Experimental Parasitology 106 (2004) 37–44

www.elsevier.com/locate/yexpr

Trypanosoma brucei: a first-generation CRE-loxPsite-specific recombination system

Brian Barrett, Douglas J. LaCount, and John E. Donelson*

Department of Biochemistry, University of Iowa, Iowa City, IA 52242, USA

Received 14 July 2003; accepted 5 January 2004

Abstract

The bacteriophage CRE-loxP system of DNA recombination is widely used to manipulate segments of the genomes of mice and

other eukaryotes for the purpose of studying the regulation and functions of their genes. Since this recombination system could have

similar applications in analyzing the genomes of trypanosomatids, we assessed the action of CRE recombinase on its loxP DNA

recognition sites in Trypanosoma brucei after inserting tetracycline-regulated CRE and two 34-bp loxP sites into the T. brucei ge-

nome. We found that when loxP sites flank in a direct orientation the transcription termination sequence (1.1 kb) of the T. brucei

GPEET/PAG3 locus, CRE recombinase deletes this termination sequence, permitting transcription and subsequent expression of a

downstream reporter gene for the green fluorescent protein (GFP). Thus, the CRE-loxP system is highly efficient in T. brucei, but the

experimental results also indicate that a better way than the existing tetracycline-regulated system is required to completely silence

expression of CRE in the T. brucei genome when it is not needed before the full range of CRE-loxP applications currently used in

mice can be exploited in African trypanosomes.

� 2004 Elsevier Inc. All rights reserved.

Index Descriptors and Abbreviations: Trypanosoma brucei; CRE recombinase; Procyclin; Transcription termination; Gene deletion; GPEET, glycine–

proline–glutamic acid–glutamic acid–threonine pentapeptide repeat protein (one of the two forms of the protein previously called PARP or

procyclin); EP, glutamic acid–proline dipeptide repeat protein (one of the two forms of the protein previously called PARP or procyclin); PAG,

procyclin-associated gene; TERMGPEET , the transcription terminator sequence of the GPEET/PAG3 locus of T. brucei; FACS, fluorescent activated

cell sorter; GFP, green fluorescent protein; Otet, tetracycline operator; PGPEET and PrRNA, promoters for the T. brucei GPEET/PAG3 locus and rRNA

gene locus, respectively; PT7, promoter for T7 RNA polymerase; rRNA, ribosomal RNA; UTR, untranslated region; VSG, variant surface glyco-

protein

1. Introduction

CRE recombinase is a protein encoded by bacterio-

phage P1 that mediates recombination between two

non-tandem 34-bp loxP sites. Depending on the orien-tation and location of the two loxP sites, CRE can

mediate DNA deletions, inversions, and cross-overs.

Through the pioneering work of Brian Sauer and

colleagues, the technology of this prokaryotic recombi-

nation system has been transferred to eukaryotic cells—

first to yeast cells (Sauer, 1987) and subsequently to

mammalian cells (Sauer and Henderson, 1988). The

CRE-loxP system now forms the basis of many elegant

* Corresponding author. Fax: 1-319-335-9570.

E-mail address: john-donelson@uiowa.edu (J.E. Donelson).

0014-4894/$ - see front matter � 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.exppara.2004.01.004

approaches for genomic DNA modifications and gene

rearrangements in mice, including conditional gene

knockouts (reviewed by Le and Sauer, 2001; Mills and

Bradley, 2001; Wilson and Kola, 2001). We decided to

introduce this DNA recombination system into the Af-rican trypanosome, Trypanosoma brucei, in anticipation

that it could be used for purposes as diverse as: (i)

forcing recombination between different VSG expression

sites, (ii) deleting large tandem arrays of homologous

genes such as the heat shock protein (HSP) or MSP

(GP63 zinc metalloprotease) gene families, (iii) deleting

or inverting large polycistronic, non-homologous gene

clusters, (iv) targeting different foreign DNA insertionsto the same genomic location so that positional effects

are eliminated during comparison of different gene ex-

pression cassettes, and (v) removing excess selectable

markers for reuse in subsequent constructs.

38 B. Barrett et al. / Experimental Parasitology 106 (2004) 37–44

To test the CRE-loxP system in T. brucei we utilizedthe gene for the green fluorescent protein (GFP) and a

transcription termination sequence located downstream

of the GPEET/PAG3 gene cluster of T. brucei that was

identified and characterized by Pays and colleagues

(Berberof et al., 1996). This gene locus was previously

called the PARP A locus (PARP, procyclic acidic re-

petitive protein or procyclin), but in accordance with the

need for a standardized nomenclature of trypano-somatid genes (Clayton et al., 1998), it was renamed

GPEET/PAG3 (Roditi and Clayton, 1999). The

GPEET/PAG3 transcription terminator sequence (ab-

breviated here as TERMGPEET ) has been shown to cause

an orientation-dependent termination of transcription

from promoters for the GPEET/PAG3 genes, the rRNA

genes and the VSG gene expression sites, all of which are

promoters for either RNA polymerase I or an enzymebearing RNA polymerase I-like activity with respect

to inhibition by a-amanitin (Berberof et al., 1996). This

TERMGPEET does not terminate transcription by

T. brucei RNA polymerase II (Berberof et al., 1996).

We first cloned theCRE coding sequence in a T. brucei

expression vector so that it was under the control of two

adjacent tetracycline operators (Otet). When this vector

was integrated into the genome of T. brucei cells engi-neered to possess the tetracycline repressor protein (pro-

cyclicT. brucei 29–13 cells;Wirtz et al., 1999), we expected

CRE expression to be repressed until it was induced by

addition of tetracycline to the culture medium. Next we

generated a PCR amplification product that contains a

1.1-kb TERMGPEET sequence flanked by direct repeats of

the 34-bp loxP sites and cloned this PCRproduct between

the promoter for the GPEET/PAG3 locus (PGPEET ) andGFP in another T. brucei plasmid. We anticipated that

when this second plasmidwas integrated into theT. brucei

29–13 genome containing CRE, expression of GFP from

the PGPEET would be prevented by the presence of the

intervening TERMGPEET until this terminator sequence

was excised by the activity of CRE acting on the loxP sites

flanking it. Thus, the functional presence of the CRE-

loxP system in T. brucei could be assayed by a ‘‘gain ofactivity,’’ i.e., the gain ofGFP expression, which is amore

convincing demonstration of a targeted genomic DNA

rearrangement event than the loss of an activity. We

found: (i) byWestern blot that CRE expression in theseT.

brucei cells was induced by tetracycline, and (ii) by se-

quence determination of PCR products from the cells

which acquired green fluorescence that CRE did indeed

trigger deletion of the TERMGPEET between the directloxP repeats. However, even when CRE expression is

suppressed by two upstream tandem Otet elements, it

appears that a few CRE molecules are made, which can-

not be detected on a Western blot but are sufficient to

catalyze a deletion of the 1.1-kb segment between two

loxP sites. Therefore, the CRE-loxP system is clearly very

efficient in T. brucei but a better method of completely

inhibiting expression of CRE until its activity is neededmust be identified before the full power of the CRE-loxP

system can be exploited to manipulate the trypanosome

genome inways similar to its current use formanipulating

other eukaryotic genomes.

2. Materials and methods

2.1. Trypanosome cells

The procyclic Trypanosoma brucei brucei cell line 29–

13, which constitutively expresses T7 RNA polymerase

and the tetracycline repressor (Wirtz et al., 1999), was

provided by G.A.M. Cross. Procyclic cells were cultured

in Cunningham�s SM medium supplemented with 20%

heat-inactivated fetal calf serum, 12 lgG418/ml and 25 lghygromycin/ml, and were stably transfected with plas-

mids via electroporation as described previously (Hill

et al., 1999). When the cells were grown in the presence

of additional antibiotics, the following concentrations

were used: phleomycin (2.5 lg/ml) and noursothricin

(100 lg/ml).

2.2. Construction of plasmids

Plasmid pLEW100 was the gift of G.A.M. Cross. The

luciferase coding region (LUC) was excised from

pLEW100 by cleavage withHindIII and BamHI, and the

recessed 30 ends filled in by a DNA polymerase reaction

and ligated together in a DNA ligase reaction to generate

the circular plasmid pLEW100/DLUC (plasmid A in

Fig. 2). The CRE coding sequence (GenBank AccessionNo. X03453) was PCR-amplified from plasmid pBS185

(purchased from Gibco-BRL/Life Technologies, Carls-

bad,CA) using a forward primerwith a 50HindIII site and

a reverse primer with a 50 BamHI site. The resulting PCR

product of HindIII-CRE-BamHI was cloned into Hin-

dIII-/BamHI-digested pLEW100 so that CRE replaced

LUC to generate plasmid pLEW100/CRETi (plasmid B in

Fig. 2).Plasmid pSK1-1/loxP (TERMGPEET )loxP-GFP (plas-

mid D in Fig. 2) was generated using several DNA cas-

settes and plasmids as follows.ClaI- andHindIII-digested

pXS2-pac (plasmid 1, provided by J. Bangs and described

in Bangs et al., 1996) was used as the basic parent plasmid

in which to clone the PCR product of ClaI–loxP(-

TERMGPEET )loxP–HindIII (cassette 1) to generate plas-

mid 2. The primers used to generate cassette 1 are showninFig. 1.AscI–GFP-30-50ACT–AscI (cassette 2)was PCR-

amplified from pHD496:GFP (plasmid 3, described in

Hill et al., 2000) and ligated into AscI-digested plasmid 2

to generate plasmid 4. The fragment HindIII–SAT–

BamHI (cassette 3), which confers resistance to nour-

seothricin via streptothricin acetyltransferase (SAT), was

PCR-amplified from pIR1–SAT (plasmid 5, Joshi et al.,

Fig. 2. Diagrams of plasmids used in this work. Plasmid A is pLEW100 (Wirtz et al., 1999) in which the luciferase coding region (LUC; not shown)

between the indicatedHindIII and BamHI sites was removed by cleavage with these two enzymes followed by a DNA polymerase reaction to ‘‘fill-in’’

the ends and a blunt-end ligation reaction to regenerate a circular plasmid. Plasmid B is pLEW100 in which the LUC coding region between the

HindIII and BamHI sites was replaced with a PCR-amplification product of the CRE coding region. Plasmids A and B were linearized with NotI and

individually electroporated into procyclic T. brucei 29–13 cells (which contain the genes for the tetracycline repressor and T7 RNA polymerase; Wirtz

et al., 1999), selecting for integration into the rRNA gene locus by expression of phleomycin (bleomycin, BLE) resistance. Phleomycin-resistant

clones of each of these transfections were obtained and used in the subsequent experiments. Plasmids C and D were constructed as described in

Section 2. Plasmid D is identical to plasmid C except that it possesses a 1.1-kb TERMGPEET flanked by the 34-bp loxP sites inserted between the

indicated ClaI and HindIII sites of plasmid C. This [loxP(TERMGPEET )loxP] segment was generated using the PCR primers shown in Fig. 1.

Plasmids C and D were linearized with NotI and individually electroporated into cloned procyclic T. brucei 29–13 cells containing either plasmid A or

plasmid B already integrated into their genome. Linearized plasmids C and D integrate into the intergenic region of the b-tubulin gene locus and cells

containing the integrated plasmids were selected for by resistance to nourseothricin (SAT). The puromycin-resistance gene (PUR) shown in the

diagrams of plasmids C and D is a component of the parent plasmid, but was not used in these experiments. PT7 is the bacteriophage promoter

recognized by T7 RNA polymerase. PGPEET and PrRNA are the promoters for the T. brucei GPEET/PAG3 locus and the rRNA genes, respectively.

Otet is the tetracycline operator. The indicated regions flanking the various genes depicted by the rectangles are the corresponding 30 or 50 flankingregions of the following T. brucei genes: 30 and 50 ACT (actin gene), 50 EP (EP gene), 30 ALD (aldolase gene) and 30 TUB (tubulin gene). Dotted lines

indicate the locations of insertions used to generate plasmids B and D. The loxP sites in plasmid D are indicated by the open horizontal arrows.

Abbreviations for restriction sites are indicated in the upper right. The PCR primers used to generate the data shown in Fig. 4A are indicated by the

circled letters F and R above plasmid C, and their sequences are shown on the bottom right.

Fig. 1. Sequences of the forward (top sequence) and reverse (bottom sequence) PCR primers used to amplify a 1.1-kb GPEET/PAG3 terminator

sequence (TERMGPEET ) from the T. brucei genome. The boxed regions are the 34-bp loxP sequence, which consists of two 13-bp inverted repeats (the

CRE recognition sites, indicated by horizontal arrows) separated by an 8-bp asymmetrical spacer that is responsible for the directional nature of the

loxP site (Le and Sauer, 2001; Sauer, 1987). The 50 ends of the primers contain restriction sites for ClaI or HindIII and the 30 ends are the sequencesflanking a T. brucei TERMGPEET (Berberof et al., 1996).

B. Barrett et al. / Experimental Parasitology 106 (2004) 37–44 39

1995) and ligated into HindIII–BamHI-digested plasmid

3. PCR amplification was then used to amplify XhoI-

rRNA promoter-50EP-SAT-30ACT-XhoI (cassette 4),

which was then ligated into XhoI-digested plasmid 4 (to

generate plasmid 6). The resultantKpnI–NotI fragment of

plasmid 6 (cassette 5) was excised and ligated into KpnI-/

NotI-digested pSK1 (described in LaCount et al., 2002) to

produce pSK1-1/loxP (TERMGPEET )loxP-GFP (plasmid

D in Fig. 2). pSK1-1/GFP (plasmid C in Fig. 2) is plasmid

D without cassette 1.

40 B. Barrett et al. / Experimental Parasitology 106 (2004) 37–44

2.3. Flow cytometric analysis

Analiquot of 106 T. brucei cells was centrifuged at 800g

for 5min and the cell pellet gently suspended in 1ml of

a PBS solution containing 0.25% paraformaldehyde.

Green fluorescence was analyzed on a fluorescence

activated cell sorter (FACScan from Becton–Dickinson,

San Jose, CA) at the Flow Cytometry Facility of the

University of Iowa.

2.4. Other procedures

PCR amplifications were conducted for 30 cycles of

the following temperature changes: 95 �C for 45 s, 55 �Cfor 60 s, and 72 �C for 90 s. After the final cycle the re-

action was incubated at 72 �C for an additional 10min

and placed at 4 �C until applied to a 1% agarose gel.Western blots were conducted as described (Hill et al.,

2000) and were probed with a 1–30,000 dilution of

polyclonal rabbit anti-CRE antiserum (Novagen, Mad-

ison, WI) and a 1–7500 dilution of donkey anti-rabbit

secondary antiserum linked to horse radish peroxidase

(Amersham Biosciences, Piscataway, NJ). The relative

amounts of CRE protein were determined by densi-

tometry of the Western blot signals.

Fig. 3. A plot of the appearance of CRE recombinase in extracts of

procyclic T. brucei 29–13 cells containing integrated plasmid B bearing

CRE (see Fig. 2), as detected by Western blots, after induction with the

following concentrations of tetracycline: 1lg/ml (open circles), 0.1 lg/ml (closed circles) or 0.01lg/ml (closed triangles). Cell extracts were

prepared prior to tetracycline addition (0 time) and 1, 2, 4, and 8 h

after tetracycline addition. The insert shows an example of a Western

blot containing extracts of cells induced with 1 lg tetracycline/ml and

probed with anti-CRE antiserum.

3. Results

3.1. Expression of CRE in T. brucei procyclic cells

CRE recombinase has been widely used in a number of

eukaryotic cells to catalyze the excision or inversion of a‘‘floxed’’ DNA segment, i.e., a DNA segment flanked by

the 34-bp loxP elements. If the two flanking loxP ele-

ments are in a direct orientation, CRE catalyzes the de-

letion of the floxed DNA segment, leaving a single loxP

site. If the flanking loxP elements are in an inverted ori-

entation, CRE inverts the DNA segment, rather than

excising it. Furthermore, a single loxP element in the

genome can serve as a target site for insertion of exoge-nously plasmid DNAs containing a loxP element, greatly

enhancing the chances that different transfected plasmids

always insert at the same genomic location (Le and Sauer,

2001; Mills and Bradley, 2001; Wilson and Kola, 2001).

To examine whether CRE can be expressed in T.

brucei, the CRE coding sequence was PCR-amplified

from a commercially available plasmid (see Section 2)

and used to replace the luciferase coding region in T.

brucei expression plasmid pLEW100 (Wirtz et al., 1999).

This new plasmid, pLEW100/CRETi (plasmid B in

Fig. 2), contains PGPEET followed by two adjacent Otet

elements and the CRE coding region. The plasmid was

linearized with NotI, which cleaves once in the plasmid�srRNA spacer sequence, and electroporated into T.

brucei procyclic 29–13 cells, selecting for phleomycin

resistance. T. brucei 29–13 cells are engineered to con-stitutively express the tetracycline repressor (as well as

T7 RNA polymerase) (Wirtz et al., 1999), so we antici-

pated that in the absence of tetracycline, transcription

from PGPEET would be blocked at the two Otet elements

and CRE would not be expressed until tetracycline was

added to the procyclic culture medium. The transfected

T. brucei cells were cloned by serial dilution, and several

clones were selected for growth in liquid culture andfurther study. PCR amplifications of the genomic DNAs

from these clones revealed that the linearized plasmid

had integrated into the rRNA locus (not shown). Clones

of T. brucei 29–13 cells containing integrated pLEW100/

DLUC (plasmid A in Fig. 2) were also generated for use

in subsequent experiments as control cells lacking CRE

(see below).

A clone of T. brucei 29–13 cells containing integratedplasmid B (pLEW100/CRETi) was grown to mid-loga-

rithmic stage (3� 106 cells/ml) in the absence of tetra-

cycline and the culture divided into aliquots to which

differing amounts of tetracycline were added. Samples

were taken at subsequent time points and cell extracts

prepared to determine the amount of CRE protein via

Western blots probed with anti-CRE antiserum. As

shown in Fig. 3, CRE protein could not be detectedprior to tetracycline addition, but was present 1 h after

addition and continued to increase in amount for at

least 8 h. The extent of CRE induction was identical for

both 1 and 0.1 lg tetracycline/ml, but no detectible in-

duction occurred in 0.01 lg/ml. This result and the ki-

netics of induction are consistent with an earlier analysis

of tetracycline induction of this expression system

using luciferase as a reporter gene, in which 1 and 0.1 lg

B. Barrett et al. / Experimental Parasitology 106 (2004) 37–44 41

tetracycline/ml gave similar levels of induction and0.01 lg/ml gave much less (Wirtz et al., 1999).

During these experiments two additional observa-

tions were made concerning the growth of the cells in

culture. First, in the absence of tetracycline, T. brucei

29–13 cells containing plasmid B consistently grew

about 10% slower than did the same cells containing

either pLEW100 or pLEW100/DLUC (plasmid A in

Fig. 2) integrated at the same rRNA locus. This resultsuggests that, even though it cannot be detected on a

Western blot (Fig. 3), a low level of CRE expression

may occur in the absence in tetracycline that retards cell

growth and is not blocked by the existence of two Otet

elements immediately preceding CRE. Secondly, after

adding either 1 or 0.1 lg tetracycline/ml, the growth rate

of cells containing plasmid B diminished still further, the

cells stopped growing by 24 h and they eventually died.In contrast, tetracycline had no effect on the growth of

T. brucei 29–13 cells containing integrated pLEW100 or

plasmid A. Likewise, when tetracycline was removed

after 1 h from cells bearing plasmid B by extensively

washing the cells with medium without tetracycline and

then re-suspending them in tetracycline-free medium at

their former concentration, the cells returned to their

previous growth rate. Thus, prolonged induction ofCRE appears to be deleterious to the cells.

3.2. Detection of CRE recombinase activity at loxP sites

introduced into the T. brucei genome

Having generated tetracycline-inducible CRE re-

combinase in T. brucei, we now wanted to establish an

assay for CRE-catalyzed DNA recombination in T.

brucei that would result in the acquisition, rather than

loss, of an activity since this would be a positive dem-

onstration of the recombination event, rather than a

negative one. Thus, we constructed a plasmid bearing

GFP preceded by a cassette containing the 1.1-kb

TERMGPEET sequence flanked by directly oriented loxP

elements, i.e., plasmid pSK1-1/loxP (TERMGPEET )loxP-

GFP (plasmid D in Fig. 2). The recombination assaystrategy was based on the fact that initially transcription

from PGPEET in this integrated plasmid should be ter-

minated by TERMGPEET before it reaches GFP. Then,

after CRE-catalyzed excision of TERMGPEET , tran-

scription through GFP should occur, resulting in GFP

expression that can be readily detected visually.

The T. brucei 29–13 cells containing CRE in inte-

grated plasmid B are resistant to neomycin, hygromycin,and phleomycin, so it was necessary to build into plas-

mid D a fourth antibiotic resistance gene that is not

downstream of TERMGPEET . Thus, plasmid D has a

cassette bearing the SAT gene (nourseothricin resis-

tance) under the control of PrRNA, and the PUR gene

(puromycin resistance) downstream of the terminator

sequence in the plasmid is not used. Plasmid C in Fig. 2

is a control plasmid in which the cassette of [loxP(TERMGPEET )loxP] in plasmid D has been removed so

that transcription can proceed directly from PGPEET to

GFP. Plasmids C and D are designed to integrate into

the intergenic region of the TUB locus when linearized

with NotI.

Linearized plasmids C and D were transfected into

T. brucei 29–13 cells containing either previously inte-

grated plasmid A or B, and cloned transfectant cells wereselected by resistance to nourseothricin. As expected, in

the absence of tetracycline, cells containing either inte-

grated plasmid A or B initially fluoresced green when

transfected with plasmid C and did not when transfected

with plasmidD (not shown). This result is consistent with,

and confirms, the original finding of Pays and colleagues

that the TERMGPEET region diminishes transcription

from PGPEET by 7–20-fold (Berberof et al., 1996), anddemonstrates that the flanking loxP elements do not in-

terfere with this termination of transcription.

After several days in liquid culture, however, a few

cells in the cloned populations containing plasmids B

and D began to fluoresce green, even though tetracycline

was not present in the culture medium. In contrast,

green fluorescent cells did not appear in the cloned

populations containing plasmids A and D. The per-centage of cells in the cloned populations containing

plasmids B and D that fluoresced green varied some-

what from one population to another, but in all cases

this percentage of green cells increased with continued

passage of the cells in culture. These results are consis-

tent with the possibility suggested above that a low level

of CRE expression occurs from integrated plasmid B,

even in the absence of tetracycline.To test this possibility, PCR amplifications of the re-

gion surrounding the cassette of [loxP(TERMGPEET )

loxP] were conducted on genomic DNA templates iso-

lated from several cloned populations bearing different

integrated plasmids (Figs. 4A and B). The locations of the

binding sites for primers F and R used in these PCR

amplifications are indicated by the corresponding circled

letters in the diagrams of plasmid C in the Fig. 2 andplasmid D in Fig. 4B. The forward primer (primer F)

contains aClaI recognition sequence at its 30end so that it

will prime at the only segment in the genome containing a

ClaI site at a specific location downstream of a PGPEET .

The reverse primer (primer R) primes within the single

integrated copy of GFP in the genome. Examples of the

resultant PCR amplification products are shown in

Fig. 4A. All of the PCR products shown in Fig. 4A werecloned and sequenced to confirm their identity. Lanes 1, 2,

5, and 6 show the PCRamplification products of template

genomicDNAs from control cells. Lanes 1 and 2 show the

1.05-kb PCR product resulting from amplification of

template DNA from two cloned cell lines containing in-

tegrated plasmids B and C. In these cases, plasmid C does

not contain the [loxP(TERMGPEET )loxP] cassette and the

Fig. 4. (A) Photograph of an ethidium bromide-stained agarose gel containing the PCR products amplified with primers F and R (shown as circled

letters in Figs. 2 and 4B) from template genomic DNA isolated from procyclic cells containing the indicated plasmids (see Fig. 2) integrated into their

genome. White arrows point to the 2.2-kb and 1.08-kb PCR products generated with primers F and R. The genomic DNAs used as templates for the

PCR amplifications were obtained from independently derived T. brucei clones containing integrated plasmids B and C (lanes 1 and 2), plasmids B

and D (lanes 3 and 4), plasmids A and D (lane 5) or plasmids A and C (lane 6). (B) Diagrams depicting the sequences in the 2.2-kb PCR product

derived from the integrated plasmid D, shown in lanes 3 and 4 of Fig. 4A, and in the 1.08-kb PCR product derived from the deleted plasmid D, also

shown in lanes 3 and 4.

42 B. Barrett et al. / Experimental Parasitology 106 (2004) 37–44

distance between the primer binding sites is about 1.05 kb.

In another control, lane 5 shows the 2.2-kb PCR ampli-fication product derived from genomic template DNA

of a cloned cell line containing integrated plasmids A

and D. In this case, plasmid D contains the [loxP(-

TERMGPEET )loxP] cassette but plasmid A does not en-

code CRE. Thus, excision of TERMGPEET should not

occur, and indeed sequence determination of the single

2.2-kb PCR product confirms that it contains the entire

[loxP(TERMGPEET )loxP] cassette. The control in lane 6shows the 1.05-kb PCR product resulting from amplifi-

cation of templateDNA froma cloned cell line containing

plasmids A and C. In this case, neither CRE nor the

[loxP(TERMGPEET )loxP] cassette is present, so a single

1.05-kb PCR product is expected and was found.

Lanes 3 and 4 show the results when the genomic

DNA templates for the PCR reaction are from two,

independently derived, cloned cell lines containingplasmids B and D that had been passaged in liquid

culture for about five weeks. Two PCR products occur,

one of 2.2 kb and the other of 1.08 kb, i.e., slightly larger

than the 1.05-kb fragment in the adjacent lanes 1 and 2.

Each of these PCR products was cloned and sequenced.

The 2.2-kb product was found to contain the same se-

quence as the single 2.2-kb PCR product in lane 5,

demonstrating it is derived from the intact [loxP(-TERMGPEET )loxP] cassette of the integrated plasmid D.

The 1.08-kb PCR product was found to possess the

same sequence as the 1.05-kb product derived from

plasmid C, except that it also contains one copy of loxP

(34-bp) between the ClaI and HindIII sites at which the

[loxP(TERMGPEET )loxP] cassette was originally in-

serted during the construction of plasmid D, as illus-

trated in Fig. 4B. Thus, the 1.08-kb PCR product isderived from genomic DNA in which TERMGPEET has

been excised, leaving a single loxP, which is consistent

with its excision by CRE recombinase. This result

demonstrates that the cell populations whose genomic

DNAs yielded the PCR products in lanes 3 and 4, andwhich were cloned populations initially, have become

two detectable populations with respect to their genome

after several weeks of passage in culture. One popula-

tion�s genomic DNA has retained the 2.2-kb region

bearing the [loxP(TERMGPEET )loxP] cassette, whereas

the other population�s genomic DNA has lost this cas-

sette to yield a 1.08-kb PCR product bearing a single

loxP at the original site of the cassette, despite the factthat CRE expression was not induced by tetracycline.

The cloned cell lines examined by PCR amplifications

(Fig. 4) were also analyzed for GFP expression by the

fluorescent activated cell sorter (FACS), as shown in

Fig. 5. Panels A, B, and C of Fig. 5 show the FACS

analyses of the same cell populations whose PCR anal-

yses are shown in lanes 3, 4, and 5 of Fig. 4A, respec-

tively. For example, lane 3 of Fig. 4A shows that aboutequal amounts of the 2.2- and 1.08-kb PCR products

were generated from the genomic DNA of this mixed

population. Correspondingly, panel A of Fig. 5 shows

that 49% of these same cells fluoresce green (open ar-

rowhead) and 51% do not (closed arrowhead), when the

trace is gated to resolve the two peaks. A comparison of

the PCR analysis in lane 4, Fig. 4A, and the FACS

analysis in panel B, Fig. 5, on the other independentlyderived mixed population of cells also yields consistent

results. The PCR analysis shows that the 1.08-kb PCR

product is more abundant than the 2.2-kb product,

whereas the gated FACS analysis reveals that 78% of the

cells fluoresce green and 22% do not. Similarly, lane 5,

Fig. 4A, contains a single 2.2-kb PCR product and panel

C, Fig. 5, shows that none of these cells fluoresce green.

The cells analyzed in lane 1, Fig. 4A, are the same cellsanalyzed in panel D, Fig. 5. In this case, all of the cells

should fluoresce green because they contain integrated

plasmid C, which does not have TERMGPEET between

Fig. 5. Analysis of GFP fluorescence in cloned procyclic T. brucei cells

containing the indicted plasmids integrated in their genome. One

million cells were collected, suspended in 1ml and their green fluo-

rescence determined by flow cytometry. The graphs are histograms of

relative GFP fluorescence plotted against frequency of events per

channel. The arrow with the closed arrowhead indicates the cell pop-

ulation that does not display green fluorescence and the arrow with the

open arrowhead indicates the cell population that fluoresces green. The

cells contained the indicated plasmids (see Fig. 2) integrated into their

genome. The cells analyzed in Panels A, B, and C are the same cells

analyzed by the PCR amplifications in lanes 3, 4, and 5 of Fig. 4A,

respectively. The cells analyzed in panel d are the same as those ana-

lyzed in lane 1 of Fig. 4A.

B. Barrett et al. / Experimental Parasitology 106 (2004) 37–44 43

PGPEET and GFP. As described above, only one PCR

product of 1.05 kb is obtained from these cells (lane 1,Fig. 4A), consistent with the presence of integrated

plasmid C and the fact that all of these cells should be

green. However, FACS analysis reveals that only about

90% of the cells fluoresce green, a determination born

out by visual inspection of the cells in a fluorescence

light microscope. We do not know the reason for this

slight discrepancy, but the simplest interpretation is that

some of these cells have lost GFP while retaining BLE,which was used to select and maintain the cells trans-

fected with plasmid C.

4. Discussion

Only one transcription terminator sequence in T.

brucei has been reported to date (Berberof et al., 1996).This sequence of about 2 kb occurs immediately after a

5-kb cluster of four tightly packed protein-encoding

genes that are transcribed from PGPEET by an RNA

polymerase I-like activity (Berberof et al., 1996; Roditiand Clayton, 1999). This 2-kb terminator region was

originally found to contain at least three independent

elements capable of terminating transcription from

PGPEET . Each of these elements is, in turn, comprised of

multiple weaker elements. No obvious sequence simi-

larities exist in the three main elements. We chose the 30-distal of these three elements to use in the experiments

here, i.e., a 1.1-kb region that occurs between NruI andXhoI sites in the larger 2-kb sequence. We found that

when this 1.1-kb sequence is placed between PGPEET and

GFP, it does indeed prevent GFP expression, similar to

the original studies that showed when this region was

inserted between PGPEET and the gene for chloram-

phenicol acetyltransferase (CAT), it diminished CAT

transcription by 7–20-fold (Berberof et al., 1996). We

did not quantitate the extent to which it blocked GFP

expression from PGPEET in our system, but we did find

that it: (i) prevents the cells from appearing green visu-

ally in the fluorescent light microscope, and (ii) permits a

clear distinction to be made via FACS analysis between

cells that fluoresce green and those which do not because

of the presence of the terminator (Fig. 5).

In addition to the experiments described here, several

other experiments were conducted whose results areconsistent with those described. First, a plasmid identical

to plasmid D (Fig. 2) was constructed, except

TERMGPEET was not flanked by loxP elements. When

this plasmid was inserted into the genome of cells bearing

plasmid B, no evidence was obtained for the possible ex-

cision of TERMGPEET by CRE, i.e., the permanently

transfected cells did not fluoresce green under any con-

ditions nor did green cells appear with time. This resultindicates that the flanking loxP elements are required for

excision of the terminator sequence by CRE. Secondly, in

another set of experiments plasmid D was inserted into

the T. brucei 29–13 genome lacking plasmid B, and then

the cells were transiently transfectedwith circular plasmid

B without selection for plasmid B�s antibiotic resistance

gene (BLEO). Transiently transfected plasmids can enter

as many as 80% of the cells during electroporation (La-Count et al., 2000), but they do not replicate in the cells as

a circular plasmid and are unlikely to integrate into the

genome, so they are lost as the cells divide. None of the

cells bearing plasmidDdisplayed green fluorescence prior

to the transient transfection with plasmid B. Two days

after the transient transfection, however, an occasion

green cell (�1 in 500) was observed visually in the fluo-

rescent microscope. No green cells were observed whenplasmid A was the transiently transfected plasmid. In the

plasmid B case, the few green cells persisted as the culture

was passaged but their percentage did not increase with

time, suggesting that transient expression of CRE from

transiently transfected plasmid B caused the excision of

the loxP-flanked TERMGPEET in these few cells. It should

be noted that this result occurred without tetracycline

44 B. Barrett et al. / Experimental Parasitology 106 (2004) 37–44

induction, again consistent with the interpretation that:(i) low-level expression of CRE occurs from plasmid B in

the absence of tetracycline, and (ii) this low-level expres-

sion is sufficient to excise the loxP-flanked TERMGPEET .

This result indicates that a short-term, low-level expres-

sion of CRE is sufficient to trigger a recombination event

at loxP sites.

Thus, the experiments summarized here demonstrate

that CRE recognizes the exogenously added 34-bp loxP

sequence in the T. brucei genome and can conduct pre-

cise excision of a DNA region flanked by directly ori-

ented loxP elements. From a practical standpoint,

however, it appears that the presence of too much CRE

is deleterious to the T. brucei cells, killing them when

CRE�s expression is induced by tetracycline from

PGPEET on a pLEW100-based plasmid (plasmid B in

Fig. 2). Since these T. brucei experiments were con-cluded, high-level expression of CRE has also been

shown to be highly toxic to mammalian and Drosophila

cells, due apparently to pseudo-loxP sites in the genome

that can serve substrates for CRE and lead to severe

chromosomal damage (Andreas et al., 2002; Heidmann

and Lehner, 2001; Loonstra et al., 2001). It is likely that

the same toxic recombination events occur in T. brucei.

With these recentmammalian andDrosophila results inmind, it is clear that for the CRE-loxP system to be useful

for analysis of gene function and regulation in T. brucei,

either: (i) a ‘‘low-level’’ expression system for CRE must

be found in which CRE expression can be completely

turned off until transiently required, and/or (ii) amodified

form of CRE is identified that has much less recombina-

tion activity on pseudo-loxP sites. Experimental muta-

tions of the PGPEET in pLEW100 to reduce its efficiency orsimilar mutations of the CRE coding sequence to reduce

its activity may be necessary to achieve this goal. Alter-

natively, another site-specific recombination system

might be used, such as the yeast flipase (FLP) system in

which the FLP recombinase has recently been shown to

have only 10% of the recombination activity at its target

sequence compared toCREat loxP (Andreas et al., 2002).

The observation that during five weeks of culturedgrowth, 22–50% of the cells containing CRE underwent

CRE-mediated excision at the loxP sites (Figs. 4 and 5),

even though CRE expression was repressed by two adja-

cent Otet elements, demonstrates that only small amounts

of a recombinase are needed to catalyze site-specific re-

combination in T. brucei.

Acknowledgments

We thank Dr. Justin Fishbaugh of the Flow Cy-

tometry Facility at the University of Iowa for expert

advice and technical assistance. This work was sup-

ported in part by National Institutes of Health (NIH)

Grants AI10512 and AI40591.

References

Andreas, S., Schwenk, F., K€uter-Luks, B., Faust, N., K€uhn, R., 2002.

Enhanced efficiency through nuclear localization signal fusion on

phage UC31-integrase: activity comparison with Cre and FLPe

recombinase in mammalian cells. Nucleic Acids Res. 30, 2299–

2306.

Bangs, J.D., Brouch,E.M.,Ransom,D.M.,Roggy, J.L., 1996.A soluble

secretory reporter system in Trypanosoma brucei. Studies on endo-

plasmic reticulum targeting. J. Biol. Chem. 271, 18387–18393.

Berberof, M., Pays, A., Lips, S., Tebabi, P., Pays, E., 1996.

Characterization of a transcription terminator of the procyclin

PARP A unit of Trypanosoma brucei. Mol. Cell. Biol. 16, 914–924.

Clayton, C., Adams, M., Almeida, R., Baltz, T., Barrett, M., Bastien,

P., Belli, S., Beverley, S., Biteau, N., Blackwell, J., Blaineau, C.,

Boshart, M., Bringaud, F., Cross, G., Cruz, A., Degrave, W.,

Donelson, J., El-Sayed, N., Fu, G., Ersfeld, K., Gibson, W., Gull,

K., Ivens, A., Kelly, J., Lawson, D., Lebowitz, J., Majiwa, P.,

Matthews, K., Melville, S., Merlin, G., Michels, P., Myer, P.,

Norrish, A., Opperdoes, F., Papadopoulou, B., Parsons, M.,

Seebeck, T., Smith, D., Stuart, K., Turner, M., Ullu, E.,

Vanhamme, L., 1998. Genetic nomenclature for Trypanosoma

and Leishmania. Mol. Biochem. Parasitol. 97, 221–224.

Heidmann, D., Lehner, C.F., 2001. Reduction of Cre recombinase

toxicity in proliferating Drosophila cells by estrogen-dependent

activity regulation. Dev. Genes Evol. 211, 458–465.

Hill, K.L., Hutchings, N.R., Grandgenett, P.M., Donelson, J.E., 2000.

T Lymphocyte Triggering Factor of African trypanosomes is

associated with the flagellar fraction of the cytoskeleton and

represents a new family of proteins that present in several

eukaryotic organisms. J. Biol. Chem. 275, 39369–39378.

Hill, K.L., Hutchings, N.R., Russell, D.G., Donelson, J.E., 1999. A

novel protein targeting domain confers flagellar pocket localization

in African trypanosomes. J. Cell Sci. 112, 3091–3101.

Joshi, P.B., Webb, J.R., Davies, J.E., McMaster, W.R., 1995. The gene

encoding streptothricin acetyltransferase (sat) as a selectable

marker for Leishmania expression vectors. Gene 156, 145–149.

LaCount, D.J., Barrett, B., Donelson, J.E., 2002. Trypanosoma brucei

FLA1 is required for flagellum attachment and cytokinesis. J. Biol.

Chem. 277, 17580–17588.

LaCount, D.J., Bruse, S., Hill, K.L., Donelson, J.E., 2000. Double-

stranded RNA interference in Trypanosoma brucei using head-to-

head promoters. Mol. Biochem. Parasitol. 111, 67–76.

Le, Y., Sauer, B., 2001. Conditional gene knockout using Cre

recombinase. Mol. Biotechnol. 17, 269–275.

Loonstra, A., Vooijs, M., Beverloo, H.B., Allak, B.A., van Drunen, E.,

Kanaar, R., Berns, A., Jonkers, J., 2001. Growth inhibition and

DNA damage induced by Cre recombinase in mammalian cells.

Proc. Natl. Acad. Sci. USA 98, 9209–9214.

Mills, A.A., Bradley, A., 2001. From mouse to man: generating

megabase chromosome rearrangements. Trends Genet. 17, 331–339.

Roditi, I., Clayton, C., 1999. An unambiguous nomenclature for the

major surface glycoproteins of the procyclic form of Trypanosoma

brucei. Mol. Biochem. Parasitol. 103, 99–100.

Sauer, B., 1987. Functional expression of the cre-lox site-specific

recombination system in the yeast Saccharomyces cerevisiae. Mol.

Cell. Biol. 7, 2087–2096.

Sauer, B., Henderson, N., 1988. Site-specific DNA recombination in

mammalian cells by Cre recombinase of bacteriophage P1. Proc.

Natl. Acad. Sci. USA 85, 5166–5170.

Wilson, T.J., Kola, I., 2001. The LoxP/CRE system and genome

modification. Methods Mol. Biol. 158, 83–94.

Wirtz, E., Leal, S., Ochatt, C., Cross, G.A.M., 1999. A tightly

regulated inducible expression system for conditional gene knock-

outs and dominant-negative genetics in Trypanosoma brucei. Mol.

Biochem. Parasitol. 99, 89–101.