Live Monitoring of Strain‐Promoted Azide Alkyne Cycloadditions in Complex Reaction Environments by Inline ATR‐IR Spectroscopy

Abstract The strain‐promoted azide alkyne cycloaddition (SPAAC) is a powerful tool for forming covalent bonds between molecules even under physiological conditions, and therefore found broad application in fields ranging from biological chemistry and biomedical research to materials sciences. For many applications, knowledge about reaction kinetics of these ligations is of utmost importance. Kinetics are commonly assessed and studied by NMR measurements. However, these experiments are limited in terms of temperature and restricted to deuterated solvents. By using an inline ATR‐IR probe we show that the cycloaddition of azides and alkynes can be monitored in aqueous and even complex biological fluids enabling the investigation of reaction kinetics in various solvents and even human blood plasma under controlled conditions in low reaction volumes.

Abstract: The strain-promoted azide alkyne cycloaddition (SPAAC) is ap owerful tool for forming covalentb onds between molecules even under physiological conditions, and therefore found broad application in fields ranging from biological chemistry and biomedical research to materials sciences. For many applications,k nowledge about reaction kinetics of these ligations is of utmost importance. Kinetics are commonly assesseda nd studied by NMR measurements. However,t hese experiments are limited in terms of temperature and restricted to deuterated solvents. By using an inline ATR-IR probe we show that the cycloaddition of azides and alkynes can be monitored in aqueous and even complex biological fluids enabling the investigation of reaction kinetics in various solvents and even human blood plasma under controlled conditions in low reaction volumes. The 1,3-dipolar cycloaddition of organic azides and alkynes, first reported by Huisgen in 1960, [1] has been re-emerging since the developmento fc opper-catalyzed click chemistry by Sharpless [2] and Meldal, [3] which has foundb road application in many fields and become ar obusta nd efficient tool for bioconjugation. [4][5][6] However,d ue to the need for cytotoxic copper, these reactions are only of limited suitability for in vivo applications. [7] Also decades ago, in 1961 Wittig and co-workersa lready reportedt hat cyclooctyne and phenyla zide reacte xtremely fast at room temperature forming as ingle product. [8] Based on these findings Bertozzi and co-workersd eveloped the concept of copper-free and thus bioorthogonal click chemistry. [9] Due to the strained triple bond, cyclooctynes are already distorted towards transition state geometry,w hich significantly lowers the activation energy. [10,11] Strain-promoted azide alkyne cycloaddition (SPAAC) reactions thus proceed already at room temperature without the use of any catalyst. [12] Several cyclooctyne derivatives have been prepared in the last decadet oi mproveb oth reactivity and stability of these bioorthogonal compounds. [13][14][15][16] In addition, the influence of different azides was investigated. [9,[13][14][15][17][18][19][20][21] Reported second order rate constants range from 2.4 10 À3 m À1 s À1 up to 34 m À1 s À1 . [9,22] Knowledge about the kinetics of bioorthogonal ligations in complex environments such as biological fluids is of utmost importance considering respective applications in vitro and in vivo. However,commonly used methods do not offer ag eneral approachf or the measurementi nb iological fluids of any SPAAC reaction.
Reaction kinetics are commonly assessed and investigated by NMR measurements. [9,13,16,21,[23][24][25][26] In this case the reaction partnersa re mixed in deuterated solvents in an NMR tube and the reaction is monitoredb yc onsecutive scans at defined time points. The advantage of this methodi st he ability to easily follow every involved species, assuming separated signals. However, there are several drawbacks. Controlo ft he reaction temperature is difficult and limited, and the need for deuterated solvents renders measurements in complex biological fluids impossible. In addition to NMR, UV-Vis has been successfully used to study SPAAC reaction kinetics. [27][28][29] While this approach can be used for live reaction monitoring with the possibilityo f temperature control and the use of aw ide variety of solvents, structuralr equirements are imposed on the reactionp artners, and solvents with highly interfering background, such as biological fluids, cannotb eu sed. Furthermore, fluorescencem easurementsh ave been appliedf or the investigation of click kinetics. [30] However,these methods depend on fluorogenic reactants or fluorescence quenching duringt he formation of the ligation product, and can thus not be used as general analytical tools. Very recently,S teflova et al. have reported the stepwise investigation of SPAAC by capillary electrophoresis. [31] Herein we presentamethodf or the live monitoring of SPAAC ligationsa td ifferent temperaturesa nd in various solvents,i ncluding humanb lood plasma, using inline ATR-IR spectroscopy ( Figure 1). IR spectroscopy offersamonitoring of the reaction progress by following the characteristic absorption of the azide moiety at around2 100cm À1 ,w hich is usually well separated from signals of other functional groups and solvents. Figure 2a shows the IR spectra of phenylacetylene( 1), phenyl azide (2)a nd the respective click product 3,w ith as eparated azide doubleb and [32] around 2100 cm À1 .D etermination of reaction rates of azide cycloadditions using IR spectroscopy was already performed by Huisgen et al. in 1967. [33] They were able to determine the reactivity of over 40 different alkynes and alkenes in the 1,3-dipolar cycloaddition with azides. More recently,v an Delft and co-workers used transmission-FTIR measurements to investigate the significantly higher reactivity of electron-deficient aryl azides towards aliphatic cyclooctynes. [19] However,a lthought hey have been successful to study reactions carried out in a9 :1 mixture of THF and H 2 O, measurements at higher water content or in other solvents like methanol failed.
To addresst he limitations of currently used methods we aimed to design an ew system and strategy enabling the monitoring of azide cycloadditionsi na queous and more complex solutions with full control of the reaction temperature. To this end, aR eactIR 15 system (Mettler To ledo) equipped with an ATR-IR SiComp probe was used, which features as ilicon crystal for ATRt hat (in contrast to diamond crystals)e xhibits only low absorption around2 100 cm À1 .T his setup not only allows for temperature control, but furthermore the use of an inert gas atmosphere and stirring, by simply immerging the probe into ar eaction solution within at empered and sealed vessel. For this study,w eh ave used as pecialf lask (Figure2b) to enable temperature-controlled measurements in low reactionv olumes. It consists of ad ouble-walled tube connected to at hermostat,a nd two NS14 glass joints arranged in a4 5 8 angle, one on top for insertingt he ATR-IR probe and as econd one for a temperature sensor and/or the addition of reagents. With this setup av olume of 0.5mLi ss ufficient for reactionm onitoring while stirring. In case stirring is not required even lower volumes can be used.
The first evaluation of the setup was done by measuring benzyla zide (2)i na cetonitrile at concentrationsr anging from 10 to 100 mm providing excellent linear correlationb etween the peak heighta nd area of the azide band to the concentration of 2 (Figure2c).
We next applied this setup to the monitoring of aS PAAC in ar eaction volume of 1mL. Before starting the measurement, the background of pure solventw as acquired. The azide solution (0.9 mL) was then placedi nt he flask and upon temperature equilibrationa10-fold concentrated solutiono ft he cyclooctyne (0.1 mL) was added to obtain an equimolar mixture of both reagents. Monitoring of the reaction was startedb efore adding the second reactant and reaction was followed by consecutive inline ATR-IR spectroscopy.T he intervalb etween the scans was chosen based on the reaction rate, ranging from 15 sf or fast reactionst o1min for slower conversions. For detailed description of the used settings see Supporting Information. For evaluation of this setup the reactionb etween benzyl azide (4)a nd cyclooctyne (5)i nD MSO at 37 8Cw as monitored (Figure 3a). Data was recorded and pre-processed using the iC IR 4.3.27 software( Mettler To ledo) and analyzed in Prism 6 (GraphPadS oftwareI nc.). First, the background spectrum was subtracted followed by baselinec orrection. Then ap eak region was assigned to the area of the azide signal (2100 cm À1 ). Peak area andp eak height were analyzed over time and rate constants were determined by linearization and subsequentl inear fitting (Figure3b). Reaction monitoringu sing the signalh eight showedl ess noise and ab etter correlation. In addition, the peak heighti sn ot dependento nt he width of the assigned peak region. [30] Therefore, peak height was chosen for further measurements. Despite the different signal-to-noise ratios similar results were obtained for the calculated rate constants (peak height:1 .52 10 À2 m À1 s À1 vs. peak area:1 .46 10 À2 m À1 s À1 ).
Measurements at differentc oncentrations, ranging from 10 to 100 mm werec onducted (see Supporting information). Althoughm easurement at low mm concentrations (< 25 mm)i s possible, the signal-to-noise ratio is reaching the limit for accurate analysis. Therefore, as tartingc oncentration of 50 mm or higheri sr ecommended for reliable and reproducible measurements.
Ab ig advantage of ours etup is the ability to easily monitor reactions at different temperatures. We have been able to monitort he SPAAC ligationo f4 and 5 at different temperatures, rangingf rom 0t o6 0 8C( Figure3c). Second order rate constantsw ere determined to be in the range from 5 10 À4 m À1 s À1 at 0 8Ct o5 .83 10 À2 m À1 s À1 at 60 8Cs howinga n overall increaseo ft he reactionr ate of approximately1 50-fold.
In addition, we have compared our results for monitoring by using inline ATR-IR to commonly used 1 HNMR measurements (Figure 3d)bystudying the reactionbetween 4 and 5 in acetonitrile at 37 8C. Both methods gave very similarr esults and comparable rate constants (1.11 10 À2 m À1 s À1 determined by NMR and 1.35 10 À2 m À1 s À1 determined by ATR-IR). To evaluate  the system's performance for the monitoring of faster ligations, measurements of the reactionb etween the highlyr eactive cyclooctyneb icyclo[6.1.0]non-4-yn-9-ylmethanol (BCN, 7, endoisomer) [16] and benzyla zide (4)w ere performed. The second order rate constant for this reactioni nD MSO at 37 8Cw as determined to be 0.15 m À1 s À1 (Figure 4).
To assesst he applicability of our methodf or the monitoring of SPAAC in water and even complex biological fluids, IR spectra of water-soluble 2-azidoethanol (8)i nb oth water (see Supporting Information) andh uman blood plasma (Figure 5a) were measured at differentc oncentrations. Both peak height and peak area of the azide band showedv ery good linear cor-relationf or concentrations ranging from 20 to 200 mm in both solvents.
Since the results did not reveal as ignificant difference between the performance in water andb lood plasma, we proceeded to investigate the monitoring of SPAAC in human blood plasma.T herefore, the reactiono f2 -azidoethanol (8) and the water-soluble PEGylated BCN derivative 9 at 100 mm concentration and 20 8Cw as followed ( Figure 5b). Despite a lower signal-to-noise ratio (in comparison to organic solvents), the decrease of azide signalc ould reliably be used for the determination of the rate constant (peak height:0 .19 m À1 s À1 vs. peak area:0 .21 m À1 s À1 ). Data for reaction monitoring at 50 mm and 37 8C( peak height: k= 0.57 m À1 s À1 )i sp rovided in the Supporting Information.
In summary,w eh ave developed am ethodf or inline ATR-IR kinetic measurements of strain-promoted azide alkyne cycload-  (1), phenyl azide (2)and the respective click product 3,h ighlighting the separated azide double band [31] of 2.b)Reactionflask for inlineA TR-IR measurements (1:double walled tube,2:t hermostat, 3: ATR-IR probe is insertedt hrough the NS14 glassj oint on top,4:temperature sensor (bluec able) and/or addition of reagents,5:magnetic stirringo fr eaction volumes as low as 0.5 mL). c) Inline ATR-IR measurements showing the azide band of 2 at differentc oncentrations (DMSO,378C, gray areas indicateSD, n = 3), and the resulting correlations of peak height and area to the concentration of 2.  ditions enabling live reaction monitoring even in complex biological fluids such as blood plasma. The setup can be used for measurements at different temperatures and low reaction volumes. Data acquisition is possible during the addition of the reactionp artner and therefore first data points are obtained right after the start of the reaction. Ar elatively short intervalo f 15 sb etween the scans enables the monitoring of fast SPAAC reactions (k > 0.1 m À1 s À1 ), whereas longer intervals can be used for slow conversions.E ven though limited to azide concentrations > 10 mm,t his method can be used for kinetic investigation of fast bioorthogonal SPAAC ligations in complex reaction mixturesa nd environments, providing important information on the reactivity of the used reaction partners. Hence,w e expect this methodt of ind application in the fields of bioorthogonal chemistry and bioconjugation, and provide valuable insights regarding the kinetics of strain-promoted azide alkyne cycloadditions.