Selenium catalysis enables negative feedback organic oscillators

The construction of materials regulated by chemical reaction networks requires regulatory motifs that can be stacked together into systems with desired properties. Multiple autocatalytic reactions producing thiols are known. However, negative feedback loop motifs are unavailable for thiol chemistry. Here, we develop a negative feedback loop based on the selenocarbonates. In this system, thiols induce the release of aromatic selenols that catalyze the oxidation of thiols by organic peroxides. This negative feedback loop has two important features. First, catalytic oxidation of thiols follows Michaelis-Menten-like kinetics, thus increasing nonlinearity for the negative feedback. Second, the strength of the negative feedback can be tuned by varying substituents in selenocarbonates. When combined with the autocatalytic production of thiols in a flow reactor, this negative feedback loop induces sustained oscillations. The availability of this negative feedback motif enables the future construction of oscillatory, homeostatic, adaptive, and other regulatory circuits in life-inspired systems and materials.


Materials and Methods
Unless otherwise noted, all chemicals were purchased from Sigma-Aldrich, Acros Organics, Alfa Aesar, and Merck.All solvents were purchased from Sigma-Aldrich and Acros Organics.D2O was purchased from Tzamal d-chem; all other NMR solvents were purchased from Cambridge Isotope Laboratories.All chemicals, including solvents, were used without further purification.LC-MS grade water was used in all kinetic experiments.NMR spectra were measured on a Bruker AVANCE III-300 spectrometer at 300 MHz for 1 H, at 73.7 MHz for 13 C{1H}, on a Bruker AVANCE III-400 spectrometer at 400 MHz for 1 H, at 100.6 MHz for 13 C{1H}, on a Bruker AVANCE III HD-500 spectrometer at 500 MHz for 1 H, and at 125.8 MHz for 13 C{1H}.Chemical shifts for 1 H and 13 C are given in ppm relative to TMS, and for 77 Se relative to Se(Me)2. 1 H and 13 C spectra were calibrated using a residual solvent peak as an internal reference (D2O 1 H NMR: δ = 4.79 ppm; CDCl3 1 H NMR: δ = 7.26 ppm, 13 C NMR: δ = 77 ppm; DMSO-d6 1 H NMR: δ = 2.50 ppm, 13 C NMR: δ = 39.52 ppm).

Synthesis
Supplementary Figure 1.The procedure for the synthesis of catalyst 2 Tert-butyl 4-selenocyanatobenzoate was prepared following Nakamura et al. 1 First, 37% HCl (5.625 mL) was added dropwise to a 250 mL flask charged with tert-butyl 4-aminobenzoate (7,5 g, 0.039 mol) and water (50 mL) in an ice-water bath.After 5 min of stirring at 0°C, a NaNO2 (2.22 g, 0.039 mol) solution in H2O (5 mL) was slowly added dropwise.The reaction was stirred for 10 more minutes.A freshly prepared KSeCN (4.66 g, 0.039 mol) solution in H2O (5 mL) was added dropwise, keeping the temperature below 5°C.The reaction was stirred at this temperature for 30 min, and then at room temperature for 3 h.The mixture was extracted with Et2O, washed with water (2x20 mL) and brine (1x20mL), then dried with Na2SO4 and concentrated under reduced pressure.The crude product was purified by column chromatography eluting with hexane/EA (v/v = 40/1) in 53% yield (5.83 g, 20.7 mmol).

Supplementary Figure 2. The procedure for the synthesis of catalyst 3
To the solution of di-tert-butyl 4,4'-diselanediyldibenzoate (0.197 g, 0.38 mmol) in Et2O (3 mL) was added a scoop of powdered Zn, water (1 mL), and HCl 37% (2 mL).A mixture was closed with a septum, bubbled with Ar and left for 30 min.After the organic phase was decolorized, it was extracted with Et2O (3x2 mL) into a flask charged with Na2SO4 under Ar flow.
2. To a flask containing the solution of the corresponding alcohol (0.76 mmol) and triphosgene (0.113 g, 0.38 mmol) in dry THF (3 mL), a pyridine (0.061 mL, 0.76 mmol) solution in THF (0.5 mL) was added.After having been stirred for 40 minutes, the resulting solution was directly used in the third step reaction.
3. To a flask with the product of step 2, the ether solution from step 1 was added.Then the pyridine (0.124 mL, 1.52 mmol) solution in DCM (1 mL) was added dropwise and stirred for 2h.Next, the reaction mixture was diluted with DCM, washed with 1M HCl (1x10 mL), Na2CO3 (1x10mL), water (1x10 mL), and brine (1x10 mL), and then dried with Na2SO4.
4. The solution of the above compound was dissolved in 3 mL DCM, and TFA (3 mL) was added at 0°C.After 15 min of stirring at 0°C and 45 min at room temperature, the reaction mixture was evaporated until dryness, resulting in the desired compound, 4 or 5 almost in quantitative yield.

The standard protocol for oscillations in flow
Four syringes were filled with the required solutions as described below: • 163 mg of thiouronium salt 8 and 64 mg of tBuOOH were dissolved in 2.5 mL of HPLCgrade water; next, this solution was transferred into a syringe.An additional 0.5 mL of water was used to wash the vial in which the solution was prepared, and the volume in the solution in the syringe was precisely filled to 3 mL.The same method was used to control the total volume of the solutions when filling the other syringes.The final concentration of 8 in the syringe was 168 mM, and the final concentration of tBuOOH in the syringe was 218 mM.
• Different selenium catalysts in various amounts were dissolved in 20 μL DMF and 2 mL water.The solution was transferred to a syringe and its volume was adjusted to 3 mL.The final concentrations in the syringe were 0.6 mM for 2, 1.5 mM for 3, 3.0 mM for 4, and 6.0 mM for 5.
• Tris-buffer (3M) was used to prepare the solution for the third syringe.The buffer was prepared to pH 7.7 after a threefold dilution to a concentration of 1 M. Thus, to obtain 100 mL of a buffer solution, 8.395 g (0.0693 mol) of anhydrous Tris-amine and 36.311g (0.2304 mol) of anhydrous Tris-HCl were placed in a 100 mL volumetric flask and filled with HPLC grade water up to 100 mL.This Tris-buffer was used to prepare a 3 mL solution of 203 mg of cystamine dihydrochloride salt in a third syringe.The final concentration of cystamine in the syringe was 300 mM for all experiments.
• To fill the last syringe, two solutions were prepared separately.First, 10 mL of HPLC grade methanol was used to dissolve 164 mg (414 mmol) of 5,5'-dithiobis-(2-nitrobenzoic acid) (Ellman's reagent) and 15 mL of HPLC grade water was used to dissolve 0.5 g of KH2PO4.Those two solutions were combined and a 25 mL glass syringe was filled with solution through a 0.22 µm syringe filter.The final concentration of Ellman's reagent in the syringe was 16.6 mM All 5 mL syringes (with 3 mL of solution each) were installed in the syringe pump system and connected to the CSTR using 0.5 mm internal diameter PTFE tubing.The CSTR outlet tubing was connected to the microfluidic mixer where the content of CSTR was mixed with Ellman's reagent.For experiments with catalysts 3-5, the flow rate for syringes 1-3 was 640 µL/h.For the experiment with catalysts 2, the flow rate for syringes 1 and 3 was 640 µL/h, whereas the flow rate for syringe 2 was 400 µL/h.In all experiments, the flow of the solution of Ellman's reagent was three times higher than the flow from CSTR.After mixing, the flow passed through the flow cell, where absorbance at 412 nm was measured.

1 H NMR kinetics experiments with proinhibitors 3-5
The kinetics of the interaction between proinhibitors 3-5 and cysteamine was monitored by 1 H NMR. In these experiments, we prepared a solution of a proinhibitor (3, 4, or 5) in 1 M Tris buffer pH 7.5, D2O.The amounts of 3-5 were calculated to produce a 5 mM solution after dilution to a final volume of 600 µL.Next, we prepared the solution of cysteamine hydrochloride (0.68 mg, 10 mM after dilution to 600 µL) in the same buffer.The solutions were mixed, and 1 H NMR spectra were recorded every minute.The concentrations of 3-5 during the reactions (Figure 3b  and kh = 0.00013 ± 1 .10 -5 s -1 .

Kinetics of the oxidation of cysteamine by tBuOOH without selenium derivatives
Generally, the solution of cysteamine in Tris buffer pH 7.5 was mixed with the solution of tBuOOH.The initial cysteamine concentration was in the range 0.5-10 mM and the initial tBuOOH concentration was in the range 3-39 mM.To follow the concentration of cysteamine in this reaction, we used the following experimental protocol: A freshly prepared solution of Ellman's reagent 2 mM in phosphate buffer solution (pH = 7.0; 200 mM) was used to fill UVvis cuvettes with 2 mL of solution in each cuvette.Next, 20 μL aliquots of the reaction mixture were taken every few seconds and mixed with Ellman's solution.The absorbance at 412 nm was measured by a UV-Vis spectrometer and converted to the total concentration of thiols.
To obtain the required plots, the absorbance data were analyzed as follows: First, the absorbance data were converted to cysteamine concentrations in the reaction mixture using an extinction coefficient of 14150 M -1 cm -1 and a dilution factor of 101 (20 μL aliquot vs 2020 μL of Ellman's solution plus aliquot).Next, the concentration was plotted against time, and the initial linear regions identified and fitted with a linear function.The slopes of these lines represent the initial reaction rates.Measurements for each reaction rate point are repeated three times, and average values were used to make the required curves and to determine the reaction orders and the rate constants.

Kinetics of the oxidation of thiocholine by tBuOOH without selenium derivatives
The kinetics of this reaction was studied using the same protocols as in the oxidation of cysteamine.The results of the studies are summarized in Supplementary Figure 12.As shown in Supplementary equation 2, the slope of the curve from Supplementary Figure 12 corresponds to 2 .k .

Supplementary
[tBuOOH]; therefore, k = 0.143 ± 0.003 M -1 s -1 .We noted that the value of 0.143 M -1 s -1 is an estimate; the actual values for rate constants under the conditions of the oscillatory experiments might differ significantly.
for the first exchange in the context of the oxidation of thiols by peroxides.The oxidation takes seconds to minutes and this equilibrium is established instantly using this time scale.
Although the averaging of signals for all compounds except cystamine prevents direct calculation of Keq1 and Keq2, some information on these constants can be obtained from equilibrium cystamine concentrations obtained from NMR integration.The information is summarized in Supplementary Table 2.

Starting composition
Cystamine concentration from integration 1 (10 mM), cysteamine (5 mM) 1.5 mM The numerical model of the oscillation in CSTR (Supplementary Figure 17) was built using COPASI software. 6Seven assumptions were made in this model.( 1) All disulfide exchanges have the same rate constants.(2) The six disulfides were reduced to three "halves of disulfides".
Therefore, the disulfide exchanges are also simplified to three exchange reactions.Accordingly, the total concentration of disulfides in the model was twofold higher than the experimental value.
(3) The two stages of the thiol-assisted amination of the thiouronium salts were not separated in this model; instead, they were treated as one irreversible reaction.( 4) Based on the experimental data for cysteamine and thiocholine, the non-catalytic oxidation of all three thiols was assumed to proceed with same rate constant, kOx.(5) The catalytic oxidation was assumed to follow Supplementary equation 22. (6) We assumed no difference between thiols in the reaction releasing 4-carboxyselenophenol from pre-catalysts 2-4.(7) We neglected the rearrangement of thiocarbonate formed in the reaction of 2-4 with cysteamine into carbamate.
The behavior of the system was described by a set of ODEs (Supplementary Figure 18) and was solved numerically by using the COPASI program.

3
Additional studies of kinetics of oxidation of cysteamine by tBuOOH 4.4 Additional studies of kinetics of oxidation of thiocholine by tBuOOH5.Derivation of the rate equation for the oxidation of thiols catalyzed by Mass spectra were measured on a Waters Xevo G2-XS QTof mass spectrometer with an electrospray ionization (ESI) source spectrometer.All spectra were acquired in the mass range of 50-2000 m/z.The mass errors of the analyzed spectra are not more than 5.0 ppm.The absorbance in the flow experiments was constantly monitored by a Cary 60 UV-VIS spectrometer, manufactured by Agilent Technologies.The flow cell used in these experiments was a hand-made cell.The reactants in the flow experiments were supplied by NEMESYS Low Pressure Syringe pumps produced by CETONI GmbH.

1 Supplementary Figure 4 .
Flow set-upThe flow set-up consisted of four main components: (i) syringe pumps, (ii) a micro continuously stirred tank reactor (CSTR), (iii) a microfluidic mixer (or two mixers), and (iv) a flow cell.We used NEMESYS Low Pressure Syringe pumps from CETONI in all our experiments.Calibration of the flow cell.We used the calibration curves to convert the absorption to the concentration values and thus, to quantify the UV data from the oscillatory reactions in CSTR.To obtain the calibration curves, we mixed one volume of the mercaptoethanol solution of a corresponding concentration in Tris-buffer pH 7.7 with three volumes of the water/methanol (1.5/1) solution of Ellman's reagent (16.6 mM) also containing KH2PO4 (146 mM).Then, we set the absorbance to 0 for pure Tris-buffer, passed the resulting solution through our self-made flow cell and recorded the absorbance of the solutions.The whole sequence of preparing the solution and measuring the absorbance was repeated three times for each concentration of mercaptoethanol.The mean values were plotted and fitted with the following function:Y = a(1 -b X ),where Y is the absorbance, X is the concentration of thiols, and a, b are the parameters.Calibration curves of the thiol concentration in a home-made flow cell using Ellman's reagents.The background was corrected by subtracting the absorbance of Tris-buffer.Nonlinear fitting of the asymptotic function Y = a(1 -b X ) was performed using Origin 9.The fitted parameters are as follows: a = 0.54941 ± 0.01215 and b = 0.649 ± 0.0.01218.The error bars represent the standard deviations from three independent measurements.

Supplementary Figure 5 .
main text) were calculated by integrating signals in the aromatic region, then plotted and fitted with Supplementary equation 1 for a second order reaction with the non-equivalent initial concentrations of the reactants: where A0 is the initial concentration of 3-5, B0 is the initial concentration of cysteamine, k is the second order rate constant, and t0 is the correction factor for uncertainty in the starting time of the reaction.During fitting, we fixed A0 at 0.005 and B0 at 0.01.Experiments were conducted in triplicate.The experimental and fitting results are summarized in Supplementary Figure 5-7.Kinetics of the reaction between 3 and cysteamine.Parameters obtained by fitting with Supplementary equation 1: left (k = 1.59 s -1 M -1 ; t0 = -63 s), middle (k = 1.92 s -1 M -1 ; t0 = -57 s), and left (k = 1.74 s -1 M -1 ; t0 = -57 s).H2O, 1 M Tris buffer pH 8, 25 o C; [8] = 20 mM; [9] = 40 mM.The model parameters obtained by fitting the experimental data are as follows: kSS = 5 s -1 M -1 (fixed), kL = 1.98 ± 0.02 s -1 M -1 ,

Figure 12 .
Determination of the second order rate constant for the reaction between thiocholine and tBuOOH.Reaction conditions: Tris buffer, pH 7.5, 1M, 25 o C, [tBuOOH] = 7.5 mM, [Thiocholine] = 0.5 -10 mM.The error bars represent the standard deviations from three independent measurements.

Table of Contents
The last three reactions were used only for oscillators with pre-catalysts 2-4.The parameters involved here are as follows: kH is the rate of the thiouronium salt hydrolysis; kL is the rate of the ligation; kSS is the rate of the disulfide exchange; kOx is the rate of the noncatalytic oxidation of thiols by tBuOOH; kcat, kOOH, and kinh correspond to k2 * /2, k2 * /k1, and 0.5k-1/k1 from Supplementary equation 22, and kAct is the second order rate constant for the release of 4-carboxyselenophenol.This kAct constant differs between substrates 2-4.All reactions except for the catalytic oxidation are assumed to follow the mass-action law.The rate law for the catalytic oxidation is described by equation 22.The rate law for flow supply and removal is described as the FvV × concentration.