A Versatile Nanotrap for Biochemical and …

1 Supplemental Material can be found at: /DC1. Research A Versatile Nanotrap for Biochemical and functional studies with fluorescent fusion Proteins* S. Ulrich Rothbauer , Kourosh Zolghadr , Serge Muyldermans , Aloys Schepers , M. Cristina Cardoso , and Heinrich Leonhardt **. Green fluorescent proteins (GFPs) and variants thereof ants thereof became popular tools to determine protein are widely used to study protein localization and dynam- localization and, in combination with fluorescence photo- Downloaded from at MPIs MARTINSRIED BIBLIOTHEK on February 7, 2008. ics. We engineered a specific binder for fluorescent pro- bleaching techniques, provided unique information on protein teins based on a 13-kDa GFP binding fragment derived dynamics in living cells (1 4). Necessary additional informa- from a llama single chain antibody. This GFP-binding pro- tion on DNA binding, enzymatic activity, and complex forma- tein (GBP) can easily be produced in bacteria and coupled tion can be obtained with various methods including chroma- to a monovalent matrix.

2 The GBP allows a fast and effi- cient (one-step) isolation of GFP fusion proteins and their tin immunoprecipitation (ChIP) and affinity purification (5, 6). interacting factors for Biochemical analyses including These methods, however, are hampered by the limited avail- mass spectroscopy and enzyme activity measurements. ability of specific antibodies. Those limitations are often by- Moreover GBP is also suitable for chromatin immunopre- passed by fusing the protein of interest to specific epitope or cipitations from cells expressing fluorescent DNA-binding protein tags including hemagglutinin, c-Myc, FLAG, or GST (7, proteins. Most importantly, GBP can be fused with cellular 8). Curiously GFP, the most widely used labeling tag in cell proteins to ectopically recruit GFP fusion proteins allow- biology, is rarely used for Biochemical analyses, although ing targeted manipulation of cellular structures and pro- cesses in living cells. Because of the high affinity capture various mono- and polyclonal antibodies have been de- of GFP fusion proteins in vitro and in vivo and a size in the scribed (9, 10).

3 This may be due in part to limited availability lower nanometer range we refer to the immobilized and specificity as well as co-eluted heavy and light antibody GFP-binding protein as GFP- Nanotrap . This Versatile chains that interfere with subsequent analyses. An alternative GFP- Nanotrap enables a unique combination of micro- to conventional antibodies are variable single domain anti- scopic, Biochemical , and functional analyses with one body fragments, also referred to as VHH, derived from heavy and the same protein. Molecular & Cellular Proteomics chain antibodies of Camelidae (11). These VHHs, which pres- 7:282 289, 2008. ent the smallest intact antigen-binding units with a molecular mass of about 15 kDa are highly soluble and stable and can be efficiently produced in heterologous systems (12, 13). VHH. After the identification of most components of the cell, fragments have been used like conventional antibodies for further insights into their regulation and function require infor- various immunological applications (14 16).

4 Mation on their abundance, localization, and dynamic inter- actions. Green fluorescent proteins (GFPs)1 and spectral vari- Here we describe a novel application of a 13-kDa GFP. binding fragment derived from a llama single chain antibody (17). This GFP-binding protein (GBP) has a small ( From the Munich Center for Integrated Protein Science, (CiPSM). and Department of Biology, Ludwig Maximilians University Munich, nm) barrel-shaped structure and can easily be produced in 82152 Planegg-Martinsried, Germany, Department of Molecular and bacteria. We immobilized the GBP to generate a GFP-nano- Cellular Interactions, VIB and Laboratory of Cellular and Molecular trap, which enables a fast and efficient isolation of GFP fusion Immunology, Vrije Universiteit Brussel, 1050 Brussels, Belgium, De- proteins and their interacting factors for Biochemical and ChIP. partments of Gene Vectors, GSF-National Research Center for Envi- ronment and Health, 81377 Munich, Germany, and Max Delbrueck analyses.

5 Moreover we demonstrated that the GFP-binding Center for Molecular Medicine, 13125 Berlin, Germany protein can be fused with structural proteins to ectopically Received, July 26, 2007, and in revised form, October 15, 2007 recruit GFP fusion proteins and interacting factors at defined Published, MCP Papers in Press, October 21, 2007, DOI regions in living cells. 1. The abbreviations used are: GFP, green fluorescent protein; GBP, GFP-binding protein; ChIP, chromatin immunoprecipitation; VHH, variable domain of heavy chain antibody; HEK, human embryonic protein; PCNA, proliferating cell nuclear antigen; H2B, histone H2B;. kidney; IgG, immunoglobulin G; YFP, enhanced yellow fluorescent Dnmt1, DNA methyltransferase I; PBD, PCNA binding domain;. protein; CFP, enhanced cyan fluorescent protein; DsRed, Discosoma HMGA1a, high mobility group protein A1a; Igf, insulin-like growth genus red fluorescent protein; mRFP, monomeric red fluorescent factor; PML, promyelocytic leukemia protein.

6 282 Molecular & Cellular Proteomics 2008 by The American Society for Biochemistry and Molecular Biology, Inc. This paper is available on line at A GFP-binding Protein for Biochemical and functional studies EXPERIMENTAL PROCEDURES M NaCl, 1 g of DNase I, 5 mM MgCl2, and SDS. After a centrifugation step (10 min at 20,000 g at 4 C) the supernatant was Expression and Purification of the GBP Llama immunization, VHH. adjusted with dilution buffer (20 mM Tris/HCl, pH , 150 mM NaCl, library construction, and selection of the GBP were described previ- mM EDTA, 2 mM PMSF) to 1 ml. 20 l (2%) were added to ously (17). The coding sequence of the GFP-binding VHH domain was SDS-containing sample buffer (referred to as input). 1 g of purified cloned into the pHEN6 (18) vector using the NcoI and NotI restriction GBP or 2 g of anti-GFP antibodies were added and incubated for sites adding a C-terminal histidine (His6) tag, and chemically compe- 5 60 min on an end-over-end rotor at 4 C.

7 For pulldown of immu- tent Escherichia coli BL21 cells were transformed. For expression and purification a 500-ml E. coli culture was induced with 1 mM isopropyl nocomplexes 25 l of an equilibrated mixture of protein A/G-Sepha- -D-1-thiogalactopyranoside for 20 h at room temperature. Bacterial rose (Amersham Biosciences) were added, and incubation continued cells were harvested by centrifugation (10 min at 5000 g), and the for 60 min. After a centrifugation step (2 min at 5000 g at 4 C). pellet was resuspended in 10 ml of binding buffer (1 PBS, pH , supernatant was removed, and 2% was used for SDS-PAGE (referred M NaCl, 20 mM imidazole, 1 mM PMSF, 10 g/ l lysozyme). The to as flow-through). The bead pellet was washed two times in 1 ml of cell suspension was incubated for 1 h at 4 C in a rotary shaker and dilution buffer containing 300 mM NaCl. After the last washing step the then sonified (6 10-s pulse) on ice. After centrifugation (20 min at beads were resuspended in 2 SDS-containing sample buffer and 20,000 g) soluble proteins were loaded on a preequilibrated 1-ml boiled for 10 min at 95 C.

8 Downloaded from at MPIs MARTINSRIED BIBLIOTHEK on February 7, 2008. HiTrap column (GE Healthcare) and purified. The His-tagged GBP Column-based GFP Purification 1 mg of purified GBP was co- was eluted by a linear gradient ranging from 20 to 500 mM imidazole. valently coupled to 1 ml of N-hydroxysuccinimide-Sepharose (GE. Elution fractions containing the GBP were pooled and dialyzed into Healthcare), according to the manufacturer's instructions, generating PBS. Protein concentration was adjusted to 1 g/ l. The yield of the GFP- Nanotrap . Subsequently 50 l of the GFP- Nanotrap was purified GBP/liter of bacterial culture was in the range of 10 15 mg. transferred to a 1-ml column (MoBiTec) and preequilibrated with 5 ml Expression Plasmids For bacterial expression and purification of of dilution buffer. Protein extracts of GFP-producing cells were pre- GFP we used the bacterial expression plasmid pRSet5D containing pared as described above, and the soluble protein fraction was the GFP coding sequence with a C-terminal His6 tag kindly provided loaded onto the column.

9 The protein solution (1 ml) passed through by B. Steipe. For mammalian cells we used the following expression the column at 100 200 l/min flow rate. Subsequently the column vectors encoding different fluorescent fusion proteins: pEGFP-C1, material was washed (two times) with 1 ml of dilution buffer containing pEYFP-C1, pECFP-C1, GFP- -actin (all from Clontech), GFP-Lamin 300 mM NaCl, and bound proteins were eluted with 100 l M. B1 (19), GFP-PCNA (20), mRFP-PCNA, GFP-Dnmt1, and GFP- glycine, pH 2% of the input and flow-through and 10% of bound Dnmt1(P1229W) (21). We also used mammalian expression vectors material were resuspended in 2 SDS-containing sample buffer and containing cDNAs coding for mRFP1 (22), mCherry, or mOrange (23) analyzed by SDS-PAGE and Coomassie Blue and by immunostaining. kindly provided by R. Tsien. Overlap extension PCR was performed to Enzymatic Activity Test Soluble protein extracts of HEK 293T. construct the CFP I147N and the CFP T154M using mutagenic prim- cells producing GFP-Dnmt1 or GFP-Dnmt1(P1229W) were prepared, ers CFP-I147N-FW (5 -GAGTACAACTACAACAGCCACAACGTCTA- and immunoprecipitation was performed as described above.)

10 Beads TA-3 ) in combination with CFP-I147N-Rev (5 -AACGTCTATATCAT- containing the GBP-GFP-Dnmt1 complexes were washed extensively GGCCGACAAGCAG-3 ) and CFP-T154M-FW (5 -AACGTCTATATC- in dilution buffer containing 300 mM NaCl, and after centrifugation the ATGGCCGACAAGCAG-3 ) in combination with CFP-T154M-Rev (5 - beads were resuspended in 500 l of assay buffer (100 mM KCl, 10. CTGCTTGTCGGCCATGATATAGACGTT-3 ), respectively, and the fl- mM Tris, pH , 1 mM EDTA, 1 mM DTT). Beads were washed (two anking primer CFP FW (5 -GATCCGCTAGCGCTACCGGTCGCCAC- times) in assay buffer, and after centrifugation (1000 g for 1 min) 30. CATG-3 ) and CFP-Rev (5 -AACGTCTATATCATGGCCGACAAGCA- l of methylation mixture ( Ci of S-[3H]adenosylmethionine (Am- G-3 ). The PCR fragment was cloned into the NheI/BsrGI site of the ersham Biosciences), pmol/ l hemimethylated double-stranded pECFP-C1 vector. For GBP-Lamin B1, the Lamin B1 was amplified 35-bp DNA (50 pmol/ l), 160 ng/ l BSA) were added.