Rac Regulates Giardia lamblia Encystation by Coordinating Cyst Wall Protein Trafficking and Secretion

ABSTRACT Encystation of the common intestinal parasite Giardia lamblia involves the production, trafficking, and secretion of cyst wall material (CWM). However, the molecular mechanism responsible for the regulation of these sequential processes remains elusive. Here, we examined the role of GlRac, Giardia’s sole Rho family GTPase, in the regulation of endomembrane organization and cyst wall protein (CWP) trafficking. Localization studies indicated that GlRac is associated with the endoplasmic reticulum (ER) and the Golgi apparatus-like encystation-specific vesicles (ESVs). Constitutive GlRac signaling increased levels of the ER marker PDI2, induced ER swelling, reduced overall CWP1 production, and promoted the early maturation of ESVs. Quantitative analysis of cells expressing constitutively active hemagglutinin (HA)-tagged GlRac (HA-RacCA) revealed fewer but larger ESVs than control cells. Consistent with the phenotype of premature maturation of ESVs in HA-RacCA-expressing cells, constitutive GlRac signaling resulted in increased CWP1 secretion and, conversely, morpholino depletion of GlRac blocked CWP1 secretion. Wild-type cells unexpectedly secreted large quantities of CWP1 into the medium, and free CWP1 was used cooperatively during cyst formation. These results, in part, could account for the previously reported observation that G. lamblia encysts more efficiently at high cell densities. These studies of GlRac show that it regulates encystation at several levels, and our findings support its coordinating role as a regulator of CWP trafficking and secretion. The central role of GlRac in regulating membrane trafficking and the cytoskeleton, both of which are essential to Giardia parasitism, further suggests its potential as a novel target for drug development to treat giardiasis.

mic reticulum (ER) exit sites (ERES) (10), a process dependent on COPII and the small GTPase Sar1 (16). After approximately 8 h, most of the CWM is contained within ESVs. At this point, ESVs are no longer associated with COPII or Sar1 but are associated with Arf1, a GTPase whose activity as a molecular switch leads to the recruitment of the coat proteins COPI and clathrin (16). Arf1 activity is required for subsequent ESV maturation and CWP trafficking out of the cell (13,16). ESVs are sites of CWP processing, which includes several posttranslational modifications (17)(18)(19). As ESVs mature they grow in size, and approximately 12 h into the encystation process, a major processing event leads to the formation of a condensed core with an outer fluid phase. Core condensation is catalyzed by cleavage of pro-CWP2 into N-and C-terminal fragments by a cysteine protease (20,21). The fragments are then partitioned so that the fluid phase is composed of the N-terminal fragment of CWP2 in addition to CWP1, whereas the core contains CWP3 and the C-terminal part of CWP2. At the same time, a GalNAc carbohydrate homopolymer is produced and trafficked by distinct carbohydratepositive vesicles that are deposited at the surface of encysting cells ahead of cyst wall protein secretion (22,23). The curled fibrils of the GalNAc carbohydrate serve as a scaffold for direct binding of CWP1 via a lectin domain (22). The cyst wall is thought to be formed by a rapid secretion event that releases CWP1 and the N-terminal part of CWP2, likely within minutes, to form the first layer of the cyst wall; this is followed by slower deposition of the condensed core cargo, CWP3, and the C-terminal part of CWP2 (9). While considerable progress has been made in describing these sequential events of encystation, the signaling events which trigger secretion remain elusive. Since CWM secretion is necessary for the formation of environment-resistant infectious cysts, uncovering the mechanisms that regulate and temporally coordinate secretion of CWM could potentially identify potential drug targets of Giardia and also uncover conserved principles of protozoan encystation.
Rho GTPases are potential candidates for regulating CWP secretion, as they have important roles in coordinating vesicle trafficking and the cytoskeleton in plants and animals (24)(25)(26)(27)(28). Rho family GTPases have undergone extensive gene duplication and functional diversification in most eukaryotic lineages (copy number in humans, 22; in Arabidopsis thaliana, 11; in Saccharomyces cerevisiae, 5). However, the Giardia genome contains just a single Rho family GTPase, GlRac, and the entire signaling system appears to be minimalistic compared to that of mammals (see Table S1 in the supplemental material) (29)(30)(31). Interestingly, Rac has been reported to be the evolutionary founding member of the Rho family GTPases (32). Therefore, studies of G. lamblia, which is itself placed in a critical, albeit controversial, deep-branching position on the tree of life (33)(34)(35), may provide insight into the ancestral function of Rho family GTPases.
Here, we set out to determine if GlRac has a role in regulating membrane trafficking in Giardia. We report that GlRac is associated with the ER in trophozoites and with both the ER and the Golgi complex-like ESVs during encystation. GlRac is crucial for production of CWP1, maturation of ESVs, and most importantly, for secretion of CWP1, which is used in a cooperative manner to form viable cysts. These roles in Giardia indicate a conserved and ancient role for Rac homologs in membrane trafficking.

GlRac associates with the ER and encystation-specific vesicles.
We previously observed that expression of a constitutively active GlRac mutant (tetracycline/doxycycline-inducible Q47L HA-Rac; HA-Rac CA , equivalent to Q61L Rac1 [see Fig. S1 in the supplemental material]), alters actin organization and, sometimes, results in formation of large vesicular structures in nonencysting trophozoites (31). The latter result suggested a possible role for GlRac in endomembrane organization. To further examine this possibility, we determined GlRac localization by endogenously tagging the protein at the N terminus with a triple hemagglutinin (HA) tag (HA-Rac). Localization of HA-Rac in trophozoites, the proliferative stage that colonizes the host intestine, revealed a pattern similar to that reported for the ER (36). We therefore examined its location relative to protein disulfide isomerase 2 (PDI2), an ER lumenal enzyme that catalyzes disulfide bond formation and protein folding (36). We observed considerable overlap between the two signals ( Fig. 1A), indicating that a portion of GlRac is ER associated. To determine whether GlRac localization might be altered during the stage conversion to cysts, we induced encystation by exchanging standard medium for encystation medium and then examined the localization of HA-Rac 12 h into the encystation process. We found that HA-Rac was associated with the perimeter of CWP1-positive vesicles known as ESVs (Fig. 1B). The localization of GlRac to the ER and ESVs could indicate a role for Rac in regulating protein trafficking in Giardia.
Constitutive GlRac signaling impairs ER function, CWP1 production, and cyst viability. To assess whether GlRac activity affects the ER, we tested the impact of inducing HA-Rac CA expression on PDI2 localization 24 h after the addition of doxycycline ( Fig. 2A). We observed a striking increase in PDI2 staining. Quantification verified a 4-fold increase in PDI2 levels as a result of induced HA-Rac CA expression above that of the uninduced control (Fig. 2B). To verify a functional role for GlRac in encystation, we assessed the impact of constitutive GlRac signaling on CWP1 levels and the production and viability of cysts. We chose to follow  CWP1 because it is the best-characterized cyst wall component  and the only marker for which a commercial antibody is available. HA-Rac CA expression led to over a 70% reduction in CWP1 levels 24 h into the encystation process, compared with its levels in uninduced control cells, as measured by quantitative immunoblotting ( Fig. 2C and D). Consequently, the number of cysts formed by HA-Rac CA -expressing cells compared to the uninduced control was reduced by 49%. To compare the quality of the cyst walls formed by these two populations, we assessed the ability of the formed cysts to exclude the vital dye trypan blue after treatment with distilled water. Cysts with defective walls lyse in distilled water and stain blue. We observed a significant increase in trypan blue-positive cysts in the HA-Rac CA -expressing population (Fig. 2E).
Since misregulation of GlRac signaling impacted the ER, CWP1 levels, and cyst production, we questioned whether GlRac levels change during the encystation process. SAGE data in Giar-diaDB reported that GlRac levels (GL50803_8496) decreased upon the initiation of encystation and then increased above the levels of nonencysting trophozoites 7 h into encystation (37). We used quantitative reverse transcription-PCR (RT-PCR) to analyze the time course of GlRac expression in wild-type (WT) cells at 0, 2, 8, 16, and 20 h post-induction of encystation (p.i.e.) relative to expression of the housekeeping gene GAPDH (38). In contrast to the SAGE and recently published transcriptome sequencing (RNA-Seq) data (8, 37), we did not observe reduced GlRac expression upon initiation of encystation, but we did observe that GlRac levels increased significantly at 8 h p.i.e. (Fig. 2F) (P Ͻ 0.05). Together, our studies indicate that GlRac is transcriptionally upregulated from~7 to 12 h p.i.e.
While GlRac localizes to the ER, it may be mostly inactive there. Rho GTPases act as molecular switches by changing conformation based upon their nucleotide binding state. The active GTP-bound state functions by recruiting effectors to carry out specific activities, and it then becomes inactive after GTP hydrolysis and disassociation with effector proteins. Rho GTPases are activated by guanine nucleotide exchange factors (GEFs) through exchange of GDP for GTP (39). Therefore, the ER could be a source of sequestered GlRac positioned to associate with budding vesicles. In support of the notion that GlRac is largely inactive at the ER, constitutively active GlRac increased PDI2 levels, indicating ER stress (40). To further examine the impact of constitutive GlRac signaling on organization of the ER and other endomembranes, we analyzed the impact of HA-Rac CA at the level of electron microscopy. In transmission electron micrographs (TEMs) of WT cells, ESVs are vesicles with a uniform electron density, and the ER appears as thin tubules ( Fig. 2G; see also Fig. S2 in the supplemental material). In doxycycline-induced HA-Rac CA encysting cells, we observed marked changes in endomembrane organization that likely correlated with the extent of HA-Rac CA expression. In severely perturbed cells, we observed atypical ER morphology, with extensive swelling potentially indicating an ER exit defect. Our TEM analysis, in the absence of any markers, did not allow us to determine which if any of the vesicular structures observed in the most severely disrupted cells were ESVs. In any case, we observed electron-dense staining of vesicular content that we have not observed in wild-type controls (see Fig. S2 in the supplemental material). We interpret the electron-dense staining within vesicles to indicate defective CWM processing, and this may account for the reduction of CWP1 in HA-Rac CA -expressing i.e. were set to 100%, and the relative amount of CWP1 was calculated based on normalization to the tubulin loading control. Statistical significance was evaluated by using the t test. ****, P Ͻ 0.0001. (E) Cyst production was quantified for induced HA-Rac CA and control cells. Additionally, the cells were stained with the vital dye trypan blue to assay cyst wall integrity. Cysts production and viability data were acquired from three independent experiments, each with Ն200 cysts. Statistical significance was evaluated by using the t test. **, P Ͻ 0.01. ( cells. Taken together, we conclude that GlRac has a critical regulatory role in encystation, based on four observations: (i) GlRac associates with the ER and ESVs; (ii) GlRac is temporally upregulated during encystation; (iii) constitutive GlRac signaling causes ER swelling and reduces cellular CWP1 levels; and (iv) constitutive GlRac signaling reduces encystation rates as well as disrupts cyst wall integrity. GlRac promotes ESV maturation. To better understand the temporal role of GlRac in encystation, we encysted doxycyclineinduced HA-Rac CA and the endogenously tagged HA-Rac cell lines. Cells were collected every 2 h from 2 to 24 h p.i.e. and stained for CWP1 and HA (see Fig. S3 in the supplemental material). Consistent with the results described above, in which GlRac expression levels were shown to increase by 8 h p.i.e., we observed a striking difference between HA-Rac and HA-Rac CA cells at 8 h p.i.e. ESVs in HA-Rac CA -expressing cells were fewer, larger in volume, and more often displayed a ring-like CWP1 distribution than did the HA-Rac control ( Fig. 3A; see also Fig. S3). The ringlike CWP1 distribution is characteristic of mature ESVs that have undergone processing and condensation and is not typically observed until 12 to 16 h p.i.e. (9). We quantified the number and volume of ESVs in the cells via three-dimensional (3D) image segmentation. The number of ESVs formed per cell was reduced by 73.6% in HA-Rac CA -expressing cells compared to HA-Rac cells ( Fig. 3B and C). Additionally, the average volume of individual ESVs was 50% larger in HA-Rac CA -expressing cells than in HA-Rac cells (Fig. 3B and D). Due to leaky expression (see Fig. S4A in the supplemental material), uninduced HA-Rac CA cells displayed an intermediate phenotype (see Fig. S4B and C), which is why we used HA-Rac cells as a negative control here. Nevertheless, the doxycycline-induced HA-Rac CA cells had a significant increase in volume over uninduced HA-Rac CA cells. These findings indicate that GlRac signaling can promote increased ESV size and precocious maturation, a result that corresponds well with its transcriptional increase at 7 to 12 h p.i.e. in wild-type Giardia.
Rac drives secretion of CWP1, which can be utilized by the entire population of cysts. Cyst wall formation is thought to be temporally regulated, because CWM is held in the Golgi complexlike ESVs for processing and sorting before its concerted secretion to form the cyst wall (9). Our observations of precocious ESV maturation and reduced intracellular CWP1 may be consistent with a switch from regulated to constitutive secretion of CWM. Therefore, we expected that if HA-Rac CA caused secretion in a temporally uncontrolled manner, we would detect free CWP1 in the medium, as secreting cells would not yet be competent to bind CWP1. This was indeed found to be the case. Cells expressing HA-Rac CA displayed an increased ratio of CWP1 exported into the medium compared to the CWP1 remaining in the cells (Fig. 4A and B). While we had expected to find free CWP1 in the medium of HA-Rac CA -expressing cells, we were surprised to find such high levels of CWP1 in the medium of the control. This prompted us to ask whether the CWP1 detected in the medium was functional (i.e., could be incorporated into the cyst wall of any cell in the population), a possibility that would implicate encystation as a cooperative process. Furthermore, cooperative production of CWP1 might in part account for previous observations that encystation rates improve with greater parasite densities (5). To test this, we conducted a medium swap experiment. Beginning with confluent cultures (~1 ϫ 10 6 cells/ml), CWP1-mCherryexpressing cells or WT cells were incubated in encystation me-dium for 24 h. The medium in which CWP1-mCherry cells were encysted (CWP1-mCherry medium) was cleared of cells by centrifugation and filtration (0.22 m) before being used to replace the medium of encysting WT cells. Following incubation for a further 24 h, we found that 100% (50/50) of the WT cysts were mCherry positive and, as expected, 0/50 WT cells from the plain medium condition had detectable red fluorescence under identical image acquisition and scaling (Fig. 4C). Consistent with the idea that only cells at a specific stage (i.e., the cells at later stages of encystation, displaying GalNAc homopolymers) are competent to bind CWP1 (22), we only found mCherry-CWP1 on rounded-up late-stage encysting cells or cysts and never on trophozoite-shaped cells (Fig. 4C). These results indicate that secreted CWP1 can be utilized at the population level and help to start explaining the improved encystation rates associated with greater parasite density (5).
To further validate the role of GlRac in secretion, we sought to inhibit GlRac signaling through loss-of-function. The ability to knock out genes in Giardia with genome editing has yet to be accomplished, and RNA interference is ineffective. Therefore, we sought to deplete GlRac by using translation-blocking morpholinos (41). In order to monitor GlRac protein levels without a custom antibody, we developed an endogenously tagged morpholino-sensitive HA-GlRac construct (HA-Rac MS ) by including the first 27 bp of the Rac coding region in front of the HA tag (see Fig. S5 in the supplemental material). Without this short sequence, the 5= HA tag would render the tagged copy of GlRac morpholino resistant (42). HA-Rac MS localization was indistinguishable from that of the previously used endogenously tagged HA-Rac (see Fig. S5D). Western blot analysis confirmed that GlRac expression was reduced by 72% at 24 h after morpholino electroporation (see Fig. S5). Depletion of GlRac in trophozoites led to defects in morphology, polarity, and cytokinesis (see Fig. S5D). In agreement with the notion that GlRac activity is not required for the initial stages of encystation, GlRac depletion did not cause an appreciable change in cellular CWP1 content ( Fig. 4D and E); it did, however, impair CWP1 secretion, as indicated by the significantly decreased amount of CWP1 detected in the medium during the later stages of encystation ( Taken together, our findings demonstrate two major functions for GlRac in encystation. GlRac signaling promotes an increase in ESV size and maturation and is required for CWP1 secretion.

DISCUSSION
We set out to determine whether GlRac plays a critical role in regulating membrane trafficking and, more specifically, if it has a role in regulating the trafficking of CWM. The initial production of CWM is dependent on the ER (10). Enrichment of GlRac at the ER in nonencysting and encysting trophozoites (Fig. 1) is consistent with a role in regulating ER structure and/or function. CWP1 release from the ER depends on functional ERES and is a necessary first step in ESV biogenesis (10). The accumulation of PDI2 and the swollen ER phenotype associated with HA-Rac CA expression suggest an ER exit defect ( Fig. 2B and G; see also Fig. S2 in the supplemental material). Whether GlRac has a regulatory role for ERES remains to be elucidated. Interestingly, however, upregulation of GlRac expression coincides with the completion of CWP loading into nascent ESVs around 8 h p.i.e. (16). There is also precedence for Rho GTPase regulation of ERES. SPIKE1, an Arabidopsis guanine nucleotide exchange factor (GEF), localizes to ERES, where it activates the Rac homolog Rho of plants 2 (At-ROP2) (29,43). Expression of AtROP2 dominant-negative or constitutively active mutants resulted in upregulation of ER stress response genes (43), which is similar to our observation that constitutive GlRac signaling increased PDI2 levels, a known ER stress response (40). We speculate that constitutive GlRac signaling caused ER trafficking defects that resulted in reduced CWP1 levels as a consequence of overall reduced capacity to produce and sort CWP1. Precedent also exists for Rho family GTPases controlling secretory traffic through coordination of the cytoskeleton and membrane trafficking systems (25). We previously demonstrated that GlRac has a role in regulating the actin cytoskeleton in Giardia (31). Furthermore, actin was shown to localize to late-stage ESVs (31). Since GlRac is enriched around ESVs (Fig. 1B), it is tempting to speculate that GlRac works to coordinate the cytoskeleton and membrane trafficking systems during ESV maturation and secretion. For example, GlRac could trigger actin-based translocation of smaller ESVs in order to bring them into proximity for fusion. A general role in coordinating the cytoskeleton and membrane trafficking systems would be in line with Rho GTPase functions in other eukaryotes. Considering that Rac is the evolutionary founding member of the Rho family GTPases (32) and that Giardia belongs to a potentially early-branching group of eukaryotes (34,44), our results suggest that the ancestral Rac homolog was capable of regulating multiple processes and that the functional diversification of the expanded Rho family GTPases is a modern adaptation.
Model of GlRac regulation of encystation. Giardia's parasitic life cycle is dependent on stage conversion from trophozoites to infectious water-resistant cysts, which preserves their viability in the outer environment and passage through stomach acid. Encystation features sequential vesicle trafficking events and pulsed secretion of CWM to the surface of the newly formed cyst (9). Our results and the current understanding of the process are summarized in Fig. 5, an updated model for encystation (9,45). Newly produced CWPs destined to form the impermeable cyst wall are produced in the ER and sorted into nascent ESVs (15,46). ESV neogenesis is dependent on functional ERES (10), which use the small GTPase Sar1 and COPII to form export vesicles (13,16). Our data suggest that Rac signaling is inhibitory at this stage of trafficking ( Fig. 2; see also Fig. S3 in the supplemental material). During early encystation (around 8 to 12 h p.i.e.), we observe the transition from many small ESVs to fewer larger ESVs; this process is promoted by GlRac signaling (Fig. 3). Coincidentally, Arf1 begins to associate with ESVs, and its activity is required for ESV maturation and ultimately CWP secretion (13,16), around the this study, we propose a model of encystation where GlRac coordinates the encystation process. Newly synthesized CWM is produced in the ER and exported from the ER into nascent ESVs (15,46). GlRac signaling disrupts export from the ER, a process dependent on active ERES and the small GTPase Sar1 (10,16). Active GlRac drives maturation of ESVs (black arrows). Mature ESVs are composed of the condensed core (con), containing CWP3 and the C-terminal part of processed CWP2, and the fluid outer phase (fl), composed of CWP1 and the N-terminal part of processed CWP2 (9). Maturation of ESVs is a consequence of CWP2 processing by a cysteine protease (20). Doublephase ESVs decondense to be sorted and sequentially secreted. Ahead of CWP secretion, GalNAc homopolymer is exported to the cyst surface by distinct carbohydrate-positive vesicles (ECVs) (22,23). CWP1 secretion is regulated by GlRac signaling (black arrows) and is crucial to form the first protein layer of the impermeable cyst wall. CWP1 binds the GalNAc carbohydrate fibrils present at the surface of the encysting cell (22) but can also diffuse away and be bound by other "primed" encysting cells (black arrows). By affecting CWP1 levels in the cell, temporal regulation of ESVs maturation, and CWP1 secretion, GlRac signaling coordinates the sequential CWP1 production and trafficking events necessary for the production of viable cysts. same time GlRac transcript levels rise. It is tempting to speculate that these events are related, as cross talk between Arf and Rho GTPases has been observed for Arf and Rho GTPase homologs in plants and animals (reviewed in references 47 and 48). CWM processing occurs during the transition to larger ESVs and is apparent as an outer fluid phase and an inner condensed phase. The outer fluid phase is composed of CWP1 and the N-terminal processed portion of CWP2, while the condensed core is composed of CWP3 and C-terminal part of CWP2 (9). The mechanism by which GlRac signaling promotes an increase in ESV volume and concurrent CWM processing remains to be elucidated. We speculate that the mechanism will involve fusion of early ESVs and vesicles containing the protease responsible for the proteolytic cleavage of CWP2 (20,21). During the later stages of encystation, in a second sorting event, the fluid and condensed phases of mature ESVs are separated and secreted in sequential steps to form a doublelayered cyst wall (9). First, the fluid phase, composed of CWP1 and the N-terminal region of CWP2, is secreted in a GlRacdependent manner (Ն16 h p.i.e.). CWP1 has been proposed to contain a lectin domain that interacts with GalNAc homopolymer fibrils present at the surface of the forming cyst (22). Our results indicate that this is a cooperative process where CWP1 cannot only bind the cell it is secreted from but also contribute toward cyst wall formation of neighboring encysting cells (Fig. 5). In the final stages of encystation (Ն20 h p.i.e.), the cyst wall is "sealed" with CWP3 and the C-terminal part of CWP2; these proteins are exported from the subpopulation of ESVs that inherited condensed cores (9). Overall, Giardia's sole Rho family GTPase appears to play a central role in coordinating maturation and secretion of CWP1. We hypothesize that GlRac's roles in ESV maturation (Fig. 3), formation of aberrant vesicles (Fig. 4) (31), and CWP secretion (Fig. 5) could all be explained by a role for GlRac in promoting membrane fusion. The specific effectors activated by GlRac signaling remain to be identified.

Conclusion.
Our data indicate that the sole Rho family GTPase in Giardia has a critical role in regulating the encystation process. GlRac is key to temporal coordination of CWP1 production, ESV maturation, and CWP1 secretion, all of which are required for cyst formation and viability. The relative simplicity of Giardia's membrane trafficking and Rho GTPase signaling system, whether evolutionarily ancient or derived, may be a powerful tool for uncovering core principles of cellular signaling, membrane trafficking, and the encystation process. In addition, Rho GTPases are druggable targets (49,50), so the modulation of GlRac activity may be a potent tool to suppress Giardia infection and transmission. Indeed, rational drug design has been successfully used to target human Rac1 while sparing the related Rho family GTPases RhoA and Cdc42 (49)(50)(51). Since human Rac1 is more similar to human Cdc42 (70% identity) than to GlRac (50% identity), this divergence should allow for Giardia-specific inhibitors to be developed, and thus we suggest GlRac as a promising new candidate for drug development.
Encystation and viability assay. To allow observation of different encystation stages, cells were synchronized by using the two-step encystation protocol (53). Confluent cultures of~1 ϫ 10 6 cells/ml were cultivated for 24 h in preencystation medium (without bile). Encystation was induced as described previously (54) in medium with a pH of 7.8 and supplemented with 10 g/liter bovine/ovine bile. Trophozoites were encysted 12 to 24 h post-morpholino treatment or induction of Rac CA expression. Two days after water treatment, cysts were stained with trypan blue and counted by using a hemocytometer.
Vector construction. Construction of the tetracycline-inducible WT and Q74L HA-GlRac (HA-Rac CA ) vectors is described in reference 31. N-terminally HA-tagged GlRac (GL50803_8496) and morpholinosensitive GlRac under endogenous promoters (HA-Rac MS ) were constructed as follows. The HA-Rac fragment was cleaved with NcoI and EcoRI from a tetHA-Rac.pac vector and ligated into pKS-HA.pac enriched with an NcoI site. The NcoI restriction site was introduced into the pKS-HA.pac vector by using adapter oligonucleotides NcoI F= (GATCC CCATGGG) and NcoI R= (AATTCCCATGGG). To create an Nterminally tagged endogenously expressed HA-Rac that would remain sensitive to morpholinos, we introduced the first 27 bp of the Rac coding sequence before the HA tag (see Fig. S5A in the supplemental material). Native promoter-only or native promoter plus the first 27 bp (morpholino binding site for Rac MS ) of the Rac gene flanking XbaI and NcoI sites was PCR amplified from Giardia genomic DNA using the primers XbaRacprom_F (5= AAATCTAGAGTGCCGAGGCGGGATTGCTC 3=) and RacpromNcoI_R (5= AAACCATGGTTTTTAATTTTTGTAACATG 3=) for GlRac and XbaRacprom_F (5= AAATCTAGAGTGCCGAGGCGG GATTGCTC 3=) and NcoRacMORF_R= (5= ATTCCATGGCTGTATCCT CATTTCCTGTAC 3=) for Rac MS , respectively. The HA-Rac MS plasmid was linearized for homologous recombination into the Giardia genome by using PstI.
To construct CWP1-mCherry, pKS-3HA.neo was digested with BamHI and EcoRI to remove the 3ϫ HA tag. CWP1, including endogenous promoter sequence, was PCR amplified from Giardia genomic DNA by using the primer CWP1prom_F (5= AAAGGATCCAAGCTTCT AGCCACGCATGGGCTG 3=) with flanking BamHI site and the CWP1_R (5= CCGACCGGTccCCAAGGCGGGGTGAGGCAG 3=) primer with flanking AgeI site. mCherry was PCR amplified from a donor vector by using mCherry_F (5= GGGACCGGTTGGAGGCGGAGGGAG CGGCGGGGGCGGAAGCATGGTGAGCAAGGGCGAG 3=) primer flanking AgeI site and a spacer to ensure proper folding of the fusion protein and mCherry_R (5= AAAGAATTCTCACTTGTACAGCTC GTCC 3=) primer flanking the EcoRI site. The three fragments were ligated, and the sequence was verified before being transfected for episomal expression.
Transmission electron microscopy. For TEM, cells attached to carbon-coated sapphire disks were fixed with 2.5% glutaraldehyde in 0.1 M Na/K-phosphate (pH 7.4) for 1 h, washed with 0.1 M Na/Kphosphate, postfixed with 1% osmium tetroxide in 0.1 M Na/K-phosphate for 1 h, and dehydrated in a series of ethanol starting at 70%. After two changes in acetone, the processed cells were embedded in Epon at room temperature followed by polymerization at 60°C for 2.5 days. Sections of 60 to 80 nm in thickness were stained with uranyl acetate and lead citrate and analyzed in a transmission electron microscope (CM12; FEI, Eindhoven, the Netherlands) equipped with a charge-coupled-device camera (Ultrascan 1000; Gatan, Pleasanton, CA) at an acceleration voltage of 100 kV.