2. Experimental Section

2.1. Membrane Synthesis
Cellulose-bound peptide arrays were prepared according to standard protocols using a MultiPep-RSi-Spotter (INTAVIS, Köln, Germany) employing Fmoc solid phase peptide synthesis according to the SPOT synthesis method and the manufacturer’s instructions. Peptides were synthesized either on amino-functionalized cellulose membranes (Whatman™ Chromatography paper Grade 1CHR, GE Healthcare Life Sciences #3001-878, Little Chalfont, UK) which were prepared by modifying cellulose paper by introducing Fmoc-β-Alanine as the first spacer residue [21], or on already-derivatized commercially available membranes (Amino-PEG500-UC540 Sheets optimized for use with the MultiPep instruments, INTAVIS). Cellulose-bound peptides were grown on the membranes from their C-terminus by using 0.6 M solutions of Fmoc-amino acid-OPfp (GL Biochem Ltd., Shanghai, China) in N-methyl-2-pyrrolidone (NMP, Sigma-Aldrich, Gillingham, UK, #494496) activated with 1-hydroxibenzotriazol monohydrate (HOBt, AGTC Bioproducts, Hessle, UK, #AGHOBT-H2O) and N-N′-diisopropylcarbodiimide (DIC, Fluka distributed by Sigma-Aldrich, #38370) and spotted on the membrane in 100 nL aliquots per spot using a robotic syringe. Residual amino functions were capped by acetylation between amino-acid block depositions with a 5% solution of acetic anhydride (Fisher Scientific, Loughborough, UK #A/0480/PB08) in NMP. Membranes were then washed several times with ethanol (EtOH) and NMP, before Fmoc groups were cleaved with 20% piperidine in N,N-dimethylformamide (DMF/Piperidine (80/20), AGTC, #AGDMFPIP). These steps were iteratively repeated for each amino acid within each peptide sequence. Following the last coupling step, the acid-labile protection groups of the amino acid side chains were cleaved by using a mixture of 95% trifluoro-acetic acid (TFA, Sigma-Aldrich, #T6508), 3% tri-isopropyl-silane (ACROS Organics distributed by Fisher Scientific, #214920100) and 2% water for 2 h. Membranes were then washed 4 × 30 s with dichloromethane (CH2Cl2, ATGC Bioproducts, Hessle, UK, #AGBC7002), 4 × 2 min with NMP and 2 × 2 min with 100% EtOH (absolute ethanol, Sigma-Aldrich, #32221) and were left to dry overnight. In order to ensure that each membrane had been synthesized properly, ultraviolet light (UV, λ = 280 nM) was used to ensure that all SPOTed peptides were present.

2.2. Choice of Membrane Material
Cellulose is the most common material used to prepare peptide arrays as initially described in the 90s [4]. In order to avoid “ring spot” (or “corona”) effects, where binding occurs mainly on the rim of the SPOT and not in its center, it has been suggested to reduce peptide density starting from about 10 nmol/cm2 and optimizing further [25]. Typically, amino-PEG500-UC540 membranes optimized for use with the MultiPep system have a density of about 400 nmol/cm2, as determined by UV quantification after coupling of Fmoc-Alanine. We found that acetyl-lysine carrying peptides synthesized on these membranes tend to saturate the readout signal very quickly, suggesting that large amounts are deposited during synthesis. This proved particularly useful when probing for very weak interactions between acetyl-lysine modified peptides and bromodomains, however strong interactions or interactions that involved binding of multiple domains to one peptide resulted in rapid signal saturation and un-interpretable results. We decided to use in-house generated membranes by directly functionalizing a cellulose surface (regular Whatman™ paper) allowing more flexibility in reducing the loading of each SPOT in our arrays by using different concentrations of Fmoc-β-Alanine during the functionalization of the membrane as previously described [26].

2.3. Control of SPOT Density
Commercially available pre-modified cellulose membrane sheets are typically used, however they do not offer the flexibility of in-house cellulose-functionalized membranes, particularly in reducing SPOT density. We functionalized membranes by esterification using Fmoc-β-Alanine and DIC (Fluka, distributed by Sigma-Aldrich, #38370) as previously reported [4,27]. Small (10 × 15 cm) sheets of Whatman™ cellulose (Whatman™ Chromatography paper Grade 1CHR, GE Healthcare Life Sciences, #3001-878) were typically incubated overnight (or at least for 2 h) with 10 mL of a solution comprising 0.64 g of Fmoc-β-Alanine dissolved in NMP, complemented with 374 µL of DIC and 317 µL of N-methylimidazole (NMI, Fisher Chemicals, Loughborough, UK, #M/4930/PB05). We found that the solution should cover the entire membrane without any bubbles forming in order to obtain consistency in the downstream array generation. Membranes were washed the following day 3 times with NMP for at least 30 s each, before being incubated for 20 min in fresh NMP.
Washed membranes were Fmoc de-protected by incubation (2 times 5 min each) in 20% Piperidine in DMF, followed by wash with NMP and (4 times for 30 s) then twice with EtOH. At this stage membranes were dried and stored at −20 °C or immediately used. Extra care was taken when handling in-house functionalized membranes as they tend to be much softer than the commercially available pre-functionalized ones. We found that they are also very easy to tear when manipulated with pincers. To avoid ring SPOT effects we found that (at least in the case of acetylation-dependent interactions) using 0.3 g of Fmoc-β-Alanine in 10 mL of solution produced the best results without any significant loss of signal.

2.4. Membrane Blocking
In order to avoid background binding of a protein/antibody outside the physical boundary of SPOTed peptides, membranes are typically blocked overnight with an appropriate agent. The choice of blocking agent is also critical for the sensitivity of the antibody used for detection. We have tested different blocking agents, including 5% skimmed milk, 5% bovine serum albumin (BSA, Fisher Scientific, #70955) and 5% alkali-soluble casein (Novagen, distributed by Merck-Millipore, Felyham, UK, #70955). We obtained optimal signal to noise ratio using BSA. It is important to note that BSA is also preferred over milk when detecting phosphorylated proteins, although care must be taken to avoid BSA which contains tyrosine phosphorylations resulting in high background when used with anti-phosphotyrosine antibodies.

2.5. Probing Protein:Peptide Interactions
Membranes were pre-wetted by rinsing several times with ethanol followed by 3 × 5 min washes with PBST buffer (3.2 mM Na2HPO4, 0.5 mM KH2PO4, 1.3 mM KCl, 135 mM NaCl, 0.1% Tween 20, pH 7.4). They were subsequently blocked with 5% BSA in PBST buffer for 8 h at room temperature in order to reduce non-specific binding. After 2 washes with PBST buffer (5 min each) followed by a single wash with PBS buffer (3.2 mM Na2HPO4, 0.5 mM KH2PO4, 1.3 mM KCl, 135 mM NaCl, pH 7.4) for 5 min, His6-tagged BRDs were added at a final concentration of 1 µM before an overnight incubation in PBS buffer at 4 °C. Each membrane was washed 3 times in PBST buffer to remove any unbound protein, blocked for 1 h with 5% BSA in PBST buffer, and washed again 3 × 5 min with PBST buffer to remove the excess of BSA. His-tag® Antibody HPR conjugated (Novagen, distributed by Merck-Millipore, #71841) was added in 1% BSA/PBST solution at a dilution of 1:3000. After 1 h incubation, membranes were washed 3 × 20 min in PBST buffer to remove any excess antibody. We found the large number of washing steps necessary in order to avoid low signal to background ratio. All incubation and washing steps were performed using a PMR-30 Compact Fixed-Angle Platform Rocker which was set to 30 oscillations per minute.

2.6. Quantification and Visualization
Pierce® ECL Western blotting Substrate (Thermo Scientific, distributed by Fisher Scientific, #32106) was used to reveal the bound antibody and chemiluminescence was detected with an image reader (Fujifilm LAS-4000 ver.2.0, GE Healthcare Life Sciences) typically using an incremental exposure time of 5 min for a total of 80 min (or until saturation was reached, in the case of very strong signal). The resulting SPOT intensities were quantified with the Kodak 1D V.3.6.2 Scientific Imaging System. Two profiles were generated for each SPOT, one covering the outer boundary of each SPOT (large profile) and one smaller (co-centric to the large profile) covering ~50% of the large profile (small profile). Numeric values were imported in Microsoft Excel, profiles were averaged and the intensity was normalized throughout the membrane between 0 and 100. We found that we could identify “corona” effects when the smaller profile intensity was 80%–85% of the large profile intensity, requiring manual adjustment of the data and further experiments. Data were binned using arbitrary derived values for each set of experiments by visual inspection of the membranes resulting in classification of SPOTs as “weak” (typically for normalized intensity values between 5% and 20%), “medium” (typically for normalized intensity values between 20% and 60%) and “strong”(typically for normalized values between 60% and 100%). Visualization was carried out with a simple VBScript within Microsoft Excel taking into account the intensity bins described.

2.7. Membrane Stripping
In principle, “stripping” of a membrane from the interacting protein it has been exposed to should be possible, resulting in regeneration of the membrane for subsequent use. It has been suggested that membranes can be regenerated up to 20 times [28], however if the binding to the SPOTed peptides is very strong, stripping may not yield a re-usable membrane. Several stripping protocols have been reported [10,27,29] employing different chemicals as well as different washing steps. In principle membranes are treated with chemicals aiming to denature the bound protein (e.g., β-mercaptoethanol, urea, guanidinium) together with a detergent (SDS or Triton X-100) which helps lift the protein from the membranes. Affinity chromatography matrixes, such as Ni-NTA agarose beads (Qiagen, Manchester, UK) or Talon Metal Affinity Resin (Clontech, Mountain view CA, USA), may help trapping the unbound protein, thus avoiding re-deposition onto the membrane. Subsequent treatment with TFA (ranging from diluted to highly concentrated) for up to 12 h, has been used to regenerate commercially available membranes (which seem to be quite resistant to acidic pH). Lower acid concentrations are required for in-house functionalized membranes which tend to be more prone to hydrolysis and loss of the ester used to attach the peptides on the sheet support. It is important that membranes remain wet following probing of interactions, in order to perform stripping steps. In addition, regeneration should be monitored by repeating the detection step in the absence of any protein sample in order to ensure that the membranes are clean and ready for re-use. We had mixed results with the regeneration of membrane carrying acetylated peptides, sometimes resulting in incomplete removal of the bound proteins. For this reason we typically used new in-house functionalized membranes for every experiment in order to avoid incomplete stripping.

2.8. Biolayer Interferometry (BLI)
In order to further assess binding to SPOTed peptides we employed biolayer interferometry against a commercially available set of biotinylated histone peptides (AltaBioSciences, York, UK Histone array, Set 4 Histone Acetyl-Lysine library), covering the same modifications studied in our membrane SPOTs, using the Octet RED384 system (FortéBio, Portsmouth, UK). Experiments were performed at 25 °C in BLI buffer (20 mM HEPES, pH 7.5, 150 mM NaCl and 0.5 mM TCEP) using the FortéBio data acquisition software V.7.1.0.100. Biotinylated peptides were first immobilized onto Super Streptavidin biosensors (SuperStreptavidin (SSA) Dip and Read Biosensors for kinetic #18-0011, FortéBio, Portsmouth, UK), pre-equilibrated in the BLI buffer then quenched in a solution of 5 µM Biotin (baseline equilibration 60 s, peptide loading for 240 s, quenching for 60 s, 1000× rpm shake speed, at 25 °C). The immobilized peptides were subsequently used in association and dissociation measurements performed within a time window of 600 s (base line equilibration 60 s, association for 600 s, dissociation for 600 s, 1000× rpm shake speed, at 25 °C). Interference patterns from peptide-coated biosensors without protein were used as controls. After referencing corrections, the subtracted binding interference data were analyzed using the FortéBio analysis software (FortéBio data analysis software V.7.1.0.38) provided with the instrument following the manufacturer’s protocols.

2.9. Experimental Protocol
The precise step-by-step protocol used in our experimental pipeline is given bellow, capturing all important observations that lead to reproducible and good quality SPOT assays in our lab (buffer recipes are also provided at the end of the protocol):
Day 1: Blocking and hybridization of the membrane (TIMING ~9 h) 
1| Rehydrate the membrane: rinse several times with 100% EtOH then equilibrate 3 × 5 min in PBST (see below).
2| Block the membrane with 10 mL 5% BSA in 1X PBST buffer for at least 8 h at room temperature (use a rocking table).
▲ CRITICAL: Always adjust the amount of solution according to the size of the membrane (and the container used to handle it). The membrane must be covered by the solution on every step.
▲ CRITICAL: Milk can be used for blocking but gives much higher background.
3| Wash the membrane 2 × 5 min in 1X PBST buffer then 1 × 5 min in PBS buffer at room temperature.
4| Add 1 µM (final concentration) of the protein of interest diluted in 10 mL 1X PBS buffer (volume adjustable); leave overnight on a rocking table at 4 °C.
Day 2: Development and reading (TIMING ~5 h) 
5| Wash the membrane 3 × 5 min with 1X PBST buffer to remove any unbound protein.
6| Block the membrane with 10 mL of 5% BSA in 1X PBST buffer for 1 h at room temperature.
7| Wash the membrane 3 × 5 min with 1X PBST buffer.
8| Add HPR-conjugated His antibody (1:3000 dilution) in 1% BSA PBST buffer and incubate for 1 h at room temperature.
9| Wash the membrane 3 × 15-20 min each in 1X PBST buffer.
▲ CRITICAL: These washes are critical to avoid high background.
10| Place the membrane quickly on drying paper to remove any liquid excess.
▲ CRITICAL: The membrane should not be left to dry completely.
11| Develop using an ECL kit; mix an equal amount of the peroxide solution and the Luminol Enhancer solution (toxic so always wear gloves) and cover the membrane. Let the reaction develop for one minute.
12| Remove the excess of the ECL solution and read the result on a chemiluminescence compatible imaging station (we use the ImageQuant LAS-4000 camera set on chemiluminescence settings) first for 1 min to check that the reaction develops properly, then for 80 min with 5 min increments.
13| Save the resulting image data in the TIF graphic format.
Day 2: Stripping—Part I (TIMING ~3 h) 
14| Rinse the membrane 3 times in distilled water.
15| Incubate the membrane in the Restore Western Blot Stripping Buffer for 30 min at 37 °C.
16| Wash the membrane 3 × 10 min in distilled water at room temperature.
17| Incubate 2 × 45-60 min in Stripping buffer A (see below) at room temperature.
18| Leave overnight at room temperature in Stripping buffer A with 500 µL of NiNTA beads to trap the released His-tagged protein.
Day 3: Stripping—Part II (TIMING ~4 h) 
19| Incubate 2 × 30 min in Stripping buffer B (see below) at room temperature.
20| Wash 3 × 10 min in distilled water at room temperature.
21| Wash 3 × 10 min in distilled water at 60 °C.
22| Wash 1 × 10 min in distilled water at room temperature.
23| Wash 1 × 30 min in 90% TFA at room temperature.
▲ CRITICAL: TFA is a very strong corrosive acid. Always wear gloves and protective goggles and leave the membrane in a fume hood during the reaction.
24| Wash the membrane 2 × 10 min in distilled water at room temperature.
25| Re-equilibrate the membrane 3 × 10 min in 1× PBST buffer.
■ PAUSE POINT If the membrane needs to be used the following day, it can be left overnight into 1X PBST buffer.
■ PAUSE POINT If the membrane is no longer needed, it can be stored at −20 °C for several months. In this case, the membrane needs to be dehydrated.
26| Wash the membrane 2 × 10 min in 20% EtOH at room temperature.
27| Wash the membrane 2 × 10 min in 50% EtOH at room temperature.
28| Wash the membrane 2 × 10 min in 95% EtOH at room temperature.
29| Let the membrane dry at room temperature overnight, wrap it in aluminum foil and store at −20 °C in a zip bag.
Day 4: Quality control of the stripping (TIMING ~5 h) 
30| Repeat steps 6 to 13.
31| If the membrane has been properly stripped (no signal apart from the positive controls), repeat steps 14 to 16, then go to step 25 or 26.
32| If the membrane still displays SPOTs in addition to the positive controls, the stripping process can be repeated (steps 14 to 30), but it is possible that a new membrane will have to be re-synthesized.
Solution/buffer recipes 
PBS Buffer (20×, 2L): Prepare by mixing 320 g of NaCl, 8 g of KCl, 57.6 g of Na2HPO4, 9.6 g of KH2PO4 in deionized water for a total volume of 2000 mL. Adjust pH to 7.4 with HCl.
PBST Buffer (20×, 2L): Prepare by mixing 320 g of NaCl, 8 g of KCl, 57.6 g of Na2HPO4, 9.6 g of KH2PO4, 20 mL of Tween-20 in diionized water for a total volume of 2000 mL. Adjust pH to 7.4 with HCl.
Stripping buffer A (1×, 1L): Prepare by mixing 573.18 g of guanidinium HCl salt (6M final concentration) and 10 mL Triton X-100 (1% final concentration) in distilled water for a total volume of 1000 mL.
Stripping buffer B (1×, 1L): Prepare by mixing 34.38 g of imidazole (500 mM final concentration), 29.22 g of NaCl (500 mM final concentration) and 20 mL of 1 M TRIS.HCl pH 7.5 (20 mM final concentration) in distilled water for a total volume of 1000 mL.
Materials needed His-tag® Antibody HPR conjugated, Novagen, distributed by Merck-Millipore, #71841.Pierce® ECL Western blotting Substrate, Thermo Scientific distributed by Fisher Scientific, #32106Bovine Serum Albumin Heat Shock Reagent grade powder pH7.0, Fisher Chemical, BPE 1600-100.Restore Western Blot Stripping Buffer, Thermo Scientific, #21059.