Application of an enzymatic cascade reaction for the synthesis of the emeraldine salt form of polyaniline

The synthesis of the emeraldine salt form of polyaniline (PANI-ES) from aniline with Aspergillus sp. glucose oxidase (GOD), d-glucose, dissolved O2, and horseradish peroxidase isoenzyme C (HRPC) in the presence of large unilamellar vesicles of AOT (sodium bis-(2-ethylhexyl)sulfosuccinate) as templates at pH = 4.3 and T ~ 25 °C was investigated in a systematic way. In this cascade reaction mixture, the oxidation of aniline is catalyzed by HRPC with H2O2 that is formed in situ as byproduct of the GOD-catalyzed oxidation of d-glucose with O2. Under the elaborated experimental conditions which we considered ideal, the formation of PANI-ES products is evident, as judged by UV/Vis/NIR and EPR measurements. Comparison was made with a reference reaction, which was run under similar conditions with added H2O2 instead of GOD and d-glucose. Although the reference reaction was found to be superior, with the cascade reaction, PANI-ES products can still be obtained with high aniline conversion (> 90%) within 24 h as stable dark green PANI-ES/AOT vesicle dispersion. Our results show that the in situ formation of H2O2 does not prevent the inactivation of HRPC known to occur in the reference reaction. Moreover, the GOD used in the cascade reaction is inactivated as well by polymerization intermediates.


Introduction
Although there are well-established procedures for the chemical (Huang et al. 1986;Stejskal and Gilbert 2002;Surwade et al. 2009;Stejskal et al. 2015) or electrochemical (Huang et al. 1986;Okamoto and Kotaka 1998;Stejskal et al. 2015) synthesis of the electroconductive polyaniline (PANI), the potential of using oxidative enzymes as biocatalysts is also being explored (Xu et al. 2006;Bouldin et al. 2010;Cruz-Silva et al. 2011;Walde and Guo 2011;Otrokhov et al. 2013;Ćirić-Marjanović et al. 2017), most often by applying a heme peroxidase with hydrogen peroxide (H 2 O 2 ) as oxidant (Liu et al. 1999a, b;Sakharov et al. 2003;Caramyshev et al. 2005;Rumbau et al. 2007;Junker et al. 2012Junker et al. , 2013 or a high potential laccase in the presence of molecular oxygen (O 2 ) (Karamyshev et al. 2003;Shumakovich et al. 2012;Junker et al. 2014;Zhang et al. 2016;Walde et al. 2019). The enzymatic approach allows carrying out the synthesis under milder, environmentally friendlier conditions as compared to chemical methods, at least if the acidity of the reaction medium is compared: pH ≈ 3.5-4.5 in the case of the enzymes, pH < 2.5 for the chemical approach (Stejskal et al. 2015). Moreover, in the case of the enzymatic approach with H 2 O 2 or O 2 as oxidant, the side product is H 2 O (Junker et al. 2012(Junker et al. , 2014, while oxidation with the most popular chemical oxidant, ammonium peroxydisulfate (APS) (Stejskal and Gilbert 2002), yields sulfuric acid as side product (Stejskal et al. 2015). Obviously, for such comparison the costs of the enzymes-as compared to APS, for example-should not be ignored (Junker et al. 2012(Junker et al. , 2014. Only if cheap, industrial grade enzymes can be used, the enzymatic approach might compete economic-wise with the traditional chemical methods. On the other hand, the chemical structure, average molar mass, polydispersity, and "solubility" or dispersibility of the PANI products obtained may not only depend on the synthesis method used, but also on the precise experimental conditions at which the synthesis is carried out with one chosen method (Kashima et al. 2018). With the milder enzymatic approach, products with superior, desired properties-as compared to the chemical or electrochemical methods-might be obtained. For exploring this possibility, detailed studies with enzymes as catalysts are worth pursuing. The work we present here is one of such studies.
In terms of "desired properties," the emeraldine salt form of polyaniline (PANI-ES) is of particular interest for many possible applications, as it is the only (or most relevant) electroconductive form of PANI (Chiang and MacDiarmid 1986;Focke and Wnek 1988;Stejskal et al. 2015;Petrova et al. 2012). In Scheme 1, the chemical structure of the repeating unit (r.u.) of an ideal PANI-ES chain is shown. It consists of four linearly connected aniline building blocks that are bonded via the nitrogen atoms and the carbon atoms in para-position to the amino group. Moreover, two of the four aniline building blocks in the PANI-ES r.u. are oxidized and protonated and two are reduced (and not protonated). This half-oxidation state of the r.u. may result in the formation of the polaron forms of PANI-ES with two unpaired electrons per r.u. and all four benzene rings with fully delocalized π electrons (benzenoid structure). One usually distinguishes polaron pairs from separated polarons (Scheme 1) (Wallace et al. 2009;Dmitrieva and Dunsch 2011). In the bipolaron form of the PANI-ES r.u., unpaired electrons do not exist (benzenoid and quinoid structures). From many previous studies, it is known that the nature of the counter ion ("dopant" A − in Scheme 1) has an influence on the conductivity of PANI-ES in the solid state (Stejskal et al. 2015).
PANI-ES has characteristic absorptions in the near-infrared (NIR) region, with maximal absorbance at λ max ≥ 800 nm (previously assigned to the π → polaron transition), at λ max ≈ 420 nm (polaron → π* transition) and at λ max ≈ 300 nm (π → π* transition) (Huang and MacDiamid 1993;Nekrasov et al. 2001;do Nascimento and de Souza 2010;Bilal et al. 2015). Furthermore, low absorbance at λ = 500-600 nm is indicative for a low extent of chain branching (Liu et al. 1999b) or low extent of phenazine formation . Recently, the above widely accepted assignments of the electron transitions at λ ≥ 800 nm and λ ≈ 420 nm to polaron transitions (Huang and MacDiamid, 1993) have been questioned (Lin et al. 2017;Mills et al. 2019). It is the observed low EPR signal intensity of PANI-ES-like oligomers that led to the suggestion that the strong absorption band in the near-infrared region of the spectrum might actually originate from transitions involving bipolarons and not polarons (Lin et al. 2017;Mills et al. 2019). Nevertheless, and independent from the correct NIR absorption band assignment, the mentioned characteristic spectroscopic features are typical for conductive PANI-ES, and they are often taken as a measure for optimizing reaction conditions in cases where the reaction products do not precipitate but remain dissolved or dispersed in solution and therefore can be analyzed by in situ UV/Vis/NIR absorption measurements. This was the case in several studies carried out in the past on the peroxidase/H 2 O 2 -or laccase/ O 2 -catalyzed oxidation of aniline in the presence of so-called "templates" (Liu et al. 1999a;Walde and Guo 2011;Junker et al. 2012).
Toward possible applications of enzymatically prepared polyaniline, in a recent study it was shown that PANI-ES obtained from aniline in the presence of AOT vesicles with horseradish peroxidase (HRP) and H 2 O 2 might by suitable as photothermal agent for tumor theranostics (Zhang et al. 2020). In another area of research, an artificial cell-like reproduction system is being developed, which is based on the AOT vesicle-guided, HRPC/H 2 O 2 -catalyzed synthesis of PANI-ES products (Kurisu et al. 2019). In that work, chemistry and soft matter physics are linked. More specifically, micrometer-sized giant AOT vesicle membranes work as templates for the synthesis of PANI-ES products, and the PANI-ES products simultaneously work as templates for the incorporation of AOT molecules from the bulk solution into the AOT membranes, which leads to vesicle membrane growth. Based on such mutual catalytic coupling between AOT membrane growth and PANI-ES product formation, AOT-based vesicles not only show growth in size but they also divide, i.e., vesicle reproduction occurs (Kurisu et al. 2019). In that work, the main focus is not on the properties of the PANI-ES products obtained, but on the physical analysis of changes of vesicle size and morphology in a constant chemical environment: AOT and H 2 O 2 molecules are continuously supplied to a giant target vesicle which is built of mainly AOT molecules.
In all these examples, the terminal oxidant, H 2 O 2 or O 2 , was added at-or was present from-the beginning of the reaction Junker et al. 2012;Fujisaki et al. 2019;Kashima et al. 2019), or it was added in portions during the course of the reaction (H 2 O 2 ) (Samuelson et al. 1998;Liu et al. 1999aLiu et al. , b, 2002Sakharov et al. 2003). The aim of the work presented here was to investigate whether PANI-ES formation from aniline with horseradish peroxidase isoenzyme C (HRPC) and AOT vesicles as templates is also possible with in situ formation of H 2 O 2 from d-glucose with glucose oxidase (GOD) and O 2 from air according to the following equation: (Bankar et al. 2009), see Scheme 2. Since such GOD-catalyzed in situ formation of H 2 O 2 in the presence of HRPC and aniline resulted in the formation of PANI-ES products, the two questions we wished to answer were the following. (a) How do the progress of the reaction and the spectroscopic properties of the obtained PANI-ES products compare with those of the PANI-ES products obtained under similar conditions without GOD and d-glucose but addition of H 2 O 2 at the beginning of the reaction? (Junker et al. 2012) And (b) Is there an advantage of using the enzymatic cascade reaction involving GOD and HRPC over the use of a single enzyme (HRPC), for example in terms of HRPC inactivation caused by H 2 O 2 ? The inactivation of heme peroxidases by high amounts of H 2 O 2 is well known since many years (Arnao et al. 1990;Hiner et al. 1996;Mao et al. 2014). With a controlled enzymatic in situ formation of H 2 O 2 , exposure of HRPC to high H 2 O 2 concentrations can be avoided or minimized.
In a broader context, enzymatic cascade reactions not only are essential for the efficient transformation of metabolites in each living cell, but they are also of great interest for the sustainable industrial synthesis of organic compounds (Wohlgemuth 2010;Ricca et al. 2011;Choi et al. 2015;Sigrist et al. 2015;Schmidt-Dannert and Lopez-Gallego 2016).

AOT vesicle preparation and CVC determination
AOT vesicle dispersions were prepared as described before with the polycarbonate membrane extrusion method (Junker et al. 2012). Solid AOT (89.0 mg) was added to a 250-mL round bottom flask and then dissolved in 5 mL chloroform, followed by the formation of a thin AOT film upon removal of chloroform with a rotary evaporator. For complete chloroform removal, the AOT film was put under high vacuum overnight. The AOT film was then hydrated and dispersed by using 10 mL of a 100 mM NaH 2 PO 4 solution (pH = 4.3) or of a 100 mM NaH 2 PO 4 solution containing a defined concentration of d-glucose (pH = 4.3). For most of the experiments, [d-glucose] = 100 mM (see below). The obtained 20 mM AOT vesicle dispersions were frozen by placing the round bottom flask in liquid nitrogen, followed by thawing in a water bath heated to 60 °C. This procedure was repeated ten times. The vesicle dispersions were first extruded ten times through 200-nm-pore-size nucleopore polycarbonate membranes, and then, another ten times through 100-nm pore-size membranes, using "The Extruder" from Lipex Biomembranes (Vancouver, Canada). The resulting AOT vesicle dispersions mainly contained large unilamellar vesicles (LUVs) with an average diameter of about 80-100 nm Junker et al. 2012). The vesicle dispersions were stored at room temperature (T ~ 25˚C) and used within 14 days after preparation.
The critical concentration for vesicle formation (CVC) of AOT in 100 mM NaH 2 PO 4 solution (pH = 4.3) and in 100 mM NaH 2 PO 4 /100 mM d-glucose solution (pH = 4.3) was estimated by turbidity measurement, as described before . A 20 mM AOT LUV dispersion that was extruded five times through 200-nm-pore-size nucleopore polycarbonate membranes was first diluted with 100 mM NaH 2 PO 4 solution (pH = 4.3) to different desired AOT concentrations. After incubation for 24 h at T ~ 25˚C, the turbidity determined as absorbance at the arbitrarily chosen wavelength λ = 400 nm, A 400 , was measured for each Scheme 2. Illustration of the two-enzyme cascade reaction used for the oxidation of aniline monomers to anilino radicals that undergo follow-up reactions (Junker et al. 2012). The two reaction steps are (1) the formation of H 2 O 2 (3) and d-glucono-δ-lactone (4) from d-glucose (1) and dissolved O 2 (2), whereby 4 is hydrolyzed to d-gluconic acid (5). In step (2), 3 oxidizes the heme enzyme HRPC in a two electron oxidation reaction to [HRPC-Fe(IV) = O] •+ (the socalled compound I, with a radical cation localized on the porphyrin ring), followed by two one-electron oxidations of two molecules of aniline (6) to two anilino radicals (7), via compound II (not shown) (Junker et al. 2012(Junker et al. , 2013. Oligomeric and polymeric products are obtained from 7, 3, and 6, most likely without direct involvement of the enzymes (Junker et al. 2012). Not shown is the role "template" molecules play that are present in the reaction mixture (AOT in the present work): (i) for guiding the reactions, (ii) for acting as counter ions, and (iii) for preventing product precipitation. FAD stands for flavin adenine dinucleotide, the cofactor of GOD diluted sample by using a V-670 UV/Vis/NIR spectrophotometer from JASCO (Japan), and quartz cells with optical path lengths of L = 1 mm. A 400 values were plotted vs. AOT concentration. The CVC value was taken as the lowest AOT concentration at which A 400 clearly deviated from 0 (~ 0.4 mM), see Fig. S-1. This value is in agreement with previous determinations .

"Cascade reaction": synthesis of PANI-ES products from aniline with d-glucose, GOD, dissolved O 2 , HPPC, and AOT vesicles
The conditions for the "cascade reaction" run in 5-mL polypropylene Eppendorf tubes were the following. All components of the reaction mixture, except the GOD solution, were added to 300 µL of 100 mM NaH 2 PO 4 solution (pH = 4.3) containing defined concentrations of d-glucose: 75 µL AOT LUV dispersion (20 mM in 100 mM NaH 2 PO 4 solution containing a defined concentration of d-glucose), 50 µL aniline solution (40 mM in 100 mM NaH 2 PO 4 solution, pH adjusted to 4.3 with H 3 PO 4 ) and 25 µL HRPC solution (18.4 µM, prepared in 100 mM NaH 2 PO 4 solution, pH = 4.3, and spectrophotometrically determined, see above). After gentle mixing, the reaction was triggered by quick addition of 50 µL of a GOD solution of defined GOD concentration (prepared in 100 mM NaH 2 PO 4 solution, pH = 4.3, spectrophotometrically determined by using ε 450 (GOD) = 2.82 × 10 4 M −1 cm −1 ) (Swoboda and Massey 1965), followed by gentle mixing and closure of the caps of the reaction tubes and sealing with Parafilm (to avoid leakage of solution and air exchange). The tubes were then placed in the holes of a home-made rotary mixing device and continuously rotated at ~ 90 rpm (see Fig. S-2) during the reaction, unless indicated otherwise. The initial reaction conditions were as follows: 3.0 mM AOT, 4.0 mM aniline, 0.92 µM HRPC and defined concentrations of GOD (0.2 μM for the "ideal conditions," see below), and d-glucose (100 mM for the "ideal conditions," see below), in 100 mM NaH 2 PO 4 solution (pH = 4.3), reaction volume = 0.50 mL, T = 25˚C, t = 24 h. The desired d-glucose concentration in the prelaid 300 μL d-glucose solution was obtained by mixing defined volumes of a d-glucose stock solution of either 250 mM or 1 M (prepared in 100 mM NaH 2 PO 4 , pH = 4.3) and an appropriate volume of 100 mM NaH 2 PO 4 solution (pH = 4.3). The d-glucose stock solutions were prepared at least 24 h before use to allow d-glucose to reach equilibrium between the α-and β-forms.

UV/Vis/NIR and EPR absorption measurements
All UV/Vis/NIR absorption measurements were carried out with a V-670 UV/Vis/NIR spectrophotometer from JASCO (Japan) using quartz cuvettes of either L = 1 mm or 10 mm optical path length (Junker et al. 2012). Electron paramagnetic resonance (EPR) measurements of the reaction mixtures were measured with a Bruker EMX X-band spectrometer equipped with a TM cavity. The spectra were measured at X-band microwave frequency with a modulation frequency of 100 kHz and modulation amplitudes of 1 G at room temperature (Junker et al. 2012).

Quantification of aniline and H 2 O 2 in the "cascade reaction mixtures" during the reaction
The amounts of remaining aniline and formed H 2 O 2 in the "cascade reaction mixtures" were determined by using the same methods as described before (Junker et al. 2012). For the quantification of remaining amounts of aniline, 30 µL of the reaction mixture was added to 1470 µL acetonitrile inside a 2-mL polypropylene Eppendorf tube. After centrifugation, the UV/Vis/NIR absorption spectrum of the supernatant solution was recorded. From the absorption intensity at λ = 238 nm, the concentration of aniline was calculated based on ε 238 (aniline) = 1.01 × 10 4 M −1 cm −1 , which was determined from a calibration curve obtained with known amounts of aniline ( Fig. S-3). For the quantification of H 2 O 2 , a spectrophotometric method was used that is based on the reaction between added oxo [5,10,15,20-tetra(4-pyridyl) porphyrinato]titanium(IV)-Ti-TPyP-and H 2 O 2 present in the "cascade reaction mixture" (Takamura and Matsumoto 2009). The following procedure was used: 125 µL of diluted or undiluted reaction mixture was added to 125 µL of a 4.8 M perchloric acid solution placed inside. Afterwards, 125 µL of a Ti-TPyP solution (50 µM in 50 mM HCl) was quickly added. After gentle mixing and standing for 5 min at T ~ 25 °C, 875 µL of deionized water was added and the solution was centrifuged. The absorption intensity at λ = 432 nm of the supernatant solution, A 432 , was measured, then this value was compared to a blank sample which did not contain H 2 O 2 . The concentration of H 2 O 2 was determined from a calibration curve obtained with known amounts of H 2 O 2 , plotted against ΔA 432 = A 432 (blank)-A 432 (sample). See

Determination of the HRPC activity in the "cascade reaction mixtures" during the reaction
The activity of HRPC in samples withdrawn from the "cascade reaction mixtures" was determined spectrophotometrically with ABTS 2− as reducing substrate and H 2 O 2 as oxidant in 10 mM MOPS buffer solution (pH = 7.0) according to the protocol described by Ghéczy et al. (2016) (Ghéczy et al. 2016). The following solutions were added in the sequence given to a 1 mm quartz cuvette at T ~ 25 °C, followed by measuring the linear increase in A 414 during the first 60 s (formation of ABTS •− ): 960 μL MOPS buffer solution (10 mM, pH = 7.0), 10 μL reaction mixture containing HRPC, 20 μL ABTS 2− stock solution (50 mM in MOPS buffer solution, pH = 7.0), and finally 10 μL H 2 O 2 stock solution (20 mM in deionized water). After gentle mixing, the slope ΔA 414 /Δt was taken as measure for the activity of HRPC. The total assay volume was 1.

Determination of the GOD activity in the "cascade reaction mixtures" during the reaction
The activity of GOD in samples withdrawn from the "cascade reaction mixtures" was determined spectrophotometrically with d-glucose and ABTS 2− in 10 mM MOPS buffer (pH = 7.0) according to the protocol described before (Ghéczy et al. 2016), with slight modifications.
The following solutions were added in the sequence given to a 1 mm quartz cuvette at T ~ 25 °C, followed by measuring the linear increase in A

Quantification of d-glucose in the "cascade reaction mixtures" during the reaction
Remaining d-glucose in the "cascade reaction mixtures" was determined spectrophotometrically by withdrawing small volumes from the "cascade reaction mixtures" and then using the GOD/HRP/ABTS 2− assay, as described above. To eliminate GOD and HRPC present in the "cascade reaction mixtures", the two enzymes were first separated from the reaction mixtures by ultrafiltration using Amicon ® Ultra 0.5-mL centrifugal filters (regenerated cellulose, nominal molecular weight limit of 10 kDa) from Merck (Germany). The purified, enzyme-free d-glucose solutions were then analyzed for d-glucose content by adding the following solutions in the sequence given to a 1-mm quartz cuvette at T ~ 25 °C, followed by measuring the linear increase in

Synthesis of PANI-ES from aniline with HRPC and H 2 O 2 in the presence of AOT vesicles ("reference reaction")
From our previous investigations with HRPC and H 2 O 2 , the following initial conditions turned out to be ideal for obtaining PANI-ES products in the presence of AOT vesicles as templates in aqueous solution (Junker et al. 2012): 3.0 mM AOT, 4.0 mM aniline, 0.92 µM HRPC, 4.5 mM H 2 O 2 , pH = 4.3 (100 mM NaH 2 PO 4 ), T = 25 °C, t = 24 h. With these conditions of our "reference reaction", (i) the initially added aniline is oxidized to about 90-95% (aniline conversion, i.e., only 5-10% of the added aniline remain at the end of the reaction), (ii) the initially colorless reaction mixture becomes first blue and finally dark green with absorption maxima at λ ≈ 1000, 420, and 300 nm, and (iii) the absorption at λ ≈ 500 nm is relatively low ( Fig. S-8). These phenomena are characteristic for the formation of PANI-ES, see the Introduction. The formation of products that are rich in PANI-ES r.u. is further supported by the fact that the reaction products produce an EPR spectrum (indication of the presence of unpaired electrons, as expected for the polaron form of PANI-ES) (Junker et al. 2012;Fujisaki et al. 2019). Moreover, the PANI-ES products remain dispersed (no precipitation), and the chosen aniline concentration is such that the progress of the reaction can easily be followed by in situ UV/Vis/NIR absorption measurements using conventional quartz cells with an optical path length of L = 0.1 cm (Junker et al. 2012). We call the conditions mentioned above "optimal" for the "reference reaction" since they fulfill several criteria we set for the reaction course and results (Junker et al. 2012;Fujisaki et al. 2019). The "reference reaction" is usually carried out in closed reaction tubes without any stirring or tube rotation. As shown in Fig. S-9, continuous rotation of the reaction tube during the reaction has no influence on the UV/Vis/NIR absorption spectrum after t = 24 h, i.e., it has no significant influence on the outcome of the "reference reaction." The only clear drawback of the reaction is the high amount of HRPC required for high aniline conversion (0.92 µM for 4.0 mM aniline) (Junker et al. 2012). Nevertheless, the electrochemical properties of the obtained PANI-ES products are very promising (easy coating of glassy carbon electrodes and excellent capacitive properties) .
For the aim of the present work, we replaced the initial addition of H 2 O 2 with a mixture of d-glucose and GOD (and dissolved O 2 ), but kept all other conditions the same.

Reaction optimization upon replacing H 2 O 2 by d-glucose and GOD ("cascade reaction")
A few initial trials showed that incubation of d-glucose and GOD with aniline (4.0 mM) and HRPC (0.92 µM) in the presence of AOT vesicles (3.0 mM AOT) at pH = 4.3 (100 mM NaH 2 PO 4 ) at T = 25 °C leads to a change in the color of the reaction mixture from initially colorless to dark blue and finally dark green. This was a first qualitative indication that PANI-ES formation from aniline through the enzymatic cascade reaction with d-glucose, GOD and HRPC is possible. Therefore, in a next step the concentrations of d-glucose and GOD were varied in a systematic way and the UV/Vis/NIR absorption spectra of the different reaction mixtures were recorded after t = 24 h for determining  Fig. 1a, Figs. S-10 and S-11, we concluded that a substantial amount of d-glucose is required for obtaining reaction products that have similar UV/Vis/NIR spectra like the ones of the "reference reaction" shown in Fig. S-8. For all three GOD concentrations, 100 mM d-glucose seems to be a good compromise, although with 500 mM d-glucose the A 1000 /A 500 value is higher than in the case of 100 mM, 2.3 vs. 1.9 (Fig. 1b). In a next step, several reaction mixtures were prepared in which the initial concentration of d-glucose was kept constant at [d-glucose] 0 = 100 mM, and the concentration of GOD was varied between 0 and 0.40 µM in order to find the ideal GOD concentration. Again, the UV/Vis/ NIR spectrum was recorded after t = 24 h and taken as a measure for the formation of PANI-ES products, see In b and c, the lines between the data points are drawn for guiding the eyes. In c, the error bars represent standard deviations from two or three measurements ▸ are summarized. A comparison of the UV/Vis/NIR spectrum of this "ideal reaction mixture," recorded after t = 24 h, with the spectrum of the "reference spectrum," again recorded after t = 24 h, is shown in Fig. 2d. Although both spectra are similar, they are clearly different. For the "reference spectrum" the intensity of the absorption around λ = 1000 nm is higher and at about λ = 700 nm lower, if compared to the "ideal" d-glucose/GOD system.

Spectroscopic changes and aniline consumption for the cascade reaction
Changes of the UV/Vis/NIR absorption spectrum of the reaction mixture were measured during the reaction for the "ideal reaction" run in rotating tubes under the elaborated optimal conditions listed in Table 1, see Fig. 3. During the first phase of the reaction, there was a rapid increase in absorbance at λ max ≈ 750 nm (Fig. 3a, c), correlating with the observed appearance of a blue color at the initial stage of the reaction. The intensity of this absorption band then slowly decreased with reaction time with simultaneous increase in absorption in the NIR region, centered around λ max ≈ 1000 nm (Fig. 3b, d), correlating with the developing green color of the reaction mixture. Overall, these changes are very similar to the changes observed for the "reference reaction" (Junker et al. 2012). There is, however, also a notable, clear difference between the two systems. The initial increase in A 750 is much faster for the reference reaction mixture with added H 2 O 2 (4.5 mM, no d-glucose, no GOD) (Junker et al. 2012) than for the d-glucose/GOD system. Furthermore, while the spectrum for the reaction mixture with added H 2 O 2 remained relatively stable after t = 24 h (Junker et al. 2012;Fujisaki et al. 2019), the spectrum of the cascade reaction mixture with d-glucose and GOD still changed with time after t = 24 h, with a decrease in A 1000 and a considerable broadening of the absorption into the visible as well as the NIR region, see Fig. S-13. Continuous rotation of the reaction tube during the reaction has a clear positive effect on the reaction (Fig. S-14). After a reaction time of t = 24 h, the intensity of the absorption spectrum of the reaction mixture which was not rotated was much lower than in the case of tube rotation. The effect of letting the reaction tube standing, without any rotation, is also obvious from the UV/Vis/NIR spectra recorded during the reaction (Fig. S-15).
The cascade reaction was also run in air-exposed reaction vessels (open tube or flask to maintain a constant O 2 concentration) (Fig. S-16). Obviously, in this case rotation of the reaction vessels in the way done for the closed tubes ( Fig.  S-2) was not possible. They were left standing at T ~ 25 °C for t = 24 h. As a result from these measurements, there was no improvement in terms of A 1000 for the standing and open reaction mixture if compared to the rotating reaction mixture in the closed tube ( Fig. S-16). For this set of experiments, the pH value in the reaction mixtures was also determined. It dropped in all cases from (unbuffered) pH = 4.3 at the start of the reaction to pH = 3.9 (for the reaction mixture which was kept standing in an open vessel), pH = 3.8 for standing and closed, and pH = 3.6 for rotating and closed (Table S-1). This difference in measured pH values correlates with the absorption intensity of the reaction mixtures in the Vis/NIR region of the spectrum: lower pH value for reaction mixtures with higher absorbance at t = 24 h, i.e., higher reaction yield (as a result of higher amounts of H 2 O 2 and d-gluconic acid formed).
The amount of remaining aniline in the cascade reaction mixture was measured for reaction mixtures that were kept inside closed tubes which were rotated and then compared with the same reaction mixtures placed inside closed tubes without rotation. Without rotation, the reaction was slower and leveled off with considerably lower aniline conversion: for [GOD] = 0.2 μM, about 50% of the initially added aniline remained after t = 24 h without tube rotation, as compared to about 10% with tube rotation. (Fig. S-17). This is different to what we observed for the reference reaction with initial addition of H 2 O 2 (no d-glucose, no GOD), where reaction tube rotation did not have a significant influence on the outcome of the reaction (see Fig. S-9). It seems that tube rotation facilitates the reaction most likely through a more efficient molecular oxygen exchange between the air in the tube and the reaction mixture. Unless specified, all further measurements were carried out with reaction tube rotation.

Quantification of in situ formed H 2 O 2 for the cascade reaction
The concentration of H 2 O 2 that (i) formed from d-glucose and O 2 with GOD and (ii) simultaneously was consumed during the HRPC-catalyzed oxidation of aniline was determined for aliquots taken from the "ideal cascade reaction mixture" (Table 1) during the course of the reaction (Fig. 5).
The quantification was carried out with the spectrophotometric Ti-TPyP assay, using oxo [5,10,15,20-tetra(4-pyridyl) porphyrinato]titanium(IV), as described before (Junker et al. 2012), see "Experimental". Very different from the "reference reaction"-where the concentration of the initially added H 2 O 2 was as high as 4.5 mM-the determined concentration of H 2 O 2 in the "ideal cascade reaction mixture" was always below 30 μM (Fig. 5). This was also the case for

Changes in GOD and HRPC activities and quantification of remaining d-glucose for the cascade reaction
For the "ideal cascade reaction mixture," the activities of HRPC and GOD were measured during the course of the reaction by withdrawing small volumes from the reaction mixture and analyzing them with either ABTS 2− /H 2 O 2 (HRPC assay) or d-glucose/O 2 /ABTS 2− /HRP (GOD assay), see Fig. 6 (HRP activity) and Fig. 7 (GOD activity). For HRP, the activity decreased rapidly during the formation of PANI-ES products, very similar to the case of the reference reaction (Fig. 6). The latter data are in good agreement with what was observed before (Junker et al. 2012). The inactivation of HRPC is a consequence of the oxidation of aniline (Junker et al. 2012). Without this reaction, HRPC is stable (Fig. S-5B). In the case of GOD, the activity also dropped rapidly during the formation of PANI-ES (Fig. 7). After 10 min reaction time, no GOD activity could be detected anymore with the method used. Like in the case of HRPC, GOD inactivation is a consequence of  Table 1). The reaction times were t = 0 s (1), 15 s (2), 30 s (3), 1 min (4), 3 min (5), 10 min (6), 60 min (7), 5 h (8) and 24 h (9). a, b Measured spectra for t = 0-10 min and 10 min-24 h, respectively. c, d Time-dependent changes of A 1000 , A 750 , A 500 , A 420 , and A 300 . In c and d, the lines between the data points are drawn for guiding the eyes. Please note that the spectra measured for t = 15 s and t = 30 s are only approximate spectra due to the rapid changes during the first phase of the reaction, also taking place during the recording of the spectra ▸ the oxidation of aniline, since without aniline, GOD is stable ( Fig. S-6B). The rapid loss in enzyme activities in the "ideal cascade reaction mixture" can be compared with the observed rapid consumption of aniline during the initial phase of the reaction (Fig. 4). The enzymatic oxidation of aniline (see Scheme 2) is completed after less than 30 min, while spectroscopic changes due to non-enzymatic processes still take place for several hours (Fig. 3). The rapid loss of GOD activity is in agreement with the low amounts of H 2 O 2 detected in the samples withdrawn from the "ideal reaction mixture" (Fig. 5). Without catalytically active GOD, no H 2 O 2 can form from d-glucose.
The change of the initially present amount of d-glucose in the "ideal cascade reaction mixture" (100 mM) was  Table 1), as determined by analyzing samples withdrawn from the reaction mixture, see section "Experimental". The lines between the data points are drawn for guiding the eyes. The error bars represent standard deviations from two or three measurements  Table 1) during the course of the reaction, as determined by analyzing samples withdrawn from the reaction mixture, see section "Experimental". The error bars represent standard deviations from three measurements  Fig. 6 Changes in the relative activity of HRP during the course of the reaction for the "ideal cascade reaction mixture" (filled symbols, see Table 1) and of the "reference reaction mixture" (empty symbols), as determined by analyzing samples withdrawn from the reaction mixtures. Hundred percent activity refers to the initial rate of reaction at time zero. The error bars represent standard deviations obtained from three measurements, see section "Experimental". The lines between the data points are drawn for guiding the eyes Changes in the relative activity of GOD during the course of the reaction for the "ideal cascade reaction mixture" (see Table 1), as determined by analyzing samples withdrawn from the reaction mixture, see section "Experimental". Hundred percent activity refers to the initial rate of reaction at time zero. The error bars represent standard deviations obtained from three measurements. The lines between the data points are drawn for guiding the eyes determined during the reaction by withdrawing small volumes from the reaction mixture, then eliminating GOD and HRP by ultrafiltration, and finally analyzing the ultrafiltrate for d-glucose content, see "Experimental". As seen from the results shown in Fig. 8, after an initial consumption of 5 mM d-glucose, the d-glucose concentration in the cascade reaction mixture remained at about 95 mM. This initial d-glucose consumption and the discontinuity of d-glucose oxidation correlate well with the observed inactivation of GOD (see Fig. 7). In summary, from previous investigations of the "reference reaction," it is clear that HRPC inactivation occurs during the formation of PANI-ES products from aniline (Junker et al. 2012). For GOD, the situation is basically the same. Native GOD is inactivated during PANI-ES product formation.

EPR spectrum of the PANI-ES products obtained from aniline with the cascade reaction
Based on the UV/Vis/NIR absorption measurements, it is expected that the PANI-ES products obtained from aniline with the enzymatic cascade reaction contain unpaired electrons (see Introduction). Therefore, the products should show EPR spectra. This was confirmed by analyzing three different cascade reaction mixtures of the composition shown in Table 1 after t = 24 h, with the only exception that the GOD concentration was either 0.09, 0.15, or 0.30 μM instead of 0.20 μM, the "ideal" concentration. There is no doubt that the PANI-ES products of the enzymatic cascade reaction contain paramagnetic centers. The spectrum for [GOD] = 0.15 μM measured after t = 24 h is shown in Fig. 9, together with the EPR spectrum of the "reference reaction mixture," also measured after t = 24 h. Both spectra are very similar, with only slightly different G-factors (2.0063 for the cascade reaction and 2.0064 for the reference reaction, respectively). The EPR signal intensity for the PANI-ES products obtained with the reference reaction was higher than in the case of the PANI-ES products obtained with the cascade reaction (Fig. 9). The EPR signal intensities correlate with the higher absorption in the NIR region for the reference reaction mixture, as compared to the cascade reaction mixture: A 1000 (reference, L = 1 mm, t = 24 h) = 1.2 (Fig. 2d) and A 1000 (cascade, L = 1 mm, t = 24 h) = 0.8 (see curve 5 in Fig. S-10), respectively.

Comparison of the PANI-ES products obtained with the "ideal cascade reaction" and with the "reference reaction"
The question we asked ourselves at the beginning of the investigation was: can PANI-ES products be obtained in the presence of AOT vesicles as templates at pH = 4.3 in high yield as stable aqueous suspension from aniline with d-glucose, GOD and HRPC (and dissolved O 2 )? And, if the answer would be "yes": How do the spectroscopic properties of the PANI-ES products obtained with this cascade reaction under "ideal" conditions compare with those resulting from a similar (optimized) reaction without using d-glucose and GOD but addition of H 2 O 2 at the beginning of the reaction ("reference reaction")? Although both reactions involve the  Table 1), as determined by analyzing samples withdrawn from the reaction mixture, see section "Experimental". Error bars represent standard deviations obtained from three measurements use of enzymes as catalysts, the key difference between the two systems is the in situ formation of H 2 O 2 from d-glucose with GOD in the case of the cascade reaction, as compared to the manual addition of H 2 O 2 at the beginning in the case of the reference reaction.
In this work, we have shown that conditions can be found for the desired GOD-and HRPC-catalyzed synthesis of PANI-ES products with an aniline conversion of ~ 90% (Fig. 4), similarly to the "reference reaction" (Junker et al. 2012). With the experimental conditions which we considered ideal for the cascade system (Table 1), the products obtained had UV/Vis/NIR absorption and EPR spectra that are typical for PANI-ES products (Figs. 2d,9). The key spectroscopic differences between the two systems can be summarized as follows: (i) The absorption of the obtained PANI-ES products at λ ~ 750 nm, A 750 , was higher for the cascade reaction than for the reference reaction. In contrast, A 420 and A 1000 for the cascade reaction were lower than for the reference reaction (Fig. 2d).
(ii) Although the EPR spectra were very similar, the EPR signal intensity for the PANI-ES products obtained with the cascade reaction was lower than in the case of the reference reaction (Fig. 9).
The observations (i) and (ii) support the previous conclusions that electronic PANI-ES transitions involving unpaired electrons are centered around λ = 420 and λ ≥ 800 nm, see Introduction.

Conclusions
There is no doubt that PANI-ES products can be obtained in aqueous 0.1 M NaH 2 PO 4 solution (pH = 4.3) in the presence of AOT vesicles in a two-enzyme cascade reaction from aniline, d-glucose, GOD and HRPC (and dissolved O 2 ) at T ~ 25 °C within t = 24 h. However, there is some drawback of reaction efficiency: GOD is rapidly inactivated during the reaction (Fig. 7). High concentrations of GOD (0.2 μM) and d-glucose (100 mM) are required for achieving high aniline conversion (Table 1). Substantial stabilization of GOD via chemical modification(s) is unlikely. HRPC (0.92 μM) is also rapidly inactivated during the reaction (Fig. 6) (Junker et al. 2012), and attempts to significantly stabilize the enzyme were unsuccessful (Junker et al. 2013). Further studies on the enzymatic synthesis of conducting PANI, for example by using an AOT/SDS (sodium dodecylsulfate) bicontinuous cubic phase as template (Jin et al. 2020), are certainly worth pursuing. Depending on application, enzyme inactivation might not be a crucial problem (Zhang et al. 2020), because enzymes are biological and biodegradable macromolecules. Preliminary experiments showed that the cascade reaction we elaborated is a useful extension to the reference reaction applied so far in the study of artificial cell-like systems, as briefly mentioned in the introduction (Kurisu et al. 2019).
Our investigations can be compared to previous reports on the synthesis of PANI from aniline with d-glucose/GOD (no HRP, no template) (Kausaite et al. 2009;German et al. 2017) or from aniline with d-glucose and a mixture of GOD and hemoglobin immobilized on a glassy carbon electrode consisting of adsorbed carboxylic acid functionalized polystyrene (Zhang et al. 2013). A direct comparison is, however, difficult. The aim of this previous work and the reaction conditions were very different from ours. Moreover, the products obtained did not have the characteristic spectra of PANI-ES, and the PANI products obtained precipitated as micrometer-sized particles (Kausaite et al. 2009;German et al. 2017). This is very different from the work we describe here, where prevention of product precipitation was set as one of the goals of the work, successfully achieved by the use of AOT vesicles as soft, dispersed additives that promote and guide the enzymatic cascade reaction (Serrano- ).