Simplified inducible system for Trypanosoma brucei

Abstract

Nowadays, most reverse genetics approaches in Trypanosoma brucei, a protozoan parasite of medical and veterinary importance, rely on pre-established cell lines. Consequently, inducible experimentation is reduced to a few laboratory strains. Here we described a new transgene expression system based exclusively on endogenous transcription activities and a minimum set of regulatory components that can easily been adapted to different strains. The pTbFIX vectors are designed to contain the sequence of interest under the control of an inducible rRNA promoter along with a constitutive dicistronic unit encoding a nucleus targeted tetracycline repressor and puromycin resistance genes in a tandem “head-to-tail” configuration. Upon doxycycline induction, the system supports regulatable GFP expression (170 to 400 fold) in both bloodstream and procyclic T. brucei forms. Furthermore we have adapted the pTbFIX plasmid to perform RNAi experimentation. Lethal phenotypes, including α-tubulin and those corresponding to the enolase and clathrin heavy chain genes, were successfully recapitulated in procyclic and bloodstream parasites thus showing the versatility of this new tool.

About the Authors:

Gabriela T. Niemirowicz

Roles Formal analysis, Funding acquisition, Investigation, Methodology, Resources, Visualization, Writing – original draft, Writing – review & editing

Affiliation: Instituto de Investigaciones Biotecnológicas (IIB) Dr. Rodolfo A. Ugalde, Universidad Nacional de San Martín (UNSAM), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), San Martín, Buenos Aires, Argentina

Juan J. Cazzulo

Roles Funding acquisition, Resources, Writing – review & editing

Affiliation: Instituto de Investigaciones Biotecnológicas (IIB) Dr. Rodolfo A. Ugalde, Universidad Nacional de San Martín (UNSAM), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), San Martín, Buenos Aires, Argentina

Vanina E. Álvarez

Roles Funding acquisition, Resources, Writing – review & editing

Affiliation: Instituto de Investigaciones Biotecnológicas (IIB) Dr. Rodolfo A. Ugalde, Universidad Nacional de San Martín (UNSAM), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), San Martín, Buenos Aires, Argentina

León A. Bouvier

Roles Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Resources, Supervision, Visualization, Writing – original draft, Writing – review & editing

* E-mail: [email protected]

Affiliation: Instituto de Investigaciones Biotecnológicas (IIB) Dr. Rodolfo A. Ugalde, Universidad Nacional de San Martín (UNSAM), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), San Martín, Buenos Aires, Argentina

ORCID logo http://orcid.org/0000-0002-8973-7425

Introduction

Genetic manipulation tools for Trypanosoma brucei, a protozoan parasite of medical and veterinary importance, have appropriately been tailored to suit the peculiar regulation of gene expression found in kinetoplastids. In the genome, unrelated protein coding genes are arranged in long units and regulation takes place mainly at the post-transcriptional level [1, 2]. Transcription by RNA polymerase II (RNAPII) produces polycistronic RNA molecules [3, 4] and transcribed intergenic sequences participate in the trans-splicing between the downstream gene and the cap bearing spliced leader, which is transcribed elsewhere in the genome. In addition this processing is also responsible in a concerted manner for the polyadenylation of the upstream coding sequence, thus resulting in the maturation of each mRNA [5, 6]. Furthermore, the remnants of the transcribed intergenic sequence, namely the 5´ and 3´ untranslated regions determine messenger stability and steady state levels as well as translation efficiency through the binding of other cellular components [7–10]. Hence in genetic constructs transgenes are usually placed between intergenic sequences of housekeeping genes. Given this post-transcriptional regulation of gene expression, in principle any transcription activity with enough processivity can be employed in the expression of transgenes. For constitutive expression close to physiological levels, constructs have been inserted into genomic loci to be transcribed by RNAPII readthrough [11, 12]. Alternatively higher expression levels can be obtained placing, into the genetic constructs, any of the available and well-defined RNA polymerase I (RNAPI) promoters. These include those that transcribe the variant surface glycoprotein (VSG) [13] and procyclin acidic repetitive protein (PARP) [14] genes and that are mutually exclusively active in vertebrate and insect specific stages, respectively. On the other hand, transcription irrespective of stage can be obtained with ribosomal RNA (rRNA) promoters [15]. However, since the discovery that bacteriophage RNA polymerases are capable of high processivity transcription in trypanosomes [16] they have been adopted in most regulatable gene expression systems. The arguments supporting this choice involve the fact that T7 RNA polymerase offers reliable high levels of transcription and that unlike other trypanosome transcription activities it is not likely the subject of endogenous regulation [17].

Different approaches for controlling the abundance of gene products have been described. Recently a system based on a drug inducible riboswitch that fused to the mRNA can affect its stability was reported [18]. However, in contrast to the physiological post-transcriptional regulation mechanism, the most widely employed strategy for exogenous control of gene expression involves regulation at the transcriptional level. Endogenous RNAPI or bacteriophage T7 promoters were modified by insertion of bacterial operators enabling accurate control on transcription activity by the drug dependent binding of the corresponding bacterial repressors. Although regulatable expression vectors based on the vanillic [19] and cumate [20] repressors have recently been reported, the most widely used and best characterised inducible systems rely on the tetracycline repressor (tetR) [21].

To date, state of the art inducible systems for trypanosomes can be thought of as being composed of two separate components. The first consists of suitable host strains that have to be established in advance to express the required exogenous regulatory elements. In earlier systems this was accomplished by sequential transfection of two vectors, each one specialized in the constitutive expression of T7 RNA polymerase and bacterial repressor [17, 22]. More recent designs, on the other hand, have managed to express appropriate levels of both these regulatory elements from the same construct and thus the generation of the desired cell line requires a single transfection event [19, 20, 23]. The second component corresponds to vectors in which the sequence of interest is cloned downstream of an inducible promoter. To prevent RNAPII read-through, which would result in higher basal transgene expression, inducible vectors are targeted by homologous recombination to transcriptionally silent regions of the genome. The non-transcribed spacer of ribosomal RNA loci [21] as well as the 177 bp repeats on minichromosomes [24, 25] are two alternative destinations frequently used for this purpose. Finally, for selection of transfected cells these molecules also harbour a constitutively expressed antibiotic resistance gene.

Due to their great success and broad acceptance, inducible systems for T. brucei have served as models for the development of analogous tools for other kinetoplastid organisms [26–31]. However, there are some drawbacks associated with all of these two component systems. During the establishment of suitable hosts, clonal variability in regulation capabilities has been reported. Usually these differences become manifest only after transfection of an inducible vector, and consequently the following screening steps become laborious [32, 33]. Furthermore, this dependence on pre-establishing cell lines hinders the implementation of inducible system based methodologies in other strains. In this sense, successive transfection and selection steps as well as lengthy periods in culture can affect the natural wild type properties sought after in field isolates.

On the other hand, given the limited number of antibiotic resistance genes available for trypanosome selection, inducible systems that minimize the required number of these elements are highly desirable. To some extent, this issue has been dealt with in works that extend the possibilities for further genetic manipulation methodologies [19, 20, 23]. Moreover, the recent development of an episome-based CRISPR-Cas9 system has allowed genome edition without the insertion of a resistance marker [34], underscoring the interest in methodologies that reduce or circumvent the use of these elements.

In this work we have developed an alternative simplified inducible system based entirely on a single plasmid. It has been designed to contain the sequence of interest along with all other required regulatory elements avoiding the need for pre-established cell lines.

Materials and methods

Plasmid construction

For the construction of pTbFIX and pTbFIX-PARP several fragments were PCR-amplified, cloned into pGEM-T Easy (Promega Corporation, Madison, WI, USA) and sequenced (Macrogen, Seoul, Korea). The tetR gene was derived from pLEW13 [17] and the sequence for the c-terminal NLS was added in frame to the 3´ primer. The pLEW100v5 vector (Dr. George Cross Lab, [17]) obtained from Addgene (Cambridge, MA, USA) was used as template for the amplification of the constitutive and inducible versions of the rRNA promoter, as well as the ribosomal spacer targeting sequence and aldolase 3´ intergenic region. Other fragments (pUC min backbone vector, PARP promoter, mEGFP gene and actin and α/β-tubulin intergenic sequences) were derived from already available plasmids. A detailed description of the performed cloning steps can be obtained upon request. The sequences of both vectors were deposited in GenBank under accession numbers MH936667 and MH936668.

Constructs for RNAi gene silencing: For the construction of pTbFIX-αT, Lister 427 genomic DNA was used as template for the PCR-amplification of 550 bp and 864 bp segments from the α-tubulin genes. In each case the same forward (AAGCTTGCTAGCCACTACACCATTGGTAAGGAG; HindIII site underlined and NheI site in bold) primer and two alternative reverse primers (GGGCCCGAGACAGAGAGTTGCTCGTGG and GGGCCCTGTGGTCAATACGAGCGAAC; PspOMI site underlined) were used respectively. After cloning into pGEM-T Easy and being sequenced, the HindIII/PspOMI shorter and the NheI/PspOMI longer segments were inserted into HindIII/SpeI digested pTbFIX producing pTbFIX-αT by means of a three fragment ligation. A 1553 bp HindIII/PstI segment from the latter was transferred to pTbFIX-PARP thus giving rise to pTbFIX-PARP-αT. An equivalent strategy was employed for the generation of constructs allowing the RNAi silencing of enolase and clathrin heavy chain genes (pTbFIX-ENO and pTbFIX-CLH). With the same forward primer (AAGCTTGGATCCACGATCCAGAAAGTTCACGGTC; HindIII site underlined and BamHI site in bold) and two reverse primers (TCTAGACCTCAGAACCCATACGCAG and TCTAGAAGCGGCTCATTAATGTCTTTG; XbaI site underlined) 553 bp and 662 bp long fragments from the enolase gene were amplified. Likewise with the forward (AAGCTTGGATCCGATTAACCTGAAGCATTCCCAT; HindIII site underlined and BamHI site in bold) and both reverse (TCTAGAACACGTTAATTAGTTTGTCCGCA and TCTAGAGTTCCTGCACCTGCCCAAC; XbaI site underlined) primers 377 bp and 503 bp segments from the CLH gene were cloned. The three fragment ligations into HindIII/BamHI digested pTbFIX involved the shorter HindIII/XbaI fragment with the corresponding BamHI/XbaI longer fragment. All primers used in this work were ordered from Macrogen.

Trypanosome culture

Pleomorphic bloodstream trypanosome EATRO 1125 cell line (BSF) was cultured at 37°C under 5% CO2 in HMI-9 medium (Sigma-Aldrich, St. Louis, MO, USA) [35] supplemented with 20% fetal bovine serum. Lister 427 procyclic cells (PCF) were maintained at 28°C in SDM-79 medium supplemented with 10% fetal bovine serum.

Generation of transgenic trypanosome cell lines

For each transfection, approximately 2×107 BSF trypanosomes grown to mid log phase (~ 0.8–1.0x106 cells/mL) in HMI-9 medium were collected, washed once with cytomix [36] (2 mM EGTA, 120 mM KCl, 0.15 mM CaCl2, 10 mM K2HPO4/KH2PO4 pH 7.6, 25 mM HEPES pH 7.6, 5 mM MgCl2.6H2O) modified with 0.5% glucose, 100 μg/mL BSA and 1 mM hypoxanthine at room temperature, and resuspended in 500 μl of the same buffer. About 5–15 μg of linearized plasmid was mixed with the cell suspension and transferred to a BTX 0.2 cm electroporation cuvette. Electroporation was immediately performed at a single pulse with a BTX 600 Electro Cell Manipulator (BTX Inc., San Diego, CA, USA) at 1.1 kV, 25 μF and R6 resistance (186 Ω). The entire cell-DNA mixture was transferred to a flask containing 5 mL of fresh HMI-9 medium and incubated for 6 hours at 37°C without selection. Cells were diluted to a density of 2x105 cell/mL and 1 mL cell aliquots were distributed in 24-well plates. Selection was achieved by the addition of 0.1 μg/mL of puromycin (InvivoGen).

PCF cells (approximately 2×107 trypanosomes) were washed and resuspended in 500 μl of modified cytomix medium at 4°C. Parasites were electroporated using a single pulse at a voltage of 1.4 kV, 25 μF capacitance and 24 Ω resistance. Selection was achieved by the addition of 1 μg/mL of puromycin.

Electrophoresis and immunoblotting

Approximately 8x106 BSF or 4x106 PCF cells were resuspended in 1X Laemmli sample buffer, heated in boiling water for 5 min and separated by SDS-PAGE (10 or 12.5% acrylamide). Gels were transferred to a nitrocellulose Hybond ECL membrane (GE Healthcare, Pittsburgh, PA, USA) for probing with anti GFP mouse clones 7.1 and 13.1 antibody diluted 1:1000 (Sigma-Aldrich). Polyclonal rabbit anti enolase was used 1:1000 [37]. Polyclonal anti tetR antibodies were raised in mice (1:500). Monoclonal anti α-tubulin 1:5000 clone B-5-1-2 (Sigma-Aldrich) was used as loading control.

Blots were probed with Alexa Fluor 790 AffiniPure Goat Anti-Mouse IgG (H+L), Alexa Fluor 680 AffiniPure Goat Anti-Mouse IgG (H+L) or Alexa Fluor 680 AffiniPure Goat Anti-Rabbit IgG (H+L) secondary antibodies (Jackson Immunoresearch Laboratories, West Grove, PA, USA). Signal intensities were detected using an Odyssey laser-scanning system and quantified with Image Studio software (LI-COR Biosciences, Lincoln, NE, USA). Prestained Protein Molecular Weight markers used were from Pierce (Rockford, IL, USA).

Fluorescence microscopy

Cells were fixed with 1% paraformaldehyde, washed with PBS and allowed to bind to poly-L-lysine coated coverslips. GFP fluorescence was monitored with an Eclipse 80i microscope (Nikon, Shinagawa, Japan). 4′,6-diamidino-2-phenylindole (DAPI) was used to visualize the nucleus and kinetoplast.

Flow cytometry

Mid-log phase density PCF (6x106 cells/mL) or BSF (1x106 cells/mL) cells were analyzed with and without doxycyline (DOX) in a CyFlow space cytometer (Partec, Germany). Subsequent data analysis was performed using the FlowJo VX0.7 software (FlowJo LLC, Ashland, OR USA).

Results

Plasmid description

The system conceived for fast inducible expression in T. brucei is based on vector pTbFIX (Fig 1). The plasmid contains two transcription units. The first drives the expression of a green fluorescent protein (GFP) reporter gene through an inducible rRNA promoter. Downstream, a constitutive dicistronic unit encoding a nucleus targeted tetR and puromycin resistance genes was inserted. To study the effect of different tetR abundances on GFP expression regulation, the original constitutive rRNA promoter in pTbFIX was replaced by that of PARP, thus giving rise to pTbFIX-PARP. After linearization at a unique NotI restriction site inside a targeting sequence, the construct is directed to the non-transcribed spacer of ribosomal RNA gene repeats for integration by single crossover “ends-in” homologous recombination.

[Figure omitted. See PDF.]

Fig 1. Schematic diagram of pTbFIX vectors.

The plasmids contain two transcription units. GFP expression is driven by an inducible rRNA promoter modified to contain 2× tetracycline operators (tetO). The downstream constitutive dicistronic unit encodes the tetracycline repressor (tetR) fused to a nuclear localization signal (NLS) and the puromycin resistance gene (PAC). Constitutive transcription of this unit depends on a second rRNA promoter in pTbFIX or a procyclic acidic repetitive protein promoter in pTbFIX-PARP. The intergenic regions for mRNA maturation are represented by small rectangles: SAS, procyclin splice acceptor site; Ald, aldolase 3´ intergenic region; A SAS, actin splice acceptor site; α/β, tubulin intergenic region; Act, actin intergenic sequence. The pTbFIX vectors are linearized at a unique NotI restriction site that targets the constructs to the non-transcribed spacer of ribosomal RNA gene repeats. Propagation in E. coli cells is accomplished by a 1.7 kbp long pUC19 derived minimum backbone plasmid (pUC min).

https://doi.org/10.1371/journal.pone.0205527.g001

Inducible GFP expression in PCF

Both pTbFIX and pTbFIX-PARP vectors where transfected into wild-type Lister 427 procyclic form trypomastigotes and transgenic cell lines were easily obtained. After selection, the resulting parasite populations were cultured with 1 μg/mL of doxycycline which corresponds to the highest concentration employed for other tetR based inducible systems in order to obtain maximal induction [17, 21]. Doubling times of induced and uninduced cells were similar to those of the wild-type parental strain indicating that the constructs had no effect on cellular viability (S1 Fig). Microscopic examination showed a homogeneous population of fluorescent cells in samples derived from induced cultures (Fig 2A). In contrast, no GFP signal could be detected in uninduced equivalents. Protein samples were analyzed by Western blot with antibodies raised against tetR. In uninduced conditions the constitutive rRNA promoter produced higher (4.2-fold) repressor amounts than those obtained with the PARP equivalent. Interestingly, upon doxycycline addition both constructs expressed similar quantities of tetR (1.3-fold higher for pTbFIX). This suggests that the unrepressed inducible rRNA promoter is capable of driving transcription of all three downstream genes (Fig 2B).

[Figure omitted. See PDF.]

Fig 2. GFP expression in procyclic cells harbouring the pTbFIX vectors.

A) Images of Lister 427 pTbFIX and pTbFIX-PARP cells induced for 24 h with 1 μg/mL of DOX. Scale bar 5 μm. B) Western blot probed with anti-tetR antibodies showing the levels of tetR expression in pTbFIX transfected cells (4x106 parasites per lane). For comparative purposes the parental strain (Lister 427) was included in the analysis. C) Doxycycline dose-dependent GFP expression in procyclic cells. GFP expression was measured by Western blot analysis 24 h post induction. D) Fixed induced (DOX) or non-induced parasites (NI) were analysed by flow cytometry. Approximately 50000 events were captured. Mean fluorescence intensity values can be found in S1 Table. E) PCF cultures were diluted to 1x106 cells/mL and induced with 1 μg/mL of DOX. After 48 h parasites were extensively washed with DOX free medium and incubated for additionally 96 h, followed by DOX re-addition. GFP expression levels were monitored daily by Western blot analysis in total cell extracts (4x106 cells per lane). Tubulin was used as a loading control. Cultures were diluted every two days to 1x106/mL to maintain cell density within a range that supports exponential growth.

https://doi.org/10.1371/journal.pone.0205527.g002

In order to determine their regulation properties, transgenic cell lines were grown with different doxycycline concentrations. Western blot analysis showed that both vectors produced similar induction patterns. Abundance of the GFP reporter increased in a dose dependent manner reaching maximum levels with inducer concentrations greater than 50 ng/mL (Fig 2C) similar to that reported for other inducible systems [17, 21, 38]. In these conditions, band quantification indicated ~250 and ~350 fold increase of gene expression in cells transfected respectively with pTbFIX and pTbFIX-PARP.

To confirm these findings and additionally to assess population homogeneity, samples were analyzed by flow cytometry (Fig 2D). In both cases, fluorescence histograms of uninduced transgenic cells overlapped that of the parental wild-type strain indicating tight repression of GFP transcription while addition of doxycycline to 1 μg/mL conduced to a uniform fluorescence increase. These properties were retained even after three months of continuous uninduced culture. In fact, with doxycycline concentrations above 50 ng/mL at least 95% of the cells exhibited fluorescence signal indicating a homogeneous population response (S2 Fig). Within intermediate concentrations the main differences between both cell lines became evident. In the presence of inducer concentrations ranging from 1 to 10 ng/mL, fewer parasites transfected with pTbFIX displayed comparatively lower GFP fluorescence than those transfected with pTbFIX-PARP. This was in agreement with the higher tetR abundance found in the former.

Finally, as can be seen in Fig 2E doxycycline removal can be used to repress transgene expression in these cell lines. In fact GFP abundances decline and reach undetectable levels after four days of culture in inducer free medium. Nevertheless, the fully induced state can be recovered by doxycycline readdition.

Inducible GFP expression in BSF

Transfections with both plasmids were performed on wild-type EATRO 1125 strain BSF trypomastigotes. All attempts involving pTbFIX-PARP were unsuccessful, probably due to the downregulation of the procyclin promoter and thus lower expression levels of the puromycin resistance gene. In contrast, viable clones were easily obtained with pTbFIX. Three independent clones were chosen in order to evaluate possible differences in expression patterns and regulation properties. Basal GFP expression levels in all three cell lines were close to undetectable (Fig 3A). However, reporter induction already became evident after six hours of growth with 1 μg/mL of doxycycline. Maximum GFP abundances were reached 24 hours post induction and fluorescence was easily detected by microscopic examination (Fig 3B). No significant deviations of the doubling times were found between induced (~6.4±0.4 h), uninduced (~6.8±0.3 h) and the wild-type parental strain (~6.5±0.5 h) (S3 Fig). However, at this point differences between the clones were already noticed with fold induction levels ranging ~170 (clone A3) and ~400 (clone D2). On the other hand, tetR abundances in the three parasite lines were equivalent and depending on doxycycline presence 1.5-fold or 2.1-fold higher than those in the well-established 90:13 strain (Fig 3C). These results suggested that the upstream inducible rRNA promoter also could contribute to repressor expression in BSF cells.

[Figure omitted. See PDF.]

Fig 3. GFP expression in EATRO 1125 cells.

A) GFP expression was analysed by Western blot (4x106 cells per lane) in three different clones. Parasites were induced with 1 μg/mL of DOX. Anti-tubulin antibody was used as loading control. B) Images from cultures induced for 24 h with 1 μg/mL of DOX (clone D2). Scale bar 5 μm. C) Western blot probed with anti-tetR antibodies showing the levels of tetR expression in wild type EATRO 1125 cells and three independent pTbFIX clones (8x106 cells per lane). For comparative purposes Lister 427 90:13 parasites were included in the analysis. D) Doxycycline dose-dependent GFP expression in BSF cells. Parasites were diluted to 3x104 cells/mL and induced for 48 h with indicated concentrations of DOX (clone B6). E) Fixed induced (DOX) or non-induced cells (NI) were analysed by flow cytometry. Approximately 50000 events were captured. Mean fluorescence intensity values can be found in S2 Table. E) BSF cultures were diluted to 2x104 cells/mL and induced with 1 μg/mL of DOX. After 48 h parasites were diluted in DOX free medium and incubated for additionally 5 days, followed by DOX re-addition (day 8). GFP expression levels were monitored by Western blot analysis in total cell extracts.

https://doi.org/10.1371/journal.pone.0205527.g003

In this cellular context the system was very sensitive to inducer concentration reaching maximum expression with 10 ng/mL of doxycycline (Fig 3D). On one hand, flow cytometry of these cell lines confirmed tight repression since fluorescence histograms of transgenic uninduced clones overlapped that of the parental wild-type strain (Fig 3E). However, it also revealed additional differences between the clones. Fluorescence of clone A3 was less than that of clone D2. For its part, clone B6 contained the parasites showing the highest fluorescence. Similar to what was found for PCF, induction in the three clones depended on the presence of doxycycline in the medium. Two successive passages in inducer free conditions were enough to fully repress GFP expression, while doxycycline readdition was capable of restoring former expression levels (Fig 3F).

Finally, to asses construct stability these cell lines were kept in culture in an uninduced state for three months with three weekly passages. Induction with different doxycycline concentrations was later analysed by flow cytometry (S4 Fig). Tight repression was retained in all three clones and maximal induction continued to be achieved with doxycycline concentrations above 10 ng/mL. No significant differences were found in population composition of clones A3 and D2, however clone B6 now contained equal amounts of responsive and unresponsive cells.

System performance in RNAi gene silencing

Inducible gene expression systems are frequently adapted for RNA interference mediated endogenous gene downregulation. To test if the vectors developed in this work could be used for this methodology both pTbFIX and pTbFIX-PARP were modified. The GFP coding sequence was replaced with two α-tubulin gene derived segments inserted in opposite orientations. Thus, transcription driven by the inducible rRNA promoter would produce a double stranded RNA molecule in a “stem-loop” configuration. The resulting plasmids (pTbFIX-αT and pTbFIX-PARP-αT) were transfected into Lister 427 PCF cells. No particular difficulties were experienced during the selection period and doubling times of the resulting transgenic lines (12.9±2.5 h and 11.2±0.7 h for pTbFIX-αT and pTbFIX-PARP-αT respectively) were comparable to that of the wild-type parental strain (12.9±0.8 h). However, upon doxycycline addition growth rates declined rapidly indicating impaired cellular viability. The effect was significant with both plasmids but more apparent in cells transfected with pTbFIX-αT 48 hours post induction (Fig 4A). Microscopic examination revealed the reported “FAT” phenotype [39] associated with α-tubulin downregulation in at least 90% of the cells in both cultures (Fig 4B).

[Figure omitted. See PDF.]

Fig 4. The pTbFIX vectors support RNAi knockdown in PCF cells.

A) Growth analysis was carried out in triplicate in cells transfected with the pTbFIX-αT and pTbFIX-PARP-αT plasmids. Values are means ± SD. B) The characteristic FAT phenotype of PCF cell lines expressing α-tubulin dsRNA is visualized by DIC optics after DOX addition (1 μg/mL). Parasite DNA was stained with DAPI. Scale bar 5 μm.

https://doi.org/10.1371/journal.pone.0205527.g004

Regarding BSF, all attempts to transfect pTbFIX-αT into EATRO 1125 parasites were unsuccessful. This limitation might be explained by the already reported sensitivity of BSF even to the slightest basal downregulation of α-tubulin genes [40, 41]. Alternatively, to test if the system could still be used for RNAi silencing of essential genes, two additional pTbFIX derivatives were obtained. With the same RNA “stem-loop” configuration pTbFIX-ENO and pTbFIX-CLH were constructed to target enolase and clathrin heavy chain genes respectively. Given the clonal variability in GFP expression patterns observed for pTbFIX, clones obtained after these transfections were initially screened for loss of viability associated with RNAi induction. In the case of pTbFIX-ENO, no proliferation could be detected in three out of four clones tested with the addition of 1 μg/mL of doxycycline. Two of these cell lines were chosen for further analysis. As can be seen in Fig 5A no growth rate defects could be detected in uninduced conditions. However, 24 hours after RNAi induction proliferation starts to decline and by 48 hours almost no further growth could be detected. Western blot analysis showed differences between the clones regarding enolase abundance (Fig 5B). In uninduced conditions, respective enzyme amounts in clones A1 and A3 were 78% and 93% that of the wild-type parental strain, suggesting tighter RNAi repression in the latter (Fig 5C). After 48 hours of induction values drop to 37% and 19% respectively, explaining the milder growth defects of clone A1.

[Figure omitted. See PDF.]

Fig 5. RNAi knockdown in EATRO 1125 BSF cells.

A) Growth analysis (triplicate) corresponding to the inducible expression of enolase RNAi in two independent clones. B) Western blot analysis upon enolase RNAi induction. Approximately 5x105 cells were probed with anti-enolase polyclonal antibodies. Tubulin was used as a loading control. C) Enolase quantification signal after DOX addition (1 μg/mL). Values are means ± SD. *P <0.05 (ANOVA). D) Inducible expression of clathrin heavy chain (TbCLH) RNAi in EATRO 1125 BSF cells. Parasite growth was monitored by cell counting in a Neubauer chamber. E) Gallery of images from cultures expressing clathrin RNAi (DOX 1 μg/mL, 24 h). The BigEye morphology is visualized by DIC optics (black arrow). Parasite DNA was stained with DAPI. Scale bar 5 μm.

https://doi.org/10.1371/journal.pone.0205527.g005

On the other hand, the initial analysis of two clones obtained for pTbFIX-CLH showed that only one experienced loss of viability when cultured with doxycycline (Fig 5D). Again no effect in proliferation were observed in uninduced conditions, however RNAi induction had even more severe growth defects than those obtained with pTbFIX-ENO. Furthermore, the reported “BigEye” phenotype associated with clathrin heavy chain gene downregulation [42] could be found by microscopic examination of the induced cells (Fig 5E).

Discussion

This work was focused on the development of a simple genetic manipulation tool that would allow easy implementation of regulated transgene expression in Trypanosoma brucei. The main objective consisted in avoiding the dependence on pre-established cell lines, thus easily extending the range of strains on which to apply these methodologies. This could be accomplished with a single DNA molecule containing the sequences of interest as well as those that code for the required exogenous regulatory elements. To keep the design simple, the system was based exclusively on endogenous transcription activities and a minimum set of regulatory components. This was supported by previous works showing that T7 RNA polymerase is not a prerequisite for efficient inducible transgene expression in T. brucei [24, 43]. Besides, apart from the PAC gene required for selection, the only other exogenous coded element was tetR. The arrangement of inducible and constitutive transcription units in pTbFIX was similar to that of other dual promoter systems [33]. Based on reported poor regulation properties, this configuration was later abandoned in favour of back-to-back alternatives [17]. However, unlike our design, in these pioneering works tetR was expressed from other genomic loci and did not contain a nuclear localization signal which is a feature included in more recent developments [19, 20, 23]. Indeed, the tandem “head-to-tail” configuration was chosen considering that eventual unregulated transcription by the leftmost rRNA promoter in pTbFIX would contribute to tetR expression in a way reminiscent of a negative feedback mechanism. The increase in repressor abundance upon doxycycline addition, most evident in PCF transfected with pTbFIX-PARP, supports this possibility. Finally, during the design phase particular care was taken in order to minimize alterations to the physiological transcription patterns of the cells. In this sense, promoters in pTbFIX vectors and those of the targeted rRNA loci share the same orientation after construct integration.

Transgenic PCF cells experienced equally tight regulation of GFP expression with pTbFIX and pTbFIX-PARP, indicating that tetR levels produced by the former were significantly above those required to repress rRNA transcription. Consequently, further system improvements should be focused on other aspects than increasing repressor amounts. On the other hand, in the fully induced state both constructs reached similar GFP expression levels and cell populations were notably homogeneous. For practical purposes both vectors are equally well suited for conducting regulatable transgene expression based methodologies in PCF cells. However, pTbFIX corresponded to the most versatile tool since it can also be used in BSF. Regulation properties in this cellular background were similar to those observed in PCF. The most apparent difference related to clonal variability, which has already been reported for other constructs [32, 33]. The fact that transgene induction and repression are reversible in all these PCF and BSF lines implies that the system developed can be employed in conditional knockout methodologies.

In principle, pTbFIX has good construct stability properties similar to those reported for other rRNA loci targeted molecules [44]. After three months of continuous culture both PCF cell lines exhibited almost no change in regulation properties and population composition. For their part only one of the three BSF clones experienced loss of GFP expression in a fraction of the cells and, most importantly, all clones retained high repression in uninduced conditions.

It can be speculated that soon after pTbFIX integrates into the genome some unregulated transgene expression might take place until repressor abundances increase. Yet, even with this delay in repression, it was still possible to obtain viable cell lines with constructs for the RNAi downregulation of essential genes in both PCF and BSF parasites. Interestingly, while in uninduced PCF populations obtained with both pTbFIX-αT vectors no cells were found with altered morphology, they showed a noticeably homogeneous response towards induction of RNAi. The former could be due to rapid counter selection of cells with loose repression.

The BSF clonal variability observed not only for GFP expression constructs but also for those intended for RNAi downregulation of enolase as well, can be explained by the already reported position effects associated to alternative site of insertion among different clones [32, 33, 45]. In this respect, besides their use as general purpose transgene expression tools, the pTbFIX vectors might be employed to perform comparative integration site studies in different strains.

Finally, all RNAi constructs in this work were obtained with conventional restriction enzyme based techniques, since these correspond to the most widely available alternative. Nevertheless, pTbFIX vectors can easily be made compatible with other cloning technologies and thus more amenable for high-throughput methodologies [46–48].

Supporting information

[Figure omitted. See PDF.]

S1 Fig. Growth analysis of Lister 427 PCF cells transformed with pTbFIX vectors.

PCF cultures were diluted to 1,25x106 cells/mL and induced with 1 μg/mL of DOX. Parasites were counted in a Neubauer chamber every 24 h and diluted every two days to maintain cell density within a range that supports exponential growth. A) Parental strain (Lister 427), B) pTbFIX and C) pTbFIX-PARP cell lines with or without DOX. Doubling times (mean ± SD) are indicated.

https://doi.org/10.1371/journal.pone.0205527.s001

(TIFF)

S2 Fig. Doxycycline dose-dependent GFP expression in procyclic cells harbouring the pTbFIX vectors.

PCF cultures were diluted to 2x106 cells/mL and induced for 24 h with the indicated concentrations of DOX. Flow cytometry analysis was performed in a CyFlow space cytometer. Approximately 50000 events captured per induction. Mean fluorescence intensity values can be found in S1 Table.

https://doi.org/10.1371/journal.pone.0205527.s002

(TIFF)

S3 Fig. Growth analysis of EATRO 1125 BSF cells transformed with pTbFIX vectors.

Parasite cultures, initially diluted to 2x104 cells/mL, were counted in a Neubauer chamber every 24 h. A) For comparative purposes the parental strain (EATRO 1125) was included in the analysis. B, C, D) Growth curves of three independent clones induced or non-induced with DOX (1 μg/mL). Doubling times (mean ± SD) are indicated.

https://doi.org/10.1371/journal.pone.0205527.s003

(TIFF)

S4 Fig. Doxycycline dose-dependent GFP expression in BSF cells.

BSF cultures were diluted to 2x104 cells/mL and induced for 48 h with the indicated concentrations of DOX. Approximately 50000 events captured per induction. Mean fluorescence intensity values can be found in S2 Table.

https://doi.org/10.1371/journal.pone.0205527.s004

(TIFF)

S1 Table. Arithmetic and geometric mean fluorescence intensity values for procyclic cells harbouring the pTbFIX vectors.

https://doi.org/10.1371/journal.pone.0205527.s005

(DOCX)

S2 Table. Arithmetic and geometric mean fluorescence intensity values for BSF clones.

https://doi.org/10.1371/journal.pone.0205527.s006

(DOCX)

Acknowledgments

GTN, JJC, VEA and LAB are members of the research career of the Argentinean National Research Council (CONICET). We are grateful to Dr. Carla Cristi D.C. Avila, Universidade Federal de São Paulo, São Paulo, SP, Brazil, for kindly providing the EATRO 1125 BSF strain. We thank Dr. Javier De Gaudenzi for the critical reading of the manuscript.

Citation: Niemirowicz GT, Cazzulo JJ, Álvarez VE, Bouvier LA (2018) Simplified inducible system for Trypanosoma brucei. PLoS ONE 13(10): e0205527. https://doi.org/10.1371/journal.pone.0205527

References

1. Clayton CE. Networks of gene expression regulation in Trypanosoma brucei. Mol Biochem Parasitol. 2014;195(2):96–106. Epub 2014/07/06. S0166-6851(14)00078-4 [pii]. pmid:24995711.

2. Clayton CE. Gene expression in Kinetoplastids. Curr Opin Microbiol. 2016;32:46–51. Epub 2016/05/14. S1369-5274(16)30053-4 [pii] pmid:27177350.

3. Imboden MA, Laird PW, Affolter M, Seebeck T. Transcription of the intergenic regions of the tubulin gene cluster of Trypanosoma brucei: evidence for a polycistronic transcription unit in a eukaryote. Nucleic Acids Res. 1987;15(18):7357–68. Epub 1987/09/25. pmid:3658696; PubMed Central PMCID: PMC306253.

4. Muhich ML, Boothroyd JC. Polycistronic transcripts in trypanosomes and their accumulation during heat shock: evidence for a precursor role in mRNA synthesis. Mol Cell Biol. 1988;8(9):3837–46. Epub 1988/09/01. pmid:3221866; PubMed Central PMCID: PMC365442.

5. LeBowitz JH, Smith HQ, Rusche L, Beverley SM. Coupling of poly(A) site selection and trans-splicing in Leishmania. Genes Dev. 1993;7(6):996–1007. Epub 1993/06/01. pmid:8504937.

6. Matthews KR, Tschudi C, Ullu E. A common pyrimidine-rich motif governs trans-splicing and polyadenylation of tubulin polycistronic pre-mRNA in trypanosomes. Genes Dev. 1994;8(4):491–501. Epub 1994/02/15. pmid:7907303.

7. Berberof M, Vanhamme L, Tebabi P, Pays A, Jefferies D, Welburn S, et al. The 3'-terminal region of the mRNAs for VSG and procyclin can confer stage specificity to gene expression in Trypanosoma brucei. EMBO J. 1995;14(12):2925–34. Epub 1995/06/15. pmid:7796818; PubMed Central PMCID: PMC398412.

8. Lee MG. The 3' untranslated region of the hsp 70 genes maintains the level of steady state mRNA in Trypanosoma brucei upon heat shock. Nucleic Acids Res. 1998;26(17):4025–33. Epub 1998/08/15. gkb644 [pii]. pmid:9705515; PubMed Central PMCID: PMC147808.

9. Hotz HR, Hartmann C, Huober K, Hug M, Clayton C. Mechanisms of developmental regulation in Trypanosoma brucei: a polypyrimidine tract in the 3'-untranslated region of a surface protein mRNA affects RNA abundance and translation. Nucleic Acids Res. 1997;25(15):3017–26. Epub 1997/08/01. doi: gka497 [pii]. pmid:9224601; PubMed Central PMCID: PMC146859.

10. Kolev NG, Ullu E, Tschudi C. The emerging role of RNA-binding proteins in the life cycle of Trypanosoma brucei. Cell Microbiol. 2014;16(4):482–9. Epub 2014/01/21. pmid:24438230; PubMed Central PMCID: PMC3974610.

11. ten Asbroek AL, Ouellette M, Borst P. Targeted insertion of the neomycin phosphotransferase gene into the tubulin gene cluster of Trypanosoma brucei. Nature. 1990;348(6297):174–5. Epub 1990/11/08. pmid:2172836.

12. Eid J, Sollner-Webb B. Stable integrative transformation of Trypanosoma brucei that occurs exclusively by homologous recombination. Proc Natl Acad Sci U S A. 1991;88(6):2118–21. Epub 1991/03/15. pmid:2006150; PubMed Central PMCID: PMC51180.

13. Zomerdijk JC, Ouellette M, ten Asbroek AL, Kieft R, Bommer AM, Clayton CE, et al. The promoter for a variant surface glycoprotein gene expression site in Trypanosoma brucei. EMBO J. 1990;9(9):2791–801. Epub 1990/09/01. pmid:1697265; PubMed Central PMCID: PMC551989.

14. Lee MG, Van der Ploeg LH. Homologous recombination and stable transfection in the parasitic protozoan Trypanosoma brucei. Science. 1990;250(4987):1583–7. Epub 1990/12/14. pmid:2177225.

15. Horn D, Cross GA. Position-dependent and promoter-specific regulation of gene expression in Trypanosoma brucei. EMBO J. 1997;16(24):7422–31. Epub 1998/02/21. pmid:9405371; PubMed Central PMCID: PMC1170342.

16. Wirtz E, Hartmann C, Clayton C. Gene expression mediated by bacteriophage T3 and T7 RNA polymerases in transgenic trypanosomes. Nucleic Acids Res. 1994;22(19):3887–94. Epub 1994/09/25. pmid:7937108; PubMed Central PMCID: PMC308385.

17. Wirtz E, Leal S, Ochatt C, Cross GA. A tightly regulated inducible expression system for conditional gene knock-outs and dominant-negative genetics in Trypanosoma brucei. Mol Biochem Parasitol. 1999;99(1):89–101. Epub 1999/04/24. S016668519900002X [pii]. pmid:10215027.

18. Cruz-Bustos T, Ramakrishnan S, Cordeiro CD, Ahmed MA, Docampo R. A Riboswitch-based Inducible Gene Expression System for Trypanosoma brucei. J Eukaryot Microbiol. 2018;65(3):412–21. Epub 2017/12/22. pmid:29265590; PubMed Central PMCID: PMC5935546.

19. Sunter JD. A vanillic acid inducible expression system for Trypanosoma brucei. Mol Biochem Parasitol. 2016;207(1):45–8. Epub 2016/04/12. S0166-6851(16)30029-9 [pii]. pmid:27063979; PubMed Central PMCID: PMC4920640.

20. Li FJ, Xu ZS, Aye HM, Brasseur A, Lun ZR, Tan KSW, et al. An efficient cumate-inducible system for procyclic and bloodstream form Trypanosoma brucei. Mol Biochem Parasitol. 2017;214:101–4. Epub 2017/04/26. S0166-6851(17)30054-3 [pii] pmid:28438458.

21. Wirtz E, Clayton C. Inducible gene expression in trypanosomes mediated by a prokaryotic repressor. Science. 1995;268(5214):1179–83. Epub 1995/05/26. pmid:7761835.

22. Bringaud F, Robinson DR, Barradeau S, Biteau N, Baltz D, Baltz T. Characterization and disruption of a new Trypanosoma brucei repetitive flagellum protein, using double-stranded RNA inhibition. Mol Biochem Parasitol. 2000;111(2):283–97. Epub 2001/02/13. S0166685100003194 [pii]. pmid:11163437.

23. Poon SK, Peacock L, Gibson W, Gull K, Kelly S. A modular and optimized single marker system for generating Trypanosoma brucei cell lines expressing T7 RNA polymerase and the tetracycline repressor. Open Biol. 2012;2(2):110037. Epub 2012/05/31. [pii]. pmid:22645659; PubMed Central PMCID: PMC3352093.

24. Kelly S, Reed J, Kramer S, Ellis L, Webb H, Sunter J, et al. Functional genomics in Trypanosoma brucei: a collection of vectors for the expression of tagged proteins from endogenous and ectopic gene loci. Mol Biochem Parasitol. 2007;154(1):103–9. Epub 2007/05/22. S0166-6851(07)00099-0 [pii] pmid:17512617; PubMed Central PMCID: PMC2705915.

25. Wickstead B, Ersfeld K, Gull K. Targeting of a tetracycline-inducible expression system to the transcriptionally silent minichromosomes of Trypanosoma brucei. Mol Biochem Parasitol. 2002;125(1–2):211–6. Epub 2002/12/07. S0166685102002384 [pii]. pmid:12467990.

26. Yan S, Myler PJ, Stuart K. Tetracycline regulated gene expression in Leishmania donovani. Mol Biochem Parasitol. 2001;112(1):61–9. Epub 2001/02/13. S0166-6851(00)00345-5 [pii]. pmid:11166387.

27. Wen LM, Xu P, Benegal G, Carvaho MR, Butler DR, Buck GA. Trypanosoma cruzi: exogenously regulated gene expression. Exp Parasitol. 2001;97(4):196–204. Epub 2001/06/01. S0014-4894(01)94612-0 [pii]. pmid:11384163.

28. Yan S, Martinez-Calvillo S, Schnaufer A, Sunkin S, Myler PJ, Stuart K. A low-background inducible promoter system in Leishmania donovani. Mol Biochem Parasitol. 2002;119(2):217–23. Epub 2002/01/30. S0166685101004182 [pii]. pmid:11814573.

29. Taylor MC, Kelly JM. pTcINDEX: a stable tetracycline-regulated expression vector for Trypanosoma cruzi. BMC Biotechnol. 2006;6:32. Epub 2006/07/11. 1472-6750-6-32 [pii] pmid:16824206; PubMed Central PMCID: PMC1544328.

30. Yao C, Luo J, Hsiao CH, Donelson JE, Wilson ME. Leishmania chagasi: a tetracycline-inducible cell line driven by T7 RNA polymerase. Exp Parasitol. 2007;116(3):205–13. Epub 2007/02/27. S0014-4894(07)00003-3 [pii] pmid:17320870; PubMed Central PMCID: PMC2231517.

31. Kraeva N, Ishemgulova A, Lukes J, Yurchenko V. Tetracycline-inducible gene expression system in Leishmania mexicana. Mol Biochem Parasitol. 2014;198(1):11–3. Epub 2014/12/03. S0166-6851(14)00166-2 [pii]. pmid:25461484.

32. Alsford S, Kawahara T, Glover L, Horn D. Tagging a T. brucei RRNA locus improves stable transfection efficiency and circumvents inducible expression position effects. Mol Biochem Parasitol. 2005;144(2):142–8. Epub 2005/09/27. S0166-6851(05)00250-1 [pii] pmid:16182389; PubMed Central PMCID: PMC3833055.

33. Biebinger S, Wirtz LE, Lorenz P, Clayton C. Vectors for inducible expression of toxic gene products in bloodstream and procyclic Trypanosoma brucei. Mol Biochem Parasitol. 1997;85(1):99–112. Epub 1997/03/01. S0166685196028150 [pii]. pmid:9108552.

34. Vasquez JJ, Wedel C, Cosentino RO, Siegel TN. Exploiting CRISPR-Cas9 technology to investigate individual histone modifications. Nucleic Acids Res. 2018. Epub 2018/06/19. 5038281 [pii]. pmid:29912461.

35. Hirumi H, Hirumi K. Continuous cultivation of Trypanosoma brucei blood stream forms in a medium containing a low concentration of serum protein without feeder cell layers. J Parasitol. 1989;75(6):985–9. Epub 1989/12/01. pmid:2614608.

36. van den Hoff MJ, Moorman AF, Lamers WH. Electroporation in 'intracellular' buffer increases cell survival. Nucleic Acids Res. 1992;20(11):2902. Epub 1992/06/11. pmid:1614888; PubMed Central PMCID: PMC336954.

37. Hannaert V, Albert MA, Rigden DJ, da Silva Giotto MT, Thiemann O, Garratt RC, et al. Kinetic characterization, structure modelling studies and crystallization of Trypanosoma brucei enolase. Eur J Biochem. 2003;270(15):3205–13. Epub 2003/07/19. 3692 [pii]. pmid:12869196.

38. Kolev NG, Ramey-Butler K, Cross GA, Ullu E, Tschudi C. Developmental progression to infectivity in Trypanosoma brucei triggered by an RNA-binding protein. Science. 2012;338(6112):1352–3. Epub 2012/12/12. 338/6112/1352 [pii]. pmid:23224556; PubMed Central PMCID: PMC3664091.

39. Ngo H, Tschudi C, Gull K, Ullu E. Double-stranded RNA induces mRNA degradation in Trypanosoma brucei. Proc Natl Acad Sci U S A. 1998;95(25):14687–92. Epub 1998/12/09. pmid:9843950; PubMed Central PMCID: PMC24510.

40. Alibu VP, Storm L, Haile S, Clayton C, Horn D. A doubly inducible system for RNA interference and rapid RNAi plasmid construction in Trypanosoma brucei. Mol Biochem Parasitol. 2005;139(1):75–82. Epub 2004/12/22. S0166-6851(04)00275-0 [pii] pmid:15610821.

41. Price HP, Peltan A, Stark M, Smith DF. The small GTPase ARL2 is required for cytokinesis in Trypanosoma brucei. Mol Biochem Parasitol. 2010;173(2):123–31. Epub 2010/07/24. S0166-6851(10)00139-8 [pii]. pmid:20653091; PubMed Central PMCID: PMC2913242.

42. Allen CL, Goulding D, Field MC. Clathrin-mediated endocytosis is essential in Trypanosoma brucei. EMBO J. 2003;22(19):4991–5002. Epub 2003/10/01. pmid:14517238; PubMed Central PMCID: PMC204465.

43. Sunter J, Wickstead B, Gull K, Carrington M. A new generation of T7 RNA polymerase-independent inducible expression plasmids for Trypanosoma brucei. PLoS One. 2012;7(4):e35167. Epub 2012/04/19. PONE-D-11-25566 [pii]. pmid:22511983; PubMed Central PMCID: PMC3325195.

44. McLatchie AP, Burrell-Saward H, Myburgh E, Lewis MD, Ward TH, Mottram JC, et al. Highly sensitive in vivo imaging of Trypanosoma brucei expressing "red-shifted" luciferase. PLoS Negl Trop Dis. 2013;7(11):e2571. Epub 2013/11/28. PNTD-D-13-01086 [pii]. pmid:24278497; PubMed Central PMCID: PMC3836995.

45. Alsford S, Horn D. Single-locus targeting constructs for reliable regulated RNAi and transgene expression in Trypanosoma brucei. Mol Biochem Parasitol. 2008;161(1):76–9. Epub 2008/07/01. S0166-6851(08)00136-9 [pii]. pmid:18588918; PubMed Central PMCID: PMC3828046.

46. McAllaster MR, Sinclair-Davis AN, Hilton NA, de Graffenried CL. A unified approach towards Trypanosoma brucei functional genomics using Gibson assembly. Mol Biochem Parasitol. 2016;210(1–2):13–21. Epub 2016/08/09. S0166-6851(16)30108-6 [pii] pmid:27496178; PubMed Central PMCID: PMC5125862.

47. Jones NG, Thomas EB, Brown E, Dickens NJ, Hammarton TC, Mottram JC. Regulators of Trypanosoma brucei cell cycle progression and differentiation identified using a kinome-wide RNAi screen. PLoS Pathog. 2014;10(1):e1003886. Epub 2014/01/24. PPATHOGENS-D-13-02035 [pii]. pmid:24453978; PubMed Central PMCID: PMC3894213.

48. Kalidas S, Li Q, Phillips MA. A Gateway(R) compatible vector for gene silencing in bloodstream form Trypanosoma brucei. Mol Biochem Parasitol. 2011;178(1–2):51–5. Epub 2011/03/23. S0166-6851(11)00092-2 [pii]. pmid:21420443; PubMed Central PMCID: PMC3101333.