Analysis of double-emulsion droplets with ESI mass spectrometry for monitoring lipase-catalyzed ester hydrolysis at nanoliter scale

Microfluidic double-emulsion droplets allow the realization and study of biphasic chemical processes such as chemical reactions or extractions on the nanoliter scale. Double emulsions of the rare type (o1/w/o2) are used here to realize a lipase-catalyzed reaction in the non-polar phase. The surrounding aqueous phase induces the transfer of the hydrophilic product from the core oil phase, allowing on-the-fly MS analysis in single double droplets. A microfluidic two-step emulsification process is developed to generate the (o1/w/o2) double-emulsion droplets. In this first example of microfluidic double-emulsion MS coupling, we show in proof-of-concept experiments that the chemical composition of the water layer can be read online using ESI–MS. Double-emulsion droplets were further employed as two-phase micro-reactors for the hydrolysis of the lipophilic ester p-nitrophenyl palmitate catalyzed by the Candida antarctica lipase B (CalB). Finally, the formation of the hydrophilic reaction product p-nitrophenol within the double-emulsion droplet micro-reactors is verified by subjecting the double-emulsion droplets to online ESI–MS analysis. Graphical abstract Supplementary Information The online version contains supplementary material available at 10.1007/s00216-022-04266-2.


2) Expression and purification of lipase CalB
The lipase CalB was expressed and purified with an N-terminal Asp5-tag and a C-terminal His6-tag adopting previously published procedures [1], [2]. The Asp5 sequence enhances soluble expression and secretion of the enzyme. The gene coding for the amino acid sequence specified in Figure S1 was synthesized by ThermoFisher (GeneArt) with codon optimization for E. coli expression and NcoI and XhoI restriction cleavage sites at the 5' and 3' ends, respectively. The synthesized gene was inserted into the expression vector pET26b using these restriction sites. The construct is expressed with a pelB signal peptide from the pET26b plasmid for secretion into the cell medium. E. coli BL21(DE3) cells were transformed with the expression plasmid. For preparative expression, the cells were grown at 37°C in LB medium to an OD600 of 0.6, and expression was induced with an isopropyl-β-D-thiogalactopyranoside (IPTG) concentration of 0.2 mM. The culture was incubated at 20 °C and 220 rpm overnight. On the following day, the cells were pelleted for 30 min at 5000x g. The supernatant was supplemented with 20 mM NaH2PO4 and 15 mM imidazole and applied to a HisTrap FF crude 5 mL column. All chromatography steps were conducted at an Äkta pure (Cytiva) FPLC instrument. After washing with 8 column volumes (CV) binding buffer (20 mM NaH2PO4 pH 7.4; 0.5 M NaCl; 15 mM imidazole), the protein was eluted with 5 CV of a linear gradient of binding buffer to elution buffer (20 mM NaH2PO4 pH 7.4; 0.5 M NaCl; 500 mM imidazole). Fractions containing CalB were pooled, concentrated to 1.5 mL, and applied to Superdex75 (16/60) size exclusion chromatography using a buffer of 20 mM Tris-HCl pH 8.0 with 200 mM NaCl). Active fractions were pooled and concentrated to a protein concentration of 2.3 mgmL -1 . Protein concentrations were determined using the absorption at 280 nm (measured at a ThermoFisher NanoDrop 2000c instrument) with a calculated extinction coefficient [3] of 41285 M -1 cm -1 or 1.175 mL mg -1 cm -1 .

3) Gene coding for the lipase CalB
Sequence of the lipase CalB construct used in this study. The expressed protein possesses an N-terminal Asp5-tag (red) and a C-terminal His6-tag (green). Shown in blue is a linker sequence to improve the accessibility of the polyhistidine tag. The gene for expression of this construct was synthesized by ThermoFisher (GeneArt) with codon optimization for E. coli expression and NcoI and XhoI restriction cleavage sites at the 5' and 3' ends, respectively.  Figure S2 shows that the double emulsion droplets stayed intact when they transited from the 1/16'' OD 300 µm ID PTFE tubing in a silanized 360 µm OD 150 µm ID FS capillary. Occasionally, however, a DED splits into two or more daughter droplets due to shear forces. This occurred when the DEDs transitioned from the PTFE tube into a non-silanized FS or PTFE capillary or after migrating a few centimeters in these capillaries. It becomes clear that the addition of the lipase CalB led to a distortion of the droplets' shape when Novec 7500 oil with 0.5% PicoSurf was used as oil phase o2 ( Figure S3). As Figure S3 shows, the shape of the DEDs was deformed during their downstream transport in the PTFE tubing. While the shape of the aqueous shell phase was approximately spherical when the DEDs detached from the FS capillary, it was elongated as the DEDs traveled through the PTFE tubing, resulting in a plug shape. The shape of the n-dodecane core droplets seemed not to be affected. Furthermore, the regular oil spacing between the DEDs was lost as the Novec 7500 oil flowed alongside the DEDs. Moreover, the encapsulation of the DEDs' core n-dodecane droplets got lost after a certain traveled distance. The n-dodecane droplets seemed to be only attached to the aqueous phase droplet instead of being encapsulated by the aqueous phase. Presumably, the observed deformation in shape was caused by the similar densities of the aqueous shell phase w and the continuous oil phase o2. Adding lipase to the aqueous shell phase caused an increase in the aqueous phase's density. The deformation in shape was not observed when the oil FC-40

7) Reproducibility of droplet diameter
Video S1 Double emulsion generation over 1 min. We got uniform droplets with respectively 1 one-core droplet, 30 two-core droplets, and 2 three-core droplets. Scale bar measured on the 150 µm OD fused silica capillary.

8) Fluigent RayDrop double emulsion droplet generator
In the following, the advantages and disadvantages of the Fluigent RayDrop and our set-up are compared. With our set-up, which consists of the fusion of a T-junction and a co-flow, DEDs can be produced reproducibly with one or multiple core droplets. The advantage of this system is that the sizes of the capillaries and thus also the shell and core droplets can be easily exchanged. The droplets can also be changed due to the different hydrophilic and hydrophobic surfaces. Therefore, even small changes to the system can result in a difference in the droplets. In method development, for example, greater flexibility is desirable.
On the other hand, the commercial Fluigent RayDrop device has a fixed diameter for generating the DEDs. This means there is only a limited volume where the DEDs can be produced reproducibly. The core droplets are often large in volume relative to the shell phase. The generation of smaller droplets (see Fig. S4C) is complex. In addition, droplets with one core are possible with the RayDrop device due to the structure of two simultaneously generating co-flows droplet generators (The nozzle in Fig. S4A and B). On the other hand, the RayDrop system is more robust concerning relocation of the device or exchange of the capillaries.

10) Calibration and reaction monitoring of p-nitrophenol
For the calibration of p-nitrophenol, an internal standard of nitrocatechol was used. Nitrocatechol can mainly compensate the influence of the ESI spray. As seen in Fig. S5A, the signals have a high standard deviation. As known, a high surfactant concentration influences the signals [6]. Therefore, a higher product concentration should be preferable for calibrating our signals. As proof of concept, a reaction monitoring over 120 min showed increased product/internal standard intensity (Fig. S5B). However, after 120 min, the product concentration was still too low for quantitative detection.

11) Direct injection measurement in a 1.5 mL reaction vessel with n-decane
To compare our double emulsion droplet reaction with a direct injection method, we performed the lipase-catalyzed reaction in a 1.5 mL reaction vessel (RotiLab® Microtube). For this purpose, 0.068 mg/mL lipase and 10 µM nitrocatechol as internal standard were diluted in a 10 mM ammonium acetate water solution. Only 10 µM nitrocatechol in water was diluted for the blank solution. The educt 5 mM p-nitrophenyl palmitate was dissolved in n-decane and 0.2 mL of it was mixed with 0.2 mL of the water phase. After overnight incubation in a shaker at 37 °C, the product and internal standard in the water phase were detected.

Fig. S5
A: Calibration curve of p-nitrophenol and internal standard 10 µM nitrocatechol. More than 60 droplets were analyzed per data point. In B: Reaction monitoring over 120 min of the heterogenic catalyzed reaction by the lipase CalB with the product p-nitrophenol. More than 17 droplets were analyzed per data point. The peak height was used for the calculation of both graphs. Furthermore, we used the Grubbs's test to identify droplets that were sprayed irregularly.