IDENTIFICATION OF NOVEL FANCD2 INTERACTING PROTEINS VIA IMMUNOPRECIPITATION AND MASS SPECTROMETRY

Fanconi Anemia (FA) is a rare genetic disease caused by biallelic mutations in one of sixteen genes involved in the FA-BRCA DNA damage repair pathway. The proteins encoded by these genes function cooperatively in a common pathway which resolves lesions caused by interstrand crosslinks (ICLs). A critical step in this pathway is the monoubiquitination and chromatin targeting of FANCD2 and FANCI. The mechanism by which these proteins are targeted to chromatin is not understood. FANCD2 is known to interact with several downstream proteins while associated with chromatin. Finding new FANCD2 interacting proteins is critical to understanding how FANCD2 functions and how it is regulated within the cell. I have identified several candidate interacting proteins by immunoprecipitations (IPs) coupled with mass spectrometry. Candidates include transcription factors, chromatin remodeling complex components and proteins involved in chromosome maintenance and stability. These interactions are being validated and functionally characterized using a variety of techniques.


STATEMENT OF THE PROBLEM
FANCD2 is a critical protein in the FA pathway, which is monoubiquitinated by the core complex and localized to chromatin. Abrogation of either of monoubiquitination or chromatin localization results in the defective repair of damage caused by interstrand crosslinks. Both the regulation and function of FANCD2 are poorly understood. Identification of novel FANCD2 interacting proteins is critical to fully understanding the role that FANCD2 plays in the cell. The goal of the project is to identify novel FANCD2 interaction protein candidates using immunoprecipitation in tandem with mass spectrometry and then to validate and functionally characterize these interactions. The hope is that this data may shed light on the biochemical process in the cell and open new opportunities for patient treatment.

Fanconi Anemia
Fanconi anemia is a rare genetic disease characterized by congenital defects, genomic instability, and a predisposition to bone marrow failure and cancer (Walden 2014). Biallelic mutations in one of sixteen bona fide FA genes cause this disease.
These gene products act cooperatively in the FA-BRCA pathway to recognize and repair ICLs within the DNA (Walden 2014). ICLs are highly genotoxic complex lesions which prevent DNA strand separation required for both replication and transcription (Deans 2011). The inability to correctly repair these lesions may lead to replication fork stalling or arrest, deleterious repair, or cell death particularly in white blood cells (Deans 2011). Most FA patients present anemia and early onset of bone marrow failure. Hypersensitivity to DNA crosslinking agents is a hallmark of FA patient cells (Kim 2012). This combination of increased propensity for hematological abnormalities and cancers, with fewer effective treatment options, makes the understanding of the molecular pathology critical for development of new treatment opportunities.

FA Activation
Much work has been done in recent years to help elucidate the mechanistic action of the FA proteins. The recognition and stabilization of the lesion is accomplished by FANCM, FAAP24, MHF1, and MHF2, which prevent replication fork collapse (Meetei 2005, Ciccia 2007, Zhijiang 2010. Following lesion recognition the core complex is recruited which is composed of FANCA, FANCB, FANCC, FANCE, FANCF, FANCF, FANCG, FANCL, and FANCM. The core complex monoubiquitinates FANCD2 and FANCI on K561 and K523 respectivly, which leads to chromatin targeting and the recruitment of downstream DNA repair proteins (Garcia-Higuera 2001, Sims 2007, Smogorzewska 2007. The monoubiquitination is seen as a marker of pathway activation. The loss of any core complex member, other than FANCM, abolishes FANCD2 and FANCI ubiquitination (Bakker 2009).

FANCD2
FANCD2 is an orphan protein which has until recently been largely uncharacterized. The closest protein to FANCD2 is its paralog FANCI, which together form a heterodimer known as the ID2 complex. FANCD2 has four recognized domains, the EDGE, PIP, CUE, and NLS domains (Montes de Oca 2005, Howlett 2009, Rego 2012, Boisvert 2013). FANCD2 has also been implicated in binding directly to chromatin though the mechanism for this remains unknown. Different forms of DNA, such as circular dsDNA, dsDNA, fragments and, ssDNA, have been shown to specifically increase the association between chromatin and FANCD2 (Sareen 2012). It has also been proposed that FANCD2 harbors both nucleosome chaperone activity and the ability to promote site-specific transcriptional activation (Sato 2012, Park 2013. Despite these domains being characterized the functional and mechanistic role of this protein is still largely unknown.

FANCD2 Monoubiquitination
The monoubiquitination of FANCD2 and FANCI has been used as a marker for activation of the FA-BRCA pathway. The monoubiquitination event is key for the recruitment of FANCD2 and FANCI to chromatin. FANCD2 mutants that harbor a K561R mutation are not able to become ubiquitinated and do not rescue sensitivity to crosslinking agents (Garcia-Higuera 2001). The primary function of the FA core complex is this ubiquitination reaction. The ubiquitination is mediated through the E3 ligase activity of FANCL and UBE2T, an E2 ubiquitin conjugating enzyme specific for FANCL (Meetei 2003, Machida 2006). This ubiquitination is easily measurable by western blotting as the ubiquitin conjugation causes a detectable shift in the masses of FANCD2 and FANCI. FANCB, FANCL and FAAP100 have been shown to form a subcomplex which is able to ubiquitinate FANCD2 in vitro indicating that the other core complex members have some other unknown functions which are required for efficient ubiquitination in vivo (Rajendra 2014). This ubiquitination conjugation is a reversible reaction using the deubiquitinating enzymes USP1 and UAF1 (Nijman 2005, Cohn 2007).

FA and DNA Repair
The monoubiquitination event of FANCD2 and FANCI leads to the eventual recruitment of downstream repair proteins. The downstream proteins consist of FANCD1/BRCA2, FANCJ/BRIP1, FANCN/PALB2, FANCO/RAD51C, FANCP/SLX4, and FANCQ/XPF. It was originally thought that FANCP/SLX4 and FAN1 were recruited to the sites of damage through a specific interaction with ubiquitinated FANCD2, however it is now known that the nucleases are recruited through a mechanism independent of FANCD2 monoubiquitination (Kimiyo 2011, Kim 2011, Shereda 2010, Chaudhury 2014, Lachaud 2014. FANCP/SLX4 acts as a recruitment platform for the nuclease complexes XPF-ERCC1, MUS81-EME1 as well as SLX1, which are able to contribute to resolving the complex structure generated by the ICLs (Zhang 2014). The incisions created by the nucleases are critical for crosslink unhooking and enabling initiation of translesion synthesis (TLS) as well as double strand breaks (DSBs) formation (Walden 2014).

TLS and Homologous Recombination
Once the crosslink has been unhooked TLS polymerases are recruited to the site to bypass the lesion. The TLS polymerases are able to synthesize across the lesion and allow for resumption of normal replication (Sharma 2013). FANCC has been shown to promote the recruitment of factors involved in switching to the error prone TLS replicative mechanism (Niedzwiedz 2004). Once the ICL lesion has been synthesized over by TLS, the DSBs that were created during nucleotide excision must be repaired in an error free manner to avoid the loss of genetic material and an increase in genomic instability. The remaining downstream proteins in the FA-BRCA pathway FANCD1/BRCA2, FANCJ/BRIP1, FANCN/PALB2, and FANCO/RAD51C promote RAD51 loading onto single stranded DNA. RAD51 is critical to the homologous recombination (HR) pathway by its ability to coordinate with a homologous DNA sequence and promote strand invasion (Mazón 2010). This process prevents the loss of genetic information by using the paired strand as a template for the damaged strand ensuring accurate hybridization across the previously damaged region.
FA patients are thought to have difficulty in promoting HR over nonhomologous end joining (NHEJ) repair which, in contrast, is an error prone mechanism. The deregulation of this error prone repair pathway has been implicated as a possible cause for the genomic instability found in FA patients (Adamo 2010, Pace 2010, Bunting 2010. It is important for patient treatment that both the mechanics and regulation of the FA-BRCA pathway are well understood.

Unbiased Screening for FANCD2 Interactors
The discovery of novel FANCD2-interacting proteins is essential for a more complete understanding of the role that FANCD2 plays in the cell. Using a variety of methods FANCD2 has been shown to interact with several proteins including, but not limited to, FANCI, FANCE, FANCJ/BRIP1, MEN1, CtIP, and the MCM helicase proteins (Sims 2007, Smogorzewska 2007, Pace 2002, Gordon 2003, Chen 2014, Jin 2003, Murina 2014, Unno 2014, Lossaint 2013. Biochemical based studies have resulted in a huge increase of the understanding in how FA works. The first six FA genes were identified through a positional cloning approach (Lo Ten Foe 1996, Fanconi anaemia/Breast cancer consortium 1996, Strathdee 1992, de Winter 1998, de Winter and Léveillé 2000. Biochemical approaches such as co-immunoprecipitations (co-IPs) have been instrumental in the discovery of the ten most recently discovered FA genes as well as several interacting partners (Meetei 2004, Howlett 2002, Dorsman 2007 (Howlett 2002, Pichierri 2004. Immunoprecipitation coupled with mass spectrometry is a proven technique within the FA field, contributing directly to the discovery of FANCN/PALB2 (Xia 2006). Similar techniques were recently used to identify the interaction between FANCD2 and the MCM helicases (Lossaint 2013). Using large scale systems combined with the sensitivity of modern biochemical techniques we hope to discover novel FANCD2 interacting proteins using an unbiased system. The maternal allele contains a frameshift mutation resulting in a severe truncation, and the paternal allele has a missense mutation, causing the cells to express highly diminished amounts of FANCD2 protein which fail to correct ICL sensitivity  (EGF, Peprotech AF-100-15), 1 mg/mL hydrocortizone (Sigma Aldrich H-0888), 1 mg/mL cholera toxin (Sigma Aldrich C-8052), 10 mg/mL insulin (Sigma Aldrich I-1882), 500 units/mL penicillin, and 500 µg/mL streptomycin (Gibco 15070-063).
Induction of shRNA expression was achieved by treating with 0.01µg/mL of doxycycline (Sigma Aldrich D-9891) for 72 hours and confirmed by GFP expression.
Cells were cultured at 37 o C with 5% CO 2 and subjected to 0.05% trypsin EDTA dissociation solution (Gibco 25300-054) for maintenance. Cells were treated with mitomycin C (MMC, Sigma Aldrich M-0503) to induce crosslink formation.

Cellular Fractionation
Cells underwent a subcellular fractionation to enrich for chromatin associated proteins. Following treatment, cells were collected in 0.05% trypsin EDTA dissociation solution and DMEM, and pelleted in a centrifuge at 1200xg at 4 o C over 4 minutes. The pellet was resuspended and washed in phosphate buffered saline pH 7.4 (Gibco 10010-023) and pelleted again. Cells were first lysed in ice cold cytoskeletal (CSK) buffer (10 mM PIPES pH 6.8, 300 mM sucrose, 100 mM NaCl, 3 mM MgCl 2 , 1 mM EGTA, and 0.5% v/v Triton-X-100) for 10 minutes at 4 o C. The remaining pellet was washed once with CSK buffer and subjected to either nuclease reaction buffer (NRB) (20 mM Tris HCl, 100 mM KCl, 2 mM MgCl 2 , 1mM CaCl 2 , 0.3M Sucrose, 0.1% v/v Triton-X-100, Roche protease inhibitor cocktail tablet) or ice cold ATM lysis buffer (Sun 2009) (20 mM HEPES pH 7.4, 150 mM NaCl, .2% Tween20, 1.5 mM MgCl, 1 mM EGTA, 2 mM DTT, 50 mM NaF, 500 µM Na 3 O 4 V, 1 mM PMSF, Roche protease inhibitor cocktail tablet). The sample was subjected to sonication for ten seconds at 10% amplitude using a Fisher Scientific Model 500 Ultrasonic Dismembrator then supplemented with 0.1% v/v micrococcal nuclease (New England Biolabs M0247S) for twenty minutes at room temperature or 0.05% v/v benzonase (Sigma Aldrich E-1014) for thirty minutes on ice, respectively. Samples were spun at 16100xg for two minutes and the supernate containing the chromatin associated proteins were placed in a new ice cold tube. Chromatin fractions were quantified using the bicinchoninic acid (BCA) assay (PI-23227, Fisher) to allow for normalization.

V5 Agarose Immunoprecipitation
V5 immunoprecipitations were performed with V5 conjugated agarose beads (ab1229, Abcam). Beads were washed three times for five minutes in NETN100 (20 mM Tris-HCl, 100mM NaCl, 1 mM EDTA, 1mM Na 3 O 4 V, 1 mM NaF, Roche protease inhibitor cocktail tablet) + 1% bovine serum albumin (BSA) (BP1600-100, Fisher) than once more for fifteen minutes to block the beads. The beads are than washed three times in NETN100 buffer. Samples were brought up to an equal volume using the ATM lysis buffer. V5 conjugated beads were added and allowed to incubate for two hours at 4 o C while nutating. Beads were separated from the supernatant by gentle centrifugation. The supernatant was aspirated and the beads were washed four times with ice cold NETN100 for five minutes. After the final wash the beads were aspirated dry using a needle. The beads are eluted into 2x LDS (NP-0008, NuPage) with 5% β-mercaptoethanol by heating at 95 o C for fifteen minutes.

Gel Staining, Tryptic Digestion and Mass Spectrometry
The SDS-PAGE matrix was stained using SimblyBlue SafeStain (LC6060, Invitrogen) or mass spec compatible silver stain (24600, Thermo). Staining was performed as described in the respective protocols. Tryptic digestion was performed using the in-gel tryptic digestion kit (89871, Thermo). Protocol was performed as described in the manual using TCEP for reduction and Iodoacetamide for alkylation.
When possible, digestion was performed inside a biosafety cabinet to reduce the prevalence of keratin contamination. Tryptic digestion was incubated at 30 o C overnight. The supernate was removed and placed into a clean new tube. The remaining gel pieces were subjected to a secondary extraction using 50% acetonitrile and 5% formic acid. The secondary extraction was collected and combined with the primary extraction than stored in the -20 o C freezer. Samples were submitted to Dr.
James Clifton at Brown University EPSCoR proteomics facility. Samples were submitted as low complexity samples to be run through LC coupled MS/MS on the LTQ Orbitrap Velos mass spectrometer.

Spectral Analysis
Analysis of spectral data was performed in multiple iterations using different systems. Initial analysis was performed using the MASCOT database which uses a probability based peptide finger printing method to automatically identify and evaluate spectral data (Perkins 1999). ProteoIQ v2.2 was used to remove hits that fell below the 5% protein false discover rate threshold. ProteoIQ was used in tandem with manual curation of the raw MASCOT data to find samples that were differentially detected in the experimental and control samples. Exclusions were made based on relative abundance in the control samples, repositories of known contaminants, and subcellular location, as a result of the prerequisite fractionation (Mellacheruvu 2013). Manual curation relied heavily on the UniProt database though several cases required rigorous searches of the literature to determine protein function, possible validity, and relevance (Apweiler 2004).

FANCD2 Interacting Proteins can be Detected by Mass Spectrometry.
To determine if the system was workable, several pulldowns were performed to enrich for FANCD2 immune complexes. Using a commercially available V5 antibody it was possible to enrich for both FANCD2 and monoubiquitinated FANCD2 ( Figure 1). The remaining eluate from the immunoprecipitation did not present a differential banding pattern when stained (Figure 2). Lack of visual detection does not preclude detectable differences between samples, so gel pieces from the SDS-PAGE were excised and examined by mass spectrometry. The mass spectrometry results confirmed the presence of FANCD2, however it did not identify any strong candidates for further analysis ( Table 1).

.2 FANCD2-V5 is Functionally Incorporated into the Cell
Using the pLenti FANCD2-V5 vector to correct the cells shows a functional incorporation of the protein into the cellular machinery. The FANCD2-V5 undergoes ubiquitination following exposure to MMC (Figure 3) and is also localized to chromatin (Figure 4). To adequately correct the sensitivity to crosslinking agents FANCD2 must undergo monoubiquitination, chromatin localization, and discrete nuclear foci formation. Monoubiquitination demonstrates that the protein is at least interacting with the components of the core complex which ubiquitinate the protein.

Fractionation Allows Isolation of FANCD2 Immune Complex Members from Endogenous Systems
Isolating the chromatin associated portion of cells has shown to enrich activated FANCD2. However the fractionation method required large volumes of cells to generate large enough protein volumes to successfully IP from. Using transformed cells such as HeLa cells allows for the rapid generation of large volumes of high protein content cells. Using unmodified HeLa cells removes the option to pull down using the V5 epitope tag. The cells were fractionated and then underwent a FANCD2 IP using an antibody against FANCD2 to enrich for FANCD2 immune complexes ( Figure 5). The eluate was resolved by SDS-PAGE and stained to observe differential banding patterns which could be excised and submitted for analysis by mass spectrometry (Figure 6). The analysis revealed a number of functionally diverse candidates with varying detection strengths ( Table 2). One of the difficulties with the endogenous system is the lack of specificity with the antibody used. The best available antibody for FANCD2 has multiple targets within the cell. The issue of nonspecific binding is exacerbated by the PIS control which also binds several proteins nonspecifically and may invalidate or mask interacting partners. While the HeLa system is able to identify several candidates the large amount of contaminants and nonspecific interactions in the data demonstrate the need for a more sensitive system to screen for interacting proteins.

Figure 5. Enrichment of FANCD2 from HeLa cells using FANCD2
immunoprecipitation HeLa cells were grown to confluency and half were treated with 250 nM MMC for sixteen hours. The cells were fractionated using CSK buffer and NRB supplemented with micrococcal nuclease. 1 mg of the chromatin associated fraction was incubated with a FANCD2 antibody (NB100-182, Novus Biologicals) and pulled down using magnetic beads. The sample was also incubated with rabbit pre immune serum and magnetic beads as well as just the magnetic beads. Only the MMC treated The IPs are shown in the blot above as they were the only samples to be submitted to proteomic analysis. The immunoblot for FANCD2 (sc-20022, Santa Cruz) demonstrates that monoubiquitinated chromatin associated FANCD2 is present in the FANCD2 IP and is not pulled down nonspecifically by rabbit serum or the magnetic beads. The Immunoblot for H4 shows that the chromatin fractionation was successful.  May play a role in mRNA splicing 0 3 TP63 transcriptional activator or repressor 0 3

Cells Identifies Several Interacting Candidates
Previous work with the PD20 patient cells showed it was difficult to generate a large enough volume of the chromatin associated fraction to be able to enrich enough FANCD2 immune complexes to detect by silver stain. Large volumes of cells were cultured to perform fractionations and subsequent IPs on. The process was shown to enrich for chromatin associated monoubiquitinated FANCD2 (Figure 7). The silver stain showed sixteen detectable differential bands or regions between the IP samples which were analyzed by mass spectrometry (Figure 8). The mass spectrometry analysis revealed several candidate proteins, which have been broken down into six different groups which includes nucleosome remodeling, nuclear matrix, DNA repair, transcription regulation, chromosome maintenance, and a miscellaneous proteins group (Table 3-8). All of these tables were generated using ProteoIQ 2.2 analysis. All results were subjected to a 5% false discovery rate validation. Proteins that were detected that are found solely in the cytoplasm or extra cellular matrix of the cell were deemed as nonspecific because of the prefractionation before the IP which should have removed these interactions. Common contaminants of this method were detected and removed from candidate lists. Immune complex candidate proteins which were detected at an equal or higher rate in the sample lacking any V5 epitope were removed from the list and seen as a nonspecific interaction.   Recruits NuRD degrades p53 2 6

Table 4. Nuclear matrix proteins in FANCD2-V5 immune complexes
The nuclear matrix proteins identified largely fall into two groups relating to the nuclear envelope. The Lamin proteins were detected with high spectral counts but were also somewhat abundant in the control sample, though there is a clear increase in the number of spectra detected in the V5 sample. The other large group relates to the nuclear pore complex. Several nuclear pore proteins were identified in the screen and in past screens which has been previously published on (Boisvert 2013 interacting proteins such as PCNA in the screen the strongest candidates identified in this group are VCP and SFPQ (Howlett 2009     Degrades K48 ubiquitin linkages causes accumulation of p53 (Dayal 2009) 0 5

Validating Structural Maintenance of Chromosomes Candidates
One of the strongest groups of candidates from the mass spectrometry screen was the structural maintenance of chromosomes proteins. SMC1A, SMC2, SMC3, SMC4 and SMC6 were detected in the large scale patient cell IP (Table 7). SMC3 was also detected in the endogenous IP system ( Table 2). SMC1 and SMC3 are component of the cohesin complex along with STAG1 which was detected in the endogenous system ( Table 2). The cohesin complex sister chromatids together which allows for the identification of sister chromatids (Rudra 2013). SMC2 and SMC4 are members of the condensing complex. CAPD2 is also a member of condensin which was detected in the PD20 patient screen ( Table 7). The condensin complex is required for the proper condensation and segregation of chromosomes (Hirano 2012). The SMC5-SMC6 complex plays a role in DNA damage repair though this may be mediated through cohesin recruitment (Potts 2006). Components of these complexes were detected in the FANCD2 immune complexes by immunoblotting (Figure 9). The input in Figure 9 shows that there may be less SMC protein expressed in FANCD2 hypomorphic cells following exposure to MMC, however expression of these proteins does not apper to be largely affected by FANCD2 or MMC (Figure 10). The dynamic of the interaction between FANCD2 and these proteins require further characterization. Because these proteins are involved in genomic stability it is possible that these proteins may have some overlapping functions with maintaining genomic stability. Using MAGI to query the cancer genome atlas shows that mutations in both the cohesin (5.96%) and condensin (5.42%) networks are common in cancer.

FANCD2 may Interact with Components of the SWI/SNF Complex
Several SWI/SNF proteins were also identified in the mass spectrometry screen. SMARCA4, SMARCA5, SMARCB1, and SMARCC2 were all identified in the patient cells as members of FANCD2 immune complexes along with several other nucleosome remodeling proteins (Table 3). Among the other proteins discovered was SUPT16H which is known to interact with SWI/SNF proteins in complexes ( Table 3).
The SWI/SNF proteins are important for transcriptional regulation, determination of cellular fate, and tumor suppression (Lu 2013 (Figure 11). The relationship that SWI/SNF proteins have with FANCD2 may be more related to the canonical function of SWI/SNF proteins as chromatin remodelers and transcription regulators. It is possible that the interaction between FANCD2 and SWI/SNF proteins may be upstream of transcription or it may be involved in stabilizing FANCD2 protein levels.

Identifying FANCD2 Interacting Candidate Proteins
Using mass spectrometry to identify novel interacting partners has been a challenging and rewarding technique. There are many difficulties in working with a large low abundance protein, but the data generated from the mass spectrometry screen has shown many new interesting results. The technique was also able to detect some already known interactions such as PCNA ( The screen has identified several candidates from diverse functional groups which may indicate a larger functional role for FANCD2 and the FA-BRCA pathway.
The numerous strong candidates identified by this method will require subsequent validation and characterization. This process has been started on several proteins however the large volume of data will require a systematic and methodical approach to utilize the data generated from the screen. The unbiased nature of the screen also does not discriminate against upstream or downstream proteins and does not occlude subtle events. Upstream events can be easily assayed by looking at FACND2 monoubiquitination, chromatin localization, and nuclear foci formation, in response to interstrand crosslinking agents which are all well developed assays within the lab.
Genomic instability can be assayed through metaphase spreads to determine if a protein interaction is involved in promoting repair downstream of FANCD2. But the need for new methods may be most realized while investigating subtle effects, which may only be viewable during specific conditions, or may require more information from experiments than is currently recorded as data. As more is known about FANCD2 it is increasingly likely that newly identified interacting proteins may have a less visible effect on the cell, but that does not bar these new interactors from playing a profound or underappreciated role within the pathway.

Proteins
The mass spectrometry data show a strong likelihood that FANCD2 is involved with the structural maintenance of chromosomes proteins in some way.
SMC3 is one of the few proteins that were detected in both the endogenous and patient systems (Table2 , Table 7). Cohesin has been shown to be enriched at sites of stalled forks and the SMC proteins SMC1A, SMC3, SMC4, SMC5, and STAG2, have previously been detected as FANCD2 interacting candidates (Lossaint 2013).
Knockdown for FA proteins and SMC proteins share a genomic instability and improper segregation phenotype (Nalepa 2014, Hirano 2012. With the connection to genomic instability the cohesin and condensin complexes as well as the SMC5-SMC6 complex make excellent potential candidates for FANCD2 interactors. These proteins are also highly involved in chromatin architecture restructuring which may be a critical component to understanding how the FA-BRCA pathway becomes activated upon damage detection. While the SMC proteins have been implicated in DNA repair the mechanism by which they act to repair DNA is currently unknown.

FANCD2 may Interact with the SWI/SNF Proteins
The SWI/SNF proteins are known as both nucleosome remodelers and transcription factors (Lu 2013). As a known tumor suppressor and the large number of complex members identified (Table 3), the SWI/SNF proteins make strong candidates for evaluation. The affect that the SWI/SNF proteins have on FANCD2 protein levels may be of great importance and it may provide a mechanism by which the SWI/SNF cells act as tumor suppressors. It is important to continue to evaluate the relationship between FANCD2 and the SWI/SNF proteins. There are still many metrics for characterization to look at such as chromatin localization and FANCD2 nuclear foci formation.

FANCD2 and Histones
Another notable interaction identified by this study is the interaction between FANCD2 and histones. Several histone variants were identified as the candidates however it is difficult to determine the strength of this interaction as the peptides also showed several spectral counts in the control sample, albeit at a lower frequency (Table 3). There are several possible reasons for this perceived interaction with histones, it may be an indirect interaction between nucleosome remodeling proteins that interact with FANCD2 or it may be an artifact of FANCD2 associating with chromatin. However this does not preclude the possibility that FANCD2 may be directly interacting with histones.
The idea that FANCD2 may be interacting with histone is supported by the premise that FANCD2 acts as a histone chaperone (Sato 2012). This idea is further evidenced by the discovery of a possible histone binding domain within FANCD2 (unpublished data). With the difference in some of the histone peptides detected in this experiment exceeding a fourfold change in some cases it is reasonable to believe that chromatin associated FANCD2 may have a direct interaction with histones.
Enrichment is shown for Histone H1, Histone H2A, Histone H2B, Histone H3, and Histone H4 (Table 3). It is important to realize that DNA repair occurs within the context of chromatin and while detecting interactions with histones the abundance of posttranslational modifications and histone variants remains unknown and is not likely to be decoded with such a broad screening method.
The diversity and prevalence of histones and histone remodeling proteins within the data allude to the idea that FANCD2 may be involved in binding specific chromatin marks or in causing changes in the profile of chromatin modifications. The FA-BRCA pathway may be regulated by the histone code or alternatively may be actively involved in writing the histone code. With several of the proteins in the FA core complex being understudied it is possible that the FA core complex along with FANCD2 may have a role in recognizing and modifying structures in chromatin either is conjunction with or separate from the ability for the FA-BRCA pathway to promote repair of the highly genotoxic ICL lesions.

Conclusions
This project has demonstrated a successful method to enrich for chromatin associated active FANCD2 immune complexes, and has given large clues as to which proteins compose the activated FANCD2 immune complexes. While this data shows a dramatic increase in the information known about possible FANCD2 interacting proteins, the data still requires a great deal of investigation to be performed. The methods and data described here show the massive power associated with proteomics techniques. The discovery of several strong candidate interacting proteins as well as preliminary data to characterize the interaction should contribute to understanding the activity and regulation of FANCD2 and the larger FA-BRCA pathway function in a meaningful way. Hopefully the information generated in the screen can contribute to the improvement for healthcare options of FA patients and increase the understanding of how DNA repair is mediated within the cell.