Efficient biotinylation and single-step purification of tagged transcription factors in mammalian cells and transgenic mice Ernie de Boer*, Patrick Rodriguez*, Edgar Bonte*, Jeroen Krijgsveld†, Eleni Katsantoni*, Albert Heck†, Frank Grosveld*, and John Strouboulis*‡ *Department of Cell Biology, Erasmus Medical Center, P.O. Box 1738, 3000 DR Rotterdam, The Netherlands; and †Department of Biomolecular Mass Spectrometry, Bijvoet Center for Biomolecular Research and Utrecht Institute for Pharmaceutical Sciences, Utrecht University, Sorbonnelaan 16, 3584 CA Utrecht, The Netherlands Communicated by Gary Felsenfeld, National Institutes of Health, Bethesda, MD, May 1, 2003 (received for review February 26, 2003) Proteomic approaches require simple and efficient protein purifi- tinylated proteins has led to the development of sequence tags that cation methodologies that are amenable to high throughput. can be biotinylated in bacterial, yeast, insect, and mammalian cells Biotinylation is an attractive approach for protein complex purifi- (7–11). Biotinylation can occur either by the cell’s endogenous cation due to the very high affinity of avidin͞streptavidin for protein–biotin ligases or through the coexpression of an exogenous biotinylated templates. Here, we describe an approach for the biotin ligase, in most cases that of the bacterial BirA enzyme. These single-step purification of transcription factor complex(es) based tags, however, are large in size (at least 63 aa) and may thus affect on specific in vivo biotinylation. We expressed the bacterial BirA the structure of the proteins they are fused to. In addition, that these biotin ligase in mammalian cells and demonstrated very efficient tags can be recognized by endogenous enzymes excludes applica- biotinylation of a hematopoietic transcription factor bearing a tions where biotinylation of the tagged protein may need to be small (23-aa) artificial peptide tag. Biotinylation of the tagged regulated. Furthermore, biotinylation using these tags is not very transcription factor altered neither the factor’s protein interactions efficient, particularly in mammalian cells (10, 11). Another ap- or DNA binding properties in vivo nor its subnuclear distribution. proach, so far only demonstrated in bacteria (12–15), utilizes small Using this approach, we isolated the biotin-tagged transcription (Ͻ23-aa) artificial tags that have been selected through multiple factor and at least one other known interacting protein from crude rounds of screening combinatorial peptide libraries for specific nuclear extracts by direct binding to streptavidin beads. Finally, biotinylation by BirA biotin ligase (16). These tags have been shown this method works efficiently in transgenic mice, thus raising the to be biotinylated in vitro with kinetics comparable to those of prospect of using biotinylation tagging in protein complex purifi- natural biotin acceptor sequences (17) and may thus serve as cation directly from animal tissues. Therefore, BirA-mediated bio- excellent substrates for efficient biotinylation in cells by coexpressed tinylation of tagged proteins provides the basis for the single-step biotin ligases. purification of proteins from mammalian cells. Simple generic affinity purification methodologies have been increasingly applied for the purposes of large-scale proteomic n the postgenome-sequencing era, focus has shifted toward the studies, particularly in yeast (18, 19). However, these often use Iidentification and characterization of the protein complement of reagents of variable affinities (e.g., antibodies) that increase back- cells, the proteome. A crucial aspect of this effort is the utilization ground and͞or intermediate steps (e.g., prepurification or affinity of simple and efficient methodologies that are amenable to high- tag removal) that restrict their simple application in more complex throughput approaches for the purification of proteins and protein protein sources such as mammalian cells. Considering the advan- complexes (1, 2). As a result, a number of generic affinity-based tages of biotinylation, we explored the feasibility of using it for the methodologies have been developed for these purposes, based simple high-affinity one-step purification of tagged proteins from primarily on the use of specific antibodies or affinity tags that are mammalian cells. To these ends, we coexpressed bacterial BirA fused to the protein of interest (3, 4). biotin ligase and hematopoietic transcription factors tagged by an Prominent among the affinity-based purification methodologies N-terminal fusion of a small artificial peptide previously shown to ͞ is the biotin avidin system. Biotin is a naturally occurring cofactor be biotinylated by BirA (15–17). We show that tagged proteins can for metabolic enzymes, which is active only when covalently at- be very efficiently and specifically biotinylated in mammalian cells tached to the enzymes through the action of specific protein–biotin and transgenic mice and can be efficiently purified in a single step ligases (5). Any biotinylated substrate can be bound very tightly by by binding to streptavidin beads. the proteins avidin and streptavidin. Biotin͞avidin binding is the ϭ Ϫ15 strongest noncovalent interaction known in nature (Kd 10 M), Experimental Procedures several orders higher than that of commonly used antibodies or Constructs. The coding region of the Escherichia coli birA biotin– ͞ other affinity tags. As a result, the biotin avidin affinity system has protein ligase gene (20) was cloned from genomic DNA by PCR as numerous applications in modern biological techniques (6). For the an Ϸ1-kb fragment by using Deep-VENT DNA polymerase (New purposes of protein purification, in particular, biotinylation offers a England Biolabs) and verified by sequencing. The birA gene was number of advantages. For example, the high affinity of biotin for recloned into the BglII site of the erythroid expression vector ͞ avidin streptavidin allows purification of the biotinylated protein pEV-puromycin, which consists of the human ␤-globin Locus under high stringency conditions, thus reducing background bind- Control Region (miniLCR), the human ␤-globin promoter, and the ing often observed with other affinity tags that elute more easily. In ␤-globin second intron (21). GATA-1 cDNA cloned in pEV- addition, there are very few naturally biotinylated proteins, thus Neomycin was N-terminally tagged by introducing into the NcoI site reducing the chance for crossreaction when using biotinylation in at the start codon, an oligonucleotide linker with NcoI overhangs protein purification, as opposed to antibodies that may crossreact with several species. The potential advantages of biotinylation tagging in protein Abbreviations: MEL, mouse erythroleukemic; HRP, horseradish peroxidase; EKLF, erythroid purification have not gone unnoticed (7). The characterization of Kru¨ppel-like factor. the minimal amino acid sequence requirements of naturally bio- ‡To whom correspondence should be addressed. E-mail: [email protected]. 7480–7485 ͉ PNAS ͉ June 24, 2003 ͉ vol. 100 ͉ no. 13 www.pnas.org͞cgi͞doi͞10.1073͞pnas.1332608100 Downloaded by guest on September 30, 2021 coding for the 23-aa biotinylation tag (16). Tagged erythroid U.K.) operating in positive ion mode. A nanoLC system was Kru¨ppel-like factor (EKLF) cDNA was constructed in pBluescript coupled to the Q-Tof essentially as described (23). Peptide mixtures by cloning sequentially into the NcoI site start codon firstly an were delivered to the system by using a Famos autosampler (LC oligonucleotide coding for three copies of the hemagglutinin tag, Packings, Amsterdam) at 3 ␮l͞min and trapped on an Aqua C18RP followed by an oligonucleotide coding for the 23-aa biotinylation column [Phenomenex (Belmont, CA); column dimensions 1 cm ϫ tag. Tagged EKLF was then subcloned into pEV-Neomycin. 100 ␮m ID, packed in-house]. After flow splitting down to 150–200 nl͞min, peptides were transferred to the analytical column (Pep- Mouse Erythroleukemic (MEL) Cell Transfections. MEL cells (22) were Map, LC Packings; column dimensions 25 cm ϫ 50 ␮m ID, packed initially electroporated with linearized BirA͞pEV, and stable in-house) in a gradient of acetonitrile (1% per min). Fragmentation clones were selected under puromycin. Clones were screened for of eluting peptides was performed in data-dependent mode, and BirA RNA expression by Northern blot analysis. A selected BirA͞ mass spectra were acquired in full-scan mode. Database searches pEV MEL clone was transfected with tagged GATA-1͞pEV, and were performed by using MASCOT (www.matrixscience.com). stable clones were double selected for puromycin and neomycin (Invitrogen). Chromatin Pull-Down Assays. MEL cells were treated with 1% formaldehyde in 40 mM Hepes, pH 7.9, at room temperature for 10 Nuclear Extract Preparation. MEL cells cultured in DMEM supple- min, followed by quenching with 0.125 M glycine. Crosslinked mented with 10% FCS at 37°C were induced to differentiate into chromatin was fragmented by sonication [10 ϫ 30-s bursts at mature erythroblasts with 2% DMSO for at least 3 days (22). Cells amplitude 5, followed by 10 ϫ 30-s bursts at amplitude 8 by using were harvested by centrifugation at 640 ϫ g and washed once with a Sanyo Soniprep (Loughborough, U.K.) 150 sonicator] and ali- cold PBS. The cell pellet was resuspended in 2.2 M sucrose in 10 quots corresponding to 4–10 OD260 units were snap-frozen. Pull- mM Hepes, pH 7.9͞25 mM KCl͞0.15 mM Spermine͞0.5 mM downs were carried out by incubating