Aug 26, 1994 - Martin C.Taylor, Rudo Kieft and Piet Borst'. The Netherlands Cancer Institute, Plesmanlaan 121,. 1066 CX Amsterdam, The Netherlands.
The EMBO Journal vol.13 no.22 pp.5470-5482, 1994
VSG gene expression site control in insect form Trypanosoma brucei
Gloria Rudenko, Patricia A.Blundell, Martin C.Taylor, Rudo Kieft and Piet Borst' The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands
'Corresponding author Communicated by P.Borst
When the African trypanosome Trypanosoma brucei is taken up from mammals by a tse-tse fly, it replaces its variant surface glycoprotein (VSG) coat by a procyclin coat. Transcription of VSG genes stops in the fly, but transcription of sequences derived from the promoter area of the VSG expression site(s) remains high. Whether this is due to continuing high activity of one promoter or to low activity of many promoters was unclear. We have used the small differences between the sequences of different expression sites to show that multiple expression site promoters are active in insect form trypanosomes. This is confirmed by the low expression of single copy marker genes introduced into the transcribed area. However, if the expression site promoter is removed from the genomic location of the expression site and inserted in the non-transcribed spacer of the ribosomal DNA (rDNA), it is derepressed. Derepression of transcription can also be accomplished by replacing the promoter of an expression site by an rDNA promoter. We conclude that the down-regulation of VSG gene expression site promoters in insect form trypanosomes is affected by both the DNA sequence of the promoter and the genomic context in which it resides. Key words: genomic context/life cycle-specific regulation/ Trypanosoma bruceilVSG expression sites
Introduction The African trypanosome Trypanosoma brucei is a flagellated protozoan that is transferred from the bloodstream of one mammalian host to another by the tse-tse fly insect vector. In the bloodstream, the trypanosome relies on changing its variant surface glycoprotein (VSG) coat to avoid elimination by the immune system (Vickerman, 1978, 1985). This coat is encoded by one of up to 1000 different VSG genes (Van der Ploeg et al., 1982), which is expressed in one of many discrete telomeric locations known as expression sites. Switching between VSG genes usually involves replacing the active VSG gene with a copy of an inactive gene via gene conversion, but can also involve an in situ switch between expression sites. An important question which is still completely unresolved, is how is the bloodstream form trypanosome capable of rigorously silencing all but one of its many expression
sites? The different facets of antigenic variation have been reviewed in Borst (1986), Donelson (1988), Pays and Steinert (1988), Cross (1990), Van der Ploeg (1990) and Pays (1992). In bloodstream form Tbrucei, expression sites are transcribed as large (40-60 kb) polycistronic transcription units containing various expression site-associated genes (ESAGs) upstream of the telomeric VSG gene (Johnson et al., 1987; Kooter et al., 1987; Pays et al., 1989b; Lips et al., 1993). During differentiation to the insect (also known as procyclic) form, transcription of virtually all of the VSG expression site is turned off (Kooter and Borst, 1984; Kooter et al., 1987) and the VSG surface coat is rapidly replaced by a coat of procyclic acidic repetitive protein (PARP; also called procyclin; Roditi et al., 1989; Dorn et al., 1991; Pays et al., 1993). Transcription of the VSG expression sites is cx-amanitin- and Sarkosylresistant, as is transcription of the ribosomal DNA (rDNA) and PARP genes, giving rise to the speculation that all of these genes may be transcribed by RNA polymerase (pol) I (Kooter and Borst, 1984; Rudenko et al., 1991, 1992; Zomerdijk et al., 1991a,c; Chung et al., 1992). The VSG expression site promoter is not bloodstream form-specific, but is capable of functioning in the insect form of the parasite. Although most of the VSG expression site is not transcribed in insect form trypanosomes, there is transcription of the immediate promoter area of at least one expression site as determined using run-on transcription analysis (Pays et al., 1989a, 1990; Zomerdijk et al., 1990). Additionally, expression site promoters have been shown to be capable of driving high levels of expression of the chloramphenicol acetyl transferase (CAT) gene in transient transfections in both insect- and bloodstream form trypanosomes (Zomerdijk et al., 1990; Jefferies et al., 1991). Levels of activity are at least as high as those from the rDNA or PARP promoters (Zomerdijk et al., 1991c; Gottesdiener et al., 1992). Deletion mapping localized the expression site promoter to a small fragment (-60 to +77) still containing maximal promoter activity (Jefferies et al., 1991; Zomerdijk et al., 1991b). These types of experiment led to the proposal that expression site promoters in insect form Tbrucei are constitutively active and unmodulated, and that all life cycle-specific regulation of mRNA levels occurs posttranscriptionally (Roditi et al., 1989; Pays, 1992). Regulation of gene expression during the life cycle of Tbrucei does not appear to be mediated to a great extent at the level of transcription, but primarily at the posttranscriptional levels of trans-splicing, polyadenylation, RNA stability and possibly translation (Jefferies et al., 1991; Erondu and Donelson, 1992; Hug et al., 1993; Revelard et al., 1993; reviewed in Clayton, 1992; Pays, 1992). An exception to this is the PARP genes, where
5470 57) Oxford University Press
VSG gene expression site control in Tbrucei
nascent RNA analysis of the PARP genes in the bloodstream form indicated that life cycle-specific shut-off of these genes must be at least 5- to 10-fold at the level of transcription (Rudenko et al., 1989; Pays et al., 1990). We therefore addressed the issue of life cycle-specific regulation of the VSG expression site transcription unit in insect form T7brucei. We first determined exactly where transcription of the expression site promoter area occurs. Next we isolated cDNA copies of transcripts derived from this region to establish if multiple expression site promoters were active simultaneously. After integrating promoterless constructs in this transcribed area, we attempted to determine the level of expression of a single expression site promoter. We show that VSG expression site promoters appear to be constitutively active in insect form trypanosomes, but only at a moderate level. The level of transcription from an expression site promoter increases if it is removed from the genomic context of the expression site. Expression from an rDNA promoter introduced into the expression site (resulting in the deletion of the endogenous expression site promoter) is at least 30-fold higher than that from an expression site promoter inserted into the same location.
Results Transcription of the expression site promoter region in insect form Tbrucei Earlier studies in insect form (procyclic) Tbrucei EATRO 1125 showed a high level of transcription predominantly upstream of the first ESAG (ESAG 7), with minor readthrough into this gene (Pays et al., 1989a). As this analysis was performed at a fairly low resolution, we have mapped transcription attenuation more accurately in our Tbrucei 427-060 strain. For our analysis we used subclones of the expression site promoter area of the dominant expression site (DES) from 11 8a' trypanosomes (Figure 1 A; Zomerdijk et al., 1991b). The DES promoter is not duplicated in the 118a' trypanosome variant (see below), but has characteristic 50 bp repeats upstream of a single promoter (Zomerdijk et al., 1990, 1991b; Gottesdiener et al., 1991, 1992). The subclones used extended from the 50 bp repeat array into ESAG 7. These were transferred to nitrocellulose filters and hybridized to nascent RNA from insect form Tbrucei nuclei made in the presence or absence of 1 mg a-amanitin per ml (Figure 1B). The control panel in Figure 1 B shows nascent RNA hybridization to various Tbrucei transcription units. As expected, transcription of the rDNA and PARP clones is resistant to a-amanitin, whereas transcription of the polymerase II transcription units (tubulin and mini-exon) and polymerase III transcription units (5S) is sensitive. There was no significant hybridization to DNA of the negative controls, the Bluescript vector or a construct containing the VSG 118 gene. In the right-hand panel in Figure l B with the expression site subclones, no transcription was detectable with DES 15 or DES 14, which contain sequences found upstream of the expression site promoter localized by Zomerdijk et al. (1990), Pays et al. (1990) and Gottesdiener et al. (1991). DES7, containing -120 bp of transcribed sequences, hybridized at a high level with nascent RNA, as did DES8, DES9 and DES 10. There was a low but significant amount
A
50 tp repeats l
SHSci
;-
..I
.:1I
.
ESAG 7
>
Sir
I
...
E Sl Sphl Rs!
Stl
|
H D I _EME
15
14
10 0.5 kb
B
Expression site subclones
Controls
_
rDNA
N
N
+
Blue
DES15 Blue
14
tub
7
-
-
5S
8
_
_
VSG 118
PARP
9
_
_
m-
10 11
minl4xon
12
Fig. 1. (A) Transcription of subclones of the expression site promoter area from the DES. In the trypanosome variant that these DES subclones were isolated from (118a') there is no duplication of the promoter repeat. The promoter is indicated with a flag, the hatched boxes indicate the 50 bp repeat arrays and the black filled rectangle the most upstream of the Expression Site Associated Genes (ESAG 7). The lines under the map indicate the various subclones used in the run-on transcription experiment in (B). Only relevant restriction enzyme sites are indicated. SI, Sall; SpI, SpeI; H, HindIll; Scl, ScaI; E5, EcoRV; RsI, RsaI; StI, StuI; D, DraI. (B) Run-on transcription analysis of expression site subclones in procyclic Tbrucei. Nascent RNA made in the presence (+) or absence (-) of 1 mg a-amanitin per ml was hybridized to DNA immobilized on slot-blots. The abbreviations used are: rDNA, ribosomal DNA; Blue, Bluescript SK; tub, tubulin; PARP, procyclic acidic repetitive protein. A description of the plasmids used is in Materials and methods.
of transcription of clones DES 11 and DES 12. Transcription of the DES 12 clone appeared to increase, which can be explained by the presence of an almost exact duplication of these sequences in the downstream ESAG 6 gene (Zomerdijk et al., 1991 b). Presumably a low level of readthrough transcription extends into ESAG 6. Despite the larger size of the DES 10 insert (390 bp) compared with DES8 (279 bp) and DES9 (297 bp), nascent RNA hybridization was reduced to 15-20%, indicating that some transcription appeared to be terminating in this fragment. If the amount of transcription per 100 bp is calculated, then DES 10 is transcribed at 10-15% of DES8, and only 1-5% of the transcription of clone DES8 extended into clones DES 11 and DES 12. As expected, all transcription was resistant to a-amanitin. Similar run-on transcription results (data not shown) were obtained with two other 427 procyclic strains (118 clone I in Rudenko et al., 1990), an original 427-060 strain, plus an independent trypanosome isolate [Tbrucei JIO in Gibson and Garside (1990); this data not shown refers to all strains]. Our transcription analysis used subclones from a 'simple' expression site promoter with 50 bp repeats
5471
G.Rudenko et al.
upstream of the promoter and ESAG 7 downstream. However, approximately half of the expression sites in Tbrucei (Gottesdiener et al., 1992; R.Kieft and G.Rudenko, unpublished results) have a more complicated arrangement in which a second expression site promoter is duplicated upstream (Gottesdiener et al., 1991, 1992; Zomerdijk et al., 1991b). Transcription analysis of sequences 3' of the upstream DES promoter, analogous to clone DES11, showed a similar decrease in the level of transcription (unpublished results). This indicates that transcription from upstream expression site promoters, similar to transcription from 'simple' or downstream expression site promoters, also terminates close to the promoter.
Multiple VSG expression site promoters are simultaneously active in procyclic Tbrucei Pays et al. (1989a) have proposed that transcription of the VSG expression site promoter area in procyclic Tbrucei could be explained as being due to a high level of transcription from a single expression site, this being the expression site that was last active in the bloodstream form trypanosome before differentiation into the procyclic form. To verify this, we cloned cDNAs derived from the expression site promoter region from procyclic trypanosomes recently differentiated from bloodstream form trypanosomes that had used known expression sites. In our Tbrucei 427-060 stock, the active VSG 221 gene is normally transcribed in an expression site located on a 2 Mb chromosome (band 15 in Gottesdiener et al., 1990). In contrast, the expressed copy of the VSG 118 gene is normally located in an expression site on a 1.5 Mb chromosome (band 10) which is referred to as the DES. In Figure 2A we compare the sequence of the promoter area of these two expression sites. The sequences are highly similar but not identical, and the differences should enable us to determine if one or many expression sites are transcribed in procyclic trypanosomes. By taking procyclic trypanosomes with a known history we should be able to determine whether there is predominantly transcription of the expression site that was previously active in the bloodstream form. We took bloodstream form trypanosomes (221a from Bernards et al., 1984) expressing the 221 VSG gene in the 221 expression site (221ES) and differentiated these in vitro into the procyclic form. RNA was isolated from these newly differentiated 221 procyclics (-2 months in culture); cDNAs were made from the expression site promoter region transcripts via reverse transcription (RT) and amplified using PCR. The oligonucleotide sequence of the primer used for reverse transcription was identical in sequence to both the 221 ES and the DES (labelled RT in Figure 2A). The sequences of the primers used to PCRamplify the cDNAs were identical to the two expression sites (Sp) or very similar (3p). In Figure 2B the sequences of cDNAs derived from the VSG expression site promoter region are shown compared with the sequences of the promoter regions of the 221 ES (Esl) or the DES (Es2). Seven different cDNA sequences derived from the 221 procyclics are shown underneath prefixed by 2p (2p-Es 1 to 2p-Es8). Only one cDNA is identical to the sequence from the 221ES (2p-Esl), all of the other cDNAs differ in multiple nucleotides, and most
5472
of these variant cDNAs appear to be more similar to the DES than the 221 ES (see for example 2p-Es3 or 2p-Es6). This RT-PCR experiment was repeated with RNA isolated from procyclic trypanosomes (118 clone I procyclics) differentiated from bloodstream form trypanosomes that had last expressed the 118 VSG gene in the DES (Rudenko et al., 1990). Of these cDNAs (8p-Esl to 8p-Es13 in Figure 2B), two out of the 10 cDNAs were identical to the DES (8p-Es2), one was identical to the 221ES (8p-Esl) and the remainder were significantly different. One of the different expression site sequences (8p-Es3) had been found independently in duplicate cDNAs generated from the 221 procyclics (2p-Es3). Here, as in the above experiment, duplicate cDNAs were derived from independent PCRs. In contrast to the procyclic RNA, the RNA from 221 a bloodstream form trypanosomes yielded mainly one sequence. Of 35 cDNAs sequenced, 33 were identical or differed in only a single random nucleotide (attributed to low fidelity of Taq polymerase) from the 221ES (BsEsl in Figure 2B), while two cDNAs were significantly different (Bs-Es9). These corresponded to a previously identified expression site sequence (8p-Es9). The main concern with this type of PCR-based approach is the possibility of PCR amplification of genomic DNA that has not been removed by the DNase I treatment of the RNA. To control for this we PCR-amplified mock RT reactions (no addition of reverse transcriptase) from the various DNase I-treated RNA preparations. Similarly, we PCR-amplified reactions where reverse transcriptase had been added but no RNA. These control reactions were run on an agarose gel; no PCR product was visible either by ethidium bromide staining or after Southern blotting and hybridization with an expression site promoter probe. In contrast, an abundant product of the correct size was always present using either procyclic or bloodstream form Tbrucei RNA preparations dependent on the presence of reverse transcriptase. A second concern is the possibility of the generation of sequence differences during the PCR amplification step. Although we have occasionally observed single nucleotide changes in PCR-derived clones, we have never observed multiple mutations in a relatively short DNA stretch comparable with the 130 nucleotides we have analysed. In addition, the vast majority of the bloodstream formderived cDNAs were either identical or had a random single nucleotide difference from the 221ES (Bs-Esl). It was notable that two independent bloodstream formderived cDNAs differed in 10 nucleotides from the 221 ES (Bs-Es9), and this identical expression site sequence had been independently found among the 118 procyclic cDNAs. The isolation of two non-221 expression sitederived cDNAs from the bloodstream form is intriguing, as the trypanosomes appear to be exclusively 221 expressors by immunofluorescence. It is possible that a minor percentage of the population is a double expressor for another expression site. The sequences of these procyclic cDNAs suggest that multiple (all?) expression site promoters are active in procyclic trypanosomes. Despite the limitations of RTPCR, whereby transcripts with 100% homology to the primers will be preferentially amplified, we obtained a range of transcript sequences. It appears therefore that in
VSG gene expression site control in Tbrucei
A
1 22 les Des
v
5p
v
90
CATATCCCTATTACCACACCAGTTTATATTACAGGGGAGGTTATTACAGAAATCTAGGATATCAGACTCAAGGTG1GGCGGACGTCTCGA ..................... T.A ............... C .
.G.. C.............. C ...................
91
180
22 les ACCGATCGTGAT
CCTGAGCGTGCCGTCGTTACCGGAGTGCCCGTATTCGATTTATGTCGACTGTGTGGCTGCGAAAACCCTTAGCG
Des
..........................................................................................
22 les
TTACCACTGTGACGATTGCGTGCGATGCAAAGAGACCGTTGTCGGTTCATGCATGACTTTAATCTCATCTACGCTTTATCCCGTGGTTCC
181 Des
270
.......
A.. T .CC.A.
.
271
360
22les TTTGTCCATTGAGGT=TCTCTCTGGGACGGGCTGAGCGTATTTTAGAGGGCTGGTGTGTTGTTTTGCATGCGAAGAGCAGGGG*TGCAAC Des .GC . G.A .C. T.A. C. .C.
3p
361
450
22les Des
GGAAGAGTGAAAGGATTAGAGGAAATGGTGGAATT CATGATCGGAGATTGTGAGGGTTA AAAAACGTGAAGAGGGAAATGGTAACT ........... .G... .....AG. .T. GC.
221es
451 477 RT jGGACATCGTGGAGAGCGG
Des
B
22les-Esl
I
2p-EsS 2p-Es6 2p-Es7 2p-Es8 8p-Es2 8p-Esl 8p-Es3 8p-Es9 8p-EslO 8p-Esll 8p- Esl2 8p-Esl3 Bs-Es 1 Es- Es9
130
CGCTTTATCCCGTGGTTCCTTTGTCCATTGAGG1=CTCTC
Des -Es2 2p-Es1 2p-Es3 2p - Es4
CC.A
TGGGACGGGCTGAGCGTATTTTAGAGGG TTTT_TOCATGC GAAGAGCAGGGG*TGCAACGGAAGAGTGAAAGGATTAGAG .C..C.C.G..G
..GC
C.GA.................G.. ..G.T.A. C.
C....
..................C C
GA G.A
C
A .T.A.
G.A
G..
C
C.C
.G.T.A.
C
G.A
.............. C............... C
.0..T...
T. .G .G .T G.G. .T...
*. *
T.A..C..C
A............
.A CC C
T. C..... A
....C.. ....
.T .A
.C
C
..... .C...
C
G
G
T
.G
G
OG
.GO............ I.
............ ............ C..... C . .......... G-A ................................A ..
...............
.............. ................................T
2
G
G
GTA
G .G .A
2
.G..T...
T
C.
. CC.A.GC
G.G. '... G.T...
C
C......... G.
T.A.
..........................
.C
N
G.A.C.TA
.T
2
T
.G .T T
.. ...........G.T.A
C..
..............A
GO..
..T...
33 2
Fig. 2. (A) Sequence comparison of the VSG expression site promoter region of the 221 expression site (22les) and the DES. Dots indicate identical nucleotides and an asterisk indicates a gap in the sequence. The inverted 'v's above the sequence indicate the 5' ends of the two expression site promoter area transcripts mapped by Zomerdijk et al. (1990). The boxes indicate the position of the various oligonucleotides used for RT-PCR. RT indicates the position of the oligonucleotide used for reverse transcription of the VSG expression site transcripts, and 5p and 3p the positions of the respectively 5' or 3' oligonucleotides used to amplify by PCR the cDNAs generated by reverse transcription. The sequence of the oligonucleotides was chosen to be identical (RT and 5p) or highly similar (3p) to both of these expression sites. (B) Sequence alignment of cDNAs made from transcripts of the VSG expression site promoter region and amplified by PCR. The sequences of the 221 expression site (221es-Esl) and the DES (Des-Es2) promoter regions are indicated above. cDNAs indicated with 2p were generated from RNA from 221 procyclic trypanosomes. cDNAs indicated with 8p were generated from RNA isolated from 118 clone I procyclics. cDNAs indicated with 'B's were generated from RNA isolated from bloodstream form VSG 221a trypanosomes. The different expression site sequences generated are indicated with the numbers Esl -Es13. The number of cDNAs with a given sequence is indicated on the right under 'N', and was one unless otherwise indicated. cDNAs generated from the bloodstream form 221 a RNA and deviating from the 221 expression site sequence in only a single random point mutation were considered as identical and are included in the total of 33. Identical nucleotides are indicated with dots, and a gap in the sequence with an asterisk.
procyclic trypanosomes the expression site last active in the bloodstream form does not appear to be the template for the bulk of the transcripts from the expression site promoter region.
Integration of a selectable marker in expression sites To determine the level of transcription from a single expression site promoter, we integrated unique sequences in this location. This was performed with promoterless constructs targeting a single copy marker gene behind an endogenous expression site promoter (Figure 3A). The hygromycin resistance gene was flanked 5' by a DNA segment containing an cx-tubulin gene splice site and 3' by a tubulin intergenic region including the a-tubulin gene polyadenylation signal. After integration via homologous recombination, the construct was inserted into the SphI site between DES8 and DES9 (Figure 1A) which we have shown to be in the highly transcribed region. The high level of sequence homology (>95%) between different
expression site promoter sequences (Pays et al., 1990; Zomerdijk et al., 1990; Gottesdiener et al., 1991, 1992) would presumably make it possible to target behind several expression site promoters.
With this construct, the hygromycin resistance gene targeted with a high efficiency into different expression sites in procyclic trypanosomes. In most of the experiments at least one in 103 trypanosomes counted 48 h after electroporation was transformed. After cloning the transformants by limiting dilution in microtitre dishes, we selected three transformants for further analysis (Figure 3B). The chromosomal location of the hygromycin resistin the transformed trypanosomes was determined using pulsed field gel electrophoresis (PFG; data not shown). Transformant enES-1 had the hygromycin resistance gene integrated in chromosomal band 2; enES2 and enES-3 both had the hygromycin resistance gene in chromosomal band 10, which has a single expression site (the DES) with a duplicated promoter structure in our procyclic trypanosome strain (Zomerdijk et al., 1991b; ance gene
5473
G.Rudenko et al.
A
Run-on transcription ESAG 7
Spi
HpI
H
e enES
E5 Si
Sphl
I
ll.X
hygro
7
H
D
Rsl
B chrom. hygro
50 bp repeats
enES- 1
D
ESAG 7
RI N
D PH RI
hygro RI N
ESAG 7
2
enES-2
p
enES-3
RI
N
H D
hygro HDfS
RIN
D
DU-R
_
10
ESAG 7
RI
HE5
N
10 500
bp
Fig. 3. (A) Integration of a selectable marker behind an endogenous expression site promoter. The map indicates a typical expression site promoter region (here from the DES). The flag indicates initiation of transcription, and the dashed line above the map indicates the extent of most of the nascent RNA transcription. The black filled rectangle indicates the most upstream ESAG (ESAG 7). Underneath the map is indicated the construct used for targeting behind an endogenous expression site promoter (enES), with the relationship of the target fragments to the expression site indicated with dashed lines. The white boxes in the construct indicate 5' the tubulin trans-splice acceptor site, and 3' the tubulin intergenic region containing the polyadenylation signals. The black and white striped box indicates the hygromycin resistance gene. Only relevant restriction enzyme sites are indicated, and are abbreviated as indicated in the legend to Figure IA. (B) Maps of transformants having the hygromycin resistance gene integrated behind expression site promoters. On the right is indicated the chromosomal band that the selectable marker has integrated into. Transformant enES-1 has the hygromycin resistance gene integrated behind an unduplicated expression site promoter in chromosome 2. Simple sequence (presumably 50 bp repeats) is upstream of the promoter and is indicated here with hatched boxes. enES-2 has the hygromycin gene integrated behind a downstream expression promoter in the DES in chromosome 10. The -13 kb sequence between the duplicated promoter repeats is only shown in part, and is indicated by a discontinuity in the line. Transformant enES-3 has the marker integrated behind an upstream promoter in the same expression site. The probes used in Figure 3 are indicated below: h is the -600 bp NcoI-HindIII fragment from the hygromycin resistance gene, d is the insert from DESI I indicated in Figure IA and k is the 350 bp PstI-KpnI fragment specific for the upstream expression site promoter. Relevant restriction enzyme sites are as indicated in Figure IA, with the addition of RI as EcoRI, N as NcoI, P as PvuII and B as BamHI.
Gottesdiener et al., 1992). Numbering of the chromosomal bands is according to Van der Ploeg et al. (1989) and Gottesdiener et al. (1990). Linkage of the hygromycin resistance gene to expression site sequences was shown by Southern blotting (Figure 4A). Maps of the integration events are shown in Figure 3B. All marker genes integrated either behind expression site promoters flanked upstream by simple sequence DNA, presumably 50 bp repeats, or into expression sites with a duplicated promoter structure (Zomerdijk et al., 1990, 1991 b; Gottesdiener et al., 1991, 1992). Fragments hybridizing with the hygromycin probe (probe h in Figure 3B) also hybridized with expression site sequences upstream of ESAG 7 (probe d) for enES- and enES-2, or sequences specific for the upstream promoter repeat of the DES (probe k) for enES-3. For a more detailed description of the linkage analysis, see Materials and methods. 5474
The level of expression of the selectable marker in all three transformants was determined by Northern analysis. In Figure 4B, levels of hygromycin mRNA from enES-1 (lane 2), enES-2 (lane 3) and enES-3 (lane 4) were compared with levels of hygromycin mRNA from a trypanosome transformant where the hygromycin gene was inserted into a tubulin transcription unit [lane 5; transformant tubhy constructed essentially as described in Ten Asbroek et al. (1990), but with the replacement of an ax-tubulin gene; G.Rudenko and P.Borst, unpublished results]. The various expression site transformants showed consistent differences in promoter strength. This range in levels of expression is presumably due to differences in promoter sequence and exact genomic context (distance from various repeat arrays) in the different expression sites targeted. Figure 4B shows comparable signals from the enES transformants (lanes 2-4) and the tubhy transformant (lane 5) in which the hygromycin gene is tran-
. ,~ I
VSG gene expression site control in T.brucei
A
B 1
h d
3
r-
h d _
h
1
2
3
4
5
k w
"
r ygrO
"
Fig. 4. (A) Linkage of the hygromycin resistance gene to expression site sequences. Panels 1-3 show NcoI-HindIII digests of enES-l (panel 1), enES-2 (panel 2) and enES-3 (panel 3) genomic DNA. Lanes marked with h were hybridized with the downstream hygromycin probe (h indicated in Figure 2B). Lanes marked with d were hybridized with the DES 11 probe upstream of ESAG 7 indicated in Figures 1 and 2B (probe d). The lane marked with k was hybridized with the DES upstream promoter-specific probe PstI-KpnI fragment indicated in Figure 2B (as probe k). (B) Relative expression levels of a single copy marker gene integrated behind different expression site promoters compared with the same gene integrated into the tubulin transcription unit. Northern analysis of 20 ig total RNA separated in a 1% formaldehyde agarose gel. Lane 1, RNA from wild-type insect form trypanosomes; lane 2, RNA from transformant enES-1; lane 3, RNA from enES-2; lane 4, RNA from enES-3; lane 5, RNA from tubhy, a transformant where the hygromycin gene was integrated into a tubulin transcription unit. The panel was hybridized with a hygromycin gene probe. The lower strip shows the same blot hybridized with a tubulin-specific probe to control for amounts of RNA loaded.
scribed by RNA polymerase II. As the tubulin genes are transcribed at a much lower rate than VSG gene expression sites (see below), the expression site promoters of procyclic trypanosomes must therefore be partly repressed.
Expression site promoter manipulations To test whether the partial repression of the expression site promoter in insect form trypanosomes was dependent on the sequence of the promoter, we replaced the expression site promoter by an rDNA promoter. rDNA promoters resemble expression site promoters based on similarities in their transcriptional behaviour in the presence of axamanitin or Sarkosyl (Van der Ploeg, 1991; Chung et al., 1992; Rudenko et al., 1992). We designed a construct which deletes the endogenous expression site promoter upon integration into the expression site, and replaces it with an rDNA promoter directing expression of a hygromycin resistance gene (RPX 1; Figure 5). An analogous construct was made using an expression site promoter (ESX 1). Both types of construct integrated very efficiently (at least one in 103 trypanosomes counted 48 h after electroporation was transformed). Integration into the correct genomic location and chromosomal localization of the selectable marker were determined in clonal lines using Southern blot and PFG analysis (results not shown; see Table I in Materials and methods for the exact genomic location of the hygromycin gene in the various transformants). Two rDNA promoter exchange transformants were analysed for levels of hygromycin gene expression (Figure 6, lanes 2 and 3), compared with two transformants having an expression site promoter replacement (lanes 4 and 5).
50 bp repeats
UZZECCMZEECCCCCL--
2) Inversion
,
II
ESAG 7 9
Sti Stl
N
N
N, 1 ~~~~~~ ~~RPX1
/ t
/
Si
HSd
/
1) Replacement
3) Duplication
>d ~54 phi Rsi E~SSl
HE S
Sl
Si
1>
ESX1
ESXO
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
,,
ESAG 7
Fig. 5. Schematic diagram of constructs addressing the effect of promoter replacement, inversion or duplication in the expression site. The top map indicates the expression site, and the various constructs are shown underneath with the derivation of their target fragments indicated with dashed lines. The black flag indicates the rDNA promoter, the white flag the expression site promoter, the boxes with black and white stripes the hygromycin resistance gene, and the thick black line plasmid sequences present after integration. The vertical arrow in front of the promoter of the last construct indicates linearization by the unique restriction enzyme HpaI before transformation. The symbols used in the map and abbreviation of restriction enzymes are as indicated in Figure 1A.
RPX l-1 can be compared directly with ESX 1-1 (lanes 2 and 4), as the constructs had integrated into the same genomic location. Clearly, the rDNA promoter generates at least 20- to 50-fold more hygromycin mRNA than the expression site promoter. The rDNA promoter does not appear to be significantly less active when integrated into the expression site than in its usual location in the rDNA
5475
G.Rudenko et al. 1
2 34 5 6
A
7 8 9
EX1
WT
RPX
4.*
+ shy
ashy
B hygro
BI 15 14 7
im
a
m
9 10
-
21
-_
mo 4m _s
-m
-. _-
- -m_
__
short ex.
w..~~~&, so
.., "
:,
tub
6.i.
*"""
rD
BI
H
tub
5S
_
__
-
_
-
_
Fig. 7. Run-on transcription analysis using nuclei isolated from ESX II (EX 1) and RPX I-I (RPX) trypanosomes. WT indicates the experiment performed with wild-type insect form Tbrucei. (A) Slotblots of M 13 single-strand DNA containing the sense strand (shy) or antisense strand (ashy) of the hygromycin resistance gene. (B) The same expression site subclones used in Figure I B. (C) Hybridizations
Fig. 6. Northern analysis of expression site transformants. 20 .g total RNA were analysed per lane. Lane 1, RNA from wild-type trypanosomes; lane 2, transformant RPXl-l; lane 3, RPX1-2; lane 4, ESXI-l; lane 5, ESXI-2; lane 6, ESXO-I; lane 7, ESXO-2; lane 8, scES-1; lane 9, scES-2. (See Figure 5 for a schematic of the constructs.) The top panel was hybridized with a hygromycin gene probe. The panel marked 'short ex' was exposed for a twelfth of the time (2 h) than the upper panel was (24 h). To check for loading of equal amounts of RNA, the strip marked 'tub' shows the same blot hybridized with a tubulin-specific probe.
rDNA; Bl, Bluescript; tub, tubulin; 5S, 5S rDNA gene. The experiments were performed in the presence (+) or absence (-) of 500 ,ug oa-amanitin per ml.
(G.Rudenko and P.Borst, results not shown). As the rDNA and expression site promoters have comparably high levels of activity when they are compared directly in transient transfections (Zomerdijk et al., 1991b), we infer that the expression site promoter is repressed in the genomic location of the expression site. This repression cannot be overcome by simply replacing the promoter, as expression site promoter exchange transformants (ESX 1) showed the same low levels of expression as transformants where the selectable marker was targeted behind endogenous expression site promoters (enES). We next determined whether the expression site promoter could escape repression by inversion or duplication in the expression site. The constructs designed to invert the expression site promoter (ESXO) or duplicate it after integration via a single cross-over targeting event (scES) are shown in Figure 5. These manipulations did not appear to significantly affect promoter activity, as the hygromycin mRNA level was not consistently higher in these transformants (Figure 6, lanes 6-9). We can directly compare ESX with ESXO- I (Figure 6, lanes 4 and 6), and scES-l (Figure 6, lane 8) with enES-2 (Figure 4B, lane 3), as the chromosomal locations were identical. All the constructs targeted into expression site promoter areas integrated into different types of expression site promoter,
but preferentially into expression sites present on chromosomal bands 2 and 10 (unpublished results). This appears to be caused by a preference for targeting into expression site sequences that are most homologous with the target fragments. If we use target fragments derived from another expression site with a divergent sequence we obtain preferential targeting into a different set of chromosomes, but the levels of expression of the marker gene are analogous to the results shown here (manuscript in preparation). The Northern blot in Figure 6 shows relative levels of steady state mRNA obtained from the various transformants. Run-on transcription analysis, however, gives a better indication of relative promoter activities. We compared levels of promoter activity in the promoter exchange transformants (ESX 1-1 and RPX 1-1) by isolating nascent RNA and hybridizing it to DNA immobilized on nitrocellulose filters (Figure 7). The controls for this experiment (Figure 7C) are essentially as described in Figure 1. Nascent RNA from transformant ESX 1-1 hybridized weakly to the antisense strand of the hygromycin resistance gene in the single-strand MP18 vector (ashy), compared with the strong signal seen using nascent RNA from RPXl-1. There was no clear hybridization with the sense strand (shy) in either transformant, and transcription of the single copy marker gene was cx-amanitin-resistant as
l-l
5476
with several of the control transcription units used in Figure I B. rD,
VSG gene expression site control in Tbrucei
expected. Hybridization to the slot-blots was quantitated and, after using the control transcription units to compensate for specific activities of the nascent RNA probe, we calculated that transcription of the hygromycin gene in the expression site was at least 30-fold higher when it was directed by the rDNA promoter (RPX1-1) compared with the expression site promoter (ESXl-l). As we had mapped a putative site of transcription attenuation in the fragment corresponding to DES10 in Figure 1, we were interested to see if transcription directed by the rDNA promoter would also terminate at the same place. We can compare transcription of the expression site subclones in wild-type Tbrucei (Figure 7B, see WT) and in the RPX1-1 transformant (RPX). It is clear that transcription of clones DES 10, DES 11 and DES 12 is higher in the RPX 1-1 transformant than in wild-type Tbrucei. Quantitation of the signals (see Materials and methods) showed that no significant termination of transcription occurred downstream of the rDNA promoterdriven hygromycin resistance gene.
An expression site promoter removed from the expression site is derepressed As inverting the orientation or duplicating the expression site promoter appeared to have no significant effect on expression of the marker gene, we wondered if inserting it in another genomic location would allow the escape of repression. Zomerdijk et al. (1993) inserted a selectable marker driven by an expression site promoter in the inverse orientation into a tubulin transcription unit, and reported a very low level of expression of the selectable marker which was undetectable by run-on analysis. However, as pointed out by the authors, this experimental design was complicated by the possibly adverse effects of large amounts of antisense tubulin RNA if the expression site promoter operated at a high rate. We therefore integrated a construct driven by a VSG expression site promoter into one of the few known regions in trypanosomes where transcription is undetectable: the non-transcribed spacer of the ribosomal DNA. The ribosomal transcription unit of Tbrucei is -10 kb long, as shown by nascent RNA analysis combined with UV irradiation mapping (White et al., 1986; Johnson et al., 1987; Coquelet et al., 1991). We have not detected transcription of the spacer DNA between rDNA transcription units using run-on transcription analysis. The expression site promoter directing expression of the hygromycin gene was targeted in front of an rDNA transcription unit via a single cross-over targeting event (Figure 8, rDES 1). By inverting the target fragment in the construct it was possible to integrate the selectable marker in the inverse orientation relative to the ribosomal transcription unit (rDESO). Both of these approaches worked and transformants were generated with high efficiencies. The marker genes were shown to be linked to DNA encoding the 18S rRNA, and were in chromosomal bands already shown to include rDNA (results not shown). Using Northem analysis we determined the levels of expression of the hygromycin gene in two clonal rDES1 transformants and one clonal rDESO transformant (Figure 9). Levels of hygromycin mRNA were increased at least 10- to 20-fold (see lanes 5 and 6) compared with hygromycin mRNA levels in the transformant containing the
18S _H
Hc
Rli
rDES 1
18S
s
Hc C
P P PH
C
Hc
r
rDESO Hc C
I
b~ ~ ~ ~ ~ ~ ~ -s
H
P
18S
P PP
I
C
...
Hc
Ir
H
P
Fig. 8. Integration of the expression site promoter into the nontranscribed spacer of the rDNA. The black flag indicates the beginning of the rDNA transcription unit, and the large black box the 18S rRNA gene. The circle underneath the map shows the construct with the 1.3 kb HincIl target fragment and the unique ClaI site used for targeting of the constructs via a single cross-over targeting event. rDES1 contains the expression site promoter (indicated with a white flag) in the same orientation as the downstream rDNA promoter, and rDESO the expression site promoter in the inverted orientation. The arrows above the integrated constructs and the rDNA transcription units indicate the direction of transcription. The filled black box with white stripes indicates the hygromycin resistance gene; the white box with thin black stripes indicates the fragment containing the expression site promoter. The thick black line indicates the plasmid vector. Hc, Hincdl; C, ClaI; P, PstI; H, Hind!ll.
hygromycin gene directed by the expression site promoter in an expression site (ESX I-1 in lane 3 or enES-1 in lane 8). Inverting the construct containing the expression site promoter and hygromycin gene in front of the ribosomal array (lane 7) did not have a significant effect on the level of expression of the selectable marker, indicating that the hygromycin mRNA seen was not derived from transcription through the ribosomal spacer. To compare the transcriptional activity of the expression site promoter inside (ESX 1-1) and outside (rDES 1) the expression site, we performed run-on transcription analysis (Figure 10). The control panels were essentially as shown in Figure 1, only in this experiment the Bluescript control is transcribed in an x-amanitin-insensitive fashion in the rDES 1 transformant. This is because the plasmid vector is located downstream of the expression site promoter after the constructs are integrated via a single cross-over targeting event (Figure 8). As there are sequence elements in common between Bluescript and MP18 (the lacZ and lacI genes), this x-amanitin-resistant transcription of Bluescript also contributes to the hybridization of the single-strand MP18 DNA in these rDES1 transformants. This background level of hybridization therefore needs to be subtracted before determining the relative levels of transcription in the transformants. The run-on transcription experiment results parallel the Northern results. In the ESX1-1 transformant, where the expression site promoter is in the genomic context of the expression site (Figure IOA, EXI), there is a very low but significant amount of transcription detected using the antisense strand of the hygromycin gene as a probe. In the rDES 1-1 transformant, where the expression site promoter has been moved into the spacer of the rDNA (Figure 1 OA, rD I), the level of transcription of the antisense strand of the hygromycin gene is much higher,
5477
G.Rudenko et al.
even after subtracting the background due to vector transcription. We quantitated levels of hybridizing nascent RNA, subtracted the background transcription of the sense strand control caused by transcription of the Bluescript vector, and compensated for the specific activity of the nascent RNA probe using the control transcription units. We estimate that the transcriptional activity of the expression site promoter is at least 10-fold higher when it is removed from the genomic location of the expression site and placed into the rDNA spacer.
234 5 6 7 8 9
1
Discussion
Pt _
hygro
W
a
short
ex. short ex.
w gg
" w p " * _
tu b
Fig. 9. Northern analysis of transformants comparing transcription of the expres sion site promoter integrated in either its natural location or the
non-trn anscribed
spacer of the rDNA. Lane 1, wild-type insect form
ane 2, RPXl-l; lane 3, ESXl-l; lane 4, ESX0-1; lane 5, 1-1; lane 6, rDESl-2; lane 7, rDESO-1; lane 8, enES-l; lane 9, tubhy. 20 ,g total RNA were loaded (see Figure 7 for a schematic of ucts). The top panel was hybridized with a hygromycin gene probe. The- strips showing the short exposure (10 min at -70°C compared with 3 h) and the rehybridization of the blot with a tubulinspecific pr^obe are indicated below. Tbrucei;
1
rDES
the constn
EX1
WT A
_
_
rD
_ _
BE tub 5s
are
Only
rg
froments
expression site and placed in another genomic location, did the promoter have high levels of activity. Although the cDNA analysis establishes that multiple expression sites are active in the trypanosome population, it does not exclude the possibility that only one site is on in individual trypanosomes, but not the same site in each trypanosome. This is intrinsically unlikely because it
_ K
_ -
_
_
promoters would occur at a much higher rate in insect form than in bloodstream form trypanosomes to account for the profound cDNA heterogeneity observed in trypano*D _ h1^_somesonly 2 months after the transition from bloodstream _toinsect form. This model is also unlikely because transfection efficiencies with hygromycin constructs _ targeted into the expression site were so high. For example, the enES construct integrated with a transformation efficiency of one in 103 trypanosomes (counted 48 h after electroporation). This integration efficiency is equivalent
Fig. 10. R,un-on transcription analysis comparing levels of transcripti(on directed by the expression site promoter in different genomic e nvironments. The experiments were performed with nuclei isolated friom wild-type trypanosomes (WT), the ESXl-l transformant (EX1, see also Figure 7) and the transformant with the expression site promoter iintegrated in the right orientation in front of the rDNA (rDES1 in dicated here as rDl). (A) The sense strand of the hygromyciin resistance gene (shy) or antisense strand (ashy) cloned in M13. (B)'The transcription of several control transcription units (Figure 6C ). These experiments were performed in the presence (+) or absence (--) of 500 kg a-amanitin per ml.
5478
rDNA promoter are not. Only in targeting experiments e d e site promoteris re the th from the is removed site expression where
would require that switching of the partially repressed
shy
3shy
B
rDl
We have mapped transcription of the expression site promoter area in insect form Tbrucei and have shown that most transcription terminates in a region -700 bp downstream from the promoter. Using RT-PCR we have isolated cDNA copies of transcripts from this region from insect form trypanosomes with a known history. We show that multiple expression site promoters appear to be simultaneously active and that there does not appear to be a significantly higher level of transcription of the expression site last active in the bloodstream form. We have succeeded in targeting hygromycin resistance genes with high efficiencies into this transcribed region of the expression site. The level of transcription of a single copy gene in a marked expression site is low. In contrast, the rDNA promoter inserted in the same expression site location directs transcription at a level that is at least 30fold higher. Even though the expression site and rDNA promoters appear to have comparable levels of activity in transient transfection experiments (Zomerdijk et al., 199 lb), constructs introducing the expression site promoter into the genomic context of the expression site appear to be down-regulated, whereas constructs introducing the
to
port
Thisintegra
tionsefficien
ivalen
to the most efficient transformations performed in our
laboratory using fragments targeted into the tubulin transcription unit [pUCTbneo3 Sac/Sal in Ten Asbroek et al.
( 993)]. We think that if only one in 20 expression sites
were transcriptionally active, a much lower transformation
efficiency would have been found. These experiments measure the levels of expression of
a single copy sequence which is at the same time a
VSG gene expression site control in T.brucei
selectable marker. A concern with this experimnental approach is that drug selection was used to obtain trypanosomes with integrated marker genes. We think, however, that we have not 'pried' open repressed expression site promoters upon selection with hygromycin. The high transformation efficiencies mentioned above make it unlikely that we have targeted behind a small subset of freak promoters which had enough transcriptional activity to be forced open. The hygromycin gene is an appropriate selectable marker for determining the level of expression of transcription units transcribed at a low level. Transformants having the hygromycin gene inserted in the gene encoding the large subunit of RNA polymerase I (which is transcribed at an extremely low rate) have a comparable IC50 to transformants where the gene is transcribed at a much higher level (G.Rudenko and P.Borst, unpublished results). Presumably very small amounts of protein encoded by the hygromycin gene are enough to give maximal resistance to the drug. Our data are therefore compatible with a life cycle stage-specific deregulation operating on all expression site promoters in insect form trypanosomes. The regulatory mechanism that differentiates between active and inactive expression sites in bloodstream form trypanosomes is presumably absent. All expression site promoters become comparable and are active at moderately low levels. Preliminary experiments indicate that the level of expression from inactive expression sites in the bloodstream form is considerably lower than the levels we describe here in the insect form (results not shown). This means that in the insect form transcription increases from most expression site promoters relative to the bloodstream form, and the previously active expression site is down-regulated. Further experiments need to be performed, however, to verify what happens exactly during the transition from bloodstream to insect form. In nature the insect stage of the trypanosome life cycle is short and the trypanosome has to get out within weeks or die with the fly. It is therefore not possible to extrapolate from the stable procyclic trypanosome lines used in our experiments to the situation directly after the transition from mammal to fly. The down-regulation of the expression site promoters that we observe in procyclic T.brucei appears to operate in the genomic location of the expression site and in an orientation-independent fashion. Such orientation independence is typical of silencers (reviewed in Cowell, 1994), but also of silencing at telomeres in yeast, mediated by DNA-associated proteins altering the chromatin structure (Aparicio et al., 1991; Chien et al., 1993; Renauld et al., 1993). Shut-off of VSG expression site transcription upon cooling of bloodstream form Tbrucei occurs very rapidly, and moves from the telomere towards the promoter (Kooter et al., 1987). This could mean that VSG expression sites, which are invariably on telomeres, could be silenced by attenuation combined with a modified telomere position effect operating over a larger distance than has been observed in yeast. It will be necessary to recreate the silencing phenomenon operating in the expression site in another location to distinguish between the various
alternatives. In contrast to transcription from the expression site promoters mapped in run-on transcription experiments,
most transcription from the rDNA promoter does not appear to terminate in the attenuation region of the expression site. This difference could be due to the inability of the RNA polymerase poised at the expression site promoter to make elongation-competent transcription complexes (reviewed in Krumm et al., 1993). The RNA polymerase could pause, as has been described for the Drosophila heat-shock genes (Rougvie and Lis, 1988), or could be 'predestined' to drop off after picking up a factor, only binding to the expression site promoter in a certain genomic context. Alternatively, the sensitivity for certain termination sequences could be influenced by the strength of the promoter (Meulia et al., 1993). For example, the rDNA promoter through sheer strength (rapid reinitiation of RNA polymerase I molecules), could allow transcription to push through attenuator sequences that are effective for a weaker (down-regulated) expression site promoter. This method of down-regulation puts the critical point of regulation at the level of transcription initiation. A remarkable feature of trypanosomes is their ability to 'remember' the last active VSG expression site when they enter the fly, allowing its preferential re-expression when the trypanosomes move back into mammals (Hajduk and Vickerman, 1981; Delauw et al., 1985). Continued high activity of the single expression site last active before fly entry in the insect stage seemed to provide a plausible explanation for this 'memory' (Pays et al., 1989a). Our results show that this explanation is incorrect, however, and that the trypanosome must remember the last active expression site, even though many (or all) expression sites are switched on in the insect. This suggests that memory must involve stable marking of an expression site. The methods used here to tag VSG expression sites should make it possible to solve this key problem in antigenic
variation.
Materials and methods Strains The trypanosomes used for the experiments were procyclic trypanosomes of the strain 427-060 (Brun and Schonenburger, 1979). The run-on transcription experiments were repeated with nuclei isolated from two different 427-060 strains with lower passage numbers: (i) strain 118 clone I which had been differentiated recently from bloodstream form trypanosome variant 118 clone I (Lee and Van der Ploeg, 1987; Rudenko et al., 1991); and (ii) an original 427-060 procyclic strain provided by W.Gibson, University of Bristol, UK. Run-on experiments were also performed with an independent procyclic trypanosome strain Tb.brucei J1O [MCRO/ZM/73/J10(CLONE A) in Gibson and Garside (1990)].
Nuclear run-on analysis Slot-blots. Plasmids containing inserts specific for various trypanosomal transcription units were denatured with 0.3 M NaOH and transferred to nitrocellulose membranes using a Schleicher & Schuell slot-blotter (Minifold II) according to the instructions of the manufacturer. 2 jg plasmid DNA or 5 jg single-strand DNA were loaded per slot. Control DNA. rDNA (plasmid pR4) contained an 18 kb PstI fragment with an entire ribosomal DNA repeat (Kooter and Borst, 1984). Blue corresponded to the plasmid vector Bluescript SK+ (Stratagene). The plasmid with the tubulin coding region (tub) contained a 3.7 kb HindIll fragment with a single tubulin repeat (Thomashow et al., 1983). The 5S rRNA plasmid had a 700 bp HindIII-EcoRI fragment encoding a 5S RNA transcription unit (Lenardo et al., 1985). VSG1 18 encoded part of a VSG 118 cDNA clone (TcV18.2; Bernards et al., 1981). PARP (plasmid CPT4) contained a PARP cDNA (Rudenko et al., 1989). Miniexon was the clone CLI03 described by Laird et al. (1985). The hygromycin resistance gene (Kaster et al., 1983) was cloned into MP18
5479
G.Rudenko et al. in both orientations, and single-stranded DNA was isolated containing the sense (shy) or antisense strands (ashy). Expression site constructs. Plasmids DES7-15 contain subclones from the DES promoter area cloned by Zomerdijk et al. (1991 b). Progressing from upstream to downstream: DES15 contains a 1.3 kb SalI-SpeI fragment; DES14 a 581 bp Spel-HindlIl fragment; DES7 a 510 bp ScaI-Sall fragment; DES8 a 279 bp EcoR5-SphI fragment; DES9 a 297 bp SphI-RsaI fragment; DESIO a 390 bp RsaI-StuI fragment; DESII a 459 bp StuI-DraI fragment; and DES12 a 576 bp DraI-HindIII fragment. Run-on transcription analysis. Nascent RNA analysis was performed basically according to Rudenko and Van der Ploeg (1989). Nuclei were harvested from procyclic Tbrucei grown to mid-log densities (1.52.5x 107 trypanosomes/ml) using a Stansted cell disruptor and were resuspended at a concentration of 1 x 109/100 [t. Run-on reactions were incubated at 28°C for 8 min before RNA extraction. Hybridization to the filters was as described, except that the post-hybridization washes were at an end stringency of 0.3x SSC and 0.1% SDS at 650C. Hybridization intensities were quantitated using a Molecular Dynamics Phosphorlmager using Imagequant software. To determine if transcription attenuation was occurring downstream of the rDNA promoter, we corrected for specific activity of the nascent RNA probe using the control transcription units. To determine the level of transcription from the single expression site targeted with the strong rDNA promoter, we subtracted the quantitated nascent RNA signal from clones DES9-12 found in the ESX l-l transformant, from the signal quantitated in the RPXl-l trypanosomes. After calculating the additional signal arising from the rDNA promoter after compensating for the relative sizes of the DNA fragments, we showed that there was no considerable transcription attenuation downstream of the rDNA promoter-driven hygromycin resistance gene.
RT-PCR amplification of procyclic expression site transcripts RNA was isolated from logarithmic phase procyclic trypanosome cultures (221 procyclics or 118 clone I procyclics) or 221a bloodstream form trypanosomes using the guanidinium thiocyanate/cesium chloride cushion procedure (Sambrook et al., 1989). 50 ,ug of RNA were treated with 20 U of RNase-free DNase I (Boehringer Mannheim) for 30 min at 37°C in 10 mM Tris pH 7.5, 10 mM MgCl2, 1 mM DTT. After phenol/ chloroform extraction, 10 gg of RNA were ethanol precipitated in the presence of 0.5 ,tM of the antisense RT primer (5'-CGCGGATCCCGCCATCCTCGCGCACGATGTCC-3'). The primer extension was performed essentially according to Ausubel et al. ( 1991 ). The precipitate was dissolved in a Ix aqueous buffer (1 M NaCI, 0.17 M HEPES, 0.3 mM EDTA); after denaturation at 80°C it was annealed at 300C overnight. The annealed transcripts were precipitated with ethanol and cDNAs were generated using AMV reverse transcriptase (Boehringer Mannheim) for 90 min at 42°C in the buffer supplied by the manufacturer. After RNase I treatment, 10 ,ug glycogen were added as a carrier, and the cDNAs were ethanol-precipitated. One sixth of the primer extension reaction was amplified using PCR utilizing 0.5 jM of the antisense 3p primer (5'-CGCGGATCCTAACCCTCACAATCTCCGATCATGC-3') and the sense Sp primer (5'-CCGGAATTCCGGACGTCTCGAACCGATCGTGAG-3') according to Sambrook et al. (1989) in a buffer containing 1.5 mM MgCl2 for 30 cycles (1 min at 94°C, 1 min at 400C, 1 min at 720C) using Taq polymerase (Gibco-BRL). Experimental negative controls included amplification of primer extension reactions without the addition of reverse transcriptase, and PCRs without the addition of cDNA. Southern blot analysis confirmed that these reactions were indeed negative. The PCR products were digested with EcoRI and BamHI (the restriction enzyme sites incorporated in the Sp and 3p oligonucleotides) and cloned into an M 13 vector. Five individual PCRs were performed on each cDNA preparation and the products were cloned independently to minimize the bias introduced during the amplification of a nonabundant transcript. The DNA sequence of individual cDNAs was then determined using di-deoxy sequencing.
(Ligtenberg et al., 1994). The hygromycin gene and tubulin processing signals were cloned using PCR and sequenced before use. All of the expression site constructs used target fragments derived from the DES promoter area, as cloned by Zomerdijk et al. (199lb) from a trypanosome variant (11 8a') with a 'simple' expression site lacking the promoter duplication described by Gottesdiener et al. (1991, 1992). Plasmid enES has as upstream target a 185 bp SalI-SphI fragment located directly downstream of the expression site promoter. The downstream target is the adjacent 688 bp SphI-StuI fragment. Constructs ESXI, ESXO and RPX have as upstream target fragment, the 1.9 kb Sall-HindIlI fragment, extending from inside the 50 bp repeats to 500 bp upstream of the expression site promoter. The downstream target is the 874 bp SalI-Stul fragment extending from the Sall site immediately downstream of the promoter. Constructs ESX1 and ESXO have as the promoter fragment the 510 bp ScaI -Sall fragment containing the DES expression site promoter driving expression of the hygromycin resistance gene with the tubulin RNA processing signals described above. ESX1 has the promoter and the hygromycin resistance gene in the same orientation, and ESXO has the promoter and hygromycin gene inverted relative to the target fragments. A number of restriction enzyme sites are introduced between the upstream target fragment and the promoter sequences. RPX is analogous to ESX1, but instead of the expression site promoter fragment it contains the rDNA promoter on a 518 bp AluI fragment as described in White et al. (1986). All of the above constructs were transfected as isolated fragments. In the construct scES 1, the 1101 bp Spel-Sall fragment was cloned in front of the hygromycin gene with tubulin RNA processing signals, and transfected into trypanosomes after linearization with HpaI. For the targeting of the expression site promoter in the non-transcribed spacer of the ribosomal DNA, the hygromycin gene was flanked by the same tubulin RNA processing signals used for all the above constructs, and the expression site promoter was on the 510 bp Scal-SalI fragment contained in DES7. In the rDES I construct the target fragment was the 1.3 kb HinclI fragment located directly upstream of the rDNA promoter in the non-transcribed spacer DNA of the ribosomal array (White et al., 1986). The construct was linearized with ClaI and targeted via a single cross-over recombination event. In construct rDESO the target fragment was cloned in an inverse orientation, which after targeting integrated the entire construct in an inverse orientation relative to the rDNA transcription unit. Transformation conditions. Trypanosomes were electroporated as described in Zomerdijk et al. (1990) using 5 jig fragments isolated using GeneClean glass milk (BiolOl Inc.) or linearized plasmid. After 3648 h, the surviving trypanosomes were counted and limiting dilution was performed by diluting the culture with successive steps of 3-fold to a concentration of 20 trypanosomes/ml (100 trypanosomes per flask). Wild-type trypanosomes were added to a concentration of 2 x 106 trypanosomes/ml; hygromycin B was added to a final concentration of 25 jig/ ml. Transformation frequencies were typically between one in 103 and one in 104 trypanosomes counted 36-48 h after electroporation.
Table I. Genomic location of the hygromycin resistance gene in transformants Transformant
Chromosome location
Promoter type
enES- 1 enES-2 enES-3
2 10 10 10 10 10 2 10 2 10 2
simple downstream promoter upstream promoter downstream promoter upstream promoter downstream promoter downstream promoter downstream promoter upstream promoter upstream promoter upstream promoter
RPX1-l
Stable transformation of Tbrucei
RPX 1-2 ESX 1-1 ESX 1-2 ESXO-1 ESXO-2 scESscES-2
Plasmid constructs. All of the constructs used for stable transformation contained the hygromycin resistance gene coding sequence (Kaster et al., 1983) flanked by tubulin RNA processing signals cloned into Bluescript KS (Stratagene). The spliced leader acceptor site was present on a 240 bp fragment from the intergenic region upstream of the a-tubulin gene. The polyadenylation signals were present on a 330 bp fragment containing the intergenic region downstream of the a-tubulin gene
Note that chromosomal band 2 comprises at least two different chromosomes (and expression sites). The presence of two different expression sites was confirmed by a BamHI polymorphism in the ESAG 7 gene adjacent to the hygromycin resistance gene of enESI compared with ESXI-2 (results not shown). Chromosome 10 has one expression site corresponding to the DES.
5480
VSG gene expression site control in T.brucei After several passages trypanosomes were analysed. All gene expression
experiments were performed using trypanosomes cloned by limiting dilution in 96-well plates (0.25 trypanosomes/well) according to the same selection procedures.
Analysis of the Tbrucei transformants Genomic DNA was digested with restriction enzymes, electrophoresed on 0.6% agarose gels and analysed using Southern blotting essentially according to Sambrook et al. (1989). Chromosomal location of the introduced selectable markers was analysed using PFG. Total RNA was separated in 1 % formaldehyde gels and blotted onto nitrocellulose membranes according to Sambrook et al. (1989). Southern blot ancalvsis of the eniES tranisformants. Transformant enES- I (Figure 4A, panel 1) had a single band hybridizing with the hygromycin probe (Figure 3B, probe h). This fragment also hybridized with expression site sequences (Figure 3B, probe d), indicating that the hygromycin gene was linked to sequences upstream of ESAG 7. Digests with other enzymes showed restriction enzyme sites already found in ESAG 7 (Figure 3B). Upstream of the expression site promoter area is an area with simple sequence DNA. This presumably consists of 50 bp repeat arrays known to be located upstream of expression site promoters (Zomerdijk et al., 1990, 1991 b; Gottesdiener et al., 1991, 1992). Transformants enES-2 and enES-3 had the hygromycin resistance gene integrated into chromosomal band 10, which has a single expression site (the DES) with a duplicated promoter structure in our procyclic trypanosome strain (Zomerdijk et al., 1991b; Gottesdiener et al., 1992). Transformant enES-2 appeared to have the hygromycin resistance gene integrated behind the 'downstream' DES promoter, as the hygromycin resistance gene could be linked to sequences upstream of ESAG 7 (Figure 4A, probe d in panel 2). Additionally, hybridization of a probe specific for the upstream part of the hygromycin gene indicated the same upstream restriction enzyme sites mapped by Gottesdiener et al. ( 1992) around a 'downstream' DES promoter. Transformant enES-3 (Figure 4A, panel 3) also had a single copy band hybridizing with the hygromycin gene probe; this fragment did not hybridize with probe d, but instead with a probe specific for the upstream promoter repeat of the DES (Figure 3B, probe k). Upstream of the expression site promoter directing expression of the hygromycin gene in this transformant was simple sequence DNA characteristic of 50 bp repeats.
Acknowledgements We are grateful to Drs Wilbert Bitter, Mike Cross, Fred van Leeuwen, Richard McCulloch, Ronald Plasterk and Alfred Schinkel for helpful discussions and critical comments on the manuscript. We thank Keith Gottesdiener for 'upstream expression site promoter'-specific probes and for communicating unpublished results on transcription of this region; Wendy Gibson, Hui-min Chung and Lex Van der Ploeg for providing trypanosome strains; and Hein te Riele for the construct containing the hygromycin resistance gene. We thank Frank Baas and the Academisch Medisch Centrum for help with the Phosphorlmager quantitation and the laboratory of Rene Bernards for the use of computer facilities. This research was supported with financial support from the Netherlands Organization for Scientific Research (NWO). M.C.T. was funded by a Wellcome Trust Travelling Fellowship.
References Aparicio,O.M., Billington,B.L. and Gottschling,D.E. (1991) Cell, 66, 1279-1287.
Ausubel,F.M., Brent,R., Kingston,R.E., Moore,D.D., Seidman,J.G., Smith,J.A. and Struhl,K. (1991) Current Protocols in Molecular Biology. John Wiley and Sons, New York. Bernards,A., Van der Ploeg,L.H.T., Frasch,A.C.C. and Borst,P. (1981) Cell, 27, 497-505. Bernards,A., De Lange,T., Michels,P.A.M., Liu,A.Y.C., Huisman,M.J. and Borst,P. (1984) Cell, 36, 163-170. Borst,P. (1986) Anniu. Rev, Biochenm., 55, 701-732. Brun,R. and Schonenburger,M. (1979) Acta Tropica, 36, 289-292. Chien,C., Buck,S., Sternglanz,R. and Shore,D. (1993) Cell, 75, 531-541. Chung,H.-M.M., Lee,M.G.-S. and Van der Ploeg,L.H.T. (1992) Parasitol. TodaY, 8, 414-418. Clayton,C. (1992) Prog. Nucleic Acid Res. Mol. Biol., 43, 37-66. Coquelet,H., Steinert,M. and Pays,E. (1991) Mol. Biochem. Parasitol., 44, 33-42.
Cowell,I.G. (1994) Trends Biochem. Sci., 19, 38-42. Cross,G.A.M. (1990) Annu. Rei. Immunol., 8, 83-110. Delauw,M.-F., Pays,E., Steinert,M., Aerts,D., Van Meirvenne,N. and Le Ray,D. (1985) EMBO J., 4, 989-993. DonelsonJ.E. (1988) The Biology of Parasitism. Alan R.Liss Inc., New York. Dorn,P.L., Aman,R.A. and Boothroyd,J.C. (1991) Mol. Biochem. Parasitol., 44, 133-140. Erondu,N.E. and Donelson,J.E. (1992) Mol. Biochem. Parasitol., 51, 111-118. Gibson,W. and Garside,L. (1990) Mol. Biochem. Parasitol., 42, 45-54. Gottesdiener,K.. Garcia-Anoveros,J., Lee,M.G.-S. and Van der Ploeg,L.H.T. (1990) Mol. Cell. Biol., 10, 6079-6083. Gottesdiener,K., Chung,H.-M., Brown,S.D., Lee,M.G.-S. and Van der Ploeg,L.H.T. (1991) Mol. Cell. Biol., 11, 2467-2480. Gottesdiener,K.M., Goriparthi,L., Masucci,J.P. and Van der Ploeg,L.H.T. (1992) Mol. Cell. Biol., 12, 4784-4795. Hajduk,S.L. and Vickerman,K. (1981) Parasitology, 83, 609-621. Hug,M., Carruthers,V.B., Hartmann,C., Sherman,D.S., Cross,G.A.M. and Clayton,C. (1993) Mol. Biochem. Parasitol., 61, 87-96. Jefferies,D., Tebabi,P. and Pays,E. (1991) Mol. Cell. Biol., 11, 338-343. Johnson,P.J., Kooter,J.M. and Borst,P. (1987) Cell, 51, 273-281. Kaster,K.R., Burgett,S.G., Rao,R.N. and Ingolia,T.D. (1983) Nucleic Acids Res., 11, 6895-6911. Kooter,J.M. and Borst,P. (1984) Nucleic Acids Res., 12, 9457-9472. Kooter,J.M., Van der Spek,H.J., Wagter,R., d'Oliveira,C.E., Van der Hoeven,F., Johnson,P.J. and Borst,P. (1987) Cell, 51, 261-272. Krumm,A., Meulia,T. and Groudine,M. (1993) BioEssays, 15, 659-665. Laird,P.W., Kooter,J.M., Loosbroek,N. and Borst,P. (1985) Nucleic Acids Res., 12, 4253-4266. Lee,M.G.-S. and Van der Ploeg,L.H.T. (1987) Mol. Cell. Biol., 7, 357-364. Lenardo,M.J., Dorfman,D.M., Reddy,L.V. and Donelson,J. (1985) Gene, 35, 131-141. Ligtenberg,M.J.L., Bitter,W., Kieft,R., Steverding,D., Janssen,H., Calafat,J. and Borst,P. (1994) EMBO J., 13, 2565-2573. Lips,S., Revelard,P. and Pays,E. (1993) Mol. Biochem. Parasitol., 62, 135-138. Meulia,T., Krumm,A. and Groudine,M. (1993) Mol. Cell. Biol., 13, 5647-5658. Pays,E. (1992) In Broda et al. (eds), The Eukarvotic GenomeOrganisation and Regulation. Society for General Microbiology, Symposium 50, pp. 127-160. Pays,E. and Steinert,M. (1988) Annu. Rev. Genet., 22, 107-126. Pays,E., Coquelet,H., Pays,A., Tebabi,P. and Steinert,M. (1989a) Mol. Cell. Biol., 9, 4018-4021. Pays,E., Tebabi,P., Pays,A., Coquelet,H., Revelard,P., Salmon,D. and Steinert,M. (1989b) Cell, 57, 835-845. Pays,E., Coquelet,H., Tebabi,P., Pays,A., Jefferies,D., Steinert,M., Koenig,E., Williams,R.O. and Roditi,. (1990) EMBO J., 9,3145-3151. Pays,E., Hanocq-Quertier,J., Hanocq,F., Van Assel,S., Nolan,D. and Rolin,S. (1993) Mol. Biochem. Parasitol., 61, 107-114.
Renauld,H., Aparicio,O.M., Zierath,P.D., Billington,B.L., Chablani,S.K. and Gottschling,D.E. (1993) Genes Devi, 7, 1133-1 145. Revelard,P., Lips,S. and Pays,E. (1993) Mol. Biochem. Parasitol., 62, 93-102. Roditi,. et al. (1989) J. Cell Biol., 108, 737-746. Rougvie,A.E. and Lis,J.T. (1988) Cell, 54, 795-804. Rudenko,G. and Van der Ploeg,L.H.T. (1989) EMBO J., 8, 2633-2638. Rudenko,G., Bishop,D., Gottesdiener,K. and Van der Ploeg,L.H.T. (1989) EMBO J., 8, 4259-4263. Rudenko,G., Le Blancq,S., Smith,J., Lee,M.G.-S., Rattray,A. and Van der Ploeg,L.H.T. (1990) Mol. Cell. Biol., 10, 3492-3504. Rudenko,G., Chung,H.-M.M., Pham,V.P. and Van der Ploeg,L.H.T. (1991) EMBO J., 10, 3387-3397. Rudenko,G., Lee,M.G.-S. and Van der Ploeg,L.H.T. (1992) Nucleic Acids Res., 20, 303-306. Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Ten Asbroek,A.L.M.A., Ouellette,M. and Borst,P. (1990) Nature, 348, 174-175. Ten Asbroek,A.L.M.A., Mol,C.A.A.M., Kieft,R. and Borst,P. (1993) Mol. Biochem. Parasitol., 59, 133-142. Thomashow,L.S. Milhausen,M., Rutter,W. and Agabian,N. (1983) Cell, 32, 3543. Van der
Ploeg,L.H.T. (1990) In Hames,B.D. and Glover,D.M. (eds),
5481
G.Rudenko et al. Gene Rearrangement. Oxford University Press, Oxford, UK, pp. 51-97. Van der Ploeg,L.H.T. (1991) New Biol., 3, 324-330. Van der Ploeg,L.H.T., Valerio,D., De Lange,T., Bernards,A., Borst,P. and Grosveld,F. (1982) Nucleic Acids Res., 10, 5905-5923. Van der Ploeg,L.H.T., Smith,C.L., Polvere,R.I. and Gottesdiener,K.M. (1989) Nucleic Acids Res., 17, 3217-3226. Vickerman,K. (1978) Nature, 273, 613-617. Vickerman,K. (1985) Br Med. Bull., 41, 105-114. White,T.C., Rudenko,G. and Borst,P. (1986) Nucleic Acids Res., 14, 9471-9489. Zomerdijk,J.C.B.M., Ouellette,M., Ten Asbroek,A.L.M.A., Kieft,R., Bommer,A.M.M., Clayton,C.E. and Borst,P. (1990) EMBO J., 9, 2791-2801. Zomerdijk,J.C.B.M., Kieft,R. and Borst,P. (1991 a) Nature, 353, 772-775. Zomerdijk,J.C.B.M., Kieft,R., Duyndam,M., Shiels,P.G. and Borst,P. (1991b) Nucleic Acids Res., 19, 1359-1368. Zomerdijk,J.C.B.M., Kieft,R., Shiels,P.G. and Borst,P. (199 l c) Nucleic Acids Res., 19, 5153-5158. Zomerdijk,J.C.B.M., Kieft,R. and Borst,P. (1993) Mol. Biochem. Parasitol., 57, 295-304. Received on February 17, 1994; revised on August 26, 1994
5482
