Pathogen Receptor Discovery with a Microfluidic Human Membrane Protein Array

Pathogen receptor discovery with a microfluidic human membrane protein array

Yair Glicka,1,Ya’ara Ben-Aria,1, Nir Draymanb, Michal Pellacha, Gregory Neveuc,d, Jim Boonyaratanakornkitc,d, Dorit Avrahamia, Shirit Einavc,d, Ariella Oppenheimb, and Doron Gerbera,2 aMina and Everard Goodman Faculty of Life Sciences, Bar Ilan University, 5290002, Israel; bFaculty of Medicine, Hebrew University, Jerusalem, 9112001, Israel; cDepartment of Medicine, Division of Infectious Diseases and Geographic Medicine, Stanford University School of Medicine, Stanford, CA 94305; and dDepartment of Microbiology and Immunology, Stanford University School of Medicine, Stanford, CA 94305

Edited by Stephen R. Quake, Stanford University, Stanford, CA, and approved March 1, 2016 (received for review September 20, 2015)

The discovery of how a pathogen invades a cell requires one to and translate into proteins in situ (10, 11). This approach has determine which host cell receptors are exploited. This determina- enabled the study of the Pseudomonas aeruginosa outer mem- tion is a challenging problem because the receptor is invariably a brane protein for immunity (12). membrane protein, which represents an Achilles heel in proteomics. Combining integrated microfluidics with microarrays and We have developed a universal platform for high-throughput ex- in vitro transcription and translation (TNT) systems may over- – pression and interaction studies of membrane proteins by creating a come all of the above mentioned difficulties (Fig. 1A) (13 16). microfluidic-based comprehensive human membrane protein array The integrated microfluidic device allows smart liquid manage- (MPA). The MPA is, to our knowledge, the first of its kind and offers ment in very low volumes, partitioning, and process integration a powerful alternative to conventional proteomics by enabling the (i.e., protein expression, immobilization, and interaction exper- simultaneous study of 2,100 membrane proteins. We characterized iments). Microarray technology provides the means for pro- direct interactions of a whole nonenveloped virus (simian virus 40), gramming thousands of different experiments (17). In vitro TNT as well as those of the hepatitis delta enveloped virus large form expression systems allow protein biosynthesis and are compatible with high throughput (18). Such systems are commercially avail- antigen, with candidate host receptors expressed on the MPA. Se- able and benefit from fast protein expression, low reaction vol- lected newly discovered membrane protein–pathogen interactions umes, and short reaction times and enable expression of synthetic were validated by conventional methods, demonstrating that the proteins with inserted epitope tags. Adding microsomal mem- MPA is an important tool for cellular receptor discovery and for branes enable the correct folding of membrane proteins and understanding pathogen tropism. support posttranslational modifications, such as glycosylation (9). In short, the microfluidic platform facilitates using in vitro pathogen–host interactions | membrane protein array | TNT systems to produce a reliable membrane protein array receptor discovery | integrated microfluidics (MPA) from DNA with high sensitivity, low material and protein consumption, and compatibility with membrane proteins. he human genome contains ∼21,000 distinct protein-coding In this study, we used a microfluidic platform to combine Tgenes (1), out of which ∼5,360 code for membrane proteins microarray technology, cell-free protein expression, and integrated (2). Membrane proteins are critical for many cellular processes, microfluidics, allowing high-throughput screening of pathogens such as signaling, transport, cell–cell communication, and also with a human membrane proteome library. As a proof-of-concept, interaction with pathogens leading to various cellular responses. we screened two pathogens differing in structure and physiology. It is not surprising that 60% of drugs currently in the market The first was simian virus 40 (SV40), a nonenveloped polyomavirus target proteins at the cell surface (3). Mapping molecular inter- containing circular double-stranded DNA, which can induce actions of membrane proteins is, therefore, of utmost importance. membrane invaginations, similarly to other polyomaviruses. It Pathogen–host recognition involves surface interactions regulated causes infections of the kidney and possibly also other tissues by membrane proteins. Many interactions between membrane proteins and pathogens are unknown, partly because of the low Significance sensitivity and limited compatibility of current methodologies with membrane proteins (4). These limitations pose a major obstacle in In this work, we report, to our knowledge, the first in vitro tool understanding pathogen tropism, a public health concern in view for host–pathogen screening that encompasses thousands of of emerging diseases, e.g., severe acute respiratory syndrome and functional insoluble proteins—primarily transmembrane pro- Ebola. There is therefore a need for new approaches that would teins—immobilized within a microfluidic device. We discovered recapitulate the pathogen–host molecular recognition, let alone in previously unknown protein–pathogen interactions, and then the context of intact pathogens. selected interactions were further validated by conventional Mapping protein–protein interaction (PPI) is a major chal- methods. Considering the tremendous difficulty in discovering lenge in proteomics. Many molecular interactions are transient pathogen receptors, this in vitro high-throughput approach is and weak, leading to low yield of bound material and thus de- extremely important and efficient for receptor discovery and manding highly sensitive detection methods. Current methods understanding pathogen tropism, with relevance to emerging for characterizing PPI networks suffer from several basic disad- human diseases. vantages: low sensitivity, leading to high false negative rate; low specificity, leading to high false positive rate (5–7); low coverage Author contributions: Y.G., Y.B.-A., N.D., D.A., S.E., A.O., and D.G. designed research; Y.G., of known interactions; and high variability, even in screens from Y.B.-A., N.D., G.N., J.B., and S.E. performed research; A.O. contributed new reagents/ the same species (4). Protein arrays could potentially override analytic tools; Y.G., Y.B.-A., N.D., M.P., G.N., J.B., D.A., S.E., A.O., and D.G. analyzed data; and Y.G., Y.B.-A., M.P., A.O., and D.G. wrote the paper. such limitations (8), but suffer from a purification bottleneck and limited functionality after deposition (9). These difficulties are The authors declare no conflict of interest. even more pronounced with membrane proteins. Membrane pro- This article is a PNAS Direct Submission. teins are usually in low abundance; in addition, they are in- 1Y.G. and Y.B.-A. contributed equally to this work. compatible with high-throughput methods (e.g., yeast two-hybrid) 2To whom correspondence should be addressed. E-mail: [email protected]. and are difficult to purify in functional form (e.g., protein ar- This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. rays). One partial solution to these obstacles was to print DNA 1073/pnas.1518698113/-/DCSupplemental.

4344–4349 | PNAS | April 19, 2016 | vol. 113 | no. 16 www.pnas.org/cgi/doi/10.1073/pnas.1518698113 Fig. 1. Membrane protein array generated by integrated microfluidic platform. (A) An integrated microfluidics platform (Left) was used for on-chip expression of membrane proteins, to serve as “baits” for protein interactions or modifications (29). The device consists of two polydimethylsiloxane (PDMS) layers, a flow layer with 64 × 64 unit cells array (gray), and a control layer with micromechanical valves (colored) that manipulate the flow of fluids in the experiment (Center). The sandwich valves (pink) separate neighboring unit cells; the neck valves (green) divide each unit cell into a DNA compartment and a reaction compartment. The button valves (blue) enable surface patterning to promote binding of proteins to an antibody surface. The button valves serve as mechanical traps of molecular interactions (MITOMI) and allow measurement at equilibrium concentration. MITOMI increases the sensitivity of the system, facilitating detection of weak and transient interactions (SI Appendix, Fig. S6). Combining the microfluidic platform with microarray technology enables BIOPHYSICS AND programming of the device with up to several thousand spotted genes (Right). Using assembly PCR (SI Appendix, Fig. S1), we added c-Myc (N-terminal) and COMPUTATIONAL BIOLOGY His6 (C-terminal) tags to the ORFs, creating synthetic genes. On-chip in vitro protein expression, following the synthetic gene programming, combined with the corresponding antibody surface patterning, facilitates the self-assembly of a MPA using cell-free TNT (rabbit reticulocyte). The immobilized bait proteins are labeled with fluorescent antibodies and quantified by using a microarray scanner. Expressed proteins form a green circle below the button valve (Right). (B) We expressed 2,686 different human membrane proteins on chip. The z axis represents the average expression level (n = 4); x and y axes represent the position on-chip in the 2D protein array. Membrane proteins were immobilized on the surface by their C terminus (B) or N terminus (C). Unspotted (red) chambers were used as negative controls. Levels of expression that were 2 SDs above the negative controls were considered positive. Color gradients are for easier visualization only. (D) The SE of protein expression was calculated for each protein, and for most proteins it was relatively uniform, with an error of 10% or below. The data represents both intrachip and interchip variability. such as the mesothelium, and it has also been associated with On average, 10% of the genes were misaligned and thus could cancer (19, 20). The second pathogen was the hepatitis delta virus not be expressed. After surface functionalization of the chip, the (HDV). We screened the large-form delta antigen (L-HDAg), a genes were expressed on-chip by using rabbit reticulocyte TNT prenylated protein essential for HDV virion assembly by allowing and microsomal membranes. We immobilized the expressed interaction with the envelope proteins of the hepatitis B helper membrane proteins to the slide surface via either their N or C virus (21, 22). Understanding interactions of both these viruses terminus, using an antibody. Only full-length proteins were labeled with host membrane proteins may shed light on important patho- with fluorescent antibodies on their corresponding C (c-Myc tag) genic processes. or N (6-histidine tag) termini (Fig. 1). Protein expression levels We created a library of ∼2,700 synthetic genes, which we ranged from no expression (0 a.u.) to high expression (>1,000 arrayed and expressed within the microfluidic platform using a a.u.). Chambers without spotted DNA were used as negative con- cell-free protein expression system. Then, we screened the as- trols (Materials and Methods). No nonspecific protein labeling was sembled MPA for interactions with each pathogen and identified observed by incubating the MPA with a nonspecific fluorescent new interactions. A bioinformatics approach was used to identify antibody (SI Appendix,Fig.S4). More than 2,100 proteins (80%) biological processes associated with the newly found interactions. were properly expressed on-chip, regardless of the direction by Specific interactions of interest were chosen for further valida- which they were immobilized to the surface (Fig. 1B, SI Appendix, tion by either coimmunoprecipitation (Co-IP) or protein–frag- Table S1,andDataset S1). Similar results were obtained when we ment complementation assay of luciferase activity (PCA). We expressed soluble protein arrays on a chip (SI Appendix,Fig.S5). In successfully demonstrated the effectiveness of using a micro- contrast, the dynamic range of expression was much lower for fluidic platform for performing a high-throughput screen of membrane proteins than for soluble proteins. This finding is con- pathogen–membrane protein interactions. sistent with the literature and is most likely due to the addition of microsomal membranes, which become the limiting factor in the Results and Discussion reaction (9, 26, 27). Protein Expression/MPA Protein. To create a human MPA, we Several possible factors can affect protein expression, including constructed ∼2,700 linear synthetic genes encoding for human DNA concentration, protein length, and membrane topology. We membrane proteins, using an assembly PCR approach (Materials examined possible bias in protein expression due to spotted DNA and Methods; SI Appendix, Fig. S1) (14, 23–25). The trans- concentration (0–0.2 μM) or protein length (60–1,400 aa). We membrane domain frequency of these synthetic genes matched found no correlation between protein expression and DNA con- the predicted frequency in the human membrane proteome (SI centration or protein length (SI Appendix, Figs. S7 and S8). Appendix,Fig.S2) (2). A microfluidic device was then programmed We were concerned that transmembrane domain topology with this ORF array (SI Appendix,Fig.S3). The programming may affect the efficiency of protein expression and that bias to- (alignment of microarray to the device) was performed manually. ward certain topologies may exist. To confirm whether such a

Glick et al. PNAS | April 19, 2016 | vol. 113 | no. 16 | 4345 bias existed, we divided the proteins according to the number of Indeed, the identified membrane interactome achieved for their transmembrane domain. These ranged from 0 to 14 trans- L-HDAg was distinct from SV40, with only seven interacting membrane domains. The number of transmembrane domains did proteins common to both pathogens, demonstrating high speci- not affect average expression levels, which hence were not affected ficity of the microfluidic platform. by protein topology (SI Appendix,Fig.S9). We also demonstrated Among our discovered interactions, for example, we detected that both sides of the membrane proteins were accessible for in- binding of SV40 to SCRB2, a close homolog of SCRB1, which teraction (SI Appendix,Fig.S10). We attributed this accessibility had been previously reported to bind SV40 (28). Another pre- to microsomal membrane instability (i.e., vesicle fusion) due to viously unidentified interaction was with TNR12, a protein that changes in surface tension following protein expression (8, 22, 23). is highly expressed in heart, placenta, and kidney. It is a close To demonstrate that the membrane proteins expressed using homolog of both TNR3 and TNR5, both previously reported to the cell-free expression procedure were functional, selected bind SV40 (28). These interactions point to a previously sug- membrane proteins with previously established ligand interac- gested common SV40 binding domain to TNR proteins (28). tions were expressed via the same expression procedure and SV40 traffics via the endosomal pathway into the ER (29) and is immobilized onto a chip (SI Appendix, Fig. S11). Interactions of known to bind to ganglioside GM1, present in caveolar/lipid raft 13 human receptors and their corresponding fluorescently la- domains at the plasma membrane (30, 31). Notably, 23% beled ligand were examined, and 10 pairs (∼75%) demonstrated (19 of 85) of our discovered interactions are located in lipid rafts, specific interactions, indicating a likelihood that a significant caveola, and endoplasmic reticulum (ER) membrane (SI Ap- proportion of the 2,700 membrane proteins were properly folded pendix, Table S2 and Dataset S2). This percentage is much and functional. higher than their representation in MPA (12%). The mechanism by which SV40 crosses the ER membrane is unknown (32), with On-Chip Virus Screening. As proof of concept for host–pathogen possible involvement of these interactions. SV40 interaction with recognition, we screened the MPA for potential viral receptors CD20, CD116, CD140b, and BST, an IFN-induced antiviral host using a fluorescently labeled SV40 as well as L-HDAg (Fig. 2). restriction factor, may play a role in cell entry of SV40 as well as For SV40 we used a two-step strategy—a wide screen followed in its immune evasion. by a randomized focused screen. The second screen included the For interacting proteins of L-HDAg, we used the Database for positive hits from the wide screen, a control set of known posi- Annotation, Visualization and Integrated Discovery functional tives (from the literature) and a set of negative controls for which annotation tool (https://david.ncifcrf.gov) to establish the most no interaction was reported or observed in the first screen. A relevant biological processes [Gene Ontology (GO) terms] as- total of 99 interactions were observed (SI Appendix, Table S2 and sociated with the proteins in our interaction list. A total of 63 of Dataset S2), including 14 of the 22 positive controls (Fig. 2B). the identified interactions are associated with 13 significantly We determined the sensitivity and specificity of this system to enriched biological processes, including glycoprotein metabolic SV40-membrane protein interaction to be 64% and 71%, re- process, immune response, carbohydrate biosynthetic process, or spectively (SI Appendix, Table S3). ER to Golgi vesicle-mediated transport (SI Appendix, Fig. S12). As with SV40, the protein array library for screening L-HDAg For both pathogens, their interacting membrane proteins were was expressed on-chip by using cell-free TNT. In parallel, analyzed by using Search Tool for the Retrieval of Interacting L-HDAg, the prey viral protein, was expressed in-tube by using Genes/Proteins (STRING; string-db.org), based on known and the same cell-free procedure and microsomal membranes, cre- predicted PPIs, to examine physical or functional associations ating synthetic viroid-like particles. After assembly of the MPA, among the proteins in the interaction list. STRING provided a L-HDAg was loaded onto the device and subsequent incubation graphic representation of a network of interactions between period of isolated unit cells allowed possible host-L-HDAg in- some identified pathogen-interacting proteins (Fig. 3). Interestingly, teractions to reach equilibrium. The surface-bound protein array although the biological function of most of the observed interac- and the trapped L-HDAg were then incubated with fluorescently tions remains to be explored, both SV40 and L-HDAg were shown labeled antibodies for the different tags on the respective pro- to interact independently with proteins that are components of the teins. Fluorescent signals were measured by using a microarray same functional complex. We observed, for example, SV40 binding laser scanner, and the extracted data was analyzed for specific to TMED1, TMED5, TMED10, and KDELR3 proteins. These interactions. A total of 150 interactions were observed and fur- have a potential role in vesicular protein trafficking, implying in- ther analyzed (Fig. 2C, SI Appendix, Table S4, and Dataset S3). volvement in virion trafficking. L-HDAg interacts with USE1 and

Fig. 2. Pathogen receptor discovery with MPA, an original tool for de novo characterizing viral tropism. (A, Left) Schematic overview of the experimental procedure. Once bait proteins were immobilized on the surface, the virus was introduced into MPA and incubated with the MPA for 30 min. (A, Right)Im- ages of two representative unit cells depicting bait protein (green) and pathogen interaction (red) with positive binding (Upper) and negative control with no binding (Lower). (B and C) Cy5-labeled SV40 virions (B) and L-HDAg (C) were applied to the MPA to test for pathogen–membrane protein interactions. Positives in the large screen were collected and reprinted in a different arrangement to avoid tech- nical biases. Approximately 100 novel SV40–protein interactions and ∼150 L-HDAg–protein interactions were discovered via the MPA screen. Signals below the cutoff (SI Appendix, Fig. S13 and Materials and Methods) and above the background were filtered out. Interactions are presented as the average ratio of pathogen binding to membrane protein expression (n = 4): new interactions are in blue; known SV40 partners detected by MPA are in red.

4346 | www.pnas.org/cgi/doi/10.1073/pnas.1518698113 Glick et al. to SV40 was negligible. The cutoff for the positive signal was three SDs above the average of all five negative controls. The bead analysis was performed by fluorescence microscopy because of the limited amount of membrane protein sample available, which further emphasized the advantage of using MPA. Validation of the L-HDAg proteomic results was performed for selected proteins by PCA, because very high nonspecific binding was observed with co-IP. Because of the role of L-HDAg in viral assembly and secretion, we selected identified host proteins that are involved in vesicular trafficking for further verifi- cation. We also selected several representative proteins from other biological processes, several of which are represented in the STRING analysis (Fig. 3). The luciferase activity of the 21 L-HDAg–protein interactions we found was measured, and results were expressed as normalized luminescence ratios. Seven interactions gained a relative luciferase activity signal that was at least three times greater than their respective negative controls (Fig. 4B). The lower success rate of L-HDAg validation compared with SV40 could be attributed to the differences in validation techniques and/or that its lipid-bilayer envelope reduces the MPA specificity. STX18, PREB, SCAMP4, and CAV-1 are proteins involved in vesicular trafficking. SEC63 is involved in translocation of proteins across the ER membrane, whereas DNAJB11 is involved in the folding of proteins entering the ER. ZDHHC11 is a member of the palmitoyltransferase family. In conclusion, we have presented the application of our human functional MPA toward discovery of pathogen–host interactions. MPA impact in pathogen tropism research is especially

important for investigating emerging diseases. Our results dem- BIOPHYSICS AND

onstrated that the microfluidic affinity assay is a powerful tool COMPUTATIONAL BIOLOGY for identifying PPIs in general, and, more specifically, pathogen– transmembrane protein interactions. We estimate the functionality of our MPA to be >65%, based on the number of validated interactions (35 of 54). The proteomics establishment of inter- twined relations between many of the identified interacting proteins provide new perspectives on viral behaviors during infection and pave the way for further research in various direc- tions. Furthermore, this platform may also serve for testing for drug design, antibody specificities, or orphan receptors/ligands as well as in a wide spectrum of other research areas that involve membrane proteins. Materials and Methods Mold Fabrication. The microfluidic devices were designed in a similar manner as described (14, 23, 33). Details are described in SI Appendix, SI Materials and Methods).

Device Fabrication. The microfluidic devices were fabricated on silicone molds Fig. 3. Known and predicted PPI between SV40 (A) and L-HDAg (B) iden- casting silicone elastomer polydimethylsiloxane (SYLGARD 184, Dow Corn- tified pathogen-interacting proteins using STRING. Disconnected proteins ing). Details are described in SI Appendix, SI Materials and Methods. (nodes) were excluded. Production of Human Synthetic Genes Library via Assembly PCR. Synthetic linear human genes were generated by using two-step assembly PCR (SI syntaxin 18 (STX18), both components of a complex involved in Appendix, Fig. S1). As a template for the first PCR, we used a library of the Golgi to ER retrograde transport that directly interact with each Human ORFome in gateway-donor GW223 plasmids (Open Biosystems). A other. L-HDAg also interacts with PDGFRA and PDGFRB, high-fidelity hot start DNA polymerase (Phusion II; FINNZYMES) was used for which upon ligand binding at the plasma membrane, form a all PCR procedures. In the first PCR step, two epitope tags were added to heterodimer and initiate several signaling cascades, depending each protein, c-Myc at the N terminus and 6-His at the C terminus. The tags on the nature of the bound ligand. were added by using the primers 5′GW223–c-Myc and 3′-GW223–His. A reaction mix with a total volume of 20 μL was prepared by using 0.8 units of Validation of Proteomic Results. SV40 protein interactions were DNA polymerase for each reaction. The PCR assay was performed in a PCR validated by co-IP, and we independently validated 25 interac- 96-well plate in 25 cycles with annealing temperature of 64 °C. The exten- tions including negative controls. These interactions included sion time ranged from 45 to 300 s in 72 °C depending on the ORF length. The representative proteins from different biological processes, some first PCR product served as a template for the second PCR. In addition, two different pairs of primers were used for the second PCR step by adding the of which are observed in the STRING diagram (Fig. 3). Of 20 5′ UTR (T7 promoter) and 3′ UTR (T7 terminator) for each gene. The reaction interactions detected by MPA, 18 were verified by co-IP (Fig. 4A mixture with a total volume of 50 μL was prepared containing 1.5 units of and SI Appendix, Fig. S14), demonstrating very high correlation DNA polymerase. The first extension primer pair, containing 85 and 95 bp, (90%) between the MPA-detected interactions and co-IP ex- was added to the mixture in a low concentration (2.5 nM). After 10 cycles, periments. The failure of two proteins in co-IP with SV40 may be the second primer pair (5′ final and 3′ final) was added to the PCR mixture attributed to the higher sensitivity of our MPA. Interactions were (0.2 μM) for an additional 25 cycles, completing the PCR process. The PCR measured relative to five negative controls, four of which binding products were filtered in multiwell 10k filter plates (AcroPrep; PALL) and

Glick et al. PNAS | April 19, 2016 | vol. 113 | no. 16 | 4347 within the microfluidic device was chemically modified. This surface chemical modification also facilitates the self assembly of a protein array on the surface. Biotinylated BSA (1 μg/μL; Thermo) was flowed through the device for 30 min, binding the BSA to the epoxy surface. On top of the biotinylated BSA, 0.5 μg/μL Neutravidin (Pierce) was added for 30 min. The “button” valve was then closed, and biotinylated PEG (1 μg/μL; Nanocs) was flowed through the chip for 30 min, passivating the rest of the device. After passivation, the button valve was released and a flow of 0.2 μg/μL penta-His (Qiagen) or c-Myc (Cell Signaling). Biotinylated antibodies was applied. The antibody bound to the exposed Neutravidin, specifically to the area under the button, creating an anti-His tag or c-Myc tag array. Hepes (50 mM; Biological In- dustries) was used for washing unreacted substrate between each of the different surface chemistry steps.

Protein Expression. Proteins were expressed on the device by using rabbit reticulocyte quick-coupled TNT reaction (Promega). Microsomal membranes (Promega) were added to the extract to express membrane proteins (including L-HDAg). The expression of the proteins from the spotted synthetic genes on the device created an array of proteins ready for use in a binding screen. By opening the “neck” valves, 12.5 μL of the expression mix was flowed through the device into the gene chamber. Next, the “sandwich” valves were closed, leaving each unit cell separated from its environment, and the device was incubated on a hot plate for 2.5 h at 32 °C. Expressed proteins were then diffused through the gene chamber to the reaction chamber, binding their C terminus His tag to the anti-His antibody or their N terminus c-Myc tag to the c-Myc antibody under the button valve, immobilizing the protein. Proteins were labeled with a c-Myc (Sigma- Aldrich) or penta-His (Qiagen) Cy3 antibody, which bound to its corresponding epitope and labeled it. Unspotted chambers were used as negative controls. Because no DNA was spotted in these chambers, we expect that no proteins can be expressed. Thus, any signal from these chambers is nonspecific background. Nonspecific labeling of the fluorescent antibody was – Fig. 4. Validated pathogen–protein interactions. (A) Co-IP of SV40 and 25 determined by using an anti V5-FITC antibody (Cell Signaling). Protein ex- different membrane proteins using magnetic beads. All membrane proteins pression levels were determined with a microarray scanner (LS Reloaded; were tagged similarly to the microfluidics assay and expressed by using cell- Tecan) using a 532-nm laser and 575-nm filter. free TNT. Proteins were immobilized to T1 magnetic beads coated with α-His antibody and scanned by fluorescent microscopy for either protein expres- Receptor–Ligand Interactions. Human receptors with N terminus c-Myc and C sion, following α-c-Myc Cy3 antibody labeling, or for Cy5-labeled SV40, after terminus T7 tag were expressed by using rabbit reticulocyte quick-coupled co-IP. Image analysis revealed that SV40 coimmunoprecipitated with 18 of 20 TNT reaction in the presence of microsomal membrane (Promega). The ex- proteins (blue). These results demonstrate high correlation with the MPA pression was performed in a final volume of 12.5 μL, including 1 μg of DNA. screen (90%). The interactions of 17 proteins were highly significant, and a The tube was incubated at 32 °C for 2.5 h with agitation (600 rpm). In mild interaction was observed for one protein. *P < 0.01; **P < 0.00001. parallel, we prepared the surface chemistry and α-T7 biotinylated antibodies SV40 failed to co-IP with two proteins (light blue). In addition, four negative (Qiagen) bound under the button. Next, each expressed receptor was at- control proteins of five showed negligible binding to SV40 virions. The fifth tached to a specific part of the device and bound to its corresponding an- demonstrated some low nonspecific binding. (B) Validation of L-HDAg in- tibody, creating a receptors array. Finally, His-tagged recombinant ligands teractions by PCA. Combinations of plasmids encoding a pair of proteins, a (CXCL10, CTLA4, ZP3, EGF, CD40L, TNF-α,TNF-β, and CD48; purchased from human membrane protein and L-HDAg each fused to a fragment of the Prospec) and FASL (Peprotech) were applied to the device. By closing the G. princeps luciferase protein (Gluc1 and Gluc2, respectively) were cotrans- sandwich valves, each unit cell separated from its environment. Next, the fected into 293T cells. Combinations with empty Gluc1 or Gluc2 plasmids button valves opened, exposing the receptor array. Ligands were allowed to were used as negative controls. Results were expressed as a luminescence incubate with the receptor array for 30 min at 32 °C. Proteins were then ratio, representing the average luminescence signal detected in cells trans- labeled with α-c-Myc Cy3 (Sigma) and α-HIS Alexa Fluor 647 (Qiagen) anti- fected with both vectors divided by that measured in control wells. bodies. Protein interactions were determined with a microarray scanner (LS Reloaded; Tecan) using a 633-nm laser and 695-nm filter for Cy5 and a 535-nm laser and 595-nm filter for Cy3. eluted with 40 μL of double-distilled water. The yield of gene product was verified twice, at the end of the first PCR step and after 1.5% agarose gel Production, Purification, and Labeling of SV40 Virions. SV40 virions were electrophoresis. In addition, PCR products were transferred to 384 UV- produced and purified as described (25). Next, SV40 virions were labeled by ’ readable plates, and concentrations were measured by using a UV plate using commercial kits, according to the manufacturer s instructions (Molec- reader (Synergy 4 Hybrid Microplate Reader; BioTek). ular Probes). All experiments were performed in accordance with regula- tions approved by the Bar Ilan University Pathogen Oversight Committee. DNA Arraying and Device Alignment. A series of synthetic genes were spotted – onto epoxy-coated glass substrates (CEL Associates) with a MicroGrid 610 (Bio MPA Pathogen Host Large Screen. For MPA screening, 2,686 membrane proteins expressed on-chip as above. Cy5-labeled SV40 virions (15 nM) were Robotics) microarrayer by using SMT-S75 silicone pins (Parallel Synthesis). applied to the device and incubated with the protein array for 30 min at 32 °C. Column and row pitch corresponded to the specific device. The device we After incubation, buttons were closed and unbound viruses were washed used contains 65 columns and 64 rows with a pitch of 281.25 by 562.5 μm, out. Interaction levels were determined with a microarray scanner (LS respectively. Each sample solution contained 0.125% of polyethylene glycol Reloaded; Tecan) using a 633-nm laser and 695-nm filter for Cy5 and nor- (PEG; Sigma Aldrich) and D-trehalosedihydrate (Sigma-Aldrich) at a concen- malized to protein expression level. tration of 12.5 mg/mL in dH2O to prevent irreversible binding of the DNA to the glass, as well as for visualization during alignment. Finally, the arrays MPA Pathogen–Host Second Screen. After the MPA large screen, a second were aligned to the microfluidic device by hand under a stereoscope and microfluidic device programmed with membrane proteins binders from the bonded for 4 h on a heated plate at 80 °C. first screen, including 22 positive controls and 50 negative controls. Proteins expressed on-chip as described. Cy5-labeled SV40 virions (15 nM) were applied Surface Functionalization. To prevent nonspecific adsorption and to achieve to the device and incubated with the protein array for 30 min at 32 °C. After suitable binding orientation of expressed proteins, all accessible surface area incubation, buttons were closed and unbound viruses were washed out.

4348 | www.pnas.org/cgi/doi/10.1073/pnas.1518698113 Glick et al. Interaction levels were determined with a microarray scanner (LS Reloaded; Image and Data Analysis. For MPA expression experiment, images were an- Tecan) using a 633-nm laser and 695-nm filter for Cy5 and normalized to alyzed with GenePix7.0 (Molecular Devices). The image (cy3 channel) was protein expression level. A cutoff of 2 and 4 SDs were used for the SV40 (SI used to determine bait (on ChIP-expressed protein) expression level. Rows Appendix, Fig. S13) and the L-HDAg screens, respectively. and columns without DNA array were used to assess nonspecific binding of “ ” Human synthetic genes were created with c-Myc (N-terminal) and six labeled antibodies to the surface. Each row and column was then normalized histidine (C-terminal) tags and expressed in vitro as described above. The ex- by subtracting the nonspecific baseline signal. A signal that was 2 SD above pression was performed in a final volume of 12.5 μL, including 1 μg of DNA and the average noise was considered successful protein expression. incubated at 32 °C for 2.5 h with agitation (600 rpm). Expressed proteins were For interaction two images (Cy3 and Cy5/GFP channels) were analyzed with then incubated either with anti–c-Myc-Cy3 antibody (Sigma-Aldrich) or with GenePix (Molecular Devices). Rows without DNA array were used to assess Cy5-labeled SV40 virions (15 nM). Next, the proteins were immobilized to T1 magnetic beads (Invitrogen) coated by α-His–biotinylated antibody (Qiagen). nonspecific binding to the surface. Columns with no prey were used to assess The beads were washed with PBS and scanned by Nikon Eclipse fluorescent nonspecific binding of labeled antibodies to the surface proteins. Each row microscope for either protein expression or for SV40 co-IP. Images were ana- and column was then normalized by subtracting the nonspecific baseline lyzed by Nikon’s NIS elements software. Single beads’ median intensities were signal. The “interaction ratio” (Cy5/Cy3 or GFP/Cy3) was calculated, and the measured for either protein expression or SV40 co-IP. Statistical significance highest ratio was normalized to 1. was determined by calculating P values for bead intensities of each protein For each MPA-expressed protein, the L-HDAg interaction signal was calcu- compared with bead intensities of all negative controls. lated as the average of median signals obtained from the respective quadruplet. The average of median signals obtained from unit cells in the array L-HDAg Interaction Validation by PCA. Two engineered Gateway vectors, with no spotted DNA was used to calculate a control background signal. A pGluc1-Nter-Gateway and pGluc2-Nter-Gateway, encode for two fragments of histogram of the L-HDAg interaction signal was plotted and allowed de- the Gaussia princeps luciferase protein. A total of 21 ORFs encoding the human termination of a cutoff at 4 SD above the array control background signal as proteins were picked from the Human ORFeome library (34) (Open Biosystems) the threshold above which the interactions were considered positive (SI Ap- and inserted as described (35, 36) into the pGluc1-Nter-Gateway plasmid, and pendix,Fig.S2). Each quadruplet above that threshold was examined, and L-HDAg was inserted into pGluc2-Nter-Gateway (22, 37). A human membrane interactions with false signal that resulted from visible aggregates on the glass protein and L-HDAg fused to each plasmid, respectively, or control empty slide or inside the microfluidic device or high SD within the quadruplet- vectors were cotransfected into 293T or Huh-7.5 cells plated in 96-well plates in triplicates. At 24 h after transfection, cells were lysed and subjected to standard resulting inconsistent L-HDAg interaction signal were discarded. luciferase reporter gene assays by using the Renilla luciferase assays system (Promega). Results were expressed as luminescence ratio, which represents the ACKNOWLEDGMENTS. This work was supported by European Research average luminescence signal detected in cells transfected with both Gluc1 and Council Starter Grant 309600 (to D.G.); Israel Science Foundation Grant 715/11 (to D.G.) and Israel Science Foundation Grant 291/12 (to A.O.); Gluc2, divided by the average of luminescence measured in negative control BIOPHYSICS AND American Cancer Society Grant RSG-14-11 0-0 1-MPC (to S.E.); Doris Duke wells transfected with Gluc1 and an empty Gluc2 vector and those transfected Charitable Foundation Grant 2013100 (to S.E.). G.N. was supported by COMPUTATIONAL BIOLOGY with Gluc2 and an empty Gluc1 vector. A combination with TSG101, which has Child Health Research Institute, Lucile Packard Foundation for Children’s been identified as a L-HDAg partner in a small-scale microfluidic screen (33), Health, and Stanford Clinical and Translational Science Award Grant was used as a positive control for the assay. UL1 TR000093.

1. Lander ES (2011) Initial impact of the sequencing of the human genome. Nature 20. Ewers H, et al. (2010) GM1 structure determines SV40-induced membrane in- 470(7333):187–197. vagination and infection. Nat Cell Biol 12(1):11–18, 1–12. 2. Almén MS, Nordström KJV, Fredriksson R, Schiöth HB (2009) Mapping the human 21. Chang FL, Chen PJ, Tu SJ, Wang CJ, Chen DS (1991) The large form of hepatitis delta membrane proteome: A majority of the human membrane proteins can be classified antigen is crucial for assembly of hepatitis delta virus. Proc Natl Acad Sci USA 88(19): according to function and evolutionary origin. BMC Biol 7(1):50. 8490–8494. 3. Overington JP, Al-Lazikani B, Hopkins AL (2006) How many drug targets are there? 22. Glenn JS, Watson JA, Havel CM, White JM (1992) Identification of a prenylation site in Nat Rev Drug Discov 5(12):993–996. delta virus large antigen. Science 256(5061):1331–1333. 4. Jäger S, et al. (2012) Global landscape of HIV-human protein complexes. Nature 23. Gerber D, Maerkl SJ, Quake SR (2009) An in vitro microfluidic approach to generating 481(7381):365–370. protein-interaction networks. Nat Methods 6(1):71–74. 5. Brückner A, Polge C, Lentze N, Auerbach D, Schlattner U (2009) Yeast two-hybrid, a 24. Maerkl SJ, Quake SR (2007) A systems approach to measuring the binding energy powerful tool for systems biology. Int J Mol Sci 10(6):2763–2788. landscapes of transcription factors. Science 315(5809):233–237. 6. Koegl M, Uetz P (2007) Improving yeast two-hybrid screening systems. Brief Funct 25. Meier M, Sit RV, Quake SR (2013) Proteome-wide protein interaction measurements Genomics Proteomics 6(4):302–312. of bacterial proteins of unknown function. Proc Natl Acad Sci USA 110(2):477–482. 7. Yu H, et al. (2008) High-quality binary protein interaction map of the yeast inter- 26. Kubick S, Gerrits M, Merk H, Stiege W, Erdmann VA (2009) In vitro synthesis of actome network. Science 322(5898):104–110. posttranslationally modified membrane proteins. Curr Topics Membr 63:25–49. 8. Zhu H, et al. (2001) Global analysis of protein activities using proteome chips. Science 27. Fenz SF, Sachse R, Schmidt T, Kubick S (2014) Cell-free synthesis of membrane pro- 293(5537):2101–2105. teins: tailored cell models out of microsomes. Biochim Biophys Acta 1838(5): 9. Sachse R, et al. (2013) Synthesis of membrane proteins in eukaryotic cell-free systems. 1382–1388. Eng Life Sci 13(1):39–48. 28. Drayman N, et al. (2013) Pathogens use structural mimicry of native host ligands as a 10. Ramachandran N, et al. (2004) Self-assembling protein microarrays. Science 305(5680): mechanism for host receptor engagement. Cell Host Microbe 14(1):63–73. 86–90. 29. Noach-Hirsh M, et al. (2015) Integrated microfluidics for protein modification dis- 11. He M, et al. (2008) Printing protein arrays from DNA arrays. Nat Methods 5(2): covery. Mol Cell Proteomics 14(10):2824–2832. 175–177. 30. Campanero-Rhodes MA, et al. (2007) N-glycolyl GM1 ganglioside as a receptor for 12. Montor WR, et al. (2009) Genome-wide study of Pseudomonas aeruginosa outer simian virus 40. J Virol 81(23):12846–12858. membrane protein immunogenicity using self-assembling protein microarrays. Infect 31. Tsai B, et al. (2003) Gangliosides are receptors for murine polyoma virus and SV40. Immun 77(11):4877–4886. EMBO J 22(17):4346–4355. 13. Einav S, et al. (2008) Discovery of a hepatitis C target and its pharmacological in- 32. Walczak CP, Ravindran MS, Inoue T, Tsai B (2014) A cytosolic chaperone complexes hibitors by microfluidic affinity analysis. Nat Biotechnol 26(9):1019–1027. with dynamic membrane J-proteins and mobilizes a nonenveloped virus out of the 14. Glick Y, Avrahami D, Michaely E, Gerber D (2012) High-throughput protein expression endoplasmic reticulum. PLoS Pathog 10(3):e1004007–e1004007. generator using a microfluidic platform. J Vis Exp 23(66):e3849. 33. Ben-Ari Y, et al. (2013) Microfluidic large scale integration of viral-host interaction 15. Zheng C, et al. (2012) High-throughput immunoassay through in-channel microfluidic analysis. Lab Chip 13(12):2202–2209. patterning. Lab Chip 12(14):2487–2490. 34. Rual JF, et al. (2004) Human ORFeome version 1.1: a platform for reverse proteomics. 16. Pelham HR, Jackson RJ (1976) An efficient mRNA-dependent translation system from Genome Res 14(10B):2128–2135. reticulocyte lysates. Eur J Biochem 67(1):247–256. 35. Cassonnet P, et al. (2011) Benchmarking a luciferase complementation assay for de- 17. Carr PA, et al. (2012) Enhanced multiplex genome engineering through co-operative tecting protein complexes. Nat Methods 8(12):990–992. oligonucleotide co-selection. Nucleic Acids Res 40(17):e132. 36. Neveu G, et al. (2012) Identification and targeting of an interaction between a ty- 18. Goshima N, et al. (2008) Human protein factory for converting the transcriptome into rosine motif within hepatitis C virus core protein and AP2M1 essential for viral as- an in vitro-expressed proteome. Nat Methods 5(12):1011–1017. sembly. PLoS Pathog 8(8):e1002845. 19. Gazdar AF, Butel JS, Carbone M (2002) SV40 and human tumours: Myth, association 37. Glenn JS, Taylor JM, White JM (1990) In vitro-synthesized hepatitis delta virus RNA or causality? Nat Rev Cancer 2(12):957–964. initiates genome replication in cultured cells. J Virol 64(6):3104–3107.

Glick et al. PNAS | April 19, 2016 | vol. 113 | no. 16 | 4349