The nuclear mRNA export receptor Mex67-Mtr2 of Trypanosoma brucei contains a unique and essential zinc finger motif


For correspondence. E-mail; Tel. (+41) 31 631 46 52; Fax (+41) 31 631 46 84.


Trypanosoma brucei is the causative agent of Human African Trypanosomiasis. Trypanosomes are early diverged protozoan parasites and show significant differences in their gene expression compared with higher eukaryotes. Due to a lack of individual gene promoters, large polycistronic transcripts are produced and individual mRNAs mature by trans-splicing and polyadenylation. In the absence of transcriptional control, regulation of gene expression occurs post-transcriptionally mainly by control of transcript stability and translation. Regulation of mRNA export from the nucleus to the cytoplasm might be an additional post-transcriptional event involved in gene regulation. However, our knowledge about mRNA export in trypanosomes is very limited. Although export factors of higher eukaryotes are reported to be conserved, only a few orthologues can be readily identified in the genome of T. brucei. Hence, biochemical approaches are needed to identify the export machinery of trypanosomes. Here, we report the functional characterization of the essential mRNA export factor TbMex67. TbMex67 contains a unique and essential N-terminal zinc finger motif. Furthermore, we could identify two interacting export factors namely TbMtr2 and the karyopherin TbIMP1. Our data show that the general heterodimeric export receptor Mex67-Mtr2 is conserved throughout the eukaryotic kingdom albeit exhibiting parasite-specific features.


Gene expression in eukaryotic cells involves a cascade of processes ranging from initiation of transcription, over mRNA maturation and control of transcript stability to protein synthesis by translation. All of these steps can be regulated to adapt the protein content of a given cell to different environmental conditions.

Trypanosoma brucei, a unicellular parasite causing African sleeping sickness in humans and Nagana in cattle, shuttles between an insect and mammalian host and undergoes drastic changes in morphology, surface coat composition and energy metabolism during its complex life cycle. Due to the lack of individual gene promoters and polycistronic transcription of large precursor mRNAs, in trypanosomes gene regulation occurs mainly post-transcriptionally (recently reviewed by Fernández-Moya and Estévez, 2010 and Kramer and Carrington, 2011). Microarray studies and high-throughput RNA sequencing revealed that the steady-state levels of around 5–30% of all mRNAs differ significantly between the two life cycle stages (Siegel et al., 2011). To understand how post-transcriptional gene regulation takes place in trypanosomes, a focus was set on the identification of regulatory sequence elements in differentially expressed mRNAs, and several trans-acting factors binding to these regions could be identified. For instance, Alba proteins and TbZFP3 were shown to affect translation by interacting with regulating loop structures in the 3′ UTR of procyclin mRNA (Hehl et al., 1994; Furger et al., 1997; Walrad et al., 2009; Mani et al., 2011). Furthermore, work on individual RNA-binding proteins revealed co-regulation of genes coding for functionally related proteins such as membrane proteins or factors involved in replicative and metabolic processes (Estevez, 2008; Archer et al., 2009; Stern et al., 2009; Das et al., 2012).

In higher eukaryotes it has been shown that regulated mRNA export can also contribute to differential gene expression and adaptation to environmental changes. A striking example is the heat shock response in yeast. Upon heat stress, bulk mRNA is retained in the nucleus and only mRNAs encoding heat shock proteins are exported (Saavedra et al., 1996; 1997). Regulation of mRNA stability and translational efficiency are certainly the main processes governing gene regulation in trypanosomes. However, we believe that nuclear mRNA export might represent an additional way of post-transcriptional gene regulation in these organisms. Therefore, we started to investigate this process in T. brucei. The active translocation of mRNA from the nucleus to the cytoplasm is well understood in higher eukaryotic model organisms ranging from yeast to human (for reviews see Kohler and Hurt, 2007; Carmody and Wente, 2009; Stewart, 2010). In Saccharomyces cerevisiae, the Mex67-Mtr2 heterodimer associates with mature mRNPs and facilitates the export through the nuclear pore complex by direct interactions with nucleoporins. However, Mex67-Mtr2 has only limited RNA affinity, thus it needs to be recruited to mRNPs by adaptor proteins such as Yra1 (Strasser and Hurt, 2000; Stutz et al., 2000). Yra1 is a member of the transcription-export (TREX) complex which further consists of the DEAD-box RNA helicase Sub2, the conserved protein Tex1 and the THO complex (Chavez et al., 2000; Strasser et al., 2002).

Until recently, nuclear mRNA export was only studied in model organisms belonging to the major supergroup Opisthokonta in which the key proteins seem to be conserved. T. brucei, together with the other Tritryp parasites Leishmania major and Trypanosoma cruzi as well as other important pathogens like Giardia sp. and Trichomonas sp., belong to the supergroup of Excavata and recent comparative genomics have shown that only a very limited number of export factors of Opistohokonts can be readily identified by homology in Excavates (Serpeloni et al., 2011b). One such protein is the trypanosome orthologue of Sub2. A functional analysis showed that it is an essential mRNA export factor. Its association with RNA polymerase II, suggesting co-transcriptional recruitment to nascent mRNA, further supports its homology with yeast Sub2 (Serpeloni et al., 2011a). A second characterized export factor is the orthologue of yeast Mex67. Downregulation of TbMex67 showed that it is essential for cell growth and nuclear mRNA export (Schwede et al., 2009).

Here, we report the functional characterization of TbMex67. Mass spectrometric analysis provides evidence that it contains an N-terminal zinc finger motif. Point mutations of the zinc finger lead to rapid cell death. Furthermore, we could identify two additional essential mRNA export factors by tandem affinity purification of TbMex67.


TbMex67 contains an essential zinc finger motif

TbMex67 is encoded by Tb927.11.2370 (previously Tb11.22.0004; Schwede et al., 2009). The coding sequence published on GeneDB (Logan-Klumpler et al., 2012) translates into a 517-aa-long protein with a calculated molecular weight of 56 kDa. However, it was proposed that TbMex67 has an extended N-terminus leading to a 63 kDa protein composed of 575 aa (Kramer et al., 2010). An analysis of the length of the 5′ UTR (Kolev et al., 2010) and the Spliced Leader (SL) acceptor site (Nilsson et al., 2010) revealed that following the SL there is indeed an ATG start codon further upstream leading to the N-terminally extended TbMex67 which we defined as the wild-type version in this study. Moreover, this 63 kDa protein contains a C-X7-C-X4-C-X3-H zinc finger, a motif that can also be found in Mex67 of T. cruzi and L. major, but is absent from all other Mex67 orthologues known (Kramer et al., 2010). In order to investigate whether the zinc finger is needed for TbMex67 function, we created in a first step a cell line allowing for tetracycline-inducible RNAi directed against the 3′ UTR of the TbMex67 mRNA. After 2 days of tet-induction, the steady-state level of TbMex67 mRNA dropped to 15% (Fig. 1, middle panel) and cells stopped dividing 3 days after induction of RNAi (Fig. 1A, upper panel). Total mRNA fluorescence in situ hybridization (FISH) showed a nuclear retention of poly(A) mRNA (Fig. 1A, lower panel) in cells with downregulated TbMex67 confirming previous results (Schwede et al., 2009).

Figure 1.

Analysis of cell growth and nuclear mRNA export after altering the expression levels of wild-type and mutated versions of TbMex67. The results are representative for at least two independent clones of each cell line.

A. Downregulation of TbMex67 is lethal and results in nuclear accumulation of total poly(A) mRNA.

B. Overexpressed TbMex67 (Mex67-WT) fully complements the TbMex67 RNAi phenotype.

C. N-terminally truncated TbMex67 (Mex67-ΔN) cannot complement the lack of wild-type TbMex67.

D. Mutation of the zinc finger of TbMex67 (Mex67-C29R) is lethal. Growth rates of non-induced cells (− Tet, black line) and induced cells (+ Tet, grey line) were monitored for 5 days (upper panels). The steady-state levels of endogenous (end.) TbMex67 and ectopically expressed (ect.) versions of TbMex67 were analysed by Northern blot. For quantification, the signals were normalized using 18S ribosomal RNA (rRNA) and the levels of endogenous TbMex67 of non-induced cells (day 0) was arbitrarily set to 1 (middle panels). After 2 days of tetracycline-induction, untreated (− Tet) and induced cells (+ Tet) were analysed by mRNA in situ hybridization using fluorescently labelled oligo-(dT). DAPI was used to stain the DNA of the nucleus and the kinetoplast. An overlay of the mRNA and DAPI signals (merge) is shown as well as whole-cell images made with differential interference contrast (DIC) microscopy.

E. Western blot analysis of cell lysates before and after 2 days of tetracycline-induction. Endogenous TbMex67 and ectopically overexpressed versions are detected by anti-Mex67. Anti-Sub2 is used to demonstrate equal loading.

In a second step, this cell line was stably transfected with pLEW100 (Wirtz et al., 1999) derived plasmids for simultaneous tet-induced ectopic expression of TbMex67 variants. In these constructs the sequences encoding the TbMex67 variants are joined to the aldolase 3′ UTR and are therefore resistant to RNAi directed against the wild-type mRNA. Overexpression of wild-type TbMex67 (approximately eightfold, compared with endogenous steady-state mRNA levels) fully rescued the growth phenotype when the endogenous mRNA level was downregulated to less than 10%. Accordingly, the total mRNA distribution in non-induced and induced cells showed no difference with mRNA dispersed all over the cytoplasm and only weak signals in the nucleus (Fig. 1B).

In contrast, an N-terminally truncated version omitting the first 37 aa with the CCCH zinc finger was not able to rescue the parasites (Fig. 1C). Growth arrest occurred already after 1 day of induction, in which the endogenous TbMex67 level was downregulated to 20% whereas the ectopic truncated copy was fourfold overexpressed. The severe growth phenotype is most likely the consequence of mRNA trapped in the nuclei of induced cells as shown by poly(A) FISH. This result, however, does not allow to conclude that the CCCH motif is essential for the function of TbMex67. Therefore, we established another cell line in which endogenous TbMex67 was downregulated to 20% while at the same time, a version with a mutated zinc finger was overexpressed 14-fold. The CCCH to CCRH mutated Mex67 was not able to complement the loss of endogenous TbMex67 and cells stopped growing after 1 day of induction (Fig. 1D). Again, nuclear retention of poly(A) mRNA is the most likely cause for this severe phenotype. For all cell lines, the correct and stable expression of wild-type and mutant versions of TbMex67 was analysed by immunoblot (Fig. 1E). These results led to the conclusion that the zinc finger motif is indispensable for the correct function of TbMex67.

A comparison of the growth curves shows that RNAi against TbMex67 and simultaneous overexpression of truncated or point-mutated TbMex67 lead to growth phenotypes that appear to be more severe than the one caused by RNAi alone. This indicates a dominant-negative effect after overexpression of these Mex67 versions. To test this we transfected cells with only the constructs for inducible overexpression of wild-type TbMex67 and the two mutants. While 10-fold overexpression of wild-type TbMex67 had no effect on cell growth and intracellular mRNA distribution (Fig. 2A), growth phenotypes similar to the ones observed with simultaneous RNAi mediated knock-down became apparent after overexpression of mutated TbMex67 (Fig. 2B and C). Cell growth ceased after 2 days of induction and poly(A) mRNA was found accumulating in the nuclei. Again, the overexpression of all three versions of TbMex67 was monitored by immunoblot (Fig. 2D). A possible explanation for the observed phenotypes is that TbMex67 is only functional in association with interacting proteins and that these putative interactors are sequestered by the overexpressed mutants.

Figure 2.

Dominant-negative effects of overexpressed mutants of TbMex67.

A. Cell growth after 10-fold overexpression of wild-type TbMex67.

B. Overexpression of N-terminally truncated TbMex67 is lethal and affects mRNA export.

C. High levels of single point mutated TbMex67 inhibit cell growth and lead to nuclear mRNA accumulation. Expression levels of wild-type and mutated versions of TbMex67 were analysed as described in the legend of Fig. 1. mRNA FISH was used to monitor the subcellular localization of mRNA.

D. Overexpression of wild-type and mutant versions of TbMex67 upon addition of tetracycline is analysed by immunoblot. TbSub2 serves as loading control.

Two proteins co-purify with PTP-tagged TbMex67

In order to identify interacting proteins we employed tandem affinity purification using PTP-tagged TbMex67. To test if epitope tagging interferes with protein function, we first generated a cell line in which one allele of TbMex67 was replaced by a hygromycin resistance gene. While the single allele knockout did not interfere with cell growth (data not shown) we failed in our attempts of tagging the remaining allele in situ with either the N- or the C-terminal PTP tag suggesting that tagging of TbMex67 resulted in a non-functional protein. Cell lines with one wild-type allele and one either N- or C-terminally tagged TbMex67 were viable. We chose to proceed with C-terminal tagged TbMex67 that, although not fully functional, correctly localized to the nucleus (Fig. 3A). After tandem affinity purification and SDS-PAGE of input, TEV protease cleaved fraction and final eluate we were able to excise two co-purified proteins with apparent molecular weights of 80 kDa and 12 kDa in addition to the bait protein TbMex67 (Fig. 3B). Mass spectrometry unequivocally identified them as Tb927.9.13520 and Tb927.7.5760 respectively. Moreover, the analysis of TbMex67 ultimately proved the existence of peptides containing the N-terminal zinc finger.

Figure 3.

Identification of proteins interacting with TbMex67.

A. Subcellular localization of Mex67-PTP. Antibodies directed against the protein A domains of the PTP-tag were used to detect Mex67-PTP. DNA is stained with DAPI and whole cells are visualized by DIC.

B. Tandem affinity purification of Mex67-PTP. A fraction containing 0.01% of the total-cell lysate (input), 5% of the total TEV digested fraction (TEV) and the entire final eluate (elu) were separated on a 4–20% SDS-polyacrylamide gel. After SDS-PAGE the gel was stained with Coomassie-blue and visible bands were excised and analysed by mass spectrometry which identified the protein C tagged TbMex67 (Mex67-P), TbIMP1 and TbMtr2. The positions of molecular weight marker bands (kDa) are indicated on the left.

C. Immunoblot analysis of co-immunoprecipitation. A cell line expressing Mex67-HA was transfected for stable expression of PTP-Mtr2. After lysate preparation, PTP-tagged proteins were immobilized on IgG-sepharose and input (inp), supernatant (SN) and immunoprecipitate (IP) were analysed by Western blot using antibodies directed against the PTP-tag and the HA-tag. To exclude unspecific binding, the blot was reprobed for the highly abundant translation elongation factor 1α (EF1α). The relative cell equivalents loaded in input, supernatant and IP are indicated. Assays were performed with lysates left either untreated (− RNase) or after treatment (+ RNase) with 10 μg of RNase A.

Tb927.7.5760 encodes a 135 aa long protein with a calculated molecular weight of 15.2 kDa and a theoretical pI of 5.2. It contains a NTF2 domain belonging to the superfamily SSF54427. Searching the yeast proteome for sequences similar to Tb927.7.5760 did not lead to any significant hit. In contrast, a screen of the human proteome identified p15 (UniProt ID Q9UKK6) with the highest P-value (2e-07). The region of highest similarity ranges over ∼ 105 amino acids (Tb927.7.5760 positions 9–114, human p15 positions 12–115) in which 26% of the amino acids are identical. Human p15 (alternatively named NTF2-related export protein 1 or NXT1) forms the stable heterodimer TAP-p15, in which TAP is the human orthologue of yeast Mex67. The yeast functional analogue of p15 is called Mtr2. Since all export factors described so far in trypanosomes were denominated according to the yeast nomenclature, we named the protein encoded by Tb927.7.5760 TbMtr2. To ultimately prove a stable interaction of TbMex67 with TbMtr2, we generated cells expressing HA-tagged TbMex67 (Mex67-HA) and transfected them in a second step for the stable expression of PTP-tagged TbMtr2 (PTP-Mtr2). In co-immunoprecipitation experiments with IgG-coated sepharose, PTP-Mtr2 efficiently pulled down Mex67-HA (as well as wild-type TbMex67, data not shown), both in the presence and in the absence of RNA in the cell lysates (Fig. 3C).

According to the annotation in GeneDB, Tb927.9.13520 codes for a conserved hypothetical protein of 939 aa with a calculated molecular weight of 103 kDa and a theoretical pI of 4.77. A look at recent RNA-seq data (Nilsson et al., 2010), however, revealed that due to a gene-internal splice site the protein sequence comprises only 908 aa leading to a molecular weight of 99 kDa and a pI of 4.73. An InterPro scan for domains and motifs discovered an Armadillo-type fold (SSF 48371) and two HEAT repeat domains (positions 504–541 and 694–731). blast searches in the yeast and human proteomes detected members of the family of karyopherins as high-score hits. Specifically, the highest similarity was observed between Tb927.9.13520 and human transportin-2 (TRN2) with 27% identical amino acids. Human TRN2 was shown to be involved in mRNA export by interacting with the Mex67 orthologue TAP (Shamsher et al., 2002) and a function in the nuclear import of proteins was demonstrated (Guttinger et al., 2004). Because orthologues of Tb927.9.13520 were shown to be importins and Tb927.9.13520 was identified as an interactor of the Mex67 protein, we named the trypanosome protein TbIMP1. A cell line expressing IMP1-PTP was generated to confirm the interaction with TbMex67 by immunoprecipitation. However, in contrast to PTP-Mtr2, immobilized IMP1-PTP does not immunoprecipitate TbMex67 (data not shown) although TbIMP1 was identified by tandem affinity purification of TbMex67 in similar buffer conditions. Attempts to generate cells for exclusive expression of in situ PTP-tagged TbIMP1 in a single knockout background failed (data not shown) suggesting that PTP-tagging interferes with the function of TbIMP which might explain the failed immunoprecipitation of TbMex67.

TbMtr2 is essential for nuclear mRNA export

The cell line that coexpresses HA-tagged TbMex67 and PTP-Mtr2 (Fig. 3C) was used to analyse the intracellular localization of PTP-Mtr2. Immunofluorescence showed that PTP-Mtr2 is localized mainly in the nucleus as expected for a component of the Mex67-Mtr2 heterodimer (Fig. 4A). Colocalization studies of Mex67-HA and PTP-Mtr2 to further corroborate a stable interaction of TbMex67 and TbMtr2 were inconclusive because the antibodies used for detection of the HA-tag also bind the protein A moiety of the PTP-tag (data not shown). Next, we generated a cell line for inducible RNAi of TbMtr2. After 1 day of tetracycline induction, the mRNA of TbMtr2 was undetectable and the cells stopped growing, indicating that TbMtr2 is an essential protein (Fig. 4B). FISH analysis showed that downregulation of TbMtr2 leads to an accumulation of poly(A) mRNA in the nucleus (Fig. 4C). In uninduced cells the mRNA is localized mainly in the cytoplasm. In contrast, a reduced signal strength in the cytoplasm and a nuclear retention of mRNA became visible already after 1 day of RNAi induction. After 2 days most cells showed an abnormal morphology (Fig. 4C). We conclude that TbMtr2 is essential for nuclear export of bulk mRNA.

Figure 4.

Functional analysis of TbMtr2.

A. Nuclear localization of PTP-Mtr2 as detected by immunofluorescence. The PTP-derived signal, the DAPI-stained DNA and an overlay (merge) are shown in addition to a picture of the whole cells (DIC).

B. TbMtr2 is essential for cell growth. A cell line for inducible RNAi against TbMtr2 was cultured in medium lacking (− Tet) or containing tetracycline (+ Tet) and cells were counted and diluted daily. Steady-state levels of TbMtr2 mRNA in non-induced and cells induced for 24 h were determined by Northern blot. Equal loading is demonstrated by ethidium bromide-stained ribosomal RNA (rRNA).

C. Subcellular localization of bulk mRNA in non-induced cells (no Tet) or cells induced for 1 and 2 days. Cy3-labelled oligo-d(T) was hybridized to polyadenylated RNA and signals were detected by fluorescence microscopy. DNA is stained with DAPI, images of whole cells were taken by DIC.

TbIMP1 plays a role in nuclear mRNA export

TbIMP1 was identified by tandem affinity purification of TbMex67 (Fig. 3B). Therefore, we decided to analyse a possible function of TbIMP1 in nuclear mRNA export. We generated a cell line for tetracycline-inducible RNAi of TbIMP1, in which the steady-state level of TbIMP1 mRNA could be downregulated significantly as shown by Northern blot (Fig. 5A). Loss of TbIMP1 led to a growth arrest 2 days after RNAi induction as compared with untreated cells. Again, the growth arrest is most likely due to an mRNA export defect as demonstrated by mRNA FISH (Fig. 5B). These results demonstrate that TbIMP1 is involved in nuclear mRNA export. Human TRN2, the putative orthologue of TbIMP1 was shown to participate in mRNA export (Shamsher et al., 2002). It was proposed that TRN2 interacts with mRNA-bound TAP (the human orthologue of TbMex67) and facilitates its export. However, another study presents the conclusion that TRN2 works as importin (Guttinger et al., 2004) and the negative effect on mRNA export after downregulation of TRN2 is only indirect. In order to test a putative influence of TbIMP1 on the localization of TbMex67, we examined the subcellular distribution of TbMex67 in a cell fractionation assay (Fig. 5C) using cells capable of RNAi against TbIMP1. Mex67-PTP was almost exclusively found in the nucleus (see Fig. 3A). In contrast, wild-type TbMex67 appears to be equally distributed in the nuclear and cytosolic fractions in cells cultured in the absence of tetracycline. We noticed a certain leakage of nuclear material into the supernatant as indicated by the histone H3 signals which might explain a possible over-representation of TbMex67 in the cytosolic fraction. Interestingly, we could observe an increased TbMex67 signal in the pellet fraction of induced cells indicating that the loss of TbIMP1 leads to a nuclear accumulation of TbMex67. We also tested the distribution of TbSub2, another experimentally confirmed nuclear mRNA export factor (Serpeloni et al., 2011a and data not shown). While downregulation of TbIMP1 led to higher nuclear signals for TbMex67, the distribution of TbSub2 does not seem to be affected significantly. Taken together, these data indicate that TbIMP1 plays an essential role in nuclear mRNA export by specifically affecting the subcellular localization of TbMex67.

Figure 5.

Functional analysis of TbIMP1.

A. Downregulation of TbIMP1 by RNAi leads to a growth phenotype. Cell growth was monitored for 4 days in the absence (− Tet) or presence (+ Tet) of tetracycline. Efficiency of RNAi is demonstrated by Northern blot. Ethidium bromide-stained rRNA serves as control for loading.

B. mRNA FISH of non-induced cells (no Tet) and cells induced for 2 or 3 days demonstrates an export defect upon loss of TbIMP1.

C. Cell fractionation assay to determine the subcellular localization of TbMex67 before (− Tet) and after 3 days (+ Tet) of induction of RNAi against TbIMP1. Same cell equivalents of total cells (T), supernatant (SN) and pellet fraction (P) after lysis and separation by centrifugation were subjected to SDS-PAGE and immunoblot. The same blot was probed for histone H3 (serves as control for efficient pelleting of nuclear components), EF1α (cytoplasmic marker), TbSub2 (nuclear mRNA export factor) and TbMex67.


Nuclear mRNA export is a central process in gene expression connecting nuclear events of mRNA production and processing with the cytoplasmic event of protein synthesis by translation. Regulation of mRNA export in trypanosomes might play a role in post-transcriptional gene regulation in trypanosomes. However, very little is known about export factors in these parasites. In contrast, export factors of higher organisms (all belonging to the eukaryotic supergroup of Opisthokonta) are well studied and are reported to be highly conserved within this group. Trypanosomes belong to the distantly related group of Excavata and only a few described export factors of higher eukaryotes reveal orthologues in the trypanosomal genome by bioinformatics. This raises the question: is the mRNA export machinery of T. brucei simplified or highly complex with so far unknown parasite-specific factors? In this study, we started to investigate trypanosomal export factors by a functional characterization of TbMex67. In general, nuclear mRNA export factors of the NXF1 family are not very similar to each other. Pair-wise sequence alignments revealed the highest percentage of 28% identical amino acids shared between human TAP and insect NXF1 whereas TbMex67 and human TAP are only 10% identical (Fig. 6A). Nevertheless, an InterPro search reveals a similar modular architecture of orthologous proteins of the higher eukaryotes belonging to the Opisthokonta (Fig. 6B). The N-terminal part including leucine-rich (LRR) domains mediates binding to RNA and proteins. The NTF2-like domain is needed for heterodimerization with Mtr2/p15 and the C-terminal ubiquitin-associated (UBA) domain together with the NTF2-like domain interacts with FG nucleoporins to facilitate movement through the nuclear pores (reviewed by Stewart, 2010). In contrast, TbMex67 seems to lack the NTF2-like and the UBA-domain. Instead, our results proved the existence of an N-terminal zinc finger motif. This element can only be found in Mex67 of kinetoplastids, but not in any other eukaryotic species with known Mex67 orthologues (Kramer et al., 2010) indicating parasite-specific features of nuclear mRNA export in trypanosomes. We could clearly demonstrate that the presence of this domain is crucial for parasite growth and nuclear mRNA export. Furthermore, overexpression of TbMex67 with a mutated zinc finger caused a dominant-negative effect on cell propagation and nuclear mRNA export indicating that mutants sequestered interacting factors essential for TbMex67 function. Using tandem affinity purification, we could identify two such proteins. We could show that TbMtr2 stably interacts with TbMex67 independent of the presence of RNA, that it is essential for parasite growth and that downregulation leads to a defect in the export of bulk mRNA. This indicates that trypanosomes also need a heterodimeric export receptor made of Mex67 and Mtr2 like higher eukaryotes. It is currently unclear how this receptor comes into contact with export competent mRNAs. The affinity of Mex67 for mRNA is rather low and binding is enhanced by adaptor proteins. Known yeast adaptor proteins for the recruitment of Mex67-Mtr2 are Npl3 (Lei and Silver, 2002), Nab2 (Green et al., 2002; Iglesias and Stutz, 2008) and Yra1 (Strasser and Hurt, 2000). Although all 3 adaptors have been shown to be essential for mRNA export in yeast, no orthologues in trypanosomes could be identified by bioinformatics (Serpeloni et al., 2011b). Yra1 interacts with Sub2 in the TREX complex (Strasser and Hurt, 2000) and in addition, ALY (REF) – the metazoan orthologue of Yra1 – is part of the exon junction complex (EJC) (Le Hir et al., 2000). Neither TbSub2 nor the trypanosomal EJC component TbY14 pull down Yra1-like proteins (B. Schimanski, unpubl. results; Bercovich et al., 2009).

Figure 6.

Sequence comparison of TbMex67 and orthologues of higher eukaryotes.

A. Pair-wise sequence alignments were performed and the percentage of identical and similar amino acids are shown. The analysis included TbMex67, Saccharomyces cerevisiae Mex67 (Sc, UniProt ID Q99257), Drosophila melanogaster NXF1 (Dm, UniProt ID Q9U1H9) and Homo sapiens TAP (Hs, UniProt ID Q9UBU9).

B. An InterPro scan was performed to identify domains and motifs. Detected and drawn to scale are the CCCH-type zinc finger domain (CCCH), leucin-rich repeats (LRR, represented by white circles), NTF2-like domains (NTF2) and UBA-like domains (UBA).

It is possible that the zinc finger in TbMex67 enhances the affinity for RNA. Thus, adaptor proteins might not be needed. Furthermore, various adaptor proteins bind subsets of mRNAs to then provide a platform for the sequence-independent recruitment of the Mex67-Mtr2 heterodimer. In trypanosomes, each mature mRNA molecule contains the identical 39 nucleotides long Spliced Leader sequence at the 5′ end as a consequence of trans splicing (reviewed by Gunzl, 2010). This sequence element common to all mRNAs might be sufficient to sequester the TbMex67-TbMtr2 heterodimer in the absence of adaptor proteins.

In addition to TbMtr2, we identified TbIMP1 in a tandem affinity purification of Mex67-PTP. Downregulation of TbIMP1 by RNAi led to a severe growth phenotype and an inhibition of nuclear export of bulk mRNA. Similar export defects were observed in HeLa cells after loss of Kap β2B (TRN2) – the human orthologue of TbIMP1 (Shamsher et al., 2002). This indicates a conserved function of TbIMP1/TRN2 in nuclear mRNA export. Our data let us conclude that TbIMP1 plays a role in the export of TbMex67 since loss of TbIMP1 leads to a nuclear accumulation of TbMex67. However, we cannot exclude the possibility that the involvement of TbIMP1 in the subcellular localization of TbMex67 is indirect. For instance, it can be that TbIMP1 functions as importin to mediate the nuclear import of so far unidentified exporters of TbMex67. Given the overall conserved architecture of the nuclear pore complex (NPC) (DeGrasse et al., 2009), it is not surprising that the general export receptor that guides mRNPs through the NPC shows conservation in its heterodimeric composition. However, the domain structure of TbMex67 with the N-terminal zinc finger motif differs greatly from the orthologues of higher eukaryotes. The presence of this motif might compensate for the apparent lack of adaptor proteins indicating a simplified mRNA export in trypanosomes. However, the export machinery could also be highly complex with export factors too divergent to be identified based on sequence homology. Therefore, further investigations of the functions and interacting proteins of the identified export factors will shed more light on the nature of the mRNA export machinery in trypanosomes.

Experimental procedures

DNA sequences

All inserts for plasmids were generated by PCR from genomic DNA of T. brucei 427. See Table 1 for a complete list of oligonucleotide primers used in this study.

Table 1. Oligonucleotides used in this study (recognition sites for restriction endonucleases are underlined)
 Oligonucleotide sequence (5′ to 3′)

pBS25, pBS28 and pBS43 are derivatives of pSLcomp1 (Mani et al., 2011) for inducible RNAi directed against TbIMP1, TbMtr2 and TbMex67 respectively. Inserts were amplified with primer combinations 1–38/1–39 (TbIMP1), 1–46/1–47 (TbMtr2) and 1–17/1–18 (TbMex67). RNAi targets ORF positions 831–1337 of TbIMP1, the entire ORF of TbMtr2 and the first 518 nucleotides of the TbMex67 3′ UTR.

pBS44, pBS45 and pBS49 are based on pLEW100 (Wirtz et al., 1999) modified to contain a blasticidin resistance gene and were used for overexpression of wild-type TbMex67, N-terminally truncated TbMex67 and a mutant zinc finger version of TbMex67 respectively. Inserts for pBS44 and pBS45 were generated with primer combinations 2–9/2–10 and 2–11/2–10, respectively, digested with HindIII and BglII and ligated into HindIII and BamHI of the pLEW100 derivative. For pBS49, primers 2–34 and 2–37 were annealed and extended with Taq polymerase. The fragment was then PCR-amplified with 2–9 and 2–38, digested with HindIII and XhoI and ligated into these sites of pBS45.

For pBS13, TbMex67 ORF positions 1387–1725 were amplified with primers 1–21 and 1–22, digested with ApaI and EagI and ligated into ApaI and NotI of pC-PTP-NEO (Schimanski et al., 2005). The same fragment was ligated into ApaI and NotI of pRPA31-HA-BLA to generate pBS20.

pBS21 contains TbMtr2 ORF positions 3–403 amplified with 1–42 and 1–43 and ligated into NotI and ApaI of pN-PURO-PTP (Schimanski et al., 2005).

pBS22 was generated by ligating a PCR product amplified from genomic DNA with 1–40 and 1–41 comprising TbIMP1 ORF positions 2001–2724 into ApaI and NotI of pC-PTP-NEO.

Trypanosome cell lines and transfection

All cell lines were cultured in SDM-79 (Brun and Schonenberger, 1979) containing 10% fetal calf serum. Procyclic T. brucei 427 cells were stably transfected with BsmI-linearized pBS13 or pBS20 to generate cell lines Mex67-PTP and Mex67-HA respectively. The latter was further transfected with EcoRI-linearized pBS21 or AvaI-linearized pBS22 to generate cells expressing PTP-Mtr2 or IMP1-PTP, respectively, in a Mex67-HA background. Mex67-RNAi, Mtr2-RNAi, IMP1-RNAi and cells for inducible expression of Mex67-WT, Mex67-ΔN or Mex67-C29R were obtained after stable transfection of the host strain 29-13 (Wirtz et al., 1999) with NotI-linearized pBS43, pBS25, pBS28, pBS44, pBS45 or pBS49 respectively. Mex67-RNAi cells were further transfected with pBS44, pBS45 or pBS49 to allow simultaneous induction of RNAi against endogenous TbMex67 and overexpression of wild-type Mex67, N-terminally truncated TbMex67 or TbMex67 with a mutated zinc finger respectively. Expression was induced by addition of 1 μg ml−1 tetracycline to the culture media. Transfectants were selected in the presence of 25 μg ml−1 hygromycin, 15 μg ml−1 G418, 2.5 μg ml−1 phleomycin, 1 μg ml−1 puromycin or 10 μg ml−1 blasticidin.

RNA and protein analysis

Total RNA was isolated with acid guanidinium thiocyanate (Chomczynski and Sacchi, 1987). Agarose gel electrophoresis, Northern blotting, radioactive probe labelling, hybridization and detection were done as described (Haenni et al., 2006). For Western blotting, proteins were separated by SDS-PAGE and semi-dry transferred onto PVDF membranes. Polyclonal anti-Mex67 (kindly provided by Mark Carrington, diluted 1:2000) was raised in rabbits against the recombinant full-length protein. PTP tagged proteins were detected with PAP-reagent (Sigma, 1:5000), HA tagged Mex67 was detected with anti-HA 3F10 (Roche, 1:5000). The polyclonal anti-Sub2 (diluted 1:10 000) was raised in rats immunized with recombinant full-length TbSub2. Anti-EF1α (1:10 000) and anti-Histone H3 (1:2000) are commercially available from Santa Cruz Biotechnology and Abcam respectively. Chemiluminescence generated by peroxidase-labelled secondary antibodies (diluted 1:5000) was detected using SuperSignal® West Pico Chemiluminescent Substrate (Thermo Scientific).


A total of 109 cells were harvested, washed twice with PBS and resuspended in 900 μl of buffer E (150 mM sucrose, 20 mM potassium-l-glutamate, 20 mM Tris-HCl pH 7.7, 3 mM MgCl2, 0.5 mM DTT, Roche Complete Mini EDTA-free protease inhibitor) and allowed to swell for 10 min on ice. Nonidet-P40 was added to a final concentration of 0.1% and cells were lysed for 2 min on ice. One hundred microlitres of extraction buffer (1500 mM KCl, 20 mM Tris-HCl pH 7.7) was added and insoluble material was pelleted by centrifugation at 16 000 g for 10 min. A fraction of the supernatant corresponding to 2.5 × 108 cell equivalents was incubated with 30 μl (settled volume) of IgG-coated sepharose beads (GE Healthcare) for 1 h at 4°C. After the supernatant was taken off, the beads were washed five times with buffer PA-150 (150 mM KCl, 20 mM Tris-HCl pH 7.7, 3 mM MgCl2, 0.1% Tween 20) and proteins were eluted by boiling the beads in SDS-PAGE loading buffer.

Tandem affinity purification (TAP) of Mex67-PTP

TAP of PTP tagged TbMex67 was performed with an extract from 1.7 × 1010 cells exactly as described (Schimanski et al., 2005). Eluted proteins were separated by denaturing electrophoresis in Mini-PROTEAN® precast 4–20% gradient gels (Bio-Rad). After staining with blue silver colloidal Coomassie (Candiano et al., 2004), individual bands were excised and proteins were identified by mass spectrometry (LC-MS/MS).

Fluorescence in situ hybridization (FISH) and immunofluorescence microscopy

Cells were harvested by centrifugation, washed with PBS, resuspended in PBS, allowed to settle on coverslips for 20 min and fixed with 4% paraformaldehyde in PBS. Poly(A) RNA FISH was performed according to Cassola et al. (2007). Briefly, cells were permeabilized for 1 h in blocking buffer (PBS containing 0.5% saponin, 2% BSA, 10 U ml−1 RNase inhibitor). Permeable cells were pre-hybridized for 2 h with hybridization solution (2% BSA, 5× Denhardt's solution, 4× SSC, 5% dextran sulphate, 35% formamide, 0.5 μg μl−1 tRNA, 10 U ml−1 RNase inhibitor). Cy3-labelled oligo-(dT)30 (2 ng μl−1 in hybridization solution) was hybridized overnight at room temperature. Afterwards cells were washed with decreasing concentrations of SSC and mounted with moviol containing DAPI (1.5 μg ml−1).

For immunofluorescence, fixed cells were permeabilized with Triton X-100 (0.2% in PBS) and blocked with PBS/2% BSA. Cells were incubated for 1 h with rabbit anti-protein A (Sigma) diluted 1:1000 in PBS/2% BSA. The secondary Alexa Fluor® 568 goat anti-rabbit antibody was diluted 1:1000 and incubated for 1 h. Cells were mounted with moviol containing DAPI (1.5 μg ml−1). Images were taken with a Leica DFC360FX monochrome CCD camera on a Leica DM5500 B microscope and analysed using LAS AF software (Leica).

Cell fractionation

A total of 2.5 × 108 cells were harvested, washed in PBS and finally resuspended in 600 μl of PBS. Five hundred microlitres of cell suspension was added to 10% Nonidet-P40 to obtain a final concentration of 0.1% Nonidet-P40. Immediately after mixing, the cell suspension was centrifuged for 1 min at 2300 g to separate the supernatant containing cytosolic proteins from the pellet fraction containing the nuclei. SDS-PAGE loading buffer was added to the pellet and an aliquot of the supernatant. Samples corresponding to same cell equivalents from whole cells, supernatant and pellet fractions were subjected to SDS-PAGE and immunoblot.


B.S. is supported by the Ambizione programme of the Swiss National Science Foundation (Grant No. 131959). We thank the Institute of Cell Biology and Isabel Roditi for lab space and support. Mark Carrington (University of Cambridge, UK) kindly provided anti-Mex67. Gaby Schumann-Burkard is thanked for critical reading of the manuscript. We are grateful to Sophie Braga Lagache and Manfred Heller for excellent mass spectrometry.