Affinity‐Directed Site‐Specific Protein Labeling and Its Application to Antibody‐Drug Conjugates

Abstract Chemically modified proteins have diverse applications; however, conventional chemo‐selective methods often yield heterogeneously labeled products. To address this limitation, site‐specific protein labeling holds significant potential, driving extensive research in this area. Nevertheless, site‐specific modification of native proteins remains challenging owing to the complexity of their functional groups. Therefore, a method for site‐selective labeling of intact proteins is aimed to design. In this study, a novel approach to traceless affinity‐directed intact protein labeling is established, which leverages small binding proteins and genetic code expansion technology. By applying this method, a site‐specific antibody labeling with a drug, which leads to the production of highly effective antibody‐drug conjugates specifically targeting breast cancer cell lines is achieved. This approach enables traceless conjugation of intact target proteins, which is a critical advantage in pharmaceutical applications. Furthermore, small helical binding proteins can be easily engineered for various target proteins, thereby expanding their potential applications in diverse fields. This innovative approach represents a significant advancement in site‐specific modification of native proteins, including antibodies. It also bears immense potential for facilitating the development of therapeutic agents for various diseases.


Analytical HPLC analysis
For reversed-phase liquid chromatography (RPLC), each sample (3 μg) was analyzed using Agilent 1260 Infinity I with Poroshell 300SB-C8 (Agilent, 1.0 × 75 mm, 5 μm) at a flow rate of 1 mL/min.Initially, 80% of mobile phase A (0.1% TFA in water) was kept for 2 min.Mobile phase B (0.1% TFA in can) then increased from 20% to 60% at 2 min to 8 min.For an additional 2 min (from 8 min to 10 min), mobile phase B increased from 60% to 98% to elute all the samples in the column.The samples were detected by a variable wavelength detector (VWD) at 280 nm.
For hydrophobic interaction chromatography (HIC), each sample (10 μg) was analyzed using the same HPLC with MAbPac HIC-Butyl HPLC (Thermo Fisher Scientific, 4.6 × 100 mm, 5 μm) at a flow rate of 1 mL/min.Initially, 100% of buffer A (50 mM sodium phosphate, pH 7.0, 1.5 M ammonium sulfate) was kept for 1 min.Buffer B (50 mM sodium phosphate, pH 7.0, 20% IPA(v/v)) then increased from 0% to 100% at 1 min to 15 min.For an additional 5 min, 100% of buffer B was kept, eluting all the samples in the column.The samples were detected by VWD at 280 nm.

Mass analysis of Fc fragment and trastuzumab
Each sample (30 μg) was cleaned up with PD SpinTrap G-25 with ammonium bicarbonate buffer (50 mM) and treated with 30 units of IdeS enzyme prior to the analysis.The samples were analyzed using a linear ion-trap Orbitrap mass spectrometer (LTQ Orbitrap XL, Thermo Fisher Scientific) with a combination of ultra-high performance liquid chromatography, Ultimate 3000 (UHPLC, Thermo Fisher Scientific), using ACQUITY UPLC Protein BEH C4 Column (Waters, 300 Å, 2.1 × 50 mm, 1.7 μm).For the UHPLC method, mobile phase A (0.1% TFA in water) was initially kept at 80% for 1 min.Mobile phase B (0.1% TFA in ACN) then increased from 20% to 90% at 1 min to 10 min.Orbitrap-MS was detected with an extended mass range from 500 to 4000 m/z.Deconvolution of the peaks was conducted using UniDec software. [1]rification of ADC using HIC To purify trastuzumab-(VC-PABC-MMAE)2, a HIC column prepacked with Skillpak Phenyl-650S HIC resin (Tosoh Bioscience, 0.8 × 10 cm, 5 mL) was employed and connected to AKTA pure 25 (Cytiva) (Advanced Bio-interface Core Research Facility).The entire process was conducted at room temperature.Prior to sample injection, the column was equilibrated with 20 column volumes of buffer A (50 mM sodium phosphate, pH 7.0, 2 M NaCl).To load onto the column, 0.25 mL of unpurified ADC (1 mg/mL in formulation buffer) was mixed with 0.25 mL of buffer A. This resulting mixture (total volume of 0.5 mL) was then injected into the column and eluted using a linear gradient from 100% buffer A to 100% buffer B (50 mM sodium phosphate, pH 7.0, 20% IPA(v/v)).

S5
NMR 1 H NMR spectra were recorded on Bruker 400 MHz spectrometers at ambient temperature and spectra were processed using MestReNova 6.0.2 using the automatic phasing and polynomial baseline correction capabilities.
Otherwise, 1 H NMR spectra were recorded on Varian 400 MHz spectrometers.Routine 13 C NMR spectra were recorded on Bruker 400 MHz or Bruker 700 MHz spectrometers with protons fully decoupled. 13C Resonances are reported in ppm relative to solvent residual peaks for CDCl3 (77.16 ppm).

FTIR
Infrared spectra were recorded on a JASCO FT/IR-4600 spectrometer and νmax are partially reported in cm -1 .Highresolution mass spectra were acquired on a JEOL JMS-700 instrument with an FAB mode.Analytical thin-layer chromatography was performed using 60 Å Silica Gel F254 pre-coated plates (0.25 mm thickness).TLC plates were visualized by irradiation with a UV lamp.Normal-phase column chromatography was performed using 60 Å Silica Gel (32-62 micron) with an appropriate mobile phase composition and gradient.Normal-phase high-performance liquid chromatography was performed using an Agilent 1260 series instrument equipped with a diode array detector and columns from COSMOSIL.

Protein expression, ncAA incorporation, and purification in E. coli
To obtain Z-DM protein, the Z-DM gene was created by amplifying the Z-domain WT gene, which was from a commercial source by gene synthesis, and overlapping with strep-tag sequence (GWSHPQFEK) at the N-terminus and a His6-tag at the C-terminus.The Z-DM gene was then inserted between the NdeI and EcoRI sites of pET20b to generate pET20b-Z-DM.The plasmid was then transformed into Escherichia coli BL21(DE3) cell and amplified overnight at 37 °C in 6 mL LB broth with ampicillin (100 μg/mL).Amplified cells (1 mL) were inoculated to defined media [2] (50 mM Na2HPO4, 50 mM KH2PO4, 25 mM (NH4)2SO4, 2 mM MgSO4, 0.1% trace metals, 0.5% glycerol, 0.05% glucose, 0.2% lactose, and 5% amino acids) with the same amount of ampicillin, and the cells were incubated overnight at 37 °C.Cells were harvested by centrifugal force (10000 rpm, 5 min) and cell pellets were frozen at -20 °C for further purification.The pellets were thawed and purified with Ni-NTA agarose resin according to the manufacturer's protocol.The concentration of purified proteins was measured by UV absorbance at 280 nm and purity was inspected by 8-12% SDS-PAGE.The molar extinction coefficient of the protein was calculated from Biomol Protein Extinction Coefficient Calculator (http://www.biomol.net/en/tools/proteinextinction.htm).
The Z-AFB gene was also from a commercial source by gene synthesis and a His6-tag was attached at the Cterminus using overlapping PCR.The Z-M gene was created from the minimized Z-domain of protein A [3] with the third helix of Z-domain WT.Also, a His6-tag was added to the C-terminus of the gene.The Z-AFB gene was inserted between NcoI and KpnI sites of pBAD and the Z-M gene was inserted between NdeI and EcoRI of pET20b, generating pBAD-Z-AFB and pET20b-Z-M, respectively.pBAD-Z-AFB variants (K4, K28, F32TAG), and pET20b-Z-M variants (M8, H19, K33, D37TAG) were generated by site-directed mutagenesis.For pBAD-Z-AFB_F32TAG, pET20b-Z-M_M8TAG, and D37TAG, the nearest lysines were substituted to Arg (CGT codon) in order to avoid the self-labeling of the probe, creating pBAD-Z-AFB_K28CGT_F32TAG, pET20b-Z-M_K5CGT_M8TAG, and pET20b-Z-M_K33CGT_D37TAG, respectively.The plasmids were then co-transformed with pEvol-AzFRS [4] or pEvol-AzKRS [5] in DH10B or BL21 cells.Transformed cells were expressed and purified the same way as described above with ncAA (100 μM of AzF or 1 mM of AzK) and chloramphenicol (35 μg/mL).Concentration and purity were also inspected by the same methods.

Expression and purification of IgG1 Fc fragment
The Fc fragment gene was amplified by PCR and then inserted between the Xho1 and BamH1 sites of pCDNA3.4 to generate pCDNA3.4-Fcfragment.The plasmid containing the Fc fragment was transiently transfected into FreeStyleTM 293 cells using PEI-Transfection reagent according to the manufacturer's protocol.For expression, the supernatant of FreeStyleTM 293 cells transiently transfected with the respective Fc fragment construct was harvested by centrifugation at 8,000 rpm for 20 min at 4 °C.The expressed Fc fragment was purified using a Hitrap Protein A HP column according to the manufacturer's protocol.The column was equilibrated with buffer A (1.8 mM KH2PO4, 10 mM K2HPO4, 137 mM NaCl, 2.7 mM KCl, pH 7.4).The supernatant containing the Fc fragment was loaded onto the column, and after washing to remove unbound impurities, the purified Fc fragment was eluted with buffer B (100 mM citrate, 100 mM NaCl, pH 3.0).Purified Fc fragments were dialyzed against A buffer three times to remove any residual purification buffers.The concentrations of Fc fragments were determined by measuring absorbance at 280 nm using the molar extinction coefficient (7.157 × 10 4 cm -1 M -1 ) of the Fc fragment.

Procedures for protein labeling
Labeling of NASA-FL to Z-DM Z-AFB variants containing PyOx (30 μM) were mixed with Z-DM protein (15 μM).NASA-FL (75 μM) was then added to the mixture and incubated for 4 h in HEPES buffer (50 mM, pH 7.2) at 37 °C.The reaction was quenched by a mixture of L-Lys and PyOx-Butyne at final concentrations of 5 mM.Labeled products were desalted with ZipTip with 0.6 µL C18 resin according to the manufacturer's protocol and the labeling yields were determined by MALDI-TOF MS analysis (matrix: sinapic acid).To further isolate labeled Z-DM from Z-AFB, purification using Strep•Tactin ® Superflow™ Agarose (Novagen, Germany) resin could be used according to the manufacturer's protocol.

Figure S6 .
Figure S6.MALDI-TOF/TOF tandem MS analysis of the peptide fragment 1 in Figure S5 (the fragment from unlabeled Z-DM).

Figure S7 .
Figure S7.MALDI-TOF/TOF tandem MS analysis of the peptide fragment 2 in Figure S5 (the fragment from labeled Z-DM), demonstrating that K18 was labeled as expected (The spectrum shows the extended mass range of Figure 2D and was included for comparison with the peptide fragment 1).

Figure S12 .
Figure S12.MALDI-TOF MS analysis of Z-M_ncAA, and Z-M_ncAA-PyOx.For Z-M_M8AzK and Z-M_D37AzK, K5 and K33, respectively, were mutated to Arg to prevent the self-labeling of the probes.

Figure S14 .
Figure S14.Comparison of structural hindrance regarding the target lysines for H19, K33, and D37.In the case of K33 and D37, the target lysines remain unobstructed by neighboring amino acids, facilitating direct contact between PyOx and the lysines.In contrast, reaching the target lysine from H19 presents greater difficulty due to structural hindrance from nearby residues.

Figure S16 .
Figure S16.Protease (Trypsin) digestion analysis of WT, FL-labeled, and Bt-labeled Fc fragment.The peptide fragment 3 contains the expected target lysine (K248) and was shifted to show the mass of probe-conjugated forms (peptide fragment 4 and 5).

Figure S18 .
Figure S18.MALDI-TOF/TOF tandem MS analysis of the peptide fragment 4, demonstrating the labeling of biotin at K248.

Figure S19 .
Figure S19.MALDI-TOF/TOF tandem MS analysis of the peptide fragment 5, demonstrating the labeling of FL at K248.(The spectrum shows the extended mass range Figure 3D and was included for comparison with the peptide fragment 3).

Figure S22 .
Figure S22.MALDI-TOF/TOF tandem MS analysis of the peptide fragment 7, demonstrating the labeling of biotin at K248.(The spectrum shows the extended spectrum of Figure 4B and was included for comparison with the peptide fragment 6).

Figure S28 .
Figure S28.MALDI-TOF/TOF tandem MS analysis of peptide fragments 8, confirming the labeling of azide at K248.Particularly, the fragments exhibited the cleavage of azide to form an imine bond and a loss of 28.01 Da, attributed to the high-energy conditions during tandem MS analysis.[8]

Figure S30 .
Figure S30.(A) Western blot of HER2 expression in MDA-MB-231, MDA-MB-453, and SK-BR-3 cells (n = 3).(B) Quantification result of the HER2 western blots relative to those of GAPDH.Bar graphs represent the average and the error bars represent the average ± standard deviation.Statistical analysis was conducted by one-way ANOVA analysis and Tukey's multiple comparison test (n = 3, ** p < 0.01, *** p < 0.001).

Figure S31 .
Figure S31.Binding assay of trastuzumab WT and the ADC product against HER2 using SPR.

Figure S32 .
Figure S32.Binding assay of trastuzumab WT and the ADC product against FcRn/β2m Heterodimer at pH 6.0 using SPR.