PEGylation and Dimerization of Expressed Proteins under Near Equimolar Conditions with Potassium 2-Pyridyl Acyltrifluoroborates.

The covalent conjugation of large, functionalized molecules remains a frontier in synthetic chemistry, as it requires rapid, chemoselective reactions. The potassium acyltrifluoroborate (KAT)-hydroxylamine amide-forming ligation shows promise for conjugations of biomolecules under aqueous, acidic conditions, but the variants reported to date are not suited to ligations at micromolar concentrations. We now report that 2-pyridyl KATs display significantly enhanced ligation kinetics over their aryl counterparts. Following their facile, one-step incorporation onto the termini of polyethylene glycol (PEG) chains, we show that 2-pyridyl KATs can be applied to the construction of protein-polymer conjugates in excellent (>95%) yield. Four distinct expressed, folded proteins equipped with a hydroxylamine could be PEGylated with 2-20 kDa 2-pyridyl mPEG KATs in high yield and with near-equimolar amounts of coupling partners. Furthermore, the use of a bis 2-pyridyl PEG KAT enables the covalent homodimerization of proteins with good conversion. The 2-pyridyl KAT ligation offers an effective alternative to conventional protein-polymer conjugation by operating under aqueous acidic conditions well suited for the handling of folded proteins.

Introduction

Chemical methods for the conjugation of large molecules (>10,000 MW), including polymers to proteins, present a continuing challenge in organic synthesis.[1,2] As noted by many researchers, the union of large molecules with near stoichiometric amounts of coupling partners poses a particular problem for organic chemistry, as it requires not only chemoselective reactions but also very fast second order kinetics.[3−5] This constraint arises from the difficulty in obtaining concentrated solutions of larger molecules, due to their high molecular mass and limited solubility. While most organic reactions operate optimally at around 200 mM, even small proteins are often soluble only at 20 μM or less: a 10,000-fold difference in concentration between small molecule and protein conjugation chemistry.

As part of our efforts to develop new, chemoselective amide-forming ligations for the conjugation of large molecules, we now document the successful application of the potassium acyltrifluoroborate[6−8] (KAT) ligation[9,10] to protein PEGylation and dimerization at near equimolar stoichiometry. Key to the success of this project was the identification of 2-pyridyl KAT reagents,[11] which display enhanced reaction rates[11] and are easily introduced onto commercial PEG reagents. These studies establish the KAT ligation as suitable for the conjugation of PEG–KATs to proteins bearing a hydroxylamine within hours at micromolar concentrations, without any nonspecific reactions or interactions. Notably, this process proceeds with folded proteins under aqueous, acidic conditions that complement the neutral or basic conditions typically employed for protein PEGylation.

Experimental Design

Background

Very few chemoselective, intermolecular reactions are fast enough to conduct site-specific protein–polymer conjugations with near equimolar amounts of reagents. The most successful and widely used chemistry for this “grafting to” approach[12,13] is the venerable addition of thiols[14]—usually in the form of cysteine side chains—to polymers bearing maleimide groups. A majority of commercially approved PEGylated therapeutic proteins[15,16] and antibody–drug conjugates[17,18] have been prepared using this chemistry. Its widespread use stems from its exceptionally fast reaction rate,[5] making it well-suited for protein modification. However, its limitations include the reversibility of the conjugate addition,[19] the formation of stereoisomers, and hydrolysis of the succinimide ring to give regioisomeric products.[20,21] Other “grafting to”[22,23] approaches to site-specific protein–PEGylation include the development of specialized PEG reagents that readily participate in other bioorthogonal reactions, such as oxime ligations, copper-catalyzed azide alkyne cycloadditions (CuAAC), strain-promoted azide alkyne cycloadditions (SPAAC), palladium-catalyzed cross couplings, Staudinger-phosphite ligations, and tetrazine-trans-cyclooctene ligations.[24]

Although not directly applicable to PEGylation, “grafting from”[25] methods for the preparation of protein–polymer conjugates have recently emerged as an alternative approach. These processes typically begin with the introduction of a radical initiator onto a protein side chain followed by RAFT or ATRP with olefin monomers. However, limitations of these methods include the use of additional initiator reagents, the requirement for thoroughly deoxygenated solvents/buffers, and the formation of hetereogeneous polymer dispersities.[26]

Implementation

At the outset of our own studies, we focused on the specific problem of identifying reaction conditions and PEG-KAT reagents suitable for conjugation to a single hydroxylamine on a protein. Our primary goal was to execute the KAT-mediated PEGylation of folded proteins under near stoichiometric reaction conditions. Provided these studies were successful, we also sought to apply this chemistry to protein dimerization, as the requisite multivalent PEG-KATs are easily prepared in one step.[10,27]

We envisioned four approaches to incorporating a hydroxylamine into a target protein: (1) chemical modification[28−30] of a pre-existing side chain, such as a solvent exposed cysteine residue;[31,32] (2) protein expression with unnatural amino acids,[33−36] such as pyrrolysine surrogates;[37] (3) site-specific modification (i.e., oxidation or acylation) of an N-terminal residue[38] or an internal lysine residue;[39] or (4) chemical protein synthesis with a preinstalled hydroxylamine side chain.[10] Preliminary results showed all of these approaches to hydroxylamine incorporation to be possible, but this step will be relevant only if the PEGylation or other conjugation onto the protein hydroxylamine confers advantages over other strategies. We elected to adopt established approaches of cysteine modification to generate protein hydroxylamines for the initial conjugation studies, although our long-term efforts favor other methods. Related site-specific modification[40−42] of other side chains, including tyrosine,[43−48] tryptophan,[49] and methionine,[50] is becoming increasingly common, and one could easily imagine preparing suitable reagents for incorporating hydroxylamines at these sites.

Experimental Results

Preparation of Protein Hydroxylamines

As a model protein for initial studies, we selected a mutant (S147C) of superfolder green fluorescent protein (sfGFP),[51] bearing a single accessible thiol that has been shown to react selectively[52] with various thiophilic reagents.[53] This S147C mutant was readily expressed from Escherichia coli host DH10B and isolated in high purity following sequential Ni-affinity and anion exchange chromatography purifications. We initially prepared a hydroxylamine–malemide reagent for protein modification, and successfully used this reagent for protein PEGylation, but found that the studies were complicated by byproducts and retroadditions. The inherent two-step nature of our approach allowed us to consider more stable cysteine adducts. We selected methylsulfonephenyl-oxadiazoles, which have been shown by Barbas and co-workers to react chemoselectively with cysteine residues in proteins.[54,55] The resulting adducts possess appreciable serum stability and are resistant to both hydrolysis and trans-thioetherification. Bifunctional methylsulfonephenyl-oxadiazole–hydroxylamine 4 was prepared in 3 steps from N-Boc protected 3-bromopropyl-O-carbamoyl hydroxylamine 2 (Scheme A).

Scheme 1

Synthesis of (A) Bifunctional Reagent 4 and (B) Its Site-Selective Labeling of sfGFP-(S147C)

Reagents and conditions: (a) 1 (1.0 equiv), 2 (1.1 equiv), K2CO3 (1.1 equiv), DMF, 18 h, 23 °C (94%). (b) mCPBA (3.0 equiv), CH2Cl2, 18 h, 0 to 23 °C. (c) TFA:CH2Cl2 (1:2.3), 5 h, 0 °C (63% over 2 steps). (d) 5 (1.3 mg/mL, 50 μM), 4 (10 equiv), 0.1 M K-Phos pH 7.8, 2 h, 0 °C. Boc = tert-butoxycarbonyl, mCPBA = meta-chloroperoxybenzoic acid, TFA = trifluoroacetic acid, DMF = N,N-dimethylformamide.

To introduce hydroxylamine 4 onto sfGFP-(S147C) 5, we pretreated 5 with DTT to reduce any disulfide homodimers. After removal of excess DTT, 5 was alkylated with 10 equiv of 4 in a potassium phosphate (K-Phos) buffer at pH 7.8. Characterization by electrospray ionization mass spectrometry (ESI-MS) showed full conversion to the hydroxylamine-tagged bioconjugate 6 (Scheme B). Excess 4 was conveniently removed by spin diafiltration, and the modified protein was stable to storage for several days at room temperature in a variety of buffers, including phosphate, HEPES, Tris, and glycine.

Protein PEGylation with PEG-KAT Reagents

We have previously reported the synthesis of 10 kDa, methyl-capped PEG chain (mPEG) p-phenyl KAT 7(10) and its application in peptide PEGylation, albeit at a concentration of 1 mM. We utilized this reagent for our initial attempts at protein PEGylation, with the anticipation that the PEGylated product 8 should be readily separable—and therefore quantifiable—from unreacted protein hydroxylamine 6 by both SDS polyacrylamide gel electrophoresis (SDS–PAGE) and fast protein liquid chromatography (FPLC).[56,57] When 6 was treated with an excess of the 10 kDa KAT reagent 7 (4.0 equiv) at 25 μM in a K-Phos buffer[58] at pH 6.0, we observed very poor ligation efficiency[59] (8%) after 12 h, with little conversion to the desired PEGylated product 8. Increasing the amount of 7 to 10 equiv and the reaction time to 24 h resulted in only a slight improvement (Table , entries 2 and 3). In order to improve the ligation efficiency, we sought to use a lower pH that should still be suitable for most folded proteins. Glycine (Gly) is a common bulking agent used in the formulation of therapeutic proteins, has been shown to effectively stabilize folded proteins via the preferential exclusion mechanism,[60,61] and is recognized as a solubilizing agent for proteins under acidic conditions (effective pH range: 2.2–3.6).[62] We initially had some concern about the combination of KAT reagents and a large excess of Gly, as they can undergo imine formation.[63] However, upon performing the ligation of 6 with 4.0 equiv of 7 in a Gly-HCl buffer at pH 3.6,[64,65] we observed an increase in ligation efficiency (61%) after 12 h (Table , entry 4).

Table 1

Screening and Optimization of Reaction Conditions for the PEGylation of sfGFP-(S147C) with 10 kDa mPEG-KATs 7 and 9ca

entry	X	KAT equiv	bufferb	pH	concn (μM)	time (h)	ligation efficc (%)
1	CH	4.0	K-Phos	6.0	25	12	8
2	CH	10	K-Phos	6.0	25	12	17
3	CH	10	K-Phos	6.0	25	24	35
4	CH	4.0	Gly-HCl	3.6	25	12	61
5	N	4.0	Gly-HCl	3.6	25	0.5	89
6	N	4.0	Gly-HCl	3.6	25	1.0	97
7	N	1.5	Gly-HCl	3.6	25	12	96
8	N	1.5	Gly-HCl	3.6	25	1.0	75
9	N	1.5	Gly-HCl	3.6	25	5.0	96
10	N	1.5	citrate	3.6	25	1.0	57
11	N	1.5	citrate	3.6	25	12	92
12	N	4.0	K-Phos	6.0	25	1.0	17
13	N	4.0	K-Phos	6.0	25	12	78
14	N	4.0	histidine	3.6	25	12	65
15	N	4.0	Gly-HCl	3.6	5.0	12	97
16	N	1.5	Gly-HCl	3.6	5.0	12	74
17	N	1.1	Gly-HCl	3.6	25	12	90

Ligations were performed at 23 °C.

Buffers were formulated to a concentration of 50 mM and contained 50 mM KF along with DMF (5% by volume).

Ligation efficiencies were estimated from the FPLC traces as a ratio of the peak area of the PEGylated product over areas of all peaks at 490 nm.

Although encouraging, these results were still far from ideal, especially with the goal of carrying out protein dimerization or PEGylation with near equimolar amounts of coupling partners. Earlier work from our group identified alternative acyl boronates that display faster ligation kinetics at higher pH.[66] Unfortunately, these reagents—such as acyl MIDA compounds[67]—are unstable in aqueous environments and are not suitable for equimolar reaction conditions. Parallel studies have identified marked substrate effects in the KAT ligation, with 2-pyridyl KATs emerging as significantly faster ligation partners at both lower and higher pH.[11] To take advantage of this finding, we required the corresponding 2-pyridyl mPEG-KAT reagent 9, which was synthesized in good yield by a nucleophilic aromatic substitution (SNAR) reaction of mPEG 10 with an excess of 5-fluoropyridin-2-yl KAT 11 (Scheme ). The same procedure was used to prepare 2-pyridyl mPEG-KATs in varying size (9a–d) and valency (12) from the corresponding mPEG or PEG diol 10.

Scheme 2

Synthesis of 2-Pyridyl mPEG-KATs 9a–d and Bis 2-Pyridyl PEG KAT 12

Reagents and conditions: 10 (1.0 equiv), 11 (3.0 equiv for 9, 6.0 equiv for 12), NaH (2.0 equiv for 9, 4.0 equiv for 12), DMF, 12 h, 23 °C.

With 9c in hand, we revisited the 10 kDa PEGylation of 6. Upon treatment with 4.0 equiv of 9c in a Gly-HCl buffer at pH 3.6, we observed exceptional ligation efficiency (89%) after only 30 min and essentially full conversion to 13c after 1.0 h; no unreacted protein was detected by FPLC (Table , entries 5 and 6). We could also use lower ratios of coupling partners 6 and 9c: with only 1.5 equiv of 9c we observed 75% conversion within 1 h (entry 8) and essentially full conversion to 13c after 5 h (Table , entry 9). Using 4.0 equiv of 9c, we could raise the pH to 6.0 in a K-Phos buffer. Although the ligation was not complete after 1 h (17%, entry 12), good conversion was obtained after 12 h (78%, entry 13). These observations are consistent with a second order rate constant k2 of 14 M–1 s–1 at pH 3.6 and 0.50 M–1 s–1 at pH 6.0. To put this into context, other relevant bioorthogonal click reactions with comparable rates include strain-promoted 1,3-dipolar cycloadditions of azides with biarylazacyclooctynones (k = 0.9–1.6 M–1 s–1),[68] 1,2-aminothiol-2-cyanobenzothiazole ligations (k2 = 2.7–59.7 M–1 s–1),[69] the hydrazino-Pictet Spengler ligation (k2 = 1–40 M–1 s–1),[70] strain-promoted alkyne–nitrone cycloadditions (SPANCs) (k2 = 1.5–39 M–1 s–1),[71,72] and the inverse electron-demand Diels–Alder reaction of tetrazines with norbornenes (k2 = 0.12–8.0 M–1 s–1).[73]

To further evaluate the effect of the reaction medium on the ligation efficiency, we tested citrate as an alternative buffer under otherwise identical conditions (pH 3.6, 25 μM protein concentration). This system gave slightly lower, but still overall good, conversions (Table , entries 10 and 11). This suggests that Gly may help overcome some nonspecific interactions between the PEG KAT and lysine side chains. In contrast, when the PEGylation of 6 with 9c was carried out in a histidine buffer at pH 6.0, a diminished conversion to 13c was observed in comparison to the same reaction carried out in K-Phos buffer (Table , entry 14). These results suggest that, at higher pH ranges, amino acid based buffers may be detrimental to the efficiency of KAT ligations, possibly due to substrate deactivation through reversible imine condensations.

Utilizing our optimized conditions, we tested the ligation of 6 and 9c at higher dilution and were pleased to observe essentially full conversion to 13c with 4.0 equiv of 9c at a concentration of 5 μM, and a 74% ligation efficiency with 1.5 equiv of 9c after 12 h (Table , entries 15 and 16). When only 1.1 equiv of 9c was employed, we could still observe an impressive conversion to 13c (90%) under our optimized conditions (Table , entry 17).

Protein PEGylation with Various KAT Reagents

We investigated the conjugation of sfGFP-(S147C) with PEG chains of varying size including 2, 5, and 20 kDa mPEG (see Scheme for syntheses). We selected two of our optimal conditions for these experiments: condition A, 4.0 equiv of mPEG KAT 9, 1.0 h reaction time (Table , entry 6), and condition B, 1.5 equiv of 9, 5.0 h reaction time (Table , entry 9). The 2-pyridyl mPEG KATs 9a–d were applied to the PEGylation of 6 under both conditions, which were analyzed by FPLC and SDS–PAGE (Scheme ). Exposure of 6 to our 2, 5, 10, and 20 kDa 2-pyridyl mPEG KATs 9a–d under conditions A and B (25 μM) gave, in all eight cases, excellent conversions (>90%) to the corresponding monoPEGylated conjugates 13a–d. Upon purification, the identities of 13a–d were verified by matrix-assisted laser desorption–ionization time-of-flight (MALDI-TOF) mass spectrometry.

Scheme 3

Site-Specific MonoPEGylation of sfGFP-(S147C) with 2–20 kDa 2-Pyridyl mPEG KATs

Conditions: 6 (1.3 mg/mL, 25 μM), 50 mM Gly-HCl, 50 mM KF, DMF (5%), pH 3.6, 23 °C. Condition A: 9a–d (4.0 equiv) 1.0 h. Condition B: 9a–d (1.5 equiv), 5.0 h. SDS–PAGE traces are shown for each reaction, with the unpurified reactions for each condition and mPEG used shown on the left and the purified products shown on the right.

Conversions were calculated by gel densitometric analysis.

Ligation efficiencies were estimated from the crude FPLC chromatograms though peak integration fractionations at 490 nm.

PEGylation of Other Protein Targets

To test the generality of the PEGylation, we selected three other protein targets that could be equipped with a hydroxylamine using bifunctional reagent 4. The ubiquitin-conjugating protein Ubc9 is involved in the conjugation of small ubiquitin-like modifier (SUMO) onto various proteins through their lysine residues.[74] This 20 kDa E2 ligase contains 4 cysteine residues, one of which, Cys93, is involved in the enzyme’s active site enabling SUMOylation through a glycyl thioester intermediate. As part of an unrelated project, we expressed a C93A mutant. Examination of the crystal structure revealed that Cys43 and Cys75 were buried within the folded protein, leaving a single, solvent-accessible cysteine (Cys138).[75] Expressed Ubc9-(C93A) was treated with 4 in K-Phos buffer (10 equiv, pH 7.8); after 2.0 h we observed quantitative conversion by ESI-MS to give O-carbamoyl hydroxylamine functionalized bioconjugate 14. Exposure of 14 to both our 10 and 20 kDa 2-pyridyl mPEG KATs 9c-d under the optimized conditions A and B (25 μM) gave, in all four cases, excellent conversions (>90%) to the corresponding monoPEGylated bioconjugates 15c-d as determined by SDS–PAGE (Scheme a). Upon purification by cation exchange chromatography, the identities of 15c-d were verified by MALDI-TOF mass spectrometry.

Scheme 4

Site-Specific MonoPEGylations of Additional Protein Substrates

Ligations were carried out in a 50 mM Gly-HCl pH 3.6 buffer with 50 mM KF and DMF (5%) at 23 °C.

Conversions were calculated by gel densitometric analysis. (a) 10 and 20 kDa PEGylation of Ubc9-(C93A). (b) 10 and 20 kDa PEGylation of T4L-(V131C). Condition A: 9c-d (4.0 equiv) 1.0 h. Condition B: 9c-d (1.5 equiv), 5.0 h. (c) 5.0 kDa PEGylation of HTP-(S261C) at 10 μM. Lane 1: HTP-(S261C). Lanes 2–5: unpurified reactions after 1.0, 2.0, 4.0, and 6.0 h, respectively. SDS–PAGE traces are shown for each reaction with the unpurified reactions for each condition and mPEG used shown on the left and the purified products shown on the right for (a) and (b).

The previously reported V131C mutant of T4 lysozyme (T4L)[76] contains a single, solvent-exposed cysteine residue that has been shown to readily react with both small molecules[77] and polymeric reagents[78] bearing maleimide end groups. The mutant T4L-(V131C) was obtained following subcloning of the gene encoding the mutant protein into a pET 28b-(+) vector, expression from E. coli host BL21 (DE3), and purification by cation exchange chromatography. T4L-(V131C) was incubated with 4 (10 equiv, pH 6.5) to quantitatively yield the O-carbamoyl hydroxylamine bioconjugate 16. This adduct was treated with 9c-d under the optimized conditions A and B. In all four cases, we were pleased to observe near quantitative conversions (>96%) to the corresponding monoPEGylated bioconjugates 17c-d by SDS–PAGE (Scheme b). Upon purification by cation exchange chromatography, the identities of 17c-d were verified by MALDI-TOF mass spectrometry.

As a fourth protein target, we expressed a mutant of the 48 kDa enzyme human thymidine phosphorylase (HTP). HTP plays a primary catabolic role in the metabolism of pyrimidine nucleosides,[79] and its overexpression and elevated activity have been found in the plasma of cancer patients[80] and in solid tumors linked to poor prognosis.[81] Through site-directed mutagenesis, an S261C mutation was found to create a single, solvent exposed Cys residue that readily reacts with thiophilic electrophiles. Following transformation of a pYSBLIC plasmid encoding the gene for the mutant into Rosetta-(DE3)-pLysS, HTP-(S261C) was expressed[82] and purified through sequential Ni-affinity and anion exchange chromatography. HTP-(S261C) was treated with 4 (10 equiv, pH 7.8) to yield the corresponding O-carbamoyl hydroxylamine conjugate 18, quantitatively by ESI-MS. Upon treatment of 18 with the 5.0 kDa 2-pyridyl mPEG KAT reagent 9b (4.0 equiv) under significantly dilute conditions (10 μM) in a Gly-HCl pH 3.6 buffer, we were pleased to observe formation of the corresponding monoPEGylated bioconjugate 19b in 82% conversion by SDS–PAGE[83] (Scheme c).

Protein Dimerization

A particularly challenging area of chemical protein modification is the formation of homo- and heterodimeric proteins. Protein dimerization is ubiquitous in biology and plays a vital role in signal transduction, gene expression, and enzymatic regulation.[84,85] The ability to control this phenomenon has widespread applications in various fields of research by enabling the manipulation of physical structure and regulatory function in organisms.[86,87] Dimerization creates avidity, resulting in a higher local concentration, and often in greater biological activities as compared to the corresponding monomeric protein.[88] There are few synthetic methods for the effective construction of dimeric protein–protein conjugates, and fast, site-specific technologies are imperative to their chemical synthesis. Significant advances in this area include the works of Maynard,[89,90] Francis,[91] Bertozzi, Nogales and Rabuka,[92] Kluger,[93] and Thordarson.[94]

Based on the outstanding results of protein PEGylation with KAT reagents under near stoichiometric conditions, we envisioned that a KAT-mediated homodimerization of a protein could be realized, utilizing one equivalent of a bis 2-pyridyl PEG KAT linker and two or more equivalents of protein hydroxylamine. The linker should be of sufficient length as to avoid steric and electrostatic clashes.[94] We prepared a 3.4 kDa bis 2-pyridyl PEG KAT 12 in an analogous fashion to our 2-pyridyl mPEG KAT reagents 9a–d (Scheme ).

We initially investigated the homodimerization of T4L-(V131C) with linker 12. One equivalent of 12 was treated with 2.0 equiv of freshly prepared O-carbamoyl hydroxylamine-tagged T4L 16 in a Gly-HCl buffer at pH 3.6 (Scheme a). After 30 min at a protein concentration of 50 μM, we observed the formation of two new higher molecular weight bands by SDS–PAGE (Scheme a, lane 2), which we attributed to monoligated T4L-PEG intermediate (middle) and homodimer 20 (upper), with the latter being formed in 52% conversion by gel densitometric analysis. By increasing the reaction time to 5 h, we observed that 20 was formed in 72% conversion (Scheme a, lane 4); prolonging the reaction time to 24 h showed no significant increase in conversion (not shown). By increasing the reaction concentration to 100 μM we could further improve the conversion to 82% (Scheme a, lane 5). Through a combination of sequential size exclusion and cation exchange chromatography, homodimer 20 was purified and its identity was verified by MALDI-TOF mass spectrometry (Scheme a, lower right).

Scheme 5

Homodimerization of (a) T4L-(V131C) and (b) sfGFP-(S147C) with Bis 2-Pyridyl PEG KAT 12

Reactions were carried out at 23 °C.

Concentrations (μM) of each reaction are given with respect to 16 or 6.

Conversions were calculated by gel densitometric analysis. (a) T4L-(V131C) dimerization. Lane 1: T4L-(V131C). Lanes 2–4: unpurified reactions after 0.5, 1.0, and 5.0 h, respectively, at a concentration of 50 μM. Lane 5: unpurified reaction after 5.0 h at a concentration of 100 μM. Lane 6: purified 20. Lower right: MALDI-TOF spectrum of 20. (b) sfGFP-(S147C) dimerization. Lane 1: sfGFP-(S147C). Lanes 2–4: unpurified reactions after 0.5, 1.0, and 5.0 h, respectively, at a concentration of 50 μM. Lane 5: unpurified reaction after 5.0 h at a concentration of 100 μM.

Following our success with the homodimerization of T4L-(V131C), we investigated the dimerization of sfGFP-(S147C) with linker 12. Under identical conditions for the T4L dimerization, 1.0 equiv of 12 was treated with 2.0 equiv of freshly prepared 6 in a Gly-HCl buffer at pH 3.6. After only 30 min at a protein concentration of 50 μM, we were pleased to observe a conversion of approximately 48% to the corresponding homodimer 21 (Scheme b, lane 2). Again, by increasing the reaction time to 5.0 h and the protein concentration to 100 μM, we could push the conversion to 21 up to 72% (Scheme b, lane 5).

Conclusions

We have demonstrated that the KAT ligation can be readily applied to protein conjugation under dilute, near-equimolar conditions. Through a two-step protocol, we have achieved PEGylation and covalent homodimerization of various recombinant proteins. Our success with 2-pyridyl KATs should lead to the discovery of ligation partners with even further enhanced ligation kinetics and ultimately to chimeric fusion architectures through the development of novel multifunctional linkers that will facilitate the covalent, selective dimerization or multimerization of different biomolecules.