Skip to main content
NCBI home page
FEMS Microbiology Letters logoLink to FEMS Microbiology Letters
. 2025 Sep 3;372:fnaf088. doi: 10.1093/femsle/fnaf088

Golgi_traff phylogeny reveals ancient eukaryotic genes with recent surprises: replication and diversification of HID1 domain-containing protein unique to Schizosaccharomyces

Mark A Hooks 1,✉, Abdulrahman Alasmari 2, Mohammed Alshehri 3, Lysiane Brocard 4, Katarzyna B Hooks 5, Mathilde Julien 6, Ramsay J McFarlane 7
PMCID: PMC12416282  PMID: 40899782

Abstract

Golgi_traff is a Pfam clan containing two members, Dymeclin (DYM) and HID1 domain-containing protein (HID). Interrogation of over 900 eukaryotic genomes with sequence models showed that both are ancient eukaryotic genes, which have exhibited different paths of gene loss, including from major taxonomic groups. For example, the Metazoa have both genes, whereas the Viridiplantae and Dikarya have lost HID and DYM, respectively. A unique replication event occurred within the genus Schizosaccharomyces in that all sequenced species possess three HID-encoding paralogs, whereas its nearest fungal relatives and other eukaryotes are almost exclusively monogenic. A phylogenetic analysis of yeasts revealed that the Golgi-resident paralog Human ortholog 3 (SPAC17A5.16) is more similar to the HID of other yeasts than to its paralogs. Transmission electron microscopy revealed that the SPAC17A5.16 mutant lacks a stacked Golgi apparatus (GA) form, suggesting a role in maintaining GA structure. Altered proliferation of the SPAC17A5.16 mutant in response to GA disrupting chemical agents indicated a perturbation of GA-related functions. Structural models suggest SPAC17A5.16 has a long, disordered N-terminal region that may facilitate anchoring to GA membranes. A modification to Schizosaccharomyces HID nomenclature is proposed to reflect their evolutionary and functional characteristics. The potential of the Golgi_traff clan to serve as a model for the diversification of protein function according to the concepts of sub/neofunctionalization is discussed.

Keywords: dymeclin, Golgi apparatus, Hidden Markov Model, high-temperature-induced Dauer formation protein, neofunctionalization, Schizosaccharomyces pombe

A gene replication event has created a gene family unique to fission yeasts indicating a diversification of roles in Golgi apparatus function and cellular protein trafficking.

Introduction

HID1 domain-containing protein (HID1) was isolated in a screen to identify factors driving the switch of larvae into Dauer development in Caenorhabditis elegans (Ailion and Thomas 2003). The name DYMECLIN (DYM) originated from the human genetic disease Dyggve–Melchior–Clausen syndrome (El Ghouzzi et al. 2003). In animals, both DYM and HID1 localize to the Golgi apparatus (GA) network through N-terminal myristoylation, but retain a strong cytoplasmic presence due to dynamic retargeting (Dimitrov et al. 2009, Wang et al. 2011). Elimination or mutations of either DYM or HID can disrupt vesicular traffic and mistarget other proteins (Osipovich et al. 2008, Hummer et al. 2017). Mutation of DYM or HID1 can cause severe genetic disease in humans (Schänzer et al. 2021), but both have been postulated to provide protective effects against certain cancers (Harada et al. 2001, Aydin and Arga 2019). Thus, due to sequence similarity and corresponding cellular and biological features, the protein database Pfam has paired these proteins into the clan Golgi_traff (Mistry et al. 2021). Since work on both proteins has focused almost exclusively on animal systems, it would prove advantageous for functional studies to determine if they are ubiquitous to cellular organisms and to find models possessing one or the other protein. Therefore, we undertook a comprehensive interrogation of eukaryotic genomes to determine the presence of DYM and HID. This revealed a nonuniform distribution of each gene among eukaryotes including a very rare, if not unique, HID replication within the genus Schizosaccharomyces. For Schizosaccharomyces pombe, Pombase (www.pombase.org; Rutherford et al. 2024) lists three orthologs of human HID1. Human ortholog 1 (SPAP27G11.12) and Human ortholog 2 (SPBP19A11.07c) remain of unknown function, whereas Human ortholog 3 (SPAC17A5.16/Ftp105) is GA-associated and appears to be necessary to target the deubiquitinase Ubp5 to the GA (Kouranti et al. 2010). The appearance of a gene family within this single genus raises interesting questions regarding retention of duplicate genes in relation to the diversification of protein function. Our experimental and computational work suggests a potential role of SPAC17A5.16 in maintaining the stacked structure of the GA.

Materials and methods

Computational analyses of HID1 and DYMECLIN orthologous sequences

Initial Hidden Markov Models (HMMs) for each protein were based on seed sets corresponding the Pfam clan Golgi_traff (https://www.ebi.ac.uk/interpro/set/pfam/CL0456) that contained 17 and 56 sequences for HID and DYM, respectively. The DYM seed set was filtered by eliminating all HID sequences to yield a set of 28 true DYM proteins. The initial HID and DYM HMMs were created from these seed sets using HMMER v3.1b1 with default parameters (Eddy 2011) and based on MAFFT v7.49 alignments (Katoh and Standley 2013). These were used to interrogate the annotated proteomes of the species specified in Barlow et al. (2018). Potential HID and DYM sequences were confirmed by BLASTP and DELTA-BLAST (Boratyn et al. 2012). Alignments and HMMs were created as above for the expanded protein and corresponding coding sequences (CDS) sets (File S1) and used to interrogate 923 eukaryotic Refseq proteomes and genomes available at NCBI. For cDNA-based interrogations, hits with an E-value below the inclusion threshold <0.01 were concatenated if originating from the same sequence/scaffold, the identity confirmed by TBLASTX and aligned to the HMM profile and corresponding protein sequence to check completeness. Visualization of species from Barlow et al. (2018) was performed according to NCBI taxonomic numbering using the R package Taxize v 0.0.9.7 (Chamberlain and Szöcs 2013). Construction of the phylogenetic tree for fungal HIDs was done with RAxML-ng v1.0.2 (Kozlov et al. 2019). Tree topologies were visualized using Figtree v1.4.4 (Rambaut 2018). Calculation of amino acid identify between HIDs was done using MAFFT-based alignments. The divergence time for each of the three HID proteins for each species was derived from an independent tree. Each analysis comprised of one of the five Schizosaccharomyces orthologs, the Hid from Pneumocystis carnii, and ECM30 from Saccharomyces cerevisiae. Estimation of divergence times using RelTime-ML in MEGAX (Stetcher et al. 2020) was done with initial MAFFT alignments and best-fit RAxML-ng trees. Tree branch lengths were calculated using the Maximum Liklihood (ML) method and Jones-Taylor-Thorton matrix-based substitution model (Jones et al. 1992). Tests for positive selection were conducted using PAML v4.9 (Yang 2007). The guide tree was created in RAxML-ng v1.0.2 and corresponding coding sequence alignments were created using Revtrans v2.0 (Wernersson and Pederson 2003). Coding sequence alignments were visually inspected in MEGAX to remove gaps with no sequence. Predicted protein structures were created by AlphaFold2 downloaded from the AlphaFold Protein Structure Database (alphafold.ebi.ac.uk; Jumper et al. 2021) and pairwise and multiple structural alignments were conducted remotely at RCSB Protein Data Bank (RCSB.org; Bittrich et al. 2024). Modeling using either rigid (JFATCAT, jCE), flexible (jCE-CP), or reside-to-residue (TM-align) alignment algorithms yielded nearly identical results.

Generation of mutant strains and treatment with chemical stress agents

Mutant strains were generated from the standard multiauxotrophic strain (BP90: h-  ade6-M26 leu1-32 ura4-D18) by recombination-mediated gene replacement (Bähler et al. 1998). PCR products used for transfection were generated using the template plasmid pFA6a-natMX6 and selection for positive transformants was conducted using cloneNat (Werner Bioagents, Jena, Germany). The lists of strains and primers employed in the study are given in Table S2(A) and (B), respectively. For the chemically induced stress treatments, all strains were grown for 12 h in 5 ml YEL medium (0.5% w/v yeast extract, 3 0% w/v glucose, supplements at 225 mg/l) at 30°C with shaking at 200 rpm and diluted to a cell count of 5 × 105 cells/ml into 5 ml of fresh media containing the chemical agent at the concentration given in the legend to Fig. 4. Cells were counted after 12 h to calculate the generation time (GT).

Figure 4.

Figure 4.

Open in a new tab

The effect on mutants of pharmaceutical treatments potentially affecting GA function. The effect on strain GT was determined by counts of cells incubated in YEL with the corresponding agent over a period of 12 h. (A) The effect BFA. Cells were incubated at the specified concentrations of BFA for 12 h. DMSO was added to cultures of BP90 at the same volume used for adding BFA to treated cultures. (B) The effect of GCA. Treatment was done at 50 µg/ml GCA (dissolved in DMSO) for a period of 12 h with control cultures receiving only DMSO. Symbol representations: Isolate 1, open bars; Isolate 2. For all treatments, Student’s t-tests for unequal variance were used to determine statistical significance between all pairs of strains for a particular treatment. The number of samples for each strain encompassing both isolates 1 and 2 were ≥4. The error bars give the standard deviation and the horizontal lines the median values. Asterisks represent statistical significance in the following manner: *, P < .05; **, P < .01; and ***, P < .001.

Preparation of strains for transmission electron microscopy (TEM)

Cells for TEM were grown overnight in 5 ml of YEL under the conditions specified above. The cells were diluted to 5 × 105 cells/ml in fresh YEL medium and grown to a cell density of 5 × 107 under standard conditions at which point they were harvested. Cells were harvested from one ml of culture by centrifugation for 2 min at 1150 × g in a bench-top centrifuge. The cells were subjected to freeze substitution with on a Leica AFS2 Leica with 1% OsO4 in acetone. The substituted cell mixture was washed four times with acetone kept at ambient temperature. The cells were impregnated with increasing proportions of EPON resin as follows: 30% EPON in acetone for 12 h, 66% EPON in acetone for 8 h, 100% EPON for 12 h, and 100% EPON for 10 h. The cells in resin were polymerized in 200 µm flat molds at 60°C for 24 h. The blocks were cut using a Leica UC7 microtome and the slices incubated in a mixture of lead citrate/uranylacetate to increase contrast. Cells were visualized on a Tecnaï Spirit 2 FEI operating at 120 kV.

Results

Asymmetric distribution of HID and DYM within eukaryotes

Standard similarity searches by BLAST revealed that neither HID nor DYM were present in prokaryotes or archea and that eukaryotic organisms could have either, neither, or both proteins. Therefore, we undertook a systematic analysis of sequenced eukaryotic genomes to assess the distribution of both proteins. It was necessary to conduct this iteratively, whereby an initial HMM was created from the Pfam seed set for each of the proteins for interrogation of a small, but diverse, group of organisms for which GA proteins have been assembled (Barlow et al. 2018). Subsequently, protein and CDS HMMs were generated from this group and used to interrogate the available eukaryotic Refseq genomes and proteomes. The use of both protein and DNA interrogation of proteomes and genomes served to reveal the number of protein isoforms and the number of genes, respectively.

The topology of the taxonomic tree derived from NCBI numeric identifiers for the small group of organisms was nearly identical to that shown in Barlow et al. (2018) and presented a diverse set of eukaryotic organisms to visually present the presence or absence of HID and DYM (Fig. 1; Table S2C). Both DYM and HID were represented across clades, but exhibited distinct patterns of presence within the various phyla. It revealed the lack of DYM in the Fungi, Ascomycota and Basidiomycota, and in a large group of the Alveolata. HID was absent from certain phyla of the Alveolata, the clades Haptista and Rhizaria, and the kingdom Viridiplantae.

Figure 1.

Figure 1.

Open in a new tab

Presence of HID and DYM in diverse eukaryotic organisms. The organisms were taken from Barlow et al. (2018), for which the proteins sequences for HID and DYM were through HMMER interrogation of Refseq proteomes and genomes and confirmed by BLAST. The topology of the taxonomic tree was established on the protein sequences in R using the TAXISE package. Color-coding of branches: black, neither HID nor DYM; cyan, HID only; orange, DYM only; dashed cyan–orange, both HID and DYM; gray, diverged S. cerevisiae orthologue; dashed orange–gray, HID-fusion, respectively. Species in red do not have a stacked GA structure.

More robust protein and cDNA HMMs were created from the species of Barlow et al. (2018) to overcome the limited, and often misannotated, Pfam seed sets. These were used to interrogate 923 eukaryotic Refseq proteomes and genomes present in NCBI on the date of retrieval. The global distribution of genes in the Refseq set mirrored that shown in Fig. 1 with the Metazoa having both HID and DYM, the Dikarya having HID, and the Viridiplantae having only DYM (Table S2D). The interrogation of Refseq proteomes revealed many species to have multiple DYM or HID protein accessions, whereas Refseq genome interrogations generally retrieved only one gene. In organisms, where duplicate genes were discovered, these could be traced to ploidy/genome duplication events, such as for the Blastocladiomycota (Amses et al. 2022) and the Mucuromycota (Ma et al. 2009). The more robust HMMs provided evidence through partial sequences for the presence of DYM or HID in various taxa and retrieved protist sequences that showed no similarity to proteins other than resident DYM or HID paralogs. The latter included examples of the explosion in pseudogene formation, such as that observed for Trichomonas vaginalis (Table S2C; Woehle et al. 2014).

Replication and retention of three Schizosaccharomyces HID genes

The S. pombe genome sequencing project (Wood et al. 2002) and subsequent Schizosaccharomyces sequencing projects had revealed three orthologs of HID1, although they have been misannotated as DYM for S. japonicus, S. cryophilus, and S. octosporus. Thus, the species of this genus are unusual among eukaryotic organisms as each has three distinct paralogs. A comparison of the S. pombe sequences with human and other fungal HIDs is given in Fig. S1. The evolutionary divergence of the genus was well-reflected in the HIDs. Schizosaccharomyces pombe and S. japonicus genomes share ~55% amino acid identity among proteins (Rhind et al. 2011), but SPAP27G11.12, SPBP19A11.07c, and SPAC17A5.16 were ∼34%, ∼20%, and ∼37% identical to the corresponding S. japonicus ortholog, respectively (Table S2E). Furthermore, there has been substantial diversification of primary sequence of S. pombe HIDs with those of later appearing species (Table S2E) and they have undergone shuffling with both chromosome location and orientation varying amongst S. japonicus, S. pombe, and S. octosporus (Fig. S2).

To understand better the relationship among the three paralogs and the orthologs of other fungi, we conducted a phylogenetic analysis, which comprised 84 HID proteins from the fungal species of Barlow et al. (2018) and Liu et al. (2009) that were present also in the set of Refseq genome set. For all Schizosaccharomyces species, Ortholog 3 appeared to be more similar to the HID of the Pneumocystis than to either of its paralogs (Fig. 2; Liu et al. 2009). Despite the diversification of sequences, the Schizosaccharomyces HIDs grouped according to ortholog with Ortholog 2 being most divergent. This was investigated further by estimating the divergence time among the five Schizosaccharomyces species according to protein orthologs using RelTime in MEGAX (Stetcher et al. 2020). By rooting each tree to ECM30 of S. cerevisiae and applying only the node calibrations of the Pneumocystis and Schizosaccharomyces split at 467 million years ago (MYA) (Beimforde et al. 2014) and the S. octosporus/S. cryophilus at 29.4 MYA (Jia et al. 2023), Ortholog 2 split at ~395 MYA, and orthologs 1/3 split between 291 and 363 MYA (Fig. S3). In addition, the branch-site model of PAML was employed to test for positive selection among each of the three Schizosaccharomyces hids, whereby, in turn, the branch for each ortholog was selected as foreground with the background comprised of hids from within the taphrinomycotina fungal subphylum. The guide tree is shown in Fig. S4. For a description of the models and their output, we refer you to the reference Zhang et al. (2005). For the clade containing SPAC17A5.16, 83% and 16% of sites showed positive selection in relation to conserved (M0) and neutrally evolving codons (M1), respectively. Comparison to the null model with the dN/dS ratios of M2 set to 1, yielded a 2ΔLnL (2x the difference in log-likelihood values) of 47 with a probability value for the log ratio test (LRT P-value) of 0 (Table S2F). This suggests the clade to be under positive selection. The subsequent Bayes Empirical Bayes analysis identified two potentially, positively selected codons (P > .95) corresponding to amino acid positions 539S and 585R. The corresponding analyses for the coding sequences of SPBP19A11.07c (2ΔLnL = 18.7, P-value = 1.5 × 10–5) and SPAP27G11.12 (2ΔLnL = 21.9, P-value = 2.9 × 10–6) also suggested clades under positive selection, but no specific sites were identified.

Figure 2.

Figure 2.

Open in a new tab

Phylogenetic representation of the evolutionary relationships among fungal Hids. Sequences were aligned using MAFFT and included 84 taxa and was comprised of 2351 amino acid positions. The values in parenthesis refer to the number of species grouped within the given genus. The maximum-likelihood tree was constructed using RAxML-ng and a LG + Gamma(4 m) model. The scale bar (dashed rectangle) gives the apparent rate of amino acid substitution among species. The support values superimposed onto the best-fit tree represented 800 bootstrap replicates (two-set convergence at a weighted Robinson–Foulds value = 1%). The tree was rooted to the fungal DYM sequences shown and the branch lengths between the HID and DYM clades shortened by a substitution rate = 1.0 for the purpose of tree presentation.

GA stacking appears lost in S. pombe lacking human Ortholog 3

Within the set of organisms employed by Barlow et al. (2018), they included 12 that do not have a classic stacked GA structure (Fig. 1). Interestingly, 10 of these 12 organisms did not have HID or had a greatly modified sequence, such as the ECM30 of S. cerevisiae (Fig. S1). Since SPBP19A11.07c and SPAP27G11.12 do not appear to be GA-associated (Matsuyama et al. 2006), SPAC17A5.16 was the likely candidate to contribute to GA structure. To test this idea, TEM was conducted on freeze-substituted cells of the spac17a5.16Δ (Ortholog 3) mutant with BP90 (WT) and spap27g11.12Δ (Ortholog 1) mutant as controls (Fig. 3). For WT and spap27g11.12Δ isolates, a stacked GA-like structure with multiple cisternae was clearly visible and corresponded to previous reports (Osumi 2012, Eckler et al. 2013). In contrast, no stacked GA structure was observed in cells of spac17a5.16Δ isolates and the only corresponding structure was a single, enlarged and elongated structure with associated vesicles. This single structure was not observed in BP90 or spap27g11.12Δ isolates. To examine further the potential consequences of Hid removal, the various strains were tested for proliferation by pharmaceutical treatments used to perturb GA structure and function. In our hands, the commonly used multiauxotrophic strain, which we have referred to as BP90 (see Table S2A), has a GT of just under 4 h (3.9 ± 0.7, n = 10). This was ~2 h longer than we found for the wild-type strain (h- strain 972) (1.8 ± 0.6, n = 6), but both fell within the attributed GT for S. pombe (Vyas et al. 2021). Treating with increasing concentrations of Brefeldin A (BFA) led to an increase in GT for BP90 from a median time of 3.8–4.3 h, but not for either spac17a5.16Δ or spap27g11.12Δ (Fig. 4A). Strain spac17a5.16Δ without BFA had a median GT of 5.4 h, whereas that for spap27g11.12Δ was similar to BP90. Treatment with Golgicide A (GCA) resulted in increased GTs for all strains with spac17a5.16Δ being most profoundly affected (Fig. 4B). The median GT for spac17a5.16Δ increased by about 6 h compared to around 2 h for both BP90 and spap27g11.12Δ. Again, the GT for spac17a5.16Δ under nontreated conditions was about 2 h longer than that for BP90.

Figure 3.

Figure 3.

Open in a new tab

TEM of thin layer sections of S. pombe cells. EPON resin slices described in the section “Materials and methods” were incubated with lead citrate/uranylacetate to increase contrast and visualized on a Tecnaï Spirit 2 FEI operating at 120 kV. Cell sizes relate to the scales given for WT samples, except for the spap27g11.12Δ iso 2 enhanced magnification image. Abbreviations: CW, cell wall; ER, endoplasmic reticulum; N, nucleus; and V, vacuole.

Although there are no experimentally determined structures of either DYM or HID, the Computed Structural Models yield potential explanations for the targeting of eukaryotic HIDs, particularly the variation observed among those of S. pombe. The global topologies of human HID1 and DYM are very similar (Fig. 5A), which likely explain their colocalization, even when the GA is dissociated upon treatment with BFA (Wang et al. 2011). The major difference between the two proteins is the presence of a C-terminal region of ∼120 residues in HID1, for which AlphaFold2 was not able to determine a structural feature. This C-terminal disordered region is also present in SPAC17A5.16 (Fig. 5B). In addition, SPAC17A5.16 has a predicted N-terminal disordered region that is about 80 residues longer than that predicted for human HID1. The C-terminal disordered region appears to be present in SPAP27G11.12, but not the N-terminal disordered region, and both disordered regions are greatly reduced in SPBP19A11.07c compared to SPAC17A5.16 (Fig. 5C). The prediction that these features are also present in the corresponding paralogs of S. japonicus suggest that these features appeared and were modified before diversification of the species (Fig. S5). There is some difference in size of the disordered regions, but this would be expected noting that membrane proteins differ between the two species due to the different membrane fatty acid content (Makarova et al. 2020).

Figure 5.

Figure 5.

Open in a new tab

Comparison of Computed Structural Models for Golgi_traff proteins. All predicted models were downloaded from the AlphaFold Protein Structure Database. The alignments shown were produced by TM-align. Images have been positioned to present the disordered regions. (A) Alignment of HID1 (Q8IV36, orange) and DYM (Q7RTS9, blue). (B) Alignment of SPAC17A5.16 (O13776, orange) and HID1 (blue). (C) Alignment of the three Hids from S. pombe SPAC17A5.16 (orange), SPBP19A11.07c (Q9HDV3, blue), and SPAP27G11.12 (Q9P7M6, green).

Discussion

The extensive distribution for both DYM and HID, particularly encompassing single-cell eukaryotes, suggests a more fundamental roles in cellular/GA function than attributed by work on multicellular organisms. However, a major impediment to understanding their functions is knowing the breadth of organisms in which they are present. We sought to resolve this through a direct search of available eukaryotic Refseq genomes. Due to limited initial seed sets in Pfam and the frequent misannotation of HID or DYM within genome sequencing projects, it was necessary to conduct the search in an iterative process to obtain robust protein and CDS HMMs. For example, taxa representing the Chlorophyta, which contain the higher plants and algae were not present in the DYM pfam seed group even though they possess DYM, and almost half of the sequences in the pfam DYM seed set were HIDs. In addition, the usefulness of the two-prong approach to interrogate both the genome and annotated proteome of each organism was exemplified through the searches within the ciliate Stramenopiles. The proteome search revealed two HIDs and a DYM for Paramecium tetraurelia, a HID for Tetrahymena thermophila, and neither HID nor DYM for Ichthyophthirius multifiliis. In contrast, searches of the genomes for the corresponding genes using the CDS HMMs were negative, except for T. thermophila, for which a matching sequence was suitable to confirm the presence of the gene and for reconstructing the coding sequence for the HID. The lack of positive hits of DNA sequences for the P. tetraurelia and I. multifilliis were due to the presence of multiple internal eliminated sequences in the HID genes (Arnaiz et al. 2012) and the lack of HID, respectively. Multiple HIDs in P. tetraurelia and the lack of HID in I. multifilliis stem from a whole genome duplication event and genome reduction, respectively, compared to the genome of T. thermophila (Coyne et al. 2011). In contrast, the ability of the HMM-based search to retrieve partial CDS sequences, some of which could not be reconfirmed by BLAST, may reflect either the power of this approach to identify genes and pseudogenes in organisms with rudimentary annotation or to identify gene fusions and chimeras. Overall, this interrogation revealed the broad distribution and ancient eukaryotic origins of the Golgi_traff proteins, and which may have been GA-associated in accordance with the early development of the GA (Barlow et al. 2018). The two genes have diverged in their patterns of loss including from major taxonomic groups as shown by the loss of HID from the Viridiplantae and Haptista and the loss of DYM from the subkingdom Dikarya. Interestingly, DYM from Arabidopsis has nearly equal identity to human DYM as does DYM from C. elegans (Table S2E), but no role has been attributed to the Arabidopsis protein (Luhua et al. 2013). DYM has not been studied in the context of single-celled eukaryotes.

HID has been retained within the Dikarya as monogenic except for species, which have descended from whole genome duplication events, such as those of the Allomyces and Phycomyces. The presence of three HIDs within the genus Schizosaccharomyces appears to be unique as they likely arose from single-gene duplications, since the Schizosaccharomyces do not show evidence of genome duplication (Wood et al. 2002). We surmise that a gene ancestral to SPAC17A5.16/SPAP27G11.12 duplicated to yield SPBP19A11.07c and the encoded protein eventually lost its GA localization. Subsequently, the SPAC17A5.16/SPAP27G11.12 underwent tandem duplication to yield SPAC17A5.16 and SPAP27G11.12. This evolutionary sequence occurred before species divergence within the genus as each of the five sequenced species have all three HIDs and our estimated divergence dates fall well before the estimated split of S. japonicus and S. pombe at 207 MYA (Shen et al. 2020). Why three paralogs have been retained in each species of this genus remains an interesting question regarding gene duplication and subsequent diversification of protein function. When compared with other members of the taphrinomycotina, it appears that each has acquired a biological role to serve as a driving for of positive selection to retain each of the Hids despite genome rearrangements and sequence diversification. According to various models of gene retention (Innan and Kondrashov 2010), diversification of Schizosaccharomyces HIDs could include a role in GA structure for one of the paralogs. The key outcome of the TEM analysis was an inability to find a stacked structure in spac17a5.16Δ compared to other strains. The precise identity of the observed elongated structure observed in spac17a5.16Δ remains unknown, but possible explanations for this structure are a remaining cis-GA, where the medial- and trans-GA have dissociated, or the entire GA has collapsed into a single structure. The response of spac17a5.16Δ to both BFA and GCA indicate that GA related processes have been affected. Our result on the lack of sensitivity to BFA and the apparent resistance of the mutant to BFA in a genome-wide screen (Rodríguez-López et al. 2023) is consistent with the known effect of BFA to disrupt GA structure. Furthermore, the apparent hypersensitivity of spac17a5.16Δ to GCA may also reflect an altered GA structure. GCA would likely bind and inhibit the Arf-GEF protein Gea1, which is located at the cis-GA. Elimination of Gea1 does not affect GA structure, but septum formation is affected (Eckler et al. 2013). The dramatic increase in GT that we observe for spac17a5.16Δ in the presence of GCA may be due to disruption of any GA remnant in the mutant. The presence of the medial- and trans-GA may be exerting a protective effect regarding cis-GA integrity in SPAC17A5.16+ strains. The cytosolic redirection of Ubp5 in spac17a5.16Δ (Kouranti et al. 2010) may be due to the morphological disruption of the GA. If HID contributes to the maintenance of GA structure in S. pombe, this role may extend to other eukaryotes.

Our subsequent exploitation of recent advancements in protein structure prediction revealed features in SPAC17A5.16 that may contribute to GA localization. High confidence in the predicted structures stems not only from the high pLDDT scores for the structured regions, but the overall topology of each protein is similar to other membrane-associated proteins, such as the α-importins that are necessary for nuclear membrane formation (Hachet et al. 2004). The global topology of animal HID and DYM is a core of alpha helices that were initially believed to be membrane spanning. Subsequent experiments demonstrated that both proteins are peripheral and whose attachment is facilitated by N-terminal myristoylation (Dimitrov et al. 2009, Mesa et al. 2011). Although mutating the myristoylation site of HID1 disrupts localization of HID1 to the GA in humans, it is unlikely that myristoylation is solely responsible for targeting HID1 to the GA (Bhatnagar and Gordon 1997). For example, yeast Arf1 is N-terminal myristoylated and employs an N-terminal amphipathic helix of basic and hydrophobic residues to bind to the GA membrane (Antonny et al. 1997, Balali-Mood et al. 2005). All three S. pombe HIDs possess both the N-terminal myristoylation motif MGXXXS/T as part of an amphipathic helix. Since both SPBP19A11.07c and SPAP27G11.12 are not localized to the GA, another feature must be responsible for targeting SPAC17A5.16 to the GA. A likely feature is the extended N-terminal disordered region of SPAC17A5.16, which is not present in the other two proteins. Disordered proteins or regions have become key elements of investigation for protein–membrane interactions (Jamecna and Antonny 2021). At the start of the region is a polybasic sequence preceeded by tryptophan residues that may facilitate binding to the GA membrane (Khemaissa et al. 2021, Alavizargar et al. 2024). The analyses for positive selection revealed two amino acids 539S and 585R in SPAC17A5.16 that merit further functional characterization. Based on the predicted structure of the protein, both amino acids reside at hinge regions linking two α-helices. Amino acid 539S would be exposed an so a possible target of phosphorylation, whereas 585R may be interacting with an unstructured region comprised of amino acids 756–773.

Based on both our computation and experimental findings, we propose a modification to current nomenclature regarding Schizosaccharomyces HIDs. We suggest that SPAC17A5.16, SPBP19A11.07c, and SPAP27G11.12 be referred to as Hid1, Hid2, and Hid3. SPAC17A5.16 becoming Hid1 is logical as it is the paralog confirmed to be GA-localized, it is the paralog most similar in sequence to that of other fungi, and it appears to have retained structural features found in HIDs of more distant eukaryotes. Although SPAC17A5.16 is referred to also as Ftp105, renaming it to Hid1 would draw it into the currently proposed nomenclature system for fission yeast (Lera-Rameriz et al. 2023) and simplify future references to other reports on HIDs. SPBP19A11.07c and SPAP27G11.12 would become Hid2 and Hid3, respectively, according to the divergence model proposed above.

The Golgi_traff clan represents an attractive model to study homologous function in relation to protein evolution. It has only two members with small family sizes and a broad eukaryotic distribution. It exhibits the various combinations of HID and DYM retention to provide models for unique functions and potential interactions. For example, noting the dynamic Golgi association of HID1 in animals, it is interesting to speculate the sub/neofunctionalization of HIDs in Schizosaccharomyces involves the maintenance of GA structure by SPAC17A5.16/Hid1. Exploring the link between HID presence and GA structure in closely related fungal species is one way to test this hypothesis. Another interesting evolutionary aspect relates to the asymmetric distribution of HID and DYM. There are advantages for some organisms to have both proteins, but for others the loss of one or the other did not appear to compromise progeny fitness. It is possible that for organisms that have lost both proteins, they may have evolved species-specific proteins arising to adopt GA functions as proposed by Barlow et al. (2018). Understanding the fundamental roles of both proteins in cellular trafficking and metabolism, will require comparative study involving more distant species, particularly single-cell organisms.

Supplementary Material

fnaf088_Supplemental_Files
fnaf088_supplemental_files.zip (1.6MB, zip)

Acknowledgments

We thank Dr Macha Nicolski and Dr Aurelien Barré at the Centre de Bioinformatique for use of the server and resident programs and Brigitte Battailler at the Bordeaux Imaging Centre for assistance with the TEM. We also thank Professor Partrick Babin of the Unit for Rare Diseases Genetics and Metabolism at the University of Bordeaux for his critical evaluation of the work.

Contributor Information

Mark A Hooks, University of Bordeaux/CNRS, UMR 5200 Laboratory of Membrane Biogenesis, 71 Avenue Edouard Bourlaux, 33882 Villenave d’Ornon, France.

Abdulrahman Alasmari, University of Bordeaux/CNRS, UMR 5200 Laboratory of Membrane Biogenesis, 71 Avenue Edouard Bourlaux, 33882 Villenave d’Ornon, France.

Mohammed Alshehri, University of Bordeaux/CNRS, UMR 5200 Laboratory of Membrane Biogenesis, 71 Avenue Edouard Bourlaux, 33882 Villenave d’Ornon, France.

Lysiane Brocard, Bordeaux Imaging Centre, 71 Avenue Edouard Bourlaux, 33882 Villenave d’Ornon, France.

Katarzyna B Hooks, University of Bordeaux/CNRS, UMR 5200 Laboratory of Membrane Biogenesis, 71 Avenue Edouard Bourlaux, 33882 Villenave d’Ornon, France.

Mathilde Julien, University of Bordeaux/CNRS, UMR 5200 Laboratory of Membrane Biogenesis, 71 Avenue Edouard Bourlaux, 33882 Villenave d’Ornon, France.

Ramsay J McFarlane, North West Cancer Research Institute, Bangor University, Bangor, Gwynedd LL57 2DG, United Kingdom.

Conflict of interest

None declared.

Funding

We would like to thank the Saudi Arabia Embassies of the UK and France for funding the PhD studies of A.A. and M.A.

References

  1. Ailion  M, Thomas  JH.  Isolation and characterization of high-temperature-induced Dauer formation mutants in Caenorhabditis elegans. Genetics. 2003;165:127–44. 10.1093/genetics/165.1.127. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Alavizargar  A, Gass  M, Krahn  MP  et al.  Elucidating the membrane binding process of a disordered protein: dynamic interplay of anionic lipids and the polybasic region. ACS Phys Chem Au. 2024;4:167–79. 10.1021/acsphyschemau.3c00051. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Amses  KR, Rabern-Simmons  D, Longcore  JE  et al.  Diploid-dominant life cycles characterize the early evolution of fungi. Proc Natl Acad Sci USA. 2022;119:e2116841119. 10.1073/pnas.2116841119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Antonny  B, Beraud-Dufour  S, Chardin  P  et al.  N-terminal hydrophobic residues of the G-protein ADP-ribosylation factor-1 insert into membrane phospholipids upon GDP to GTP exchange. Biochemistry. 1997;36:4675–84. 10.1021/bi962252b. [DOI] [PubMed] [Google Scholar]
  5. Arnaiz  O, Mathy  N, Baudry  C  et al.  The Paramecium germline genome provides a niche for intragenic parasitic DNA: evolutionary dynamics of internal eliminated sequences. PLoS Genet. 2012;8:e1002984. 10.1371/journal.pgen.1002984. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Aydin  B, Arga  KY.  Co-expression network analysis elucidated a core module in association with prognosis of non-functioning non-invasive human pituitary adenoma. Front Endocrinol. 2019;10:361. 10.3389/fendo.2019.00361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bähler  J, Wu  JQ, Longtine  MS  et al.  Heterologous modules for efficient and versatile PCR-based gene targeting in Schizosaccharomyces pombe. Yeast. 1998;14:943–51. 10.1002/(SICI)1097-0061(199807)14:10<943::AID-YEA292>3.0.CO;2-Y. [DOI] [PubMed] [Google Scholar]
  8. Balali-Mood  K, Harroun  TA, Bradshaw  JP.  Membrane-bound ARF1 peptide: interpretation of neutron diffraction data by molecular dynamics simulation methods. Mol Membr Biol. 2005;22:379–88. 10.1080/09687860500220148. [DOI] [PubMed] [Google Scholar]
  9. Barlow  LD, Nývltová  E, Aguilar  M  et al.  A sophisticated, differentiated Golgi in the ancestor of eukaryotes. BMC Biol. 2018;16:27. 10.1186/s12915-018-0492-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Beimforde  C, Feldberg  K, Nylinder  S.  et al.  Estimating the phanerozoic history of the Ascomycota lineages: combining fossil and molecular data. Mol Phylogenet Evol. 2014;78:386–98. 10.1016/j.ympev.2014.04.024. [DOI] [PubMed] [Google Scholar]
  11. Bhatnagar  RS, Gordon  JI.  Understanding covalent modifications of proteins by lipids: where cell biology and biophysics mingle. Trends Cell Biol. 1997;7:14–20. 10.1016/S0962-8924(97)10044-7. [DOI] [PubMed] [Google Scholar]
  12. Bittrich  S, Segura  J, Duarte  JM.  et al.  RCSB Protein Data Bank: exploring protein 3D similarities via comprehensive structural alignments. Bioinformatics. 2024;40:btae370. 10.1093/bioinformatics/btae370. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Boratyn  GM, Schäffer  AA, Agarwala  R.  et al.  Domain enhanced lookup time accelerated BLAST. Biol Direct. 2012;7:1–14. 10.1186/1745-6150-7-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Chamberlain  SA, Szöcs  E.  taxize: taxonomic search and retrieval in R. F1000Res. 2013;2:191. 10.12688/f1000research.2-191.v1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Coyne  RS, Hannick  L, Shanmugam  D  et al.  Comparative genomics of the pathogenic ciliate Ichthyophthirius multifiliis, its free-living relatives and a host species provide insights into adoption of a parasitic lifestyle and prospects for disease control. Genome Biol. 2011;12:R100. 10.1186/gb-2011-12-10-r100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Dimitrov  A, Paupe  V, Gueudry  C  et al.  The gene responsible for Dyggve-Melchior-Clausen syndrome encodes a novel peripheral membrane protein dynamically associated with the Golgi apparatus. Hum Mol Genet. 2009;18:440–53. 10.1093/hmg/ddn371. [DOI] [PubMed] [Google Scholar]
  17. Eckler  AM, Wilder  C, Castanon  A  et al.  Haploinsufficiency of the Sec7 guanine nucleotide exchange factor gea1 impairs septation in fission yeast. PLoS One. 2013;8:e56807. 10.1371/journal.pone.0056807. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Eddy  SR. Accelerated profile HMM searches. PLoS Comput Biol. 2011;7:e1002195. 10.1371/journal.pcbi.1002195. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. El Ghouzzi  V, Dagoneau  N, Esther  K  et al.  Mutations in a novel gene DYMeclin (FLJ20071) are responsible for Dyggve-Melchior-Clausen syndrome. Hum Mol Genet. 2003;12:357–64. 10.1093/hmg/ddg029. [DOI] [PubMed] [Google Scholar]
  20. Hachet  V, Köcher  T, Wilm  M.  et al.  Importin α associates with membranes and participates in nuclear envelope assembly in vitro. EMBO J. 2004;23:1526–35. 10.1038/sj.emboj.7600154. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Harada  H, Nagai  H, Tsuneizumi  M.  Identification of DMC1, a novel gene in the TOC region on 17q25.1 that shows loss of expression in multiple human cancers. J Hum Genet. 2001;46:90–95. 10.1007/s100380170115. [DOI] [PubMed] [Google Scholar]
  22. Hummer  BH, de Leeuw  NF, Burns  C  et al.  HID1 controls formation of large dense core vesicles by influencing cargo sorting and trans-Golgi network acidification. MBoC. 2017;28:3870–80. 10.1091/mbc.e17-08-0491. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Innan  H, Kondrasho  F.  The evolution of gene duplications: classifying and distinguishing between models. Nat Rev Genet. 2010;11:97–108. 10.1038/nrg2689. [DOI] [PubMed] [Google Scholar]
  24. Jamecna  D, Antonny  B.  Intrinsically disordered protein regions at membrane contact sites. Biochim Biophys Acta Mol Cell Biol Lipids. 2021;1866:159020. 10.1016/j.bbalip.2021.159020. [DOI] [PubMed] [Google Scholar]
  25. Jia  GS, Zhang  WC, Liang  Y  et al.  A high-quality reference genome for the fission yeast Schizosaccharomyces osmophilus. G3. 2023;13:28. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Jones  DT, Taylor  WR, Thornton  JM.  The rapid generation of mutation data matrices from protein sequences. Comput Appl Biosci. 1992;8:275–82. [DOI] [PubMed] [Google Scholar]
  27. Jumper  J, Evans  R, Pritzel  A  et al.  Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596:583–9. 10.1038/s41586-021-03819-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Katoh  K, Standley  DM.  MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol Biol Evol. 2013;30:772–80. 10.1093/molbev/mst010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Khemaissa  S, Sagan  S, Walrant  A.  Tryptophan, an amino-acid endowed with unique properties and its many roles in membrane proteins. Crystals. 2021;11:1032. 10.3390/cryst11091032. [DOI] [Google Scholar]
  30. Kouranti  I, McLean  J, Feoktistova  A.  A global census of fission yeast deubiquitinating enzyme localization and interaction networks reveals distinct compartmentalization profiles and overlapping functions in endocytosis and polarity. PLoS Biol. 2010;8:e1000471. 10.1371/journal.pbio.1000471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kozlov  AM, Darriba  D, Flouri  T  et al.  RAxML-NG: a fast, scalable and user-friendly tool for maximum likelihood phylogenetic inference. Bioinformatics. 2019;35:4453–5. 10.1093/bioinformatics/btz305. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Lera-Ramírez  M, Bähler  J, Mata  J  et al.  Revised fission yeast gene and allele nomenclature guidelines for machine readability. Genetics. 2023;225:iyad143. 10.1093/genetics/iyad143. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Liu  Y, Leigh  JW, Brinkmann  H  et al.  Phylogenomic analyses support the monophyly of Taphrinomycotina, including Schizosaccharomyces fission yeasts. Mol Biol Evol. 2009;26:27–34. 10.1093/molbev/msn221. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Luhua  S, Hegie  A, Suzuki  N  et al.  Linking genes of unknown function with abiotic stress responses by high-throughput phenotype screening. Physiol Plant. 2013;148:322–33. 10.1111/ppl.12013. [DOI] [PubMed] [Google Scholar]
  35. Ma  LJ, Ibrahim  AS, Skory  C  et al.  Genomic analysis of the basal lineage fungus Rhizopus oryzae reveals a whole-genome duplication. PLoS Genet. 2009;5:e1000549. 10.1371/journal.pgen.1000549. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Makarova  M, Peter  M, Balogh  G  et al.  Delineating the rules for structural adaptation of membrane-associated proteins to evolutionary changes in membrane lipidome. Curr Biol. 2020;30:367–80. 10.1016/j.cub.2019.11.043. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Matsuyama  A, Arai  R, Yashiroda  Y  et al.  ORFeome cloning and global analysis of protein localization in the fission yeast Schizosaccharomyces pombe. Nat Biotechnol. 2006;24:841–7. 10.1038/nbt1222. [DOI] [PubMed] [Google Scholar]
  38. Mesa  R, Luo  S, Hoover  CM  et al.  HID-1, a new component of the peptidergic signaling pathway. Genetics. 2011;187:467–83. 10.1534/genetics.110.121996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Mistry  J, Chuguransky  S, Williams  L  et al.  Pfam: the protein families database in 2021. Nucleic Acids Res. 2021;49:D412–9. 10.1093/nar/gkaa913. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Osipovich  AB, Jennings  JL, Lin  Q  et al.  Dyggve-Melchior-Clausen syndrome: chondrodysplasia resulting from defects in intracellular vesicle traffic. Proc Natl Acad Sci USA. 2008;105:16171–6. 10.1073/pnas.0804259105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Osumi  M.  Visualization of yeast cells by electron microscopy. J Electron Microsc. 2012;61:343–65. 10.1093/jmicro/dfs082. [DOI] [PubMed] [Google Scholar]
  42. Rambaut  A.  FigTree v1.4.4. Edinburgh: Institute of Evolutionary Biology, University of Edinburgh, 2018. First accessed 2020-05-18. http://tree.bio.ed.ac.uk/software/figtree/. [Google Scholar]
  43. Rhind  N, Chen  Z, Yassour  M  et al.  Comparative functional genomics of the fission yeasts. Science. 2011;332:930–6. 10.1126/science.1203357. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Rodríguez-López  M, Bordin  N, Lees  J.  et al.  Broad functional profiling of fission yeast proteins using phenomics and machine learning. eLife. 2023;12:RP88229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Rutherford  KM, Lera-Ramírez  M, Wood  V.  PomBase: a global core biodata resource—growth, collaboration, and sustainability. Genetics. 2024;227:iyae007. 10.1093/genetics/iyae007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Schänzer  A, Achleitner  MT, Trümbach  D  et al.  Mutations in HID1 cause syndromic infantile encephalopathy and hypopituitarism. Ann Neurol. 2021;90:143–58. 10.1002/ana.26127. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Shen  XX, Steenwyk  JL, LaBella  AL  et al.  Genome-scale phylogeny and contrasting modes of genome evolution in the fungal phylum ascomycota. Sci Adv. 2020;6:ABD0079. 10.1126/sciadv.abd0079. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Stecher  G, Tamura  K, Kumar  S.  Molecular evolutionary genetics analysis (MEGA) for macOS. Mol Biol Evol. 2020;37:1237–9. 10.1093/molbev/msz312. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Vyas  A, Freitas  AV, Ralston  ZA.  et al.  Fission yeast Schizosaccharomyces pombe: a unicellular ‘micromammal’ model organism. Curr Protoc. 2021;1:e151. 10.1002/cpz1.151. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Wang  L, Zhan  Y, Song  E  et al.  HID1 is a peripheral membrane protein primarily associated with the medial- and trans-Golgi apparatus. Protein Cell. 2011;2:74–85. 10.1007/s13238-011-1008-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Wernersson  R, Pedersen  AG.  RevTrans: multiple alignment of coding DNA from aligned amino acid sequences. Nucleic Acids Res. 2003;31:3537–9. 10.1093/nar/gkg609. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Woehle  C, Kusdian  G, Radine  C  et al.  The parasite Trichomonas vaginalis expresses thousands of pseudogenes and long non-coding RNAs independently from functional neighbouring genes. BMC Genomics. 2014;15:906. 10.1186/1471-2164-15-906. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Wood  V, Gwilliam  R, Rajandream  M-A  et al.  The genome sequence of Schizosaccharomyces pombe. Nature. 2002;415:871–80. 10.1038/nature724. [DOI] [PubMed] [Google Scholar]
  54. Yang  Z. PAML 4: Phylogenetic Analysis by Maximum Likelihood. Mol Biol Evol. 2007;24:1586–1591. 10.1093/molbev/msm088. [DOI] [PubMed] [Google Scholar]
  55. Zhang  J, Nielsen  R, Yang  Z.  Evaluation of an improved branch-site likelihood method for detecting positive selection at the molecular level. Mol Biol Evol. 2005;22:2472–9. 10.1093/molbev/msi237. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

fnaf088_Supplemental_Files
fnaf088_supplemental_files.zip (1.6MB, zip)

Articles from FEMS Microbiology Letters are provided here courtesy of Oxford University Press

RESOURCES