Chapter 2. Spatiotemporal effects of a bioautocatalytic chemical wave revealed by
2.3. Results and discussion
2.3.2. Implementation of an isotopically labeled trigger
In order to provide a solid evidence for the higher speed of enzyme-accelerated transduction relative to passive transduction – and overcome the reproducibility problems mentioned above – we subsequently implemented isotopically labeled ATP as a trigger of the chemical wave.
When 30 μL 13C10-ATP is injected to the inlet port of the 600 μL section of the drift cell, it initiates the bienzymatic autocatalytic process. However, since the substrate AMP (0.1 mM) contains 12C isotope, the newly produced ADP and ATP molecules are not labeled. Thus, we can consider the chemical wave produced by the reaction as a carrier of unlabelled ADP and ATP. The labeled 13C10-ATP migrates mainly due to passive transduction (convection, diffusion, advection (bulk movement of substance due to flow of fluid)), and its role as the reaction trigger ceases as populations of the newly synthesized unlabelled ADP and ATP molecules grow. In this experiment, following the injection of 10-2 M 13C10-ATP, the two waves – corresponding to unlabelled and labelled ADP/ATP species – were recorded.
Experimental data were treated according to the following formulae:
Rel. yield ( C12 ) =𝐼 𝐼ADP+𝐼ATP
AMP+𝐼ADP+𝐼ATP+𝐼13CAMP+𝐼13CADP+𝐼13CATP (eq. 2.4) Rel. yield ( C13 ) =𝐼 𝐼13CADP+𝐼13CATP
AMP+𝐼ADP+𝐼ATP+𝐼13CAMP+𝐼13CADP+𝐼13CATP , (eq. 2.5)
smoothed, normalized, and displayed in time domain (Figure 2.7). The normalization was done by scaling the values of the data points to the maximal value within the displayed time range. Interestingly, even at the high concentration of the trigger ATP (10-2 M), a difference in the propagation speed could be observed (Figure 2.7, top). In this particular experiment, accelerated transduction led to a sharp increase of the signal (blue trace), whilst passive transduction of labeled ATP led to a gradual increase of the ionizable species (red trace). At a lower concentration of the 13C-ATP, used as the trigger, the edges of the accelerated and the passive fronts were less sharp (Figure 2.7, bottom) but the difference between half-maxima of the normalized profiles was greater: 740 vs. 93 s in the case of 5 × 10-3 and 10-2 M trigger, respectively (Figure 2.7). The differences between the accelerated and the passive transduction could be observed in all the replicate experiments using isotopically labeled trigger (Figure 2.8) even though both fronts shifted due to the above-mentioned imperfections of the experimental setup. Based on this, one may conclude that – in the present experimental setup – the onset of the increase is determined by the advective transport, and the chemical
28
wave has a relatively small influence when compared with the advection. It is also worthwhile noting that the inspection of the unnormalized datasets also led to the observation of a higher speed of the bioautocatalytic chemical wave, as compared with the passive transduction (Figure 2.9). Nonetheless, the difference is less evident than in the plot of the normalized data (Figure 2.7).
Figure 2.7 Transduction of labelled and unlabelled ATP along the drift cell. Concentration of the 13C10-ATP
trigger: 10-2 M (top) and 5 10-3 M (bottom). See Figure 2.8 for replicates of the concentration 5 10-3 M. Figure 2.9 presents unprocessed extracted ion currents from the experiment with the
concentration 5 10-3 M. Exponential smoothing with a time constant of 4.1 s has been applied, and followed by normalization (scaling to the maximal value). Dashed blue line denotes the time lapse between half-maxima of the normalized curves (0.5 level) corresponding to the passive and accelerated chemical transduction: 93 and 740 s in the case of 10-2 M and 5 10-3 M trigger solution, respectively.
29
Figure 2.8 Transport of labelled and unlabelled ATP along the drift cell. Concentration of the 13C10-ATP trigger:
5 10-3 M. These three results are replicates of the result shown in Figure 2.7 (lower panel). Exponential smoothing with a time constant of 4.1 s has been applied, and followed by normalization (scaling to the maximal value).
30
Figure 2.9 Transduction of labelled and unlabelled ATP along the drift cell. Raw data (extracted ion currents) obtained in the same experiment as the one illustrated in Figure 2.7 (bottom). Concentration of the
13C10-ATP trigger: 5 10-3 M.
2.4. Summary
The results presented above highlight the possibility of using enzyme-aided chemical waves as a means to transmit binary signals. However, to increase practicality of this scheme, one may consider incorporating a “quenching” mechanism, so that the system can return to its initial state (low level of ADP and ATP in the drift cell). This might be achieved by incorporating a mechanism to hydrolyze ADP and ATP to AMP once the signal reaches the
31
outlet end of the drift cell. It is also worthwhile noting that the experiment using the isotopically labelled trigger ATP does not require additional control because spatiotemporal behaviour of two populations of ATP are recorded in one run.
It was previously reported that hydrodynamic instabilities can be induced by chemical reactions, and subsequently influence convective motion.58 The current experimental model shows that a chemical reaction itself can produce an apparent “motion” of reactant molecules, which does not seem to be directly connected with advection (that certainly exists in the current experimental system). Chemical waves are often discussed in the context of oscillating reactions, with the most prominent example being the Belousov-Zhabotinsky reaction,62 which produces oscillating chemical waves composed of iron complexes and other species.
Unlike Belousov-Zhabotinsky reaction, the chemical wave reported here has a character of a moving front of elevated concentration, without apparent oscillations (except those attributed to imperfections of the experimental setup and formation of eddy currents). The proposed system is suitable for operation at the room temperature. In addition, unlike many oscillating reaction schemes, the system reported here uses an aqueous buffer at neutral pH, which renders it biocompatible. We believe more chemical wave reaction schemes can be demonstrated in future, for example, implementing various ways of molecular amplification.
However, not every amplification reaction can offer the advantage of sending a chemical signal in a seamless fashion. For instance, the most famous one – polymerase-chain reaction (PCR)63 – requires periodic changes of temperature, which certainly eliminates the advantage of sending chemical signals without the need for actuation by periodic input of thermal energy.
The current results show the possibility to apply time-resolved mass spectrometry in the monitoring of chemical waves. In this demonstration, a relatively low temporal resolution (seconds) was sufficient to observe the between the passive transduction and reaction front.
While in previous work we applied optical and mass spectrometric detection in the monitoring of a convection process,64 here time-resolved mass spectrometry61 was successfully used to show the difference in the migration speed of chemical waves due to bioautocatalytic reaction and accompanying non-chemical processes.
Modelling biological and chemical processes has been of interest to many scholars.
Relevant insights on spatially and temporally resolved processes (such as oscillating reactions) were presented by Arthur Winfree.65 It might be of potential interest to study the bioautocatalytic process described here in view of the published models of reaction diffusion systems, such as Kolmogorov-Petrovsky-Piskounov’s56 or Fisher’s57 equation. With such
32
models one could potentially disentangle the contributions of enzymatic reaction and passive dispersion of molecules on the propagation of the moving front of ADP/ATP. However, the current experimental setup has limitations that preclude the use of data in quantitative analysis required to study the mechanism of the chemical wave induced by the bioautocatalytic process. This is due to: (i) the use of an ambient ionization system without separation of analytes; (ii) presence of various components (e.g. buffer salt) in the aqueous reaction mixture;
(iii) the use of ion trap mass analyzer; (iv) formation of a temperature gradient in front of the mass spectrometer’s orifice. Overcoming these drawbacks in future should augment quantitative capabilities of the spatiotemporal monitoring by time-resolved mass spectrometry.
However, it is also appealing to implement other detection system when quantitative information is required: for example, fluorescence resonance energy transfer probes, immobilized on the wall of the drift cell. Nonetheless, such optical assays would eliminate the possibility of using isotopic labels to distinguish between the trigger ATP and the ATP produced in the course of the reaction. Therefore, in future studies, it would be interesting to use both mass spectrometry as well as optical detection systems as complementary tools.
In summary, the above results show the possibility of transmitting a chemical signal (ADP/ATP) over the range of centimetres with a speed exceeding that of other transport phenomena occurring under the same conditions. We believe the chemical wave propagating ADP and ATP (the energy carriers in biochemistry) can act as a turn-on/off trigger, and be used to initiate various processes in bioengineered systems incorporating tissues, cells, (bio)nanorobots, or various biocatalysts (e.g. kinases). Coupling this scheme with other biomolecular constructs (e.g. enzyme systems supplying the AMP and PEP substrates) and microfluidics (e.g. parallel capillary channels), and using anti-convective media (e.g.
hydrogels), may further enable multiplexed transmission of binary data. Since the speed of chemical waves – propagating according to the proposed scheme – is related to the kinetic properties of the enzymes involved in the bienzymatic process (cf. Figure 2.2), modification of these enzymes (for example, through directed evolution), or taking advantage of non-enzymatic amplification schemes,66 might further increase the speed of wave propagation, which would augment practicality of this approach for real-world applications.
33 biotechnology.67-69 Being biocatalysts, enzymes speed up chemical reactions by lowering the activation energy.70, 71 Enzymes provide improved efficiencies by increasing the reaction rates.
Due to their specificity, they ensure high purity of the synthesized products. Appropriate assays need to be developed in order to characterize catalytic properties of enzymes. Ideally, such assays should enable reaction monitoring within certain periods of time, so that any loss of catalytic activity can be easily noted. Due to the high price of enzymes, it is also desirable that these methods consume as little sample (biocatalyst) as possible.72 As outlined in
Chapter 1, gas chromatography (GC)-based methods have been in common use since the
1950s.27, 28, 73 In order to ascertain satisfactory sensitivity and selectivity, high-performance detectors are normally connected to outlets of GC columns. For instance, mass spectrometry (MS) is one of the popular detectors hyphenated with GC systems because it can assist identification of the eluting molecules based on their mass-to-charge ratios (m/z) and fragmentation patterns. Owing to their versatility and ruggedness, the hyphenated GC-MS systems are used in routine analyses in environmental science,74 forensic science,75 geological sciences,76 food chemistry,32 cosmetics77 and biochemistry78 – to name just a few application areas.In general, to obtain high-quality results, and to minimize manual effort, automated sampling devices are combined with GC instruments. Autosamplers make the entire analytical procedure straightforward, and they facilitate high-throughput operation. Clearly, superior repeatabilities can be obtained with robotised autosamplers79 as compared with the manual injection. GC instruments (sometimes equipped with autosamplers) are often used in discrete analyses – one sample at a time. This mode of operation can provide data on the yield of chemical or biochemical reactions at a given time point.80, 81 To obtain data revealing progress
34
of a chemical reaction, aliquots of the reaction mixture are withdrawn from the reaction vessel, and transferred to the GC apparatus.82 This “low-tech” sampling step conspicuously curtails the advantage of utilizing automated systems since the key part of the analytical procedure (i.e.
sampling routine) is performed by humans. Moreover, the involvement of manual operation makes it difficult to time the sampling, thus reducing usefulness of the resulting data in post-run kinetic evaluations. Hence, there is a need to develop automated sampling tools which could easily be coupled with inlets of GC instruments, and enable assaying chemical processes in real time. Such systems could provide optimum conditions (e.g. temperature control, stirring) for the studies of dynamic processes, and reduce the delay time between sampling and analysis.
Microcomputers and microcontrollers are vital for the construction of analytical instrumentation.83, 84 In fact, most commercial instruments incorporate customized microcontroller circuits. In the past few years, many universal microcontrollers have become available. One prominent example is the Raspberry Pi – a single printed circuit board (PCB) microcomputer, which was introduced to the market in 2012.85, 86 The original purpose of the Raspberry Pi was to enhance education in computer science, electronics, automation, and robotics.87 Plethora of practical uses of this universal platform have emerged soon after the sales had started.88 Typical applications include: data logging, sensing, controlling displays, and automation of household appliances.89 The introduction of the Raspberry Pi – as well as other platforms such as Arduino,90 Netduino,91 mbed,92 or Beaglebone93 – has inspired home-grown innovators, and fostered creativity of students. These versatile and easy-to-use electronic circuits nowadays find applications in scientific instrumentation, including the construction of analytical systems.94-96 In this study, we aimed to develop a facile automated injection system for GC to accommodate monitoring of dynamic samples, which would incorporate minimum number of mechanical parts, and be controlled by the Raspberry Pi microcomputer. The system comprises control panel, peristaltic pump, two pinch valves, and GC-MS apparatus (Figure 3.1). We further demonstrate the applicability of this system in the monitoring of transesterification reaction catalysed by single microbeads containing immobilised lipase enzyme as well as extraction of natural products from small amounts of plant tissues.
3.2. Experimental section
3.2.1. Samples and chemicals
A widely available thermostable recombinant lipase (from Candida antarctica, expressed in
35
Aspergillus niger; E.C. 3.1.1.3) adsorbed on macroporous acrylic resin (specific activity: ≥ 5,000 U g-1, microbeads) was used as the main model sample (cf. section 3.3.3.). The samples of lemon and kumquat (cf. section 3.3.4.) were obtained from local grocery shops. Small fragments (~ 2 mm, < 5 mg) of the fruit peel were obtained using stainless steel blade.
Isopropenyl acetate, LC-MS grade acetonitrile, D-limonene, β-pinene, γ-terpinene, and thymol were all purchased from Sigma-Aldrich (St Louis, MO, USA). 1-Butanol was purchased from Merck (Darmstadt, Germany).
3.2.2. Interface
The interface comprises a Y-junction (P-779, PEEK NanoTight Union, OD 1/16-inch; IDEX Health & Science, Oak Harbor, WA, USA) and two solenoid pinch valves (12 V DC, P/N 075P2NO12-01S and P/N 075P2NC12-01S; Bio-Chem Fluidics, Boonton, NJ, USA): out of which one is “normally closed” (NC) while the other one is “normally open” (NO; Figure
3.1). The valves were attached onto a flat support made of 1-cm-thick plywood with plastic
holder and a small amount of modelling clay (Sugru; FormFormForm, London, UK), supported by four stainless steel breadboard-type posts (TR75/M-P5, 1.2 cm, 7.5-cm long).The sample was delivered by a peristaltic pump (Ismatec ISM936D, IPC Series, 8 channels;
IDEX Health & Science). The pump pushed the liquid along a 35.4-cm-long section of Tygon tubing (ID 0.25 mm, OD 2.07 mm, cat. No. SC002; IDEX Health & Science), connected to a 28.6-cm-long section of polytetrafluoroethylene (PTFE) tubing (OD 1.58 mm, ID 0.3 mm,cat.
No. 58702; Supelco, Bellefonte, PA, USA), and – further downstream – to the Y-junction (via a 4.8-cm-long section of polyimide-coated fused silica capillary; ID 150 μm, OD 375 μm, cat.
No. 1010-32132; GL Sciences, Tokyo, Japan). The other port of the Y-junction was connected (via another 4.8-cm-long section of polyimide-coated fused silica capillary, 15-cm section of PTFE tubing, and a 31.4-cm-long section of silicone tubing) to the NO pinch valve. The outlet end of the silicone tubing – passing the pinch valve – was dipped into the waste collector (20-mL glass vial). The outlet port of the Y-junction was connected via a 6-cm-long section of silicone tubing to a 1/16-inch union (IDEX Health & Science). The other side of this union was fitted with a metal needle removed from a discarded solid-phase microextraction (SPME) fibre assembly (23 ga, cat. No. 57284-U; Supelco). The NC pinch valve was mounted on the 6-cm silicone tubing joining the Y-junction and the 1/16-inch union. While on stand-by, the liquid – delivered by the pump – was directed to the waste reservoir because the NC pinch valve prevented its entry to the stainless steel needle. However, when the NC valve got open, and the NO valve got closed, the liquid was directed towards the metal needle, and further on
36
to the injection port of the GC instrument.
The sample vial was placed inside a small thermoshaker (Vortemp 56 EVC; National Labnet Company, Woodridge, NJ, USA) to assure optimum conditions for the studied process (e.g. reaction, extraction). To ascertain seamless operation of the system it is recommended that the Tygon tubing is replaced every day, and the system rinsed before operation with the solvent used in the analysis for
15 min. Note that the room temperature during the
experiments described here was 20 C; it is expected that huge variations of ambient
temperature could potentially affect reproducibility of the described method.Figure 3.1 System for online sampling and sample introduction to gas chromatography – mass spectrometry using pinch valves: (A) device layout; (B) view of the assembled device. NO – normally open valve. NC – normally closed valve. Note well: The commercial autosampler (left-hand side) was not used in this study. The lid of the mini-thermoshaker (upper right-hand side) was closed during the experiment.
3.2.3. Electronic control system
In order to control operations of the on-line GC-MS injection system, we implemented a single board microcomputer – Raspberry Pi (Type B) with a 700-MHz central processing unit (ARM1176JZF-S) and 512-MB random-access memory (Figure 3.2). It was set up with a
37
Linux-related operating system (Debian/Raspbian Wheezy, 2012) loaded from an 8-GB secure digital (SD) memory card (Transcend, Taiwan). The microcomputer was connected to a multifunction input/output extension PCB (Pi PLUS; web4robot.com, Parkville, MD, USA) via its general-purpose input/output (GPIO) interface to facilitate sending out digital control signals, and further development of the system. For convenience of operation and programming, it was also fitted with standard peripherals (keyboard, mouse, WiFi dongle) and a miniature (~ 4.4 inch) liquid crystal display (LCD) monitor (connected to the RCA-type port of the Raspberry Pi PCB). The control program – written in C – was referenced to the Pi PLUS function codes and header files provided by the manufacturer (web4robot.com), compiled with the GNU compiler on the Raspberry Pi, and deployed. The “sleep()” command was used to time the injection. The Pi PLUS extension PCB was connected via digital interface pins to three relay boards (2 relays each; Songle Relay, Ninbo, China). The relays controlled the main functions of the system – including peristaltic pump (on/off, change of direction), operation of the pinch valves, and sending a trigger signal to the GC-MS instrument. All the PCBs and the monitor were fitted into an acrylic stand (Muji, Tokyo, Japan). The device also featured a “START” button which activated the functions of the program. The flow rate was regulated by adjusting the electric potential on the pin No. 5 of the peristaltic pump interface – this was enabled by a 50 k potentiometer (RPOT) in a voltage divider circuit (Figure 3.2). For convenience, and to avoid electrical interferences, the main components were powered from separate power supplies – Raspberry Pi: 5 V, 1 A;Pi PLUS: 9 V, 1.3 A; pinch valves: 12 V, 1.5 A; monitor: 9 V, 1 A.
Figure 3.2 Layout of the electronic connections of the device microcontroller unit incorporating the Raspberry Pi microcomputer, general purpose input / output extension board, and relay boards.
38
3.2.4. Apparatus and method
The testing as well as the subsequent analyses were performed using a commercial gas chromatograph (TRACE GC; Thermo Fisher Scientific, Waltham, MA, USA) coupled with a single quadrupole mass spectrometer (ISQ; Thermo Fisher Scientific). Capillary column with non-polar phase comprising 5% phenyl methyl polysiloxane (TRACE SPB-5; length: 60 m;
ID 0.53 mm; film thickness: 1.5 μm) was used. The eluting compounds were ionized by EI source operating at 70 eV. Internal calibration of the mass spectrometer was conducted using perfluorotributylamine (Alachua, FL, USA). Helium gas was used as the mobile phase with the constant flow rate of 1 mL min-1 at a pressure of 10 kPa. The injector was kept at 280 C.
Split injection was used with the split ratio of 50. In the first part of the study (enzymatic reaction), the following separation conditions were used: Column temperature was initially set to 40
C. Following 3 min of separation, the temperature was gradually increased reaching
60 C after 2 min (ramp: 10 C min-1), then increased to 100C (ramp: 35 C min
-1). In the final stage of the run, temperature was increased to 150 C (ramp: 20C min
-1). It was kept constant for 5 min (till the end of the analysis). The mass spectrometer was set to record ions in the range of 42-250 u e-1. The analyte (butyl acetate) species were monitored at the m/z 43 u e-1while the internal standard (limonene) species were monitored at the m/z 93 u e-1. In the second part of the study (extraction of plant tissue samples), the following separation conditions were used: Column temperature was initially set to 40 C, and after 3 min, the temperature was increased to 120C (ramp: 30 C min
-1). Subsequently, temperature was increased up to 180 C (ramp: 15 C min-1), and it was kept constant for 7 min (till the end of analysis). In this experiment, the mass spectrometer was set to record ions in the range of 40-250 u e-1. The analyte (limonene, pinene, terpinene) species were monitored at the m/z 93 u e-1while the internal standard (thymol) species were monitored at the m/z 135 u e-1.3.2.5. Data treatment
The data were acquired using the Xcalibur software (ver. 2.1.0 SP1.1160; Thermo Fisher Scientific), and saved in the “.raw” files. The raw data (extracted ion currents, EICs) were copied from the Xcalibur software (ver. 2.1.0 SP1.1160, Qual Browser) to Excel software (ver.
The data were acquired using the Xcalibur software (ver. 2.1.0 SP1.1160; Thermo Fisher Scientific), and saved in the “.raw” files. The raw data (extracted ion currents, EICs) were copied from the Xcalibur software (ver. 2.1.0 SP1.1160, Qual Browser) to Excel software (ver.