A Humanized Yeast Model for Studying TRAPP Complex Mutations; Proof-of-Concept Using Variants from an Individual with a TRAPPC1-Associated Neurodevelopmental Syndrome

Variants in membrane trafficking proteins are known to cause rare disorders with severe symptoms. The highly conserved transport protein particle (TRAPP) complexes are key membrane trafficking regulators that are also involved in autophagy. Pathogenic genetic variants in specific TRAPP subunits are linked to neurological disorders, muscular dystrophies, and skeletal dysplasias. Characterizing these variants and their phenotypes is important for understanding the general and specialized roles of TRAPP subunits as well as for patient diagnosis. Patient-derived cells are not always available, which poses a limitation for the study of these diseases. Therefore, other systems, like the yeast Saccharomyces cerevisiae, can be used to dissect the mechanisms at the intracellular level underlying these disorders. The development of CRISPR/Cas9 technology in yeast has enabled a scar-less editing method that creates an efficient humanized yeast model. In this study, core yeast subunits were humanized by replacing them with their human orthologs, and TRAPPC1, TRAPPC2, TRAPPC2L, TRAPPC6A, and TRAPPC6B were found to successfully replace their yeast counterparts. This system was used for studying the first reported individual with an autosomal recessive disorder caused by biallelic TRAPPC1 variants, a girl with a severe neurodevelopmental disorder and myopathy. We show that the maternal variant (TRAPPC1 p.(Val121Alafs*3)) is non-functional while the paternal variant (TRAPPC1 p.(His22_Lys24del)) is conditional-lethal and affects secretion and non-selective autophagy in yeast. This parallels defects seen in fibroblasts derived from this individual which also showed membrane trafficking defects and altered Golgi morphology, all of which were rescued in the human system by wild-type TRAPPC1. This study suggests that humanized yeast can be an efficient means to study TRAPP subunit variants in the absence of human cells and can assign significance to variants of unknown significance (VUS). This study lays the foundation for characterizing further TRAPP variants through this system, rapidly contributing to disease diagnosis.


Introduction
Rare diseases are considered those that affect 200,000 individuals or fewer in the United States [1] and affect 3.5-5.9% of the world population [2], with an estimation of around 30 million affected people in the U.S. and E.U. [3,4].Despite their low prevalence individually, more than 10,000 rare diseases are known worldwide [5], and with the rapid improvement of genetic technologies, new diseases are continuously being identified [2].Rare variant analysis through exome sequencing in families with individuals who suffer from undiagnosed diseases is a potent method to discover 'orphan' gene versions [6].While it is difficult to define the precise cause of a rare disorder, most of them are generated by single-gene mutations [4].The burden that rare conditions pose for the patients, caregivers, and society is tremendous [7,8].Because of their chronic and degenerative nature, these diseases extensively affect quality of life.In addition, rare disorders take time to diagnose, particularly as new genes are identified and novel variants in the new and known candidate genes are classified as variants of unknown significance (VUS).Such variants pose scientific and ethical dilemmas for clinicians, researchers, and families [9].
The compartmentalization of eukaryotic cells confines metabolic activities into individual organelles.Macromolecules of the endomembrane system are transported to their proper destination in a process referred to as membrane trafficking [10].An increasing number of disorders linked to defects in membrane traffic proteins have been reported in recent years and the number of genes known to cause these predominantly monogenic disorders has grown from 49 genes in 2011 to over 300 within a decade [11].Diseases caused by defects in these proteins frequently involve multiple systems but can also be manifested as more tissue or organ-specific [12].Approximately 70% of rare monogenic disorders are manifested with neurodevelopmental symptoms [13,14].By studying these rare Mendelian disorders, we can gain important insights into the function of membrane traffic genes and their complex molecular interactions.
TRAPP complexes are multi-subunit tethering factors initially identified in the yeast Saccharomyces cerevisiae that are highly conserved throughout eukaryotes.Two forms of the complex called TRAPPP II and TRAPP III have been reported in both yeast and human cells.Both complexes in yeast are composed of a common 'core' of subunits that include Bet5p, Trs20p, Bet3p, Trs23p, Trs31p, and Trs33p [15].With the exception of Trs33p, all of the other core subunits are essential for cell viability [16].In addition to this core, TRAPP II has four additional subunits (Tca17p, Trs65p, Trs120p, and Trs130p) [17][18][19][20], while TRAPP III has one additional subunit (Trs85p) [21].Although the exact function of these extra subunits is still ambiguous, they are thought to stabilize the complexes by giving them substrate and compartmental specificity [22].Although their role as bona fide tethers remains to be proven, it is thought that they carry out their function indirectly through their activity as Rab guanine nucleotide exchange factors (GEF) [23].TRAPP II is known to function in the late Golgi, regulating processes like endocytosis and post-Golgi trafficking [19,24], while TRAPP III is implicated in ER-to-Golgi trafficking and autophagy [25][26][27].The latter complex mediates its functions in ER-to-Golgi traffic and autophagy by activating the small GTPase RAB1/Ypt1p.This GTPase recruits distinct effectors for each process, including p115 for membrane trafficking and ATG1 for autophagy [28,29].Acting downstream of these effectors are SNARE proteins in membrane trafficking and ATG8/LC3 in autophagy.The TRAPP II complex also activates a GTPase called RAB11/Ypt31/32p with a distinct set of effectors.Thus, TRAPP complexes play early roles in these processes, at the level of GTPase activation.
Table 1.Mammalian and yeast TRAPP subunits.Mammalian subunits together with their yeast counterparts are listed and their presence in TRAPP II and/or TRAPP III is shown.Essential yeast TRAPP subunits are indicated as is the percentage of identity and similarity between yeast and human orthologs for the core subunits.Orthologs between yeast and mammals are listed on the same row.* TRAPPC13 is part of TRAPP III in mammals and Trs65p is part of TRAPP II in yeast.The detrimental effects of mutations in genes that encode TRAPP subunits are proof of the importance of TRAPP complexes in humans [31].In recent years, numerous mutations have been reported in TRAPP genes that are collectively termed TRAPPopathies, and the number of diseases associated with these complexes is expanding [15].Thus far, mutations in several core TRAPP proteins, including TRAPPC2 [32], TRAPPC4 [33,34], and TRAPPC6A/B [35][36][37][38], have been reported and presumably affect the functions of both TRAPP II and III complexes.Mutations in proteins specific to TRAPP II and III have also been reported [39][40][41][42][43][44][45][46].Disorders caused by variants in TRAPP proteins fall mainly into three categories: neurodevelopmental disorders, muscular dystrophies, and skeletal dysplasias [15,32,42].

Mammalian
One of the simplest sources for studying cellular phenotypes related to rare diseases is human dermal fibroblasts (HDF) [47].These are easily obtained and maintained and are genetically stable [48,49].HDFs have been used in many studies as cellular models for characterizing mutations that are linked to TRAPPopathies [39,43,50].Although patientderived cells make an ideal model system for studying the molecular mechanism of many diseases, the procedure for obtaining these cells is invasive.Therefore, the number of patients who agree to undergo a skin biopsy is relatively small [51].This poses a limitation when it comes to characterizing disease-causing mutations in humans and analyzing VUS.Thus, other approaches that do not rely on patient cells offer new possibilities for studying these disorders.
Saccharomyces cerevisiae has long been used as a model system for studying human biology due to its evolutionarily conserved genes and pathways that are very similar to those in humans [52,53].Roughly 30% of disease-associated genes in humans have orthologs in yeast and can functionally replace their yeast counterparts [54].An efficient way to use yeast as an experimental tool to study human diseases is by humanizing yeast cells.Laurent and colleagues considered five degrees of yeast humanization: (degree 0) non-humanized yeasts as systems for analyzing human-specific processes, (degree 1) heterologous expression of human genes in yeast, (degree 2) humanization of specific sites in yeast genes, (degree 3) humanization of complete yeast genes, (degree 4) and humanization of entire pathways in yeast [55].
One way of studying human mutations in yeast is the replacement of the entire yeast gene with the open reading frame (ORF) of its human ortholog.In the last 30 years, there have been many attempts to humanize yeast genes [53,[55][56][57], mainly due to the simplicity and effectiveness of this process [52].The development of CRISPR/Cas9 technology in yeast has proved to be an extremely useful tool [58].Unlike conventional methods, it has enabled the efficient replacement of yeast genes with their human orthologs in a scar-less way and with no need for a selectable marker [59].This approach consists of generating single plasmids that express both the Cas9 nuclease and the guide RNA, which will target any desired locus in the yeast genome.These plasmids, when transformed together with a repair template (i.e., human ORF), enable the efficient replacement of yeast genes with their human orthologs [59,60].
The main objective of this study is to create a humanized yeast strain that can be used as a platform to study TRAPP core subunit variants, particularly VUS, especially when there are restrictions in obtaining cells from an affected individual.Using the CRISPR/Cas9 editing tool, we show that several yeast TRAPP core genes can be singly humanized while others cannot.We then use this system to characterize the first known TRAPPC1 gene variants.Since the affected individual has compound heterozygous VUS, our system was able to reveal that both variants compromise function, moving these variants into the realm of likely disease-causing.Additionally, defects at the yeast cellular level with the humanized yeast are similar to those seen in fibroblasts derived from the affected individual, confirming this yeast system as a useful model system to study TRAPP gene variants when human cells are not available.

Materials and Methods
A complete list of plasmids and yeast strains used or created in this study are presented in Tables 2 and 3, respectively.

Plasmid
Construct or Backbone

Preparation of sgRNAs That Target the Yeast TRAPP Genes
All the single guide RNA oligonucleotides (sgRNA) targeting the yeast TRAPP genes were designed using the Geneious software (2024.0)and the Doench algorithm to score sgRNA for high on-target and low off-target activity [61,62].The sequences used are shown in Table 4. Forward and reverse oligonucleotides with the sgRNA sequence and Golden Gate compatible overlaps were mixed (10 µM each) in a total volume of 20 µL and annealed with each other using a thermocycler programmed to start at 95 • C and reduce the temperature in 5 • C increments every 15 min.An amount of 2 µL of annealed oligonucleotides (1:500 dilution) was used in the Golden Gate reaction for the assembly of the sgRNA cassette into the CRISPR/Cas9 plasmid (pCEN6-Cas9-GFP-KanMX).

Preparation of Repair Template DNA
The repair templates (human ORFs) were amplified using oligonucleotides designed for each of the human TRAPP genes that contained flanking sequences from the yeast genome.The sequences are shown in Table 5.To obtain 5 µg or more of the repair template DNA needed, PCR reactions of 150 µL were set using a high-fidelity polymerase (Phusion, New England Biolabs, Ipswich, MA, USA).Each template PCR was verified with agarose gel electrophoresis and then purified to use for yeast transformation.

Yeast Transformations
Yeast strains were transformed using the Frozen-EZ Yeast Transformation II Kit (Zymo Research, Irvine, CA, USA) as per the manufacturer's protocol.An amount of 250 µL was plated on selective media (YPD + G418).

Primer Design for Confirmation via Colony PCR
PCR primers to confirm gene replacement were designed for each of the human ORFs replacing their yeast counterparts such that the forward primers annealed to the yeast 5' untranslated region (UTR) and the reverse primers annealed only to the new human ORF.The oligonucleotides are listed in Table 6.

Colony Screening via PCR
Colonies from YPD + G418 plates were picked using a pipette tip, suspended in 50 µL of water, and microwaved at the highest power for 2 min to lyse the cells for a genomic DNA (gDNA) template.Using Phire Plant Direct PCR Master Mix Kit (Thermo Fisher Scientific, Waltham, MA, USA) as the polymerase, a 10 µL PCR reaction was set up, with confirmation primers and 2 µL of the gDNA template.The colony PCR reactions were fractionated by agarose gel electrophoresis to assess successful replacement of the yeast gene with its human ortholog.

Yeast Liquid Growth Assays
A BioTek Synergy HT incubating spectrophotometer (BioTek Instruments, Winooski, VT, USA) was used to perform the liquid growth assays.Yeast cultures of 150 µL were seeded in a 96-well plate with an initial cell density of 2.5-5 × 10 5 cells/mL by normalizing them to approximately 0.01-0.02OD600 units, across three replicates.The cultures were grown and OD600 was measured in the microplate reader, with continuous double orbital shaking.Readings were taken every 10-20 min over a 30 h time period.

Autophagy Assays in Yeast
Processing of pre-Ape1p was assessed by monitoring different forms of the protein using Western blot analysis [63].Yeast cultures of 10 mL were grown overnight to an earlyto-mid-log phase at OD600 = 0.5-0.8 in YPD at 30 • C. Five OD units of cells were pelleted at 2500 rpm in a tabletop centrifuge for 3 min, and then the supernatant was removed and the pellet was resuspended in 5 mL pre-warmed YPD medium and incubated at 35 • C or 37 • C (for temperature-sensitive mutants) or 30 • C in a water bath for 1 h.For selective autophagy, 1 OD unit of cells (1 mL culture) was immediately processed for lysis (see below).For non-selective autophagy, the remaining four OD units of cells (4 mL culture) were pelleted at 2500 rpm for 3 min and resuspended in pre-warmed synthetic medium lacking nitrogen (SD-N with 0.17% yeast nitrogen base without ammonium sulphate and amino acids and 2% glucose), incubated for 4 h at 35 • C or 37 • C (for temperature-sensitive mutants) or 30 • C, and then 1 mL of the culture was processed for lysis (see below).
Processing of GFP-Atg8p was performed by monitoring the cleavage of this fusion protein using Western blot analysis.Yeast cultures of 10 mL were grown overnight to an early-to-mid-log phase at OD600 = 0.5-0.8 in uracil drop-out medium at 30 • C. Five OD units of cells were pelleted as above at 2500 rpm for 3 min, the supernatant was removed and the pellet was resuspended in 5 mL pre-warmed uracil drop-out medium and incubated at 30 • C in a water bath for 1 h.After incubation, 1 OD unit of cells (1 mL culture) was immediately processed for lysis (0 h time point) (see below).The remaining four OD units of cells (4 mL culture) were pelleted at 2500 rpm for 3 min, resuspended in pre-warmed synthetic medium lacking nitrogen (SD-N with 0.17% yeast nitrogen base without ammonium sulphate and amino acids, and 2% glucose), incubated at 30 • C, and then 1 mL of culture was taken out after 1 h, 2 h, and 4 h to process for lysis (see below).

Preparation of Yeast Cell Lysates
For pre-Ape1p and GFP-Atg8p processing, 1 mL of culture was transferred immediately on ice and 100% trichloroacetic acid (TCA) was added to 10% final concentration (100 µL), and then the samples were incubated on ice for 20 min.The precipitated proteins were pelleted by centrifugation at 15,000× g for 3 min in a microfuge.After washing two times with 1 mL of ice-cold acetone, the pellet was air-dried.The dried pellet was resuspended in 1x Sample Buffer (4x SB recipe: 250 mM Tris-HCl pH 6.8, 40% Glycerol [vol/vol], 8% SDS [w/vol], 0.2% Bromophenol Blue [w/vol]) and then an equal volume of glass beads was added, keeping the glass beads below the level of the liquid.The samples were vortexed for 5 min and incubated at 70 • C for 10 min.Unlysed cells were removed by centrifugation at 10,000× g for 1 min in a microfuge and protein extracts equivalent to an OD600 = 0.2 units of cells (20 µL) were loaded on a 12% SDS-polyacrylamide gel (for pre-Ape1p) or a 10% gel (for GFP-Atg8p), resolved by SDS-PAGE, and processed for Western blotting using anti-Ape1 antibody or anti-GFP antibody.

Invertase Secretion Assay
The invertase assay protocol was based on the method described by Munn et al. [64].Yeast cells were cultured overnight in YPD medium at 24 • C until reaching an OD600 of 0.2-0.5.For each sample, two OD units of cells were collected and washed with 1 mL of distilled water.Subsequently, two OD units of each yeast strain were resuspended in inducing YP(low D)S (0.05% glucose and 2% sucrose) and non-inducing YPD (2% glucose) media to an OD600 of 0.5.The samples were then incubated at 25 • C, 30 • C, and 37 • C for 1 h.Post-incubation, the cells were washed twice with 2 mL of ice-cold 10 mM sodium azide, then adjusted to an OD600 of 0.5 using 10 mM sodium azide.Two 0.5 mL aliquots (0.25 OD units each) were transferred to sterile microfuge tubes: one for whole cell invertase activity (external) and one for lysed cell invertase activity (total; external + internal).For whole cells, 50 µL of distilled water was added, and for lysed cells, 50 µL of 10% Triton-X was added.To lyse the cells, two freeze-thaw cycles were performed using liquid nitrogen and room temperature thawing.The assay was prepared on ice.To 20 µL of cells, 25 µL of 0.2 M sodium acetate (pH 4.9) and 12.5 µL of 0.5 M sucrose were added.The sucrose was added to the tube walls to initiate the reactions simultaneously upon centrifugation.The tubes were centrifuged briefly and placed in a 37 • C water bath for 10 min.The reaction was terminated by adding 50 µL of 100 mM potassium phosphate (pH 7), and invertase was inactivated by heating at 90 • C for 3 min.The samples were chilled on ice.Next, 500 µL of assay mix (50 µg/mL glucose oxidase from Aspergillus niger, 10 µg/mL horseradish peroxidase, 10 mM potassium phosphate buffer pH 7, 300 µg/mL o-dianisidine, 38% vol/vol glycerol) was added.The samples were incubated at 30 • C for 20 min.Color development was achieved by adding 750 µL of 6N HCl, and absorbance was measured at A540 using a spectrophotometer.

Western Blot Analysis
Samples (0.2 OD units of cells for yeast lysates and 20 µg total protein for human fibroblast lysates) were fractionated on SDS-polyacrylamide gels (different concentrations depending on the protein analyzed) for 1 h and 30 min at 110 V.The proteins were trans-ferred to a nitrocellulose membrane or PVDF membrane (for LC3-II detection) for 1 h at 100 V in transfer buffer (25 mM Tris, 192 mM glycine, and 20% methanol [vol/vol]).Membranes were blocked with 5% skim milk powder in PBS-T (PBS with 0.1% Tween 20 [vol/vol]) for 2 h and incubated overnight at 4 • C with primary antibodies diluted in PBS-T.The secondary antibodies diluted in PBS-T were added for 1 h at room temperature.After incubation with ECL reagent (Thermo Fisher Scientific, Waltham, MA, USA) for 1 min, membranes were visualized on an Amersham Imager 600 (Chicago, IL, USA).

Genetic Studies
Written informed consent was obtained from the patient's parents according to an IRB-approved protocol (WCHN).Whole-exome sequencing was performed on genomic DNA obtained from the proband's blood.The amplified DNA fragments were hybridized to the Agilent SureSelect Human All Exon V4 (51 Mb), the captured library was sequenced on a HiSeq 2000 platform, and the reads were aligned against the human reference genome (GRCh37 at UCSC).Candidate variants were independently verified by Sanger sequencing on genomic DNA samples from blood the patient and her parents.Target loci were amplified by PCR, and PCR products were purified and bidirectionally sequenced using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Waltham, MA, USA) on an ABI 3730 DNA Analyzer (Applied Biosystems).

Cell Culture
Primary fibroblasts obtained by skin biopsy from the affected individual were cultivated and propagated using Dulbecco's modified eagle medium (DMEM) (Wisent) supplemented with 10% (vol/vol) fetal bovine serum (FBS; Thermo Fisher Scientific, Waltham, MA, USA).Cells were then incubated at 37 • C in a humidified incubator with 5% CO 2 .

Autophagy Assays in Mammalian Cells
Primary fibroblasts seeded in 6 cm diameter dishes were washed twice with phosphatebuffered saline (PBS) and incubated with Earl's balanced salt solution (EBSS) (Wisent) for 0 h, 1 h, and 2 h to induce autophagy.The cells were then lysed using 150 µL lysis buffer (150 mM NaCl, 50 mM Tris pH 7.2, 1 mM DTT, 1% Triton X-100, 0.5 mM EDTA, complete protease inhibitors (Roche Diagnostics, Basel, Switzerland)) and analyzed by Western blotting for the processing of LC3.The data were normalized to the tubulin signal developed from the same membrane as the LC3 blot.

Immunofluorescence Microscopy
Fibroblast cells cultured in 12 or 6-well plates on coverslips were washed twice with PBS, fixed for 15 min at room temperature with 4% paraformaldehyde (PFA), quenched for 10 min in 0.1 M glycine, and then permeabilized for 7 min in 0.1% Triton X-100 in PBS.The cells were blocked with 5% blocking solution (5% normal goat serum in PBS) for 45 min at room temperature and then incubated with primary antibodies resuspended in 5% blocking solution overnight at 4 • C. The cells were washed twice with PBS for 10 min each time and secondary antibodies resuspended in 5% blocking solution were added for 45 min at room temperature.The coverslips were washed one time with Hoescht in PBS for 2 min and two other times with PBS for 10 min each wash.The coverslips were then dried for 2-3 min and mounted on clean slides using Prolong Gold AntiFade (Life Technologies).Nikon confocal laser scanning microscope C2 TIRF was used to take images of the cells.The Golgi fragmentation images were obtained with 0.20 µm increment size and ImageJ was used to delineate the cells for each condition.The delineated cell images were converted to IMS files and imported to Imaris, and then analyzed using parameters pre-set to recognize Golgi structures as spherical-shaped structures (fragments) no smaller than 0.550 µm in size.The number of fragments were identified by the software and the number of fragments was quantified.The N value for control cells was 75 and the N values for S1 (patient cells) and S1 + RFP-TRAPPC1 (rescued cells) were 78 and 81, respectively.

Retention Using Selective Hooks (RUSH) Assay
Fibroblasts cultured in DMEM supplemented with 10% fetal bovine serum were used to monitor ER-to-Golgi trafficking through the Retention Using Selective Hooks (RUSH) assay as previously described [65].ImageJ software, version 1.54, was used to quantify the fluorescence intensity in individual cells.First, the fluorescence intensity of the Golgi was measured in each cell to be quantified.For this, the Golgi was distinguished and outlined manually in every cell and the fluorescence intensity was measured for each time point in the movie (images were captured every two minutes).With the same Golgi outlining, a random spot was chosen as the background and this background value was used to correct the Golgi fluorescence intensity at each time point.Next, the total fluorescence intensity was measured for each time point by outlining the entire cell.With the same cell outlining, a "background" measurement was taken.For each time point, the backgroundcorrected Golgi fluorescence intensity was divided by the background-corrected total cell fluorescence intensity.The produced values (the ratio of Golgi fluorescence intensity/total cell fluorescence intensity) were plotted.

Results
3.1.TRAPPC1, TRAPPC2, TRAPPC2L, TRAPPC6A, and TRAPPC6B Successfully Replace Their Yeast Orthologs BET5, TRS20, TCA17, and TRS33, Respectively In an effort to create a platform in yeast that can be used for characterizing mutations in TRAPP subunits in the absence of patient-derived cells, and since many reported mutations affect the core subunits, we started humanizing the core of the TRAPP complex.The human TRAPP genes and proteins TRAPPC1, TRAPPC2, TRAPPC2L, TRAPPC3, TRAPPC4, TRAPPC5, TRAPPC6A, and TRAPPC6B, will be referred to as C1, C2, C2L, C3, C4, C5, C6A, and C6B, respectively.These proteins are part of the core in both yeast and humans, common to both TRAPP complexes, except for C2L (Tca17p ortholog), which is only found in the TRAPP II complex in yeast [15].Each human open reading frame (ORF) was PCR amplified as a repair template with flanking homology to the locus of interest in yeast and was assessed by replacing its yeast ortholog using CRISPR/Cas9 as an editing tool.Eight human TRAPP genes (C1, C2, C2L, C3, C4, C5, C6A, and C6B) were tested for their ability to complement their yeast orthologs (BET5, TRS20, TCA17, BET3, TRS23, TRS31, and TRS33, respectively) individually.As seen in Figure 1A, after the transformation of the yeast cells, reactions containing C1, C2, C2L, C6A, and C6B repair templates showed colony growth in the selective media, suggesting that replacement has taken place, while those containing C3, C4, and C5 did not show any colonies.To confirm the success of replacement for C1, C2, C2L, C6A, and C6B genes, colony PCR was performed using confirmation primers designed for each of the transformed loci in yeast.As shown in Figure 1B, all the replaced loci in yeast show a band of the expected size, suggesting successful integration of the repair templates.
For further verification and to confirm a scar-less replacement of the yeast gene with the human ORF, the respective humanized loci were subjected to Sanger sequencing and all the sequences confirmed the presence of human ORFs in their respective yeast loci, replacing only the sequence between the start and stop codons and not affecting the 5' or 3' untranslated regions (Figure S1).The results from these experiments suggest that C1, C2, C2L, C6A, and C6B are able to complement their yeast orthologs while complementation was not successful for C3, C4, and C5.

TRAPPC3, TRAPPC4, and TRAPPC5 Cannot Complement the Loss of BET3, TRS23, and TRS31, Respectively, in Yeast
Since the humanized yeast strains express the human gene from the endogenous yeast promotor, the reason for the unsuccessful replacements of C3, C4, and C5 might be due to their inability to compensate for the loss of the genes encoding their essential yeast orthologs BET3, TRS23, and TRS31, respectively, at endogenous levels.We therefore examined if they could compensate when overexpressed.Haploid yeast strains in which the genomic copies of BET3, TRS23, and TRS31 were disrupted (bet3∆, trs23∆, trs31∆) but the cells were kept alive with URA3-based plasmids containing BET3, TRS23, or TRS31.These plasmids can be counter-selected on plates that contain 5-fluoroortic acid (5-FOA).The bet3∆, trs23∆, and trs31∆ strains were transformed with empty vectors, or with vectors containing either BET3 (the bet3∆ strain), TRS23 (the trs23∆ strain), or TRS31 (the trs31∆ strain), or C3 (the bet3∆ strain), C4 (the trs23∆ strain), or C5 (the trs31∆ strain), under the control of different overexpression promoters (GPD or ADH1).As shown in Figure 2, none of the human TRAPP genes could compensate for the loss of BET3, TRS23, and TRS31 as indicated by the lack of growth on plates containing 5-FOA.These results indicate that C3, C4, and C5 are not able to compensate for the disruption of their orthologous genes in yeast and reinforce the previous results from the humanization experiments in yeast where C3, C4, and C5 were shown to be incapable of replacing their yeast counterparts BET3, TRS23, and TRS31, respectively.

Humanized Yeast Strains Show a Slower Growth Phenotype Compared to the Parental Yeast Strains
A previous study with humanized yeast strains revealed that although most yeast strains might be slightly affected by the process of humanization, many other humanized strains show highly reduced growth rates that can be up to 70% slower [66].Therefore, we sought to test the growth rate of the humanized yeast strains with C1, C2, C2L, C6A, and C6B genes.For this, serial dilutions of these humanized strains were spotted onto YPD plates.The plates were incubated at three different temperatures (30 • C, 35 • C, 37 • C) and, after 48 h, examined for growth.As shown in Figure 3A, yeast strains humanized with C1, C2, and C2L showed a slower growth phenotype compared to the parental strain (MSY135) at 30 • C, 35 • C, and 37 • C, which is also the optimal temperature for human genes to function.In contrast, yeast strains humanized with C6A and C6B genes appeared to be unaffected by any of the conditions.
indicated by the lack of growth on plates containing 5-FOA.These results indicate that C3, C4, and C5 are not able to compensate for the disruption of their orthologous genes in yeast and reinforce the previous results from the humanization experiments in yeast where C3, C4, and C5 were shown to be incapable of replacing their yeast counterparts BET3, TRS23, and TRS31, respectively.plasmids, respectively.These strains were transformed with empty plasmids (vector), or vectors containing either BET3 (the bet3∆ strain), TRS23 (the trs23∆ strain), or TRS31 (the trs31∆ strain) under their endogenous promoters or either C3 (the bet3∆ strain), C4 (the trs23∆ strain), or C5 (the trs31∆ strain) under the control of the ADH1 or GPD promoters.Transformed cells were plated on 5-FOA to counter-select for the URA3-based plasmids.

Humanized Yeast Strains Show a Slower Growth Phenotype Compared to the Parental Yeast Strains
A previous study with humanized yeast strains revealed that although most yeast strains might be slightly affected by the process of humanization, many other humanized strains show highly reduced growth rates that can be up to 70% slower [66].Therefore, we sought to test the growth rate of the humanized yeast strains with C1, C2, C2L, C6A, and C6B genes.For this, serial dilutions of these humanized strains were spotted onto plasmids, respectively.These strains were transformed with empty plasmids (vector), or vectors containing either BET3 (the bet3∆ strain), TRS23 (the trs23∆ strain), or TRS31 (the trs31∆ strain) under their endogenous promoters or either C3 (the bet3∆ strain), C4 (the trs23∆ strain), or C5 (the trs31∆ strain) under the control of the ADH1 or GPD promoters.Transformed cells were plated on 5-FOA to counter-select for the URA3-based plasmids.
Next, we quantitatively tested the growth rate of the humanized yeast strains in a liquid medium.As shown in Figure 3B, C1, C2, and C2L strains showed slower growth rates compared to the parental strain in all the temperatures with a 2-5 h lag phase.C6A and C6B strains showed growth rates similar to the parental strain at 30 • C and 35 • C, while at 37 • C, the growth rate of the C6B strain was slower compared to the parental strain.This was in contradiction to the previous experiments on solid plates where the C6B strain appeared to grow normally even at 37 • C.These experiments demonstrate that humanized yeast strains with C1, C2, and C2L overall show a partial reduction in fitness compared to the parental strain, while the growth of humanized yeast strains with C6A and C6B is not affected at low temperatures and show similar fitness to the parental strain, but C6B appears to be heat-sensitive.
strain.This was in contradiction to the previous experiments on solid plates where C6B strain appeared to grow normally even at 37 °C.These experiments demonstrate humanized yeast strains with C1, C2, and C2L overall show a partial reduction in fit compared to the parental strain, while the growth of humanized yeast strains with and C6B is not affected at low temperatures and show similar fitness to the parental st but C6B appears to be heat-sensitive.

A Novel Disorder Associated with Biallelic TRAPPC1 Variants
The proband who carries these biallelic C1 variants is an 11-year-old girl with a severe neurodevelopmental phenotype.She was born to healthy and unrelated parents following an uncomplicated at-term pregnancy by normal delivery.She then presented with global neurodevelopmental delay and a phenotype characterized by truncal hypotonia with lower limb spasticity, intellectual disability with severely impaired adaptive functioning, lack of speech, and autistic features with self-injurious behavior.She has some dysmorphic features including hypertelorism, long palpebral fissures, high-arched eyebrows, a bulbous nasal tip, large ears, a high-arched palate, widely spaced teeth, and a single palmar crease on the right hand.Brain MRI revealed a thin corpus callosum, hypoplastic olfactory nerves, and hypomyelination.She also has a myopathy with persistently elevated creatine kinase (CK) levels.A muscle biopsy showed minor non-specific findings and no ultrastructural anomalies.
Whole exome sequencing, initially reported as negative, was reanalyzed through the Program for Undiagnosed Rare Diseases of CIBERER (ENoD) [67].Two candidate variants were identified in the TRAPPC1 gene: NM_021210.5:c.362_363del; p.(Val121Alafs*3)/ NM_021210.5:(c.64_72del); p.(His22_Lys24del).Ulterior validation by Sanger sequencing and segregation in parental samples revealed that the patient is a compound heterozygote with biallelic variants in C1, while both parents are heterozygous carriers.In the maternally inherited variant, p.(Val121Alafs*3), a deletion of two nucleotides (c.362_363del) results in a frameshift mutation that introduces a premature stop codon, altering the C-terminal 24 amino acids that delete one of the three helices (Figure 4A,B).This variant (rs768379826) is rare (allele population frequency in the gnomAD 4.0 database of 0.0002763, with no homozygotes) and has a very high in silico prediction of deleteriousness (CADD value of 33).In the paternally inherited variant, p.(His22_Lys24del), a deletion of nine nucleotides (c.64_72del) results in an in-frame three amino acid deletion in the protein in a short loop connecting two strands of the β-sheet (Figure 4A,B).This variant is very uncommon (allele population frequency in the gnomAD 4.0 database of 0.0000012) and has a high in silico prediction of deleteriousness (CADD value of 22.4).

Humanized Yeast as a Platform for Characterizing the First Reported Mutation in TRAPPC1
We next sought to test whether the humanized platform in yeast provides an optimal system for characterizing mutations in human genes.For this, we studied the first reported mutation in C1.C1 is an essential gene in humans, encoding for a protein ortholog to Bet5p in yeast that shares 33% identity and 54% similarity (Table 1).C1 is part of the longin domain proteins, which are composed of a conserved domain of 120-140 amino acids at their N-terminus [68].Longin domains contain a five-stranded β-sheet flanked by a single α-helix on one side and two α-helices on the other side [69], and they have been proposed to regulate many membrane-trafficking processes [70].C1 and Bet5p are part of the core TRAPP subunits and, therefore, they are common to both TRAPP II and III.
There were two reasons for using the mutant C1 variants as a candidate to test our humanized platform in yeast.First was the fact that wild-type C1 can replace BET5 in yeast, and second was the presence of fibroblast cells from the patient bearing the C1 variants.This would allow us to compare the phenotype of the humanized yeast platform with that of human cells bearing these variants.As a source of each variant for our yeast work, messenger RNA (mRNA) was extracted from the patient fibroblasts and used to synthesize complementary DNA (cDNA).This cDNA was used as a template to generate the repair templates to humanize the yeast for both C1 mutations.
As a first estimate to see whether we would be able to humanize yeast strains with both C1 mutations, we examined the overexpression of each variant in yeast devoid of BET5 (bet5∆) containing a URA3-based wild-type BET5 in the presence of 5-FOA.As shown in Figure 5A, only BET5, C1, and C1 (c.64_72del) could compensate for the loss of BET5 as indicated by growth on plates containing 5-FOA, while C1 (c.362_363del) was non-functional and did not support the growth of bet5∆ yeast strains on 5-FOA.
Knowing that when overexpressed, the maternal variant C1 p.(Val121Alafs*3) (expressed from C1 (c.362_363del)) is lethal and the paternal variant C1 p.(His22_Lys24del) (expressed from C1 (c.64_72del)) is functional, we examined their functionality in the humanized yeast platform under the control of the endogenous BET5 promoter.For this, we used the CRISPR/Cas9 plasmid and either C1 (c.362_363del) or C1 (c.64_72del) as repair templates.As seen in Figure 5B, after the transformation of the yeast cells, humanized strains with C1 and C1 (c.64_72del) showed colony growth in the selective media while humanized strains with C1 (c.362_363del) could not support yeast growth.

Humanized Yeast as a Platform for Characterizing the First Reported Mutation in TRAPPC1
We next sought to test whether the humanized platform in yeast provides an optimal system for characterizing mutations in human genes.For this, we studied the first reported mutation in C1.C1 is an essential gene in humans, encoding for a protein ortholog to Bet5p in yeast that shares 33% identity and 54% similarity (Table 1).C1 is part of the longin domain proteins, which are composed of a conserved domain of 120-140 amino acids at their N-terminus [68].Longin domains contain a five-stranded β-sheet flanked by a single α-helix on one side and two α-helices on the other side [69], and they have been   Since the C1 (c.64_72del) mutant was functional and could replace the BET5 gene in yeast, we next examined the growth efficiency of this mutant in solid and liquid media using the parental strain (MSY135) and the humanized C1 strain as controls.As shown in Figure 6A, yeast strains humanized with C1 (c.64_72del) showed a slower growth phenotype compared to the parental strain (MSY135) at 30 • C, even slower growth at 35 • C, and no growth at 37 • C.These observations are consistent with the results from the liquid growth curves (Figure 6B) where the C1 (c.64_72del) strain showed a slower growth rate at 30 • C compared to the parental MSY135 strain and was essentially non-functional at 35 • C and 37 • C.These results demonstrate that the C1 (c.64_72del) humanized yeast strain is a conditionally lethal mutant, and its growth is sensitive to temperatures above 30 • C.Although the C1 (c.64_72del) mutation complements BET5 in yeast, its functionality is not comparable to the wild-type C1 at 30 • C.

Non-Selective Autophagy Is Defective in the Humanized C1 (c.64_72del) Yeast Strain
The fact that the mutant C1 (c.64_72del) was conditionally lethal presented us with the ability to characterize this mutation in our humanized platform by growing this strain at permissive and restrictive temperatures.Since the C1 subunit is part of both TRAPP complexes in humans, and knowing that TRAPP III functions in autophagy, we tested whether C1 (c.64_72del) affected autophagy.
Depending on the media in which the cells are grown, the presence of a nitrogen source (rich media) or the absence of a nitrogen source (starvation media) can help distinguish between the selective cytosol-to-vacuole (Cvt) autophagic pathway (+nitrogen) and the non-selective pathway (-nitrogen).The Cvt pathway occurs continuously during cell growth in conditions rich in nutrients, while the non-selective pathway is usually triggered by starvation [71].The marker protein Ape1p is delivered to the vacuole by the selective pathway in the presence of nitrogen, and in the absence of nitrogen it is transported to the vacuole in a non-selective manner [72].The processing of Ape1p is monitored by detecting the levels of the precursor protein and its processed form by SDS-PAGE/Western blot analysis.As shown in Figure 7A, the processing of Ape1p in the presence of nitrogen (+N panel) was completely blocked for the humanized strain with C1 and C1 (c.64_72del) compared to the parental strain, suggesting that selective autophagy is impaired in the two humanized strains.In starvation conditions, when Ape1p is processed by non-selective autophagy (−N panel), at 30 • C, there was a notable amount of precursor Ape1p and a very small amount of mature Ape1p detected in the humanized C1 (c.64_72del) strain compared to the humanized C1 strain and the parental strain, suggesting a defect in non-selective autophagy of Ape1p in the C1 (c.64_72del) strain (Figure 7A).At 35 • C, no processing of the Ape1p was detected from the precursor to the mature form in the humanized C1 (c.64_72del) strain compared to the humanized C1 strain and the parental strain, although the humanized C1 strain at this temperature showed less processing of Ape1 compared to the parental strain.At 37 • C, both humanized strains with C1 and C1 (c.64_72del) were defective in non-selective autophagy and did not process Ape1p compared to the parental strain.These results indicate that humanized strains with C1 and C1 (c.64_72del) are both defective in selective autophagy while C1 (c.64_72del) impairs the non-selective autophagy of Ape1p.
growth curves (Figure 6B) where the C1 (c.64_72del) strain showed a slower growth rate at 30 °C compared to the parental MSY135 strain and was essentially non-functional at 35 °C and 37 °C.These results demonstrate that the C1 (c.64_72del) humanized yeast strain is a conditionally lethal mutant, and its growth is sensitive to temperatures above 30 °C.Although the C1 (c.64_72del) mutation complements BET5 in yeast, its functionality is not comparable to the wild-type C1 at 30 °C.To further evaluate whether the C1 (c.64_72del) mutation affects non-selective autophagy, we examined the processing of GFP-Atg8p.Under starvation conditions, this fusion protein is delivered to the vacuole in a non-selective manner and cleaved, releasing GFP, which is resistant to the proteolytic enzymes of the vacuole.The appearance of free GFP can be used as a marker, which can be detected through Western blot analysis over time.As seen in Figure 7B, at 30 • C, in the parental strain and the humanized C1 strain, there was an increase in the amount of free GFP over time, whereas the humanized C1 (c.64_72del) strain accumulated only the full-length GFP-Atg8p but not free GFP.Collectively, these results demonstrate that C1 (c.64_72del) affects the non-selective autophagy of Ape1p and GFP-Atg8p while both the mutant and wild-type C1 appear to affect selective autophagy.C, and 37 • C, as described in the Section 2. The mutant strains atg1∆ and trs85∆ were used as controls to show defects in autophagy.Equal amounts of protein were fractionated by SDS-PAGE and analyzed by Western blot analysis using anti-Ape1p IgG.(B) The parental strain MSY135 and the humanized strains with C1 and C1 (c.64_72del) expressing plasmidbased GFP-Atg8 under the control of the endogenous ATG8 promoter were grown in nitrogen-starved medium (SD-N) to induce autophagy at 30 • C. The mutant strains atg1∆ and trs85∆ were used as controls to show defects in autophagy.At the indicated times, aliquots were removed and protein extracts were prepared and fractionated by SDS-PAGE, and analyzed by Western blot analysis using anti-GFP antibody.(C,D) Yeast cells (wild-type, sec18, humanized C1 and C1 (c.64_72del)) were grown in medium with 0.05% glucose and 2% sucrose to induce expression of the SUC2 gene that produces the secreted invertase enzyme.Induction was allowed to proceed for 1 h, at which time cells were collected and analyzed before or after lysis to determine percent external (C) and internal (D) invertase.Percentage external was calculated as (whole cell/lysed cell) × 100, and percentage internal was calculated as (lysed cell-whole cell/lysed cell) × 100.N = 8 over three biological replicates and error bars indicate SEM.Significance was calculated using a one way ANOVA with a Bonferroni post hoc correction.* indicates p < 0.05; **** indicates p < 0.0001; ns = not significant.

Secretion Is Defective in the Humanized C1 (c.64_72del) Yeast Strain
Since TRAPP III also plays a role in ER-to-Golgi trafficking and TRAPP II functions at the level of the Golgi, a defect in C1 would be expected to affect secretion.We tested this notion by examining the secretion of invertase, the product of the SUC2 gene.In high (2%) glucose, the gene is not expressed, but expression is induced in low (0.05%) glucose in the presence of the substrate sucrose [64].Yeast cells were grown under non-inducing and inducing conditions at various temperatures.We included a sec18 mutant, which served as a control for a secretion block.Invertase was assayed on whole cells (a measure of external invertase activity) and on lysed cells (a measure of total invertase activity).We then calculated the percent secreted invertase and internal invertase.As shown in Figure 7C,D, wild-type cells efficiently secreted invertase upon induction with no significant change seen at any of the temperatures examined.As expected, the percentage of invertase that was secreted in sec18 was comparable to that of control yeast cells at 25 • C but dropped precipitously as the growth temperature increased to 34 • C, with a concomitant increase in internal invertase.The C1 humanized yeast showed no significant difference in secreted invertase when compared to the wild type at any of the temperatures.In contrast, the C1 (c.64_72del) humanized strain showed a significant defect in secretion at 25 • C that was more significantly seen at 34 • C.These results indicate that the C1 (c.64_72del) mutant affects secretion and does so in a temperature-sensitive manner.

Fibroblasts from the Affected Individual with TRAPPC1 Mutations Do Not Show an Accumulation of LC3-II during Starvation
To better establish the humanized yeast system as a valid system to use in the absence of human cells, we needed to compare the data obtained from the humanized yeast to fibroblasts obtained from the affected individual with the compound heterozygous TRAPPC1 variants.We therefore characterized these cells for an autophagic defect to assess whether the results from this experiment would parallel those in yeast.For this purpose, we starved the fibroblasts and monitored the levels of LC3-II.LC3, the mammalian ortholog of yeast Atg8p, is synthesized as pro-LC3, a soluble unprocessed form that is proteolytically processed into LC3-I and finally modified into LC3-II (the phosphatidylethanolamineconjugated form), which is associated with membranes [73].
Control and patient fibroblasts were starved for up to 2 h in EBSS starvation medium to induce autophagy.Cell lysates were collected at 0 h, 1 h, and 2 h time points.As indicated in Figure 8A and quantified in Figure 8B, control cells showed an increase of LC3-II over the 2 h period.In contrast, patient cells did not show significant differences in the level of LC3-II over time compared to the control cells, while they showed stable levels of LC3-I over the same 2 h period.The failure to convert LC3-I to LC3-II suggests a defect in autophagosome formation in these cells.

Wild-Type TRAPPC1 Rescues a Membrane Trafficking Defect and Golgi Fragmentation in Patient Fibroblasts
Human TRAPP II and III are involved in ER-to-Golgi trafficking and late Golgi transport, respectively, and the C1 protein is part of the core TRAPP subunits of both complexes.Hence, patient fibroblasts were examined for defects in membrane trafficking using the RUSH system.In this assay, a fluorescent cargo protein (ST-eGFP) is kept in the ER until biotin is administered.Upon biotin addition, the cargo exits the ER and is transported to the Golgi, and the fluorescent signal is monitored and quantified as it accumulates in the Golgi [44,65].
As shown in Figure 9A, the perinuclear fluorescence intensity in the control cells increased over time as the cargo protein arrived at the Golgi.In contrast, the patient fibroblasts (S1) displayed a delay in the arrival of the cargo protein at the Golgi compared to the control cells.These results demonstrate that patient fibroblasts are defective in membrane-trafficking events along the biosynthetic pathway.This defect is due to non-functional C1 since patient cells transfected with a construct of the wild-type C1 tagged with a red fluorescent protein (RFP) rescued the trafficking defect (Figure 9A).

Wild-Type TRAPPC1 Rescues a Membrane Trafficking Defect and Golgi Fragmentation in Patient Fibroblasts
Human TRAPP II and III are involved in ER-to-Golgi trafficking and late Golgi transport, respectively, and the C1 protein is part of the core TRAPP subunits of both complexes.Hence, patient fibroblasts were examined for defects in membrane trafficking using the RUSH system.In this assay, a fluorescent cargo protein (ST-eGFP) is kept in the ER until biotin is administered.Upon biotin addition, the cargo exits the ER and is transported to the Golgi, and the fluorescent signal is monitored and quantified as it accumulates in the Golgi [44,65].
As shown in Figure 9A, the perinuclear fluorescence intensity in the control cells increased over time as the cargo protein arrived at the Golgi.In contrast, the patient fibroblasts (S1) displayed a delay in the arrival of the cargo protein at the Golgi compared to the control cells.These results demonstrate that patient fibroblasts are defective in membrane-trafficking events along the biosynthetic pathway.This defect is due to nonfunctional C1 since patient cells transfected with a construct of the wild-type C1 tagged with a red fluorescent protein (RFP) rescued the trafficking defect (Figure 9A). in mind, the fragmentation of the Golgi into small structures was quantified in the patient fibroblasts (S1) and was compared to the control cells.To visualize the Golgi, an antibody that recognizes the Golgi resident protein mannosidase-II (Man-II) was used.As shown in Figure 9B, the number of Golgi fragments in patient cells was significantly higher compared to the control.The transfection of patient cells with a construct of the wild-type C1 fused to RFP significantly rescued this phenotype, confirming that the Golgi morphological alterations are C1-dependent.Next, we examined the Golgi for alterations.Previous studies have reported changes in Golgi morphology in patients with mutations in TRAPP subunits [39,43,44].With this in mind, the fragmentation of the Golgi into small structures was quantified in the patient fibroblasts (S1) and was compared to the control cells.To visualize the Golgi, an antibody that recognizes the Golgi resident protein mannosidase-II (Man-II) was used.As shown in Figure 9B, the number of Golgi fragments in patient cells was significantly higher compared to the control.The transfection of patient cells with a construct of the wild-type C1 fused to RFP significantly rescued this phenotype, confirming that the Golgi morphological alterations are C1-dependent.

Discussion
The scarcity of human cellular models represents a major challenge for studying the globally increasing number of rare diseases in general and TRAPPopathies in particular.This can be attributed at least in part to the limited number of patients diagnosed with rare conditions who agree to provide cells through invasive techniques such as skin and deeper tissue biopsies.To overcome this difficulty, there is a need for using other models to characterize disease-causing variants in humans.The yeast Saccharomyces cerevisiae has long been used as a model organism for studying human processes due to the remarkable degree of gene conservation between the two organisms.Therefore, using CRISPR/Cas9 as an editing tool, we generated a platform in yeast that would enable the characterization of mutations in human TRAPP genes in the absence of patient-derived cells.Analyzing first the core TRAPP genes, we demonstrated that C1, C2, C2L, C6A, and C6B can complement individually their orthologs BET5, TRS20, TCA17, and TRS33, respectively, in yeast.However, C3, C4, and C5 could not replace their corresponding yeast counterparts, BET3, TRS23, and TRS31, respectively.These results were confirmed by plasmid-based assays in yeast, which proved that none of these genes, even when expressed under strong promoters, were able to compensate for the loss of their yeast orthologs.
Despite the successful and scar-less replacement of their yeast equivalents by some of the human core TRAPP genes, C1, C2, and C2L humanized strains revealed slightly compromised fitness compared to the parental yeast strain, a feature more pronounced at temperatures above 30 • C, while the C6B humanized strain was found to be heat-sensitive.The growth delays observed in the humanized strains here have also been reported in other studies that explored humanized yeast systems [66,74].Although the exact mechanisms that contribute to this decrease in growth efficiency are not known, the adverse environment in yeast, alterations in function and regulation of human genes, or failure of human orthologs to complement the moonlighting functions of a yeast protein might be contributing factors.
After generating a set of individually engineered strains with human TRAPP genes expressed at their native loci, the applicability of this system in studying human variants was investigated.Within the context of this study, the focus was to use this system as a test platform to characterize the first reported compound heterozygous mutations in C1, where deletion of two nucleotides (c.362_363del) in the maternal variant (C1 (p.Val121Alafs*3)) alters an entire α-helix at the C-terminus of this longin domain protein, while deletion of nine nucleotides (c.64_72del) in the paternal variant (C1 p.(His22_Lys24del)) affects a short loop bridging two strands of the β-sheet structure.Towards this goal, yeast humanization with both mutants individually showed that the C1 (c.362_363del) variant is lethal and cannot replace BET5 in yeast, while the C1 (c.64_72del) variant is functional and can complement its ortholog BET5.The C1 (c.64_72del) variant was tested for its growth efficiency under different temperature conditions (30 • C, 35 • C, and 37 • C) and the growth rate of this mutant was substantially compromised at 35 • C and 37 • C, indicating that it is a conditionally lethal mutant.
The phenotypes observed for these two mutants might be better explained by looking at the structure of the C1 protein and the data that come from crystallographic and cryo-EM studies that have revealed the structure of the TRAPP core.The octameric core forms an elongated rod with two flattened surfaces that accommodate the C1 and C4 proteins in the center.As two longin domain proteins, C1 and C4 also form most of the catalytic site for activating Rab GTPases.As shown in Figure 4B, C1 has three close core interactors, C4 (in light pink), C3 (in cyan), and C6 (in hot pink).In the C1 (p.Val121Alafs*3) variant (expressed from C1 c.362_363del), the entire α-helix at the C-terminus of the protein, which is in close proximity to the C6 subunit, is deleted.Therefore, the removal of this α-helix would be expected to have a detrimental effect on C1 as it would interfere with the structure of this longin domain protein.This might explain the non-functionality of the maternal variant C1 p.(Val121Alafs*3) in the yeast.On the other hand, in the C1 p.(His22_Lys24del) variant (expressed from C1 c.64_72del), the three amino acid deletion affects a portion of the protein, which is a short peptide that connects two domains of secondary structure and is close to the catalytic where C1 and C4 accommodate and activate Rab GTPases.Although previous reports have revealed the importance of loops in the folding/geometry of proteins, the partial truncation of this loop might reflect less severe effects on the geometry of C1.Indeed, previous studies have revealed that when perturbations are introduced by mutations in the structure of the protein, they might change the connection between elements of the system (atoms/residues), which in turn might provoke changes in the functional site and other sites of the structure [75][76][77].This might explain why in fact the paternal variant C1 p.(His22_Lys24del), contrary to C1 p.(Val121Alafs*3), acts as a conditionally lethal variant.
Taken together, our data highlight the efficiency of the humanized yeast as a platform to characterize mutations in human TRAPP genes and suggest a possible correlation between the C1 variant and the severe neurodevelopmental delay from which the individual bearing the C1 mutations suffers.This latter suggestion is related to the use of this system for addressing VUS.VUS are variants for which the clinical significance, and therefore the association with the risk of disease, is not clear [9].In this study, we applied this platform for addressing the significance of a compound heterozygous mutation in TRAPPC1 identified as a VUS from genetic testing data.While the maternal variant was found to be lethal and could not replace its yeast counterpart, the paternal variant manifested as a conditional phenotype.Therefore, employing this system with other TRAPP proteins will allow us to rapidly test additional VUS and determine if they are functionally significant, thereby accelerating the diagnostic odyssey for rare disease patients.
Although we tried to detect the expressed TRAPPC1 protein in yeast lysates from the humanized strains, the commercial antibody we used did not work well by Western blotting and the results were ambiguous.We suspect that the wild-type protein is indeed expressed since the yeast homolog Bet5p is essential and the TRAPP core would be predicted to not assemble in its absence based on the known structures of the complexes from multiple organisms [78][79][80][81].It is possible that the conditional phenotype of the paternal variant is unstable at elevated temperatures in both yeast and human cells, possibly accounting for the human disorder.
It was surprising that several yeast genes could not be functionally replaced by their human orthologs, particularly BET3, which shows 56% identity and 76% similarity at the protein level with human C3.It is important to note that C3 is the only core protein that is present in two copies (C3a and C3b) by flanking both sides of the core center formed by C1 and C4.This might explain why its humanization is intolerable since the single gene replacement should generate in a sense a double-protein replacement in the complex.Therefore, a major challenge in the future will be to make these genes humanizable in yeast.Since certain genes require local interactions for functional replaceability, we are exploring combinations of replacements.As we progressively humanize additional TRAPP genes in yeast, our final goal would be to assemble all these replacements in one strain, generating in this way a fully humanized TRAPP core in yeast.
Through the humanized system, we were able to characterize the first reported mutations in C1 by performing functional experiments that explore pathways where this protein is involved.The results from these experiments were supported by similar outcomes in patient fibroblasts confirming the applicability of this system as a reliable tool for characterizing other gene mutations.Since the C1 mutation in this study is compound heterozygous,

Figure 1 .
Figure1.Yeast humanization with core TRAPP genes using a CRISPR/Cas9 editing tool.(A) Yeast cells were transformed with ORFs from human C1, C2, C2L, C3, C4, C5, C6A, or C6B as repair templates in the presence of the self-containing CRISPR/Cas9 plasmid to replace their yeast orthologs-BET5, TRS20, TCA17, BET3, TRS23, TRS31, and TRS33, respectively.The upper left plate is a control of cells transformed with the empty CRISPR/Cas9 plasmid that contains no sgRNA, and no repair template is provided; therefore, it is used to evaluate the transformation efficiency of the yeast strains being used.The remaining plates show the results of the indicated sgRNA with or without (second plate from left at top as an example of the double-stranded break (DSB)-induced lethality) repair template.The presence or absence of growth in the remaining plates is an indication of rescue or no rescue from the DSB lethality of the Cas9 nuclease when an appropriate repair template is provided.(B) Colonies from plates transformed with the empty CRISPR/Cas9 plasmid and from plates transformed with C1, C2, C2L, C6A, or C6B repair templates were picked and analyzed by PCR for correct integration of the human ORF.Lanes labeled parental strain represent negative controls from the parental strain transformed with the empty plasmid.The remaining lanes represent the lysates from cells transformed with the appropriate repair templates.The expected band size for each successful transformation is 274 bp for C1, 311 bp for C2, 255 bp for C2L, 428 bp for C6A, and 276 bp for C6B.

Figure 1 .
Figure1.Yeast humanization with core TRAPP genes using a CRISPR/Cas9 editing tool.(A) Yeast cells were transformed with ORFs from human C1, C2, C2L, C3, C4, C5, C6A, or C6B as repair templates in the presence of the self-containing CRISPR/Cas9 plasmid to replace their yeast orthologs-BET5, TRS20, TCA17, BET3, TRS23, TRS31, and TRS33, respectively.The upper left plate is a control of cells transformed with the empty CRISPR/Cas9 plasmid that contains no sgRNA, and no repair template is provided; therefore, it is used to evaluate the transformation efficiency of the yeast strains being used.The remaining plates show the results of the indicated sgRNA with or without (second plate from left at top as an example of the double-stranded break (DSB)-induced lethality) repair template.The presence or absence of growth in the remaining plates is an indication of rescue or no rescue from the DSB lethality of the Cas9 nuclease when an appropriate repair template is provided.(B) Colonies from plates transformed with the empty CRISPR/Cas9 plasmid and from plates transformed with C1, C2, C2L, C6A, or C6B repair templates were picked and analyzed by PCR for correct integration of the human ORF.Lanes labeled parental strain represent negative controls from the parental strain transformed with the empty plasmid.The remaining lanes represent the lysates from cells transformed with the appropriate repair templates.The expected band size for each successful transformation is 274 bp for C1, 311 bp for C2, 255 bp for C2L, 428 bp for C6A, and 276 bp for C6B.

Figure 3 . 3 Figure 3 .
Figure 3. Growth efficiency of the humanized yeast strains in solid and liquid media.(A) Im from the plates of C1, C2, C2L, C6A, and C6B humanized yeast strains.Serial dilutions of the l cultures were spotted on YPD plates and incubated for 48 h at three different temperatures, 3 Figure 3. Growth efficiency of the humanized yeast strains in solid and liquid media.(A) Images from the plates of C1, C2, C2L, C6A, and C6B humanized yeast strains.Serial dilutions of the liquid cultures were spotted on YPD plates and incubated for 48 h at three different temperatures, 30 • C, 35 • C, and 37 • C. (B) Representation of the growth curves in liquid medium for C1, C2, C2L, C6A, and C6B humanized yeast strains in 30 • C, 35 • C, and 37 • C. Shown are the averages of OD600 for three biological replicates.

Figure 4 .
Figure 4.The compound heterozygous mutation in the C1 protein.(A) C1 p.(Val121Alafs*3) and C1 p.(His22_Lys24del) variants are predicted in the amino acid sequence of the C1 protein.Shown are the reference and the predicted sequence of the protein for both variants and the nomenclature for C1 mutations and variants.(B) C1 p.(Val121Alafs*3) and C1 p.(His22_Lys24del) are highlighted in red in the structure of the C1 protein.Interacting subunits of the TRAPP core with C1 are included partially in the figure.Shown are C1 p.(Val121Alafs*3) variant in dark red, C1 p.(His22_Lys24del) variant in light red, C1 subunit in yellow, C3 subunit in cyan, C4 subunit in light pink, and C6 subunit in hot pink.

Figure 4 .
Figure 4.The compound heterozygous mutation in the C1 protein.(A) C1 p.(Val121Alafs*3) and C1 p.(His22_Lys24del) variants are predicted in the amino acid sequence of the C1 protein.Shown are the reference and the predicted sequence of the protein for both variants and the nomenclature for C1 mutations and variants.(B) C1 p.(Val121Alafs*3) and C1 p.(His22_Lys24del) are highlighted in red in the structure of the C1 protein.Interacting subunits of the TRAPP core with C1 are included partially in the figure.Shown are C1 p.(Val121Alafs*3) variant in dark red, C1 p.(His22_Lys24del) variant in light red, C1 subunit in yellow, C3 subunit in cyan, C4 subunit in light pink, and C6 subunit in hot pink.C1 p.(Val121Alafs*3) (expressed from C1 c.362_363del) was lethal and suggested that the paternal variant C1 p.(His22_Lys24del) (expressed from C1 c.64_72del) was functional at endogenous Bet5p levels.Replacement with either wild-type C1 or C1 (c.64_72del) was confirmed by colony PCR (Figure 5C, upper panel) and DNA sequencing (Figure 5D, lower panel).These results confirm the functionality of C1 (c.64_72del) in yeast and the ability of this mutant to complement the BET5 gene.

Figure 5 .
Figure 5. C1 (c.64_72del) but not C1 (c.362_363del) can compensate for the loss of BET5 in yeast.(A) Haploid yeast strain bet5∆ kept alive with URA3-based BET5 was transformed with an empty plasmid, with a vector containing BET5 under its endogenous promoter or with vectors expressing either C1, C1 (c.362_363del), or C1 (c.64_72del) under the control of the ADH1 promoter.Transformed cells were plated on 5-FOA to counter-select for the URA3-based plasmids.(B) Yeast cells were transformed with either C1, C1 (c.362_363del), or C1 (c.64_72del) repair templates in the presence of the self-containing CRISPR/Cas9 plasmid.In one control (left-most), the cells were not transformed with guide RNA or repair template to evaluate the transformation efficiency.A second control (second from left) used only gRNA to demonstrate the lethality of a DSB.The presence or absence of growth in the remaining plates is an indication of rescue or no rescue from the DSB lethality of the Cas9 nuclease when an appropriate repair template is provided.(C) An image from the colony PCR products of the transformation of yeast strains with C1 (c.64_72del) using humanized C1 strain and parental strain as controls.The expected band size for each successful transformation is 274 bp for C1 and 265 bp for C1 (c.64_72del).(D) Electropherogram of the humanized yeast locus with C1 (c.64_72del).The start and stop codons are shown in red boxes, the location of the 9-nucleotide deletion for C1 (c.64_72del) ORF is indicated and the corresponding UTR regions for the BET5 yeast locus are underlined in red.

Figure 5 . 32 3. 6 .
Figure 5. C1 (c.64_72del) but not C1 (c.362_363del) can compensate for the loss of BET5 in yeast.(A) Haploid yeast strain bet5∆ kept alive with URA3-based BET5 was transformed with an empty plasmid, with a vector containing BET5 under its endogenous promoter or with vectors expressing either C1, C1 (c.362_363del), or C1 (c.64_72del) under the control of the ADH1 promoter.Transformed cells were plated on 5-FOA to counter-select for the URA3-based plasmids.(B) Yeast cells were transformed with either C1, C1 (c.362_363del), or C1 (c.64_72del) repair templates in the presence of the self-containing CRISPR/Cas9 plasmid.In one control (left-most), the cells were not transformed with guide RNA or repair template to evaluate the transformation efficiency.A second control (second from left) used only gRNA to demonstrate the lethality of a DSB.The presence or absence of growth in the remaining plates is an indication of rescue or no rescue from the DSB lethality of the Cas9 nuclease when an appropriate repair template is provided.(C) An image from the colony PCR products of the transformation of yeast strains with C1 (c.64_72del) using humanized C1 strain and parental strain as controls.The expected band size for each successful transformation is 274 bp for C1 and 265 bp for C1 (c.64_72del).(D) Electropherogram of the humanized yeast locus with C1 (c.64_72del).The start and stop codons are shown in red boxes, the location of the 9-nucleotide deletion for C1 (c.64_72del) ORF is indicated and the corresponding UTR regions for the BET5 yeast locus are underlined in red.

Figure 6 .Figure 6 .
Figure 6.Growth efficiency of humanized C1 (c.64_72del) strain in solid and liquid media.(A) Images from the plates of parental strain (MSY135) and humanized yeast strains with C1 and C1 (c.64_72del).Serial dilutions of the liquid cultures were spotted on YPD plates and incubated for 48 h at three different temperatures, 30 °C, 35 °C, and 37 °C.(B) Representation of the growth curves Figure 6.Growth efficiency of humanized C1 (c.64_72del) strain in solid and liquid media.(A) Images from the plates of parental strain (MSY135) and humanized yeast strains with C1 and C1 (c.64_72del).Serial dilutions of the liquid cultures were spotted on YPD plates and incubated for 48 h at three different temperatures, 30 • C, 35 • C, and 37 • C. (B) Representation of the growth curves in liquid medium for MSY135 and humanized yeast strains with C1 and C1 (c.64_72del) at 30 • C, 35 • C, and 37 • C. Shown are the averages of OD600 for three biological replicates.

Figure 7 .
Figure 7. Non-selective autophagy and secretion are affected in the humanized C1 (c.64_72del) strain.(A) The parental strain MSY135 and the humanized strains with C1 and C1 (c.64_72del) were grown in rich (+N panel) or nitrogen-starved (-N panel) media to induce autophagy at three different temperatures, 30 °C, 35 °C, and 37 °C, as described in the section 2. The mutant strains atg1∆ and trs85∆ were used as controls to show defects in autophagy.Equal amounts of protein were fractionated by SDS-PAGE and analyzed by Western blot analysis using anti-Ape1p IgG.(B) The parental strain MSY135 and the humanized strains with C1 and C1 (c.64_72del) expressing plasmid-based GFP-Atg8 under the control of the endogenous ATG8 promoter were grown in nitrogen-starved medium (SD-N) to induce autophagy at 30 °C.The mutant strains atg1∆ and trs85∆ were used as controls to show defects in autophagy.At the indicated times, aliquots were removed and protein extracts were prepared and fractionated by SDS-PAGE, and analyzed by Western blot analysis using anti-GFP antibody.(C,D) Yeast cells (wild-type, sec18, humanized C1 and C1 (c.64_72del)) were grown in medium with 0.05% glucose and 2% sucrose to induce expression of the SUC2 gene that

Figure 7 .
Figure 7. Non-selective autophagy and secretion are affected in the humanized C1 (c.64_72del) strain.(A) The parental strain MSY135 and the humanized strains with C1 and C1 (c.64_72del) were grown in rich (+N panel) or nitrogen-starved (-N panel) media to induce autophagy at three different temperatures, 30 • C, 35 • C, and 37 • C, as described in the Section 2. The mutant strains atg1∆ and trs85∆ were used as controls to show defects in autophagy.Equal amounts of protein were fractionated by SDS-PAGE and analyzed by Western blot analysis using anti-Ape1p IgG.(B) The parental strain MSY135 and the humanized strains with C1 and C1 (c.64_72del) expressing plasmidbased GFP-Atg8 under the control of the endogenous ATG8 promoter were grown in nitrogen-starved medium (SD-N) to induce autophagy at 30 • C. The mutant strains atg1∆ and trs85∆ were used as controls to show defects in autophagy.At the indicated times, aliquots were removed and protein extracts were prepared and fractionated by SDS-PAGE, and analyzed by Western blot analysis using

Figure 8 .
Figure 8. Patient fibroblasts harboring the compound heterozygous TRAPPC1 mutation show a defect in autophagy.(A) Control (CTRL#1) and patient fibroblasts (S1) were starved (EBSS medium) for up to 2 h and cell lysates were collected at 0 h, 1 h, and 2 h time points.Shown is a representative Western blot where the expected bands for LC3-I, LC3-II, and tubulin (loading control) are indicated.LC3-I migrates at ~17 kDa and LC3-II migrates at ~15 kDa.(B) Quantification of the Western blots with LC3-II signal normalized to the tubulin loading control.Quantification was performed with ImageJ.The error bars indicate SEM.

Figure 8 .
Figure 8. Patient fibroblasts harboring the compound heterozygous TRAPPC1 mutation show a defect in autophagy.(A) Control (CTRL#1) and patient fibroblasts (S1) were starved (EBSS medium) for up to 2 h and cell lysates were collected at 0 h, 1 h, and 2 h time points.Shown is a representative Western blot where the expected bands for LC3-I, LC3-II, and tubulin (loading control) are indicated.LC3-I migrates at ~17 kDa and LC3-II migrates at ~15 kDa.(B) Quantification of the Western blots with LC3-II signal normalized to the tubulin loading control.Quantification was performed with ImageJ.The error bars indicate SEM.

Figure 9 .Figure 9 .
Figure 9. Trafficking from the ER-to-Golgi is delayed and Golgi morphology is altered in patient fibroblasts harboring a compound heterozygous TRAPPC1 mutation.(A) The RUSH assay was performed on control, S1 patient fibroblasts, and S1 + RFP-C1 patient fibroblasts using ST-eGFP cargo protein.The cells were imaged every 2 min over a period of 30 min upon biotin addition, which Figure 9. Trafficking from the ER-to-Golgi is delayed and Golgi morphology is altered in patient fibroblasts harboring a compound heterozygous TRAPPC1 mutation.(A) The RUSH assay was performed on control, S1 patient fibroblasts, and S1 + RFP-C1 patient fibroblasts using ST-eGFP cargo protein.The cells were imaged every 2 min over a period of 30 min upon biotin addition, which causes the release of the cargo protein from the ER.Fluorescence intensity in the Golgi was quantified using ImageJ as described in Koehler et al., 2017.Representative images of the cells in each condition for the time points indicated are shown in Figure S2.(B) Control fibroblasts (CTRL), patient fibroblasts untransfected (S1), or fibroblasts transfected with a C1 construct fused to RFP (S1 + RFP-C1) were fixed and stained for manosidase II to visualize Golgi and Hoechst to visualize the nucleus.The number of Golgi fragments per cell was quantified using Imaris as described in the Materials and Methods section.For estimating the statistical significance, one-way ANOVA was used with post hoc Tukey HSD analysis.The error bars indicate SEM.The N values for the control, S1, and S1 + RFP-C1 are 75, 78, and 81, respectively, in three different biological replicates.** indicates p < 0.01; **** indicates p < 0.0001.

Table 2 .
Plasmids used in this study.

Table 3 .
Yeast strains used in this study.

Table 4 .
Oligonucleotides designed for each sgRNA targeting the yeast TRAPP genes.

Table 5 .
Oligonucleotides designed for each repair template.

Table 6 .
Primers used to confirm humanization of yeast.