Endosomal vesicle fusion machinery is involved with the contractile vacuole in Dictyostelium discoideum

Contractile vacuole (CV) complexes (CVCs) are morphologically diverse, but functionally comparable, osmo-regulatory organelles found in diverse freshwater protists (Patterson, 1980; Plattner, 2013). There are descriptions of CV structures in freshwater members of nearly every major eukaryotic lineage (Allen and Naitoh, 2002). Despite their morphological variation, unifying features of CV systems are apparent. In general, the CV system is formed from tubular-vesicular membrane structures, called the spongiome, carrying high levels of the vacuolar ATPase (V-ATPase). Under hypo-osmotic conditions, the V-ATPase creates an osmotic gradient across the CVC membrane driving the removal of water from the cytosol. Fusion of these tubular-vesicular compartments creates a large water containing vacuole. This subsequently undergoes fusion with the plasma membrane, thereby removing water from the cytosol and combating the effects of the hypo-osmotic environment (Allen and Naitoh, 2002; Plattner, 2013). This cycle is repeated, giving the appearance of pulsatile vacuolar swelling and contraction.

Beyond these gross morphological descriptions of various CVCs, our understanding of the molecular basis of CVC biogenesis and function remains patchy.

Aspects of CVC function have been examined in a range of model systems, most intensively in Paramecium tetraurelia and Dictyostelium discoideum, as well as Trypanosoma cruzi. However, with the exception of Rab11 paralogues, which have been shown to be important for CV function in multiple systems (Harris et al., 2001; Ulrich et al., 2011; Niyogi et al., 2014), these studies largely assess different protein families and pathways. This currently limits our ability to draw comparative conclusions. Their membrane-bounded nature and the common requirement for Rab11 does, however, suggest that CVCs have arisen through adaptations of the existing eukaryotic endo-membrane system, likely the endocytic and vacuolar branches specifically. Whether these organelles represent a conserved eukaryotic feature or multiple examples of convergent evolution remains an open question.

Extensive molecular characterisation and comparative genomics has revealed the common origin of the bulk of the membrane trafficking machinery found in extant eukaryotes (reviewed in Dacks and Field, 2018). This comparative evolutionary cell biological approach relies on broadly comparable inventories of organelle-associated gene products and knowledge of their functions across distantly related eukaryotic systems. As much of our understanding of endomembrane system organization is derived from studies in yeast and metazoan systems, organisms without CVCs, the data from which to base comparative studies of CV biogenesis and function is currently lacking. Furthermore, the diverse morphologies and kinetics of CV function, together with the paucity of genetic tools in many protistan lineages expressing these organelles make classical molecular cell biology approaches challenging. We therefore set out to use more broadly applicable transcriptomic approaches for the identification of a cohort of CV-associated genes. In this instance, we employed the best characterised and most experimentally tractable CV-expressing system, D. discoideum, in order to evaluate the relevance of the transcriptome changes observed. We show that a large-scale transcriptome response to hypo-osmotic stress includes the upregulation of nearly all known CV-associated genes in D. discoideum. Focusing on genes encoding proteins associated with the membrane-trafficking machinery, we observed that Rab GTPases and SNARE proteins were notably upregulated, whereas vesicle coat and multi-subunit tethering complex components were not. Moreover, of the differentially expressed genes, those paralogues associated with the late secretory or endocytic system were enriched versus those for the early secretory system. Furthermore, we demonstrate that this approach has predictive potential through the identification of a previously uncharacterized plasma membrane localised SNARE protein, which was found to be enriched at sites proximal to CVs. Finally, we provide phylogenetic evidence that CV-associated Rab11 paralogues from D. discoideum and T. cruzi represent lineage specific expansions of this protein family, and thus the involvement of specialized Rab11 paralogues in CV function represents a convergent evolutionary phenomenon.

In order to examine the transcript-level response of CVC system induction, we subjected D. discoideum cultures to hypo-osmotic conditions and made a comparison to cultures in isotonic medium. We analysed transcriptomes of D. discoideum (DH-1) under normal growth conditions (HL5 medium) and following 1 h of hypo-osmotic shock (H₂O). Samples were generated in biological triplicates and transcript abundance assessed by RNA-sequencing (RNA-seq). In total, abundance profiles were obtained for over 13,000 transcripts. In line with previous findings (Na et al., 2007; Mathavarajah et al., 2018), differential expression analysis using a false discovery rate (FDR) cut-off of <0.01 revealed that hypo-osmotic shock led to a large-scale alteration of transcript abundance within this timeframe. We identified over 2200 and 1900 downregulated and upregulated transcripts, respectively (Fig. 1A; Table S1).

To assess whether genes associated with CV function were under transcript-level control, we generated a test set of genes previously characterized in D. discoideum as being implicated in CV function (Table S2). Examining our comparative transcriptome dataset, we saw that the majority of genes within this test set (16 of 18) were significantly upregulated during hypo-osmotic stress (FDR <0.01) (Fig. 1B). The high level of concordance in both the appearance of transcript regulation and its direction in this test set strongly supported our premise that comparative transcriptomics could reveal genes and larger networks of genes involved in CV biogenesis and function.

As we sought to obtain insight specifically into CV-associated genes, we focused our further attention on the upregulated gene set. As has been previously reported (Mathavarajah et al., 2018), analysis of significantly enriched GO: Biological Process terms identified within the upregulated gene-set revealed a number of terms associated with ‘cellular catabolic process’, predominantly protein ubiquitylation pathways (Fig. 1C; Table S3, yellow rows) and cyclic AMP-related signalling pathways (Fig. 1C, Table S3, blue rows). In addition, and perhaps due to our increased sampling depth, we identified a substantial proportion of GO terms relating to intracellular membrane trafficking (Fig. 1C). Finally, analysis of significantly enriched GO: Cellular Component terms revealed a striking number of terms relating to the vacuole, endosome, Golgi apparatus and SNARE complex, again supporting the validity of our approach for the identification of membrane trafficking machinery with potential relevance to CV function.

As we are interested in the biogenesis of the CV and the conservation of this process, we next examined in more detail the transcriptional control of membrane trafficking machinery during hypo-osmotic stress. We took two approaches, firstly by broadly categorizing the identified upregulated membrane trafficking-associated genes by their predicted site of action. Secondly, to obtain a more broadly comparable picture of the membrane trafficking machinery, we identified D. discoideum homologues of four evolutionarily well-conserved membrane trafficking machinery families and examined their expression levels between our experimental conditions.

Our identified upregulated gene set contained ∼130 genes annotated as having a function related to intracellular membrane trafficking (Fig. 2A; Table S4). Notably, these genes included several (e.g. Rab 1A, Rab8a, Rab11a, Rab14, Sec1, Copine, Drainin, Disgorgin, LvsA and LvsD) that had been previously implicated in CV function. We then categorized these genes by their predicted site of action (Fig. 2A). As would be expected, based on the broad characterisation of CVs as having an endo-lysosomal origin, we found that the most highly represented localisation was endosomal or vacuolar. This finding gives us confidence that we are indeed seeing transcriptional regulation of genes whose products are important for CV function. Of note, within this set of endosomal and vacuolar genes, there was a strong enrichment of ESCRT components (Chmp2a1, Chmp2a2, Chmp4, Vhmp5, VPS28 and the putative ESCRT0 analogue Tom1). Previous analysis (Mathavarajah et al., 2018), and our current data, suggest upregulation of ubiquitylation machinery in D. discoideum, and it is possible that this is reflected in the observed significant modulation of the ESCRT machinery. We saw relatively few genes associated with the biosynthetic pathway [Golgi and endoplasmic reticulum (ER)] supporting the notion that the majority of CV membrane is derived from the endolysosomal system.

We carried out a secondary analysis of members of four major classes of membrane trafficking machinery, namely, Rabs, SNAREs, coat proteins and tethers. Together, these proteins make up a well-characterised and functionally well-conserved core of the membrane trafficking machinery. D. discoideum genes belonging to these classes were identified, and their expression levels assessed across our experimental conditions (Table S5). Overall comparison between these classes revealed interesting differences. Expression of coat protein and multi-subunit tethering complex genes appeared to be largely unaffected by hypo-osmotic stress, with the possible exception of the Class-C HOPS tethering complex. Four subunits of the six that make up this complex were found to be mildly downregulated. Conversely, expression of multiple SNARE and Rab protein genes appeared to be upregulated during hypo-osmotic stress. The requirements for various Rab proteins during CV biogenesis and function in D. discoideum have been explored previously (Bush et al., 1996; Harris et al., 2001; Du et al., 2008; Essid et al., 2012). Among the upregulated Rab genes were those encoding the known CV-associated Rab11a, Rab8a and Rab14 (Fig. 2B). In addition to these known CV associated Rabs, the majority of upregulated Rabs were those associated with endosomal trafficking. These included Rab5b, Rabs7b and Rab7b, and Rab32a, which have been previously implicated in CV function in T. cruzi (Ulrich et al., 2011; Niyogi et al., 2015), and several Dictyostelium-specific Rab7 paralogues (Porfírio-Sousa et al., 2022) (Fig. 2C; Table S6). These Rab7 paralogues would be predicted to function in later endosomal compartments, but this prediction has not been tested. Overall, there appeared to be a balanced upregulation of Rabs acting at both early and late endosomal compartments.

SNAREs showed a similar pattern to the Rab proteins. Both Vti1B and, to a lesser extent, VAMP7B have been shown to localise to CVs of D. discoideum, although in the case of VAMP7B the abundance at the CV was lower than that found at endosomal membranes (Wen et al., 2009). To date, there is no data available on which SNAREs are required for CV function in D. discoideum. We saw upregulation of Vti1B but not VAMP7B (Fig. 2B). In addition to Vti1B, we identified a total of 17 SNARE genes with upregulated transcripts during hypo-osmotic stress. Among these it is interesting to note that each of the four SNARE subfamilies (Qa, Qb, Qc and R SNAREs; Fasshauer et al., 1998; Malsam et al. 2008) is approximately equivalently represented (Fig. 2C; Table S7). As one of each of these subfamilies of SNAREs is required to come together to produce a fusion competent SNARE complex it seems likely that hypo-osmotic stress is inducing upregulation of specific SNARE complexes. The majority of upregulated SNAREs were orthologues of those shown to be involved in post-Golgi endosomal trafficking pathways in other systems, including among others syntaxin 16, Vti1A and Vti1B, syntaxin 6 and syntaxin 8, with little to no upregulation of early biosynthetic pathway SNARE orthologues, such as Sec20, Sec22, Use1A and Bet1A (Table S6). In terms of cognate endosomal SNARE complexes, we saw upregulation of both early (Syx7, Vti1A and Vit1B, and Syx6) and late (Syx8 and Syp7) members with the exception of the R SNARE VAMP7. We did, however, note upregulation of two homologues of the R SNARE Ykt6. Although VAMP7 is predominantly associated with endosomal fusion pathways, Ykt6 is a promiscuous R SNARE that is able to compensate for VAMP7 when VAMP7 level are insufficient (Davis et al., 2021). Thus, we saw upregulation of potential fully functional SNARE complexes associated with early and late endosomal membrane fusion events. In addition to the endosomal SNAREs, we also saw upregulation of a cognate SNARE complex that would be predicted to act in secretory events at the plasma membrane (Syn1B, NPSNB, Syp7A and Syb1). This predicted SNARE complex is based on studies from plant systems and has not been characterised in D. discoideum (El Kasmi et al., 2013).

NPSNB belongs to a poorly characterised family of SNARE proteins lost among animals and fungi, but that is widely conserved in other eukaryotes (Venkatesh et al., 2017). The NPSN SNARE family is currently only characterised in Arabidopsis, where they localise predominantly to the plasma membrane (Uemura et al., 2004). The upregulation of NPSNB in our dataset led us to ask whether its plasma membrane localisation is conserved in Dictyostelium and whether it might be associated with CV release.

To assess the localisation of NPSNB, we generated N-terminal fluorescent protein fusions and expressed them in D. discoideum AX2 cells via a non-integrating plasmid vector. Both GFP-tagged and mCherry-tagged NPSNB were predominantly plasma membrane localised (Fig. 3A; Fig. S1). The plasma membrane distribution was not homogenous, with enrichment of fluorescent signal in small regions of the plasma membrane adjacent to phase-light structures. This was particularly noticeable in cells with lower expression of the transgene (Fig. 3A). Given the increased expression of NPSNB upon CV induction, we reasoned that these structures might be plasma membrane-adjacent CVs. We therefore carried out live-cell imaging of NPSNB–GFP-expressing cells to monitor the behaviour of these NPSNB-associated phase-light structures. This showed that plasma membrane NPSNB foci were dynamic, and their appearance and disappearance corresponded to sites and times of close association between the plasma membrane and CVs (Fig. 3B; Movie 1). We then sought to quantify this association. For each GFP–NPSNB-expressing cell with a visible CV, we traced the cell periphery and identified the centroid of the CV (Fig. 3C). The fluorescence intensity around the cell periphery was measured (Fig. 3D) and plotted against the distance from the CV centroid (Fig. 3E). From these data, we extracted a mean fluorescence intensity value for each 0.5 μm of distance from CV centroid (Fig. 3F). By measuring 30 cells in this way, we were able to establish the normalised NPSNB fluorescence intensity distribution relative to the CV centroid across our cell population (Fig. 3G). We saw a clear inverse relationship between NPSNB intensity and distance from the CV centroid (Fig. 3G). We therefore suggest that NPSNB is enriched at the plasma membrane adjacent to CVs.

Beyond the cell biology in D. discoideum specifically are evolutionary questions regarding CVs across eukaryotes. On the one hand, organelles classified as CVs share an osmoregulatory function and a similar physiological cycle. By contrast, the taxonomic distribution of CVs found in freshwater members of disparate lineage with no observable CVs in non-freshwater members, as well as differences in morphological features of the organelles, leaves open the question of whether the CVs of different eukaryotes are retained from an ancient homologous CV organelle or are convergent in nature, derived independently from endolysosomal organelles in different eukaryotic lineages. Homology between a wide range of endomembrane organelles has been demonstrated most conclusively through identification of conserved machinery involved in organelle biogenesis and function across the eukaryotic diversity.

Rab11 proteins have been implicated in CV function in both T. cruzi and Dictyostelium. However, both Dictyostelium and Trypanosoma possess multiple paralogues of Rab11 (Elias et al., 2012). If the Rab11 paralogues that act at the organelle formed a clade containing both Dictyostelium and Trypanosoma orthologues, to the exclusion of the Rab11 paralogues that do not act at the CV, this would be strong evidence for a CV in the common ancestor of the taxa with organelle-specific paralogues, and thus homology of the CVs. If on the other hand, the Rab11 paralogues group by taxon, this would be most simply explained as the protein having been independently specialized to work at the organelle (i.e. convergence of CVs).

Phylogenetic analysis showed with robust support that the Rab11 paralogues group by lineage rather than by organelle (Fig. 4). Although there are several explanations for how these paralogues might have come to work at the organelle, this result means that although Rab11 paralogues can act at the CVs of Dictyostelium and Trypanosoma, they do not derive from a single primordial, pre-last eukaryotic common ancestor (LECA), CV-specific protein, and as such cannot be used as evidence for organelle homology between the Dictyostelium and Trypanosoma CVs.

One of the remaining open questions surrounding the evolutionary history of eukaryotic endomembrane organelles concerns the level of conservation versus convergence in CV origins. Answering this question has potential implications for our understanding of the cellular physiology of a huge diversity of freshwater protists of various medical, agricultural and economic importance, as well as potentially revealing details about the lifestyle of the LECA of all extant eukaryotes. In comparison with other widely distributed organelles of the endomembrane system, our understanding of the molecular mechanistic basis of CV biogenesis and function is patchy at best.

Indeed, there is, by our reckoning, only a single paralogous gene family whose involvement in CV function has been assessed in multiple evolutionarily divergent model systems, namely Rab11. Obviously, a single data point such as this is insufficient to base broad statement of organelle paralogy or otherwise. However, our phylogenetic examination revealed that CV-associated Rab11 paralogues from D. discoideum and T. cruzi both represent lineage-specific expansions of the Rab11 family and therefore is not suggestive of a hypothesis of deep organelle homology of CVs. This echoes the recent demonstration that a Qa-SNARE implicated in CV function in P. tetraurelia (Plattner, 2013) is the product of a gene duplication of a plasma membrane clade (Syntaxin PM) specific to one clade of ciliates (Kaur et al., 2022) and thus cannot be homologous to CV machinery across eukaryotes, or even all ciliates for that matter. Both data points are, nonetheless, consistent with the hypotheses of deep organelle homology (but with convergent elaboration) or of convergent evolution. Far greater sampling depth of CV-associated proteins from multiple organisms is required to expand the comparative evolutionary studies of this organelle.

The most widely used yeast and metazoan model systems do not have the capacity to form CVCs. Instead, these organelles are found in an evolutionarily diverse range or freshwater protistan lineages. Historically, studies of CV function have been carried out in relatively tractable ciliates and, for the most part, D. discoideum. However, to answer deep evolutionary questions, it is vital to sample as broad a range of organisms as possible. Most suitable organisms are not amenable to classical cellular and molecular biology approaches, due to inherent technical limitation or a lack of developed molecular biology tools. Therefore, approaches which do not require extensive genetic modification will be important in building a picture of the diversity of CVs across the eukaryotes. As such we have assessed the suitability of a transcriptomic approach using the relatively well described Dictyostelium model system.

There are some additional considerations and confounding variables that deserve recognition from the outset. Hypo-osmotic incubation and nutrient withdrawal have long been known to induce starvation responses in Dictyostelium, which in turn stimulate differentiation (Raper, 1940). Transcriptomic analysis of differentiation and development of Dictyostelium in response to nutrient withdrawal in a moderately hypo-osmotic environment has demonstrated wide ranging modulation of gene expression and established the temporal ordering of these responses (Rosengarten et al., 2015). Based on this data, we elected to use a very early timepoint (1 h post hypo-osmotic incubation) for analysis; at this point, the transcriptional response to starvation is somewhat minimised relative to later time points. It is clear, however, that any transcriptomic alterations we observe likely represent not only adaptation to hypo-osmotic conditions but also the beginning of the cellular response to nutrient withdrawal and the initiation of a differentiation and development programme. In recognition of this problem, we have not attempted to draw conclusions about overall cellular responses to hypo-osmotic stress specifically, but limited our focus to membrane trafficking machinery, which as a system is known to be required for CV biogenesis. Indeed, it has not previously been shown the extent to which CV induction is reflected by a transcriptional response. We were able to utilize existing knowledge of CV-associated genes to generate a reference set against which we could compare our observed transcriptional changes. We found that almost all of our reference set were identified as significantly upregulated during hypo-osmotic stress. Moreover, we identified a SNARE protein NPSNB as being upregulated during hypo-osmotic stress and showed that it is associated with sites of contact between CVs and the plasma membrane. This is an important and encouraging result that suggests that transcriptomic approaches might be more broadly applicable in the characterisation of CV-associated genes from organisms which remain, to date, entirely unexplored.

Employing this membrane trafficking-focused approach, we also saw interesting patterns in the regulation of membrane trafficking machinery as a whole. Vesicle coats and tethers remained largely unaffected, whereas SNARE proteins and Rabs appeared to undergo a large amount of transcriptional regulation. It is tempting to hypothesize that this might represent an increase in membrane flux, driven by increased rates of Rab-mediated membrane identity conversion and SNARE-mediated fusion, while maintaining the veracity of cargo sorting and organelle fusion through coat and tether driven recognition.

Overall, we saw a marked upregulation of Rab and SNARE proteins predicted to function at early and late endosomal membranes, and those associated with secretion at the plasma membrane. We therefore speculate that the CV membrane of D. discoideum is drawn from these sorting pathways with likely little involvement of membrane derived from the early biosynthetic pathway. We recently reported an apparent CV-related expansion of plasma membrane secretory SNAREs in ciliates, suggesting the involvement of these gene products in CV function (Kaur et al., 2022).

In summary, we have demonstrated regulation of CV-associated transcripts in D. discoideum, implicating specific sets of membrane-trafficking machinery in CV function and with implications for the organelle homology of the Dictyostelium CV. These general observations form a framework against which other organisms can be assessed, via this broadly applicable transcriptomic approach, coupled with focused assessment of membrane trafficking machinery, which has the potential to capture this regulation and provide predictive detail of novel CV-associated gene products.

For transcriptomic studies, D. discoideum DH1 (Dicty Stock Center, Northwestern University, Chicago, IL, USA) were grown axenically at 21°C in shaking culture at 150 rpm in HL5 medium, which has an osmolarity of ∼0.1 M (Gamper et al., 1999). For induction of hypo-osmotic stress, medium was replaced with Milli-Q reverse osmosis H₂O and cells cultured for 1 h at room temperature. CV induction and cell viability post hypo-osmotic exposure was confirmed by live-cell imaging (Movie 2).

The NPSNB gene (DDB G0285365) was amplified from Dictyostelium discoideum AX2 (Kay laboratory) genomic DNA using PCR. The forward (5′) primer (5′-GCAGTCAGATCTATGGCCGATTTACAAGAAAATG-3′) contained an additional BglII restriction site, and the reverse (3′) primer (5′-CT GACGACTAGTTTATTGGGTTGAACCAGTAACAT-3′) contained a SpeI restriction site. The product was ligated into non-integrating (extra-chromosomal) expression vectors (Veltman et al., 2009), amplified using Escherichia coli DH5α, and verified by Sanger sequencing. Specifically, we used the pDM1207 (for N-terminal GFP chimeras) plasmid, which drives overexpression of chimeras with the actin15 promoter and drive expression of a G418 resistance cassette with the cofillin A promoter.

D. discoideum AX2 (Kay laboratory) cells were cultured in Erlenmeyer flasks at 22°C suspended in liquid HL5 medium (Formedium, Hunstanton, UK; Ashworth and Watts, 1970) in an orbital shaker at 180 rpm. Amoebae were transformed via electroporation as described previously (Gaudet et al., 2007). For this, 3×10⁷ cells were harvested from liquid culture at a density of ∼1×10⁶–5×10⁶ cells/ml (log phase) by centrifuging at 300 g for 3 min, resuspended in H-50 electroporation buffer and put on ice. For each transformation, 5 μg of plasmid DNA was pipetted into a sterile 1 mm electroporation cuvette (at 4°C), to which 100 μl of cell suspension (also at 4°C) was added. This mixture was electroporated using a Bio-Rad gene pulser electroporator (exponential voltage curve, 750 V, 25 μF, ∞ Ω resistance). Two pulses were applied to the cells with 5 s between pulses. The cuvette was then set on ice for 5 min, and then the electroporated cells were incubated in HL5 medium in Petri dishes at 22°C overnight. Transformants were then selected using G418 at 10 μg/ml.

Prior to imaging, transformed D. discoideum AX2 cells were grown to confluence in plastic tissue-culture plates, with KK₂ buffer (16.5 mM KH₂PO₄, 3.9 mM K₂HPO₄ and 2 mM MgSO₄) containing a suspension of heat-killed Klebsiella pnumoniae at an optical density at 600 nm (OD₆₀₀) of 5. Also, 50 μg/ml tetracycline and 100 μg/ml dihydrostreptomycin were added to prevent contamination, and again 10 μg/ml G418 was added for selection. Imaging was performed using a Zeiss LSM 710 confocal microscope with a 63× (oil) objective lens. Time series (videos) were captured by sampling at 5 s intervals.

For measurement of the relationship between NPSNB fluorescence and CV positioning, a macro (available upon request) was written for FIJI to extract coordinates and associated fluorescence intensity from line plots drawn manually around the limiting membrane of the cell. The CV was identified and traced in corresponding phase images and the centroid was calculated (Fiji). These measurements were used to derive the spatial relationship between NPSNB distribution and CV position. In order to compare multiple cells with differing expression levels, NPSNB fluorescence was normalized to peak fluorescence along the line-plot and background subtracted. The mean of normalised fluorescence intensity for each 0.5 μm bin of distances was then calculated.

Total RNA from D. discoideum was extracted from six samples, with three biological replicates per condition, using the Qiagen RNeasy mini kit. RNA concentration was measured using a NanoPhotometer spectrophotometer. Samples were submitted to The Applied Genomics Core (TAGC), University of Alberta. Libraries were constructed using the Illumina TruSeq RNA Library Preparation kit and run on an Illumina NextSeq 500 system. Reads were aligned and quantified using the Galaxy web platform (https://usegalaxy.eu/). Reads were mapped to the D. discoideum genome (release 2.7) using HISAT2. Mapping quality was assessed using MultiQC. Aligned reads were assigned to genes (release 2.7.50) using FeatureCounts. Statistical analysis was carried out with Voom/Limma. A false discovery rate cut off of <0.01 was applied to assign differentially expressed transcripts. The raw reads for this project have been deposited in the NCBI Sequence Read Archive as BIOPROJECT PRJNA909995 (http://www.ncbi.nlm.nih.gov/bioproject/909995).

Amino acid sequences were aligned using MUSCLE v3.8.31 with default parameters (Edgar, 2004). Positions that did not exhibit clear homology were removed from the alignments prior to phylogenetic analysis. Models of sequence evolution were selected using ModelFinder (IQtree) (Kalyaanamoorthy et al., 2017). Maximum likelihood analysis was performed using IQtree (Nguyen et al., 2015). Bayesian analysis was performed using MrBayes (Ronquist and Huelsenbeck, 2003). For this, 10⁷ MCMC generations were run with a burnin of 25%, and a splits frequency of less than 0.01 was reached indicating convergence. IQtree and MrBayes analyses were run on the CIPRES server (Miller et al., 2010). The resulting trees were visualized using FigTree v1.4.0.

We are immensely grateful to Gareth Bloomfeld and Rob Kay (Laboratory of Molecular Biology, Cambridge) for hosting and supervising L.D.B. during the data collection of Dictyostelium NPSNB, for the mentorship that he received, and for the input on this manuscript. We wish to thank the many scientists at the LMB and UAlberta who provided access to shared and lab-based equipment that facilitated this research.