Probing the limits of Q-tag bioconjugation of antibodies† †Electronic supplementary information (ESI) available: Methods section including the synthesis of Ab-labelling reagents, Ab conjugation methods, chromatography and native MS. See DOI: 10.1039/c9cc02303h

Precise analyses reveal that, while useful in reducing heterogeneity, the use of TGases in site-selective Ab modification may still create unwanted ‘off-site’ conjugates.


Supplementary Schemes
Scheme S1. Synthesis of DFO-alkyne 5 (A) and Dfo+Il1α-alkyne 6 (B). Figure S1. Structure of DBCO-Cy3 S1 and its conjugation to azido-dg-Her 4.   conjugation of DFO caused non-specific metal uptake and peak broadening. In this case, reasonable desolvation conditions reduced the majority of non-specific adducts leaving only minor peak tailing and allowing semi-quantitative analysis of conjugation products. Five major peaks were found and structures could be assigned to four of these as shown based on relative peak heights of Herceptin-derived products;

Supplementary Figures
(b) nMS analysis was performed 7 months after synthesis. In this case, the standard deviation is higher than in the previous spectrum due to broadening of the peaks; analysis suggested little or no degradation within the limits of estimation. We speculate that the species labelled * may be an impurity in the DFO chelation reagent but is in too low abundance to be more conclusive.
(a)  The conjugation of DFO caused non-specific metal uptake and peak broadening. In this case the nMS data is less well resolved and quantitative analysis is not possible because low abundance species cannot be resolved from poorly desolvated major products. The spectrum exhibits three major peaks, with the base peak B corresponding to the expected product, [DFO·IL1α] 2 -dg-Her, suggesting 8 is the major product of conjugation. In addition, A can be assigned to the singly incorporated IL1α+DFO-alkyne and we postulate that * is a degradation product; (b) Postulated structure for *, which could be a degradation product of 8, formed via a retro-Michael elimination of the thiosuccinimide group (observed mass difference between * and A ≈  Herceptin 1
For NMR experiments, samples were dissolved in DMSO-d6 and analysed on Bruker 400 or 500 MHz spectrometers at room temperature. The chemical shifts, δ, are reported in ppm (parts per million). The residual solvent peaks have been used as an internal reference. The abbreviations for the peak multiplicities are as follows: s (singlet), d (doublet), dd (doublet of doublets), t (triplet), q (quartet), m (multiplet) and br (broad). Other abbreviations used were 'app', standing for apparent and 'ar', standing for aromatic.

Synthesis of antibody labelling reagents 5.1 Synthesis of DFO-alkyne 5
To a stirred solution of DBCO-C 6 -acid SI2 (13.5 mg, 0.04 mmol) in DMF (0.32 mL), were added HATU (23.0 mg, 0.06 mmol) and DIPEA (13.3 µL, 9.80 mg, 0.08 mmol). Separately, deferoxamine (DFO) mesylate salt (26.3 mg, 0.04 mmol) was dissolved in DMF (0.32 mL) and DIPEA (13.3 µL, 9.80 mg, 0.08 mmol) and NMM (30.0 µL) were added. Both solutions were stirred at room temperature for 1 h. During this time, the solution containing DBCO-C 6 -acid SI2 turned orange. After 1 h, the DFO-mesylate salt solution was added to the DBCO-C 6 -acid solution and the reaction mixture stirred for 2 days, at room temperature, under Nitrogen. After this time, ice cold acetone (5 mL) was added and the reaction mixture was sonicated for 5 min. This resulted in a white solid which was separated from the liquid by centrifugation (7000 rpm). The white solid was then washed with acetone (2 x 5 mL) and water (3 x 5 mL), with sonication and centrifugation in between the washes. Finally the resulting white solid was dried under high vacuum to afford 12 mg (0.03 mmol, 69%) of 5. 1

Chromatography Methods
Samples were analysed by HPLC using a separation module (Waters 2695, Watford, UK) equipped with a mass spectrometry detector (Waters micromass ZQ) and a Waters 2996 photodiode array detector.
Compounds were monitored using total ion count in electrospray positive mode from m/z 250-1500 with a cone voltage of 10V or UV absorbance at 215 nm. Mass spectrometry settings were: capillary voltage was set at 3.2 kV, extractor lens 2V, RF lens 0.3V, desolvation gas flow 450 L/h, cone gas flow 90 L/h, source temperature 120 °C, and desolvation temperature 425 °C.

LCMS analysis under reducing conditions
Aliquots (2 µg of antibody) from the reaction mixture were diluted with DTT (20 mM solution) to achieve a Minimum intensity ratios were set at 33% for both left and right. Raw data was not subjected to subtraction, smoothing or centring prior to deconvolution. Note: the faint band above between 35-50 kDa represents residual TGase from the azide incorporation reaction.

Attempted transamidation using reduced equivalents of 3 (Method E).
To a solution of deglycosylated Herceptin 2 in PBS were added polyetherazidoamine 3 (40 equivalents) and TGase enzyme (6 U/mg antibody). The reaction was incubated at 37 °C with shaking (250 rpm for 20 h. After this time, the samples were passed through a 50kDa filter membrane and washed with PBS to remove unreacted azide and the TGase enzyme and analysed by rLCMS. The LCMS indicated the expected product was formed to an extent of ~5%. Decreasing the number of equivalents of azide from 80 equivalents to 40 equivalents resulted in a higher percentage of nonfunctionalized dgHer-2. This indicated that the reaction required a higher equivalents of azide to push it to completion.

Heavy chain Light chain
Heavy chain/2 Azido-PEG functionalized heavy chain

1 Sample Preparation for native mass spectrometry experiments
Samples of antibodies were dialysed overnight against 50 mM ammonium bicarbonate the concentration adjusted to 0.5 mg/mL. Additional desalting was sometimes performed by further buffer exchange into freshly prepared ammonium bicarbonate 50 mM using P6 Biospin columns (Bio-Rad) or Zeba Spin desalting columns (7K or 40K) (Thermo Fisher). Concentration was reduced as necessary for nMS.

High Resolution Native Mass Spectrometry
High resolution native mass spectrometry was performed as described elsewhere. 2 Briefly, ions were introduced into a Q Exactive hybrid quadrupole-Orbitrap (Thermo Fisher, Bremen Germany) mass spectrometer modified for the transmission and detection of high mass ions. 4,5 All spectra were acquired in "Native Mode" with maximum RF applied to all ion optics, -3.2 kV to the central electrode of the Orbitrap and with ion trapping in the HCD cell. Ions were generated in the positive ion mode from a static nanospray source (+1.0 to +1.4 kV) using gold-coated capillaries prepared in-house. 6 Ions then passed through a temperature controlled transfer tube (40-80 o C), RF-lens, injection flatapole and bent flatapole.
After traversing the selection quadrupole, which was operated with a wide selection window (2,000-15,000 m/z), ions were trapped in the HCD cell before being transferred into the C-trap and Orbitrap mass analyser for detection. Transient times were 64 ms and AGC target was 1×10 6 . Spectra were acquired with either 1 or 10 microscans, averaged over 50-100 scans and with a noise level parameter set to 3, slightly lower than the default of 4.68. Efficient desolvation of intact antibodies was achieved through increased voltages applied in the HCD cell (150-200 V). No in-source activation was applied. The collision gas was Argon and pressure in the HCD cell was maintained to achieve a UHV pressure of approximately 1×10 -9 mbar. Data was processed using Thermo Scientific Xcalibur 2.1 and masses and S.D. calculated using in-house software (http://benesch.chem.ox.ac.uk/resources.html) using the three most abundant charge states. The reaction product abundances were calculated (where possible) based on the peak intensities on the three most abundant charge states as described previously 2 .
When performing nMS analysis on modified antibodies conditions were kept as "gentle" as possible to avoid dissociation of conjugated moieties as previously described 2 . The resulting peaks were often slightly asymmetric (with a tail to the right hand side) and peak widths were consistently around twice that expected purely from the isotopic distribution. This peak broadening is likely due to complexation with salts and other small molecule adducts that have not been completely removed during desolvation. 7 This incomplete desolvation likely explains the consistently higher measured masses we report compared to the theoretical mass. It may also be responsible to the slightly elevated standard deviation in measured mass (as adduct retention is likely to be charge state dependant, and thus increase S.D.). In the case of highly conjugated antibody constructs such as 7 and more so with 8, desolvation was particularly challenging. This is likely due to the presence of the DFO moiety which may strongly chelate various metal ions such as Fe.
Methods to achieve more exhaustive salt removal remain the subject of ongoing work.