ABSTRACT
Trypanothione reductase (TR), a flavoprotein oxidoreductase present in
trypanosomatids but absent in human cells, is regarded as a potential target
for the chemotherapy of several tropical parasitic diseases caused by
trypanosomes and leishmanias. We investigated the possibility of modulating
intracellular TR levels in
Trypanosoma cruzi
by generating transgenic lines that extrachromosomally overexpress either sense
or antisense TR mRNA. Cells overexpressing the sense construct showed a 4-10-fold increase in levels of TR mRNA, protein and enzyme activity. In
contrast, recombinant
T.cruzi
harbouring the antisense construct showed no significant difference in TR
protein or catalytic activity when compared with control cells. Although
increased levels of TR mRNA were detected in some of the antisense cells
neither upregulation nor amplification of the endogenous trypanothione
reductase gene (
tryA
) was observed. Instead, a proportion of plasmid molecules was found rearranged
and, as a result, contained the
tryA
sequence in the sense orientation. Plasmid rescue experiments and sequence
analysis of rearranged plasmids revealed that this specific gene inversion
event was associated with the deletion of small regions of flanking DNA.
The trypanosomatids, members of the order Kinetoplastida, include parasitic
protozoa of relevant importance to public health such as
Trypanosoma spp.
(sleeping sickness and Chagas' disease) and
Leishmania spp.
(visceral, cutaneous and mucocutaneous leishmaniasis). They are characterised by
complex life cycles comprising various developmental stages that alternate
between insect and human hosts. Once inside the human host they proliferate
rapidly to cause severe illness and in extreme cases, death (
1
). Since no satisfactory vaccines against trypanosomatid infections are yet
available, chemotherapy remains the only means of treatment for the millions of
infected individuals worldwide. In many cases however, available drugs are few
and their efficacy is limited due mainly to developed parasite resistance
and/or lack of specificity. Other drugs are highly toxic and cause severe side
effects (
2
). Thus the need for more efficient antiparasitic drugs.
In recent years, a rational approach to the development of new pharmaceuticals
has emerged as an alternative to random testing. It is based on the
identification of structural or metabolic cellular components present in the
target organism which are either absent in the host or are sufficiently
different to their host's counterpart to be treated as unique. In the case of
the trypanosomatids one of the most promising examples of such unique features
is the trypanothione system (
3
,
4
). These ancient eukaryotes differ from their human hosts in their ability to
conjugate two molecules of the tripeptide glutathione with one molecule of
spermidine to form the dithiol
N
1
,
N
8
-bis(glutathionyl)-spermidine, trivially known as trypanothione (
5
). Together with its corresponding oxido-reductase, trypanothione reductase (TR), trypanothione is thought to
fulfil important physiological functions including maintenance of a reduced
intracellular environment and defense against heavy metals, oxidants and
xenobiotics (
4
), roles ascribed to the glutathione system in most other organisms. Since the
trypanothione system appears to be shared between most members of the
trypanosomatid family it is likely that any rationally designed inhibitor of TR
will be potentially useful in the treatment of a wide variety of parasitic
diseases.
In attempting to develop new antiparasitic drugs against a potential cellular
target it is important to gather genetic evidence that the target in question
is essential for cell proliferation and survival. Reverse genetics techniques
such as antisense RNA and gene replacement are ideal for this purpose; they
have proved invaluable in the elucidation of the regulation of gene expression
and of gene function in a variety of biological systems (reviewed in ref. 6-9). We and others have previously reported the cloning, characterisation,
mutagenesis and heterologous overexpression of the trypanothione reductase gene
(
tryA
) from
Trypanosoma cruzi
(
10
,
11
). Here we report its homologous overexpression and, in an attempt to interfere
with the expression of the endogenous
tryA
gene, the extrachromosomal overexpression of its complementary mRNA.
A cloned epimastigote line of
Trypanosoma cruzi
(MHOM/BR/78/Silvio X10-clone 6) was grown at 28oC in RPMI-1604 medium (Gibco) supplemented with 20 mM HEPES, 0.03 mM
haemin, 0.4 % (w/v) trypticase, 10 % (v/v) foetal calf serum, 50 U/ml
penicillin and 50 [mu]g/ml streptomycin. Trypanosomes were maintained by subculture and kept at
cell densities ranging between 1 * 10
6
and 3 * 10
7
cells/ml.
Transfection of
T.cruzi
was performed mainly as described previously (
12
). Briefly, late-log phase epimastigotes were washed in PBS solution (
13
) and resuspended in electroporation buffer (10 mM sodium phosphate pH 7.1, 0.27
M sucrose) at a concentration of 1 * 10
8
parasites/ml. Three to eight pulses from a Hoefner Progenitor I electroporator
(400 V, 99 ms) were applied to 1 ml aliquots of the cell suspension in the
presence or absence of 25 or 50 [mu]g supercoiled plasmid DNA. Cells were diluted 5-fold with culture medium and allowed to recover for 24 h; a further 10-fold dilution with medium containing 0.1 mg/ml G418 was used to
select for drug-resistant trypanosomes. Clones of transgenic trypanosomes were derived by
limited dilution from established drug-resistant lines obtained in different transfection experiments.
Plasmid pBTR (
11
) was digested with
Eco
RI and
Pst
I to release a 1.5 kb fragment that contains the entire
tryA
coding sequence. Ends were rendered blunt by treatment with T4 and
E.coli
(Klenow fragment) DNA polymerases and the fragment was ligated to the
Eco
RV-linearised expression vector pTEX (
12
), which had been previously treated with alkaline phosphatase. Following
transformation,
E.coli
JM109 clones containing plasmids with the
tryA
insert in either possible orientation were identified and used to purify
plasmids pTTcTR (sense) and pTTcTRAS (antisense). The purity of plasmid DNA
solutions used in transfection experiments was confirmed by control PCR
reactions using increasing amounts of template DNA (10 pg to 500 ng) and
appropriate combinations of sense- and antisense-specific oligonucleotides.
Trypanothione reductase and alanine aminotransferase (ALAT) activities were
assayed in cell-free extracts prepared as follows: 1-3 * 10
8
T.cruzi
epimastigotes were harvested by centrifugation, washed in PBS solution and lysed
in 0.25-0.35 ml lysis buffer (10 mM potassium phosphate pH 7.2, 10 mM EDTA, 1 mM
DTT, 1% Triton X-100, 5 mM benzamidine, 5 mM phenanthroline, 0.1 mM phenylmethylsulphonyl
fluoride). Following three cycles of freeze (liquid N
2
)-thawing, crude extracts were cleared by centrifugation at 4oC and 14 000 r.p.m. for 5 min. Supernatants were transferred to
fresh tubes and kept on ice until assayed for enzymatic activity.
TR activity was monitored spectrophotometrically by following the trypanothione disulphide-dependent oxidation of NADPH at 340 nm (
14
). The reaction was initiated by adding 0.05 mM trypanothione disulphide
(Bachem) to a reaction mixture (0.5 ml final volume) containing 0.1 M HEPES, pH
7.8, 0.5 mM EDTA, 0.2 mM NADPH and 20 [mu]l of the appropriate cell-free extract. ALAT activity was assayed spectrophotometrically by a
modification of the method of Segal
et al
. (
15
). The assay is based on the ALAT-dependent synthesis of pyruvate from L-alanine and [alpha]-ketoglutarate coupled to the pyruvate-dependant oxidation of NADH by lactate
dehydrogenase, this being monitored at 340 nm. A standard reaction mixture (0.5
ml) contained 0.1 M HEPES pH 7.8, 0.5 mM EDTA, 0.17 mM NADH and 20 [mu]l cell-free extract. The reaction was initiated by the addition of L-alanine and [alpha]-ketoglutarate to final concentrations of 30 and 2
mM respectively, plus 2.5 U lactic dehydrogenase (Sigma). Both TR and ALAT
assays were performed using a Beckman DU-70 spectrophotometer fitted with a cell temperature regulator. One unit of
ALAT or TR activity is defined as the amount of enzyme required to oxidise 1 [mu]mol NADH or NADPH (respectively) per min at 27oC. Protein concentrations were determined by the method of Bradford (
16
).
Twenty five [mu]g of total protein from cell-free extracts prepared as described above were fractionated by SDS-PAGE on a 10% gel and electroblotted onto nitrocellulose using
a Mini-Protean II system (BioRad). A 1:100 dilution of a TR polyclonal antibody
prepared as described previously (
17
) was used to probe the blot; bands were visualised using an alkaline
phosphatase-coupled immunoassay (
13
).
Total DNA was prepared using the proteinase K method essentially as described (
18
), except that crude extracts were incubated at 37oC for 1 h in the presence of 70 [mu]g/ml DNAse-free RNAse prior to the addition of proteinase K. Concentration
of DNA was determined spectrophotometrically. For DNA blot analysis, 2 [mu]g aliquots were digested with appropriate restriction enzymes and, following
electrophoresis through 0.8% agarose gels, blotted onto nylon membranes (Hybond
N, Amersham) using a Vacu-aid apparatus (Hybaid). RNA was extracted in the presence of guanidinium
thiocyanate and purified by centrifugation through a 5.7 M CsCl cushion as
described (
18
). For northern blot analysis, total RNA aliquots (10 [mu]g) were electrophoresed through 1.1% formaldehyde-containing agarose gels and blotted onto nylon membranes by
capillarity. Crosslinking of nucleic acids to nylon membranes was by exposure
to UV light in a crosslinking oven (Stratagene). Hybridisation to radiolabelled
probes was carried out using standard techniques (
13
). DNA probes were labelled to high specific activity by random priming (
19
). Labelled, orientation-specific RNA probes were generated with a TransProbe T Kit (Pharmacia)
using T3 or T7 RNA polymerases and linear pBTR as a template.
Since pTEX-derived vectors replicate as DNA multimers in
T.cruzi
(
12
; our own unpublished observations) unit-size plasmid molecules (7.1 kb) were recovered from recombinant parasites
by digestion with
Xho
I, an endonuclease with a single recognition site within plasmids pTTcTR and
pTTcTRAS (Fig.
1
A). DNA aliquots (2 [mu]g) were digested with
Xho
I and size-fractionated by electrophoresis through 0.8 % agarose gels. DNA fragments
in the 6-8 kb size range were recovered from the gel by centrifugation through a
glass fibre cushion and, after concentration, treated with T4 DNA ligase for 3
h at 16oC. Self-ligated DNA was transformed into
E.coli
JM109 or XL1-Blue competent cells. Ampicillin-resistant clones were then used for plasmid preparation and
analysis.
The nucleotide sequence of
tryA
junctions in plasmid DNA was determined by the double-stranded DNA sequencing method using a Sequenase 2.0 kit (USB). DNA
synthesis was primed by synthetic oligodeoxynucleotides TcTR19(-) [5'-GCGCCAATGACAACCAAATC-3'] and TcTR1441(+) [5'-GGTGAGAAGATGGAAAAGCC-3']. The former is
complementary to the DNA sequence 19 bp downstream from the translation
initiation codon, the latter lies 19 bp upstream from the translation
termination codon and were used to read into the 5' and 3' junctions respectively.
Antisense AS1 cells were seeded at a density of 0.5 * 10
6
cells/ml in culture medium supplemented with 0.5 mg/ml G418 and incubated at 28oC to a final density of 2 * 10
7
cells/ml. Cells were diluted in drug-free medium to a density of 2 * 10
5
cells/ml and serial 1:1 dilutions were then made in a 24-well culture dish until a theoretical population density of 0.025
parasites/ml was achieved. Plates were sealed and incubated at 28oC for 6-8 weeks. AS1-0.5 clones obtained in this manner were cultured for at least
ten generations in the presence of G418 (0.1 mg/ml) before further analyses.
Plasmid pBTR has been described previously (
11
) and was used as the source of the
T.cruzi tryA
gene. The non-integrative shuttle vector pTEX (
12
) was used to clone the entire
tryA
coding region in both possible orientations. The resulting plasmids, pTTcTR
(sense) and pTTcTRAS (antisense) are depicted in Figure
1
A. Expression of
tryA
and antisense
-tryA
in these plasmids is linked to the constitutive expression of the drug
resistance marker and utilises the polyadenylation and splice acceptor signals
of the tandemly repeated glyceraldehyde phosphate dehydrogenase (GAPDH) genes
from
T.cruzi
(
12
,
20
). The orientation of
tryA
was established by restriction with the diagnostic enzyme
Kpn
I (see Fig.
1
A) and was confirmed by DNA sequencing. Following plasmid purity tests
(Materials and Methods)
T.cruzi
epimastigotes were transfected with these constructs in various independent
experiments and recombinant cells were selected in the presence of G418. Clones
TR4 (sense) and AS1 and AS2 (antisense) were randomly chosen for analysis.
To test for the integrity of transfected plasmids and to estimate plasmid copy
number per cell, DNA blot analysis of selected G418-resistant clones was carried out using
Eco
RI-digested DNA (Fig.
1
B); the target site of
Eco
RI is present once in plasmids pTEX, pTTcTR and pTTcTRAS (see Fig.
1
A). Figure
1
B confirms the presence of major, unit-size plasmid bands in the corresponding
T.cruzi
recombinant clones, which is in agreement with the drug resistant phenotype.
Comparison of band intensities between the single chromosomal
tryA
locus (Tovar
et al.,
submitted) and plasmid-borne
tryA
in Figure
1
B, lanes 3 and 5 (
tryA
panel) reveals the presence of >= 20 plasmid molecules per cell in both sense and antisense recombinant
clones. Equivalent amounts of plasmid DNA were estimated for the pTEX control
clone when a
neo-
specific sequence was used to probe the same blot (Fig.
1
B).
To assess the effect of the extrachromosomal expression of sense and antisense
TR mRNA on TR activity levels in
T.cruzi
, we assayed for enzymatic activity in log phase epimastigotes. ALAT, another
soluble, house-keeping enzyme in this organism was assayed in cell-free extracts to control for the general metabolic state of cells.
As shown in Figure
2
A, a 4-fold increase in TR specific activity was found in cells harbouring the sense construct. However, activity
levels observed in both antisense clones were not significantly different to
those of control cells suggesting either that the expression of antisense TR
mRNA is not harmful to these cells or that any potentially deleterious effect
of antisense RNA is efficiently neutralised. No significant difference in ALAT
activity was found between untransfected and recombinant clones (Fig.
2
B) indicating that replication of recombinant plasmids
per se
does not affect the general metabolic state of the cell.
To investigate whether the lack of apparent changes on TR phenotype in
recombinant antisense clones could be due to inefficient expression of
antisense RNA, total RNA was isolated from antisense clone AS1 grown under two
different selection regimes and analysed by northern blotting. High levels of
apparently correctly processed antisense TR mRNA were observed in these cells
using an antisense-specific
tryA
probe. This is indicated by the presence of a 2.1 kb transcript band in lanes 2
and 3 of Figure
3
A which is of the expected size if the trans-splicing and polyadenylation sites present in plasmid pTTcTRAS (Fig.
1
A) were used during RNA processing. Moreover, the fact that plasmid-derived sense and antisense TR transcripts are of equivalent sizes (Fig.
3
A and B) also suggests that these antisense transcripts are correctly processed.
However, the presence in the same lanes of an extra band in the 3.9 kb region,
which is also highlighted by a
neo
probe (not shown) and most likely represents bicistronic pre-mRNA, suggests that antisense RNA processing is inefficient in these
cells. Such accumulation of pre-mRNA could result from specific blocking of antisense RNA processing or,
alternatively, from inefficient RNA processing due to high expression levels.
The latter appears unlikely since cells that overexpress TR to a high level do
not accumulate multicistronic transcripts (Fig.
3
B, lane 4).
Extensive restriction analysis of total DNA isolated from antisense AS1 cells
grown at different selection levels suggested that, at high selective pressure,
structural changes had occurred in a proportion of plasmid molecules. This is
best illustrated by restriction analysis with
Kpn
I which allows determination of the orientation of
tryA
in a plasmid (see Fig.
1
A). Antisense plasmid pTTcTRAS (7.1 kb) should yield
Kpn
I fragment bands of about 3.5 and 3.6 kb; this is observed when a DNA digest
from cells grown in the presence of 0.1 mg/ml G418 is probed with either
neo-
or
tryA
-specific probes (lane 2 of Fig.
5
A and B respectively). However, digestion of DNA from cells grown at a higher
drug concentration revealed an additional set of restriction fragments, one of ~2.7 kb which hybridises to
neo
and another of ~4.4 kb which hybridises to the
tryA
probe (lane 3 of Fig.
5
A and B respectively). The fact that both new fragment bands co-migrated with those observed for pTTcTR transfected cells (Fig.
5
A and B, lane 5) suggested the presence, in this antisense clone, of rearranged
plasmid DNA carrying the
tryA
insert in the sense orientation.
Table 1
Structure of plasmid DNA rescued from transgenic trypanosomes
In an attempt to understand the mechanics of the observed DNA rearrangements the
nucleotide sequence of the
tryA
flanking regions of selected rescued plasmids was determined. Four apparently
unmodified, and four rearranged plasmids from each of the antisense clones were
analysed. As a control, four plasmids rescued from clone TR4 were also
sequenced. All of the apparently unrearranged antisense plasmids were identical
in sequence to the original transforming antisense plasmid. Likewise, the
control sense plasmids were identical in sequence to the original transforming
sense plasmid. However, of the eight rearranged plasmids analysed which
originally contained the
tryA
gene in the antisense orientation, all were confirmed to contain
tryA
in the sense orientation and, additionally, were found to contain a 41 bp
deletion at the 5' junction and a 25 bp deletion 3' to the
tryA
translation termination codon. Such deletions span the entire polylinker
sequence but do not affect the flanking untranslated regions present in the
expression vector (Fig.
6
). The accurate, identical nature of the DNA rearrangements observed in both
antisense clones AS1 and AS2 hints at the involvement of homologous
recombination.
Regulation of gene expression by antisense RNA occurs naturally in both
prokaryotic (
22
,
23
) and eukaryotic cells (
24
-
26
). Antisense RNA transcribed from transforming plasmid vectors has been used to
regulate endogenous gene expression in a number of eukaryotic systems,
including the slime mould
Dyctiostelium discoideum
, mammalian cells and plants (
27
-
29
). In parasitic protozoa however, no equivalent studies have yet been reported,
although the effect of synthetic antisense oligonucleotides on parasite
proliferation and survival has been documented (
30
-
33
).
In this study we investigated the possibility of modulating intracellular TR
levels by extrachromosomal, homologous expression of sense and antisense TR
mRNA in
T.cruzi
. Although TR specific activity was readily upregulated in cells harbouring the
sense construct, cells transformed with the antisense construct proved
recalcitrant to TR downregulation despite the presence of >= 20 copies of the antisense plasmid per cell and high expression of antisense
TR mRNA. Antisense RNA is thought to work in eukaryotic cells by formation of
hybrid, double-stranded RNA molecules that have been proposed to affect transcript
stability and/or processing (splicing/nuclear export/ribosome binding) thus
interfering with the expression of specific gene products (reviewed in ref.
34
). Various possibilities could account for our failure to downregulate the
expression of
tryA
in
T.cruzi.
Antisense transcripts may not be able to form duplex RNA hybrids with their
target molecules due to the non-complementarity of their 5' and 3' untranslated regions, including the spliced leader sequence
and poly A tail. Cellular compartmentalisation, well documented in eukaryotes
(reviewed in ref.
35
,
36
) may also play a part; formation of heteroduplex RNA may be prevented by the
absence, in the appropriate cellular compartment, of one or more of a number of
proteins that have been proposed to play a role in the process of antisense RNA
regulation, including hybrid-promoting proteins, winding and unwinding activities, and double-stranded RNAses and their modulators (ref.
34
and references therein). At a different level, other cellular mechanisms such
as transcriptional upregulation, gene amplification, or structural
rearrangement of plasmid DNA could also explain failure to regulate gene
expression by transforming antisense gene constructs. In this respect, a
peculiar plasmid DNA rearrangement event was observed in our investigation.
From its original antisense orientation, the
tryA
insert in plasmid pTTcTRAS was flipped over to the sense orientation. Although
such an event lead to increased intracellular TR mRNA levels this was not
reflected phenotypically by increased TR catalytic activity nor TR protein
suggesting that only a proportion of these transcripts were translated into
functional enzyme either as a result of RNA degradation (perhaps following
heteroduplex formation with antisense RNA) or due to inefficient or incorrect
RNA processing.
Rearrangement of plasmid DNA must have occurred in
T.cruzi
since control PCR experiments failed to detect the presence of contaminating
sense plasmid molecules in antisense plasmid DNA solutions used in transfection
experiments. This is further supported by the fact that the structure of
rearranged molecules has been shown to be different from those of the original
sense and antisense transfecting constructs and by the observation that no such
rearranged molecules are detected in DNA blots of the originally selected
antisense AS1 clone. Instead, structural analysis of rearranged plasmid
molecules suggests the involvement of short repeated sequences in the inversion
of
tryA
and associated deletion of flanking regions. In Figure
6
a sequence of events that could explain the conversion of an original antisense
plasmid into a rearranged molecule is depicted. In this model a first round of
homologous reciprocal recombination between the inverted duplications marked a
and b in the diagram would lead to inversion of the intervening sequence. As a
result the
tryA
coding region would now be in the sense orientation and the 41 bp 5' polylinker fragment would have been removed from its original upstream
location and fused to the 25 bp 3' polylinker sequence. A second round of homologous reciprocal
recombination between the newly created direct repeats marked c and d in the
diagram would then lead to the deletion of the 66 bp intervening sequence, thus
removing the entire polylinker region and producing a plasmid of structure
identical to that observed for rearranged plasmid molecules. The proposed model
faithfully accounts for our experimental observations and implies that repeated
DNA sequences as short as 4 bp can be used as substrates for homologous
recombination in
T.cruzi.
Eight base pair direct repeats have been shown to support homologous
recombination in
Leishmania mexicana
(
37
), but the minimum homology requirement for reciprocal and non-reciprocal homologous recombination in trypanosomatids has yet to be
determined experimentally.
High molecular weight plasmid concatemers that appear to be linked in a head to
tail arrangement accumulate in transgenic
T.cruzi
harbouring pTEX-derived vectors (
12
; our own unpublished observations). This type of extrachromosomal plasmid
multimers have also been observed (
38
) and extensively characterised in
T.brucei
(
39
),
Leishmania
(
40
,
41
) and
Leptomonas
(
42
). It is likely that plasmid DNA replication occurs in trypanosomatids by a
mechanism in which DNA synthesis and recombination are tightly interrelated.
Such a mechanism could facilitate the generation of plasmid DNA diversity in a
given cell population and lead to selection of advantageous plasmid variants.
At present it is not possible to test directly whether the observed plasmid
rearrangements arose solely as a result of the expression of antisense TR mRNA
since an inducible expression system for
T.cruzi
has not yet been developed. However, the high frequency of rearrangement, its
accuracy and exclusivity to the antisense clones warrant further investigation.
We thank J. Kelly for discussions and reagents (plasmid pTEX and
T.cruzi
cells expressing pTEX-CAT), M. Cunningham for help with TR assays, and S. Beverley, J. Kelly, W.
Nellen and R. Hernandez for critically reading an earlier version of this
manuscript. This work was supported by the Wellcome Trust.
CloneG418 (mg/ml)Orientation of
tryA
a
SenseAntisenseAS10.10/4343/430.510/2818/28AS20.113/3017/300.58/4032/40TR40.116/160/160.520/200/20
REFERENCES
Return