and Cross,G.A.M. (1986) Parasitology, 92, Proc. Natl Acad. Sci. USA, 90, 3745–3749.
355–367. 37 Lauhon,C.T. and Szostak,J.W. (1995) J. Am. Chem. Soc., 117, 1246–1257.
14 Turner,C.M. and Barry,J.D. (1989) Parasitology, 99, 67–75. 38 Huizenga,D.E. and Szostak,J.W. (1995) Biochemistry, 34, 656–665.
15 Bülow,R., Overath,P. and Davoust,J. (1988) Biochemistry, 27, 2384–2388. 39 Li,Y.F., Geyer,C.R. and Sen,D. (1996) Biochemistry, 35, 6911–6922.
16 Ziegelbauer,K., Multhaup,G. and Overath,P. (1992) J. Biol. Chem., 267, 40 Weiss,S., Proske,D., Neumann,M., Groschup,M.H., Kretzschmar,H.A.,
10797–10803. Famulok,M. and Winnacker,E.L. (1997) J. Virol., 71, 8790–8797.
17 Ziegelbauer,K. and Overath,P. (1993) Infect. Immun., 61, 4540–4545.
41 Pan,W., Craven,R.C., Qiu,Q., Wilson,C.B., Wills,J.W., Golovine,S. and
18 Seyfang,A. and Duszenko,M. (1993) Eur. J. Biochem., 214, 593–597.
Wang,J.-F. (1995) Proc. Natl Acad. Sci. USA, 92, 11509–11513.
19 Nolan,D.P., Jackson,D.G., Windle,H.J., Pays,A., Geuskens,M., Michel,A.,
42 Ringquist,S., Jones,T., Snyder,E.E., Gibson,T., Boni,I. and Gold,L. (1995)
Voorheis,H.P. and Pays,E. (1997) J. Biol. Chem., 272, 29212–29221.
20 Cohn,C.S. and Gottlieb,M. (1997) Parasitol. Today, 13, 231–235. Biochemistry, 34, 3640–3648.
21 Overath,P., Stierhof,Y.-D. and Wiese,M. (1997) Trends Cell Biol., 7, 27–33. 43 Barry,M.A., Dower,W.J. and Johnston,S.A. (1996) Nature Med., 2, 299-305.
22 Johnson,J.G. and Cross,G.A.M. (1979) Biochem. J., 178, 689–697. 44 Webster,P., Russo,D.C.W. and Black,S.J. (1990) J. Cell Sci., 96, 249–255.
23 Tuerk,C. and Gold,L. (1990) Science, 249, 505–510. 45 Russo,D.C.W., Grab,D.J., Lonsdale-Eccles,J.D., Shaw,M.K. and
24 Gold,L., Polisky,B., Uhlenbeck,O. and Yarus,M. (1995) Annu. Rev. Biochem., Williams,D.J.L. (1993) J. Cell Biol., 62, 432–441.
64, 763–797. 46 Steverding,D., Stierhof,Y.-D., Chaudhri,M., Ligtenberg,M., Schell,D.,
25 Ellington,A.D. and Conrad,R. (1995) Biotechnol. Annu. Rev., 1, 185–214. Beck-Sickinger,A.G. and Overath,P. (1994) Eur. J. Cell Biol., 64, 78–87.
26 Morris,K.N., Jensen,K.B., Julin,C.M., Weil,M. and Gold,L. (1998) 47 Salmon,D., Geuskens,M., Hanocq,F., Hanocq-Quertier,J., Nolan,D.,
Proc. Natl Acad. Sci. USA, 95, 2902–2907. Ruben,L. and Pays,E. (1994) Cell, 78, 75–86.
27 Hirumi,H. and Hirumi,K. (1994) Parasitol. Today, 10, 80–84. 48 Overath,P., Chaudhri,M., Steverding,D. and Ziegelbauer,K. (1994)
28 Cross,G.A.M. (1975) Parasitology, 71, 393–417. Parasitol. Today, 10, 53–58.
29 Michiels,F., Matthyssens,G., Kronenberger,P., Pays,E., Dero,B., Van Assel,S., 49 Yang,Q., Goldstein,I.J., Mei,H.-Y. and Engelke,D.R. (1998) Proc. Natl
Darville,M., Cravador,A., Steinert,M. and Hamers,R. (1983) EMBO J., 2, Acad. Sci. USA, 95, 5462–5467.
1185–1192. 50 McConville,M.J. (1996) In Smith,D.F. and Parsons,M. (eds), Molecular
30 Rice-Ficht,A.C., Chen,K.K. and Donelson,J.E. (1982) Nature, 298, 676-679. Biology of Parasitic Protozoa. IRL Press, Oxford, UK.