Hydrogen bond based smart polymer for highly selective and tunable capture of multiply phosphorylated peptides

Multisite phosphorylation is an important and common mechanism for finely regulating protein functions and subsequent cellular responses. However, this study is largely restricted by the difficulty to capture low-abundance multiply phosphorylated peptides (MPPs) from complex biosamples owing to the limitation of enrichment materials and their interactions with phosphates. Here we show that smart polymer can serve as an ideal platform to resolve this challenge. Driven by specific but tunable hydrogen bonding interactions, the smart polymer displays differential complexation with MPPs, singly phosphorylated and non-modified peptides. Importantly, MPP binding can be modulated conveniently and precisely by solution conditions, resulting in highly controllable MPP adsorption on material surface. This facilitates excellent performance in MPP enrichment and separation from model proteins and real biosamples. High enrichment selectivity and coverage, extraordinary adsorption capacities and recovery towards MPPs, as well as high discovery rates of unique phosphorylation sites, suggest its great potential in phosphoproteomics studies.


Synthesis of PNI-co-ATBA copolymer by atom transfer radical polymerization (ATRP)
A 25 mL round-bottom flask was charged with a solvent mixture containing DMF (0.5 mL), CH 3 OH (2.5 mL) and H 2 O (2.5 mL), then NIPAAm (1.13 g), ATBA (0.50 g), catalyst copper bromide (CuBr, 0.0142 g), ligand PMDETA (0.0173 g) with relative molar ratios of NIPAAm:ATBA:CuBr:PMDETA = 100:20:1:1 were added to the mixture. The flask was sealed with a rubber septum and evacuated, and back filled with argon four times. After that initiator 2-bromo-2-methylpropionic acid (0.0167 g) was added. The flask placed in a preheated oil bath was maintained at 60 o C for 24 h. Then the reaction mixture was diluted with CHCl 3 and passed through a basic alumina column to remove ATRP catalyst. The resulting solution was then concentrated and copolymer was precipitated into excess diethyl ether. The obtained copolymer was dried under vacuum to provide PNI-co-ATBA. The number-average molar mass (M n ) was 35,500 g·mol -1 and poly-dispersity index (PDI) was 1.28, determined by gel permeation chromatography (GPC) on a PL-GPC 50 system, using polystyrenes as standards for calibration and DMF as the eluent at a flow rate of 1.00 mL·min -1 at 25 o C.

Synthesis of PNI-co-ATBA thin film on silicon substrate
A clean silicon substrate was immersed in sodium hydroxide (NaOH, 0.1 mol·L -1 ) aqueous solution for 5 min and subsequently in nitric acid (HNO 3 , 0.1 mol·L -1 ) aqueous solution for 10 min to generate surface hydroxyl groups. After the substrate had been washed with an excess of water and dried under a flow of nitrogen, it was heated to reflux in toluene that contained 5 wt% ATMS for 3 h to obtain chemically bonded amine groups on the surface. The surface was rinsed with dry toluene and dichloromethane to remove remaining ATMS, then dried under a flow of nitrogen gas, and immersed in dry 10 mL dichloromethane that contained pyridine (2% w/v). Then the polymerization initiator bromoisobutyryl bromide (BIBB, 1.0 mL) was added dropwise into the solvent containing the silicon substrate at 0 o C, and the mixture was left for 1 h at this temperature then at room temperature for 12 h. The silicon substrate was cleaned with dichloromethane, and dried under a nitrogen flow. Taking advantage of a surface initiated atom transfer radical polymerization (SI-ATRP), [1] copolymerization of PNI-co-ATBA was achieved by immersing the silicon substrate with the initiator grafted on the surface in a degassed solution of NIPAAm (1.13 g), ATBA (0.50 g, 20 mol% ATBA against NIPAAm) in a degassed mixture of H 2 O (5.0 mL), methanol (5.0 mL) and DMF (1.0 mL) containing CuBr (0.0284 g, 0.2 mmol) and PMDETA (0.0346 g) for 6 h, at 60 o C. After that, the reaction was stopped by removing the substrate from the polymerization bath, the copolymer grafted substrate was washed with CH 3 OH, DMF and H 2 O, respectively, and dried under a N 2 stream. Under this condition, the thickness of the copolymer film was between 18 nm and 20 nm.

Synthesis of PNI-co-ATBA thin film on Au-coated quartz-crystal resonators or SPR sensor chips
Au-coated quartz-crystal (QC) resonator was washed with distilled water and ethanol three times. Then a monolayer of 2-mercaptoethylamine was covered on the gold surface after the Au-coated QC resonator was immersed in a solution of 2-mercaptoethylamine (1 × 10 -2 mol·L -1 ) in ethanol for at least 24 h. After that the QC resonator was rinsed with ethanol three times and dried under a flow of nitrogen gas. Then the QC resonator was immersed in dry CH 2 Cl 2 that contained pyridine (2% w/v). The polymerization initiator BIBB was added dropwise into the solvent containing the QC resonator at 0 o C, and the mixture was left at this temperature for 1 h then at room temperature for 12 h. Subsequently, the QC resonator was cleaned with CH 2 Cl 2 , and dried under a nitrogen flow, generating a bromine-modified Au surface suitable for polymerization. Then the same procedure described above was adopted in order to graft the PNI-co-ATBA copolymer thin film on the Au-coated QC resonator.
The same method was used to graft PNI-co-ATBA 0.2 on the SPR sensor chip surface.

H and 13 C NMR titration experiments
ATBA is regarded as the critical recognition unit in the smart copolymer. In order to validate the combination between ATBA and phosphates, 1 H and 13 C NMR titration experiment were performed to investigate the binding details. [2] To avoid the interference of D 2 O with strong suppression effect on hydrogen bonding, d 6 -DMSO was chosen as the solvent because both ATBA and phosphates were well soluble in it. HPO 4 2anion was prepared according to the reference [3] and tetrabutylammonium (Bu 4 N + ) works as cation. The detailed preparation method for anion is described as below: The tetrabutylammonium salts were prepared by adding 1 equiv.   were used to investigate the interaction of ATBA with an equimolar ratio of HPO 4 2in d 6 -DMSO. Similarly, the binding behavior of ATBA with phenyl phosphate anion (Bu 4 N + works as cation) [3] was also studied by 1 H NMR, as shown in Supplementary Figure 6.

Polymer lower critical solution temperature (LCST) measurement
The polymer solution was injected into a closed quartz cell and the LCST measurement could be completed within 1 h, under this condition, the solution pH value would not change remarkably. It is worth mentioning that various buffer solutions (e.g., phosphate, Tris-HCl, or ammonium formate) were not used because these buffering agents might also impact on the LCST of the copolymer. Transmittance of copolymer solution at 500 nm was measured by UV-Vis spectrophotometer at different temperature, then the effect of solution pH or addition of PP on the LCST of the copolymer were discussed. According to the dramatic change of transmittance near the LCST, the copolymer LCST was determined to be approximately 28 o C in pure water (pH 6.5). Its LCST decreased to 26.6 o C at pH 3.0 and increased to 29.1 o C at pH 10.0. In addition, upon the additions of PPs, the copolymer LCST decreased from 28 o C to 27 o C or 26.4 o C for 1pS or 4pS, respectively. These data indicated that PNI-co-ATBA 0.2 was a typical thermo-responsive polymer and its LCST was strongly influenced by the solution pH value or the addition of PPs. In this experiment, buffer solution was not used in order to eliminate its effect on the copolymer LCST.

Fluorescent titration experiment
Fluorescent titration experiment, a typical and widely adopted method for calculating the association constant (K a ) in host-guest chemistry, [4] (Supplementary Figure 7). [5] Calculation formula [5,6]  SPR measurement of association rate constant (K a ) of peptide on copolymer surface SPR is an optical phenomenon which is highly sensitive for detecting refractive index changes near the measurement surface, in particular molecular binding or release. Using the SPR phenomenon, multi-parametric SPR is a real-time and label free in vitro tool for investigating molecule-molecule interactions, providing information on the kinetics and affinity of the studied system. In this experiment, SPR sensor chips with Au coating of 50 nm were purchased from BioNavis Corp, Finland. The Au-coated sensor chips covered with PNI-co-ATBA 0.2 were prepared through the procedure described above. Initially, the chemical modified chip was washed with DMF and water several times, and then it was put into a sample chamber for SPR measurement after dried under nitrogen gas. After stabilization of fundamental resonant angle with pure water and buffer solution (80% CH 3  1pS-3pS SPR adsorption data, 1:1 binding mode was applied to fit the data; for 4pS, a 1:2 binding mode was applied to fit the data, then association rate constants (K a ) were calculated out according to the fitting curve. Each SPR adsorption experiment was repeated three times to obtain the reliable K a data.

Quartz crystal microbalance-dissipation (QCM-D) adsorption experiment
QCM-D adsorption measurement was all carried on a Q-Sense E4 system (Sweden).
Au-coated quartz crystals (QC) with intrinsic frequency (F 0 ) of 5 MHz were all purchased from Q-Sense Corp. (Sweden). The Au-coated QCs covered with our copolymers or ATBA monolayer were prepared through the procedure described above. Initially, the copolymer-modified QC was washed with DMF and water three times, and then it was put available TiO 2 microspheres (particle size: 5 m, pore size: 100 Å, ordered from GL Sciences, Tokyo, Japan) were covalently bound onto the QC surfaces (Supplementary Figure 23a).
Under this condition, serine tetra-PPs (4pS) adsorption-induced frequency change on TiO 2 surface was still lower than 180 Hz (Supplementary Figure 23b), which was still far smaller than that on our copolymer surface.

AFM experiment
AFM measurements were performed using a Multimode 8 AFM (Bruker, USA). The Au-coated QC (Sample A) covered with copolymers was prepared through the procedure

Bio-attenuated total reflection (ATR) mode
Infrared spectra were recorded on a Bruker Vertex 80v FT-IR spectrometer with a Bio-ATR cell II accessory, which is based on dual crystal technology: the top crystal is made of silicon, and the second crystal has a hemispherical design and is made of zincselenide (ZnSe). The Bio-ATR II unit is factory-preassigned; hence, no alignment is required. Model PPs and ATBA were prepared to 0.   4 3-, benzene phosphate or acetate was 290, 7.14 × 10 5 , 2.7 × 10 6 , 6.81 × 10 3 or 620 L·mol -1 , respectively. The K a (HPO 4 2-) / K a (CH 3 COO -) ratio was 1150:1. Energy dissipation (D) during the oscillation of resonator is an indication of the rigidity of the adsorbed layer. [7] The increase of dissipation indicated that the adsorbed layer became less rigid. Thus, we presumed that the copolymer chains became more relaxed and expanded in response to the MPP adsorption. This presumption was further proven by the relationships between ΔD and ΔF corresponding to the adsorption processes (Supplementary Figure 35).

Supplementary
The ΔD/ΔF curve indicates the formation of a viscoelastic and hydrated layer, a positive shift in the slope denotes the formation of a more viscoelastic and less rigid layer. [7] Based on this knowledge, 4pS-adsorption triggered a rapid and obvious expansion of the copolymer chains, accompanying with dramatic change in viscoelasticity and Frigidity of the copolymer film. In comparison, no evidential change of the copolymer film was observed in response to the Where F represents the fluorescent intensity, F 0 and F lim are the initial and ultimate fluorescent intensity, respectively, and C H and C G are the corresponding concentrations of host fluorescein-labeled thioureido-benzoic acid and anion guest, C 0 is the initial concentration of host.
[c] Reliable K a could not be obtained due to too small change in fluorescence spectra.
Where F represents the fluorescent intensity, F 0 and F lim are the initial and ultimate fluorescent intensity, respectively, and C H and C G are the corresponding concentrations of host fluorescein-labeled thioureido-benzoic acid and anion guest, C 0 is the initial concentration of host. Where F represents the fluorescent intensity, F 0 and F ∞ are the initial and ultimate fluorescent intensity, respectively, and G are the corresponding concentrations of anion guest, n is the stoichiometry.

Supplementary
[d] Reliable K a could not be obtained due to too small change in fluorescence spectra.