The endosomal TbTpr86/TbUsp7/SkpZ (TUS) complex controls surface protein abundance in trypanosomes

In trypanosomes the orthologs of human USP7 and VDU1 control abundance of a cohort of surface proteins, including invariant surface glycoproteins (ISGs) by functioning as deubiquitinases (DUBs) Silencing TbUsp7 partially inhibits endocytosis and invariant surface glycoprotein turnover. As a component of cullin E3 ubiquitin ligases, S-phase kinase-associated protein 1 (Skp1) has crucial roles in cell cycle progression, transcriptional regulation, signal transduction and other processes in animals and fungi. Unexpectedly, trypanosomes possess multiple Skp1 paralogs, including a divergent paralog designated SkpZ. SkpZ is implicated in suramin-sensitivity and endocytosis and decreases in abundance following TbUsp7 knockdown and physically interacts with TbUsp7 and TbTpr86. The latter is a tetratricopeptide-repeat protein also implicated in suramin sensitivity and located close to the flagellar pocket/endosomes, consistent with a role in endocytosis. Further, silencing SkpZ reduced abundance of TbUsp7 and TbTpr86 and many trans-membrane domain surface proteins. Our data indicate that TbTpr86, TbUsp7 and SkpZ form the ‘TUS’ complex that regulates abundance of a significant cohort of trypanosome surface proteins.


Introduction
Trypanosomiasis is a vector-borne parasitic disease caused by infection with trypanosomes and includes multiple African and American species. The trypanosome surface is the interface with the host and possesses adaptations critical for survival in multiple distinct environments; this importance is reflected in the diverse surface protein components exhibited by the major human pathogens Trypanosoma brucei and T. cruzi.
The T. brucei mammalian bloodstream trypomastigote has an efficient endocytic system that enables rapid recycling of surface proteins, antibody clearance and immune evasion mechanisms (Manna et al 2015). The vast majority of the surface is covered by the GPIanchored variant surface glycoprotein (VSG), although a considerable diversity of additional proteins also populates the surface (Gadelha et al., 2015, Shimogawa et al., 2018. While VSG itself is responsible for antigenic variation and hence immune evasion, several receptors have been characterised recently, but most surface proteins remain functionally uncharacterised. The trypanosome surface is clearly organised as the proteomes of the cell body, flagellum and flagellar pocket are distinct (Oberholzer et al., 2014, Gadelha et al., 2015. Further, this differential composition extends to internal compartments, with many proteins shared between endosomes and surface domains while other proteins have distinct locations (Gadelha et al., 2015). The mechanisms responsible for targeting likely reside within multiple determinants, including cis-elements embedded within the proteins themselves, post-translational modification, including phosphorylation, fatty acylation and ubiquitylation, together with so far uncharacterised gating mechanisms at the surface (Allen et al., 2007, Emmer et al, 2011, Gadelha et al, 2009, Mussmann et al, 2004, Graf et al., 2013, Baker et al., 2012. ISG75 is a member of an extensive superfamily of T. brucei-specific type I transmembrane proteins with large extracellular domains and comparatively short cytoplasmic domains (Allison et al 2014). ISGs, and their related intracellular glycoproteins (IGPs), are comparatively abundant and the presence of multiple paralogs for each ISG/IGP subfamily suggests a requirement for sequence diversity as well as abundance. Significantly, both ISG75 and its closest relative, ISG65, have a considerable presence within the endosomal system as well as at the surface and are believed to be receptors although physiologically relevant ligands remain unidentified. Further, both are ubiquitylated and while there are clear differences between mechanisms underlying ISG65 and ISG75 targeting and turnover, they share at least one deubiquitylating enzyme in the trypanosome ortholog of Vdu1, while Usp7 only impacts ISG75 turnover (Zoltner et al., 2015). Ubiquitylation is an 3 important regulator of the trypanosome surface but, given the complexity of ubiquitylation systems across the tree of life, it is unclear precisely how these various elements function together.
Protein ubiquitylation proceeds via several steps, depending on the mechanism of action of the ubiquitin ligase responsible for transfer of ubiquitin to the client protein (Hershko and Ciechanover 1998). Cullins are a family of scaffold proteins that support E3 ubiquitin ligases with members present across the eukaryotes, including trypanosomes (Rojas et al., 2017, Lin et al., 2017, Huysman et al., 2014, Desbois et al., 2018, Angers, etal., 2006, Iglesias et al., 2018and Nagai et al., 2018. Cullins combine with RING proteins to form cullin-RING ubiquitin ligases (CRLs) that are highly diverse and play roles in various cellular processes, most notably protein degradation by ubiquitylation. CRLs, such as the Skp1-Cul1-Fbox (SCF) complex target proteins for ubiquitin-mediated destruction and regulate multiple functions including DNA replication, glucose sensing and limb formation in metazoa. The cullin N-terminus is highly variable and, in the SCF complex, interacts with specific adaptor proteins including Skp1 (S-phase kinaseassociated protein 1), to bring substrate proteins close to the E3 ligase Rbx1 (Petroski and Deshaies 2005, Zheng et al., 2002, Goldenberg et al., 2004. Different combinations of adaptor proteins (F-box proteins in the SCF complex) binding to the cullin N-terminus are key to diversifying substrate specificity.
To date functions of Skp1 and its paralogs remain uninvestigated in trypanosomatids, despite the potential for important contributions towards cell cycle progression and differentiation pathways. Here we show that a trypanosomatid-specific Skp1-like protein (SkpZ) forms a novel complex together with TbUsp7 and a trypanosomespecific Tpr protein which we designate as the Tpr86/Ubc7/SkpZ or TUS complex. Further, using unbiased whole cell proteomics we demonstrate that TUS mediates surface transmembrane protein turnover, indicating that the complex is a major factor in controlling surface protein expression.

Results
Trypanosoma brucei possesses multiple Skp1 paralogs: To investigate the complexity and possible roles for Skp1 and paralogs in trypanosomes, we performed a genome search using human Skp1 (P63208) as well as the trypanosome Skp1 paralogs Tb927.11.6130 and Tb927.11.13330 identified previously by cullin affinity isolations (Canavate et al., in preparation) and a Skp1 paralog (Tb927.11610) identified as a suramin sensitivity-4 associated gene (Alsford et al., 2012) as BLAST queries. We searched across high quality kinetoplastid genomes and retrieved orthologs from the vast majority of these taxa.
Phylogenetic reconstruction robustly identified four clades of Skp1 paralogs, which based on associations with cullin ligases we designated as Skp1.1 and Skp1.3, together with a more divergent Skp1-like protein as SkpZ and the fourth paralog with very weak homology that we designated as Skp1-related or Skp1.4 (Figure 1, Tables S1 and S2). Only the genomes of T. cruzi (CL Brener and Dm28c), Angomonas and Strigomonas lack a SkpZ gene, probably the result of incomplete databases.
Skp1 possesses a POZ-domain at the N-terminus, consisting of a β/α sandwich double layer with a C-terminal dimerization domain essential for binding of F-box proteins.
The interaction stabilises the conformation of Skp1 and increases protein levels by preventing aggregation (Zheng et al., 2002, Yoshida andTanaka 2011) Trypanosome Skp1.1 and Skp1.3 retain this canonical Skp1 architecture, but SkpZ has an extended Nterminus ( Figure S1). Skp1.1 and Skp1.3 associate with cullin ligases in trypanosomes (Canavante et al., in preparation) but SkpZ was never detected in association with a cullin complex, suggesting a divergent function.

TbSkpZ is an endosomal protein and interacts with TbUsp7 and an 86kDa Tpr-protein.
To investigate SkpZ further we established an endogenously tagged SkpZ cell line with three copies of the HA-epitope fused at the C-terminus. While Homo sapiens Skp1 localizes to both the nucleus and cytoplasm none of the trypanosome Skp1 orthologs apparently share this localisation and SkpZ localised to the endosomal region, supporting the possibility of a novel function associated with protein trafficking (www.tryptag.org, Dean et al., 2017).
TbSkpZ HA-tagged cells were harvested and cryo-milled (Obado et al., 2016). We identified conditions for isolation of TbSkpZ together with several additional interacting proteins ( Figure 2, Table S3), which were identified by LCMSMS and MaxQuant/Perseus (Zoltner et al., 2020). Unexpectedly, TbSkpZ interacted with TbUsp7 rather than a cullin complex, with several additional proteins also identified ( Figure 2). Altogether four proteins, GRESAG4, PFC17, Tb927.11.810 and TbUSP7 were significantly enriched. GRESAG4 has multiple paralogs, is highly abundant and frequently identified in proteomics analyses which suggests that this is a contaminant, while PFC17 (paraflagellar component 17) is a small protein, which increases the potential of non-specific binding and/or detection in mass spectrometry and is a flagellar protein which, given no additional evidence for flagellar interaction, we considered unlikely to be a bona fide SkpZ partner (Subota et al., 2014).

5
Tb927.11.810 encodes an 86kDa protein with a predicted tetratricopeptide-repeat motif (TPR, IPR011990) located towards the centre of the sequence, between residues 246 and 485, and an overlapping pentatricopeptide repeat region (PPR, IPR002885) at residues 196 to 384, with several experimentally determined phosphorylation sites between 84 and 103; we designated this protein TbTpr86. There is evidence for modest upregulation in the mammalian bloodstage compared to insect forms and decreased expression in stumpy forms at the protein level. Importantly, the protein is located close to the flagellar pocket, similar to SkpZ (www.tryptag.org).
The Tpr motif consists of degenerate 34 amino acid repeats and is present in a wide variety of proteins in other organisms, including anaphase-promoting complex subunits cdc16, cdc23 and cdc27, transcription factors, the major receptor for peroxisomal matrix protein import PEX5 and others (Das et al. 1998). Tandem arrays of three to 16 Tpr motifs form scaffolds to mediate protein-protein interactions including assembly of heteromeric complexes as well as self-assembly (Blatch and Lässle 1999). In T. brucei the Tpr motif is retained by the Pex5 ortholog (Gualdrón-López et al., 2013) and predicted for numerous additional proteins.
TbSkpZ knockdown reduces TbUsp7, ISG75 and Tpr86 levels. Cells harbouring a stemloop RNAi construct specific for TbSkpZ were induced with tetracycline. Quantification revealed that knockdown of TbSkpZ led to enlargement of the flagellar pocket, the 'BigEye' phenotype originally observed for knockdown of clathrin (Allen et al., 2003). Specifically, 17% of induced cells possessed the BigEye morphology, constituting a six-fold increase in frequency when compared to uninduced controls. Furthermore, TbSkpZ knockdown resulted in a slight proliferative effect. These impacts effectively mirrored those obtained for TbUsp7 knockdowns (Zoltner et al., 2015).
A whole cell proteome derived from SDS lysates of silenced cells was compared with parental cells using stable isotope-labeling by amino acids in culture (SILAC) and LCMSMS. 3904 protein groups were detected, of which 3116 could be quantified in both replicates ( Figure 3; Table S3). Significantly, TbSkpZ knockdown reduced protein levels of TbUsp7 by 38%, ISG75 by ~50%, Tpr86 by ~42% and SkpZ itself by ~70%, which also validates the knockdown as efficient (Figure 3). We were able to reproduce the ISG75 decrease using western blotting providing orthogonal validation of the proteomics data ( Figure 3, inset). Significantly, TbUsp7 knockdown was previously shown by us to decrease both ISG75 expression (~40%) and Tpr86 (~58%) (Zoltner et al., 2015). The impact of TbSkpZ was highly biased towards surface/endosomal membrane proteins as well as some of the machinery potentially associated with endocytosis, specifically the 6 VAMP7b SNARE protein (Tb927.5.3560) ( Tb927.11.1750, also present only in kinetoplastida and which lacks an obvious TMD or GPI-anchor signal and a location at the Golgi complex, was also decreased.
Most significantly, TbSkpZ silencing impacted surface protein levels depending on their mechanism of membrane attachment. Specifically, proteins with a predicted TMD were decreased in abundance, while those with a predicted or known GPI-anchor increased. The decrease in abundance is most likely due to the absence of DUB activity via loss of TbUsp7, with a resulting increase in lysosomal delivery of ubiquitylated proteins.
Significantly, the observed changes are near identical to those previously described for TbUsp7 ( Figure 4), strongly supporting the possibility that both proteins function in the same pathway. Further, many TMD proteins affected are enriched in the T. brucei surfacelabeled proteome (Gadelha et al., 2014) suggesting that turnover of these surface proteins is controlled through ubiquitylation (Table 1).

TbTpr86 is a pan-kinetoplastid protein: TbTpr86 protein abundance is affected by TbSkpZ
RNAi and is a TbSkpZ interactor by affinity isolation. Hence, we chose to investigate TbTpr86 in more detail. Initially, we performed comparative genomics screen using the Tpr86 sequence as BLAST query at EuPathDB, NCBI as well as a HMMER search. The TbTpr86 gene is well conserved and syntenic across the kinetoplastida and extends into the bodonids ( Figure 5, Table 3). There is, however, no evidence for the protein outside of the lineage indicting an origin post-speciation from the Euglenids. Molecular modelling, using the Phyre2 server, indicates with high confidence that TbTpr86 adopts an a-helical solenoid structure along much of its length ( Figure 5).
We established an endogenously HA-tagged TbTpr86 cell line and cells were harvested and cryo-milled as before. We used essentially the same conditions as developed for the TbSkpZ affinity isolation to identify coenriching proteins with LCMSMS (Table S3). We found strong interactions between TbTpr86, TbUsp7 and TbSkpZ ( Figure   5), robustly confirming the presence of a heterotrimeric complex. We designated this complex TUS, based on a composition of TbTpr86, TbUsp7 and TbSkpZ. GRESAG4 and PFC17, enriched in the TbSkpZ affinity isolation, were not detected and confirmed as contaminants. We only detected additional proteins with significantly lower enrichment suggesting that additional interactions were not well retained under the conditions used.
TbSkpZ is required to support vesicle transport. Transmission electron microscopy revealed a significant accumulation of intracellular membranous structures in TbSkpZ silenced cells, (Figure 6). These structures appear in the region of the cell usually associated with the Golgi complex and endosomes ( Figure 6B, C) and also appeared to 7 be associated with the endoplasmic reticulum ( Figure 6D) and entirely consistent with localisations for TUS components at TrypTag. We also observed pronounced multi vesicular bodies/autophagosomes ( Figure 6D), but no obvious impact to other structures.
The contiguous appearance of novel membrane structures suggests a failure to complete budding and hence accumulate vesicular structures.

Discussion
We describe here the TUS complex, which is comprised of the trypanosome ortholog of USP7, a pan-eukaryotic deubiquitinase, together with kinetoplastida-specific proteins SkpZ, and Tpr86. TUS complex interactions are robustly identified by reciprocal isolations using all three subunits as well as in trans impacts on abundance following silencing. Significantly, genes encoding all three proteins can be identified from nearly all kinetoplastid genomes, suggesting that the TUS complex is present throughout the lineage and arose early in the evolution of kinetoplastida following speciation from the euglenids.
TUS regulates the cell surface proteome, with significant impacts on both TMD and GPIanchored proteins, but significantly affecting these protein cohorts differentially; TMD proteins are down-regulated when TUS subunits are silenced and GPI-anchored proteins increase in abundance. It is possible that the increase in GPI-anchored protein abundance is a secondary consequence; endocytosis is clearly affected by RNAi against TUS components and in turn GPI-anchored nutrient receptors are upregulated. Further, the complex is associated with the endosomal region of the trypanosome cell, and which suggests that TUS functions to deubiquitylate TMD proteins as they enter the endosomal system, permitting their recycling -silencing prevents this action and increases turnover of TMD proteins (Zoltner et al., 2015). An impact on trafficking machinery proteins, as well as MBAP, is also significant as both participate in endocytic pathways. The increase in GPIanchored protein abundance may be the result of a simple counting/density sensing mechanism that permits an increase in these proteins due to decreased density of TMD proteins. However, as the GPI-anchored VSG is sorted from other endocytic cargo within sorting endosomes by a clathrin-dependant mechanism, it is also likely that there is a direct impact on the efficiency of this process. However, how GPI-anchored proteins are specifically recognised in trypanosomes is unclear albeit that this is likely to be distinct from mechanisms in animals and fungi. Regardless, it is clear that the TUS complex has a major impact on the cell surface proteome and that TUS is an adaptation that occurred early within the kinetoplastida. It is tempting to speculate that this may be associated with the high level of surface GPI-anchored proteins and glycoconjugates in these organisms.

8
In human cells USP7 is a high abundance DUB with many roles, including control of the cell cycle via p53. USP7 is subject to complex regulatory mechanisms and itself ubiquitylated with sophisticated multi-domain architecture involving a self-activation mechanism and several Ubl domains. USP7 can also act as a deSUMOylase (Lecina et al, 2016). Significantly, of many known animal and fungal USP7-interacting proteins, none were identified here and the vast majority are also unlikely to be encoded in any kinetoplastid genome, indicating significant diversity in function. USP7 forms a cellular switch with E3 ligases to facilitate rapid responses to stimuli, coordination of protein level and multiple functions, including endocytosis (Kessler et al, 2007, Hao et al, 2015, Kim and Sixma 2017. This seems unlikely the case in trypanosomes as we failed to identify any E3 ligase, associated subunits or evidence for association of ubiquitylation machinery from isolations of members of the TUS complex. TbTpr86 and TbSkpZ are trypanosome specific, but the known roles of Skp1 proteins as substrate adaptors in association with cullin and Rbx1 containing E3 ligases suggests a possible similar role. While speculative, structure modelling for TbTpr86 confidently predicts formation of an extended α-helical rod, consisting of repeat domains, which is at least conceptually similar in architecture to cullin proteins. Hence, we suggest that the TUS complex may be an 'anti-cullin', whereby TbTpr86 provides a scaffold to facilitate TbSkpZ binding substrate and delivering it to TbUsp7 for deubiquitylation, in contrast to the opposite reaction mediated by cullin ligases (Figure 7). Hence, while not intimately associated with an E3 ligase as in metazoan organisms, USP7 may none-the-less function in kinetoplastids as part of a switch mechanism providing highly sensitive and rapid control of surface protein expression.

TCTTGGCGTGGAGAATGACTTCAAGGCTGAAGAAGAGGCTGAACTCAGGAAAGAGTA C G G A A G A AT G T C A G A G G A G A A G G G Ta c c G G G c c c C C C c t c G A G R e v e r s e :
ATCCACTAGTTCTAGAGCGGCCGCCAACATGAGGGTGTGAGGCACACTTGTTTTTGC CGATGTGCGCGTATTCGAGAACCAGGTGTGGCGCGTGTTGACGG T b T p r 8 6 F o r w a r d : with Program X-001. Cells were then transferred to tube A containing 30 ml of HMI-9 medium plus any appropriate antibiotic drug for parental cell growth. Serial dilution was performed by transferring 3 ml of cell suspension from tube A into tube B containing 27 ml of HMI-9 medium and repeated again by diluting 3 ml from tube B into tube C. One millilitre aliquots from each dilution were distributed between three 24-well plates and incubated at 37°C. After 6 hours HMI-9 containing antibiotic selection was added to the wells at the desired final concentration. Transformed cells were recovered on days five to six posttransfection. For procyclic cells, 3 x 10 7 cells per transfection were harvested at 4°C, washed in cytomix and resuspended in 400 µl cytomix. Electroporation was performed with 5-15 µg of linearised DNA using a Bio-Rad Gene Pulser II (1.5 kV and 25 µF). Cells were transferred to 9.5 ml SDM-79 medium and incubated for 16 hours after which selection antibiotics were added. The cells were then diluted into 96-well microtiter plates.

TGCGAGGTGCATCTTTACCCTGGGGTTGACGCATTGTTGCTGTCGTTGATGGCTTACT G C C G C T T T A A T T G G G A G C
Positive transformants were picked into fresh selective medium 10-15 days post transfection. Brj58 on ice. The sample was sonicated and centrifuged for 15min, 13000g at 4°C. The supernatant was incubated with anti-HA magnetic beads for two hours at 8°C. After three times washing using buffer A with 0.01% Brij58, the beads were incubated with one times SDS buffer and the supernatant collected for SDS-PAGE as described above for further MS analysis, except samples were run on an OrbiTrap Velos Pro (Thermo Scientific) mass spectrometer. All proteomics manipulations were performed using LoBind tubes (Eppendorf) for efficient protein extraction and removal from each tube.

Preparation of specimens for conventional (resin-embedded) transmission electron
microscopy: Cells were pelleted and fixed in 0.1 M Na cacodylate buffer (pH 7.2) containing 4% paraformaldehyde and 2.5% glutaraldehyde for 60 mins, at room temperature. Cells were stained with 1% OsO4 with 1.5% sodium ferrocyanide in cacodylate buffer for 60 min followed by 1% tannic acid in 0.1M cacodylate buffer for 1hr and 1% uranyl acetate in acetate buffer for 1hr. Fixates were dehydrated with a 50% to 100% ethanol series followed by 100% propylene oxide. Embedding in Durcupan resin followed standard procedures. Sections were cut on an ultramicrotome at 70-100nm and stained with 3% uranyl acetate followed by Reynolds lead citrate. Grids were imaged on a JEOL 1200EX TEM using an SIS camera.
Protein electrophoresis and immunoblotting: Proteins were separated by electrophoresis on 12.5% SDS-polyacrylamide gels and then transferred to polyvinylidene difluoride (PVDF) membranes (Immobilon; Millipore) using a wet transfer tank (Hoefer Instruments).
Non-specific binding was blocked with Tris-buffered saline with 0.2% Tween-20 (TBST) supplemented with 5% freeze-dried milk and proteins were detected by incubation with 12 primary antibody diluted in TBST with 1% milk for 1 hour at room temperature. Antibodies were used at the following dilutions: mouse monoclonal anti-HA (sc-7392, Santa Cruz) at 1:10,000, rabbit monoclonal anti-myc (7E18, Sigma) at 1:5000, mouse monoclonal anti-v5 (37-7500, Invitrogen) at 1:1000. Following three washes of 10 minutes with TBST, the membrane was incubated in secondary antibody diluted in TBST with 1% milk for 1 hour at room temperature. Commercial secondary anti-rabbit peroxidase-conjugated IgG (A0545, Sigma) and anti-mouse peroxidase-conjugated IgG (A9044, Sigma) were used both at 1:10,000. Detection was by chemiluminescence with luminol (Sigma) on BioMaxMR film (Kodak). Densitometry quantification of relative protein level was achieved using ImageJ software (NIH).
Immunofluorescence (   following the taxon name refer to accessions given in Table S1.    Abundance shifts after 48 hours silencing of SkpZ and Usp7 (Zoltner et al., 2015) were plotted against each other. Selected protein groups in the cohorts are labeled and color coded as green; predicted trans-membrane proteins, orange; predicted GPI-anchored, red; TUS complex component. Note that the proteome analysis of the SkpZ RNAi lysate is deeper as extracts were analysed on a more advanced mass spectrometer accounting for the absence of some proteins from the Usp7 data set.    In contrast to cullin 1, we propose that TUS is responsible for removal of ubiquitin and 19 which is supported by the decrease in abundance of trans-membrane domain surface proteins following individual knockdown of all TUS complex subunits.  Tables S1 and S2: Accession numbers for protein sequences included in phylogenetic analysis.