Terminally Phosphorylated Triblock Polyethers Acting Both as Templates and Pore-Forming Agents for Surface Molecular Imprinting of Monoliths Targeting Phosphopeptides

The novel process reported here described the manufacture of monolithic molecularly imprinted polymers (MIPs) using a terminally functionalized block copolymer as the imprinting template and pore-forming agent. The MIPs were prepared through a step-growth polymerization process using a melamine–formaldehyde precondensate in a biphasic solvent system. Despite having a relatively low imprinting factor, the use of MIP monolith in liquid chromatography demonstrated the ability to selectively target desired analytes. An MIP capillary column was able to separate monophosphorylated peptides from a tryptic digest of bovine serum albumin. Multivariate data analysis and modeling of the phosphorylated and nonphosphorylated peptide retention times revealed that the number of phosphorylations was the strongest retention contributor for peptide retention on the monolithic MIP capillary column.


Characterization and Evaluation Procedures
Preparation of trypsin digested proteins. Twelve proteins containing phosphorylated residues were used to prepare tryptic digest samples: Carbonic anhydrase (bovine), serum albumin (bovine), ribonuclease B (bovine pancreas), lactoglobulin (bovine), α-casein (bovine), β-casein (bovine), ovalbumin (chicken), lysozyme (chicken), alcohol dehydrogenase (Saccharomyces cerevisiae), myoglobin (whale skeletal muscle), α-amylase (Bacillus sp.) and transferrin (human). All proteins were from Sigma-Aldrich except transferrin, which was from ACE Biosciences (Odense, Denmark). Each protein was separately dissolved in 50 mM triethylammonium bicarbonate buffer to a concentration of 20 pmol/µL, reduced with 10 mM of 1,4-dithiothreitol at 56 °C for 30 min, and subsequently alkylated with 40 mM of iodoacetamide for 30 min at room temperature, protected from light. Trypsin (1 % of the protein mass) was thereafter added followed by incubation at 37 °C for 12 h. The protein digest sample was prepared by mixing peptides from each of the protein digests in equimolar ratio and diluting with 0.1 % aqueous trifluoroacetic acid to final concentrations of 2 pmol/µL for each protein. This peptide mixture was stored at -20 °C until further use.
Nitrogen Cryosorption. Extracted and dried monolith cubes with ≈ 2 mm sides (50 to 150 mg) recovered from the vials were transferred to dry 9.5 mm i.d. sample tubes and further dried for 2 h at 60 °C by a flow of dry N 2 using a Micromeritics SmartPrep degassing unit (Atlanta, GA, USA). The samples were then immediately subjected to multipoint nitrogen adsorption-desorption analysis at cryoscopic temperature on a Micromeritics TriStar 3000 automated gas adsorption analyzer. The specific surface areas of the monolithic materials were determined based on the Brunauer-Emmett-Teller (BET) model S1 using adsorption data in the relative pressure range from 0.18 to 0.35. The Barrett-Joyner-Halenda (BJH) scheme S2 was used to estimate the total pore volume, average mesopore diameters between 1.7 and 300 nm, and pore size distribution based on the desorption branches of the isotherms, shown in Figures S9c-d.
Field-Emission Scanning Electron Microscopy. Samples with fresh breakage rupture surfaces were positioned on adhesive carbon foils attached to standard aluminum sample stubs and mechanically secured to the holders by conductive glue (Ted Pella Inc., Redding, CA, USA). All samples were thereafter coated with a 10 nm of layer of platinum in a Quorum Q150TS sputter coater (Quorum Technologies, Ringmer, UK). The metal-coated samples were analyzed in a Zeiss Merlin field emission scanning electron microscope from Carl Zeiss Microscopy GmbH (Oberkochen, Germany) operated at 5 kV acceleration voltage. Images were captured from randomly chosen areas at standardized magnifications.
Fourier-transform Infrared Spectroscopy. Diffuse reflectance Fourier transform infrared spectroscopy (DRIFTS) was performed on a Bruker (Ettlingen, Germany) IFS 66 FTIR spectrometer under ≈ 400 Pa partial vacuum. Dried monoliths (≈ 10 mg) were ground manually together with ≈ 390 mg of KBr using an agate mortar and pestle followed by direct transfer to a DRA-2CI diffuse reflectance cell from Harrick Scientific Products (Pleasantville, NY, USA). The signals were recorded using a total of 256 coadded interferogram scans, transformed to obtain spectra between 4000 and 400 cm −1 at a spectral resolution of 4 cm −1 with acceptable signal-to-noise ratio.
X-ray Photoelectron Spectroscopy (XPS). Surface elemental analysis and elemental stoichiometries were determined on dried monoliths by means of XPS. A Kratos Axis Ultra DLD electron spectrometer (Kratos Analytical, Manchester, UK) equipped with a monochromatic Al K α1 Xray source (1486.6 eV) operated at 150 W and a delay-line detector was used in the measurements. The surface potential was stabilized by the charge neutralization system of the spectrometer. All binding energies (BE) are referenced to the C 1s C-H peaks of aliphatic carbon at 285.0 eV and at 284.7 eV for aromatic carbon. Processing of the spectra was accomplished with the Kratos software.
HPLC-MS/MS Method. Reversed phase HPLC was used to determine the unbound amounts of test compounds in the isotherm experiments. The column was an ACE-III C 18 (150 × 2.1 mm, 5 µm) from Advanced Chromatography Technologies (Aberdeen, UK) and separations were done in isocratic mode. When the target compound was PPA, the mobile phase consisted of 0.1 % formic acid in acetonitrile/water 1:1 (v/v), pumped at a flow rate of 200 µL/min. Mass spectrometric detection was performed on an LCQ Duo ion-trap mass spectrometer (Thermo Finnigan, San Jose, CA, USA), in negative ion mode with quantitative fragment set to m/z = 173.00 and qualitative fragment set to m/z = 79.00. For Fmoc-pY-OEt, the mobile phase consisted of 35 vol-% acetonitrile in 10 mM aqueous ammonium formate. Mass spectrometric detection was performed in positive mode, with quantitative fragment set at m/z = 501.10 and qualitative fragment at m/z = 483.70.
Solid-state Nuclear Magnetic Resonance Spectroscopy. Extracted, dried, and ground monolith powder was transferred to 4 mm inner diameter zirconium oxide (ZrO 2 ) NMR rotors with Kel-F inserts. Crosspolarization magic angle spinning 13 C NMR ( 13 C CP-MAS NMR) was performed on a Bruker (Billerica, MA) Avance III 500 MHz spectrometer, equipped with a 13 C CP-MAS probe, spun at a rate of 10.0 kHz. Spectra were acquired at 298 K with a 2.85 µs proton 90° pulse, followed by a cross polarization step using a 13 C spin lock field strength of 62.5 kHz and a ramped 1 H field strength (from 43 to 86 kHz) during 1.5 ms. 1 H decoupling was done by the SPINAL64 sequence S3 using a decoupling field strength of 88 kHz during the 6.8 ms acquisition time. The FID signals were collected with 2 s relaxation delay and 3000 scans and the accumulated FID signals were multiplied by a Gaussian apodization function prior to Fourier transform, followed by manual phase and baseline correction. Adamantane was used as an external chemical shift reference with CH 2 signals set to 38.5 ppm. All spectra processing was performed using Bruker TopSpin 4.0.6.
Determination of Polymer Composition using Nuclear Magnetic Resonance Spectroscopy. Experiments were carried out at 298 K using a Bruker AVIII 400MHz spectrometer equipped with a 5 mm SmartProbe BBF-H/D. Pluronics F127-OH and F127-PO 4 (0.100 ± 0.001 g) were dissolved in separate 500 µL aliquots of a 80:20 (v/v) mixture of acetonitrile-d3 and deuterium oxide. Spectra were referenced using the acetonitrile proton peak at 1.94 ppm. The HDO peaks appeared at 3.84 ppm with F127-OH and at 4.15 ppm with F127-PO 4 (Figures S4a and  c). This can be compared to a shift of 3.66 ppm observed by Koley and Ghosh for χ H 2 O = 0.42 S4 . Spectra were recorded by Bruker TopSpin 4.0.6, using 1024 zg30 sequence pulses at 5 s intervals with receiver gain of 90.5, 5 s relaxation delay, 10.4 μs pulse width, and 7978.7 Hz spectral width.
A detailed composition analysis of Pluronics using NMR spectrum is illustrated below.
(i) Normalize methyl proton of PPO at 1.05 ppm to 3.0 (ii) Determine the relative moles of PPO (denoted as x) [S1] (iii) Determine the relative moles of PEO (denoted as y) [S2] (iv) Calculate the weight percentages of PEO and PPO [S3] DLS and Zeta-potential Measurement. The cluster sizes and zeta potentials of the mixed porogens in a biphasic solvent mimicking polymerization condition were both measured at 60 o C using a Zetasizer Nano ZS (Malvern Instruments). A 633 nm He-Ne laser was used as light source and an avalanche photodiode (APD) as detector. The porogen cluster sizes were measured using the dynamic light scattering method, where the scattered light was collected at 173 o . The values are reported as the median cluster diameters, in plots of hydrodynamic diameter (nm) vs. intensity (%). Zeta potential measurements were based on the electrophoretic light scattering principle. A disposable folded capillary cell DTS1070 from Malvern Instruments was used and signals were collected at 12.8 o . All data were analyzed by Zetasizer Software 7.13. The Smoluchowski model was used to calculate the hydrodynamic diameters and the zeta potentials were calculated by the default method in the Zetasizer software. Measurements were made in the porogen solution prepared as described in the experimental section, to which Milli-Q water was added in ratio of 1:1.3 to mimic the polymerization system. Additionally, three solutions, each only containing one of the porogen mixture components (PPG4000, F127-OH, or F127-PO 4 ), were prepared in the same ratio to investigate their individual roles in the micelle and cluster formation.
Binding Isotherm Assay. The crushed and sieved monolithic materials (5 mg each of MIP and NIP) were separately mixed in 1.5 mL screw cap vials with 1 mL of phosphorylated probe PPA at varying concentrations in acetonitrile/water 1:1 (v/v) containing 0.1 % (w/w) formic acid, then shaken overnight on a Vibrax VXR Orbital Shaker (IKA-Werke, Staufen, Germany). The samples were then centrifuged for 3 min at 2800 × g relative centrifugal force, and the supernatants were analyzed by the RP-HPLC-MS/MS method described above to determine the concentrations of the unbound probes. The amount of bound analyte per unit surface area of polymer (B) was calculated according to [S4] where C 0 and C are the probe concentrations of the initial solution and supernatant, respectively, V is the total volume of the adsorption mixture, m is the mass, and S the specific surface area of the polymer in each vial.
Binding curves were established by plotting B against C, and fitting these by non-linear regression in OriginPro 2020 (OriginLab Corporation, USA) to a Langmuir mono-site model, [S5] where B max is the maximum amount of probe bound to each surface area unit, and K eq is the dissociation constant.
Imprinting factors (IF) were calculated by the ratios of saturated uptake of the MIP and NIP, as Capillary Liquid Chromatography. LC-MS evaluations of the capillary monolithic columns were done by loading test mixtures of non-phosphorylated and phosphorylated peptides onto a 300 μm i.d. × 5 mm C18 PepMap 100 trap column (5 µm, 100 Å, Thermo Scientific, Waltham, MA, USA), followed by transfer of these peptides onto the monolithic capillary acting as separation column. The LC-MS set-up consisted of an Ultimate 3000 HPLC system connected to an LTQ Orbitrap XL mass spectrometer, both from Thermo Scientific. The outlets from the capillary monoliths were connected directly to 20 µm PicoTip emitters (New Objective, Woburn, MA, USA) to produce a nano-electrospray. Mobile phase A (MPA) consisted of 0.1 % (v/v) formic acid (FA) in water, and mobile phase B (MPB) consisted of 95 % acetonitrile, 4.9 % water, and 0.1 % FA. In the loading step, 2 % of MPB was used to capture all peptides to the C18 trap column before eluting them onto the monolithic capillary column by 20 % of MPB for 2 min. The peptides loaded onto the capillary monolith were separated by a gradient with increasing percentage of MPB: 20 % (5 min), 20-35 % (0.5 min), 35 % (5.5 min), 35-60 % (0.5 min), 60 % (5.5 min), 60-95 % (3 min), and 95 % (5 min). The gradient is shown in Figure S15.
Two testing mixtures were designed for investigation of the selective enrichment of phosphorylated peptides onto the MF capillary monoliths. The first contained a mixture of GADDSYYTAR peptides (non-and mono-phosphorylated on tyrosine (YY, and pYY or YpY), at final concentrations of 250 fmol/µL of each) spiked into a tryptic digest of 500 fmol/µL bovine serum albumin (BSA) to act as a realistic matrix. The second test mixture, aimed to provide multivariate data, consisted of a pooled tryptic digest of twelve proteins (2 pmol/µL in total), which included numerous naturally phosphorylated tryptic peptides.
The capillary monoliths were equilibrated with at least 100 µL of 20 % MPB before each injection of 1 µL test sample dissolved in MPA. Separations were performed at a total eluent flow rate of 1 µL/min with the gradient composition given above. Blanks of MPA were injected between each sample to minimize carry-over. Mass spectrometric detection was performed in positive ion mode with data dependent acquisition (DDA, top 4), fragmentation induced by higher-energy collisional dissociation (HCD, 29 eV). The mass tolerance of the precursor ion was set to 5 ppm.
Data were acquired as RAW files by Xcalibur 2.5.5 SP1 (Thermo Scientific), from which extracted ion chromatograms (EICs) based on exact monoisotopic m/z values for the respective peptides were extracted by the msaccess module of ProteoWizard 3.0.20337 (64-bit) S5 . Carbamidomethyl (C) was selected as fixed and phospho (S, T, and Y) as variable modifications. A Savitzky-Golay filtering (3 rd order, 15 points) was implemented by the sgolayfilt function from version 0.7-6 of the signal package S6 in the 64-bit version 4.0.2 of R S7 . EICs having filtered peaks with intensity higher than 10 5 were used for further data analysis.
Multivariate Data Analysis. Modeling of peptide properties against their retention was done by projection against latent structures (PLS) in SIMCA 15.0 (Umetrics, Umeå, Sweden), using aliphatic index S8 , molecular weight, and the total number of residues with phosphorylated, basic, acidic, and hydrophilic side chains as explanatory variables. As explained variable the peptide retention times were chosen, corrected by subtracting the system dwell time determined by substituting an empty capillary for the monolith-filled one. All variables were centered and scaled to unit variance. The peptide sequences, and their given codes (where e.g. "2" means the second non-phosphorylated peptide and "58-1P" means the 58 th peptide phosphorylated at a single site; see Table S4) as Primary IDs. The PLS models were extracted with single components since further component extractions did not improve the cross-validation score, Q 2 . Biplots of loadings and scores were produced using the built-in procedure in SIMCA to visualize the connections between samples, factors, and responses in a single plot. The kernel density estimates for the non-phosphorylated and phosphorylated peptides in Figure 4 were computed separately with the density function in R, using a Gaussian kernel at the default bandwidth, followed by normalization to unit magnitude. A hierarchical cluster analysis was finally applied to analyze the separation of natural tryptic peptides on the NIP and MIP monolithic columns. Figure S1. Modifications of the capillary inner surfaces. The fused silica surface is first etched, then the siloxane groups are converted into silanol groups that react with ethoxysilanes groups to form 3-D attachment of amine to the surface (C1). Then, C1 was treated by formaldehyde or diluted trimethylol melamine pre-polymer to produce capillaries C2 or C3, respectively.     (c, d) MIP in capillaries aminated by APTES further reacted with aqueous formaldehyde solution (capillary C2; c) and diluted MF precursor (capillary C3; d). Images have been contrast adjusted to cover the entire greyscale. Figure S8. Field emission scanning electron micrographs (FE-SEM) of random fracture cross-section surfaces of (left) NIP, and (right) MIP monoliths in capillaries at 5k, 20k, and 50k times magnifications. Images have been contrast adjusted to cover the entire greyscale.

S-6
S-10 Figure S9. (a-b) Histograms of fused monolithic particle diameters (n = 200) for the NIP and MIP monoliths, respectively. The particle diameters were measured by ImageJ (NIH, USA) and fitting of normal distribution functions to the histograms was done with non-linear least squares (NLS) fitting in R, using the Gauss-Newton algorithm; (c-d) BJH pore volume plots dV/dlog(D) from the desorption branch of nitrogen cryosorption isotherms. Figure S10. FT-IR spectra of NIP and MIP monoliths. Figure S11. Survey X-ray photoelectron spectroscopic spectra of NIP and MIP monolith surfaces.