Conceived and designed the experiments: CL HGS EM XW. Performed the experiments: EM XW PWJ VS. Analyzed the data: EM XW PWJ VS HGS CL. Contributed reagents/materials/analysis tools: EM XW PWJ VS. Wrote the paper: CL EM.
Current address: Department of Molecular Cancer Research, University Medical Center Utrecht, Utrecht, The Netherlands
The authors have declared that no competing interests exist.
Nucleosome translocation along DNA is catalyzed by eukaryotic SNF2-type ATPases. One class of SNF2-ATPases is distinguished by the presence of a C-terminal bromodomain and is conserved from yeast to man and plants. This class of SNF2 enzymes forms rather large protein complexes that are collectively called SWI/SNF complexes. They are involved in transcription and DNA repair. Two broad types of SWI/SNF complexes have been reported in the literature; PBAF and BAF. These are distinguished by the inclusion or not of polybromo and several ARID subunits. Here we investigated human SS18, a protein that is conserved in plants and animals. SS18 is a putative SWI/SNF subunit which has been implicated in the etiology of synovial sarcomas by virtue of being a target for oncogenic chromosomal translocations that underlie synovial sarcomas.
We pursued a proteomic approach whereby the SS18 open reading frame was fused to a tandem affinity purification tag and expressed in amenable human cells. The fusion permitted efficient and exclusive purification of so-called BAF-type SWI/SNF complexes which bear ARID1A/BAF250a or ARID1B/BAF250b subunits. This demonstrates that SS18 is a BAF subtype-specific SWI/SNF complex subunit. The same result was obtained when using the SS18-SSX1 oncogenic translocation product. Furthermore, SS18L1, DPF1, DPF2, DPF3, BRD9, BCL7A, BCL7B and BCL7C were identified. ‘Complex walking’ showed that they all co-purify with each other, defining human BAF-type complexes. By contrast,we demonstrate that human PHF10 is part of the PBAF complex, which harbors both ARID2/BAF200 and polybromo/BAF180 subunits, but not SS18 and nor the above BAF-specific subunits.
SWI/SNF complexes are found in most eukaryotes and in the course of evolution new SWI/SNF subunits appeared. SS18 is found in plants as well as animals. Our results suggest that in both protostome and deuterostome animals, a class of BAF-type SWI/SNF complexes will be found that harbor SS18 or its paralogs, along with ARID1, DPF and BCL7 paralogs. Those BAF complexes are proteomically distinct from the eukaryote-wide PBAF-type SWI/SNF complexes. Finally, our results suggests that the human bromodomain factors BRD7 and BRD9 associate with PBAF and BAF, respectively.
Gene expression programs determine cell identity and response to endocrine stimuli, as has been demonstrated most dramatically by the generation of induced pluripotent stem cells with the Oct4, Sox2, Klf4 and c-Myc transcription factors
The C-terminal bromo domain-bearing SNF2 enzymes are found in so-called SWI/SNF multiprotein complexes and are conserved in most eukaryotes. They are implicated in transcriptional regulation and multiple DNA repair pathways
In mice and humans, at least 20 different SWI/SNF complex subunits have been reported (
Protein | alternative names | polyA mRNA |
TAP | INI1 | SS18 | SS18SSX1 | BCL7A | BCL7C | DPF2 | BRD9 | PHF10 |
BRG1 | SMARCA4 | 5.93 | 0 | 0.303 | 5.337 | 4.440 | 2.039 | 2.290 | 1.662 | 0.141 | 0.984 |
BRM | SMARCA2 | 1.45 | 0 | 0 | 1.339 | 0.730 | 0.116 | 0.280 | 0.179 | 0 | 0.028 |
BAF250A | SMARCF1, ARID1A | 1.63 | 0 | 0.027 | 5.692 | 2.079 | 0.379 | 1.766 | 1.485 | 0.027 | 0.027 |
BAF250B | ARID1B, OSA1 | 2.04 | 0 | 0 | 2.728 | 1.540 | 0.179 | 0.638 | 0.315 | 0.028 | 0 |
BAF200 | ARID2, zipzap | nd | 0 | 0.122 | 0 | 0.029 | 0.029 | 0.122 | 0 | 0 | 0.884 |
BAF180 | Polybromo-1 | 4.36 | 0 | 0.457 | 0 | 0 | 0 | 0.248 | 0 | 0 | 1.769 |
BAF170 | SMARCC2 | 9.25 | 0 | 1.532 | 2.793 | 2.360 | 0.438 | 1.432 | 1.154 | 0.084 | 1.745 |
BAF155 | SMARCC1 | 6.94 | 0 | 2.981 | 5.813 | 3.467 | 1.239 | 2.043 | 2.831 | 0.080 | 0.468 |
BAF60A | SMARCD1 | 4.01 | 0 | 0.15 | 3.037 | 3.037 | 0.784 | 2.054 | 1.477 | 0.072 | 0.630 |
BAF60B | SMARCD2 | 7.69 | 0 | 0.719 | 7.161 | 4.080 | 1.762 | 2.875 | 2.384 | 0 | 0.607 |
BAF60C | SMARCD3 | 3.02 | 0 | 0 | 1.102 | 0.346 | 0 | 0.346 | 0.16 | 0 | 0 |
BAF57 | SMARCE1 | 9.97 | 0 | 0.957 | 11.115 | 5.190 | 1.154 | 3.642 | 3.217 | 0.957 | 1.61 |
BAF53A | ACTL6A, ArpNß | 8.37 | 0.233 | 0.110 | 6.305 | 6.305 | 2.511 | 2.511 | 0.874 | 0.369 | 0.52 |
BAF53B | ACTL6B, ArpNα | 1.68 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
BAF47 | SMARCB1, INI1, SNF5 | 5.81 | 0 |
|
5.449 | 3.160 | 1.154 | 3.160 | 0.551 | 0.245 | 0.823 |
BAF45A | PHF10 | 3.08 | 0 | 0.086 | 0 | 0 | 0 | 0 | 0 | 0 |
|
BAF45B | DPF1 | 0.50 | 0 | 0 | 0 | 0.110 | 0 | 0.110 | 0 | 0 | 0 |
BAF45C | DPF3, CERD4 | 0.26 | 0 | 0 | 0 | 0 | 0 | 0.105 | 0 | 0 | 0 |
BAF45D | DPF2, REQ, UBID4 | 2.80 | 0 | 0 | 4.623 | 1.610 | 0.101 | 1.371 |
|
0 | 0 |
SS18 | SYT, SSXT | 3.73 | 0 | 0 |
|
|
0.78 | 0 | 2.16 | 1.000 | 0 |
SS18L1 | CREST | 9.92 | 0 | 0 | 0 | 0 | 0 | 0.585 | 0.585 | 0 | 0 |
BCL7A | - | nd | 0 | 0 | 0.874 | 1.310 |
|
0 | 0.585 | 0.233 | 0 |
BCL7B | Hom s 3 | 4.77 | 0 | 0 | 0 | 0.292 | 0 | 0 | 0 | 0 | 0 |
BCL7C | - | 4.18 | 0 | 0 | 0 | 1.783 | 0 |
|
0 | 0 | 0 |
BRD7 | CELTIX-1 | nd | 0 | 0.064 | 0 | 0 | 0 | 0 | 0 | 0 | 0.645 |
BRD9 | MU-RMS-40.8 | 0.18 | 0 | 0 | 0.186 | 0.668 | 0 | 0.089 | 0 |
|
0 |
SSX1 | - | nd | 0 | 0 | 0 |
|
0 | 0 | 0 | 0 | 0 |
GLTSCR1 | GSCR1 | 1.20 | 0 | 0 | 0.619 | 1.116 | 0.055 | 0.708 | 0 | 0 | 0 |
SRRM2 | - | 8.15 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
MYBBP1A | p160 | 3.72 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
NONO | NMT55, p54(nrb) | nd | 0.066 | 0 | 0 | 0.066 | 0 | 0 | 0 | 1.966 | 0.292 |
NUMA1 | - | 3.23 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0.167 | 0 |
SFPQ | PSF | 11.42 | 0.199 | 0 | 0 | 0 | 0 | 0 | 0 | 0.624 | 0.528 |
DDX3X | HLP2 | 12.18 | 0.116 | 0 | 0 | 0 | 0 | 0 | 0 | 10.159 | 0.315 |
DDX17 | p72 | 11.85 | 0.058 | 0 | 0 | 0 | 0 | 0 | 0.058 | 9.578 | 0.136 |
RBM14 | COAA | 7.10 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 4.736 | 0 |
DDX5 | p68 | 17.37 | 0 | 0 | 0 | 0 | 0 | 0 | 0.061 | 4.223 | 0.805 |
actin | actg1 | 30.37 | 4.109 | 1.371 | 4.109 | 6.499 | 2.481 | 2.831 | 0.778 | 7.254 | 3.437 |
Protein abundance is represented by the exponentially modified Protein Abundance Index
mRNA abundance was estimated from probe set fluorescence signal intensities, as recommended by Affymetrix (see
Strikingly, multiple SWI/SNF subunits function as tumor suppressors in man and mouse, adding a key medical dimension to SWI/SNF research
Another link to cancer is provided by the SS18-SSX oncofusion proteins
We found SS18 to be present in BAF-class human SWI/SNF chromatin remodeling complexes. Purification of SS18-SSX1 revealed that this oncofusion protein resides in the same complexes. Interestingly, we detected several additional putative SWI/SNF interactors
In order to define the protein complexes harboring known and suspected human SWI/SNF subunits we generated stable human embryonic kidney cell (Hek293) clones transduced with retroviral TAP-tag fusion expression constructs
(A) Silver-stained gels of the actual purified protein preparations that were analyzed by mass spectrometry. The respective TAP-tag fusion proteins are designated by black triangles. Size markers (Da) are indicated for every gel. Banding patterns differ because the gels were not all run under the same conditions. (B) Osprey interaction network
INI1 is a core subunit of SWI/SNF complexes that is also known as hSNF5, SMARCB1 or BAF47. In our hands the yield of SWI/SNF complexes obtained with INI1TAP has consistently been comparatively low. For instance, in most INI1TAP preparations we detect BRG1 but not BRM, and ARID1A but not ARID1B (
In keeping with a role as a core SWI/SNF subunit, INI1TAP purifications harbored both PBAF and BAF-specific SWI/SNF subunits (
Human Protein | Alternative human names | Human complex |
|
Fly complex |
|
|
BRG1 | SMARCA4 | Core | brahma/CG5942 | Core | SNF2 | STH1 |
BRM | SMARCA2 | Core | brahma/CG5942 | SNF2 | STH1 | |
BAF250A | SMARCF1, ARID1A | BAF | OSA/eyelid/CG7467 | BAP | SWI1 | - |
BAF250B | ARID1B, OSA1 | BAF | OSA/eyelid/CG7467 | SWI1 | - | |
BAF200 | ARID2, zipzap | PBAF | BAP170/CG3274 | PBAP | - | - |
BAF180 | Polybromo-1 | PBAF | polybromo/BAP180/CG11375 | PBAP | - | RSC1, RSC2, RSC4 |
BAF170 | SMARCC2 | Core | moira/BAP155/CG18740 | Core | SWI3 | RSC8 |
BAF155 | SMARCC1 | Core | moira/BAP155/CG18740 | SWI3 | RSC8 | |
BAF60A | SMARCD1 | Core | BAP60/CG4303 | Core | SWP73 | RSC6 |
BAF60B | SMARCD2 | Core | BAP60/CG4303 | SWP73 | RSC6 | |
BAF60C | SMARCD3 | Core | BAP60/CG4303 | SWP73 | RSC6 | |
BAF57 | SMARCE1 | Core | dalao/BAP111/CG7055 | Core | - | - |
BAF53A | ACTL6A, ArpNb | Core | BAP55/CG6546 | Core | ARP7 & ARP9 | ARP7 & ARP9 |
BAF53B | ACTL6B, ArpNa | Core | ? | ARP7 & ARP9 | ARP7 & ARP9 | |
BAF47 | SMARCB1, INI1, SNF5 | Core | SNR1/CG1064 | Core | SNF5 | SFH1 |
BAF45A | PHF10 | PBAF | e(y)3/SAYP/CG12238 | PBAP | - | - |
BAF45B | DPF1 | BAF | d4/CG2682 | ? | - | - |
BAF45C | DPF3, CERD4 | BAF | d4/CG2682 | - | - | |
BAF45D | DPF2, REQ, UBID4 | BAF | d4/CG2682 | - | - | |
SS18 | SYT, SSXT | BAF | CG10555 | ? | - | - |
SS18L1 | CREST | BAF | CG10555 | - | - | |
BCL7A | - | BAF | BCL7-like/CG17252 | ? | - | - |
BCL7B | Hom s 3 | BAF | BCL7-like/CG17252 | - | - | |
BCL7C | - | BAF | BCL7-like/CG17252 | - | - | |
BRD7 | CELTIX-1 | PBAF | CG7154 | ? | - | - |
BRD9 | MU-RMS-40.8 | BAF | CG7154 | - | - | |
actin | actg1 | actin | actin | actin | ||
RTT102 | RTT102 | |||||
SWP82 | NPL6 | |||||
HTL1 | ||||||
LDB7 | ||||||
RSC3 | ||||||
RSC30 | ||||||
RSC58 | ||||||
RSC9 | ||||||
SNF6 |
SS18TAP purifications yielded high levels of SWI/SNF (
Because the chromosomal translocation t(X;18)(p11.2;q11.2) results in production of the oncogenic SS18-SSX1 protein fusion it was of interest to compare the proteomic environments of SS18 and the SS18 oncofusions. Essentially, purification of SS18-SSX1TAP resulted in the same set of interactors as purification of SS18TAP, with the exception of peptides originating from the SSX1 moiety of the oncofusion protein (
DPF2, also known as ubi-d4 or Requiem, is ubiquitously expressed and implicated in apoptosis
(A) Domain organization of the human proteins DPF1,-2,-3; PHF10; SS18 and its paralog SS18l1/crest; and BCL7A,-B,-C. We note that while CG2682, the
The DPF2TAP purification results indicate that DPF2 resides mainly in ARID1-bearing BAF complexes, since no polybromo or ARID2 peptides were identified, whilst high confidence ARID1A and ARID1B peptides were detected (
PHF10 harbors two PHD domains but it is not a member of the DPF paralog group as it lacks a central Krüppel zinc finger motif, and harbors a SAY domain that is conserved in animals but not plants
Notably, high confidence BRD7 peptides were detected, similar to the INI1TAP purification (
Similar to multiple SWI/SNF subunits, BCL7 family members have been implicated in carcinogenesis
A well established function of bromodomains is to recognize specific acetylated lysines. The paralogous catalytic subunits of SWI/SNF, BRG1 and BRM harbor one C-terminal bromodomain that is closely related to the six bromodomains of polybromo, but quite distinct from the bromodomains of BRD7 and BRD9
A FLAG-BRD7 fusion has been reported to purify PBAF complexes
We quantified our mass spectrometry data on the basis of the exponentially modified protein abundance index (emPAI,
Crabtree and colleagues
Of the other putative novel BAF-associated proteins, we could detect NONO and its binding partner SFPQ
Paralogous human SWI/SNF subunits are known to be expressed in tissue and signal specific fashion, generating alternative SWI/SNF complex configurations that can cooperate with transcription factor networks to coordinate cell proliferation and differentiation. Here, we focus on SWI/SNF subunits that are absent from yeast but conserved in animals and plants (SS18) or only in animals (DPF, BCL7 and PHF10) (
Essentially there are two types of human SWI/SNF complexes
Our mass spectrometry analysis of affinity tag-mediated protein complex purifications confirms bipartition of SWI/SNF complexes in BAF and PBAF-class complexes. We demonstrate here that the paralogous cancer-related minor SWI/SNF subunits DPF1, -2, -3; BCL7A, -B, -C; and SS18 and SS18L1 reside in BAF-class human SWI/SNF complexes and, that PHF10 marks PBAF SWI/SNF complexes. Moreover, because quantitative analysis indicates that the chief interaction partners of PHF10, DPF2, SS18, SS18-SSX1, BCL7A and BCL7C are the other SWI/SNF subunits, we speculate that they exert their molecular action through their respective SWI/SNF complexes. It remains to be seen indeed to what extent our results, which were obtained in one human cell line, can be extrapolated to other cell types and even other organisms. Considering the congruence between our data and a recent studies on
Since we could not detect notable differences between SS18 and its oncogenic fusion products at the proteomic level, the oncogenic activity of the SS18-SSX fusions may have to be sought either at the level of SWI/SNF (dis)assembly dynamics, post-translational modifications or an affinity for specific genomic loci
Interestingly, we detected the bromodomain proteins BRD7 and BRD9 in our SWI/SNF preparations. The fly protein CG7154 is that organism's sole BRD7/BRD9 ortholog. It will be interesting to determine whether it associates with the fly brahma SNF2 ATPase, and if so, whether it is specific to the BAP or PBAP fly equivalents of BAF and PBAF. In humans, BRD7 appears to promote cellular senescence
Altogether, this work demonstrates that paralogs of SS18, BCL7 and DPF factors, which can be found in both protostome
Tandem Affinity Purification (TAP) constructs were generated by PCR using the oligomers indicated in parentheses and cloned into the XhoI and EcoRI restriction sites in the retroviral expression vector pZXN, whereby the TAP-tag sequence was fused to the coding sequences at their N-terminus
Human Embryonic Kidney (Hek293, ATCC CRL-1573) and phoenix cells were grown in Dulbecco's modified Eagles medium (Invitrogen) supplemented with 10% FCS, penicillin 100 µg/ml and streptomycin 100 U/ml (Invitrogen) at 37°C in 5% CO2. Retroviral stable cell lines were generated as previously described
Tandem Affinity Purification was performed as previously described in detail
Co-immunoprecipitations were performed on TEV cleavage eluates obtained as described above, using antibodies directed against BRM (Abcam 15597), DPF2 (Aviva systems biology ARP33221_P050) or BRD9 (Aviva systems biology ARP34803_T200), under the same conditions as the anti-MYC or TY1 immunoprecipitations in the TAPtag purification protocol. The immunoprecipitated proteins were separated by SDS-PAGE, transferred onto nylon filters and probed with anti-MYC antibodies, which recognize the transduced SS18-SSX1 (
The silver stained gel lanes were cut into small pieces. After reduction and alkylation the proteins were trypsin (Promega) digested and extracted from the gel using trifluoroacetic acid (TFA). Peptides were sequenced using a nano-high-pressure liquid chromatography Agilent 1100 nanoflow system connected online to a 7-Tesla linear quadrupole ion-trap Fourier transform (FT) mass spectrometer (Thermo Electron, Bremen, Germany) essentially as described previously
Expression profiling was performed on four Hek293 polyA mRNA samples by microarray analysis using Affymetrix human exon array 1.0 ST according to manufacturer instructions (
Mass spectrometry results, including; accession numbers, short protein descriptions, peptide sequences, associated Mascot score, peptide delta score and absolute calibrated mass relative error.
(XLS)
Quadruplate polyA mRNA expression profile of Hek293 cells determined with the Affymetrix human exon array 1.0 ST platform.
(XLS)
We are thankful to Michiel Vermeulen for communicating unpublished data early on in this project and to the members of the molecular biology department for their generous help and support.