Natural Glycoforms of Human Interleukin 6 Show Atypical Plasma Clearance

Abstract A library of glycoforms of human interleukin 6 (IL‐6) comprising complex and mannosidic N‐glycans was generated by semisynthesis. The three segments were connected by sequential native chemical ligation followed by two‐step refolding. The central glycopeptide segments were assembled by pseudoproline‐assisted Lansbury aspartylation and subsequent enzymatic elongation of complex N‐glycans. Nine IL‐6 glycoforms were synthesized, seven of which were evaluated for in vivo plasma clearance in rats and compared to non‐glycosylated recombinant IL‐6 from E. coli. Each IL‐6 glycoform was tested in three animals and reproducibly showed individual serum clearances depending on the structure of the N‐glycan. The clearance rates were atypical, since the 2,6‐sialylated glycoforms of IL‐6 cleared faster than the corresponding asialo IL‐6 with terminal galactoses. Compared to non‐glycosylated IL‐6 the plasma clearance of IL‐6 glycoforms was delayed in the presence of larger and multibranched N‐glycans in most cases

After two hours the solution was concentrated in vacuo. The glycopeptide B2 was precipitated by addition of 20 mL of cold diethyl ether, collected by centrifugation, washed with cold diethyl ether (2x 45 mL) and dried in high vacuum. The residue was dissolved in 20 % acetonitrile and purified by RP-HPLC (YMC-Pack C8, 250 x 20 mm, 5 µm, gradient from 5 to 25 % acetonitrile/water, 0.1 % formic acid).
3.12 mL of 300 mM aqueous NaBrO3 (936 µmol, 15 eq) and subsequently 7.06 mL of 130 mM aqueous Na2S2O4 (780 µmol, 12.5 eq) were added slowly under stirring. The reaction was vigorously stirred at ambient temperature until completion (16 h, tlc: CH2Cl2/acetone, 2:1, Rf = 0.48) and 100 µl of 10 % aqueous Na2S2O3 were added. The mixture was diluted with CH2Cl2 and extracted with H2O. The organic phase was dried with MgSO4, concentrated in vacuo and dried in high vacuum. The residue was suspended in 6.24 mL of n-butanol and 1.67 mL (25 mmol, 401 eq) of ethylenediamine were added. Subsequently, the mixture was stirred at 90 °C for 12 h. After completion (tlc: isopropyl alcohol/1 M ammonium acetate, 2:1 Rf = 0.38) the reaction was cooled to ambient temperature and the volatiles were evaporated in vacuo. Residual reagents were removed by addition of toluene and azeotropic distillation. The residue was dried in high vacuum and subsequently dissolved in a mixture of 2.77 mL of acetic anhydride and 5.55 mL of pyridine and stirred at ambient temperature until complete conversion (tlc: isopropyl alcohol/1 M ammonium acetate, 2:1 Rf = 0.88). The volatiles were evaporated in vacuo and residual reagents were removed by addition of toluene and azeotropic distillation (8 x). The dried residue was dissolved in CH2Cl2, extracted with 1 M HCl and 2 M KHCO3, dried with MgSO4, concentrated in vacuo and dried in high vacuum.
6.4 mg (3.66 µmol, 1 eq) of glycosyl azide 12 were dissolved in absolute methanol (200 µL) and 6.25 µL (36.5 µmol, 10 eq) of DIPEA and 22 µL (219 µmol, 60 eq) of 1,3-propanedithiol were added. After 1.5 h 100 µl of absolute methanol were added. After a total reaction time of 3.5 h the solution was concentrated in vacuo. The residue was dissolved in 720 µl of absolute methanol and the glycosylamine G7 was precipitated with 2.2 mL of cold diethyl ether and centrifuged. This procedure was repeated two more times. The pellet was dried in high vacuum.  4.0 mg (1.67 µmol, 1 eq) of glycosyl azide 13 [7] were dissolved in 186 µL of absolute methanol and 8.6 µL (50.2 µmol, 30 eq) of DIPEA and 30.4 µL (30.8 µmol, 180 eq) of 1,3propanedithiol were added. After 5 h the resulting glycosylamine G8 was precipitated with 1 mL of cold diethyl ether and collected by centrifugation. The residue was dissolved in 300 µL methanol and precipitated with 1 mL of cold diethyl ether. This procedure was repeated two more times. The pellet was dried in high vacuum.  11. Synthesis of IL-6 (43-48) tetraantennary 2,6-sialylated glycopeptide B9 Figure S21: Synthesis of B9.
B9 was prepared by enzymatic elongation of B8. 3.07 mg (0.98 µmol, 1.0 eq) of B8 and 5.8 mg (9.11 µmol, 9.3                     was purified over a HiLoad 16/60 Superdex 75 pg column (flow rate: 2 mL/min, refolding buffer). The fractions of monomeric IL-6 3 (ca. 10 mL per 60 µL aliquot) were pooled and IL-6 3 was concentrated by ultrafiltration as above.             . The fractions of monomeric IL-6 6 (ca. 10 mL per 60 µL aliquot) were pooled and IL-6 6 was concentrated by ultrafiltration as above.  The reactions were performed in an anaerobic chamber.     The reactions were performed in an anaerobic chamber.      The reactions were performed in an anaerobic chamber.

CD spectroscopy
An aliquot of the final solutions of the IL-6 glycoforms (50 µL) was exchanged to 10 mM sodium phosphate pH 7.4 (NAP-10 columns, GE Healthcare, Germany; procedure as suggested by the manufacturer). The CD spectra were recorded in a Hellma 165-QS cuvette (1.0 mm, 160 µL) with a J-600 spectropolarimeter (Jasco, Germany).

Validation of ELISA test for IL-6 glycoforms
For the detection of the glycoforms IL-6 1-9 and IL-6 from E. coli we used a commercial research ELISA sandwich test specific for IL-6 in human serum (ImmunoTools, Friesoythe, Germany). An initial quantification of IL-6 glycoforms was performed as follows: A solution containing 250 pg of each glycoform was subjected to the assay according to the manufacturer's instructions.

Analysis of IL-6 plasma clearance in vivo
Rats which have been equipped with two vascular catheters were obtained from Charles River Laboratories. Rats (250-275 g body weight) were housed under standard conditions with food and drinking water ad libitum and held under a 12h light-dark cycle. Animals had one week to adjust after arrival before the experiment started.
The IL-6 glycoforms were shipped as solutions in the buffer used for the final gel filtration.
The concentration of each glycoform was determined by absorption at 280 nm as specified in the supporting information. The solutions for injection were prepared by diluting the stocks with PBS to a final concentration of 32 µg of IL-6 glycoform/mL. Aliquots of 250 µL were made and kept at -24 °C until use.
Rats were injected with one of the IL-6 glycoforms (8 µg of IL-6 in 250 µl of sterile PBS) in one catheter and blood samples were obtained from the second catheter 0, 1, 2, 5, 10, 15, and 20 minutes after injection of IL-6. Serum was obtained from the coagulated blood samples via centrifugation, and serum samples were stored until measurement at -20°C. IL-6 was determined via a human IL-6 specific ELISA at appropriate dilution according to the manufacturer's instructions (ImmunoTools, Friesoythe, Germany). 24h after the IL-6 injection, the final blood sample for the determination of acute phase proteins was drawn, which was treated as the other blood samples described above. One week after the first experiment, rats received a different IL-6 glycoform and were treated according to the same protocol. Rats were used 3 to 4 times, depending on the patency rate of the catheter. After the last experiments, rats were sacrificed and their livers harvested for determination of induction of genes encoding acute phase response proteins. All animal experiments were approved by the local authorities (Ministerium für Energiewende, Landwirtschaft, Umwelt und ländliche Räume, Kiel, Germany, V 242-7224.121-3 (47-4/15)).

Quantitative Real-Time PCR (qPCR)
Rat livers were excised, snap-frozen in liquid nitrogen and total RNA was isolated using the Nucleospin RNA II kit (Macherey-Nagel, Düren, Germany) according to the manufacturer's instructions. 1 µg of the total RNA was reversely transcribed with oligo-dT15 primers using the RevertAid reverse transcriptase (Thermo Fisher Scientific, Waltham, MA, USA). The amount of cDNA derived from 50 ng of RNA was used for each qPCR reaction using the Power SYBR Green PCR Master Mix (Thermo Fisher Scientific, Waltham, MA, USA) according to manufacturer's instructions on a LightCycler® 480 System (Roche Applied Systems, Penzberg, Germany). One sample from each of the eight rats was measured in duplicates and normalized to GAPDH expression. Relative gene expression was calculated using the 2 -∆Ct method. The following primers were used:

Quantification of rat acute phase proteins
The levels of the acute phase proteins (rat C-reactive protein and rat alpha1-acid glycoprotein) in the sera of rats injected with the different IL-6 variants were detected using specific ELISA Kits (Abcam, Berlin, Germany), which were used according to the manufacturer's instructions.

Figure S 81: Determination of serum concentrations of a) rat C-reactive protein and b) rat
alpha1-acid glycoprotein before (0h) and after (24h) administration of specific IL-6 variants.
The increase of serum concentration for C-reactive protein was statistically not significant whereas for alpha1-acid glycoprotein an increase was found (Mann-Whitney-U test).

Biolayer interferometry (BLI)
The Octet RED96 system (ForteBio, Pall Life Science, Fremont, CA, USA) was employed for binding kinetic measurements. All steps were performed in a final volume of 200 μl at 25° C and 1000 rpm agitation. The ligand sGP130Fc (IL-6 Coreceptor GP130-Fc Fusion) [10] was immobilized on anti-human Fc biosensors (ForteBio) at 5 μg/mL in PBS (Sigma Aldrich) for 180 s. Subsequently, the tips were rinsed in PBS for 45 s. The association of the three IL-6/sIL-6R complexes (different concentrations in PBS, see Fig. S82) was measured for 300 s followed by dissociation for 3600 s (plain PBS). In each experiment a reference control was measured by incubating the captured sGP130Fc with PBS. Data fitting and analysis was performed with ForteBio data analysis software using a 1:1 model after Savitzky-Golay filtering.

Crystallization and X-ray structure determination
Synthetic IL-6 1 (GlcNAc) [1] was used at concentrations of 15 and 7.5 mg/ml for the screening of initial crystallization conditions by using the commercial Qiagen screen, JCSG+. Equal volumes of 300 nl of protein and screening solutions were pipette by a Phoenix liquid handling robot (Art Robbins) into 96-well MRC-2 crystallization plates (Molecular Dimensions) and then stored at 293 K, in a RockMaker-1000 imaging system (Formulatrix), and at 277 K. Initial 3D crystals appeared within days to weeks only at 277 K in a condition consisting of 0.24 M sodium malonate, pH 7.0 and 20% (w/v) PEG 3350. After optimization of the initial condition in 48-well MRC crystallization plates, a larger 3D crystal was transferred to a cryo-protectant solution of 0.2 M sodium malonate, pH 7.2, 18% (w/v) PEG 3350 and 25% (v/v) glycerol, flash-frozen and stored in liquid nitrogen. Diffraction measurements were carried out at the MX-14.1 beamline at Helmholtz-Zentrum Berlin. A complete data set was recorded, processed and scaled with with XDSAPP3 [12] up to 2.0 Å resolution. Phase determination was performed by Molecular Replacement with PHASER [13] using 1ALU (PDB ID) as a start model. Model building and refinement were performed by REFMAC [14] and COOT [15] . Data processing, refinement and model statistics are given in Table S1. Figure S84: gray) and mFo-DFc residual (2.25 r.m.s.d., orange) densities of the final model of IL-6 1 at the ASN-44 glycosylation site.
Model and density calculations did not contain any sugar atoms. Both densities indicate a covalent extension of the Asn side chain.