Proteomic-Based Identification of CD4-Interacting Proteins in Human Primary Macrophages

Background Human macrophages (Mφ) express low levels of CD4 glycoprotein, which is constitutively recycled, and 40–50% of its localization is intracellular at steady-state. Although CD4-interacting proteins in lymphoid cells are well characterised, little is known about the CD4 protein interaction-network in human Mφ, which notably lack LCK, a Src family protein tyrosine kinase believed to stabilise CD4 at the surface of T cells. As CD4 is the main cellular receptor used by HIV-1, knowledge of its molecular interactions is important for the understanding of viral infection strategies. Methodology/Principal Findings We performed large-scale anti-CD4 immunoprecipitations in human primary Mφ followed by high-resolution mass spectrometry analysis to elucidate the protein interaction-network involved in induced CD4 internalization and degradation. Proteomic analysis of CD4 co-immunoisolates in resting Mφ showed CD4 association with a range of proteins found in the cellular cortex, membrane rafts and components of clathrin-adaptor proteins, whereas in induced internalization and degradation CD4 is associated with components of specific signal transduction, transport and the proteasome. Conclusions/Significance This is the first time that the anti-CD4 co-immunoprecipitation sub-proteome has been analysed in human primary Mφ. Our data have identified important Mφ cell surface CD4-interacting proteins, as well as regulatory proteins involved in internalization and degradation. The data give valuable insights into the molecular pathways involved in the regulation of CD4 expression in Mφ and provide candidates/targets for further biochemical studies.


Introduction
Mass spectrometry (MS)-based identification of the components of purified protein complexes has become one of the most powerful and routinely used technologies for high-throughput detection of protein interactions [1,2]. The study of protein interactions by MS for identification of components of protein complexes gives powerful insights into protein function, binding partners and cellular pathways [3,4]. In most studies, proteins in a given complex are identified via MS analysis of in-gel tryptic digests of electrophoretically separated proteins of particular subcellular fractions (membranes, nuclei, intracellular compartments) or in co-immunoprecipitated complexes [5,6,7,8].
CD4 is the main cellular receptor used by human immunodeficiency viruses HIV-1, HIV-2 and simian immunodeficiency virus [9,10,11]. It is a type I transmembrane glycoprotein of 55 kDa expressed on the surface of Regulatory and Helper subsets of T lymphocytes and interacts with MHC class-II carrying cells [12]. CD4 increases the avidity of the low affinity interactions between the peptide-MHC complex on antigen presenting cells and the T cell receptor on the lymphocyte, and its association with the intracellular protein tyrosine kinase LCK modulates signal transduction [13]. In humans and rats CD4 is also expressed on cells of the monocyte/Mw lineage, although its function on these cells is poorly understood, and the protein expression levels are 10to 20-fold less than in T cells [14,15]. In lymphoid cells expressing LCK, 90% of CD4 is restricted to the cell surface and undergoes limited internalization [16]. Endocytosis of CD4 can occur, through clathrin-coated pits, when the cytoplasmic domain becomes serine phosphorylated, leading to its dissociation from LCK [17,18,19]. In myeloid cells, such as Mw, which do not express LCK, CD4 is constitutively internalized and 40-50% is intracellular at steady-state [16]. The pathways by which CD4 is removed from the cell surface and the protein-network involved are poorly defined. Cell surface CD4 levels can be down-regulated by exposure to gangliosides [20], soluble HIV-1 gp120 [21], phorbol esters [17,22] and during HIV-1 infection [23,24]. Moreover, down-regulation of viral receptors is a common mechanism used by most retroviruses to avoid superinfection (multiple rounds of infection) and to promote viral release. HIV-1 Nef protein accelerates CD4 internalization and degradation in the lysosomes [25], and at the late stages of HIV-1 infection, CD4 can be targeted for proteasomal degradation by HIV-1 Vpu [26,27,28].
In common with other laboratories we found that the kinetics of HIV-1 replication was modulated by the simultaneous presence of Mw and T cells in different ratios and activation states [33,34,35]. Data from our laboratory reported that HIV-1 viral production was typically slower in infected cultures in which Mw were cocultured with activated T cells. More recently, we extended these observations and showed that activated T cells produce soluble factors that selectively induce the internalization and degradation of CD4 in primary Mw, thus critically affecting HIV-1 entry in a process sensitive to the vacuolar ATPase inhibitor bafilomycin A1, and the proteasomal inhibitor, MG132 (Saraiva Raposo et al., manuscript under revision).
In this report we perform high-resolution mass spectrometry analysis of CD4 co-immunoisolates in human primary Mw, in order to characterise the CD4 containing complexes in steadystate and at different stages of CD4 internalization and degradation. The experimental strategy is shown in Fig. 1.

Conditioned media from activated T cells induces CD4 internalization and degradation in Mw
In order to effectively demonstrate the induction of CD4 internalization and degradation, we detected the expression of CD4 in Mw before and after treatment with conditioned media from activated T cells by flow cytometry. Eighteen hours posttreatment the expression of CD4 levels at the surface of Mw was barely detectable ( Fig. 2A), and the percentage of Mw expressing surface CD4 was significantly reduced by 4-fold (Fig. 2B). In addition, total CD4 expression (surface + intracellular) was diminished by 2-fold (Fig. 2C). Altogether, these data suggest the internalization and degradation of CD4 after treatment with conditioned supernatants from activated T cells.

Anti-CD4 co-immunoprecipitation sub-proteome in control Mw
We performed large-scale CD4 immunoprecipitations in normal resting primary human Mw, followed by LC-MS/MS. A representative gel of the resolved proteins after CD4 coimmunoisolation is shown in Fig. 3. In control resting Mw (condition 1), several cell surface proteins associated with CD4 were identified, including CD9, a tetraspanin-family member involved in cell adhesion, cell motility and IL-16 signalling [36,37,38,39]; CD163, involved in the clearance and endocytosis of hemoglobin/haptoglobin complexes [40,41]; integrin subunit beta (CD18), involved in cell surface adhesion and reported to interact with integrins alpha-M and alpha-X [42]; protein S100, a calcium binding protein known to be involved in phagocyte migration and infiltration at sites of wounding [43]; chemokine receptor 1 (CCR-1), a G protein-coupled receptor [44]; adaptor protein 2 (AP-2), a known adaptor protein which functions in protein transport via transport vesicles in different membrane trafficking pathways [25,45], and HLA class I, involved in antigen presentation [46]. CD4 was also found to be associated with cytoskeleton and actin-modulating proteins, such as gelsolin, tropomyosins and dynein. An unknown and uncharacterised protein, TPP1 was also identified. A summary list of interacting proteins is shown in table 1.

Anti-CD4 co-immunoprecipitation sub-proteome in induced internalization and degradation
Internalization and degradation of CD4 in Mw was induced by conditioned media from activated T cells (condition 2) and interacting proteins were identified by CD4 co-immunoprecipitation followed by LC-MS/MS. A representative gel of the resolved proteins after CD4 co-immunoisolation is shown in Fig. 3. Proteins identified included Cdc42, a small GTPase family protein involved in signal transduction and endocytosis [47,48]; proteins associated with late endocytic trafficking, such as LAMP1, a component of the lysosomal membrane [49,50]; RhoB, known to be associated with the late endosome membrane; adaptor protein 1 (AP-1), a subunit of clathrinassociated adaptor protein complex 1 [45,51,52]; Sec23B, a component of coating protein II (COPII) involved in the transport of vesicles from the Golgi apparatus to the endoplasmic reticulum, and Rab10/Rab11B, important components of vesicle recycling and protein turn-over [45,53]. Several cytoplasmic and cytoskeleton-related proteins were also identified, including fascin, myosin and tensin. Annexin A2, a calcium regulated membrane binding protein and flotillin-1, a scaffolding protein associated with caveolar membranes [54] were also identified with more than 5 unique peptides. A complete list of the uniquely identified proteins is shown in table 2.

Anti-CD4 co-immunoprecipitation sub-proteome in induced internalization and blocked degradation
In condition 3, internalization of CD4 in Mw was induced by the same conditioned media from activated T cells, as described for condition 2, and cellular degradation was blocked using the proteasome inhibitor MG132 and the vacuolar ATPase inhibitor bafilomycin A1. CD4-interacting proteins were identified by coimmunoprecipitations followed by LC-MS/MS. A representative gel of the resolved proteins after CD4 co-immunoisolation is shown in Fig. 3. CD4 was associated with a large number of proteins related to protein degradation, in particular the proteasome. Proteasome-related proteins such as the 26S regulatory subunit 6B, ubiquitin-like modifier activating enzymes E1 and E3 ubiquitin protein ligase subunit Itch [55,56,57,58] were identified. Figure 1. Strategy for the identification of CD4-complexes in human primary Mw. CD14 + monocytes were isolated from human blood by magnetic cell sorting (MACS) and cultured for 7 days in the presence of M-CSF. One hundred million day 7 fully differentiated Mw were left untreated (Condition 1, blue), treated with conditioned media from activated T cells (Induced CD4 internalization and degradation, Condition 2 red) or treated with conditioned media from activated T cells in the presence of the proteasomal inhibitor MG132 and the inhibitor of vacuolar ATPases bafilomycin (BafA1) (Induced CD4 internalization but blocked degradation, Condition 3 green). Eighteen hours later, cells were detached from tissue culture plates, lysed and large-scale anti-CD4 immunoprecipitations (IP) using monoclonal antibody against CD4 (clone QS4120) or isotype control IP were carried out. IP products were loaded onto SDS-PAGE pre-cast gels and electrophoresis were run. Protein gels were coomassie stained, gel lanes were cut into 10 equal pieces and trypsin-digested. Proteins were identified by LC-MS/MS. doi:10.1371/journal.pone.0018690.g001 Proteins associated with antigenic presentation and intracellular protein trafficking were also identified, such as MHC-I molecules (HLA-A and HLA-B), ERp29 and ERp1 (endoplasmic reticulum chaperones) [59]. Although identified with one unique peptide, but with high iProphet probability scores, we also detected 7 proteins, including components of vacuolar proton-transporting ATPases, such as V-type proton ATPase subunits D and G1. A complete list of the uniquely identified proteins is shown in table 3. Table 4 lists the proteins commonly identified in all three conditions. CD4 is internalized and degraded after treatment with conditioned media from activated T cells. Mw were treated with conditioned media from activated T cells for 18 hours or left untreated, followed by flow cytometry staining with directly conjugated mAb to CD4. A Black histogram represents the appropriate isotype control. Histograms show the intensity of the signal on the X-axis with a log 10 -scale and the percentage of maximum expression on the Y-axis. Representative staining of more than five donors tested (n.5). B Bars represent the mean percentage of Mw expressing surface CD4 with SD error bars from ten independent donors (n = 10). C Total CD4 expression levels (surface + intracellular) were determined by dividing the geometrical MFI of the antibody staining over the MFI of the isotype control. Bars represent the mean values of five independent donors (n = 5) with SD error bars. In B and C, black bar corresponds to untreated Mw and white bar corresponds to conditioned media treated Mw (T cell Sup). doi:10.1371/journal.pone.0018690.g002

Western Blotting analysis of CD4 co-immunoprecipitates in Mw
Mass spectrometry identifications of CD9, E3 ubiquitin ligase Itch and clathrin heavy chain in CD4 co-immunoisolates were confirmed by western blot analysis. As anticipated, CD4 was identified in all Mw sample conditions, but at reduced levels in condition 2. Clathrin heavy chain 1 was co-immunoisolated with CD4 in all three conditions and the E3 ubiquitin ligase subunit Itch was only co-immunoisolated with CD4 when cellular degradation was blocked. CD9 antigen was only co-immunoisolated with CD4 in the control Mw. CCR5, reported to interact with CD4 at the surface of Mw and T cells [60], was not identified by mass spectrometry in any of the conditions described above and was not detected by western blot analysis of CD4 co-immunoisolates (Fig. 4).

GO annotations
Uniquely identified protein identifications in all three conditions were exported to ProteinCenter and GO annotations were carried out. In induced CD4 internalization and degradation (condition 2) there is an over-representation of proteins associated with the endosome, vacuole and Golgi, when compared to control Mw (condition 1). Moreover, when cellular degradation is blocked (condition 3) the over-represented CD4-associated proteins are related to the proteasome, endoplasmic reticulum, organelle lumen, mitochondrion and cytosol (Fig. 5A). Proteins related to DNA and nucleotide-binding are over-represented in condition 3 and metal binding proteins are over-represented in condition 2. No proteins with structural molecular activities were uniquely identified in condition 3, in contrast to control or condition 2, where 30% and 15%, respectively, of the uniquely identified proteins fall into this category (Fig. 5B). Proteins related to cell organization and biogenesis, cell differentiation, development and transport are greatly over-represented in condition 2 over condition 3. In control Mw, proteins related to response to stimulus and defence response are over-represented over the other two. Cell motility-related proteins cluster with CD4 in control Mw and in condition 2 (Fig. 5C).

Discussion
Mass spectrometry analysis of CD4 co-immunoisolates, supplemented with GO annotations provided useful information on the clustering of CD4 molecules in resting Mw and elucidated the protein-network involved in the internalization and degradation. CD4 in resting Mw showed association with a range of molecules found in the cellular cortex and membrane rafts. Consistent with earlier reports [19,25,61], we also observed CD4 association, and confirmed by western blotting, with components of clathrinmediated endocytosis, such as clathrin heavy chain 1 and the adaptor protein AP-2, clearly suggesting that in resting Mw CD4 undergoes constitutive internalization and recycling [16,18,62]. AP-2 has been reported to be involved in the initial formation of clathrin coated pits at the plasma membrane, and it is an important mediator of receptor internalization and clathrin assembly [63]. We observed CD4 association with the tetraspanin protein CD9, and as both CD4 and CD9 are able to bind IL-16 in mast cells [36,64], this association might in fact be physiologically relevant in Mw.
In addition to CD4, HIV-1 requires CXCR4 or CCR5 to enter target cells. Xiao et al., reported a constitutive cell surface association between CD4 and CCR5 [60] and showed that the presence of gp120, leads to the clustering of CD4 and CCR5. However, they stated that it was difficult to co-immunoisolate CD4 and CCR5 in human primary Mw and CD4 + T cells in the absence of gp120, arguing that the levels of both receptors were very low and the techniques used were not sensitive enough. Employing highresolution mass spectrometry analysis on a large sample of primary Mw, a more sensitive technique than the one used by Xiao et al., we did not detect CCR5 molecules in CD4 co-immunoisolates. Although a constitutive CD4-CCR5 interaction in the absence of gp120 might still exist, our results do not support this notion.
Many reports to date have shown that in CD4 + T cells LCK binds directly to the cytoplasmic tail of CD4 [13,16,18], providing stability at the cell surface. As we did not identify any Src family protein kinases in CD4 co-immunoisolates in Mw, it seems unlikely that this kinase family plays a similarly prominent role in the regulation of CD4 in Mw, as it does in T cells. This could also explain the faster turn-over of CD4 in Mw compared to T cells.
Data from our laboratory showed that upon treatment with conditioned media from activated T cells, CD4 expression in Mw is down-regulated due to induced internalization and degradation (Saraiva Raposo et al., manuscript under revision). Under this condition, CD4 was associated with specific components of signal transduction and transport pathways, including plasma membrane-associated small GTPases, such as Cdc42, Ras-related proteins and RhoB. The small GTPases of the Ras superfamily are well known to have roles in endocytosis [65,66]. RhoB regulates endosomal trafficking, in co-operation with mDia1 and Src kinase [67], and Cdc42, which has also been connected to cell migration and cell polarity, has also been linked to the regulation of endocytosis [68]. We observed an interaction between CD4 and LAMP1, suggesting the intervention of lysosomes in the downregulation of CD4. This observation correlates with the effect induced by the phorbol ester PMA in the induction of CD4 internalization and degradation [69]. Overall, the over-representation of endosome-related proteins in this condition, clearly clusters CD4 with the endocytic pathways.
When Mw are treated with conditioned media from activated T cells in the presence of MG132 and bafilomycin A1, CD4 can still be internalized, but it is not degraded (Saraiva Raposo et al. manuscript under revision). Under this condition, CD4 was associated with several components of the proteasome, such as regulatory and activating subunits involved in the cascade of protein ubiquitination, suggesting the involvement of the proteasomal pathway. We identified the member of the E3 ubiquitin (Ub) ligase family, Itch/AIP4 to be associated with CD4 and confirmed it by western blot. Itch is a member of the HECT domain-containing E3 Ub ligases and has been implicated in the post-translational modification with Ub of CXCR4, followed by desensitization at the cell surface by engagement to its cognate ligand SDF-1a [70].
In the early stages of HIV-1 infection, the viral protein HIV-1 Nef, reported to accelerate CD4 down-regulation, avoiding viral superinfection and promoting efficient viral spread and optimal viral particle production [25], also alters the intracellular trafficking of MHC-I and MHC-II molecules [71]. HIV-1 Nefdependent reduction of surface MHC-I protects HIV-infected primary T cells from recognition and killing by HIV-specific cytotoxic T cells in vitro [72]. Schaefer et al. reported that HIV-1 Nef targets MHC-I molecules and CD4 for degradation in the lysosomes, by showing co-localization of CD4 and a subset of HLA-A2 proteins in late endosomes and multi-vesicular bodies (MVB) [73]. We showed an interaction between CD4 and components of MHC-I (HLA-A and HLA-B). Although, our system is an HIV-1 Nef-independent system, both induced pathways seem to have some degree of similarity.
Overall in resting macrophages CD4 shows association with a range of proteins found in the cellular cortex, clathrin coated pits and membrane rafts. In induced internalization the spectrum of proteins clustered with the receptor changes and CD4 becomes associated with components of signal transduction and transport. Finally, under conditions where protein degradation pathways are chemically blocked, CD4 associates with components of the proteasome and ubiquitin-modifying proteins. This is the first co-immunoisolation LC-MS/MS-based identification of CD4 complexes in human primary Mw elucidating CD4-interacting proteins and the protein-network involved in its induced internalization and degradation. Due to its importance in the context of HIV-1 infection, revealing the CD4 ''interactome'' can lead to the discovery of important proteins in the pathogenesis of the virus. In conclusion, our mass spectrometry data contribute to a better understanding of the fate of CD4 molecules in resting Mw and in induced internalization and degradation.

Ethics statement
Adult human blood was obtained from anonymous donors through the UK National Blood Service and tested negative for HIV-1, hepatitis B/C, and syphilis. Local IRB approval was sought for this work from Oxford University's Central University Research Ethics Committee (CUREC), and we were informed that specific ethical approval was unnecessary for this study, in accordance with their guidelines on the use of human blood (http://www.admin.ox.ac.uk/curec/resrchapp/faqethapp. shtml): ''CUREC does not require an ethics form for laboratory research using buffy coats. However there are occasions when the National Blood Service donating the buffy coats may require ethical approval from the University. In this instance a checklist completion will suffice. Applicants should answer Question C (8) as a 'NO'. A covering note should be sent to the Secretary of the MSD IDREC with the checklist explaining that the research uses buffy coats and the NBS requires University ethical approval.'' Although not required by NBS, we completed a checklist as indicated and received exemption from MSD IREC.

Cells and reagents
PBMC were isolated using Ficoll-Plaque Plus (GE Healthcare Life Sciences, Europe) density gradient centrifugation from

Flow cytometry
CD4 expression levels were detected by direct immunofluorescence. Mw in staining buffer (10 mg/mL human IgG (Sigma UK), 1% FCS and 0.01% NaN 3 ) were incubated with 5 mg/mL anti-CD4 specific mAb (clone RPA-T4, Becton Dickinson) or matched isotype control (IgG1k, Becton Dickinson) on ice for 30-45 min. For intracellular staining, cells were first fixed, then permeabilized with 0.2% saponin (Sigma, UK) and stained. The percentage of positive cells and the mean fluorescence intensity (MFI) were analyzed by FACS Calibur (Becton Dickinson) with 15,000-20,000-gated events collected. The data was processed using FlowJo (version 7.2.4). Protein expression levels were determined by dividing the geometrical MFI of the Ab staining over the MFI of the isotype control.

Western blotting
Adherent Mw were washed free of media, detached using ice cold 10 mM EDTA/PBS and cell pellets were lysed in ice-cold lysis buffer (50 mM Tris-HCl pH 8, 150 mM NaCl, 1% (v/v) n-Dodecyl b-D-maltoside (Sigma), 16 protease inhibitor cocktail (Roche), phosphatase inhibitor cocktail 2 (Sigma)). n-Dodecyl b-Dmaltoside is a water-soluble non-ionic detergent, shown to be a rather gentle detergent able to preserve protein activity and structure better than many commonly used agents, such as Triton X-100, NP-40, CHAPs and octyl-b-glucoside [74,75,76,77]. Lysates were centrifuged for 10 min at 4uC, 13,0006g to separate insoluble material and cleared lysate was resuspended in 16 Laemmli sample buffer (Invitrogen, UK) under reducing conditions and heated for 10 min at 90uC. Lysates were electrophoresed through SDS-PAGE gels and proteins were electroblotted to PVDF transfer membranes. Blocked membranes were incubated with one of the following primary antibodies diluted in 3% (w/v) BSA (Sigma) in 16 PBS-T (16 PBS, 0.1% (v/v) Tween-20) for 2 hours at room temperature or over-night at 4uC: rabbit polyclonal antibody anti-CD4 (clone H-370), rabbit polyclonal antibody anti-CD9 (clone H-110), rabbit polyclonal antibody anticlathrin heavy chain 1 (clone H-300), rabbit polyclonal antibody anti-E3 Ubiquitin ligase (clone H-110) (all from Santa Cruz) and mouse monoclonal antibody anti-CCR5 (clone CTC5, R&D Systems). Primary antibodies were detected using the matching LI-COR secondary antibodies and membranes were scanned using the quantitative western blotting imaging system Odyssey (LI-COR).

Immunoisolation analysis
Anti-CD4 immunoisolation reactions consisted of 10 mL of protein G-Sepharose bead slurry (4B Fast Flow, Sigma, UK) per 1610 7 lysed cells and 5-10 mg mouse monoclonal antibody anti-CD4 (clone QS4120, Santa Cruz) was incubated for 2 hours at room temperature to allow binding of the antibody to the beads. Beads were gently spun, cell lysate was added to the mixture of beads/antibody and the reactions were incubated by inversion for 3 hours at 4uC. The immunoisolates were collected by centrifugation for 5 min at 4uC, and washed three times for 5 min with lysis buffer. The final immunoisolates were resuspended in Laemmli sample buffer under reducing conditions and heated for 10 min at 90uC, before loading them onto a gel. Isotype control immunoprecipitations were also performed to identify background binding proteins.

Mass spectrometry and protein identification
Anti-CD4 or isotype control immunoisolated pellets were reduced in NuPAGE sample reducing agent (Invitrogen, UK), separated on a NuPAGE Novex 4-12% Bis-Tris gel (Invitrogen, UK) and coomassie stained. Gel lanes were excised, cut into 10 equal portions and in-gel digested with trypsin [78]. Briefly, gel bands were diced into cubes and destained in 25 mM ammonium bicarbonate in 50:50 water/acetonitrile. Proteins were reduced with 10 mM DTT and alkylated with 55 mM iodoacetamide. Gel bands were then incubated with 3 mg of trypsin (Promega, UK) in 25 mM ammonium bicarbonate over-night at 37uC. Peptides were extracted and desalted using home-made C18 tips. Mass spectrometry data were acquired on an Orbitrap mass spectrom- eter (Thermo) fitted with a nanospray source (Proxeon, Denmark) coupled to a U3000 nano HPLC system (Dionex, UK). The samples were loaded onto a 15 cm long, 100 micron ID, homepacked column manufactured by packing a Picotip emitter (New Objective, USA) with ProntoSIL C18 phase; 120 angstrom pore, 3 micron bead, C18 (Bischoff Chromatography, Germany). HPLC was run in a direct injection configuration. One hundred and twenty minute gradients were used to resolve the peptides. The Orbitrap was run in a data dependent acquisition mode in which the Orbitrap resolution was set at 60,000 and the top 5 multiply charged precursors were selected for MS/MS fragmentation. Samples were typically injected three times in order to increase the number and confidence of identifications. RAW data files were converted to mzXML format using ReAdW (version 4.2.1) and submitted to the in-house developed Central Proteomics Facilities Pipeline (CPFP) [79]. The CPFP is based on the Trans Proteomic Pipeline tools (version 4.2.1) [80] and implements automatic identification of MS/MS spectra using multiple search engines to maximise coverage of a sample. mzXML files were converted to suitable peaklist formats for submission to Mascot (Matrix Science), X!Tandem with k-score plugin [81] and OMSSA [82]. Searches are performed automatically and executed on a compute cluster, using Sun GridEngine, and the resulting peptide identifications from each search engine are validated with PeptideProphet [83]. iProphet is used to combine peptide hits from each three search engines and refines identification probabilities. ProteinProphet infers protein identifications from the resulting combined peptide list and performs grouping of ambiguous hits [84]. Protein identifications were exported from the CPFP and uploaded to ProteinCenter (Proxeon, Denmark) for filtering, annotation, classification, and interpretation. Searches were performed against a concatenated target/decoy human IPI database providing an empirical false discovery rate (FDR) and criteria for protein identification included 1% FDR and two or more unique peptides identified for each individual protein.
Proteins that were identified in the isotype control immunoprecipitations were filtered out of the final interpretation. Uniquely identified proteins were only identified in the condition tested and commonly identified proteins were identified in all conditions tested.

Statistical analysis
Statistical analysis was performed by paired t-test using GraphPad Prism (version 5.01). Stars indicate the p-value: **p = 0.01-0.001; ***p,0.001. Significance refers to difference from the controls, unless otherwise indicated. N refers to the number of blood donors tested. Figure 5. Gene Ontology (GO) annotations of the uniquely identified proteins in anti-CD4 immunoprecipitations in Mw. Protein identifications from the three different conditions were exported from the in-house developed Central Proteomics Facilities data analysis pipeline (CPFP) and uploaded to ProteinCenter software. A illustrates the percentage of protein identifications versus protein cellular localizations (GO cellular annotations); B illustrates the percentage of protein identifications versus protein molecular functions (GO molecular annotations) and C illustrates the percentage of protein identifications versus protein biological functions (GO biological annotations). Blue bars represent the percentage of unique proteins identified in condition 1 (Resting macrophages); Red bars represent the percentage of unique proteins identified in condition 2 (Induced CD4 internalization and degradation); Green bars represent the percentage of unique proteins identified in condition 3 (Induced CD4 internalization and blocked degradation). doi:10.1371/journal.pone.0018690.g005