EXAMENSARBETE INOM BIOTEKNIK, AVANCERAD NIVÅ, 30 HP STOCKHOLM, SVERIGE 2017

Proteomic Profiling of Vesicular Organelles

HANNA HASSAN

KTH SKOLAN FÖR BIOTEKNOLOGI KTH Biotechnology

Proteomic Profiling of Vesicular Organelles Hanna Hassan Supervisor: Peter Thul Examiner: Emma Lundberg 2016

Abstract The Human Atlas (HPA) is a scientific research project aiming to map the entire human proteome using an antibody-based approach. The Subcellular Atlas that is a part of the HPA provides spatial information about the subcellular location of . Currently, the Subcellular Atlas does not go beyond the annotation “vesicles” for proteins showing a dot-like staining. In this study a proteomic profiling of three vesicular-like organelles, , endosomes and lysosomes was conducted to expand the annotations in the HPA’s subcellular atlas and to identify new members of the respective proteomes. The proteomic profiling was done by indirect immunofluorescence and colocalization studies with organelle specific markers. By this method, the location of proteins that were previously described to target one of the studied organelles could be verified, while the subcellular location of so far uncharacterized proteins could be identified. In total, 91 proteins were identified to localize to the three studied vesicular organelles. This includes proteins whose subcellular location has previously been described such as PEX14 (), LAMTOR4 (lysosome) and Rabankyrin-5 (endosome), but also proteins like HEATR4 (peroxisome) and PIK3R1 (lysosome) that were not associated with any organelle before. The observations in this study serve as a validation that colocalization studies are useful when verifying a previously described subcellular location and identifying unknown locations of proteins. Moreover, the high-resolution immunofluorescent images allowed single-cell analyses showing proteins that are only expressed in a subpopulation of cells, proteins that are only localized to a subpopulation of organelles or proteins that are only present in a subcompartment of the organelle. Altogether, this study has verified known locations as well as identified the location of proteins, which will result in more precise annotations for vesicles in the HPA’s Subcellular Atlas.

Sammanfattning Human Protein Atlas (HPA) ar¨ ett vetenskapligt forskningsprojekt vars mal˚ ar¨ att kartlagga¨ det manskliga¨ proteomet med hjalp¨ av antikroppar. HPAs subcellulara¨ atlas ger information om ett proteins subcellulara¨ plats. For¨ tillfallet¨ benamns¨ bilder i denna atlas som har en prickliknande infargning¨ endast som lokaliserade till ”vesiklar”. I denna studie forekommer¨ en protein-profilering av de tre vesikel-liknande organellerna peroxisomer, endosomer och lysosomer for¨ att utoka¨ HPAs subcellulara¨ atlas och identifiera nya medlemmar av respekive proteom. Denna protein- profilering gjordes genom indirekt immunofluorescensteknik och co-lokaliseringsstudier med organell specifika markorer.¨ Tidigare kanda¨ platser for protein blev validerade och okanda¨ platser for¨ protein blev identifierade. Totalt blev 91 protein lokaliserade till de tre studerade organellerna. Detta inkluderar protein vars subcellulara¨ plats tidigare beskrivits sa˚ som PEX14 (peroxisom), LAMTOR4 (lysosom) och Rabankyrin-5 (endosom), men aven¨ protein som HEATR4 (peroxisom) och PIK3R1 (lysosom) som inte varit associerade med nagon˚ organell. Resultaten i denna studie validerar co-lokaliseringsstudier som anvandbara¨ nar¨ verifiering av ett tidigare kant¨ subcellulart¨ omrade˚ och identifiering av okanda¨ omraden˚ protein tillhor¨ studeras. De hoguppl¨ osta¨ immunofluorescerande bilderna tillat¨ enkelcell-analyser som visade protein som endast uttrycktes i en underpopulation av celler, protein som endast lokaliserade till en underpopulation av organeller eller protein som endast finns i ett delomrade˚ av en organell. Denna studie har verifierat kanda¨ platser och identifierat platser protein befinner sig i vilket kommer resultera i en mer detaljerad annotering for vesiklar i HPAs subcellulara¨ atlas.

Keywords Proteomic profiling — Colocalization — Immunostaining — Peroxisome — Endosome — Lysosome

1 Proteomic Profiling of Vesicular Organelles — 2/22

Contents

Introduction 3 Vesicles...... 3 Proteomic Profiling of Organelles...... 3 Aim...... 3 Materials and Methods 3 Data Analysis...... 3 Cell Cultivation...... 4 Immunostaining...... 4 Image Acquisition...... 4 Image Analysis...... 4 Results 4 Immunostaining of Proteins...... 4 Members of the Peroxisomal Proteome...... 4 Members of the Endolysosomal Proteome...... 5 Candidate Proteins with an Unknown Location...... 9 Discussion 9 Immunostaining of Proteins...... 9 Colocalization with Organelle Markers...... 9 Candidate Proteins with an Unknown Location...... 9 Proteomic Profiling of Organelles...... 9 Conclusion 10 Future Perspectives 10 Acknowledgments 10 References 10 Appendix 12 Proteomic Profiling of Vesicular Organelles — 3/22

Introduction Proteomic Profiling of Organelles Identifying proteins that are located in a certain organelle HE Human Protein Atlas (HPA) is a project aiming can be approached in different ways. The general way is the to map the entire human proteome through antibody- isolation of the organelle of interest followed by the identi- based proteomics and transcriptomics. The first ver- T fication of the residing proteins mass by mass spectrometry. sion of the HPA was released in 2005 and as of April 2016, The results are then analyzed by a protein database to deter- the database has been updated with its fifteenth release. [1][2] mine which proteins were found in the sample. [10] Another The atlas is divided into four subparts: normal tissue, cancer approach is to use immunofluorescence or green fluorescent tissue, subcellular and cell line. Today, the database contains protein (GFP). These approaches use fluorescence to visual- protein data covering 86% of the predictive human ize e.g. molecules and organelles. Locations of subcellular and more than 11 million images with primary data from im- proteins can be determined by labeling proteins of interest munofluorescence (IF) and immunohistochemistry. [3] The with fluorescent dyes, followed by image acquisition using HPA’s Subcellular Atlas provides spatial information about fluorescence microscopy. There are two main techniques that the subcellular location of proteins. Identifying the locations are used for labeling proteins, namely tagging with a flu- of proteins is important for a better understanding of the in- orescent protein and immunolabeling with fluorophores (also teraction and cellular functions of proteins. Proteins rarely known as immunofluorescence). This study will focus on act alone and if two proteins are located to the same location, using indirect immunofluorescence to detect proteins. Indirect it is more likely that they interact with each other or have an immunofluorescence uses a primary and secondary antibody effect on the function. [4] This information can in turn also to target and detect the protein of interest. The primary anti- be valuable to develop more effective treatments for different body targets the protein of interest and is indirectly coupled diseases by e.g. simplifying the target identification process to a fluorophore by the use of a secondary antibody. The fluo- for drug discovery. Currently, the HPA does not go beyond rophore enables the detection of the targeted protein through the annotation category “vesicles” for proteins with a dot-like visualization. [11] Colocalization is the comparison of the staining. However, there are different types of vesicles and spatial distribution of two fluorescently labeled molecules and vesicle-like organelles in the cell. Therefore a more detailed can be used to determine if two molecules are associated to annotation that shows what type of organelle this protein is the same organelle. In this method, the protein of interest is located in is needed. The aim of this study is to expand the targeted with a fluorophore and the organelle of interest with annotation category ”vesicles” in to a more specific one by another fluorophore. If the signal of the fluorophores overlap, using resources from the HPA’s subcellular atlas. This was the protein localizes to the organelle. [12] achieved by conducting proteomic profiling of three vesicular- like organelles through antibody-based techniques. Aim The main aim of this project is to expand the annotations for Vesicles proteins with a vesicle-like pattern in the HPA by identifying Vesicles, a category comprised of many different organelles, proteins located in peroxisomes, endosomes and lysosomes. are small structures enclosed by a membrane and are involved Using indirect immunofluorescence, colocalization with or- in a variety of functions such as transportation, sorting, stor- ganelle specific markers and confocal microscopy, the proteins ing and metabolism. The vesicular-like organelles that will localized to these organelles can be determined. Another aim be studied in this project are peroxisomes, endosomes and of this project is to validate HPA-antibodies by using them lysosomes. Peroxisomes are multifunctional organelles that to identify the subcellular location of proteins with a known harbour a variety of enzymes and are involved in several an- location. abolic and catabolic cellular pathways. The main function of peroxisomes is β-oxidation of long- and very long-chain Materials and Methods fatty acids. In addition, peroxisomes carry out important reactions such as phospholipid biosynthesis, chemical detox- Data Analysis ification, oxidation of purines, polyamines and some amino Using the UniProtKB database, lists of proteins with the sub- acids. Many of these reactions involve hydrogen peroxide, cellular location peroxisome, endosome or lysosome, that which is why they were named peroxisomes. [5] Endosomes were asserted experimentally were generated. The gener- are formed via endocytosis and their main function is to sort ated lists contained proteins with known subcellular locations. the material that has been taken up by the cell. There are three These lists were ran against a compiled table of HPA-antibody main types: early, late and recycling. Early endosomes mature annotations. The compiled HPA-antibody table consisted of into late endosomes that fuse with lysosomes for degradation antibodies annotated to be targeting proteins with a positive of material. Recycling endosomes are sub-compartments of vesicular staining in U-2 OS cells. The HPA-antibody table the early endosome that recycle material to the plasma mem- entries also had a FPKM value over 0.5, ensuring that the brane. [6][7] Lysosomes are organelles that contain several gene of interest was expressed in the cell line. The proteins different enzymes and are responsible for the degradation of from the UniProtKB lists that were found in the compiled molecules internalized within the cell. [8][9] HPA-table were chosen as candidates to be tested. A model Proteomic Profiling of Vesicular Organelles — 4/22 for the identification of novel peroxisomal, lysosomal and Image Acquisition endosomal proteins was built in KNIME (v. 3.1.2) [13] using Images of the immunofluorescently stained U-2 OS cells were the image segmentation data that is generated for all images in acquired manually in sequential steps (one for each dye) to the Subcellular Atlas. [14] Images of identified members were minimize spectral bleed-through. This was done in room tem- used to train the model. The number of extracted features perature using Leica SP5 DM6000 CS confocal microscope, was reduced by removing features that showed only a low a 63X/1.4 numerical aperture oil immersion objective and variance for vesicular staining. The features of the identified the software LAS AF (Leica Microsystems). The confocal members were used to define the center of cluster by k-means. settings were as follows: 16 bit acquisition, 600 Hz, line aver- A list of vesicular proteins that were closest to the cluster age 2, pixel size 80 nm. The detector gain was adjusted for center was generated. The generated list was compared to each sample to obtain an image with a high signal to noise entries in UniProtKB and the proteins that had no or a curated ratio and the maximum allowed gain was set to 800 for the subcellular location were chosen as candidates for testing. HPA-antibody channel.

Image Analysis Cell Cultivation The acquired images were visually inspected for a colocaliza- U-2 OS cells, a cell line derived from osteosarcoma, grew tion between the protein of interest and the organelle marker. in McCoy’s 5A medium (Sigma-Aldrich and Lonza) supple- The images were analyzed and merged by using ImageJ. mented with 10% Fetal Bovine Serum (FBS, Sigma-Aldrich) ◦ and 1% L-glutamine in a 37 C and 5.2% CO2 environment. 10 000 cells/well were seeded onto a 96-well glass bottom Results plate (Greiner) coated with 40 µl 12.5 µg/ml fibronectin and incubated for 24h before immunostaining, in the same envi- Immunostaining of Proteins ronment and temperature. The aim of this study was to identify proteins localized to peroxisomes, endosomes and lysosomes. This was done by using organelle markers for these three organelles together Immunostaining with HPA-antibodies targeting previously selected protein can- The growth medium was removed from the 96-well glass didates to immunostain U-2 OS cells. The manually acquired bottom plate and the cells were washed with PBS. The images of cells were visually inspected for colocalization cells were then fixed by incubation with 40 µl ice cold 4% between proteins of interest and organelle markers. The HPA- paraformaldehyde (PFA, VWR) diluted in PBS for 15 min- antibody was labeled green, the nuclei was stained blue with utes. After removal of PFA, the cells were permeabilized by DAPI and the organelle markers for peroxisomes, endosomes incubating with 40 µl PBS containing 0.1% Triton X-100 and lysosomes were labeled with different fluorophores. In the (Sigma-Aldrich) for 3x5 minutes followed by a washing step figures shown in this result section all the organelle markers with PBS. Primary rabbit HPA-antibodies were diluted to are shown in red to enable easier detection of colocalization 2-4 µg/ml in blocking buffer (PBS + 4% FBS) containing an with the green HPA-antibody. organelle specific marker. Organelle markers used: mono- The candidates were divided into one out of four different cat- clonal mouse anti-ABCD3 diluted 1:1000 (Atlas Antibodies) egories depending on the type of overlap the images showed for the peroxisome, monoclonal mouse anti-VPS26A diluted (Fig.1). The categories were: 1:800 (Atlas Antibodies) and a monoclonal mouse anti-EEA1 diluted 1:250 (BD Biosciences) for the endosome and a poly- i Complete overlap: vesicles detected by the HPA- clonal sheep anti-LAMP-1 diluted 1:80 (R&D systems) for antibody (green channel) and organelle marker (red the lysosome. Endosome and lysosome organelle markers channel) colocalize. were used together. After incubation overnight in 4◦C the cells were washed with 40 µl PBS 4x10 minutes. 40 µl blocking ii Partial overlap: not all vesicles detected by the organelle buffer containing 1 µg/ml of the secondary antibodies goat marker colocalize with the HPA-antibody staining anti-rabbit Alexa 488, goat anti-mouse Alexa 555 and donkey iii Fractional overlap: a fraction of the vesicle detected by anti-sheep Alexa 633 (Life Technologies) was added and the organelle marker colocalizes with the HPA-antibody incubated in room temperature for 90 minutes. Subsequently, staining. A fractional overlap was only seen when using the blocking buffer containing antibodies was removed and endosomal organelle markers. the nuclei of the cells were stained with the addition of 50 µl DAPI (Invitrogen) diluted 1:500 in PBS. After 10 minutes iv No overlap: vesicles detected by the HPA-antibody and DAPI was removed and the cells were washed with 40 µl organelle marker do not colocalize. PBS 4x10 minutes. The plate was then mounted with 85% glycerol in 10xPBS and sealed with foil cover. Members of the Peroxisomal Proteome Out of 48 proteins with a known location that were tested, 19 showed a complete overlap, 7 a partial overlap and 22 Proteomic Profiling of Vesicular Organelles — 5/22

Figure 1. Three categories of overlap patterns. First row represents a complete overlap, second row a partial overlap and the third row a fractional overlap. The HPA (Green) column shows the HPA-antibody and the marker (Red) column an organelle marker. of them had no signal overlapping with the peroxisomal or- Members of the Endolysosomal Proteome ganelle marker. One of the tested proteins SCP2 (non-specific lipid-transfer protein) showed a distinct staining in only some The endosomal and lysosomal candidates with a known loca- cells. (see Fig. 2 and Appendix 2.) 57 proteins with an un- tion were tested together as the organelles have no clear-cut known location were tested with the peroxisomal organelle boundary between them. Both organelles had 55 proteins marker. These proteines were predicted by a trained model selected as candidates to be tested each and 20 of these candi- in KNIME for having a similar staining pattern with proteins dates were described to be located in both locations. There- known to localize to the peroxisome and lysosome. 5 out of fore 90 candidates in total were tested for colocalization with 57 candidates showed a complete overlap and 52 showed no the LAMP-1 antibody as a marker for lysosomes and the overlapping signal with the peroxisomal marker. VPS26A antibody for endosomes. Out of these 90 candidates 12 showed a complete overlap, 17 a partial overlap and 61 Proteomic Profiling of Vesicular Organelles — 6/22

Figure 2. Immunostaining of three different proteins. HPA-antibody is shown in green, the peroxisomal marker ABCD3 in red and nuclei stained with DAPI in blue. A) IF-staining of a known peroxisome protein, peroxisomal membrane protein PEX14. B) IF-staining of an unknown peroxisomal protein, HEAT repeat-containing protein 4, showing a complete overlap with the peroxisomal organelle marker. C) IF-staining of a known peroxisomal protein SCP2, non-specific lipid-transfer protein, showing a complete overlap with one cell. This is an example of a subpopulation of peroxisomes between cells. no overlap with the lysosomal organelle marker. For the en- a complete overlap, 1 a partial overlap, 10 a fractional overlap dosomal marker VPS26A, 3 showed a complete overlap, 6 a and 47 no overlapping signal with the endosomal marker. partial overlap, 8 a fractional overlap and 73 no overlap. The 57 proteins with an unknown location were tested with the three proteins showing a complete overlap are all associated lysosomal organelle marker. These proteins were predicted with the retromer, a complex of proteins that is involved in by a trained model in KNIME for having a similar staining recycling transmembrane receptors from endosomes to the pattern with proteins known to localize to the peroxisome and trans-Golgi network, according to the UniProtKB database. lysosome. Out of the 57 candidates, 1 showed a complete The VPS26A organelle marker is therefore thought to be an an- overlap and 4 a partial overlap with the lysosomal marker. 52 tibody more suited for targeting the retromer complex rather showed no overlapping signal with the lysosomal organelle than the entire endosome. The early endosomal organelle marker. For examples of acquired images see Fig. 3. marker EEA1 was thereafter used. 60 candidates were tested In total, 91 proteins (97 antibodies) out of 179 proteins in total with this marker, 55 candidates from the first experi- (195 antibodies) were identified to localize to the three studied ment described above and 5 additional candidates (lysosomal vesicular organelles. Table 1 summarizes the amount of anti- candidates that showed a colocalization with the former endo- bodies tested for each vesicle-like organelle and what overlap somal marker VPS26A). Out of these 60 candidates 2 showed category they fall under. The number seen in parenthesis de- Proteomic Profiling of Vesicular Organelles — 7/22

Figure 3. Immunostaining of four different proteins. HPA-antibody is shown in green, the organelle marker in red and nuclei stained with DAPI in blue. A) IF-staining of a known lysosomal protein, HLA class II histocompatibility antigen DP alpha 1 chain. B) Staining of a known endosomal protein, Rabankyrin-5. C) IF-staining of a protein known to be a part of the retromer complex, vacuolar protein sorting-associated protein 35. D) IF-staining of an unknown lysosomal protein, phosphatidylinositol 3-kinase regulatory subunit alpha, showing a complete overlap with the lysosomal organelle marker scribes the amount of proteins targeted as the same protein HPA-antibodies. In Table 3 the amount of vesicular organelles could be targeted by several antibodies. If available, more than seen in the cytoplasm are presented. The number for each one antibody per protein was used in order to validate HPA- organelle is an average count of vesicles when comparing antibody stainings. However, in Table 2 only the amount of three cells from the same image. One acquired image was antibodies used is shown due to only one antibody being used chosen for each organelle. Tables containing the antibodies per protein. This is because identifying previously unknown tested for each organelle marker can be found in Appendix 1. members of the organelles’ proteomes does not validate the Proteomic Profiling of Vesicular Organelles — 8/22

Table 1. Results after testing protein candidates with a known subcellular location. The amount of antibodies tested followed by the number of proteins tested in parenthesis are shown. Category Peroxisome Endosome Lysosome Retromer Complete overlap 19(16) 2 12 3 Partial overlap 7(6) 1 17 6 Fractional overlap 0 10 0 8 No overlap 22(16) 47(44) 61(55) 73(68)

Total 48(38) 60(57) 90(84) 90(84)

Table 2. Results after testing protein candidates with an unknown subcellular location. The amount of antibodies tested are shown. Category Peroxisome Lysosome Complete overlap 5 1 Partial overlap 0 4 Fractional overlap 0 0 No overlap 52 52

Total 57 57

Table 3. The amount of vesicular organelles visualized by the organelle markers. The number shown below is the average vesicle count for each organelle when comparing three cells in the same acquired image. Organelle Amount of vesicles Peroxisome 140 Endosome 118 Lysosome 64

Table 4. Proteins with a previously unknown location colocalizing with one of the organelle markers. Protein Gene name Subcellular location HEAT repeat-containing protein 4 HEATR4 Peroxisome Annexin A10 ANXA10 Peroxisome Transmembrane protein 41B TMEM41B Peroxisome Proline-rich transmembrane protein 4 PRRT4 Peroxisome Zinc finger protein 709 ZNF709 Peroxisome Phosphatidylinositol 3-kinase regulatory subunit alpha PIK3R1 Lysosome Proteomic Profiling of Vesicular Organelles — 9/22

Candidate Proteins with an Unknown Location proteins belong to the organelle they were observed to localize Acquired images of proteins that colocalize with the used or- to. This is important since unspecific binding is a major issue ganelle markers were used to train a model for the prediction when drawing conclusions from studies based on a single of proteins that have similar features. The predicted protein antibody. Preferably, the location should be verified by an candidates were tested if they had a previously unknown lo- antibody-independent method. In this study certain antibodies cation. Out of the 57 tested candidates, 10 colocalized either targeted the same protein and as the results showed that the partially or completely with the organelle markers for the per- outcome was the same for both antibodies this could serve as oxisome and lysosome. The tested candidates that showed a another type of validation. However, to raise the validity of complete overlap with an organelle marker signal are shown the results, verifying antibody specificity would be necessary. in Table 4. Antibody specificity in IF can be validated by a reduction of the protein expression, which would correlate with a decrease Discussion of the antibody signal in IF. Methods to reduce the protein expression are for example the use of CRISPR for a gene Immunostaining of Proteins knockout, which completely removes the protein, or a knock- The aim of this study was to identify proteins localized to per- down through siRNAs. The 10 HPA-antibodies that showed oxisomes, endosomes and lysosomes by immunostaining with an overlap have not been validated through siRNA knockdown specific organelle markers. The immunostaining in this study or a gene knockout. However they have all been validated showed four different colocalization patterns: a complete, through Western blot. E.g. the used HPA-antibody targeting partial, fractional and no overlap. A complete overlap with HEAT repeat-containing protein 4 was seen to completely the specific organelle marker signal shows that the targeted overlap with the signal from the peroxisome organelle marker. protein is located in all organelles detected by the marker The Western blot results showed a single band corresponding whereas a partial overlap shows that the targeted protein is to the predicted size in kDa for this antibody. This result only located in some organelles. A fractional overlap with the makes it reasonable to concludes that the antibody used is spe- organelle marker signal, which was only seen with the endo- cific. Another example is the phosphatidylinositol 3-kinase somal organelle markers, shows that only a part of the labeled regulatory subunit alpha protein that was seen to overlap with organelle colocalizes with the HPA-antibody. This type of the lysosomal organelle marker. The antibody targeting this overlap could be due to optical or chromatographical aberra- protein showed a band of predicted size in kDa (+/-20%) with tions. However, this was not deemed to be the case as this additional bands present in the Western blot. To ensure that overlap pattern was only seen with the endosomal candidates. this antibody is specific, a better Western blot result would be This fractional overlap is thought to be due to the domain the optimal. targeted protein is localized to inside the endosome. If the organelle marker and the targeted protein have a fractional overlap, they could be localized to different domains in the Proteomic Profiling of Organelles endosome. [15][16] There are different approaches to study the proteomes of or- ganelles. The vast majority of proteomic studies are based Colocalization with Organelle Markers on mass spectrometry (MS) studies whereas this study used Out of the 91 proteins seen to colocalize with the used or- immunofluorescent labeling. Immunofluorescent labeling ganelle markers, 40 candidate proteins showed a complete on a proteome- and genome-wide level is only done by the overlap with the organelle marker signal. These include HPA. Both of these approaches have limitations when identi- known members of the three studied organelles such as PEX14 fying what proteins are members of the organelles of interest. (peroxisome), LAMTOR4 (lysosome) and Rabankyrin-5 (en- To study the peroxisome, endosome and lysosome, the MS- dosome). This supports that the observed colocalization with based approach would first need to isolate and purify these the chosen markers did show if the protein was a member of organelles from a sample and reduce the complexity. A disad- one of the organelles or not. The 88 candidates that showed no vantage with these steps is that the isolation can be challenging colocalization with the organelle markers could be due to the to succeed with and during the complexity reduction, impor- used antibody being unspecific and therefore targeting some- tant proteins can be lost and will therefore not be able to thing else. Another reason is that the previously determined be identified. [17] The immunofluorescent labeling approach subcellular location of a protein could have been detected in a does not need to isolate, purify or reduce the complexity of specific cell line or in a specific condition different from what the sample and does therefore not lose any proteomic infor- was used in this study. mation. However, this approach can cause artifacts and does require cell fixation as well as permeabilization, which can Candidate Proteins with an Unknown Location be challenging when dealing with different cell lines. [11] It Out of the 57 protein candidates with previously unknown is also limited to the antibodies that are available. The IF- locations that were tested, 10 showed either a partial or a com- approach, if colocalization is used, is not reliant on databases plete overlap with the used organelle markers. These results to identify whether a protein is located in a certain organelle need to be verified to be able to draw the conclusion that these but it is reliant on antibody specificity. The accuracy of the Proteomic Profiling of Vesicular Organelles — 10/22 location visualized of native proteins in the microscope can be ther validation of the results through antibody validation and affected by antibody specificity which can result in mislocal- super-resolution microscopy to analyze the fractional overlap ization. However, it has been shown that immunofluorescent pattern would also be encouraged as well as analyzing more tagging is a reliable technique. [18] Both of these approaches organelles through colocalization in the long run. has validation of protein identification as an obligatory step but this is not always provided in literature, making this a Acknowledgments shared limitation between the approaches. Which one is better than the other depends on what questions you want to answer I would like to express my deepest gratitude to Emma Lund- as both approaches have their advantages and disadvantages. berg who gave me the golden opportunity to work with this For example the MS-based approach can identify many pro- project, for the encouraging comments and support through- teins at once for an isolated organelle whereas the IF-approach out this process. Furthermore, I would like to sincerely thank focuses on the colocalization of one protein at a time (when vi- my supervisor Peter Thul for being so understanding and al- sualizing the results through microscopy). The MS-approach ways answering my questions, for the useful remarks and for is good for quantification, whereas the IF-approach is better at helping me through the ups and downs of this project. I would detecting finer sub-structures or overlaps such as the fractional also like to thank the Cell Profiling group for creating such a overlap observed in this study. The MS-approach would be lovely working environment, helping me when in need and more useful if proteomes were to be compared at different always making my day brighter with laughter! conditions or time points since the IF-approach would need the entire IF antibody repertoire, consequently taking more References time and resources. The IF-approach can however show if the targeted protein is also located in another organelle, a separate [1] Mathias Uhlen,´ Erik Bjorling,¨ Charlotta Agaton, Cristina compartment, if it is cell cycle dependent or expressed in only Al-Khalili Szigyarto, Bahram Amini, Elisabet Andersen, a subpopulation of the cells, which the MS-approach cannot. Ann-Catrin Andersson, Pia Angelidou, Anna Asplund, An example of this is the protein SCP2 (see Fig.2C) which Caroline Asplund, et al. A human protein atlas for normal showed an expression in only some cells where it also over- and cancer tissues based on antibody proteomics. Molec- lapped with the peroxisome organelle marker. This protein ular & Cellular Proteomics, 4(12):1920–1932, 2005. in particular has an isoform which is located in peroxisomes [2] HPA home page [Internet], 2016 [Cited May 16]. Avail- which could explain why its expression was only seen in some able from: http://www.proteinatlas.org/about/releases. cells and not all. It could also be due to cell cycle dependency [3] or that some cells have a subpopulation of peroxisomes with HPA home page [Internet], 2016 [Cited May 18]. Avail- different content. [19] able from: http://www.proteinatlas.org/about. [4] Andrew R Joyce and Bernhard Ø Palsson. The model or- Conclusion ganism as a system: integrating’omics’ data sets. Nature Reviews Molecular Cell Biology, 7(3):198–210, 2006. In this study 91 proteins localized to the studied vesicular or- [5] ganelles could be determined and the acquired images will be Vasily D Antonenkov, Silke Grunau, Steffen Ohlmeier, publicly available on the Human Protein Atlas. Many of the and J Kalervo Hiltunen. Peroxisomes are oxidative or- tested proteins were also found to not colocalize with an or- ganelles. Antioxidants & redox signaling, 13(4):525–537, ganelle marker. However, the obtained results do suggest that 2010. colocalization studies could be a good approach to validate the [6] Nature home page [Internet], 2016 [Cited May 20]. Avail- known subcellular location of proteins with organelle markers able from: http://www.nature.com/subjects/endosomes. or map the proteome of vesicular organelles. Mapping these [7] I. Dikic. Endosomes. Molecular Biology Intelligence proteomes would further expand the annotations for proteins Unit. Springer New York, 2008. located in vesicular organelles in the Human Protein Atlas. [8] This will enable users of the atlas to get a more specific subcel- E. Holtzman. Lysosomes. Cellular Organelles. Springer lular location, giving them valuable information for medical US, 2013. and biological research. [9] Nature home page [Internet], 2016 [Cited May 20]. Avail- able from: http://www.nature.com/subjects/lysosomes. Future Perspectives [10] Uwe Michelsen and Jorg¨ von Hagen. Isolation of subcel- The main aim of this study was to identify the location of lular organelles and structures. Methods in enzymology, known and unknown peroxisomal, endosomal and lysoso- 463:305–328, 2009. mal proteins. In retrospect, many antibodies were not tested [11] Ben NG Giepmans, Stephen R Adams, Mark H Ellisman, due to time constraint. Therefore a future step for this study and Roger Y Tsien. The fluorescent toolbox for assessing would be to test more antibodies to detect more proteins that protein location and function. Science, 312(5771):217– are members of these vesicular organelles’ proteomes. Fur- 224, 2006. Proteomic Profiling of Vesicular Organelles — 11/22

[12] Kenneth W Dunn, Malgorzata M Kamocka, and John H on endosomes in the recycling pathway visualized by McDonald. A practical guide to evaluating colocaliza- multicolor imaging of rab4, rab5, and rab11. The Journal tion in biological microscopy. American Journal of of cell biology, 149(4):901–914, 2000. Physiology-Cell Physiology, 300(4):C723–C742, 2011. [17] [13] Gert Lubec and Leila Afjehi-Sadat. Limitations and KNIME home page [Internet], 2016 [Cited October 11]. pitfalls in protein identification by mass spectrometry. Available from: https://www.knime.org/ Chemical reviews, 107(8):3568–3584, 2007. [14] Justin Y Newberg, Jieyue Li, Rao Arvind, Fredrik Ponten,´ Mathias Uhlen,´ Emma Lundberg, Robert F Murphy. Au- [18] Charlotte Stadler, Elton Rexhepaj, Vasanth R Singan, tomated analysis of human protein atlas immunofluores- Robert F Murphy, Rainer Pepperkok, Mathias Uhlen,´ cence images. IEEE International Symposium on Biomed- Jeremy C Simpson, and Emma Lundberg. Immunofluo- ical Imaging: From Nano to Macro, p. 1023–1026, 2009. rescence and fluorescent-protein tagging show high corre- Nature [15] Yannis Kalaidzidis, Inna Kalaidzidis, and Marino Zerial. lation for protein localization in mammalian cells. methods, 10(4):315–323, 2013. A probabilistic method to quantify the colocalization of markers on intracellular vesicular structures visualized [19] Markus Islinger, Afsaneh Abdolzade-Bavil, Sven Liebler, by light microscopy. In AIP Conference Proceedings, Gerhardt Weber, and Alfred Volkl.¨ Assessing hetero- volume 1641, page 580, 2015. geneity of peroxisomes: isolation of two subpopulations [16] Birte Sonnichsen,¨ Stefano De Renzis, Erik Nielsen, Jens from rat liver. Liver Proteomics: Methods and Protocols, Rietdorf, and Marino Zerial. Distinct membrane domains pages 83–96, 2012. Proteomic Profiling of Vesicular Organelles — 12/22

Appendix Appendix 1. Tables containing the antibodies tested for each organelle marker. Information such as what type of overlap was observed and if the location of the targeted protein was previously known or unknown can be seen below.

Table 5. Peroxisome Organelle Marker ABCD3 Antibody Ensembl id Gene name Overlap Location HPA000262 ENSG00000109436 TBC1D9 None Unknown HPA001242 ENSG00000243978 RGAG1 None Unknown HPA001493 ENSG00000125846 ZNF133 None Unknown HPA001569 ENSG00000145675 PIK3R1 None Unknown HPA002114 ENSG00000115233 PSMD14 None Unknown HPA003251 ENSG00000251369 ZNF550 None Unknown HPA003300 ENSG00000183475 ASB7 None Unknown HPA003642 ENSG00000187105 HEATR4 Complete Unknown HPA003720 ENSG00000181704 YIPF6 None Unknown HPA004909 ENSG00000115592 PRKAG3 None Unknown HPA005469 ENSG00000109511 ANXA10 Complete Unknown HPA005552 ENSG00000068366 ACSL4 None Known HPA005821 ENSG00000085982 USP40 None Unknown HPA006764 ENSG00000060971 ACAA1 None Known HPA007244 ENSG00000060971 ACAA1 Complete Known HPA008862 ENSG00000102910 LONP2 None Known HPA011861 ENSG00000107897 ACBD5 Complete Known HPA012145 ENSG00000107897 ACBD5 Complete Known HPA012571 ENSG00000095970 TREM2 None Unknown HPA014946 ENSG00000166471 TMEM41B Complete Unknown HPA015049 ENSG00000160055 TMEM234 None Unknown HPA016495 ENSG00000106609 TMEM248 None Unknown HPA017322 ENSG00000197601 FAR1 Partial Known HPA017992 ENSG00000166783 KIAA0430 None Known HPA019365 ENSG00000005469 CROT None Known HPA019527 ENSG00000242110 AMACR Partial Known HPA019556 ENSG00000143921 ABCG8 None Unknown HPA019827 ENSG00000165609 NUDT5 None Unknown HPA020099 ENSG00000197943 PLCG2 None Unknown HPA020235 ENSG00000127980 PEX1 Partial Known HPA020260 ENSG00000095321 CRAT None Known HPA020912 ENSG00000242110 AMACR None Known HPA021192 ENSG00000161533 ACOX1 None Known HPA021195 ENSG00000161533 ACOX1 Partial Known HPA021302 ENSG00000133835 HSD17B4 Partial Known HPA021311 ENSG00000133835 HSD17B4 Complete Known HPA021479 ENSG00000133835 HSD17B4 None Known HPA021575 ENSG00000141569 TRIM65 None Unknown HPA021593 ENSG00000115425 PECR Complete Known HPA022130 ENSG00000198721 ECI2 None Known HPA022904 ENSG00000197448 GSTK1 Complete Known HPA022991 ENSG00000167536 DHRS13 None Unknown HPA023157 ENSG00000197566 ZNF624 None Unknown HPA023454 ENSG00000166329 CCDC182 None Unknown HPA023561 ENSG00000171206 TRIM8 None Unknown HPA024785 ENSG00000147647 DPYS None Unknown Continued on next page Proteomic Profiling of Vesicular Organelles — 13/22

Table 5 – continued from previous page Antibody Ensembl id Gene name Overlap Location HPA027135 ENSG00000116171 SCP2 None Known HPA027317 ENSG00000116171 SCP2 Complete (not all cells) Known HPA029012 ENSG00000198198 SZT2 None Known HPA029065 ENSG00000011485 PPP5C Complete Known HPA030209 ENSG00000018510 AGPS Complete Known HPA030211 ENSG00000018510 AGPS Complete Known HPA030226 ENSG00000140548 ZNF710 None Unknown HPA031323 ENSG00000197568 HHLA3 None Unknown HPA031626 ENSG00000198721 ECI2 Partial Known HPA031630 ENSG00000143157 POGK None Unknown HPA031838 ENSG00000115956 PLEK None Unknown HPA032026 ENSG00000117528 ABCD3 None Known HPA032027 ENSG00000117528 ABCD3 Complete Known HPA032141 ENSG00000162928 PEX13 Complete Known HPA035840 ENSG00000087008 ACOX3 Complete Known HPA035870 ENSG00000055147 FAM114A2 None Unknown HPA037000 ENSG00000033178 UBA6 None Unknown HPA038052 ENSG00000110013 SIAE None Unknown HPA039324 ENSG00000087470 DNM1L None Known HPA039584 ENSG00000131379 C3orf20 None Unknown HPA040626 ENSG00000092470 WDR76 None Unknown HPA042830 ENSG00000034693 PEX3 Complete Known HPA044157 ENSG00000212123 PRR22 None Unknown HPA044175 ENSG00000258436 RNASE12 None Unknown HPA044732 ENSG00000183292 TISP43 None Unknown HPA045018 ENSG00000163807 KIAA1143 None Unknown HPA046104 ENSG00000142655 PEX14 Complete Known HPA046373 ENSG00000224940 PRRT4 Complete Unknown HPA047631 ENSG00000242612 DECR2 Partial Known HPA047844 ENSG00000172661 FAM21C None Unknown HPA048668 ENSG00000274349 ZNF658 None Unknown HPA049231 ENSG00000142655 PEX14 Complete Known HPA050036 ENSG00000176896 TCEANC None Unknown HPA050077 ENSG00000101417 PXMP4 Complete Known HPA051282 ENSG00000121691 CAT None Known HPA051966 ENSG00000162735 PEX19 Complete Known HPA052708 ENSG00000100372 SLC25A17 None Known HPA053153 ENSG00000242852 ZNF709 Complete Unknown HPA053646 ENSG00000180011 ZADH2 None Known HPA054039 ENSG00000187498 COL4A1 None Unknown HPA055664 ENSG00000131944 C19orf40 None Unknown HPA055805 ENSG00000188707 ZBED6CL None Unknown HPA055838 ENSG00000121691 CAT None Known HPA056933 ENSG00000011485 PPP5C None Known HPA057168 ENSG00000139697 SBNO1 None Unknown HPA057603 ENSG00000133805 AMPD3 None Unknown HPA057637 ENSG00000148832 PAOX None Known HPA058296 ENSG00000143502 SUSD4 None Unknown HPA058466 ENSG00000101417 PXMP4 None Known HPA059737 ENSG00000134594 RAB33A None Unknown Continued on next page Proteomic Profiling of Vesicular Organelles — 14/22

Table 5 – continued from previous page Antibody Ensembl id Gene name Overlap Location HPA060972 ENSG00000100372 SLC25A17 Complete Known HPA061006 ENSG00000152954 NRSN1 None Unknown HPA061550 ENSG00000075643 MOCOS None Unknown HPA061693 ENSG00000163125 RPRD2 None Unknown HPA066512 ENSG00000169957 ZNF768 None Unknown HPA071502 ENSG00000177990 DPY19L2 None Unknown HPA072710 ENSG00000152049 KCNE4 None Unknown HPA073653 ENSG00000172239 PAIP1 None Unknown HPA073760 ENSG00000131373 HACL1 None Known

Table 6. Endosome Organelle Marker EEA1 Antibody Ensembl id Gene name Overlap Location HPA001467 ENSG00000079950 STX7 None Known HPA003524 ENSG00000103811 CTSH None Known HPA004167 ENSG00000108774 RAB5C Partial Known HPA004426 ENSG00000197746 PSAP None Known HPA006615 ENSG00000075785 RAB7A None Known HPA006964 ENSG00000075785 RAB7A None Known HPA007728 ENSG00000185359 HGS Fractional Known HPA014717 ENSG00000170088 TMEM192 Fractional Known HPA017672 ENSG00000213088 ACKR1 None Known HPA017910 ENSG00000092871 RFFL None Known HPA017967 ENSG00000231389 HLA-DPA1 None Known HPA018156 ENSG00000164733 CTSB None Known HPA019053 ENSG00000073921 PICALM None Known HPA019204 ENSG00000010270 STARD3NL None Known HPA019513 ENSG00000082805 ERC1 None Known HPA020998 ENSG00000188186 LAMTOR4 None Known HPA023920 ENSG00000136933 RABEPK None Known HPA024235 ENSG00000029725 RABEP1 None Known HPA024705 ENSG00000155975 VPS37A None Known HPA024781 ENSG00000155975 VPS37A None Known HPA024817 ENSG00000104497 SNX16 None Known HPA025960 ENSG00000156675 RAB11FIP1 None Known HPA026531 ENSG00000136643 RPS6KC1 None Known HPA028162 ENSG00000134262 AP4B1 None Known HPA028598 ENSG00000072274 TFRC Complete Known HPA028747 ENSG00000030582 GRN None Known HPA031470 ENSG00000160179 ABCG1 None Known HPA031565 ENSG00000136643 RPS6KC1 None Known HPA034808 ENSG00000114331 ACAP2 Fractional Known HPA035584 ENSG00000068650 ATP11A None Known HPA036033 ENSG00000196455 PIK3R4 None Known HPA036162 ENSG00000158411 MITD1 None Known HPA036163 ENSG00000158411 MITD1 None Known HPA037400 ENSG00000205302 SNX2 None Known HPA037612 ENSG00000214357 NEURL1B None Known HPA037726 ENSG00000107560 RAB11FIP2 None Known HPA039734 ENSG00000136100 VPS36 None Known HPA040727 ENSG00000206418 RAB12 None Known Continued on next page Proteomic Profiling of Vesicular Organelles — 15/22

Table 6 – continued from previous page Antibody Ensembl id Gene name Overlap Location HPA040802 ENSG00000069329 VPS35 Fractional Known HPA040978 ENSG00000176428 VPS37D None Known HPA041019 ENSG00000124222 STX16 None Known HPA041138 ENSG00000124067 SLC12A4 None Known HPA042231 ENSG00000141971 MVB12A None Known HPA042629 ENSG00000123154 WDR83 None Known HPA043348 ENSG00000167987 VPS37C None Known HPA047373 ENSG00000028528 SNX1 Fractional Known HPA049374 ENSG00000129515 SNX6 Fractional Known HPA050351 ENSG00000095066 HOOK2 None Known HPA050520 ENSG00000134070 IRAK2 None Known HPA050918 ENSG00000122705 CLTA Fractional Known HPA052900 ENSG00000197746 PSAP None Known HPA057498 ENSG00000122958 VPS26A Fractional Known HPA058342 ENSG00000106460 TMEM106B Fractional Known HPA059143 ENSG00000141367 CLTC Fractional Known HPA059590 ENSG00000198689 SLC9A6 None Known HPA061447 ENSG00000149823 VPS51 None Known HPA063748 ENSG00000198720 ANKRD13B None Known HPA065849 ENSG00000185722 ANKFY1 Complete Known HPA066538 ENSG00000105355 PLIN3 None Known HPA070346 ENSG00000164342 TLR3 None Known

Table 7. Endosome (Retromer) Organelle Marker VPS26A Antibody Ensembl id Gene name Overlap Location HPA001467 ENSG00000079950 STX7 None Known HPA003524 ENSG00000103811 CTSH None Known HPA004167 ENSG00000108774 RAB5C Fractional Known HPA004426 ENSG00000197746 PSAP Fractional Known HPA006615 ENSG00000075785 RAB7A Partial Known HPA006964 ENSG00000075785 RAB7A None Known HPA007728 ENSG00000185359 HGS Fractional Known HPA014717 ENSG00000170088 TMEM192 Partial Known HPA017672 ENSG00000213088 ACKR1 None Known HPA017910 ENSG00000092871 RFFL None Known HPA017967 ENSG00000231389 HLA-DPA1 None Known HPA018156 ENSG00000164733 CTSB None Known HPA019053 ENSG00000073921 PICALM None Known HPA019204 ENSG00000010270 STARD3NL None Known HPA019513 ENSG00000082805 ERC1 None Known HPA020998 ENSG00000188186 LAMTOR4 None Known HPA023920 ENSG00000136933 RABEPK None Known HPA024235 ENSG00000029725 RABEP1 None Known HPA024705 ENSG00000155975 VPS37A None Known HPA024781 ENSG00000155975 VPS37A None Known HPA024817 ENSG00000104497 SNX16 None Known HPA025960 ENSG00000156675 RAB11FIP1 None Known HPA026531 ENSG00000136643 RPS6KC1 None Known HPA028162 ENSG00000134262 AP4B1 None Known HPA028598 ENSG00000072274 TFRC None Known Continued on next page Proteomic Profiling of Vesicular Organelles — 16/22

Table 7 – continued from previous page Antibody Ensembl id Gene name Overlap Location HPA028747 ENSG00000030582 GRN Fractional Known HPA031470 ENSG00000160179 ABCG1 None Known HPA031565 ENSG00000136643 RPS6KC1 Fractional Known HPA034808 ENSG00000114331 ACAP2 Partial Known HPA035584 ENSG00000068650 ATP11A None Known HPA036033 ENSG00000196455 PIK3R4 None Known HPA036162 ENSG00000158411 MITD1 None Known HPA036163 ENSG00000158411 MITD1 None Known HPA037400 ENSG00000205302 SNX2 Partial Known HPA037612 ENSG00000214357 NEURL1B None Known HPA037726 ENSG00000107560 RAB11FIP2 None Known HPA039734 ENSG00000136100 VPS36 None Known HPA040727 ENSG00000206418 RAB12 None Known HPA040802 ENSG00000069329 VPS35 Complete Known HPA040978 ENSG00000176428 VPS37D None Known HPA041019 ENSG00000124222 STX16 None Known HPA041138 ENSG00000124067 SLC12A4 Partial Known HPA042231 ENSG00000141971 MVB12A None Known HPA042629 ENSG00000123154 WDR83 None Known HPA043348 ENSG00000167987 VPS37C None Known HPA047373 ENSG00000028528 SNX1 Partial Known HPA049374 ENSG00000129515 SNX6 Complete Known HPA050351 ENSG00000095066 HOOK2 None Known HPA050520 ENSG00000134070 IRAK2 None Known HPA050918 ENSG00000122705 CLTA None Known HPA052900 ENSG00000197746 PSAP Fractional Known HPA057498 ENSG00000122958 VPS26A Complete Known HPA058342 ENSG00000106460 TMEM106B Fractional Known HPA059143 ENSG00000141367 CLTC None Known HPA059590 ENSG00000198689 SLC9A6 None Known HPA061447 ENSG00000149823 VPS51 None Known HPA063748 ENSG00000198720 ANKRD13B None Known HPA065849 ENSG00000185722 ANKFY1 Fractional Known HPA066538 ENSG00000105355 PLIN3 None Known HPA070346 ENSG00000164342 TLR3 None Known

Table 8. Lysosome Organelle Marker LAMP-1 Antibody Ensembl id Gene name Overlap Location HPA000262 ENSG00000109436 TBC1D9 None Unknown HPA001242 ENSG00000243978 RGAG1 Partial Unknown HPA001467 ENSG00000079950 STX7 Partial Known HPA001493 ENSG00000125846 ZNF133 None Unknown HPA001569 ENSG00000145675 PIK3R1 Complete Unknown HPA002114 ENSG00000115233 PSMD14 None Unknown HPA003251 ENSG00000251369 ZNF550 None Unknown HPA003300 ENSG00000183475 ASB7 None Unknown HPA003524 ENSG00000103811 CTSH None Known HPA003642 ENSG00000187105 HEATR4 None Unknown HPA003720 ENSG00000181704 YIPF6 None Unknown HPA004167 ENSG00000108774 RAB5C None Known Continued on next page Proteomic Profiling of Vesicular Organelles — 17/22

Table 8 – continued from previous page Antibody Ensembl id Gene name Overlap Location HPA004426 ENSG00000197746 PSAP Complete Known HPA004909 ENSG00000115592 PRKAG3 None Unknown HPA005469 ENSG00000109511 ANXA10 None Unknown HPA005821 ENSG00000085982 USP40 None Unknown HPA006615 ENSG00000075785 RAB7A Complete Known HPA006964 ENSG00000075785 RAB7A Partial Known HPA007728 ENSG00000185359 HGS Partial Known HPA008763 ENSG00000030582 GRN Partial Known HPA012571 ENSG00000095970 TREM2 None Unknown HPA014717 ENSG00000170088 TMEM192 Complete Known HPA014946 ENSG00000166471 TMEM41B None Unknown HPA015049 ENSG00000160055 TMEM234 None Unknown HPA016495 ENSG00000106609 TMEM248 None Unknown HPA017655 ENSG00000197217 ENTPD4 None Known HPA017967 ENSG00000231389 HLA-DPA1 Complete Known HPA018156 ENSG00000164733 CTSB None Known HPA019204 ENSG00000010270 STARD3NL Complete Known HPA019556 ENSG00000143921 ABCG8 None Unknown HPA019827 ENSG00000165609 NUDT5 Partial Unknown HPA020099 ENSG00000197943 PLCG2 None Unknown HPA020998 ENSG00000188186 LAMTOR4 Complete Known HPA021506 ENSG00000204386 NEU1 None Known HPA021575 ENSG00000141569 TRIM65 None Unknown HPA022991 ENSG00000167536 DHRS13 None Unknown HPA023157 ENSG00000197566 ZNF624 None Unknown HPA023235 ENSG00000104324 CPQ None Known HPA023454 ENSG00000166329 CCDC182 None Unknown HPA023561 ENSG00000171206 TRIM8 None Unknown HPA024522 ENSG00000185722 ANKFY1 None Known HPA024785 ENSG00000147647 DPYS None Unknown HPA024817 ENSG00000104497 SNX16 None Known HPA025960 ENSG00000156675 RAB11FIP1 None Known HPA026419 ENSG00000119396 RAB14 None Known HPA028162 ENSG00000134262 AP4B1 None Known HPA028598 ENSG00000072274 TFRC Partial Known HPA028747 ENSG00000030582 GRN Partial Known HPA030226 ENSG00000140548 ZNF710 None Unknown HPA031068 ENSG00000064601 CTSA Partial Known HPA031323 ENSG00000197568 HHLA3 Partial Unknown HPA031470 ENSG00000160179 ABCG1 None Known HPA031565 ENSG00000136643 RPS6KC1 Partial Known HPA031630 ENSG00000143157 POGK None Unknown HPA031838 ENSG00000115956 PLEK None Unknown HPA034597 ENSG00000136710 CCDC115 Partial Known HPA034808 ENSG00000114331 ACAP2 None Known HPA035526 ENSG00000163820 FYCO1 None Known HPA035584 ENSG00000068650 ATP11A None Known HPA035870 ENSG00000055147 FAM114A2 None Unknown HPA036033 ENSG00000196455 PIK3R4 None Known HPA036733 ENSG00000124333 VAMP7 None Known Continued on next page Proteomic Profiling of Vesicular Organelles — 18/22

Table 8 – continued from previous page Antibody Ensembl id Gene name Overlap Location HPA036923 ENSG00000138246 DNAJC13 None Known HPA037000 ENSG00000033178 UBA6 None Unknown HPA037400 ENSG00000205302 SNX2 Partial Known HPA037612 ENSG00000214357 NEURL1B Complete Known HPA037726 ENSG00000107560 RAB11FIP2 None Known HPA038052 ENSG00000110013 SIAE None Known HPA038052 ENSG00000110013 SIAE None Unknown HPA038053 ENSG00000110013 SIAE None Known HPA038421 ENSG00000114742 WDR48 None Known HPA039584 ENSG00000131379 C3orf20 None Unknown HPA039734 ENSG00000136100 VPS36 Partial Known HPA040626 ENSG00000092470 WDR76 Partial Unknown HPA040727 ENSG00000206418 RAB12 Partial Known HPA040802 ENSG00000069329 VPS35 Complete Known HPA041138 ENSG00000124067 SLC12A4 None Known HPA041995 ENSG00000169682 SPNS1 Complete Known HPA042988 ENSG00000169682 SPNS1 None Known HPA044157 ENSG00000212123 PRR22 None Unknown HPA044175 ENSG00000258436 RNASE12 None Unknown HPA044732 ENSG00000183292 TISP43 None Unknown HPA045018 ENSG00000163807 KIAA1143 None Unknown HPA046373 ENSG00000224940 PRRT4 None Unknown HPA047373 ENSG00000028528 SNX1 Partial Known HPA047844 ENSG00000172661 FAM21C None Unknown HPA048668 ENSG00000274349 ZNF658 None Unknown HPA049374 ENSG00000129515 SNX6 Partial Known HPA049876 ENSG00000101160 CTSZ None Known HPA050036 ENSG00000176896 TCEANC None Unknown HPA050520 ENSG00000134070 IRAK2 None Known HPA050918 ENSG00000122705 CLTA Complete Known HPA052900 ENSG00000197746 PSAP None Known HPA053153 ENSG00000242852 ZNF709 None Unknown HPA053478 ENSG00000109323 MANBA None Known HPA054039 ENSG00000187498 COL4A1 None Unknown HPA055489 ENSG00000116954 RRAGC Complete Known HPA055664 ENSG00000131944 C19orf40 None Unknown HPA055805 ENSG00000188707 ZBED6CL None Unknown HPA055838 ENSG00000121691 CAT None Known HPA057168 ENSG00000139697 SBNO1 None Unknown HPA057498 ENSG00000122958 VPS26A Complete Known HPA057603 ENSG00000133805 AMPD3 None Unknown HPA057966 ENSG00000163820 FYCO1 None Known HPA058296 ENSG00000143502 SUSD4 None Unknown HPA058342 ENSG00000106460 TMEM106B Partial Known HPA059143 ENSG00000141367 CLTC Partial Known HPA059381 ENSG00000176978 DPP7 None Known HPA059737 ENSG00000134594 RAB33A None Unknown HPA061006 ENSG00000152954 NRSN1 None Unknown HPA061385 ENSG00000038532 CLEC16A None Known HPA061550 ENSG00000075643 MOCOS None Unknown Continued on next page Proteomic Profiling of Vesicular Organelles — 19/22

Table 8 – continued from previous page Antibody Ensembl id Gene name Overlap Location HPA061693 ENSG00000163125 RPRD2 None Unknown HPA061701 ENSG00000156171 DRAM2 None Known HPA063891 ENSG00000164073 MFSD8 None Known HPA064306 ENSG00000141564 RPTOR Complete Known HPA065849 ENSG00000185722 ANKFY1 None Known HPA066285 ENSG00000073060 SCARB1 Partial Known HPA066512 ENSG00000169957 ZNF768 None Unknown HPA067637 ENSG00000115194 SLC30A3 None Known HPA070346 ENSG00000164342 TLR3 None Known HPA070642 ENSG00000162736 NCSTN None Known HPA070738 ENSG00000103111 MON1B Partial Known HPA071502 ENSG00000177990 DPY19L2 None Unknown HPA072710 ENSG00000152049 KCNE4 None Unknown HPA073653 ENSG00000172239 PAIP1 None Unknown

Appendix 2. Extra images of SCP2 only overlapping with the peroxisome marker in some cells. Images in the order: Merge, HPA-antibody targeting SCP2, nucleus and peroxisomal organelle marker (ABCD3).

Merge Proteomic Profiling of Vesicular Organelles — 20/22

HPA-antibody Proteomic Profiling of Vesicular Organelles — 21/22

DAPI Proteomic Profiling of Vesicular Organelles — 22/22

Peroxisomal organelle marker (ABCD3)