Disc and Actin Associated Protein 1 influences attachment in the intestinal parasite Giardia lamblia

The deep-branching eukaryote Giardia lamblia is an extracellular parasite that attaches to the host intestine via a microtubule-based structure called the ventral disc. Control of attachment is mediated in part by the movement of two regions of the ventral disc that either permit or exclude the passage of fluid under the disc. Several known disc-associated proteins (DAPs) contribute to disc structure and function, but no force-generating protein has been identified among them. We recently identified several Giardia actin (GlActin) interacting proteins at the ventral disc, which could potentially employ actin polymerization for force generation and disc conformational changes. One of these proteins, Disc and Actin Associated Protein 1 (DAAP1), is highly enriched at the two regions of the disc previously shown to be important for fluid flow during attachment. In this study, we investigate the role of both GlActin and DAAP1 in ventral disc morphology and function. We confirmed interaction between GlActin and DAAP1 through coimmunoprecipitation, and used immunofluorescence to localize both proteins throughout the cell cycle and during trophozoite attachment. Similar to other DAPs, the association of DAAP1 with the disc is stable, except during cell division when the disc disassembles. Depletion of GlActin by translation-blocking antisense morpholinos resulted in both impaired attachment and defects in the ventral disc, indicating that GlActin contributes to disc-mediated attachment. Depletion of DAAP1 through CRISPR interference resulted in intact discs but impaired attachment, gating, and flow under the disc. As attachment is essential for infection, elucidation of these and other molecular mediators is a promising area for development of new therapeutics against a ubiquitous parasite.

Introduction Giardia lamblia, a single-celled parasite, is the cause of over 280 million annual cases of the gastrointestinal disease giardiasis worldwide [1,2]. This extracellular parasite colonizes the lumen of the intestine, where its ability to maintain infection requires tight attachment to microvilli, in order to withstand the peristaltic flow of the gastrointestinal tract [3]. The primary means of attachment is via a specialized microtubule-based organelle known as the ventral adhesive disc [4][5][6]. A well-functioning disc is critical for cell survival and maintenance of infection [7], but the molecular mechanisms controlling conformational dynamics and function remain poorly understood.
Imaging of fluid flow at the disc margin identified two areas postulated to be important for tight attachment: one at the anterior and the other at the posterior of the disc [8][9][10] (Fig 1A). The former is the marginal region of the overlap zone, the area of the disc where the sheet of microtubules overlaps itself. The other is the ventral groove, a hump in the posterior region of the disc where the ventral flagella exit between the disc and cell body. Interruption of contact in these areas results in disruption of attachment and in fluid moving underneath the disc [9,11,12]. Cryo-electron tomography imaging of the disc revealed different protein densities in these areas, indicating specialized architecture within the disc itself, which may indicate that specific effector proteins control disc conformation and fluid flow [13]. In the current model, both hydrodynamic flow and suction contribute to the mechanism of attachment [9]. Attachment occurs in several distinct stages [8,9]. In early attachment, the anterior ventrolateral flange, lateral crest and ventral groove make contact with the surface while the ventral flagella beat to generate fluid flows. During this stage cells can either skim the surface or remain stationary. In the later stages of attachment, the cell body becomes more intimately associated with the surface so that the lateral shield and bare area make contact with the surface and the cells remain in place. Flagella-beating is required for initiation, but not maintenance, of attachment, indicating that fluid must be moved out from underneath the disc during the earlier stages of attachment [9,14].
These conformational changes in the disc which control attachment are not generated by microtubule turnover, as microtubule depolymerizing or stabilizing drugs do not affect attachment [15]. Though microtubules are the foundation of the disc, it is also composed of over 100 other proteins, known as disc-associated proteins (DAPs). At least five of these DAPs have been shown to be important for disc biogenesis and function [9,12]. Only three DAPs identified to date contain known microtubule-binding motifs; most are ankyrin-repeat proteins, NEK kinases, or the annexin-like alpha-Giardins [9,15]. At least 25, however, lack identifiable domains and are annotated in the Giardia genome as "hypothetical proteins" [16,17]. So far, no DAPs have been identified as motor or force-generating proteins. Thus, the proteins that mediate changes in disc conformation are unknown.
This observation led to multiple experiments attempting to test the role of GlActin in attachment using actin inhibitors. These studies yielded conflicting results because reagents that typically recognize actin or actin-binding proteins in other organisms were presumed to act similarly in Giardia [6]. Later, GlActin was determined to be one of the most divergent actins found among eukaryotes and Giardia's genome was found to lack genes encoding conventional actin-binding proteins that were reportedly localized to the disc [20][21][22]. Further research revealed that GlActin has non-conservative amino acid substitutions, predicted to disrupt actin inhibitor binding, and anti-GlActin antibodies revealed localization throughout the cell [21]. Thus, GlActin's potential role in disc-based attachment remains controversial.
We recently identified eight Giardia actin (GlActin) interactors that localize to the ventral disc [17], suggesting a means to resolve the role of actin and actin-binding proteins in ventral disc function. One such actin-interacting protein, GL50803_16844, which we denote Disc and Actin Associated Protein 1 (DAAP1), appears highly enriched in the two regions of the disc important for regulating fluid flow during attachment.
In this study, we use morpholinos and CRISPR interference (CRISPRi) to deplete GlActin and DAAP1 respectively and investigate the role of both proteins in the ventral disc. Our results show that GlActin plays a role in normal ventral disc morphology and both GlActin and DAAP1 are required for parasite attachment.

GlActin affects attachment and disc morphology
In vitro, Giardia normally grows attached to the inside of a culture tube. We used an antisense translation-blocking morpholino to deplete GlActin for 24 hours, [21] after which we assessed attachment. We compared the numbers of cells attached to the tube wall to cells unattached in the media for each condition. With an approximately 50% knockdown (S1 Fig), we found attachment to plastic culture tube walls to be significantly decreased, with an average of 23% of cells attached compared to 75% of cells treated with a control morpholino ( Fig 1B). Dead Giardia cells lose their ability to adhere to culture tubes. To account for the possibility that the majority of the unattached cells were dead, we stained the unattached cells with propidium iodide to mark dead cells, and fluorescein diacetate to mark live cells. While slightly more unattached GlActin knockdown cells were dead compared to that of the control population, this difference was not statistically significant ( Fig 1C), indicating that the defect in attachment in GlActin-depleted cells was not due to a preponderance of dead cells.
We asked if the role of GlActin in attachment could be attributed to an effect on the ventral disc. Attached and unattached cells were treated with either control or anti-GlActin morpholino, fixed after 24 h, and examined for distribution of a ventral disc marker using immunofluorescence microscopy. We used a HALO-tagged delta-giardin which is widely distributed in the disc to monitor whether the disc appeared normal or defective. Defects were then further classified as 1) mild, with misshapen/broken discs or small gaps in the disc or 2) severe, including unwound discs or severely misshapen discs (Figs 1D, 1E and S2). More GlActin knockdown cells had defective discs compared to those in the control; however, this difference was only significant in unattached cells (Fig 1D and 1E). While both attached and unattached GlActin-depleted cells had more disc defects than their counterparts treated with the control morpholino, the unattached GlActin-depleted cells had the most defective discs among all the categories. Consistent with GlActin depletion resulting in detachment, we found that GlActin levels were lower in unattached vs attached GlActin knockdown cells (S3 Fig). Notably, there were more unattached cells than cells with disc defects, suggesting that GlActin has a role in attachment beyond its contribution to disc morphology.

GlActin interacts with disc-associated protein DAAP1
We next asked if GL50803_16844, previously identified by mass spectrometry as a potential actin filament interactor at the ventral disc [17], could be involved in attachment. GL50803_16844 is conserved in all Giardia lamblia isolates and Giardia muris, but is absent from the genome of Spironucleus, a close relative of Giardia which lacks a ventral disc [23]. Hereafter, we will refer to this protein that lacks identifiable domains and homologues in other organisms as Disc and Actin Associated Protein 1 (DAAP1). We tagged DAAP1 with a C-terminal 3xHA tag and expressed it episomally in Giardia (S4 Fig). Immunoprecipitation of the episomally expressed DAAP1-3HA revealed it in complex with GlActin (Fig 2A). Episomal

Fig 2. DAAP1 is a GlActin interactor and a ventral disc-associated protein. A)
Immunoprecipitation from extracts of cells expressing DAAP1-3xHA and wild type cells, followed by western blots probed with anti-GlActin and anti-HA antibodies, revealed interaction between the two proteins. B) Image of DAAP1-mNeonGreen in a live cell. The fusion protein localizes to the disc and is enriched in the ventral groove (asterisk) and overlap zone (arrow). Scale bar denotes 5 μm. C) Immunofluorescence of GlActin (magenta), tubulin (cyan), DAAP1-mNeonGreen (green), and DAPI/nuclei (blue) during the cell cycle. DAAP1-mNeonGreen remains on the disc into telophase, is incorporated into the nascent discs by mid cytokinesis (see arrows), and localization is fully restored in newly divided cells. Pearson's correlation coefficients using Costes randomization were calculated for colocalization between DAAP1 and GlActin in each image, shown in the lower righthand corner. The Costes P-value was 1 for all images. Scale bar denotes 5 μm.
plasmids have variable copy number and expression (S4 Fig), which could complicate analyzing knockdown phenotypes. Therefore, we endogenously tagged DAAP1 with a C-terminal mNeonGreen tag and confirmed genomic integration through PCR (S5 Fig). Similar to episomal and integrated DAAP1-3HA, this fusion protein localized to the entirety of the disc, and was highly enriched in the ventral groove and the overlap zone, two areas implicated in attachment (Figs 1A and 2B). Flexion of the overlap zone can be seen during detachment from the surface, consistent with the notion that conformational changes of specific regions of the disc can regulate attachment (S1 Movie).
To investigate co-localization between GlActin and DAAP1, we calculated Pearson's correlation coefficients for these two proteins using the JACoP plugin for ImageJ [24,25]. While DAAP1-mNeonGreen does not appear to co-localize with GlActin during interphase, in mitosis DAAP1 leaves the ventral disc and was observed to co-localize with GlActin ( Fig 2C).
We also measured the intensity of DAAP1-mNeonGreen fluorescence in GlActin-depleted cells, selecting a region in the ventral groove, as that is where DAAP1 levels are the highest (Fig  3). While there was no significant difference between fluorescence levels in attached cells ( Fig  3A and 3B), this difference was significant in unattached cells (Fig 3C and 3D). Furthermore, in some GlActin-depleted cells, we observed DAAP1 loss of enrichment in the entire disc ( Fig  3C). These results indicate that GlActin influences DAAP1 disc localization.
The microtubules in the ventral disc are stable and do not undergo dynamic instability; they rely on a set of core DAPs to maintain this stability [15]. During mitosis, the disc is dismantled as the cells must detach to "swim" away from each other in cytokinesis [26,27]. New discs in the daughter cells appear during early cytokinesis and are built from microtubules originating from the mother disc and the median body [26,28]. Notably, DAAP1-mNeonGreen remains on the disc during telophase, late into disassembly, and begins to coalesce again on the daughter discs in mid cytokinesis. In newly-divided cells, DAAP1-mNeonGreen has returned to its disc localization and is once more enriched in the ventral groove ( Fig 2C). Fluorescence recovery after photobleaching (FRAP) confirmed the stability of DAAP1-mNeonGreen for disc localization, as fluorescence levels did not recover for 12 minutes after photobleaching (Fig 4) in either of the two regions of the disc where DAAP1 is enriched. Viability was confirmed by cell movement as seen in S2 Movie. This result is similar to the fluorescence recovery of other DAPs. Taken together, these data indicate that DAAP1 is a GlActin interactor and stably localized to the ventral disc in interphase.

DAAP1-depleted cells have intact ventral discs
Since GlActin is involved in both attachment and disc morphology, we asked if DAAP1 could have a similar role. We attempted depletion of DAAP1 with translation-blocking morpholinos, however, the sequence was design-resistant and the knockdown level inconsistent. We then turned to CRISRPi to investigate DAAP1's potential role in attachment and disc morphology. To monitor disc morphology as well as deplete DAAP1, we inserted the HALO-tagged delta-giardin marker into the dead Cas9-(dCas9) based CRISPRi plasmid and introduced this into a DAAP1-mNeonGreen expressing cell line. This allowed us to evaluate knockdown on both a whole population ( Fig 5A) and an individual cell basis. We analyzed the fluorescence of endogenously-tagged DAAP1-mNeonGreen in control cells expressing dCas9 with a non-specific guide RNA to those expressing a DAAP1 specific small guide RNA resulting in an average decrease in expression of 40% for the population. We once again separated cells that were attached to culture tubes from unattached cells to probe for disc defects in each subpopulation. We did observe a higher proportion  of disc defects in the DAAP1-depleted cells than in the control, but this difference was not significant (Fig 5B and 5C). These results indicate that a 40% depletion of DAAP1 did not impact gross ventral disc morphology, as assessed by light microscopy.

Inability to maintain attachment in DAAP1-depleted cells
DAAP1 knockdown cells had no visible defects, so DAAP1 may not be needed for disc assembly. To investigate a possible role in attachment, we used the aforementioned attachment assay on DAAP1-depleted cells and observed a small but significant decrease in attachment for two of three replicates relative to control cells (S2 Table). As CRISPRi is known to have variable penetrance in Giardia [29], we analyzed fluorescence levels in unattached and attached cells separately. Here we found reduced levels of DAAP1-mNeonGreen in unattached cells compared to attached ones (Fig 6A). This result prompted us to further explore the possible role of DAAP1 in attachment using a shear force attachment assay. In their normal intestinal habitat, Giardia must maintain attachment even when subjected to the flow of chyme. With this in mind, we challenged DAAP1-depleted cells with a fluid shear force. DAAP1-depleted and control cells were allowed to attach inside a flow chamber and then subjected to a flow of growth medium. Cell position and velocity were tracked using Trackmate (Fig 6B, [30]). DAAP1-depleted cells slid significantly faster under flow than their control counterparts (Fig 6C and 6D). Furthermore, more control cells remained in the field of view than DAAP1-depleted cells over time (Fig 6E), indicating a better ability to maintain attachment.

Impaired seal formation in DAAP1-depleted cells
We then hypothesized that the impaired attachment in DAAP1-depleted cells may be due to impairments of the ventral groove or overlap zone "gates" that control fluid flow and buildup of negative pressure differentials under the disc. We used fluorescent microspheres to investigate this fluid flow in live, attached cells imaged in normal growth medium. We found that spheres entered under the disc more frequently in DAAP1-depleted than in dCas9 control cells, consistent with a defect in gating flow under the disc (Fig 7A). Furthermore, we found more beads per cell in the DAAP1-depleted cells (S4 Table).

DAAP1 in Giardia attachment
Since the ability of the gates to close affects disc seal formation [8,14], we used the plasma membrane stain, CellMask Orange and Total Internal Fluorescence (TIRF) microscopy to evaluate seal formation for DAAP1-depleted and dCas9-Control cells. This approach allowed us to assess the different levels of attachment based on which portions of the cell were in contact with the surface. Since stage 1 of attachment is unattached skimming, we omitted these cells from our analysis, and only included cells in stage 2 (lateral crest contact), stage 3 (lateral shield contact), and stage 4 (bare area contact). We found more cells at stages 2 and 3 and fewer in stage 4 for DAAP1-depleted cells than in the dCas9-Control (Fig 7B), indicating impaired seal formation around the ventral disc. Taken together, these data demonstrate that DAAP1 functions at the disc gates that regulate fluid flow and ventral disc seal formation.

Discussion
Here, we have confirmed that GlActin has a role in ventral disc morphology and disc-based attachment (Fig 1). We also show that GlActin interactor DAAP1 is a stable component of the ventral disc (Fig 4). While perturbation of DAAP1 levels was not associated with visible effects on disc morphology, it resulted in decreased attachment (Figs 5-7).
The ventral disc is well known as a Giardia-specific microtubule-based structure [16,31] and our data show that the actin cytoskeleton has an important role in ventral disc assembly and function. A functional relationship between actin and microtubules has been well-documented for other structures. In Chlamydomonas, actin is required for intraflagellar transport [32,33]. Actin monomers are a key component of the flagella through interaction with the inner dynein arm [34][35][36][37] and recent evidence indicates actin is a component of the gammatubulin ring complex [38]. Actomyosin interactions influence the orientation of the mitotic spindle [39,40] and F-actin associates directly with mitotic spindles in Xenopus and starfish embryos [41,42]. Actin is also involved in microtubule-or tubulin-based host interaction in other parasites. For instance, one of Toxoplasma gondii's formins localize exclusively to the conoid [43], a latticelike organelle composed of tubulin which initiates host invasion [44]. There is evidence of coordination between the microtubules of Toxoplasma's glideosome, which is used for actin-based motility, during host invasion [45]. Furthermore, actin polymerization is believed to contribute to conoid extrusion [46,47], and the conoid-localized Myosin H is essential for entry and exit from host cells [48,49]. GlActin polymerization could have an analogous role in controlling the movement of the ventral disc to facilitate attachment.
We also observed defects in the ventral disc in our GlActin-depleted cells (Fig 1), which we characterized as either mild (misshapen, broken, or with small gaps) or severe (unwound, severely misshapen, often associated with cytokinesis defects). There are several possible explanations for these phenotypes. During mitosis, the ventral disc is dismantled and two daughter discs are rebuilt during cytokinesis [26]. One possibility is that GlActin has a role in loading proteins on the assembling disc. GlActin also likely contributes to ventral disc positioning, since actin has a role in positioning basal bodies that nucleate the disc [50,51]. However, many of the cells with defective discs have normally-positioned flagella (Figs 1 and S2), indicating defective discs are not entirely due to basal body mispositioning. While the resolution of light microscopy limits our ability to see details in the microtubule array of the disc [52], over 25% of our unattached GlActin-depleted cells lacked noticeable defects (Fig 1). Therefore, a portion of the attachment defect cannot be attributed to disc defects alone.
Another possibility is that GlActin helps maintain the structure of the disc, consistent with its role in maintaining cell and organelle shape in other organisms [18,21]. The unusual stability of disc microtubules can be at least partially attributed to certain DAPs [8,15,16], which lock the microtubules into place, preventing their turnover. GlActin interacts with at least one PLOS PATHOGENS DAAP1 in Giardia attachment of these DAPs, GL50803_5188 [17], and could be working in conjunction with it to stabilize the disc microtubules themselves. Another explanation is that GlActin works to stabilize one of the non-microtubule components of the ventral disc. The microtubules of the disc are overlaid by the giardin-composed microribbons [53], which are connected by crossbridges [13,54]. Since the spiral shape is often maintained, but unwound, in GlActin-depleted cells (Figs 1 and  S2), it may be that GlActin contributes to establishing and/or maintaining disc overlap, as the overlap zone also functions as the anterior gate of the ventral disc.
A previous study from our lab identified eight GlActin interactors which localize to the ventral disc, including DAAP1 [17]; here we confirmed complex formation between GlActin and DAAP1 through co-immunoprecipitation (Fig 2). We did not detect co-localization between DAAP1 and GlActin in interphase cells, but present evidence that indicates co-localization during mitosis when new discs are assembled (Fig 2). Polymerized GlActin may be enriched in the ventral disc at these stages, but detection may be obscured due to technical limitations. Unfortunately, use of the anti-GlActin antibody does not differentiate between filaments and pools of monomeric GlActin which likely obscures some filaments. Currently, there are no methods to specifically detect F-GlActin, as Rhodamine-Phalloidin does not bind to GlActin filaments [21]. Actin in Toxoplasma gondii was believed to be primarily monomeric and only capable of forming short filaments [55,56]. However, expression of an anti-actin nanobody revealed a previously unseen complex network of F-actin [57]. Poor detection of GlActin using the anti-GlActin antibody could also be due to steric hindrance. Actin-binding proteins could sterically block access of anti-GlActin antibodies to disc-localized GlActin, similar to low antitubulin antibody staining of the ventral disc compared to our ability to see microtubules of the disc with mNeonGreen-Tubulin [26]. Another possibility is that actin only transiently associates with the disc for specific conformational changes which would be challenging to capture in fixed cells. A live marker specific for GlActin is needed to determine the extent of actin association with conformational dynamics of the ventral disc including disc doming, bare area protrusion and disc flexion as seen in S1 Movie. It would also be interesting to assess the association of GlActin and DAAP1 with the ventral disc during encystation, when the ventral disc is broken down to make cysts for transmission, as well as during excystation when excysted cells rebuild their ventral disc to establish infection [57].
Similar to other disc-associated proteins [43], DAAP1's localization to the disc is stable during interphase, as we did not see any recovery of this protein for 12 minutes after photobleaching (Fig 4). While DAAP1 was previously identified as a potential disc interactor, it was not found in high enough abundance in the detergent-extracted disc proteome to qualify for further study as a disc-associated protein [14]. These results possibly indicating that this protein is a peripheral component of the disc, which would be expected in order for it to engage with a dynamic subset of GlActin.
Using CRISPR-interference, we depleted DAAP1 from cells, and found a mild defect in attachment, but no obvious impact on disc morphology (Figs 5-7). While a DAAP1 knockout [58] would give us better insight into how DAAP1 loss affects the disc, the ability to readily knockout genes of interest remains under development. CRISPRi and morpholino knockdown both have variable penetrance, and thus the level of depletion fluctuates among cells, and we observed normal discs even in cells with very little DAAP1 fluorescence (Fig 5). In contrast, depletion of median body protein with morpholinos results in broken discs [12]. Although an inability to attach can be attributed to disc defects [12,29], the depletion of gamma-giardin, another disc-localized protein that interacts with GlActin [17], resulted in malformed discs with normal attachment [59]. These findings indicate that some disc proteins may be involved in maintaining the structural integrity of the disc, while others are involved in controlling PLOS PATHOGENS DAAP1 in Giardia attachment conformational changes of the disc required for modulating attachment. Our data is consistent with DAAP1 being in the latter category. [12] The possibility that some disc proteins contribute to attachment but not structure may also explain the disparity between the severity of attachment defects in GlActin-depleted cells and DAAP1-depleted cells. Since many of the discs in GlActin-depleted cells were unwound or misshapen, this suggests roles in building the disc or collaboration with DAPs that maintain the disc structure. However, since over 25% of our unattached GlActin-depleted cells had apparently normal discs, the attachment defect cannot be attributed to disc defects alone. Our previous study identified eight other putative disc-localized F-GlActin interactors which could also contribute to attachment and/or disc morphology [17].
An additional explanation for the severity of attachment defects in GlActin-depleted cells is due to its involvement in supplemental mechanisms of attachment. GlActin has been demonstrated to function in building the ventrolateral flange, Giardia's membrane lamella which also contributes to attachment [60,61]. Further, the marginal plate, a rigid structure associated with the anterior axonemes, is also hypothesized to be involved in attachment [11,31]. GlActin is stably enriched in this structure [21,62] and we have described six interactors which localize there as well [17]. Given GlActin's involvement in all three structures hypothesized to augment attachment, depletion of GlActin likely has multivariant effects on attachment, contributing to a more severe phenotype.
The biology of parasites is often unlike anything seen in other organisms, shaped by the unusual environments in which they live. Giardia's ventral disc exemplifies parasitic innovation and makes this Giardia specific structure a target for drug development. Since most DAPs lack homologs, they are in theory potential drug targets; however, the lack of active sites reduces their potential [63]. Further understanding of the mechanisms of attachment and its regulation will increase the viability of developing drugs that could specifically function through regulating attachment. Actin, however, is a druggable protein and conventional actin drugs do not bind to GlActin or inhibit its function, rendering it an appealing therapeutic target.

Parasite strain and growth conditions
G. lamblia strain WB clone 6 (ATCC 50803; American Type Culture Collection) was cultured as described previously [21].

Vector construction
All constructs in this study were C-terminal fusions, made using Gibson reactions with linearized vectors and PCR products [64]. See S3 Table for primer sequences and workflow.

Live cell imaging
Cells were chilled with ice for 15 minutes to detach from the culture tube and then placed into an Attofluor cell chamber (Molecular Probes) and incubated in a GasPak EZ anaerobic pouch (BD) or a Tri-gas incubator (Panasonic) set to 2.5% O2, 5% CO2 for 90 minutes at 37˚C. For live imaging of Halo-tagged Delta-Giardin JaneliaFluor 646 HaloTag ligand (Promega, GA1120) or JaneliaFluor 549 HaloTag ligand (Promega, GA1110) were added to Attofluor cell chambers in growth media at a ratio of 1:1000 during the initial re-attachment step. Cells were then washed four times with HEPES-buffered saline (HBS: 137mM NaCl, 5mM KCl, 0.91mM Na2HpO4-heptahydrate, 5.55mM Glucose, 20mM HEPES, pH7). Live cell imaging was performed on a DeltaVision Elite microscope (GE) equipped with DIC optics, using a 100 × 1.4 NA or 60 × 1.42 NA objective, and a sCMOS 5.4 PCle air-cooled camera (PCO-TECH) and an Oko lab Bold Line stage-top, incubator, set to 37˚C, 2.5% O2, and 5% CO2.

Coimmunoprecipitation
Cell cultures of 500 mL each of Giardia wildtype and DAAP1-3xHA were grown for 3 days, then iced for 2 hours to detach, and spun for 1500xg at 4˚C for 20 minutes. Cells were then washed twice in 1X HBS with 2X HALT protease inhibitors and 10 μM chymostatin, 1 μM leupeptin, and 1 μM E64. Each pellet was resuspended to a final volume of 1.2 mL and 100 mM DSP in DMSO was added to a final concentration of 1 mM and incubated at room temperature for 30 minutes. The reaction was quenched for 15 minutes with an addition of Tris pH 7.4 final concentration 20 mM. Cells were then pelleted by spinning for 7 minutes at 700xg and resuspended in 350 μL lysis buffer (80 mM KCl, 10 mM imidazole, 1 mM MgCl2, 1 mM EGTA, 5% Glycerol, 20 mM HEPES, 0.2 mM, CaCl2, 10 mM ATP, 0.1% Triton X-100, 500 mM NaCl, pH 7.2, 2XHALT protease inhibitors and 10 μM chymostatin, 1μM leupeptin, and 1 μM E64). Cells were then lysed by sonication and cleared with a 10 minute spin at 10,000xg. A volume of 17.5 μL of equilibrated EZview Red Anti-HA Affinity gel (Sigma) was added to each tube of lysate, then incubated at 4˚C with end-over-end mixing for 1 hour. Beads were then spun at 8,200xg for 30 seconds and the supernatant was discarded, followed by a total of three washes with 750 μL wash buffer (80 mM KCl, 10 mM imidazole, 1 mM MgCl2, 1 mM EGTA, 5% Glycerol, 20 mM HEPES, 0.2 mM CaCl2, 10 mM ATP, 0.5% Tween, 500 mM NaCl, pH 7.2). Each wash consisted of end-over-end rotation for 5 minutes followed by a 30 second spin at 8,200 x g. Protein was then incubated with 50 μL of 8 M Urea at RT for 20 minutes to elute. The beads were pelleted and then sample buffer was added to the supernatant, the sample was boiled for 5 mins at 98˚C, and run on a 12% SDS-PAGE gel, followed by western blot protocol described previously [21].

Immunofluorescence microscopy
Fixation and imaging were performed as described in previous work [65], with the addition of a JaneliaFluor 646 HaloTag ligand or JaneliaFluor 549 HaloTag ligand at 1:1000 in the second immunostaining step. Image analysis was performed in FIJI [23]. DAAP1 intensity was measured by generating a circular region of interest surrounding the ventral groove and measuring fluorescence intensity at the brightest point in each image stack.

Morpholino knockdown
Morpholino knockdown was performed as described previously [66]. Cells were lysed and run on a 10% SDS-PAGE gel, followed by western blot protocol as described previously [17,21]. For each of three biological replicates, samples were divided equally into three lanes and intensity of each lane measured using FIJI and averaged. Cell counts were performed with a MoxiZ coulter counter to ensure the same number of cells were included in each sample. For morpholino sequences, see S3 Table. Culture tube attachment assay GlActin: Cells were treated with either a standard control or anti-GlActin morpholino and allowed to recover for 24 hours in a total volume of 13 mL of TYDK media at 37˚C. At 24 hours, each culture tube was inverted six times; detached cells in media were decanted and counted with a MoxiZ coulter counter (Orflo), then replaced with 13 mL ice cold media.

DAAP1 in Giardia attachment
Culture tubes were placed on ice for 30 minutes to detach the remaining cells, after which these were also counted. Cells were then imaged live or fixed as described previously [26]. Three independent replicates of each experimental and control were analyzed. DAAP1 assays were performed similarly, except knockdown was accomplished with CRISPRi rather than morpholinos (see below).

Assessment for live/dead cells
Stock solutions of dye were created by dissolving 5 mg fluorescein diacetate (Sigma-Aldrich, F-7378) in 1 ml acetone and 2 mg propidium iodide (Sigma-Aldrich, P4170) in 1 ml HBS. A 2X stain solution was prepared by diluting the stock solutions of fluorescein diacetate and propidium iodide by 1:100 in 1X HBS. Cells were pelleted by centrifugation at 500 x g and resuspended in 1 mL 1X HBS. 1mL of 2x stain solution was then added to the cell suspension, washed twice in 1X HBS by centrifugation and resuspended in 1X HBS for imaging.

Fluorescence recovery after photobleaching
FRAP was performed on DAAP1-mNeonGreen in the ventral disc to measure protein turnover. Cells were chilled with ice for 20 minutes to detach from the culture tube and then placed into an Attofluor cell chamber (Molecular Probes) and incubated in a O 2 /CO 2 incubator (Panasonic) for 1-2 hours at 37˚C. Media was replaced with 37˚C 1X HBS before imaging. All experiments were conducted using a FRAP-enabled DeltaVision Spectris equipped with an environmental chamber set to 37˚C, 405 nm 50 mW solid-state laser set to 100% laser power with 25 ms stationary pulse. Three images were taken until the bleach event for 0.5 second intervals, followed by image acquisition every 10 sec for 10 minutes. Normalized mNeonGreen fluorescence recovery was calculated by subtracting the background noise from the ROI intensity measurement; to normalize for photobleaching due to imaging, this measurement was then divided by a fluorescent control ROI intensity measurement with the background noise subtracted as well.

CRISPR interference knockdown
Constructs were designed as described previously [67]. For primer design and workflow, see S3 Table. Cells containing an integrated C-terminally mNeonGreen-tagged DAAP1 were transfected with a dead Cas9 vector which contained either a non-specific small guide RNA (dCas9-CTRL) or a small guide RNA targeting base pairs 392 through 411 of the coding region of DAAP1 (dCas9+sgRNA392-411). The CRISRPi vectors also carried a C-terminally HALOtagged delta-giardin. Upon establishment of the line (7-10 days after transfection), cells were split into new cultures to measure knockdown efficiency and preserve the fresh transformants cryogenically. Knockdown efficiency was measured using a Tecan Spark plate reader. dCas9-CTRL or dCas9+sgRNA392-411 cells were pelleted and washed with 1X PBS and then resuspended in 1X PBS such that a 150 μL sample would contain 250,000 cells. Samples were loaded into Tecan black flat 96-well plates and DAAP1-mNG levels were measured. Three measurements per sample were averaged and three independent transformations were averaged for total knockdown efficiency, normalized to the control and blanked with 1X PBS.

Shear force attachment assay
Microfluidic devices were fabricated using a silicon master, which was created using deep reactive ion etching. The silicon master contained a raised channel that had a rectangular crosssectional area, with a length of 1 cm, a width of 1 mm, and a height of 115 μm. A polydimethylsiloxane (PDMS) (Sylgard 184, Dow Corning) mold was made using a 10:1 base to curing agent ratio, which was mixed for 5 minutes, degassed for 20 minutes, poured over the master, and placed in an oven for 10 minutes at 110˚C to cure. The mold was peeled away from the silicon master and holes were punched for the inlet and outlet of the channel. Afterwards, inlet and outlet ports were created using a custom aluminum mold with pegs to insert the silicone tubing (0.040 inch ID/ 0.085 inch OD, HelixMark). Degassed 10:1 PDMS was poured into the preheated aluminum mold and cured at 110˚C for 1 hour. The inlet and outlet ports were aligned with holes on the top face of the microfluidic device and stuck together using uncured PDMS and 5 minutes of oven heat. Finally, the bottom face of the microfluidic device and a clean glass slide were plasma treated for 10 seconds and aligned together to create a water-tight seal between the two layers.
Control dCas9 or DAAP1 targeting dCas9+sgRNA392-411 trophozoites were chilled for 20 minutes on ice and then loaded into microfluidic devices placed in a GasPak EZ anaerobic pouch (BD Scientific) at 37˚C for 1 h to allow re-attachment. Growth media was drawn into a disposable syringe (BD Scientific) and then loaded onto a syringe pump (Harvard Apparatus). For experiments, the flow rate was set to 20 μl/min for 30 s to clear swimming trophozoites and then ramped to 100 μl/min over 5 s and challenged for 60 s. The challenge assays were recorded at 2 FPS using a Nikon Eclipse Ti microscope with a 40X objective, using phase illumination, Clara DR-1357 CCD camera (Andor), and a live-cell incubation chamber (In Vivo Scientific, Inc) to maintain a temperature of 37˚C. Trophozoite mean speed was quantified starting at 35s using the TrackMate ImageJ plugin via manual tracking over the first 10 seconds of 100μL/min flow [30].

TIRF microscopy
Cells were prepared for live imaging as described above, with the addition of CellMask Orange Plasma membrane Stain (ThermoFisher) according to manufacturer's instructions after media replacement with 1X HBS. Cells were imaged on a DeltaVision OMX (GE) using 60X TIRF objective in an environmental chamber set to 37˚C, 2% CO2 and 5% O2. Images were taken from randomly selected fields of view. Each dish was imaged for a maximum of 30 minutes to minimize stress on the cells. Three dishes from two independent experiments were used for each condition.

Live imaging with microspheres
Cells expressing DAAP1-mNeonGreen were transfected with a plasmid containing Halotagged delta-giardin and either the dCas9 control or dCas9+sgRNA392-411 to deplete DAAP1. Cells were grown in parallel and upon confluency were chilled with ice for 15 minutes to detach from the culture tube and then placed into an Attofluor cell chamber (Molecular Probes), then incubated in a GasPak EZ anaerobic pouch (BD) or a Tri-gas incubator (Panasonic) set to 2.5% O2, 5% CO 2 for 90 minutes at 37˚C. 646 HaloTag ligand (Promega, GA1120) was added to cells 45 minutes after incubation at 1:1000 concentration, incubated for 15 minutes, and then replaced with TYDK for 30 minutes prior to imaging. 2.5 μL of fluorescent microspheres (0.2 μm, Polysciences Fluoresbrite YO Carboxylate Microspheres catalog number 19391) were added to 0.5 mL 1xHBS and sonicated for 10 minutes using a bath sonicator to disrupt aggregates. This diluted solution was then added to cells at a 1:10 ratio to each chamber (30 μL diluted bead solution in 300 μL cells). Images of single focal planes in which the disc was visible were taken every 1 second for 15 seconds for each condition. Cells with beads which were collected under their ventral disc and cells with no beads were counted. Cells which detached or moved were excluded from analysis. Live cell imaging was performed on a