Toxoplasma and Plasmodium associate with host Arfs during infection

ABSTRACT The apicomplexans Toxoplasma gondii and Plasmodium are intracellular parasites that reside within a host-derived compartment termed the parasitophorous vacuole (PV). During infection, the parasites must acquire critical host resources and transport them across their PV for development. However, the mechanism by which host resources are trafficked to and across the PV remains uncertain. Here, we investigated host ADP ribosylation factors (Arfs), a class of proteins involved in vesicular trafficking that may be exploited by T. gondii and Plasmodium berghei for nutrient acquisition. Using overexpressed Arf proteins coupled with immunofluorescence microscopy, we found that all Arfs were internalized into the T. gondii PV, with most vacuoles containing at least one punctum of Arf protein by the end of the lytic cycle. We further characterized Arf1, the most abundant Arf inside the T. gondii PV, and observed that active recycling between its GDP/GTP-bound state influenced Arf1 internalization independent of host guanine nucleotide exchange factors (GEFs). In addition, Arf1 colocalized with vesicle coat complexes and exogenous sphingolipids, suggesting a role in nutrient acquisition. While Arf1 and Arf4 were not observed inside the PV during P. berghei infection, our gene depletion studies showed that liver stage development and survival depended on the expression of Arf4 and the host GEF, GBF1. Collectively, these observations indicate that apicomplexans use distinct mechanisms to subvert the host vesicular trafficking network and efficiently replicate. The findings also pave the way for future studies to identify parasite proteins critical to host vesicle recruitment and the components of vesicle cargo. IMPORTANCE The parasites Toxoplasma gondii and Plasmodium live complex intracellular lifestyles where they must acquire essential host nutrients while avoiding recognition. Although previous work has sought to identify the specific nutrients scavenged by apicomplexans, the mechanisms by which host materials are transported to and across the parasite vacuole membrane are largely unknown. Here, we examined members of the host vesicular trafficking network to identify specific pathways subverted by T. gondii and Plasmodium berghei. Our results indicate that T. gondii selectively internalizes host Arfs, a class of proteins involved in intracellular trafficking. For P. berghei, host Arfs were restricted by the parasite’s vacuole membrane, but proteins involved in vesicular trafficking were identified as essential for liver stage development. A greater exploration into how and why apicomplexans subvert host vesicular trafficking could help identify targets for host-directed therapeutics.

they have evolved closely with their hosts to promote replication and evade immune recognition.During invasion, the parasites isolate a portion of the host cell membrane to form an enclosure termed the parasitophorous vacuole (PV).While the parasitophorous vacuole membrane (PVM) is a physical and semipermeable barrier that protects the parasites, it also restricts the free exchange of nutrients and waste with the host cytosol (1,2).Both T. gondii and Plasmodium depend on nutrient transport across the PVM to acquire host amino acids, sterols, fatty acids, and other essential nutrients (3,4).To date, how apicomplexans recruit and transport host resources into the PV remains a critical area of study.
Previous research has uncovered how T. gondii and Plasmodium reorganize host structures during their intracellular lifecycles.Specifically, both parasites reside at the juxtanuclear position, recruit the microtubule-organizing center, and become enveloped by microtubules that traffic the endoplasmic reticulum (ER), Golgi, and lysosomes to the PV (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16).T. gondii then internalizes select host organelles through deep invagina tions of the PVM formed by host microtubules or through fusion with the intravacuolar network (15).Conversely, liver stage Plasmodium parasites do not internalize organelles but have been shown to acquire exogenous particles into the PV lumen through an unknown phagocytic mechanism (13).Taken together, these observations suggest that apicomplexans may associate with host organelles to acquire nutrients.Furthermore, the reorganization of host organelles could allow the parasites to interact with the vesicular trafficking network.
The human genome encodes five ADP ribosylation factors (Arfs) that regulate host vesicle trafficking across the endomembrane system (17).Arf proteins (Arf1, Arf3, Arf4, Arf5, and Arf6) are small GTPases responsible for recruiting vesicle coat complexes and modifying membrane lipids.These proteins have different subcellular localizations; Arf6 primarily localizes to the plasma membrane but Arf1, Arf3, Arf4, and Arf5 are found at the Golgi, endosomes, ER, and ER-Golgi intermediate compartment for trafficking.Arf-associated transport between the cis-Golgi and the ER uses coatomer coat complexes (COPI and COPII), while trafficking between the trans-Golgi network and endosomes uses Golgi localized, γ-ear-containing, Arf-binding proteins 1-3 (GGA1-3) (18).Arfs are found in the cytosol and loosely associated with membranes as a GDP-bound inactive monomer.Upon GTP exchange, Arfs integrate into organelle membranes and recruit effector proteins including coat complexes.GTP exchange is facilitated by guanine nucleotide exchange factors (GEFs) while GTP hydrolysis is supported by GTPase-activat ing proteins (18).Pathogens including Legionella pneumophila, Chlamydia trachomatis, and Salmonella enterica all subvert host Arfs but their function during apicomplexan infection remains largely unknown (19)(20)(21).Apicomplexan parasites are known to fragment and then recruit host Golgi stacks, a process correlated with their viability (6,11,12,22).Thus, Arf proteins, which have a prominent role in Golgi trafficking, are a particularly enticing host protein family to study for their possible role during parasite infection.
In this study, we interrogated host trafficking proteins during T. gondii and Plasmo dium berghei infection of nucleated cells.We found that all human Arf proteins were trafficked to the T. gondii PV during early infection and internalized by the end of the lytic cycle.We further characterized host Arf1 as it was the most abundant in the T. gondii PV.We found T. gondii preferentially internalized Arf1 during active Arf recycling and that the puncta of Arf1 inside the PV colocalized with vesicle coat complexes and exogenous sphingolipids.By contrast, P. berghei did not associate with Arf1 during the infection of hepatocytes indicating it is not recruited by all apicomplexans.Instead, P. berghei development was reduced upon gene depletion of Arf4 and the host GEF, GBF1.Arf4 and GBF1 were also shown to associate with the PVM, highlighting their critical function during the liver stage.Together, this work provides insights into host trafficking machinery hijacked during apicomplexan parasite infection and highlights distinctions between Toxoplasma and Plasmodium.Future studies exploring the possible function of this pathway in nutrient acquisition and intracellular survival could advance our understanding of critical host-parasite interactions in these deadly pathogens.

Host Arfs associate with the T. gondii PV throughout the lytic cycle
We investigated the localization of host Arfs during T. gondii infection using immuno fluorescence microscopy.To avoid the potential for antibody cross-reactivity with T. gondii Arf1 (Fig. S1), we transfected HeLa cells for 24 hours with plasmids encoding human Arfs (Arf1, Arf3, Arf4, Arf5, and Arf6) fused to hemagglutinin (HA).Twenty-four hours post-transfection, cells were infected with an RH-88 strain of T. gondii expressing mCherry (Tg-mCh).Infection proceeded for 32 hours, allowing the parasites to undergo at least five rounds of replication before fixation.As a control, HeLa cells were transfected with a plasmid encoding GFP, infected with Tg-mCh, and fixed at 32-and 48 hours post-infection (hpi).Uninfected and Tg-mCh-infected cells were stained with Hoechst before visualization by confocal microscopy.We observed no internalization of GFP into the T. gondii vacuole at either time point (Fig. S2A).On the other hand, we observed puncta of Arf1, Arf3, Arf4, Arf5, and Arf6 inside the PV that is associated with T. gondii at 32 hpi (Fig. 1A).
To further understand host Arf internalization during infection, we quantified the percentage of PVs containing at least one punctum of host Arfs at various times post-infection.We examined Arf recruitment during invasion (0.5 hpi), after invasion but before the first replication cycle (4 hpi), halfway through the lytic cycle (24 hpi), and during late lytic development when parasites prepare for egress (48 hpi).During parasite invasion (0.5 hpi), we observed a gathering of Arf6 at the PVM.This phenotype was unsurprisingly unique to Arf6 given that it is the only Arf to localize at the host cell membrane for trafficking (Fig. S2B).All Arfs then followed the same pattern of localization during the lytic cycle.Before parasite replication (4 hpi), less than 20% of T. gondii PVs had enrichment of Arfs at the host-parasite interface (Fig. 1B and C; Fig. S2).
As Tg-mCh replicated (24 hpi), intense speckles of Arfs were seen in uneven patterns between the individual parasites in about 50% of all vacuoles.By the end of the lytic cycle (48 hpi), each Arf was found in more than 80% of T. gondii PVs (Fig. 1C).Thus, as infection progressed (4-48 hpi), more T. gondii PVs internalized host Arfs, indicating that these host proteins could be important during the later stages of the lytic cycle.These results are consistent with previous studies that described Arf6 during T. gondii infection.Specifically, Vieira da Silva et al. found an accumulation of Arf6 during the invasion, and Hartman et al. reported the internalization of Arf6 into the T. gondii PV at 24 hpi (23,24).
To quantify Arf recruitment, we completed an analysis of the mean gray value of each Arf within the T. gondii PV as previously implemented by Romano et al. with the T. gondii PV (6,25).While all Arfs are internalized, we found Arf1 was the most abundant in the T. gondii PV (Fig. 1D).Importantly, we observed no linear relationship between the host expression of Arf1 and the internalization of Arf1 into the PV, suggesting that higher Arf expression did not influence our abundance analysis (Fig. S3).To focus our efforts on Arf1, we asked whether endogenous Arf1 expression levels changed during Tg-mCh infection.Previously published RNA-seq data sets have reported no significant changes in Arf1 expression levels in human foreskin fibroblasts and mouse models of acute and chronic T. gondii infection (Table S1) (26,27).To confirm, we used qRT-PCR and assessed Arf1 expression levels at 4, 24, and 48 hpi of Tg-mCh in HeLa cells (Fig. S4).Consistent with previous findings, we observed no significant change in Arf1 expression, suggesting that while T. gondii can sequester Arf1 into its PV, it does not alter its expression.

Proper recycling of Arf1 is important for its internalization into the T. gondii PV
Arf recycling from a GDP-bound to a GTP-bound state is essential for cargo sorting and vesicle formation (18).We investigated whether active Arf1 recycling was a prerequisite for its internalization using the well-characterized point mutations in the Arf-GTP binding pocket that creates a constitutively active (Arf1-Q71L) or dominant negative (Arf1-T31N) protein state (28).Since Arf1-Q71L is constitutively bound to GTP, the protein still promotes vesicle formation and budding but the vesicles cannot dock to their target membranes.As a result, Arf1-Q71L remains bound to vesicle and organelle membranes.By contrast, Arf1-T31N cannot bind GTP so it fails to associate with membranes and redistributes in the cytosol (28).
HeLa cells were transfected with Arf1-WT, Arf1-Q71L, and Arf1-T31N as HA fusion proteins before infection with Tg-mCh.At 48 hpi, the cells were fixed, stained with anti-HA, and visualized by microscopy (Fig. 2A; Fig. S5A).In uninfected HeLa cells, we observed the characteristic phenotypes of the Arf1-Q71L and Arf1-T31N mutants.Arf1-Q71L formed aggregates within the cytosol as it failed to disassociate with membranes while Arf1-T31N became dispersed.In Tg-mCh-infected cells, the number of T. gondii vacuoles that internalized Arf1 significantly decreased by approximately 50% in cells overexpressing either the constitutively active or dominant negative Arf1 protein (Fig. 2B).We also found vacuoles that internalized either mutant Arf1-Q71L or Arf1-T31N contained only a few puncta, whereas Arf1-WT was observed in a greater abundance.We calculated the mean gray value for the Arf1 signal inside the PV and found a significant decrease in the amount of Arf1 trafficked inside the vacuole for both mutant proteins (Fig. 2C).These findings indicate that without proper Arf1 recycling, internalization into the T. gondii PV is reduced.Moreover, we found a significant reduction in PV size at 48 hpi for cells overexpressing either Arf1-Q71L or Arf1-T31N when compared to Arf1-WT (Fig. S5B).

Arf1 is trafficked to the T. gondii PV independent of host GEF activity
Arf recycling and GTP exchange are facilitated by a family of host GEF proteins (18).We hypothesized that if Arf1 recycling was important for its internalization, a specific host GEF may support the association of Arf1 to the T. gondii PV.To test this, we focused on GBF1, the canonical Arf1 GEF that regulates vesicle trafficking from the Golgi to the ER (29).We began by characterizing the location of endogenous GBF1 during the T. gondii lytic cycle.Uninfected and Tg-mCh-infected HeLa cells were stained with an anti-GBF1 antibody at 4, 24, and 48 hpi.GBF1 was internalized by the T. gondii PV starting early in infection (4 hpi) with nearly 100% of T. gondii vacuoles containing at least one punctum by the end of the lytic cycle (48 hpi) (Fig. 3A and B).To confirm this recruitment, YFP-GBF1 was overexpressed in Tg-mCh-infected HeLa cells and visualized (30).We found YFP-GBF1 was also internalized into the T. gondii PV (Fig. S6).
To further explore the role of GBF1 during infection, we employed the well-character ized chemical inhibitor golgicide A (GCA) to selectively decrease GBF1-mediated Arf1 activation (31).We treated uninfected and Tg-mCh-infected HeLa cells overexpressing Arf1-HA with 3.5 µM GCA for 3 hours before fixation at 24 hpi (Fig. 3C).Cells were stained with GM130 to mark the cis-Golgi and anti-HA to visualize Arf1.In uninfected control cells, GCA caused the host cis-Golgi to disassemble and Arf1 to be released from the Golgi membrane, as expected (Fig. 3C) (31).In Tg-mCh-infected cells, we observed Golgi disruption upon GCA treatment, but parasites remained capable of Arf1 internalization.To distinguish differences in Arf1 internalization during GCA treatment, the number of T. gondii vacuoles that contained at least one punctum of Arf1 was quantified.The percentage of vacuoles that internalized Arf1 remained unchanged between DMSO and GCA treatment, suggesting that GBF1 was not required for Arf1 trafficking to the PV (Fig. 3D).Moreover, while GBF1 and Arf1 colocalized at the host Golgi, they did not colocalize in the T. gondii PV.In the parasite vacuole at 32 hpi, Arf1 existed as puncta of protein around the tachyzoites, while GBF1 appeared to associate with the parasite's apical region that houses its Golgi complex (Fig. 3E).Thus, while GBF1 might play a role during infection, its function appears to be independent of Arf1.
Arf1 also interacts with several other GEFs (i.e., BIG1 and BIG2) at the trans-Golgi network for protein sorting and endosome trafficking (18).We investigated whether these additional GEF proteins could support Arf1 recycling and internalization by employing brefeldin A (BFA), a small molecule inhibitor that indiscriminately targets the Sec7 domain of GEF proteins (31).Uninfected and Tg-mCh-infected HeLa cells overex pressing Arf1-HA were treated for 3 hours with 1 µM BFA before fixation at 24 hpi (Fig. 3C).As expected, BFA treatment caused the complete disassembly of the host cis-Golgi in uninfected cells; however, no significant change in the number of T. gondii vacuoles that internalized Arf1 was observed between DMSO and BFA treatment (Fig. 3C and  D).To determine whether host GEF activity would regulate Arf1 internalization later in infection, we treated Tg-mCh-infected cells with DMSO, GCA, or BFA from 29 to 32 hpi.Again, we observed no significant difference in the number of vacuoles that internalized Arf1 with inhibitor treatment (Fig. S7A and B).Together, these results suggest that neither the host Golgi nor GEF function is required for Arf1 internalization by the T. gondii PV.

Arf1 colocalizes with host vesicle coat complexes and sphingolipids in the T. gondii PV
To further interrogate Arf1 internalization by T. gondii, we performed colocalization studies with different host membrane markers.Arf1 primarily localizes to the host Golgi, an organelle that is well characterized to undergo fragmentation during T. gondii infection (6).To determine whether the punctum of host Arf1 observed inside the PV represented Golgi fragments, we performed co-staining for the cis-Golgi marker GM130 in HeLa cells overexpressing Arf1-HA (Fig. 4A).As previously reported, we observed the host cis-Golgi fragmented around the T. gondii PV (6).However, no Golgi fragments colocalized with Arf1-HA inside the T. gondii PV, indicating the Arf1 puncta did not stem from internalized Golgi fragments.
Alternatively, T. gondii could internalize Arf1 vesicles into its PV.Arf1 mobilizes COPI at the cis-Golgi for trafficking to the ER and recruits GGA1 for trafficking from the trans-Golgi (29).The COPI protein complex is composed of seven subunits in mammalian cells with COPI coat complex subunit beta 2 (COPB2) commonly used to represent the localization of the full complex.Given that T. gondii has homologs for multiple COPI subunits, we performed overexpression studies of COPB2 and GGA1 fused to a V5-tag to avoid the potential for antibody cross-reactivity (32).At 48 hpi, we observed puncta of COPB2 and GGA1 inside the T. gondii PV near the residual body (Fig. 4B).We found that approximately 30% of vacuoles contained COPB2 puncta while GGA1 was observed in about 70% of vacuoles at 48 hpi (Fig. 4C).We then overexpressed both Arf1-HA with COPB2-V5 or GGA1-V5 and infected HeLa cells with an untagged RH-88 strain of T. gondii.To assess colocalization, we quantified the Manders' colocalization coefficient (MCC) of COPB2 and GGA1 with Arf1 within the T. gondii PV for at least 20 PVs per biological replicate.The MCC determines the percentage of the total signal from one channel that overlaps with the signal from another such that an MCC of 0.5 indicates a 50% overlap between the two signals.Based on this analysis, a moderate colocalization of the COPB2 or GGA1 puncta to Arf1 puncta was observed (Fig. 4D and F).As a negative control, we overexpressed GFP and Arf1-HA in HeLa cells infected with untagged RH-88 strain of T. gondii and determined the MCC within the T. gondii PV (Fig. 4F; Fig. S8).This analysis indicated no colocalization (MCC <0.5) was observed with GFP and Arf1-HA, as expected.
Given the known dependence of T. gondii on host sphingolipids, we hypothesized the parasites may use Arf1 to recruit lipids (6,15,33,34).We infected HeLa cells overexpress ing Arf1-GFP with an untagged RH-88 strain of T. gondii.Cells were then treated with BODIPY TR C5-ceramide from 24 to 48 hpi before live cell microscopy.We found that BODIPY TR C5-ceramide was successfully trafficked inside the PV at 48 hpi.Furthermore, we found that Arf1-GFP puncta colocalized with BODIPY TR C5-ceramide signal within the PV (Fig. 4E).An analysis of the MCC for at least 20 PVs per biological replicate revealed a strong relationship between Arf1-GFP and BODIPY TR C5-ceramide (Fig. 4F).These findings indicate that Arf1 may facilitate the trafficking of host sphingolipids into the PV.

Arf4 and GBF1 are important for P. berghei liver stage infection
We have previously reported the importance of host vesicular trafficking machinery in the related apicomplexan P. berghei (35).Specifically, we found that GGA1 and two subunits of COPI vesicles (COPB2 and COPG1) were critical to P. berghei liver stage development and that GGA1 and COPB2 were recruited to the PVM of both P. berghei and Plasmodium yoelii at 48 hpi (35).Furthermore, a recent CRISPR-cas9 screen during the Plasmodium liver stage found significant enrichment for gene ontology terms associated with vesicular trafficking (12).Thus, we sought to investigate the function and localiza tion of host Arfs during the P. berghei liver stage.In addition to Arf1 and GBF1, COPI retrograde vesicles are regulated by Arf4.To interrogate the functional role of these host proteins during the P. berghei liver stage infection, we used two pooled siRNAs to knockdown mRNA expression of Arf1, Arf4, and GBF1 in HuH7 cells infected with luciferase-expressing P. berghei (Pb-Luc).At 48 hpi, we found that siRNA knockdown of Arf1, Arf4, and GBF1 had no significant impact on HuH7 cell viability (Fig. 5A and B).Interestingly, Arf1 knockdown had no impact on parasite load while Arf4 and GBF1 depletion led to a significant reduction in P. berghei load (Fig. 5C).We next completed this analysis with single siRNAs against Arf4 and GBF1 to confirm the phenotype.After Arf4 and GBF1 depletion, we observed a significant decrease in P. berghei load in HuH7 cells with two separate siRNAs but no impact on host cell viability (Fig. 5D through F).We then investigated whether Arf4 and GBF1 were important for P. berghei invasion and/or development.We found gene depletion of Arf4 and GBF1 had no impact on the relative infection rate at 4 hpi, indicating that parasite invasion was not hindered (Fig. 5G).However, both the relative infection rate and PV size significantly decreased at 48 hpi with Arf4 and GBF1 depletion (Fig. 5H and I).These findings indicate that Arf4 and GBF1 are critical for parasite development and survival during the liver stage.
We next studied the localization of Arf1, Arf4, and GBF1 during P. berghei liver stage infection.Since we were unable to validate a selective Arf1 antibody, we overexpressed Arf1-HA in HeLa cells and infected them with Pb-Luc.Cells were fixed at 24 or 48 hpi to represent mid-and late liver stage development, and stained for anti-HA and the PVM resident protein UIS4 (36).We observed that Arf1-HA was not internalized by the vacuole nor enriched at the P. berghei PVM at either 24 or 48 hpi (Fig. 6A).This finding is distinct from our results with T. gondii, which recruited all five Arf proteins.We next examined the location of Arf4 and GBF1, both of which were important for P. berghei liver stage development.For these experiments, we could examine the location of endogenous protein as we had validated Arf4 and GBF1 antibodies.HuH7 cells were infected with Pb-Luc, fixed at 24 or 48 hpi, and subsequently stained with anti-UIS4 and anti-Arf4 or anti-GBF1 (Fig. 6B and C).We observed that both Arf4 and GBF1 were associated with the PVM.Arf4 and GBF1 association resembled fragmented Golgi stacks as recently reported in two independent studies (11,12).This is distinct from the association previously observed for COPB2 and GGA1 which colocalized uniformly with UIS4 (35).
To further study the accumulation of endogenous Arf4 and GBF1 at the PVM, we obtained Z-stacks of Pb-Luc-infected HuH7 cells and compared a random distribution of either protein to their actual localization within the host cell.Positive attraction to the UIS4 surface was assessed when the distribution of Arf4 and GBF1 spots deviated from 98% of the simulations for a random distribution (Fig. 6D).Examples of infected cells without a positive attraction of Arf4 and GBF1 to the UIS4 surface are shown (Fig. S9).We found that approximately 80% of the cells showed a positive accumulation of Arf4 and GBF1 at the parasite vacuole, and on average, this accumulation was around 2 µm from the PVM surface (Fig. 6E).A similar analysis could not be performed with Arf1 since the protein can only be observed after Arf1-HA overexpression (i.e., not endoge nous protein).However, visualization of Arf1-HA overexpression in Pb-infected versus Tg-infected cells suggests Arf1 is only recruited to the T. gondii PV.Furthermore, our gene depletion and protein recruitment studies are consistent, suggesting the observed recruitment of Arf4 and GBF1 may support Plasmodium liver stage survival.

DISCUSSION
T. gondii and P. berghei remodel host organelles and the microtubule network to establish a protective and nutrient-rich environment [reviewed for T. gondii in (37) and P. berghei in (4)].Significant research has focused on characterizing the host Golgi given that making contact with and fragmenting the organelle appears to be critical for parasite development (5,6,11,12).Modifying the host Golgi could serve several purposes: to acquire nutrients, to inhibit the presentation of immune signaling molecules, and/or to gain proximity to host vesicular trafficking.To further interrogate host trafficking in the context of T. gondii and P. berghei infection, we characterized the location of Arf GTPases.We found that the entire Arf family was trafficked inside the T. gondii PV and validated previous reports that Arf6 accumulated around invading parasites and was internalized during parasite development (23,24).Furthermore, we discovered that Arf1 was recruited in the greatest abundance and active Arf1 recycling independent of host GEF activity promoted internalization into the PV, suggesting a putative parasite protein could regulate Arf1 activation.Interestingly, Arf movement during T. gondii infection was distinct from Plasmodium, where Arf4 (not Arf1) was recruited to the parasite vacuole but not internalized.Depletion of host Arf4 and GBF1 reduced P. berghei liver stage develop ment and survival with no impact on HuH7 viability.These host proteins may directly aid in parasite development, or their host-dependent roles are indirectly impacting parasite viability.While T. gondii and P. berghei interact with host Arfs differently, our findings build on the growing evidence that apicomplexan parasites leverage the host trafficking machinery for survival.In addition to our study of the Arf family, a second class of small GTPases, Ras-asso ciated binding proteins (Rabs), are also involved in host vesicular trafficking and have been studied in the context of apicomplexans.To date, 15 Rabs have been found inside the T. gondii PV (6,15).Rab11A puncta have been specifically characterized as bona fide vesicles inside the PV using electron microscopy.While the precise mechanism for Rab internalization is unknown, it is postulated that microtubules carrying Rab vesicles may form deep pockets in the PVM.Then through a phagocytic-like process, vesicles are internalized and engulfed by the T. gondii intravacuolar network for vesicle and cargo degradation (15).For P. berghei, there is no evidence that the parasites internalize host Rabs or vesicles.However, Rab1A, Rab11A, and Rab14 are all essential for liver stage development, and Rab7A, which localizes to late endosomes, were found to be enriched around the PVM (13,38,39).Moreover, Rab14 and nine other Rabs are upregulated during P. berghei infection, indicating that the parasite may modulate host gene expression to increase protein trafficking (40).These findings on Rab GTPases thus complement our observations.Arfs are internalized by T. gondii but restricted by the P. berghei PVM.Furthermore, like multiple Rabs, Arf4 was found to promote P. berghei survival and development without being recruited inside the PV.
T. gondii can differentially internalize specific Rab and Arf GTPases into its PV.Rabs from the recycling, secretory, and anterograde pathways are all preferentially internalized by the parasite and we found that Arf1 was internalized with the greatest frequency when compared to other Arf proteins (6,15).Likewise, T. gondii can internalize host endocytic organelles and multivesicular bodies while excluding peroxisomes and ER elements (8,15).This bias supports a parasite mechanism that can identify and divert specific host vesicles and organelles to its PV.Identifying the parasite protein(s) that support such a mechanism as well as what resources they seek to sequester will be critical next steps to further understand this host-parasite interaction.In the case of Arf1, Rab14, Rab30, and Rab43, these predicted vesicles colocalize with exogenous BODIPY TR C5-ceramide and may support the trafficking of sphingolipids and perhaps other essential nutrients and proteins into the PV (6).
Arf1 has multiple functions that could explain why it is targeted by T. gondii.Canonically, Arf1 is associated with vesicular trafficking at the Golgi (29).We found that COPB2 and GGA1, two components of vesicular coat complexes, were also internalized into the T. gondii PV.Importantly, the COPB2 and GGA1 puncta colocalized with Arf1 suggesting that these host proteins could exist as bona fide vesicles.We also observed puncta of COPB2 and GGA1 that did not colocalize with Arf1 inside the PV, which was particularly variable with COPB2.It is possible that other members of the Arf family are activating those vesicles, or the observed puncta could represent digested vesicles.The greater abundance of GGA1 over COPB2 is consistent with previous studies where Rab GTPases involved in anterograde trafficking are more abundant within the T. gondii PV than proteins participating in retrograde pathways (6,15).
Arf1 could also be recruited to supply lipids through either lipid droplet metabolism or non-vesicular trafficking of lipids.Lipid droplets store the precursors of membrane lipids and are released by Arf1 as nano-lipid droplets for further digestion (41).Recent studies have determined that T. gondii infection promotes lipid droplet formation and causes them to rearrange around the parasite before their internalization and degrada tion within the PV (42)(43)(44).Furthermore, Arf1 plays a crucial role in activating two proteins involved in non-vesicular lipid transport: ceramide transporter 1 (CERT1), which facilitates the transport of ceramides from the ER to the Golgi, and pleckstrin homology domain-containing A (PLEKHA8), which transports glucosylceramides within the Golgi and from the trans-Golgi to the plasma membrane (45).Finally, Arf1 plays a critical role in forming the ER-Golgi intermediate compartment, an essential process for transporting the major histocompatibility complexes to the cell surface and the regulation of immune mediators' secretion (46).Thus, disruption of ER-Golgi biogenesis by trafficking Arfs into the PV could impact antigen presentation and immune responses.
We also reported that Arf1 recycling between its GDP/GTP-bound states facilitated its internalization into the T. gondii PV and overexpression of either Arf1-Q71L or Arf1-T31N resulted in a significant reduction in PV size at 48 hpi.This phenotype indicates that functional Arf1 protein or Arf1-mediated functions facilitate proper T. gondii develop ment.Future studies may resolve if Arf recycling directly impacts T. gondii size or if host-dependent functions of Arf are influencing the parasite.However, the finding that Arf1 recycling did not depend on host GEF function hints at a possible parasite protein that can regulate Arf activation.In bacterial systems, secreted effector proteins that mimic Arf GTPase partners can modulate host Arfs.For instance, L. pneumophila secretes the bacterial effector protein RalF, which contains a Sec7 domain that activates host Arf1 for recruitment to its vacuole (19).Conversely, enterohemorrhagic Escherichia coli secretes a type III effector protein, EspG, that binds GTP-bound Arf1 and inhibits bidirectional Golgi trafficking (47,48).It is therefore possible that T. gondii and P. berghei could export proteins into the host cytosol to modulate host protein machinery and recruit specific vesicles to its PV.
Altogether we have demonstrated that host Arfs play a critical role during the T. gondii lytic cycle and the P. berghei liver stage, where each parasite uniquely exploits the protein family.This work builds on the evidence that Arf and Rab GTPases are hijacked by apicomplexan parasites possibly to exploit vesicular trafficking.Importantly, we provide a comprehensive analysis of Arf1 internalization into the T. gondii PV and identify trafficking machinery essential for P. berghei liver stage survival.Additional studies are critical to understand how parasites redirect different GTPases, access their cargo, and undergo nutrient exchange.Further exploration into why Arfs are manipulated by apicomplexans will uncover whether they function to support nutrient acquisition and/or prevent elimination.There is also a need to identify exported and PVM-resident proteins for both T. gondii and P. berghei as a framework to discover candidate protein(s) that may regulate these processes.Continued efforts to resolve the role of Arf GTPases will help lay the groundwork to identify essential and targetable pathways to inhibit apicomplexan parasites for disease control.

Cell and parasite culture
HeLa (Duke Cell Culture Facility) and HuH7 (Sigma, European Collection of Authenticated Cell Cultures) were cultured in DMEM (Gibco) supplemented with 10% heat-inactivated FBS (Sigma) and 1% antibiotic-antimycotic (Sigma).Cells were maintained in a standard tissue culture incubator at 37°C and 5% CO 2 .T. gondii RH-88 stably expressing mCherry (Tg-mCh), generously gifted by Prof. Laura Knolls (University of Wisconsin Madison), and an untagged T. gondii RH-88 (BEI, NR-223) were propagated in HeLa cells by serial passage.Parasites were released from host cells by syringe lysis and passed through a 0.3-µm filter.Anopheles stephensi mosquitos infected with Plasmodium berghei ANKA (Pb-Luc) stably expressing luciferase (Luc) were purchased from the New York University Langone Medical Center Insectary or the SporoCore at the University of Georgia, Athens.P. berghei sporozoites were harvested from freshly dissected A. stephensi salivary glands before experiments.Following the addition of T. gondii and P. berghei to mammalian cells, plates were centrifuged at 500 × g for 10 minutes at room temperature (RT).

Transfections and knockdowns
For transfections, 3,000 HeLa cells were seeded on 384-well glass bottom black plates (Cellvis-MSPP-P38415HN). Twenty-four hours post-seeding, plasmids (75 ng/well), see Table S2, were transfected with 0.035 µL Lipofectamine 3000 reagent (Invitrogen) per well according to the manufacturer's protocol.Protein expression was confirmed by microscopy.Cells recovered for 24 hours before infection with 6,000 T. gondii tachyzoites or 4,000 P. berghei sporozoites per well.

GEF inhibitor studies
To evaluate the role of host GEF activity on Arf1 recruitment, 3,000 HeLa cells were seeded on 384-well glass bottom plates, transfected with Arf1-HA plasmids, and infected with Tg-mCh as described above.At 21 or 29 hpi, infected cells were treated with 3.5 µM golgicide A (APExBIO; Cat# B1385), 1 µM brefeldin A (APExBIO; Cat# B1400), or 1% DMSO (vehicle).Cells were fixed as described below after 3 hours of treatment.

Sphingolipid acquisition studies
HeLa cells were seeded on 384-well glass bottom plates, transfected with Arf1-GFP, and infected with an untagged T. gondii RH-88 strain as described above.At 24 hpi, cells were incubated with 5 µM BODIPY TR C5-ceramide complexed to BSA (Invitrogen; Cat# B34400) in FBS-free DMEM.Nuclei were stained with Hoechst (1:2,000 in PBS) for 5 minutes before live cell microscopy at 48 hpi.

RNA extraction and qRT-PCR
Arf1 mRNA levels during T. gondii infection were assessed by qRT-PCR.HeLa cells seeded in 6-well plates (300,000 cells/well) were infected with 1.5 million T. gondii tachyzoites.Uninfected and T. gondii-infected cells were harvested at 4, 24, and 48 hpi with RNA lysis buffer (Zymo Research).RNA was harvested per the manufacturer's protocol using the Quick-RNA MiniPrep kit (Zymo Research).First-strand cDNA synthe sis was performed with 0.5 µg of RNA using random hexamers (Invitrogen) and the GoScript reverse transcriptase (Promega) per the manufacturer's instructions.qRT-PCR analysis was performed with oligonucleotide primers in Table S3 with 2× universal SYBR green fast qPCR (Abclonal) and a LightCycler 480 Instrument II (Roche Diagnostics).All reactions were at a final volume of 5 µL in a 384-well plate (Roche 04729749001).Cycle threshold (CT) values for the genes were normalized to CT values of the 18S human housekeeping gene rRNA (CT[target] -CT[18S rRNA] = ∆CT).Data were normalized to uninfected cells (∆CT[experimental] − ∆CT[control] = ∆∆CT) and the relative amount was calculated as 2 −∆∆CT .
The mRNA levels of genes targeted by siRNA were measured 48 hours post-reverse transfection.Six pooled wells of a 384-well plate were used to extract total RNA with the Quick-RNA MicroPrep per the manufacturer's instructions (Zymo Research).First-strand cDNA synthesis and qRT-PCR were performed as described above.

Protein expression and western blot
T. gondii RNA extraction and first-strand cDNA synthesis were performed as described above.TgArf1 was PCR amplified from cDNA using the primers in Table S2 for cloning into the pET21a(+) expression vector in frame with a C-terminal His-tag.Plasmids were verified by Sanger Sequencing (Eton Bioscience) and transformed into BL21(DE3) E. coli.Protein expression was induced at OD 600 = 0.6 with 100 µM IPTG (Chem Impex, cat #00194) overnight at 20°C.One milliliter of culture was pelleted and resuspended in 200 µL of denaturing lysis buffer (0.1 M Tris-HCl, pH 8, 0.1 M Na 2 PO 4 , 8 M urea).For HeLa lysate, confluent cells in a T75 flask were trypsinized, washed three times with PBS, and resuspended in 200 µL of denaturing lysis buffer.Both BL21-TgArf1 and HeLa lysates were cleared by centrifugation at 20,000 × g for 20 minutes, boiled, and resolved on a Novex 4-20% tris-glycine gel.Proteins were transferred to a nitrocellulose membrane using a Trans-Blot Turbo transfer system (Bio-Rad), blocked for 1 hour with 3% BSA in PBS containing 0.2% Tween 20 (PBST), and incubated with the following primary antibodies overnight at 4°C: Arf1 polyclonal antibody (Invitrogen PA1-127, 1:1000) and Arf1 monoclonal antibody (ThermoScientific 1862342, 1:1000).The membranes were washed three times with PBST before incubation with Invitrogen Alexa Fluor conjugated antibodies (A11001 and A21206, 1:1000).Membranes were washed three times before imaging with a ChemiDoc system.

P. berghei luciferase assay and cell viability
Forty-eight hours post-infection with Pb-Luc, hepatocytes were assessed for cell viability using CellTiter-Fluor (Promega).The relative fluorescence signal intensity was evaluated with an EnVision plate reader (PerkinElmer).Parasite load from the same microplate was then assessed using Bright-Glo (Promega).Immediately after reagent addition, the relative bioluminescence signal intensity was evaluated.Data were normalized to negative controls.

Image and data analysis
Image analysis to adjust brightness, contrast, and crop images was conducted with FIJI (Schindelin et al. 2012).Positive internalization of Arf proteins was assessed on the middle focal plane of the vacuole and was considered positive if the vacuole contained at least one punctum.The mean fluorescent intensity values within the PV were assessed on one focal plane using images acquired with the same laser intensity and exposure time for all channels.For internalization, mean fluorescent intensity, and MCC of the T. gondii vacuole area was defined by the outermost perimeter of the Tg-mCh rosette.
Imaris (version 9.9.1) was used to study the accumulation of Arf4 and GBF1 in the parasite vacuole.A surface of the PVM was generated by the UIS4 signal and Arf4 and GBF1 spots within the host cell boundary, determined by phalloidin staining, were identified.Positive attraction was assessed using the cumulative count plot where Arf4 or GBF1 spots deviated from 98% of simulations (1,000) for the random distribution of spots.To assess the average distance around the parasite vacuole where Arf4 and GBF1 accumulated, we used the smoothed probability density plot where the accumulation of spots around the surface was greater than the random distribution.
All experiments were completed in triplicate with more than 50 PVs analyzed per condition, except where noted.Data were analyzed using GraphPad Prism10.

FIG 1
FIG 1 Internalization of host Arfs into the T. gondii PV.Representative confocal immunofluorescence images of Tg-mCh (red)-infected HeLa cells overexpressing human Arf-HA fusion proteins.Cells were stained with an anti-HA antibody (green) and nuclei were stained with Hoechst (blue).(A) At 32 hpi, Arf1, Arf3, Arf4, Arf5, and Arf6 were observed within the T. gondii vacuole.Yellow arrows indicate the T. gondii PV.Insets (yellow boxes) of the T. gondii PV are shown to magnify Arf puncta.(B) Arf1, Arf3, Arf4, Arf5, and Arf6 were visualized in uninfected and Tg-mCh-infected HeLa cells at 4, 24, and 48 hpi.The enrichment of host Arfs at the PV was observed at 4 hpi (yellow arrows) which were internalized by the end of the lytic cycle at 48 hpi.(C) Quantification of T. gondii PVs with at least one punctum of host protein at 4 (gray), 24 (blue), and 48 (green) hpi.Data represent mean ± SEM; n = 3 biological replicates analyzing >50 PVs per each condition.(D) Plot showing the mean Arf fluorescent intensity values within the PV of infected cells at 48 hpi.n = 3 biological replicates analyzing a minimum of 20 PVs per replicate.P-values display one-way ANOVA with Dunnett's multiple comparison test for each condition compared to Arf1.**P < 0.01; ****P < 0.0001.Scale bars are 10 µm.

FIG 3
FIG 3 Arf1 and GBF1 are uncoupled in the T. gondii PV. (A) Representative confocal immunofluorescence images of uninfected and Tg-mCh (red)-infected HeLa cells at 4, 24, and 48 hpi.Cells were stained with anti-GBF1 (green) and nuclei were stained with Hoechst (blue).(B) Quantification of T. gondii PVs with at least one punctum of GBF1 at 4 (gray), 24 (blue), and 48 (green) hpi.Data represent mean ± SEM, n = 3 biological replicates analyzing >50 PVs for each condition.(C) Assay schematic showing drug administration at 21 hpi and analysis at 24 hpi.Representative confocal immunofluorescence images of uninfected and Tg-mCh (red)-infected HeLa cells overexpressing Arf1-HA and treated with DMSO, 3.5 µM golgicide A (GCA), or 1.0 µM brefeldin A (BFA).Cells were stained with anti-HA (green), anti-GM130 (magenta), and nuclei were stained with Hoechst (blue).(D) Quantification of T. gondii PVs with at least one punctum of Arf1 at 24 hpi after treatment with DMSO (gray), GCA (blue), or BFA (green).Data represent the percentage of T. gondii PVs with at least one punctum of Arf1 compared to the DMSO control ±SEM, n = 3 biological replicates analyzing >50 PVs for each condition.P-values display one-way ANOVA with Dunnett's multiple comparison test for each condition compared to DMSO.ns = non-significant.(E) Representative confocal immunofluorescence images of uninfected and Tg-mCh (red)-infected HeLa cells overexpressing Arf1-HA at 32 hpi.Cells were stained with anti-HA (green), anti-GBF1 (magenta), and nuclei were stained with Hoechst (blue).Scale bars are 10 µm.

FIG 4
FIG 4 Arf1 colocalizes with COPB2 and GGA1 in the T. gondii PV. (A) Representative confocal immunofluorescence images of the host cis-Golgi (GM130) in Tg-mCh (red)-infected HeLa cells overexpressing Arf1-HA at 48 hpi.Cells were stained with anti-HA (green), anti-GM130 (magenta), and nuclei were stained with Hoechst (blue).(B) Representative confocal immuno fluorescence images of Tg-mCh (red)-infected HeLa cells overexpressing COPB2-V5 and GGA1-V5 at 48 hpi.Cells were stained with anti-V5 (green) and nuclei were stained with Hoechst (blue).Insets (yellow boxes) of the T. gondii PV are shown to magnify the puncta (yellow arrows) of each vesicle coat complex inside of the T. gondii PV. (C) Quantification of T. gondii PVs with at least one punctum of COPB2 and GGA1 at 48 hpi.Data represent mean ± SEM, n = 3 biological replicates analyzing >50 PVs for each condition.(D) Representative confocal immunofluorescence images of untagged T. gondii-infected HeLa cells overexpressing both Arf1-HA and COPB2-V5 or GGA1-V5 at 48 hpi.Cells were stained with anti-HA (green), anti-V5 (red), and nuclei were stained with Hoechst (blue).Insets (yellow boxes) are shown to magnify the colocalization of Arfs and coat complexes within the T. gondii PV (yellow arrows).(E) Representative confocal immunofluorescent images of untagged T. gondii-infected HeLa cells overexpressing Arf1-GFP (green) at 48 hpi.Cells were treated with 5 µM BODIPY-TR C5-ceramide (red) from 24 to 48 hpi.Yellow arrows indicate puncta of Arf1-GFP and BODIPY-TR C5-ceramide that colocalize within the T. gondii PV.An inset (yellow boxes) is shown to magnify the colocalization of Arf1 and BODIPY-TR C5-ceramide.(F) The Manders' colocalization coefficient (MCC) of Arf1 with COPB2, GGA1, and BODIPY-TR C5-ceramide was determined for T. gondii PVs.Data represent COPB2 to Arf1, GGA1 to Arf1, and Arf1 to C5-cer.As a negative control, the MCC of GFP to Arf1 was determined inside of the T. gondii PV for cells overexpressing both proteins.n = 3 biological replicates analyzing >20 PVs for each condition.Scale bars are 10 µm, except for 4D and 4E insets, which are 5 µm.

FIG 5
FIG 5 Arf4 and GBF1 depletion impairs P. berghei liver stage development.(A-C) HuH7 cells were reverse transfected (15 nM) for 48 hours with a non-targeting scramble control (siCTRL, gray) or two pooled siRNAs targeting Arf4 (blue), GBF1 (green), and Arf1 (orange).(A) The relative mRNA levels of HuH7 cells treated with pooled siRNAs were determined by qRT-PCR.Samples were normalized to Hs18S and compared to cells treated with siCTRL.(B) Forty-eight hours post-siRNA transfection, cells were infected with Pb-Luc sporozoites.HuH7 cell viability was assessed at 48 hpi using a CellTiter-Fluor assay.Data are normalized to siCTRL.(C) The parasite load was assessed at 48 hpi and normalized to cells treated with siCTRL.(D-F) HuH7 cells were reverse transfected (15 nM) for 48 hours with a non-targeting scramble control (siCTRL, gray) or two different siRNAs targeting Arf4 (1 and 2, blue) and GBF1 (1 and 2, green).(D) The relative mRNA levels of HuH7 cells treated with single siRNAs were determined by qRT-PCR.Samples were normalized to Hs18S and to cells treated with siCTRL.(E) Forty-eight hours post-siRNA transfection, cells were infected with Pb-Luc sporozoites.HuH7 cell viability was assessed at 48 hpi using a CellTiter-Fluor assay.Data are normalized to siCTRL.(F) The parasite load was assessed at 48 hpi and normalized to cells treated with siCTRL.(G-I) Cells were reverse transfected with two pooled siRNAs (15 nM) for a non-targeting scramble control, Arf4, and GBF1.The (G) relative infection rate at 4 hpi, (H) relative infection rate at 48 hpi, and (I) PV size at 48 hpi were assessed for a non-targeting scramble control (siCTRL, gray), Arf4 (blue), and GBF1 (green).Data for the relative infection rates (Continued on next page)