ABSTRACT
We have constructed artificial linear mini-chromosomes for the parasitic protozoan
Trypanosoma brucei.
These chromosomes exist at
~
2 copies per cell, are indefinitely stable under selection but are lost from 50%
of the transformed population in
~
7 generations when grown in the absence of selective pressure. Consistent with
results obtained earlier with natural chromosomes in
T.brucei
, the telomeres on these artificial chromosomes grow, adding
~
1-1.5 telomeric repeats per generation. The activity of a procyclic acidic
repetitive protein (
parp)
gene promoter on these elements is unaffected by its proximity to a telomere,
implying the lack of a telomere-proximal position effect (TPE) in procyclic trypanosomes. Among other
things, these autonomously replicating dispensable genetic elements will
provide a defined system for the study of nuclear DNA replication, karyotypic
plasticity and other aspects of chromosomal behavior in this ancient eukaryotic
lineage.
The long range objective of this study was to lay the groundwork for an
investigation of chromosomal behavior in
Trypanosoma brucei
. Parasitic protozoa, such as
T.brucei
,
Plasmodium falciparum
,
Giardia lamblia
,
Trypanosoma cruzi
and
Leishmania
demonstrate an unusually plastic karyotype (
1
-
15
; for recent reviews see
16
,
17
). In a study of
T.brucei
and certain sub-species, Gibson and Borst observed a `remarkable fluidity of chromosome
organization' and saw `no stocks with an identical karyotype'(
2
). An inherited alteration in the karyotype is by definition a mutation and the
range of these changes and the tolerance shown by the organism for these
alterations seems truly remarkable. An understanding of such a process in
trypanosomatids, for which a fairly sophisticated repertoire of molecular
techniques are already available, could have general significance for other
parasitic protozoa.
Trypanosoma brucei
is one of the causative agents of African trypanosomiasis. Phylogenetic
analyses demonstrate that these organisms are among the most ancient eukaryotic
lineages known (
18
-
21
). The parasite has a digenetic life-cycle, alternating between the midgut and salivary glands of its insect
vector (
Glossina
: tsetse) and the mammalian bloodstream.
Nuclear DNA replication is an unexplored aspect of trypanosome biology. In an
earlier attempt to develop models that would aid in the delineation of critical
replication control elements, we had constructed a panel of autonomously
replicating episomes for
T.brucei
(
22
). These were made by inserting random pieces of its genomic DNA onto a molecule
that could not otherwise exist as a stable replicon in this organism. We call
the inserted pieces of DNA the plasmid maintenance sequence or PMS. One of
these plasmids (pT13-11), has been extensively characterized (
22
-
24
). It exists as a single-copy episome in procyclic (insect mid-gut form)
T.brucei
, but nonetheless demonstrates substantial mitotic stability in the absence of
selection (50% loss in ~12 generations). In addition, we have shown that autonomous plasmid
replication was dependent on interactions between the PMS and a 108 bp region
encompassing the procyclic acidic repetitive protein (Parp or procyclin) gene
promoter that serves to drive the transcription of a selectable marker on this
episome (
24
).
Parp is the major surface protein in procyclics and is transcribed from two
unlinked loci,
parp
A and B (
25
).
Parp
transcription is resistant to [alpha]-amanitin and is thought to be mediated by RNA polymerase I (pol I),
which is capable of supporting mRNA transcription in
T.brucei
(
26
,
27
; also see review by
28
).
A 39 nt spliced leader is
trans
-spliced onto each mRNA in trypanosomatids and provides the 5' cap that is a necessary feature of all eukaryotic mRNA. Although [alpha]-amanitin sensitive transcription units are known, no
RNA polymerase II (pol II) promoters have been yet identified in
T.brucei.
An important additional motivation for the construction of these linear mini-chromosomes lay in our continuing interest to develop `promoter-traps' for the isolation and characterization of pol II promoters of
T.brucei.
A previous attempt to obtain such a trap, by deleting the
parp
promoter from pT13-11 and seeking to complement it with another promoter from the
T.brucei
genome, was complicated by the essential role played by the
parp
promoter region in plasmid replication (
24
). An artificial chromosome would provide us with a means to overcome this
complication, by allowing us to direct the
parp
promoter mediated transcription on the linear element away from a promoter-less selection unit. This would avoid potential problems of read-through transcription that would occur on a circular plasmid as a
result of polycistronic transcription, which is the norm in trypanosomatids.
Culture adapted procyclic forms of
T.brucei
strain 427 were grown at 27oC in SDM-79 medium (
29
) supplemented with 10% heat-inactivated fetal bovine serum (Sigma) as previously described (
22
).
For pTacA
,
a 1530 bp fragment consisting of an
Escherichia coli
chloramphenicol acetyl transferase (
cat
) gene cartridge (787 bp
Hin
dIII fragment from plasmid pCM7; Pharmacia), flanked on its 5' end by the
parp
B promoter and splice acceptor signals (SAS) and on its 3' end by a section of the [beta]-[alpha] tubulin (
tub
) intergenic region, was inserted between the
Hin
dIII and
Spe
I restriction sites of Bluescript SK+. The promoter and SAS correspond to
sequences between -312 and +93 with respect to the initiation site of
parp
-B transcription (
30
). The intergenic region comprises the last 108 bp of the [beta]-
tub
coding sequence and 227 bp of the 3' untranslated region (
31
).
The PMS and the neomycin phosphotransferase (Neo) expression unit were derived
from a modified version of pT13-11 where the
Bgl
II restriction site within the PMS had been destroyed, and a
Kpn
I site had been introduced at the unique
Eco
NI restriction site of pT13-11, as part of a polylinker. These alterations do not affect the ability
of the plasmid to replicate autonomously (Patnaik, unpublished observations).
The PMS (~4.8 kb) consists of the entire insert from the
Sal
I restriction site to the left most
Eco
RI restriction site of pT13-11 and the Neo expression unit is the ~2.1 kb
Apa
I-
Kpn
I fragment (
22
). These were inserted at the
Xho
I restriction site, and between the
Apa
I and
Kpn
I restriction site of the Bluescript vector respectively. The PMS is in the same
orientation with respect to the Neo expression unit as it is in pT13-11.
A ~4 kb
Sac
I fragment consisting of two head to head `telomeres' (1.6 kb each) separated by
a 650 bp spacer with
Bgl
II linkers at either end, was inserted into the
Sac
I restriction site of the Bluescript vector. The 1.6 kb `telomeric' DNA
corresponds to the
Eco
RI-
Pst
I fragment from pT4 (
32
,
33
) and consists of 55 copies of the 6 bp
T.brucei
telomeric repeat, in addition to other related sequences, and ~1 kb of `sub-telomeric' DNA (
33
). The spacer was expected to inhibit recombination between the inverted
telomeric repeats (
34
).
The construct pTacB is identical to pTacA, except that three tandem copies of a
1.5 kb fragment, consisting of a hygromycin phosphotransferase (
hyg
) gene flanked by the
Parp
B SAS (+6 to +93 with respect to the start site of
Parp
B transcription) and the same [beta]-[alpha]
tub
intergenic region that abuts the 3' end of the
cat
gene, were inserted into the
Xba
I restriction site of Bluescript.
Construct pTacC is a
Sal
I deletion derivative of pTacB. This deletion removes the
Parp
B promoter and SAS flanking the
cat
coding sequence.
Construct pTacD was obtained by inserting a synthesized polylinker comprising of
sites for restriction enzymes
Sal
I,
Afl
III,
Bcl
I,
Pme
I,
Sfi
I,
Asc
I and
Sac
I into a
Sal
I partial-
Sac
I digest of pTacB. The sequence constituting these sites is 5'-GTCGACTTAAGTGATCAGTTTAAACGGCCTAACTGGCCGGCGCGCCGAGCTC-3'. The
Sal
I,
Pme
I,
Sfi
I and
Asc
I sites are unique on pTacD.
Construct pTacAm is identical to pTacA except for a 286 bp sequence (derived
from pTacB and described in the text) inserted into the
Not
I site of the Bluescript polylinker. This site is located just upstream of the
Sac
I site flanking the right sub-telomeric repeats. As there is a
Not
I site in the PMS (
22
), insertion was achieved through a partial-
Not
I digest of pTacA.
Construction of pTacE involved several steps. The
Hin
dIII-
Bam
HI fragment comprising the luciferase coding sequences in pHD16 (
35
) was replaced with an 411 bp
Hpa
I-
Stu
I fragment from the plasmid pUT56 (kindly provided by D. Drocourt, CAYLA
Laboratories, France) comprising the
Streptoalloteichus hindustanus
phleomycin resistance
(
ble
) gene (
36
). The plasmid obtained is called pHD16-ble. A ble-expression cassette was generated by polymerase chain reaction (PCR)
mediated amplification of the appropriate fragment using primers complementary
to the
T.brucei
aldolase (
ald
) SAS and PAS respectively that flanked the
ble
gene in pHD16-ble. These primers were designed with additional
Xho
I,
Pme
I (SAS primer) and
Kpn
I (PAS primer) sites. The mobilized cassette was then inserted into
Xho
I-
Kpn
I digested pEV-luc described earlier (
24
) giving rise to the plasmid pBLN (
Construct pTacE-R was derived from pTacE by inserting an ~550 bp
Bam
HI-
Hin
dIII fragment corresponding to the rRNA promoter on the plasmid polINeo (
37
) into the unique
Pme
I site of pTacE.
The only
Bgl
II sites in any of these constructs are the ones flanking the spacer that
separates the telomeric repeats.
Procyclic trypanosomes were transfected by electroporation using 10 [mu]g of
Bgl
II-linearized DNA. Electroporator settings and other conditions were as
previously described (
38
). Transformed procyclic clones were obtained following plating of cells on
media containing agarose and 50 [mu]g/ml of G418 (GIBCO-BRL), as previously described (
39
). Where appropriate, 50 [mu]g/ml of hygromycin B (Sigma) or phleomycin (CAYLA,) was substituted for
G418. Transfection efficiencies were similar to those previously determined for
episomes (
22
).
Except in the preparation of samples for FIGE gels (see later), DNA was isolated
using the commercially available reagent DNA STAT60 and the manufacturer's
protocol (TEL TEST `B' Inc., Friendswood, TX). RNA STAT60 from the same
manufacturer was used to isolate total RNA.
Restriction digests, gel electrophoresis, Southern and Northern hybridizations
were performed using standard protocols (
40
). Final washes following hybridizations were done under stringent conditions
(0.1* SSC, 1% SDS, 65oC). Separation of large DNA fragments was performed using a field
inversion gel electrophoretic (FIGE) system (Minipulse programmable polarity
switching system obtained from IBI New Haven, CT). Agarose blocks containing ~1 * 10
7
cells were prepared as previously described (
41
). These were electrophoresed through a 1.2% agarose gel in 44 mM Tris-base, 2 mM EDTA, 40 mM boric acid at an electrical gradient of 5 V/cm
using the INV output mode of the minipulse generator and a forward (fwd) and
reverse (rev) pulse interval ranging from 0.025 to 0.505 (fwd)/s and 0.01 to
0.202 (rev)/s. This constituted program #1 of the manufacturer's protocol.
To digest DNA isolated from 2 *10
8
cells, 5 U of
Bal
31 exonuclease (NEB) was used. Reactions were done in duplicate at 30oC. Aliquots were removed at 0, 5, 10, 20 and 30 min into tubes containing
EDTA (20 mM final concentration). Each of these tubes was kept on ice until all
the aliquots had been obtained after which they were placed at 68oC for 10 min to destroy the exonuclease. Following ethanol precipitation,
the DNA was dissolved in TE (10 mM Tris pH 8.0, 1 mM EDTA). This DNA was then
digested with
Bam
HI and electrophoresed through agarose, Southern transferred and probed as
shown.
Trypanosomes bearing pTacA or pTacB were subcultured (1:150 dilution at each
passage, equivalent to ~7 generations between passages) in media without G418. Aliquots were
withdrawn at each passage. Genomic DNA was isolated, digested with
Bam
HI, electrophoresed through agarose, Southern transferred and probed with a 110
bp
PflM
I fragment from pEV (
24
) corresponding to the
Parp
A promoter.
The intensities of the signals corresponding to the genomic
Parp
A locus (arrow G) and the artificial chromosome (arrow AC) were measured using a
phosphorimager.
Clones bearing pTacA or pTacB were kept in continuous culture for 3 months under
G418 selection. DNA analysis (Fig.
3
B top panel) indicated that the right arm of pTacA grew by 1.2-1.6 kb over a 3 month period (~200 generations) of continuous culture (compare lanes 1 and 4),
corresponding to a growth rate of ~6-8 bp per generation. Also, the measured growth between any two time
points suggests that this growth rate was consistent over the entire period.
In the experiment shown in the bottom panel of Figure
3
B, genomic DNA isolated at different times from a continuous culture of clones
harboring either pTacA (lanes 1 and 2) or pTacB (lanes 3 and 4) was digested
with
Bam
HI and
Sac
I before analysis.
Sac
I digestion of these chromosomes was expected to remove the telomeric DNA from
the fragment seen by the
cat
probe (see Fig.
1
). Following this digest, the artificial chromosome band visualized by the
cat
probe was identical in size to the corresponding fragment derived from the
input plasmid (compare lanes 1 and 2 with lane A and lanes 3 and 4 with lane
B). Thus, these results indicated that the growth of artificial chromosomes was
confined to their telomeric/sub-telomeric region. Although we have not shown that this growth is confined
precisely to the ends of telomeres, that is the most likely scenario. The
estimated growth rate would indicate that ~1-1.5 telomeric repeats are added at every generation.
The electrophoretic mobility of pTacA and pTacB isolated from transformed
procyclic clones, and the growth of their telomeric DNA, strongly suggests that
these are linear extrachromosomal elements. We have confirmed this by
Bal
31 sensitivity experiments (Fig.
4
), which demonstrated the presence of exonuclease accessible ends on this
artificial chromosome. As a control for non-specific exonuclease activity, the filter was stripped and reprobed with a
fragment corresponding to the
parp
A promoter. As shown, the band corresponding to the genomic
parp
A locus, and the central
Bam
HI fragment from the artificial chromosome seen by the same probe, were both
unaffected by this period of
Bal
31 exposure, indicating their chromosome-internal position.
Clones bearing pTacA or pTacB have been kept in continuous culture under G418
selection for >5 months without loss of the artificial chromosomes or any
discernible effect on trypanosome growth. To determine the stability of these
elements in the absence of selective pressure, cultures of trypanosomes bearing
pTacA or pTacB were subcultured in media without G418. Aliquots were withdrawn
after each subculture. Measurements (Fig.
5
) indicated that pTacA and pTacB were present at an initial copy number of ~2 per cell and, in the absence of selection, 50% of these molecules are
lost from a transformed population in ~7 generations. Consequently, these linear elements are less stable than the
parent episome from which they were derived.
An important motivation for the construction of these elements was for their
anticipated use as promoter-traps. pTacB-transformed procyclics are able to survive selection with either
G418 or hygromycin. A
Sal
I
deletion removed the
parp
B promoter on pTacB giving the plasmid pTacC. Procyclic trypanosomes transformed
with
Bgl
II-linearized pTacC maintain the construct as an artificial chromosome in the
presence of G418 but could not survive hygromycin selection, indicating the
absence of fortuitous transcription into the hyg coding region, and the
potential use of constructs such as pTacC as promoter traps. We further refined
this vector to improve its convenience for such a use. A
Sal
I-partial
Sac
I digest removes the entire cat-hyg expression unit, which we replaced with a polylinker (pTacD).
Construct pTacE (Fig.
1
) has a promoter-less cassette inserted at this polylinker. The cassette consists of
firefly luciferase (
luc
) and phleomycin resistance (
ble
) genes with flanking SAS and PAS derived from
T.brucei
actin (
act
) and aldolase (
ald
) loci, respectively. Very low levels of ble transcription can be selected for,
as indicated by our success in being able to introduce and select for this
marker at `silent Vsg' expression sites (Navarro and Cross unpublished data).
Consequently, pTacE should enable us to `trap' fairly weak promoters. The
luc
unit facilitates analysis of the strength of a trapped promoter in a transient
assay.
Procyclic trypanosomes transformed with
Bgl
II-linearized pTacE could not withstand ble selection, but could if
transformed with linear pTacE-R with an rRNA promoter in the polylinker of pTacE.
We were concerned that the proximity of a `trapped' promoter to a telomere on
these constructs might result in the repression of transcription, as has been
demonstrated in the yeast
Saccharomyces cerevisiae
(
43
-
45
). We addressed this concern by comparing
cat
transcription in pTacB and pTacAm-transformed procyclic trypanosomes. Construct pTacAm (Fig.
1
) is a derivative of pTacA and was obtained by the insertion of a 286 bp
fragment, consisting of the parp-SAS and the first 169 bp of the
hyg
coding sequence, downstream of the
tub
intergenic region. The 286 bp insert was derived from pTacB, and was used to
fulfill the demonstrated need of a downstream SAS in 3' end formation of mRNA in trypanosomatids (
46
-
48
).
Southern analysis of genomic DNA derived from pTacAm-transformed clones indicated that this construct, like its parent pTacA,
existed as an autonomous linear element. No integrated, or otherwise rearranged
molecules were detected (data not shown). RNA analysis indicated that the cat
mRNA derived from pTacAm and pTacB-tansformed clones were of the same size (Fig.
6
top panel), indicating that the transcripts from both clones were similarly
processed as expected. Phosphorimager analysis of band intensities showed that,
when corrected for loading differences, the amounts of cat mRNA in the two
populations were almost identical. Similar results were obtained when we
measured transcriptional initiation from the
parp
promoter driving cat expression, by doing nuclear run-on assays with lysolecithin permeabilized cells (data not shown). We
conclude that transcription from the
parp
B promoter in pTacAm is not significantly different from its level in pTacB
despite the fact that it is ~4.5 kb closer to the telomere in the former construct.
We report the development and characterization of a set of linear artificial
mini-chromosomes for
T.brucei
. A similar construct has been described for
Leishmania
(S. Beverley, personal communication). In addition, linear or supercoiled DNA
injected into the macronucleus of the ciliated protozoan
Paramecium tetraurelia
are maintained as linear molecules. The injected DNA is cleaved at random
points and sequences characteristic of
Paramecium
telomeres are added. However, neither replication nor telomere addition shows
any defined sequence requirement and the linear elements are maintained at a
copy number of ~10 000-50 000 molecules/cell (
49
). In contrast, and as in yeast where they were first constructed (
50
,
51
), the artificial chromosomes described here exist at a copy number
characteristic of natural chromosomes, and the parental episome from which
these were derived displays very specific sequence requirements for autonomous
replication (
23
,
24
). Although these elements are indefinitely stable under selection and do not
integrate into the genome, they are lost fairly quickly on the removal of
selective pressure. We do not know the basis for this loss in stability
compared to that of their parent pT13-11. However, a similar phenomenon has been documented in the yeast
Saccharomyces cerevisiae
, where circular CEN plasmids (i.e. plasmids with centromeric DNA) are
substantially more stable than the linear chromosomes derived from them (
50
). Finally, the estimated growth rate of the telomeric repeats on these
artificial chromosomes agrees with previous data on the growth of natural
chromosomes in
T.brucei
(
32
,
52
). It is interesting to note that the band corresponding to the growing
telomeric fragment in pTacA does not show substantial smearing on prolonged
culture, indicating a fairly uniform growth of telomeres over the entire
population.
As small dispensable genetic elements that bear all the cis-acting signals necessary for autonomous replication, these defined linear
mini-chromosomes should be useful models for the study of trypanosome nuclear
DNA replication and chromosomal behavior. We expect such studies in this
ancient lineage to provide a unique perspective on the evolution of the DNA
replication and segregation apparatus in eukaryotes.
A potential concern in the use of these small linear elements as promoter-traps was the possibility of telomere mediated repression (silencing) of
transcription. A telomeric position effect (TPE) whereby several pol II
promoters are silenced in the vicinity of telomeres has been extensively
studied in the yeast
S.cerevisiae
(
43
-
45
). TPE is mitotically heritable but reversible, and a telomere-proximal promoter shows a variegated pattern of activity, such that some
cells initiate transcription from this promoter while others do not. TPE is
thought to be mediated by a change in chromatin structure, which establishes a
gradient of transcriptional repression originating at the telomeres and
spreading inwards (for ~5 kb in
S.cerevisiae
) along the chromosome. The extent of this silenced domain is affected by many
factors, including the strength of the promoter (
45
).
Our observations indicate that initiation of transcription from the
parp
B promoter is unaffected by its distance from the telomere in procyclic
T.brucei
. At least three earlier studies have reported data consistent with these
observations (
53
-
55
). In each instance, a
parp
or rRNA promoter showed high activity at a telomere-proximal (~4-5 kb from sub-telomeric repeats) position in procyclic
T.brucei
. The absence of a gradient of transcriptional repression between positions that
are ~1.7 (pTacAm) to ~6.2 kb (pTacB) away from the subtelomeric repeats on these artificial
chromosomes is inconsistent with the phenomena of TPE as described in yeast
implying a putative difference in the make-up of telomeric chromatin between yeast and procyclic-form trypanosomes. However, we cannot discount the possibility that
weaker promoters than the one studied here might be subject to telomere-mediated repression. Delineation of Pol II promoters will play a key role
in the understanding of gene expression in trypanosomes.
NA thanks Mary Gwo-Shu Lee for her hospitality in making laboratory space available to
continue her research at Columbia University. This work was supported by NIH
grants AI 21729 (to GAMC), AI 21784 (to LHTVdP) and by a grant from the John D.
and Catherine T. MacArthur Foundation to LHTVdP.
REFERENCES
Return