Cilia and flagella are conserved eukaryotic organelles important for motility and sensory. The RanGTPase, best known for nucleocytoplasmic transport functions, may also play a role in protein trafficking into the specialized flagellar/ciliary compartments, although the regulatory mechanisms controlling Ran activity at the flagellum remain unclear. The unicellular parasite Trypanosoma brucei contains a single flagellum necessary for cell movement, division and morphogenesis. Correct flagellum functions require flagellar attachment to the cell body, which is mediated by a specialized flagellum attachment zone (FAZ) complex that is assembled together with the flagellum during the cell cycle. We have previously identified the leucine-rich-repeat protein 1 LRRP1 on a bi-lobe structure at the proximal base of flagellum and FAZ. LRRP1 is essential for bi-lobe and FAZ biogenesis, consequently affecting flagellum-driven cell motility and division. Here, we show that LRRP1 forms a complex with Ran and a Ran-binding protein, and regulates Ran–GTP hydrolysis in T. brucei. In addition to mitotic inhibition, depletion of Ran inhibits FAZ assembly in T. brucei, supporting the presence of a conserved mechanism that involves Ran in the regulation of flagellum functions in an early divergent eukaryote.
Protein trafficking into the membrane-bound flagellum or cilium is tightly controlled by polarized vesicular trafficking as well as a diffusion barrier located at the base of the structure (Hsiao et al., 2012), creating a privileged compartment crucial for flagellar biogenesis and functions. In recent years, several studies have described a role of the small GTPase Ran – best known for its biological function in nucleocytoplasmic protein transport (Sazer and Dasso, 2000) – in the regulation of ciliary protein targeting in various eukaryotes (Dishinger et al., 2010; Fan et al., 2011; Ludington et al., 2013). Whereas the regulation of Ran activity at the nucleus is relatively well-understood, how Ran activity is regulated at the ciliary base is not yet clear.
In Trypanosoma brucei, the etiological agent for human african trypanosomiasis (HAT; also known as sleeping sickness), a single flagellum is present in each parasite cell, required for cell locomotion, immune evasion, cell division, organelle positioning and cell length regulation (Absalon et al., 2008; Absalon et al., 2007; Broadhead et al., 2006; Engstler et al., 2007; Kohl et al., 2003; Ralston et al., 2006). The normal function of T. brucei flagellum requires flagellum attachment to the cell body, through the flagellum attachment zone (FAZ), a unique feature found in T. brucei and related trypanosomatid parasites (Vaughan, 2010). Disruption of flagellum attachment in T. brucei leads to defects in flagellum-driven cell motility and cell division as well as changes in cell morphology, often resulting in cell death (LaCount et al., 2002; Sun et al., 2013; Vaughan et al., 2008; Woods et al., 2013). Little, however, is known about the protein composition of the FAZ complex and how it is assembled together with the flagellum.
A bi-lobe structure is present at the proximal base of the single flagellum, partially overlapping with the proximal tip of the FAZ (Morriswood et al., 2012; Shi et al., 2008) and adjacent to the single Golgi complex (He et al., 2005). Previously thought to be required for Golgi duplication, recent studies suggest a more fundamental role of the bi-lobe in FAZ assembly that, consequently, affects flagellum functions in organelle inheritance, cell division and cell morphogenesis (Zhou et al., 2010; Zhou et al., 2011). Depletion of the bi-lobe proteins Centrin2 (gene accession number Tb927.8.1080) and LRRP1, both leads to defects in FAZ assembly, flagellum inheritance, flagellum-driven cell motility and cell division (Shi et al., 2008; Zhou et al., 2010). To better understand the bi-lobe and the molecular function of LRRP1 and to identify LRRP1-interacting proteins, we carried out yeast-two-hybrid screens. The results showed that LRRP1 forms a complex with Ran GTPase and a previously uncharacterized Ran-binding protein. Further functional characterization suggested a role of LRRP1 as a Ran regulator specific for flagellar functions.
Yeast-two-hybrid screening for LRRP1-interacting proteins
T. brucei LRRP1 has previously been identified in a comparative proteomics screening for flagellum-associated proteins (Zhou et al., 2010). The 713-aa protein lacks sequence homology to other organisms but is conserved in trypanosomatids. LRRP1 is stably associated with the bi-lobe structure in T. brucei, shown by stable expression of YFP-tagged LRRP1 or by immunostaining with antibody against LRRP1 (Zhou et al., 2010). LRRP1 comprises an N-terminal leucine-rich repeat (LRR) domain followed by a 50-aa-long coiled-coil region within the C-terminal part (Fig. 1A). LRR domains are generally required for protein–protein interactions (Bella et al., 2008; Kobe and Kajava, 2001). The LRR in LRRP1 belongs to the ribonuclease inhibitor (RI)-like subfamily – with an Expect (E)-value of 1.27×10−23 – whose structure has been first described in the porcine ribonuclease inhibitor (Kobe and Deisenhofer, 1993). The RI-like LRR domain is also found in all known Ran–GTPase-activating proteins (RanGAPs) such as yeast rna1 (Hillig et al., 1999), human RANGAP1 (Haberland and Gerke, 1999) and plant RanGAP (Rose and Meier, 2001). In T. brucei, the short coiled-coil domain alone was sufficient to mediate bi-lobe localization of the YFP reporter (CC–YFP) (Fig. 1C; supplementary material Fig. S1A), possibly through its protein oligomerization activity (Meier et al., 2010). The LRR domain, however, does not contain bi-lobe-targeting information and LRR–YFP was found throughout the parasite cell (Fig. 1B; supplementary material Fig. S1A).
A yeast-two-hybrid screening was performed to search for T. brucei proteins that interact with the LRR or the coiled-coil domain. One LRR-binding candidate, encoded by Tb927.10.8650, is a 337-aa hypothetical protein that is conserved in trypanosomatids but has no homolog in other eukaryotes. It contains a putative Ran-binding domain (RanBD, with an E-value of 1.29×10−10) in the C-terminal region (Fig. 1D), suggesting that it binds to Ran in T. brucei.
T. brucei contains a bona fide Ran that is essential for nuclear division
The highly conserved small GTPase Ran is a key regulator of nucleocytoplasmic transport (Sazer and Dasso, 2000; Yudin and Fainzilber, 2009). Relying on the combined action of RanGEF (RCC1) in the nucleus and RanGAP in the cytosol, a Ran gradient is established across the nuclear envelope, with high Ran–GTP levels in the nucleus and high Ran–GDP levels in the cytosol.
Ran has been identified in many protozoan parasites, including Giardia, apicomplexans and trypanosomatids (Casanova et al., 2008; Chen et al., 1994; Frankel and Knoll, 2008). In T. brucei, a single Ran homolog is encoded by Tb927.3.1120 (Casanova et al., 2008; Field et al., 1995), with its amino acid sequence sharing an overall 78.5% similarity to yeast Ran. Stable expression of a YFP fusion protein of Ran in T. brucei (YFP–Ran) resulted in distinct nuclear staining throughout the cell cycle (Fig. 2A). This distribution in T. brucei is similar to that observed in yeast and mammalian cells, but appears different to the peri-nuclear localization observed for Ran in Leishmania major (Casanova et al., 2008) and the evenly throughout the cell distributed Ran in Toxoplasma gondii (Frankel and Knoll, 2008). The nuclear localization of Ran in T. brucei was further verified by using a monoclonal antibody that crossreacts with T. brucei Ran on immunoblots (Fig. 2B,C). Ran was present in the nucleus at all times of the cell cycle, consistent with the ‘closed’ mitosis observed in this parasite. Inducible silencing of Ran by RNA interference (RNAi) led to rapid, albeit incomplete, Ran depletion (Fig. 2D) that was sufficient to arrest cell proliferation 24 h post induction (Fig. 2E). The DNA contents of Ran-RNAi cells was examined by fluorescence microscopy and the results indicated a rapid increase of cells that contain duplicated mitochondrial DNA (aka the kinetoplast) and a single, undivided nucleus (hereafter referred to as 2K1N cells). 2K1N cells rose from ∼10% in control populations to >20% at 12 h post RNAi targeting Ran and >40% at 24 h post RNAi targeting Ran, indicating efficient mitotic arrest (Fig. 2F). In later induction stages, the fraction of zoids (cells that lack nuclei) and cells that contain other aberrant DNA compositions (including 1K2N cells and multi-kinetoplast and multi-nuclei cells) also increased, consistent with dysregulated cell division after mitotic block. Together, these results indicate that T. brucei contains a bona fide Ran with a distinct nuclear localization and a strong effect on mitotic progression, similar to Ran GTPases characterized in higher eukaryotes (Sazer and Dasso, 2000).
Tb927.10.8650 encodes a Ran-binding protein
Interaction of the RanBD-containing Tb927.10.8650 with T. brucei Ran was investigated by using a pull-down approach. His-fusion to the full length Tb927.10.8650 or the C-terminal RanBD-containing region was purified and immobilized to Ni2+-NTA beads and incubated with GST–Ran purified from Escherichia coli. Both full length Tb927.10.8650 and the C-terminal RanBD interacted specifically with GST–Ran but not GST alone (Fig. 3A,B). Interactions were GTP-dependent and drastically inhibited when GST–Ran had been pre-loaded with GDP. Pre-loading of GST–Ran with non-hydrolyzable GTPγS did not significantly enhance Ran binding to the beads, possibly because of a large amount of GST–Ran already in GTP-bound form. Furthermore, His-Tb927.10.8650 did not interact with T. brucei Rab1, another small GTPase (Dhir et al., 2004), regardless of its GTP- and/or GDP-bound states (Fig. 3C). Together, these results indicate that Tb927.10.8650, despite the sequence variation within the RanBD (Fig. 1D), binds specifically and selectively to GTP-bound Ran, just like other Ran effectors (Coutavas et al., 1993).
The protein product of Tb927.10.8650 was thus named RanBP-like protein (hereafter referred to as RanBPL), in order to distinguish this protein from T. brucei RanBP1 (Casanova et al., 2008), the only other RanBD-containing protein found in T. brucei genome database. Although both contain a RanBD region (Fig. 1D), the primary sequence of RanBPL lacks similarity to RanBP1. Using a polyclonal antibody specific against RanBPL, T. brucei RanBPL was found to be present on the nucleus (supplementary material Fig. S2B); this was similar to the localization observed for YFP–RanBP1 in T. brucei (supplementary material Fig. S2A) and in Leishmania major (Casanova et al., 2008). However, unlike RanBP1, which is essential for T. brucei survival (Casanova et al., 2008), depletion of RanBPL had only a moderate effect on proliferation of T. brucei cells in culture (supplementary material Fig. S2C).
Interactions of RanBP1 and RanBPL with Ran from T. brucei cell lysates were also evaluated in pull-down assays (Fig. 3D). His–RanBP1 and His–RanBPL both were able to bind Ran and pulled down endogenous Ran protein. Interestingly, whereas the binding of RanBPL to Ran was GTP-dependent – with increased binding in the presence of GTPγS and inhibited binding in the presence of GDP – the binding of RanBP1 to Ran was less affected by the presence of GTPγS or GDP (Fig. 3E). It is not clear what may account for this less discriminative binding of T. brucei RanBP1 to GTP- and GDP-bound Ran. Only RanBPL, however, was investigated in further studies.
T. brucei Ran forms a complex with RanBPL and LRRP1
The interaction between RanBPL and LRRP1, initially identified in the yeast-two-hybrid screen, was verified by using pull-down analysis. As shown in Fig. 4A, His–RanBPL-coated beads were incubated with parasite cell lysates. His–RanBPL was able to pull down endogenous LRRP1, supporting the hypothesis that RanBPL associates with LRRP1. Interestingly, the interaction between RanBPL and LRRP1 relied on the presence of endogenous Ran. Neither full-length RanBPL nor C-terminal RanBD were able to pull down LRRP1 from cell lysates that had been induced for RNAi targeting Ran, although expression of LRRP1 was not affected in Ran-depleted cells (Fig. 4A).
We next asked whether T. brucei Ran interacts with LRRP1 via the LRR domain that was used as prey in the yeast-two-hybrids screen (see above). To test this, GST–Ran or GST only were immobilized on agarose beads and incubated with T. brucei cell lysates expressing LRR–YFP or CC–YFP (Fig. 4B). As expected, LRR–YFP but not CC–YFP was pulled down with GST–Ran. Like the Ran–RanBPL interaction (c.f. Fig. 3), the interaction between the LRR domain and Ran was also GTP-dependent and inhibited in the presence of GDP (Fig. 4C). As further control, LRR–YFP did not interact with GST–Rab1 in the presence of GTPγS or GDP (Fig. 4D).
Direct interaction between Ran and the LRR domain was verified next by using bacterially expressed GST–LRR and His–Ran. GST or GST–LRR were purified and immobilized on agarose beads, then incubated with E. coli lysates expressing His–Ran (Fig. 4E). His–Ran was pulled down by GST–LRR but not GST alone, confirming direct interaction between Ran and the LRR domain of LRRP1. Taken together, the above experiments suggest that RanBPL and LRRP1 each interact directly with Ran through the RanBD and the LRR domain, respectively, and – together – form a complex (Fig. 5A).
To further test whether endogenous Ran, LRRP1 and RanBPL forms a complex, co-immunoprecipation was performed by using anti-Ran, anti-LRRP1 or anti-RanBPL antibodies. Neither anti-Ran nor anti-LRRP1 antibodies worked for immunoprecipitation, because Ran and LRRP1 did not immunoprecipitate following the use of their respective monoclonal antibodies (Fig. 5B). Polyclonal antibody against RanBPL, however, was able to immunoprecipitate RanBPL (Fig. 5C). Interestingly, Ran and LRRP1 only co-immunoprecipitated with antibody against RanBPL when cell lysates were supplemented with GTPγS, possibly stabilizing the GTP-dependent interactions.
The formation of a complex between LRRP1, Ran and RanBPL (Fig. 5A) is, thus, reminiscent of the RanGAP–Ran–RanBP1 ternary complex that has previously been identified in yeast and mammalian cells (Seewald et al., 2002; Seewald et al., 2003). In this ternary complex, Ran interacts with RanBD of RanBP1 on one side and the LRR domain of RanGAP on the other side; and these interactions allow rapid and irreversible hydrolysis of Ran–GTP to Ran–GDP. T. brucei LRRP1, which contains an LRR domain that interacts with RanBPL in a Ran-dependent manner and interacts with Ran in GTP-dependent manner, might thus play a role in Ran regulation.
LRRP1 facilitates Ran–GTP hydrolysis
The insolubility of T. brucei Ran expressed in E. coli prevented direct in vitro enzymatic assay (data not shown). We, therefore, examined the effect of LRRP1 on Ran–GTP hydrolysis in an ex vivo Ran activation assay. In this assay, His-tagged T. brucei RanBPL immobilized on agarose beads was incubated with T. brucei cell lysate that had been depleted of LRRP1 or with lysate of cells that had been overexpressing LRR–YFP. The LRR domain of RanGAP is sufficient to mediate Ran binding and Ran–GTP hydrolysis within the ternary complex (Seewald et al., 2002; Seewald et al., 2003). Unlike full-length LRRP1–YFP, whose expression suppresses the level of endogenous LRRP1 (Zhou et al., 2010), stable expression of LRR–YFP did not affect endogenous expression of LRRP1 (supplementary material Fig. S1B). The selective binding of His–RanBPL to Ran–GTP (c.f. Fig. 3D,E) was utilized to specifically pull down Ran–GTP present in the cell lysates, allowing quantification of Ran–GTP levels in cell lysates of parasites by immunoblotting (Fig. 6A).
The total level of Ran was not affected by LRRP1 depletion or LRR–YFP overexpression, evidenced by similar Ran levels in non-induced control and LRRP-RNAi cells (Fig. 6B). Significantly more Ran–GTP was pulled down from LRRP1-RNAi lysates – and less from LRR–YFP overexpression lysates (Fig. 6B) – than from the non-induced controls. This was verified by quantification (Fig. 6C). These observations suggested that Ran–GTP hydrolysis is inhibited in cells that lack LRRP1 and enhanced in cells that express LRR–YFP, supporting a role of LRRP1 in Ran–GTP hydrolysis.
Unlike Ran, LRRP1 has only a moderate effect on the nucleus
The nucleocytoplasmic Ran–GTP gradient is crucial for protein trafficking across the nuclear membrane. LRRP1-depletion substantially inhibits flagellum-driven cell motility and cell division, but has no obvious effects on nuclear division (Zhou et al., 2010). This lack of nuclear effects in LRRP1-depleted cells appears to contradict the above observations that LRRP1 affects Ran–GTP hydrolysis and intracellular Ran distribution, and is also in contrast to the rapid mitotic arrest observed in Ran-RNAi cells (c.f. Fig. 2).
To further confirm the lack of effect LRRP1 depletion has on the nucleus, the localization of the RNA-binding protein La was monitored by using antibody against La in LRRP1-RNAi and Ran-RNAi cells (Fig. 7). La contains a nuclear localization signal (NLS) and is normally found in the nucleus (Fig. 7A) (Marchetti et al., 2000). La expression levels were not affected following RNAi targeting either LRRP1 or Ran (Fig. 7H); however, its nuclear localization was greatly inhibited in Ran-RNAi cells (Fig. 7E) but only moderately affected in LRRP1-RNAi cells (Fig. 7C). The change in La localization was quantified by intensity profiling across the nucleus (Fig. 7B,D,F) and by comparing the relative intensity of La and DNA staining over the nuclear region (Fig. 7G). The moderate effect LRRP1 has on nuclear protein import is consistent with the previous observation that RNAi that targets LRRP1 does not affect nuclear division (Zhou et al., 2010), which supports a primary, non-nuclear effect of this Ran regulator.
T. brucei Ran affects FAZ assembly
Recent studies have found a role of the Ran/importin system in flagellar protein trafficking (Dishinger et al., 2010; Fan et al., 2011; Ludington et al., 2013). In T. brucei, the presence of LRRP1, a Ran regulator at the bi-lobe essential for flagellum functions, suggested a conserved mechanism of Ran in flagellum regulation in addition to its nuclear functions (c.f. Figs S2 and S7).
To examine the effects of Ran on flagellum or FAZ biogenesis, control and Ran-RNAi cells (24 h post induction) were scored for new flagellum attachment, new flagellum length and new FAZ length during the cell cycle of T. brucei (Fig. 8). We chose to focus on 2K1N-stage cells that contain two kinetoplasts and one nucleus because this is the stage at which T. brucei cells undergo flagellum and FAZ formation. Besides, due to mitotic inhibition, 2K1N cells accounted for ∼10% of an asynchronous control population and ∼40% of the Ran-RNAi population (Fig. 2F). In control cells, new FAZ and new flagellum elongate in a coordinated way, their lengths correlating strongly (R2 = 0.74). As such, the majority (∼90%) of new flagella were attached to the cell body. By contrast, in Ran-RNAi cells >30% of the cells contained new flagella that were partially detached, and the coordinated assembly of new flagellum and new FAZ was disrupted in these cells (R2 = 0.32) (Fig. 8A,B). Furthermore, whereas the length distribution of new flagella was similar in control and Ran-RNAi cells (Fig. 8B), that of new FAZs was consistently shorter in Ran-depleted cells throughout the stages of new flagellum synthesis (Fig. 8C–G). Together, these results supported a role of T. brucei Ran in FAZ assembly.
The leucine-rich-repeat-containing protein LRRP1 has previously been characterized as a protein resident on the bi-lobe and involved in diverse cellular processes including coordinated organelle duplication, cell motility and cell division (Zhou et al., 2010), all of which require normal function of flagellum and flagellum attachment zone (FAZ). Whereas assembly of the flagellum appeared unaffected in LRRP1-depleted cells, FAZ assembly was significantly inhibited (Zhou et al., 2010).
In this study, LRRP1 was found to interact with a previously uncharacterized Ran-binding protein RanBPL (encoded by Tb927.10.8650) in a Ran-dependent manner. The small GTPase Ran interacted directly with both the RanBD in RanBPL and the LRR domain in LRRP1. Although robust interactions were readily detected with overexpressed proteins in T. brucei or in E. coli, complex formation and colocalization of the endogenous proteins were difficult to observe. Whereas LRRP1 resides on the bi-lobe, Ran and RanBPL both show primary nuclear localization. Neither Ran nor RanBPL was found associated with the bi-lobe in immunofluorescence experiments or YFP-reporter assays. Consistently, formation of the complex by endogenous Ran, RanBPL and LRRP1 was not observed unless GTPγS was also present to stabilize the GTP-dependent interactions. The LRRP1–Ran–RanBPL complex is, thus, similar to the RanGAP–Ran–RanBP1 ternary complex found in higher eukaryotes. Similarly, in yeast and mammalian cells, both RanBP1 and RanGAP are cytosolic proteins, whereas Ran is predominantly found in the nucleus (Quimby and Dasso, 2003). In vivo, direct observation of the formation of the Ran–RanBP1–RanGAP complex is also challenging, unless overexpressed proteins are used. This is probably owing to the transient, enzymatic nature of these interactions (Seewald et al., 2003).
Whether LRRP1 acts as a RanGAP remains to be experimentally confirmed. Owing to sequence divergence, identification of RanGAPs by using bioinformatics has been challenging, particularly in early-branching eukaryotes (Frankel and Knoll, 2009). Several lines of evidence suggested that LRRP1 is a Ran regulator, possibly a RanGAP: (1) the LRR domain of LRRP1 belongs to the RI-like family, like the LRRs of other previously characterized RanGAPs; (2) Ran interacts directly with the LRR domain of LRRP1 in a GTP-dependent fashion; (3) endogenous LRRP1, RanBPL and Ran form a complex; 4) LRRP1 protein expression in T. brucei affects Ran–GTP levels. Further confirmation of the RanGAP activity requires biochemical and/or enzymatic analyses, which so far – mainly because of the insolubility of T. brucei Ran expressed in E. coli – have not been successful.
Unlike most Ran regulators identified to date that are either on the nuclear envelope or in the cytosol and crucial for nuclear function (Rose and Meier, 2001; Frankel and Knoll, 2009; Casanova et al., 2008), LRRP1 of T. brucei is present at the bi-lobe structure – a cytoskeletal complex located at the proximal base of the single flagellum (Zhou et al., 2010) – possibly through oligomerization of the C-terminal coiled-coil domain. Despite perturbation of Ran–GTP hydrolysis, depletion of LRRP1 has only modest effects on nuclear division and nuclear protein import – the latter at least in the case of La. Rather, as previously reported (Zhou et al., 2010), RNAi targeting LRRP1 inhibits duplication of the bi-lobe and assembly of the FAZ, consequently inhibiting flagellum inheritance, flagellum-driven cell motility and cell division. Thus, the role of LRRP1 as a Ran regulator is likely to be restricted to flagellum-associated structures, such as the bi-lobe and the FAZ, both of which are crucial to flagellum inheritance and flagellum functions.
The lack of effect on nuclear division in LRRP1-RNAi cells is in contrast to the rapid mitotic arrest observed in the Ran-RNAi cells. It is possible that ‘closed’ mitosis and nuclear protein import in T. brucei function independently of the Ran–GTP/Ran–GDP gradient but require the presence of Ran. It is also possible that other RanGAP or Ran-regulators are present in the parasite, that are sufficient to maintain nuclear function and division. For T. brucei, a putative RanGAP (encoded by Tb927.7.1430) that contains an RI-type LRR has previously been predicted bioinformatically; however, so far there is no biochemical or functional evidence in support of its RanGAP activity (Casanova et al., 2008).
The presence of LRRP1, a Ran regulator at the bi-lobe suggested a possible Ran-mediated function at the base of the flagellum. Indeed, in addition to nuclear functions, T. brucei Ran was found to also have an effect on FAZ assembly. The FAZ is a structure uniquely observed in the trypanosomatids, whose coordinated assembly with the flagellum is crucial for correct flagellar functions (Vaughan, 2010). It is interesting to note that, although both LRRP1 and Ran affect FAZ assembly, neither protein has obvious effects on the biogenesis or structure of the flagellum (Zhou et al., 2010). How LRRP1 and Ran affect FAZ assembly remains unknown. It is, however, possible that the observed FAZ defect is a consequence of other effects, especially considering the diverse role of Ran in various cellular processes including protein trafficking and cytoskeletal organization (Sazer and Dasso, 2000). Further characterization for Ran effectors in T. brucei will be crucial to address this question.
In mammalian cells, several components of the nuclear pore complex have been localized to the transition zone region at the cilium base (Kee et al., 2012). This, together with the discovery that Ran/importin have a function in cilium biogenesis (Fan and Margolis, 2011), suggest the presence of a highly conserved nuclear-pore-like mechanism that operates at both cilium and nucleus to selectively target proteins into the ciliary or nuclear compartments. In T. brucei, in addition to LRRP1, several nuclear-pore components have been found in previously published bi-lobe proteomes (Gheiratmand et al., 2013; Morriswood et al., 2012; Zhou et al., 2010). Whether the bi-lobe contains or functions as a selective gate, regulating protein trafficking to the FAZ or the flagellum by using a mechanism that involves Ran, will be of great interest.
MATERIALS AND METHODS
Cells and vectors
All experiments were performed on procyclic YTat1.1 (T. brucei rhodesiense) (Ruben et al., 1983) or 29.13 cells (T. brucei brucei) that were genetically engineered to allow tetracycline-inducible expression and RNA interference (RNAi) (Wirtz et al., 1999). Cell concentration was kept between 106 and 107 cells/ml by dilution with fresh medium and growth curves were presented as cumulative cell density.
For LRRP1-truncation mutants, the LRR region (nucleotide 1–1074) and the coiled-coil region (CC, nucleotide 1441–1594) were amplified and fused to the N-terminus of YFP in the pXS2 vector (Bangs et al., 1993). The full-length coding sequences of RanBP1 (Tb927.11.3380) and Ran (Tb927.3.1120) were fused to the C terminus of YFP in pXS2 vector. RNAi targeting LRRP1 has been reported previously (Zhou et al., 2010). For RNAi targeting Ran or RanBPL, a 457 bp fragment (nucleotide 142–598) or a 593 bp fragment (nucleotide 347–932), respectively, was cloned into the p2T7 plasmid (Wickstead et al., 2002). Stable cell transfection was achieved by electroporation with 15 µg of linearized plasmid followed with serial dilution and antibiotic selection for clonal lines. cDNA encoding the full-length RanBP1 (Tb927.11.3380), RanBPL (Tb927.10.8650), the C-terminal RanBD (nucleotide 594–1014) or the full-length Ran (Tb927.3.1120) were cloned into to the pET-28b(+) vector (Novagen) to generate His–RanBP1, His–RanBPL, His–RanBPL/C and His–Ran fusions. GST–Ran, GST–Rab1 and GST–LRR were generated by cloning the full-length Ran cDNA (Tb927.3.1120), Rab1 cDNA (Tb927.8.4610) or LRR cDNA (Tb927.11.8950, nucleotide 1–1074) into the pGEX-6p-1 vector (GE Healthcare).
Procyclic T. brucei (YTat1.1) cDNA library was generated using a MatchmakerTM Library Construction kit (Clontech). To construct the bait plasmid either the LRR domain (1–349 aa) or the coiled-coil domain (480–530 aa) of LRRP1 was cloned into the multi-cloning site of the pGBKT7 vector (Clontech), by using EcoRI and BamHI, which allowed in-frame fusion of the LRRP1 fragments to the Gal4-DNA-binding domain in the vector. The bait constructs were then tested for transcriptional auto-activity and toxicity.
1 ml aliquot of the T. brucei cDNA library in AH109 and 5 ml (≧109 cells/ml) of Y187 transformed with bait construct were combined in a sterile 2 l flask containing 40 ml 2×YPDA and 50 µg/ml kanamycin. The contents were then incubated at 30°C for 20–24 h with gentle swirling (30–50 rpm to allow mating). After mating, the mixture was centrifuged at 1000 g for 10 min. Cell pellet was resuspended in 10 ml 0.5 XYPDA/Kan (50 µg/ml). After 5 days of incubation at 30°C, diploids that express interacting proteins appeared on the plates. To eliminate the most common group of false-positive colonies, Ade+/His+ colonies were streaked out on fresh SD/−Ade/−His/−Leu/−Trp/XαGal master plates and incubated for 4–5 days at 30°C in order to test Ade+/His+ colonies for the third reporter; MEL1, which encodes yeast α-galactosidase. Positive blue colonies were further analyzed by PCR.
Identification of cDNA inserts from positive colonies
To identify the genes (and thus proteins) responsible for the positive two-hybrid interactions, the cDNA inserts were rescued by PCR colony-screening using the Matchmaker AD LD-insert screening amplimer set (Clontech). PCR products were then analyzed by gel electrophoresis and excised from the gel for DNA purification and sequencing. The DNA sequences were blasted against the T. brucei genome database (tritrypdb.org).
Cells were washed and resuspended in phosphate-buffered saline (PBS pH 7.4), attached to coverslips for 20 min and either fixed and permeabilized 5 min in cold methanol (−20°C) (anti-MORN, L3B2 and anti-PFR2) or 20 min in 4% PFA followed by 5 min in 0.25% Triton X-100 in PBS. Fixed cells were then blocked with 3% BSA in PBS and probed for 1 h with the corresponding antibodies. Monoclonal anti-LRRP1 (1:1000) and anti-MORN1 (1:500) were used to mark the bi-lobe (Morriswood et al., 2009; Zhou et al., 2010). A monoclonal anti-Ran (1:100) antibody (Cell Biolabs) was used to label T. brucei Ran. Rabbit anti-La antibody (1:1000) recognized the nuclear La protein (Marchetti et al., 2000). FAZ was labeled with the monoclonal L3B2 antibody (1:25) which is directed against FAZ1 (Vaughan, 2010). The flagellum was labeled with a rat antibody against PFR2 (1:5000). Nuclear and kinetoplast DNAs were stained with 10 µg/ml DAPI for 20 min. All images were acquired using a Zeiss Axio Observer microscope equipped with a CoolSNAP HQ2 camera (Photometrics). Images were processed with ImageJ or Adobe Photoshop. Quantification of fluorescence intensity and flagellum/FAZ length measurements were performed by using ImageJ.
Antibody against RanBPL
To obtain anti-RanBPL antibody, His–RanBPL was expressed in E. coli and affinity purified using Ni2+-NTA (Qiagen) following manufacturer's instructions. Purified His–RanBPL was used for the production of rabbit polyclonal antibody (Abnova, Taiwan). Affinity-purified anti-RanBPL was used at 1:200 for immunofluorescence assays and 1:1000 for immunoblots.
Immobilization of His–RanBP1, His–RanBPL or His–RanBPL/C on Ni2+-NTA beads
10 ml of BL21 bacterial culture was induced for 2 h with 0.1 mM IPTG. Pelleted bacterial cells were resuspended in 1 ml PBS supplemented with protease inhibitor cocktail, and homogenized by sonication. After solubilization with 1% Triton-X-100 at 4°C for 1 h, lysates were cleared by centrifugation. The supernatant was incubated with 50 µl Ni2+-NTA beads for 2 h at 4°C. Unbound proteins were removed by washes with PBS containing 20 mM imidazole. Similar procedure was used to immobilize GST–LRR, GST–Ran, GST–Rab1, or GST-only on glutathione Sepharose 4B beads (GE Healthcare).
Preparation of cell lysates and pull down
Typically, 10 ml bacterial culture or 1×108 T. brucei cells were used for pull-down assays. The cells were resuspended in 1 ml PBS, lyzed by sonication and extracted with 1% Triton-X-100 for 1 h at 4°C. Beads coated with recombinant proteins were then incubated with centrifugation-cleared cell lysates for 1 h at 4°C to allow binding. Unbound proteins were removed in three washes with PBS and proteins bound to the beads were eluted by boiling 5 min in SDS loading buffer. When GTPγS and GDP were used as controls, bacterial lysates or cell lysates expressing GST–Ran or endogenous Ran were preloaded with 1 mM GDP or 0.1 mM non-hydrolysable GTPγS for 30 min at 30°C, prior to the binding with the His–RanBPL beads.
10 µl anti-RanBPL unpurified serum, 8 µg anti-Ran, or 10 µl anti-LRRP1 was immobilized on 50 µg of protein G magnetic Dynabeads (Novex®). Approximately 1×108 parasite cells were lyzed as described previously. When required, cleared lysates were pre-loaded with 1 mM GTPγS for 30 min at 30°C, before incubation with the beads.
For immunoblots, proteins samples were separated on SDS-PAGE and probed with the appropriate antibodies: anti-Ran (1:1000), anti-LRRP (1:2000), anti-RanBPL (1:1000), anti-GFP (1:1000), anti-La (1:1000), anti-His (1:5000) (GE Healthcare), anti-TUBα (1:5000) (Santa Cruz Biotechnology) and anti-GST (1:5000) (Santa Cruz Biotechnology).
RanGTP hydrolysis in cell lysates was evaluated by the amount of Ran–GTP pulled down with His–RanBPL, taking advantage of its discriminative binding to GTP-bound Ran. To prepare parasite lysates for each assay, 1×108 of T. brucei cells were washed twice with ice-cold PBS and resuspended in 1 ml ice-cold lysis buffer (25 mM HEPES pH 7.5, 150 mM NaCl, 1% NP-40, 10 mM MgCl2, 1 mM EDTA, 2% glycerol). Cells were lyzed by sonication and cell lysates were then centrifuged at 14,000 g for 10 min at 4°C to remove cell debris. The volume of each sample was adjusted to 1 ml with lysis buffer.
To pull down Ran–GTP in the cell lysates, His–RanBPL-conjugated to Ni2+-NTA beads were incubated with cell lysates prepared as described above. The tubes were then incubated at 4°C for 1 h with gentle mixing. The beads were washed thoroughly with lysis buffer before resuspended in reducing SDS-PAGE sample buffer. Proteins bound to the beads were eluted, denatured by boiling at 100°C for 5 min, fractionated by SDS-PAGE and immunoblotted with appropriate antibodies. Imaging and quantification of the immunoblots were performed using an ImageQuant LAS 4000 mini (GE Healthcare). All Ran–GTP measurements were normalized against the intensity of Ran–GTP pulled down from the control cells performed at the same time of each experiment.
mRNAs from T. brucei cells before and after induction of RNAi targeting RanBPL were extracted with Trizol (Invitrogen) and used as template for reverse transcription with RT Superscript II (Invitrogen). The cDNAs were then amplified by PCR with specific primers for the RanBPL and α-tubulin mRNAs.
We thank the Protein Expression Facility at the Singapore Mechanobiology Institute for technical assistance in cloning and expression of some of the recombinant proteins described in this study.
↵* Present address: Department of Microbiology and Molecular Genetics, University of Texas Medical School, Houston, TX 77030, USA.
↵‡ These authors contributed equally to this work
The authors declare no competing interests.
A.B., S.B., Z.Y., C.X.L. and Z.Q. did the experiments, analyzed the data and wrote the paper. L.B.C. and C.Y.H. conceived the project, analyzed the data and wrote the paper.
This project is funded by Singapore National Research Foundation Fellowship [grant number NRF-RF001-121].
Supplementary material available online at http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.148015/-/DC1
- Received December 13, 2013.
- Accepted August 14, 2014.
- © 2014. Published by The Company of Biologists Ltd