3D-Printed High-Density Droplet Array Chip for Miniaturized Protein

Mar 17, 2017 - *E-mail: [email protected] (Y.Z.)., *E-mail: [email protected] (S.Y.)., *E-mail: [email protected] (Q.F.). ... We fabricated high-densit...
0 downloads 9 Views 5MB Size
Download PDF
Research Article www.acsami.org

3D-Printed High-Density Droplet Array Chip for Miniaturized Protein Crystallization Screening under Vapor Diffusion Mode Yi-Ran Liang,† Li-Na Zhu,† Jie Gao,† Hong-Xia Zhao,‡ Ying Zhu,*,† Sheng Ye,*,‡ and Qun Fang*,† †

Institute of Microanalytical Systems, Department of Chemistry and Innovation Center for Cell Signaling Network, and ‡Life Sciences Institute and Innovation Center for Cell Signaling Network, Zhejiang University, Hangzhou, 310058, China S Supporting Information *

ABSTRACT: Here we describe the combination of threedimensional (3D) printed chip and automated microfluidic droplet-based screening techniques for achieving massively parallel, nanoliter-scale protein crystallization screening under vapor diffusion mode. We fabricated high-density microwell array chips for sitting-drop vapor diffusion crystallization utilizing the advantage of the 3D-printing technique in producing high-aspect-ratio chips. To overcome the obstacle of 3D-printed microchips in performing long-term reactions caused by their porousness and gas permeability properties in chip body, we developed a two-step postprocessing method, including paraffin filling and parylene coating, to achieve high sealability and stability. We also developed a simple method especially suitable for controlling the vapor diffusion speed of nanoliter-scale droplets by changing the layer thickness of covering oil. With the above methods, 84 tests of nanoliter-scale protein crystallization under vapor diffusion mode were successfully achieved in the 7 × 12 droplet array chip with a protein consumption of 10 nL for each test, which is 20−100 times lower than that in the conventional large-volume screening system. Such a nanoliter-scale vapor diffusion system was applied to two model proteins with commercial precipitants and displayed advantages over that under microbatch mode. It identified more crystallization conditions, especially for the protein samples with lower concentrations. KEYWORDS: 3D-printed chip, surface modification, protein crystallization, high density, droplet array

■

diffusion.20−25 In the microbatch method, a small volume of protein solution is first mixed with a precipitant solution. In the condition that the protein becomes supersaturated, protein starts to gather into nuclei which subsequently grow to form crystals. Microfluidic approaches allow the miniaturization of microbatch experiments into nanoliter volume by implementing them in water-in-oil droplets17,19,26,27 and microfabricated wells.28 The mixing ratio of protein and precipitant can be precisely optimized by changing their relative flow rates in droplet system or the well volumes in chip system. In the FID method, physically separated protein and precipitate solutions are connected with a microcapillary channel to allow them to diffuse into each other driven by concentration gradient. The precipitant concentration in the protein solution gradually increases until nucleation zone is reached and crystallization occurs. Thus, the FID method allows the test of a wide range of precipitant concentrations in a single experiment. Microfluidic chips perfectly fit with FID experiments.10,19 Nanoliter-volume protein and precipitant solutions can be metered using microfabricated chambers with a connecting channel. Con-

INTRODUCTION For decades, tremendous efforts have been made to determine the three-dimensional structures of proteins and their complexes, which are the keys to understanding their functions, and to rationally designing novel drug compounds.1,2 Protein crystallization is conventionally considered challenging and has remained the rate-limiting step to determine protein structures using X-ray crystallography, especially for those biologically important but scarce proteins.3,4 So far, growing well-diffracting crystals largely relies on the trial-and-error method, where over hundreds or even thousands of chemical conditions containing different salts, pH buffers, and precipitants have to be tested.5−7 Moreover, crystallization of a particular protein is rarely predictable. Facing such challenges, the microfluidic technique has been applied to protein crystallization in recent years by performing liquid metering,8,9 sample/reagent mixing,10,11 and reaction in ultrasmall volumes.12,13 It not only allows high throughput screening of thousands crystallization conditions on a single compact microchip with submilligram-scale protein samples but also improves the crystal formation and growth process with precisely controlled chemical and physical environment.14,15 Microfluidic techniques have been developed to implement protein crystallization screening under various modes including microbatch,16−18 free interface diffusion (FID),10,19 and vapor © XXXX American Chemical Society

Received: December 12, 2016 Accepted: March 17, 2017 Published: March 17, 2017 A

DOI: 10.1021/acsami.6b15933 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

testing multiple different chemical conditions. In addition, fundamental researches on the optimization and modification of 3D-printed microdevices are also lacking. Construction of inert and biocompatible surface is vital for performing chemical and biological assay in small volume. For example, Macdonald et al.49 observed that photopolymer used in stereolithography (SLA)-based 3D printing was toxic to human cell line and fish embryos. Here, we describe the implementation of nanoliter-scale vapor diffusion-based protein crystallization screening with a 3D-printed microwell array chip. The 3D-printing technique enables the fabrication of high-density microchambers with high-aspect-ratio protein and reservoir wells in a simple and low-cost way. To achieve long-term incubation of nanoliter crystallization in these chambers, we developed a two-step postprocessing method to construct fully sealed chambers and inert chamber surface. A microfluidic droplet robot, described in our previous work for microbatch-based crystallization,17,50 was employed to assemble nanoliter-volume droplet reactors in a fully automated way. Efficient and reliable vapor diffusion experiment in nanoliter droplet was confirmed by hydrating and dehydrating with lower and higher concentration salt solutions. The dehydration speed was optimized and adjusted to facilitate crystal growth. We applied the established method in the primary screening of crystallization conditions of two model proteins with commercial precipitants with different chemical compositions. More crystallization conditions were identified under vapor diffusion mode compared with those under microbatch mode.

vection-free diffusion was triggered by switching on a microvalve10 or slipping a chip19 to open the channel. The vapor diffusion method has received the widest applications, and it has contributed up to 70% of protein structures deposited in the Protein Data Bank (PDB).23 In the vapor diffusion method, a droplet containing both protein and precipitant and a larger reservoir containing only precipitant are compartmented in a sealed chamber. Initially, the droplet contains comparatively low protein and precipitant concentration. As the droplet and reservoir equilibrate, both the protein and precipitant concentrations in the droplet increase to reach the nucleation zone. Various microfluidic methods were successfully demonstrated for vapor diffusion-based crystallization.11,20,23−25 To achieve vapor diffusion, vapor permeable PDMS layer11,20,24,25 or fluorinated oil21,22 was commonly employed to separate protein and precipitant solutions. The vapor diffusion process could be precisely controlled by dynamically adjusting the salt concentration of precipitant solutions24,25 or the length of oil plug.21 More recently, vapor diffusion was demonstrated on a centrifugal chip,23 where two-level capillary stop valves were designed to meter and isolate protein and precipitant solutions. Despite these advances, microfluidic vapor diffusion systems with the ability of screening large number of crystallization conditions and fully automated operation are still in great demand. Miniaturization of traditional microwell-plate-based vapor diffusion method was rarely reported. One of the main obstacles is the lack of reliable multistep and nanoliter-scale liquid handling method for premixing protein with different precipitants and settling them close to reservoirs. Another obstacle lies in the challenge to fabricate high-aspect-ratio microchips for containing small protein droplets and large reservoirs, whose volume ratio is in the range of 1/50 to 1/200. Some specific microfabrication methods, such as multilayer SU8 lithography29 and laser ablation,30 provide possible solutions for it but suffer from high cost and complicated procedures. Recent years have witnessed tremendous development of the 3D-printing technique and the revolution it brings in the field of industry, art, food, medical device,31 chemical synthesis,32−34 and microfluidics.35−39 3D printing is characterized as an additive manufacturing technique, in which structure is produced by successively printing material layers on a base. Complicated microstructures can be designed in a computer and directly fabricated with a 3D printer in a single step, without long-time and multistep photolithography, molding, and bonding procedures. A range of different chip materials are available to meet different requirements in chemical and biological researches.29,39,40 A more recent study demonstrated micrometer-scale microstructure (4 μm channel) can be produced with a modified printer.41 The comparison and evaluation of different 3D-printing techniques for microfluidic chips were performed, and the cons and pros of each techniques were discussed.38,42 A major advantage of the 3Dprinting technique is the ability to produce complicated multiple-layer structures in single steps, such as valves,43 pumps,44 mixers, and fluidic circuits. The 3D-printing technique has been successfully applied in chemical synthesis,34 fast mixing,45 cell culture,46 biosensing,47 liquid−liquid extraction,48 and so on. In most of the above-mentioned systems, closed microchannels were usually used for liquid manipulations and reactions, which may encounter the challenge in directly accessing the nanoliter reaction products for further analysis as well as the limitation in applications for

■

RESULTS AND DISCUSSION System Design. We aimed to develop a microfluidic system capable of performing vapor diffusion-based protein crystallization in nanoliter scale while maintaining the advantages of conventional systems, including fully automated liquid handling, easy crystal harvesting, and high throughput. We employed our previously developed droplet robot system for liquid handling under the sequential operation droplet array (SODA) mode.17,50 The droplet robot can reliably meter, mix, transfer, and deposit aqueous droplets with volumes from 60 pL to 500 nL. It is fully automated and programmable under the control of a computer. It can directly generate droplet array from samples/reagents stored in a 96- or 384-well plate and achieve seamless use of commercial crystallization kits. More importantly, droplets were covered with a layer of oil and can be accessed with a capillary probe through the oil, which provide the possibility for convenient crystal harvesting from droplets. We have successfully applied the droplet robot system in large-scale protein crystallization screening under microbatch mode as well as in determination of phase diagrams of model proteins.17 However, the extension of the droplet robot to vapor diffusion-based protein crystallization was hindered by the difficulties in chip fabrication, in which microwells with different volumes (ca. 1:100 difference) and depths are required. To address this challenge, the state-of-art 3D-printing technique was used to produce microchip suitable for vapor diffusion-based crystallization. We chose a commercial multiJet printer (ProJet 3510 HDPLUS from the 3D system) because it can fabricate microstructures with both high precision (25−50 μm per 25.4 mm) and high speed (ca. 2 h for a batch of nine chips). A recent study of ProJet 5500X demonstrated its ability to print fine features as small as 250 μm.51 This multiJet printer uses a piezo print head to deposit photocurable resin or support B

DOI: 10.1021/acsami.6b15933 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 1. (a) Image of a prototype 3D-printed chip with an array of 84 chambers. (b) Schematic drawing showing the structure and dimension (mm) of each chamber including a crystallization well and a reservoir well. (c) Schematic 3D drawing of a 3D-printed chip with multiple-layer structure for crystallization incubation. (d) Image of an assembled 3D-printed chip. (e) Procedures of vapor diffusion-based protein crystallization experiment in droplet system. Protein and precipitant solution are sequentially loaded into capillary probe at a volume ratio of 1:1 to form droplet (e1). The droplet is dispensed into the crystallization well with cover oil (e2). Then, precipitant is added in the reservoir well (e3). The chambers in the chip are sealed with a glass plate with PDMS coating and incubated to allow droplet in each chamber equilibrates with precipitant and dehydrates for crystal formation (e4).

through the chip body can be efficiently eliminated. Second, a glass plate coated with a thin layer of PDMS was assembled with the 3D-printed microchip before incubation (Figure 1c,d). The elastomeric PDMS layer coupled with glass substrate could efficiently seal the microchambers on the chip with the aid of the bonding pressure provided by the clips. Using the abovedescribed methods, complete seal of microchambers was confirmed by long-term (over 70 h) monitoring of 20 nL fluorescent droplets without evident change in their volumes (Figure 3). During the droplet generation and incubation, the crystallization well was filled with a layer of mineral oil, which played vital roles in these processes. On the one hand, the cover oil can minimize the evaporation of nanoliter-scale droplets during droplet generation process and before sealing the microchambers. We observed that a 10 nL water droplet could completely evaporate within 1 min in a typical laboratory environment. As a comparison, no evident volume change could be observed over 1 h when the droplet was covered with a layer of oil with a thickness of 0.5 mm. On the other hand, the cover oil can serve as an efficient vapor barrier to reduce the vapor diffusion speed of nanoliter-scale droplets. As nucleation and crystal growth are usually slow, reduction of the droplet dehydration rate can regulate the crystallization process. Previous study also indicated that a slower dehydration rate is more likely to help to produce larger and better crystals.52 Nanoliter-scale droplet has a small volume and a high surface-to-volume ratio, therefore resulting in a high dehydration rate. The equilibration between a 20 nL droplet and a 1 μL salt solution with 2-fold higher concentration in the reservoir through vapor diffusion took approximately 1 h.

materials (casting white wax) layer by layer. Then a UV lamp flashes to solidify the layer of resin and produces a cured plastic chip. It allows the use of translucent material (VisiJet M3 Crystal in this work) for easy optical observation. The precision of the printer was further confirmed by printing several model microstructures with different widths and depths as shown in Figure S1. The SEM result indicates microstructures with wall thickness as small as 200 μm and aspect ratio of 8 can be reliably produced. It should be noted that the SLA-based 3Dprinting method also provides similar high resolution and could be used to produce the microwell array chip. The design of vapor diffusion-based crystallization chip was inspired by conventional sitting drop microwell plate. As shown in Figure 1a, 84 microchambers are integrated in one microchip, with every chamber consisting of a reservoir and a crystallization well. The reservoir well has a size of 1.5 mm (length) × 1.5 mm (width) × 1.5 mm (depth) and a volume of 3.4 μL, which guarantees a volume ratio over 100 against protein droplet (Figure 1b). The crystallization well is designed as a cylinder with the size of 1.0 mm in diameter and 1.0 mm in depth. Compared with standard 96-well plates, the well size in the present 3D-printed chip was significantly decreased by 100 times and the density was significantly increased by ca. 7 times. To achieve successful vapor diffusion crystallization, the complete seal of microchambers during long incubation process is a prerequisite to prevent uncontrollable vapor transportation to external environment. We developed two techniques to solve this problem. First, we developed a two-step postprocessing method for 3D-printed microchip including molten paraffin filling and parylene coating (see the section of Postprocessing of 3D-Printed Chips for details). The vapor transportation C

DOI: 10.1021/acsami.6b15933 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

fluorescent droplet rapidly leaked into the chip body within 30 min, indicating that the chip body is porous. To further prove our hypothesis, scanning electronic microscopy (SEM) was used to investigate the top surface (Figure 2b1) and the cross sections (Figure S2) of the chip. Micropores with sizes from 1 to 10 μm were irregularly spreading on the chip surface (Figure 2b1). The phenomenon of the leakage of fluorescent solution into the chip body implied these micropores were interconnected, which was further verified by the cross-section image of the cleaned chip (Figure S2). Such a phenomenon could be explained by the additive manufacturing format of 3D printing. As most of the 3D-printing systems, the multi-inkjetbased 3D printer produces structures point by point and layer by layer. Although the highest resolution was used in the printing process, tiny gaps still existed between these points and layers. The rapid printing process also left portion of polymer uncured. After support material and uncured photopolymer were removed, these micropores showed up. Because of the pores, the microchip also becomes opaque (Figure 2c1) and would severely interfere with the observation of protein crystal in droplets. In this work, we developed a paraffin filling method to eliminate these micropores by immersing the underside of the chip in molten paraffin and allowing the molten paraffin to permeate into the chip body by capillary force. An SEM image shows that most of the micropores are disappeared except several microgrooves with sizes smaller than 5 μm (Figure 2b2). We also used fluorescence imaging experiments to study the droplet leakage in the modified chip. As shown in Figure 2d, no evident droplet leakage could be observed, indicating that most of the micropores in the chip body are filled with paraffin. In addition, the chip body becomes translucent after paraffin filling, which is important for optical observation and crystal identification (Figure 2c2). To maintain highly stable surface, the 3D-printed chip with filled paraffin was further subjected to surface coating with parylene C. Parylene C is a type of hydrophobic poly(pxylylene) polymers with excellent chemical inertness, low permeability, and biocompatibility. Parylene coating is widely used in microfluidic systems for PCR amplification,53 implanted probe,54 electrospray,55 and chromatography separation.56 In this work, parylene C film on the surface of 3D-printed chips was produced by the chemical vapor deposition (CVD) method.57 The hydrophobic interaction between parylene monomer and Visijet material helps to form stable and durable coating. The main function of parylene coating is to prevent the paraffin in chip body from dissolving in mineral oil used as cover oil in droplet crystallization experiments. We proved this protection effect by depositing 300 nL of mineral oil in a chip well and monitoring it over 2 weeks. During this period, no oil leakage and chip swelling were observed. In addition to mineral oil, the parylene coating also exhibited protection effect against a range of organic solvents such as silicon oil, tetradecane, ethanol, and FC40. The other function of the parylene layer is to form an effective permeation barrier that sealed the crystallization microchambers. As shown in Figure 2b3, the chip surface became smoother without any observable micropores after parylene coating with a layer thickness of 10 μm. With the paraffin filling and parylene coating, the 3D-printed chip can be used for nanoliter-volume vapor diffusion experiments over 6 weeks.

Whereas, the equilibration time was extended to 19 h if the droplet was covered with 150 nL oil (corresponding to an oil layer thickness of 190 μm). In addition, the vapor diffusion speed could be easily adjusted by changing the oil volumes (see the section Effects of Salt Concentration and Cover Oil on Vapor Diffusion for details). Furthermore, the cover oil also helped to release nanoliter-scale droplets from the capillary tip by using the oil/air interface to scrape the droplet from the tip when removing the capillary tip from the oil, during droplet depositing. Compared with droplet depositing methods based on surface contacting,17 this method could be readily carried out without the high standard requirement for the moving precision and planeness of microchips. Postprocessing of 3D-Printed Chips with Paraffin Filling and Parylene Coating. To obtain completely sealed microchambers for long-term incubation, the 3D-printed chip was processed with paraffin filling followed by parylene coating (Figure 2a). Our preliminary experiments of unsuccessful liquid storage indicated that original 3D-printed chips were porous and gas permeable. As shown in Figure 2d1, a 300 nL

Figure 2. (a) Schematic drawing of the two-step postprocessing procedures for 3D-printed chip. The micropores (a1) in the chip body are filled by molten paraffin by capillary force (a2), and then the chip surface was deposited with a layer of parylene C (a3) by chemical vapor deposition. (b) SEM images of the typical top surfaces of a 3Dprinted chip before (b1) and after paraffin filling (b2), and after parylene C deposition (b3). Micropores with sizes from 1 to 10 μm irregularly spread in the chip body can be filled by molten paraffin and the layer of parylene made the surface more smooth. (c) Images of a whole 3D-printed chip before (c1) and after paraffin filling (c2). (d) Fluorescent images showing droplet leakage in a 3D-printed chip without surface modification (d1), with paraffin filling (d2), and with parylene deposition (d3). Droplet volume: 300 nL. D

DOI: 10.1021/acsami.6b15933 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces Effects of Salt Concentration and Cover Oil on Vapor Diffusion. The feasibility of the present system in vapor diffusion experiments was demonstrated by equilibrating an array of 20 nL droplets to NaCl solutions with different concentrations. The NaCl concentration in the droplets was kept at 0.5 M, while the concentrations in reservoir were varied among 0.1, 0.5, 1, and 2.5 M, forming four concentration ratios of 5/1, 1/1, 1/2, and 1/5. Sodium fluorescein with a concentration of 1 mM was added into droplets for aiding droplet visualization in a fluorescent detection system. In a vapor diffusion system of two salt solutions, the vapor pressure of the lower concentration salt solution is higher, resulting in water transport to higher concentration solution through the diffusion of vapor phase. As shown in Figures 3a and 3b, the volume of the droplets increased, unchanged, and decreased with concentration ratios of 5/1, 1/1, and 1/2 (1/5), respectively. The volume variation was rapid in the first 24 h and tended to be slow afterward. The rate of droplet dehydration also increased with the concentration difference between droplet and salt solutions (Figure 3b), which could be attributed to the difference of vapor pressures. In all of these systems, vapor transport phenomenon could be observed over several days to weeks, which is similar to that of the conventional vapor diffusion systems. The droplet dehydration rate could also be controlled by changing the thickness of cover oil. We observed that mineral oil, which is commonly used in microbatch method, could be permeable to water vapor if proper thickness was selected. In addition to paraffin coating, the chip surface was finally fluorinated using Aquapel solution. The fluorinated surface is not wettable with both water (with contact angle of 123°) and mineral oil (with contact angle of 112°). The high contact angle of mineral oil on fluorinated surface could help to maintain uniform thickness of oil layer over droplet among different wells. The effect of cover oil on the rate of droplet dehydration was investigated by varying the oil thickness from 190 to 760 μm in a vapor diffusion system with a concentration ratio of 1:2. As shown in Figure 3c, the oil thickness has a significant effect on the vapor transport speed. In the first 20 h, the droplet volume decreased by 25% with an oil thickness of 190 μm and only 5% with a thickness of 760 μm. In an ideal vapor diffusionbased protein crystallization system, the droplet dehydration should be gentle to obtain slow nuclei formation and moderate crystal growth. Thus, cover oil with a thickness of 380 μm was used in the following crystallization experiments. Nanoliter-Scale Protein Crystallization and Screening with the Vapor Diffusion Method. We first evaluated the performance of the present system in nanoliter-scale vapor diffusion-based protein crystallization by crystallizing a model protein (lysozyme) with different concentrations and comparing the result with that in microbatch method. In both methods, lysozyme solutions with concentrations of 10, 20, 30, and 50 mg/mL in 0.1 M NaAc buffer (pH = 4.6) were used, and precipitant containing 6% (w/v) NaCl in 0.1 M NaAc buffer (pH = 4.6) was used. Crystal incubation was also kept at identical conditions (16 °C) for 7 days and inspected every 2 days. The results are summarized in Table 1. In the microbatch method, lysozyme crystals could be observed only with the highest protein concentration of 50 mg/mL after a 3 day incubation, while in the vapor diffusion method, more concentration conditions from 20 to 50 mg/mL were observed to produce crystals. For lower protein concentrations, it took longer time (5 days) to produce crystals, which could be

Figure 3. Vapor diffusion experiments of droplets in the 3D-printed chip system. CCD images (a) and volume change curves (b) showing the effect of concentration ratios between droplet and precipitant. The NaCl concentration in the droplets was kept at 0.5 M, while the concentrations in reservoir well were varied among 0.1, 0.5, 1, and 2.5 M, forming four concentration ratios of 5/1, 1/1, 1/2, and 1/5. (c) Effect of oil thickness on the dehydration speed of droplet under a concentration ratio of 1/2. Relative volume indicates the volume ratio between remained droplet and original droplet.

explained by the slow droplet dehydration process in the vapor diffusion method. We also observed that the number and the size of crystals under vapor diffusion mode were bigger than those under microbath mode (Figure 4). The above results confirm that the vapor diffusion crystallization performed in a nanoliter-scale droplet system could promote crystal formation with low protein concentrations, which is similar to conventional large-volume systems. E

DOI: 10.1021/acsami.6b15933 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces Table 1. Comparison of Microbatch and Vapor DiffusionBased Protein Crystallization with Different Concentrations of Lysozymea time (day) microbatch

vapor diffusion

a

protein concn (mg/mL)

1

3

5

7

10 20 30 50 10 20 30 50

0/6 0/6 0/6 0/6 0/6 0/6 0/6 0/6

0/6 0/6 0/6 5/6 0/6 1/6 1/6 4/6

0/6 0/6 0/6 5/6 0/6 2/6 2/6 4/6

0/6 0/6 0/6 5/6 0/6 2/6 3/6 4/6

For each condition, six parallel tests were performed. Figure 5. Screening of multiple different crystallization conditions for lysozyme and trypsin. The number of identified conditions for lysozyme (a) and trypsin (b) with vapor diffusion and microbatch methods. Typical crystal images of lysozyme (c) and trypsin (d) in droplets. Droplet volume: 20 nL.

the basis of 3D printing and droplet robot techniques. We demonstrated that the present nanoliter-scale vapor diffusion system not only significantly reduced protein consumption but also allowed crystal formation from more conditions. We also demonstrated vapor diffusion speed can be controlled by adjusting the thickness of covered oil, which could benefit current vapor diffusion-based protein crystallization method. The system is fully automated without tedious manual operations and is programmable under the control of a computer. It can be scaled up to screen over hundreds or even thousands of crystallization conditions by simply increasing the microchamber number and the chip size. In the future, we will further improve the system by using a more transparent photopolymer to increase the optical property of 3D-printed chip for easier crystal observation by reducing the crystallization volumes to several nanoliter even picoliter volumes as well as by improving the droplet handling throughput with multiple capillary probes. In addition, since on-chip X-ray diffraction could enable direct data collection and structural elucidation without tedious crystal transferring and mounting procedures, we will also study and optimize the 3D-printed chip for direct X-ray diffraction of micrometer-size crystals. We also developed a two-step postprocessing method for 3D-printed chips to achieve long-term incubation of nanoliterscale droplets. For the application of microfluidics in chemistry and biology, there are increasing demands to fabricate microchips in a simple, flexible, and affordable way. We believe our method could provide an effective way for performing chip processing to eliminate liquid leakage, to minimize vapor penetration, and to provide inert and biocompatible surfaces, which would significantly broaden the use of 3D-printing techniques in microfluidics. In this work, we achieved the first use of 3D-printing technique for producing microwell array chips. We believe the 3D-printing technique will enhance the development of microwell-array chips in the following three aspects. First, the rapid prototype of the 3D-printing technique will greatly accelerate the design and optimization of microwell array chips to meet the requirements of different applications. Second, the 3D-printed structure can be used as a mold to replicate microwell array chip in a rapid way. Finally, although the

Figure 4. Typical images showing the crystallization results of lysozyme at concentrations of 20, 30, and 50 mg/mL under vapor diffusion and microbatch modes. More concentration conditions are found to be capable of generating crystals with vapor diffusion method, and crystals grown in the vapor diffusion system are bigger than that in microbatch system under the protein concentration of 50 mg/mL. Droplet volume: 20 nL.

We then applied the nanoliter-scale vapor diffusion system in screening of multiple crystallization conditions for two model proteins: lysozyme (30 mg/mL) and trypsin (40 mg/mL). Twenty-eight different precipitants were randomly selected from a commercial screen kit (Table S1), and each precipitant was tested for three times. For each test, 10 nL protein solution and 10 nL precipitant solution were mixed to form one droplet and the droplet was allowed to equilibrate with 1 μL precipitant tests, which is 20−100 times lower than that in the conventional large-volume screening system. Nanoliter-scale microbatch experiments were also performed in comparison with vapor diffusion crystallization. Crystals were confirmed and discriminated from salt crystals using needle crushing or Izit Crystal Dye methods. The results are summarized in Figure 5a, Figure S3, and Table S2. For lysozyme, four conditions were identified to yield crystals using the vapor diffusion system, and only one condition when the microbatch system was used. For trypsin, nine conditions were identified using the vapor diffusion system, with two of them were unique. Using the microbatch system, eight conditions were identified and among which only one was unique. Typical crystal images can be found in Figure 5c,d.

■

CONCLUSIONS In conclusion, we have successfully achieved protein crystallization in the nanoliter range under vapor diffusion mode, on F

DOI: 10.1021/acsami.6b15933 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

The cleaned chip was dried using high pressure nitrogen and then placed into a glass Petri dish containing a thin layer (ca. 2 mm) of molten paraffin at 105 °C for 10 min (Figure 2a). After the molten paraffin was filled into the tiny pores in the chip body by capillary force, the chip was cooled to room temperature to facilitate the solidification of paraffin. A layer of parylene C with a thickness of 10 μm was coated on the surface of the microchip by a local company (Penta Technology, Suzhou, China) (Figure 2a). Finally, the chip was soaked in Aquapel solution for 5 min to fluorinate its surface. Before experiment, the chip was washed by demineralized water and isopropanol. The cover of the chip was fabricated by spinning a layer of PDMS on a glass substrate. The PDMS and its curing agent were mixed completely at a ratio 10:1 (m/m) and degassed in a vacuum desiccator. The mixture was poured on the glass substrate and spin-coated with a spin-coater (KW-4A, Institute of Microelectronics of Chinese Academy of Sciences, Beijing, China) operated at 500 rpm (rpm) for 9 s and then 1000 rpm for 15 s to obtain a PDMS layer thickness of 40 μm. The glass substrate was placed in an oven at 90 °C for 10 h to cure the PDMS. Procedures of Protein Crystallization. Before experiment, precipitant and protein solutions were loaded in a 384-well plate, with a layer of mineral oil covering each well. The multiwell plate and microchip were mounted on the translation stage at predefined positions. Degassed water was filled into the syringe and the capillary as carrier to completely remove air bubble. A plug of mineral oil with a length of 5 mm was aspirated into the tip of the capillary to segment the carrier water and sample/reagent. The use of mineral oil can also significantly reduce the contamination between adjacent samples because the oil forms a protection layer on the inner and outer surface of capillary probe to isolate its interaction with sample solution. The aspirating and dispensing rate of mineral oil and solutions was kept at 5 nL/s using the syringe pump. Mineral oil and solutions were aspirated into capillary probe by suction force of the syringe pump. To minimize the liquid evaporation in the liquid handling process, the robot system was kept in an environment with high humidity of 80%. For vapor diffusion-based crystallization in 3D-printed chips, each crystallization well was preloaded with 300 nL of mineral oil unless stated otherwise. A crystallization droplet containing protein and precipitant was generated and deposited into crystallization well using the droplet assembling technique.17,50 Briefly, the translation stage was moved to allow the capillary probe to sequentially immerse into the wells in the multiwell plate and to aspirate definite volumes of protein and precipitate solutions into the probe tip to form a droplet reactor (Figure 1e). Afterward, the droplet was deposited into crystallization well by moving the stage to make the tip immerse in the mineral oil and push the droplet out of the tip. The droplet settled to the bottom of the well by gravity. The reservoir wells in each crystallization chamber in the chip were filled with 1 μL of precipitant solution corresponding to the composition of the precipitant in the crystallization droplet. Finally, the chip and the cover were assembled and tightened with foldback clips to achieve complete seal of each crystallization chamber (Figure 1d). The chip was incubated at 16 °C and was observed every other day using a stereomicroscope. For microbatch-based crystallization on glass chips, an array of 5 nL droplets containing different precipitants was first generated in the microwells of the chip using the droplet robot. During droplet depositing, the distance between the probe tip and the bottom of the microwell was kept at 20 μm to allow droplets to be reliably released from the tip. Then, 600 nL protein solution was aspirated into the capillary and sequentially injected into each precipitant droplet with a volume of 5 nL. The distance was kept at 30 μm to allow deposited protein droplet to contact and fuse with corresponding precipitant droplet. Every precipitant was repeated for three times as parallel experiment. The chip was sealed with a glass plate and incubated at 16 °C. The droplet array was observed using an inverted microscope (CLIPSE TE2000-S, Nikon, Tokyo, Japan) under polarized light or differential interference contrast mode.

throughput of 3D printing for array chips is relatively low in the present stage, compared with that of injection molding-based manufacture technique, the recent studies demonstrated its throughput can be significantly increased with the advance of new materials and fabrication methods. A very recent progress in 3D printing shows that its manufacture speed can be significantly increased by 100 times.58 With the rapid advance of 3D-printing techniques, it will become a universal production method for microfluidic chips.

■

EXPERIMENTAL SECTION

Materials. All solvents and chemicals were used as received unless stated otherwise, and demineralized water prepared with a water purification system (Milli-Q, Millipore, Bedford, MA) was used throughout. Mineral oil (M5904, with viscosity of 14.2−17.2 cSt at 40 °C), octadecyltrichlorosilane (OTCS), isopropanol, benzamidine hydrochloride, HEPES, sodium azide, lysozyme, and trypsin were purchased from Sigma-Aldrich (St. Louis, MO). Poly(dimethylsiloxane) (PDMS) and its curing agent (Sylgard184) were purchased from Dow Corning (Midland, MI). Solid paraffin, HF, NH4F, and HNO3 were products of Sinopharm Chemical (Beijing, China). Aquapel (Pittsburgh, PA) was used as a surface fluorination treating agent. Glass plates with chromium and photoresist coating was from Shaoguang Microelectronics (Changsha, China). Protein crystallization kit (Index HT) and the Izit Crystal Dye are products of Hampton Research Co. (Aliso Viejo, CA). Protein solutions of lysozyme (30 mg/mL in 0.1 M NaAc buffer, pH 4.6) and trypsin (40 mg/mL in 20 mM HEPES buffer containing 10 mM CaCl2 and 10 mg/mL benzamidine hydrochloride, pH 7.0) were freshly prepared before experiments. Instruments and Apparatus. Nanoliter-scale liquid handling was performed using a droplet robot system described in our previous work.17,50 Briefly, the robot consists of four parts: (1) a high precision syringe pump (PHD2000, Harvard Apparatus, Holliston, MA) for liquid delivering and metering; (2) a 10 μL syringe (1700 series, Hamilton, Reno, NV) with a capillary (150 μm i.d., 250 μm o.d., Reafine Chromatography, Handan, China) served as a probe for nanoliter liquid aspirating and droplet dispensing; (3) a 3D-printed droplet array chip and a 96/384-well microplate as sample/reagent container; (4) an x-y-z translation stage (PSA series, Zolix, Beijing, China) for motion control of the chip and multiwell plate. The syringe pump and translation stage were automatically controlled with a home-written software (Labview 8.0, National Instruments, Austin, TX). The capillary was tapered with a micropipet puller (P2000/F, Sutter instrument, Novato, CA) to obtain a tip size of 30 μm i.d. and 50 μm o.d. Its outer and inner surfaces were made to be hydrophobic with octadecyltrichlorosilane using procedures described previously.17 A scanning electron microscope (SEM, S-3700N, Hitachi, Tokyo, Japan) was used for studying the surface of 3D-printed chips. A stereomicroscope (SMZ1500, Nikon, Tokyo, Japan) with a CCD camera (DS-Vi1, Nikon, Tokyo, Japan) was used for droplet observation and crystal identification. Fabrication and Postprocessing of Protein Crystallization Microchip. The structure of a prototype 3D-printed chip for protein crystallization is shown in Figure 1a−d. The microchip contains a 2dimensional array of 7 × 12 microchambers, each of which is composed of a crystallization well and a reservoir well. The center-tocenter distances between adjacent microchambers are 4.5 mm in the x direction and 2.5 mm in the y direction. It was designed using Solidworks 2010 (SolidWorks Co., Waltham, MA) and then directly produced with a multiJet printing-based 3D printer (Projet 3510 HDPLUS, 3D Systems Co., Rock Hill, SC), which was operated at a 16 μm printing resolution mode. A translucent photopolymer (VisiJet M3 Crystal) was selected as the chip material. The support material and uncured polymer in the chip body were removed by immersing and sonicating it in cleaning agents. Compared with the standard cleaning procedures,38 we observed longer cleaning time (mineral oil for 60 min, detergent solution for 30 min, and pure water for 10 min) was required until no turbid substances were observed in the agents. G

DOI: 10.1021/acsami.6b15933 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

■

(14) Li, L.; Ismagilov, R. F. Protein Crystallization using Microfluidic Technologies Based on Valves, Droplets, and SlipChip. Annu. Rev. Biophys. 2010, 39, 139−158. (15) Puigmartí-Luis, J. Microfluidic Platforms: a Mainstream Technology for the Preparation of Crystals. Chem. Soc. Rev. 2014, 43, 2253−2271. (16) Li, L.; Du, W.; Ismagilov, R. User-Loaded SlipChip for Equipment-Free Multiplexed Nanoliter-Scale Experiments. J. Am. Chem. Soc. 2010, 132, 106−111. (17) Zhu, Y.; Zhu, L.-N.; Guo, R.; Cui, H.-J.; Ye, S.; Fang, Q. Nanoliter-Scale Protein Crystallization and Screening with a Microfluidic Droplet Robot. Sci. Rep. 2014, 4, 5046. (18) Zheng, B.; Roach, L. S.; Ismagilov, R. F. Screening of Protein Crystallization Conditions on a Microfluidic Chip Using NanoliterSize Droplets. J. Am. Chem. Soc. 2003, 125, 11170−11171. (19) Li, L.; Du, W.; Ismagilov, R. F. Multiparameter Screening on SlipChip used for Nanoliter Protein Crystallization Combining Free Interface Diffusion and Microbatch Methods. J. Am. Chem. Soc. 2010, 132, 112−119. (20) Lau, B. T.; Baitz, C. A.; Dong, X. P.; Hansen, C. L. A Complete Microfluidic Screening Platform for Rational Protein Crystallization. J. Am. Chem. Soc. 2007, 129, 454−455. (21) Zheng, B.; Tice, J. D.; Roach, L. S.; Ismagilov, R. F. A DropletBased, Composite PDMS/glass Capillary Microfluidic System for Evaluating Protein Crystallization Conditions by Microbatch and Vapor-Diffusion Methods with On-Chip X-Ray Diffraction. Angew. Chem., Int. Ed. 2004, 43, 2508−2511. (22) Zheng, B.; Gerdts, C. J.; Ismagilov, R. F. Using Nanoliter Plugs in Microfluidics to Facilitate and Understand Protein Crystallization. Curr. Opin. Struct. Biol. 2005, 15, 548−555. (23) Wang, L.; Sun, K.; Hu, X.; Li, G.; Jin, Q.; Zhao, J. A Centrifugal Microfluidic Device for Screening Protein Crystallization Conditions by Vapor Diffusion. Sens. Actuators, B 2015, 219, 105−111. (24) Zheng, B.; Ismagilov, R. F. A Microfluidic Approach for Screening Submicroliter Volumes against Multiple Reagents by Using Preformed Arrays of Nanoliter Plugs in a Three-Phase Liquid/Liquid/ Gas Flow. Angew. Chem., Int. Ed. 2005, 44, 2520−2523. (25) Zheng, B.; Tice, J. D.; Ismagilov, R. F. Formation of Arrayed Droplets by Soft Lithography and Two-Phase Fluid Flow, and Application in Protein Crystallization. Adv. Mater. 2004, 16, 1365− 1368. (26) Zhou, X.; Lau, L.; Lam, W. W. L.; Au, S. W. N.; Zheng, B. Nanoliter Dispensing Method by Degassed Poly (dimethylsiloxane) Microchannels and Its Application in Protein Crystallization. Anal. Chem. 2007, 79, 4924−4930. (27) Shim, J.-U.; Cristobal, G.; Link, D. R.; Thorsen, T.; Jia, Y.; Piattelli, K.; Fraden, S. Control and Measurement of the Phase Behavior of Aqueous Solutions using Microfluidics. J. Am. Chem. Soc. 2007, 129, 8825−8835. (28) Shim, J.-U.; Cristobal, G.; Link, D. R.; Thorsen, T.; Fraden, S. Using Microfluidics to Decouple Nucleation and Growth of Protein Crystals. Cryst. Growth Des. 2007, 7, 2192−2194. (29) del Campo, A.; Greiner, C. SU-8: a Photoresist for High-AspectRatio and 3D Submicron Lithography. J. Micromech. Microeng. 2007, 17, R81−R95. (30) Selimović, Š.; Piraino, F.; Bae, H.; Rasponi, M.; Redaelli, A.; Khademhosseini, A. Microfabricated Polyester Conical Microwells for Cell Culture Applications. Lab Chip 2011, 11, 2325−2332. (31) Ursan, I. D.; Chiu, L.; Pierce, A. Three-Dimensional Drug Printing: a Structured Review. J. Am. Pharm. Assoc. 2013, 53, 136−144. (32) Gross, B. C.; Erkal, J. L.; Lockwood, S. Y.; Chen, C.; Spence, D. M. Evaluation of 3D Printing and Its Potential Impact on Biotechnology and the Chemical Sciences. Anal. Chem. 2014, 86, 3240−3253. (33) Symes, M. D.; Kitson, P. J.; Yan, J.; Richmond, C. J.; Cooper, G. J.; Bowman, R. W.; Vilbrandt, T.; Cronin, L. Integrated 3D-Printed Reactionware for Chemical Synthesis and Analysis. Nat. Chem. 2012, 4, 349−354.

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.6b15933. SEM images of microgrooves produced by the multijet printer with different widths and depths and the cross section of a 3D-printed chip’s body; tables and in-droplet crystal images showing the screening results of two model proteins with a commercial kit (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (Y.Z.). *E-mail: [email protected] (S.Y.). *E-mail: [email protected] (Q.F.). ORCID

Qun Fang: 0000-0002-6250-252X Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Financial support from the Natural Science Foundation of China (Grants 21435004, 21227007, and 21475117), the Ministry of Science and Technology (Grant 2014CB910300), and the Natural Science Foundation of Zhejiang Province (Grant LY14B050001) is gratefully acknowledged.

■

REFERENCES

(1) McPherson, A.; Gavira, J. A. Introduction to Protein Crystallization. Acta Crystallogr., Sect. F: Struct. Biol. Commun. 2014, 70, 2−20. (2) Manjasetty, B. A.; Turnbull, A. P.; Panjikar, S.; Bussow, K.; Chance, M. R. Automated Technologies and Novel Techniques to Accelerate Protein Crystallography for Structural Genomics. Proteomics 2008, 8, 612−625. (3) Chayen, N. E.; Saridakis, E. Protein Crystallization: from Purified Protein to Diffraction-Quality Crystal. Nat. Methods 2008, 5, 147−153. (4) Lizak, C.; Gerber, S.; Numao, S.; Aebi, M.; Locher, K. P. X-Ray Structure of a Bacterial Oligosaccharyltransferase. Nature 2011, 474, 350−355. (5) Benvenuti, M.; Mangani, S. Crystallization of Soluble Proteins in Vapor Diffusion for X-Ray crystallography. Nat. Protoc. 2007, 2, 1633− 1651. (6) Giege, R. A Historical Perspective on Protein Crystallization from 1840 to the Present Day. FEBS J. 2013, 280, 6456−6497. (7) Berman, H. M.; Bhat, T. N.; Bourne, P. E.; Feng, Z.; Gilliland, G.; Weissig, H.; Westbrook, J. The Protein Data Bank and the Challenge of Structural Genomics. Nat. Struct. Biol. 2000, 7, 957−959. (8) Sauter, C.; Dhouib, K.; Lorber, B. From Macrofluidics to Microfluidics for the Crystallization of Biological Macromolecules. Cryst. Growth Des. 2007, 7, 2247−2250. (9) Song, H.; Chen, D. L.; Ismagilov, R. F. Reactions in Droplets in Microfluidic Channels. Angew. Chem., Int. Ed. 2006, 45, 7336−7356. (10) Hansen, C. L.; Skordalakes, E.; Berger, J. M.; Quake, S. R. A Robust and Scalable Microfluidic Metering Method that Allows Protein Crystal Growth by Free Interface Diffusion. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 16531−16536. (11) Hansen, C. L.; Classen, S.; Berger, J. M.; Quake, S. R. A Microfluidic Device for Kinetic Optimization of Protein Crystallization and in situ Structure Determination. J. Am. Chem. Soc. 2006, 128, 3142−3143. (12) Hansen, C.; Quake, S. R. Microfluidics in Structural Biology: Smaller, Faster... Better. Curr. Opin. Struct. Biol. 2003, 13, 538−544. (13) Leng, J.; Salmon, J. B. Microfluidic Crystallization. Lab Chip 2009, 9, 24−34. H

DOI: 10.1021/acsami.6b15933 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces (34) Kitson, P. J.; Symes, M. D.; Dragone, V.; Cronin, L. Combining 3D Printing and Liquid Handling to Produce User-Friendly Reactionware for Chemical Synthesis and Purification. Chem. Sci. 2013, 4, 3099−3103. (35) Ho, C. M. B.; Ng, S. H.; Li, K. H. H.; Yoon, Y.-J. 3D Printed Microfluidics for Biological Applications. Lab Chip 2015, 15, 3627− 3637. (36) Erkal, J. L.; Selimovic, A.; Gross, B. C.; Lockwood, S. Y.; Walton, E. L.; McNamara, S.; Martin, R. S.; Spence, D. M. 3D Printed Microfluidic Devices with Integrated Versatile and Reusable Electrodes. Lab Chip 2014, 14, 2023−2032. (37) Kitson, P. J.; Rosnes, M. H.; Sans, V.; Dragone, V.; Cronin, L. Configurable 3D-Printed Millifluidic and Microfluidic ‘Lab on a Chip’ Reactionware Devices. Lab Chip 2012, 12, 3267−3271. (38) Lee, J. M.; Zhang, M.; Yeong, W. Y. Characterization and Evaluation of 3D Printed Microfluidic Chip for Cell Processing. Microfluid. Nanofluid. 2016, 20, 1−15. (39) Amin, R.; Knowlton, S.; Hart, A.; Yenilmez, B.; Ghaderinezhad, F.; Katebifar, S.; Messina, M.; Khademhosseini, A.; Tasoglu, S. 3DPrinted Microfluidic Devices. Biofabrication 2016, 8, 022001. (40) Khoo, Z. X.; Teoh, J. E. M.; Liu, Y.; Chua, C. K.; Yang, S.; An, J.; Leong, K. F.; Yeong, W. Y. 3D Printing of Smart Materials: A Review on Recent Progresses in 4D Printing. Virtual Phys. Prototyping 2015, 10, 103−122. (41) Spivey, E. C.; Xhemalce, B.; Shear, J. B.; Finkelstein, I. J. 3DPrinted Microfluidic Microdissector for High-Throughput Studies of Cellular Aging. Anal. Chem. 2014, 86, 7406−7412. (42) Chen, C.; Mehl, B. T.; Munshi, A. S.; Townsend, A. D.; Spence, D. M.; Martin, R. S. 3D-printed Microfluidic Devices: Fabrication, Advantages and LimitationsA Mini Review. Anal. Methods 2016, 8, 6005−6012. (43) Gong, H.; Woolley, A. T.; Nordin, G. P. High Density 3D Printed Microfluidic Valves, Pumps, and Multiplexers. Lab Chip 2016, 16, 2450−2458. (44) Au, A. K.; Bhattacharjee, N.; Horowitz, L. F.; Chang, T. C.; Folch, A. 3D-Printed Microfluidic Automation. Lab Chip 2015, 15, 1934−1941. (45) Bonyár, A.; Sántha, H.; Ring, B.; Varga, M.; Kovács, J. G.; Harsányi, G. 3D Rapid Prototyping Technology (RPT) as a Powerful Tool in Microfluidic Development. Procedia Eng. 2010, 5, 291−294. (46) Bertassoni, L. E.; Cecconi, M.; Manoharan, V.; Nikkhah, M.; Hjortnaes, J.; Cristino, A. L.; Barabaschi, G.; Demarchi, D.; Dokmeci, M. R.; Yang, Y. Hydrogel Bioprinted Microchannel Networks for Vascularization of Tissue Engineering Constructs. Lab Chip 2014, 14, 2202−2211. (47) Krejcova, L.; Nejdl, L.; Rodrigo, M. A. M.; Zurek, M.; Matousek, M.; Hynek, D.; Zitka, O.; Kopel, P.; Adam, V.; Kizek, R. 3D Printed Chip for Electrochemical Detection of Influenza Virus Labeled with CdS Quantum Dots. Biosens. Bioelectron. 2014, 54, 421−427. (48) Hsieh, K.-T.; Liu, P.-H.; Urban, P. L. Automated On-Line Liquid-Liquid Extraction System for Temporal Mass Spectrometric Analysis of Dynamic Samples. Anal. Chim. Acta 2015, 894, 35−43. (49) Macdonald, N. P.; Zhu, F.; Hall, C.; Reboud, J.; Crosier, P.; Patton, E.; Wlodkowic, D.; Cooper, J. Assessment of Bbiocompatibility of 3D Printed Photopolymers Using Zebrafish Rmbryo Toxicity Assays. Lab Chip 2016, 16, 291−297. (50) Zhu, Y.; Zhang, Y.-X.; Cai, L.-F.; Fang, Q. Sequential Operation Droplet Array: an Automated Microfluidic Platform for Picoliter-Scale Liquid Handling, Analysis, and Screening. Anal. Chem. 2013, 85, 6723−6731. (51) Yang, H.; Lim, J. C.; Liu, Y.; Qi, X.; Yap, Y. L.; Dikshit, V.; Yeong, W. Y.; Wei, J. Performance Evaluation of ProJet Multi-material Jetting 3D Printer. Virtual Phys. Prototyp. 2017, 12, 95−103. (52) Li, G.; Xiang, Y.; Zhang, Y.; Wang, D.-C. A Simple and Efficient Innovation of the Vapor-Diffusion Method for Controlling Nucleation and Growth of Large Protein Crystals. J. Appl. Crystallogr. 2001, 34, 388−391.

(53) Heyries, K. A.; Tropini, C.; VanInsberghe, M.; Doolin, C.; Petriv, I.; Singhal, A.; Leung, K.; Hughesman, C. B.; Hansen, C. L. Megapixel Digital PCR. Nat. Methods 2011, 8, 649−651. (54) Takeuchi, S.; Ziegler, D.; Yoshida, Y.; Mabuchi, K.; Suzuki, T. Parylene Flexible Neural Probes Integrated with Microfluidic Channels. Lab Chip 2005, 5, 519−523. (55) Kameoka, J.; Orth, R.; Ilic, B.; Czaplewski, D.; Wachs, T.; Craighead, H. An Electrospray Ionization Source for Integration with Microfluidics. Anal. Chem. 2002, 74, 5897−5901. (56) He, Q.; Pang, C.; Tai, Y.-C.; Lee, T. D. In Ion Liquid Chromatography on-a-Chip with Beads-Packed Parylene Column, 17th IEEE International Conference on Micro Electro Mechanical Systems (MEMS 2004), IEEE: Maastricht, The Netherlands, 2004; pp 212− 215. (57) Fortin, J. B.; Lu, T. M. A Model for the Chemical Vapor Deposition of Poly (para-xylylene) (Parylene) Thin Films. Chem. Mater. 2002, 14, 1945−1949. (58) Tumbleston, J. R.; Shirvanyants, D.; Ermoshkin, N.; Janusziewicz, R.; Johnson, A. R.; Kelly, D.; Chen, K.; Pinschmidt, R.; Rolland, J. P.; Ermoshkin, A.; Samulski, E. T.; DeSimone, J. M. Continuous Liquid Interface Production of 3D Objects. Science 2015, 347, 1349−1352.

I

DOI: 10.1021/acsami.6b15933 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX