Conformable self-assembling amyloid protein coatings with genetically programmable functionality

See allHide authors and affiliations

Science Advances  20 May 2020:
Vol. 6, no. 21, eaba1425
DOI: 10.1126/sciadv.aba1425


Functional coating materials have found broad technological applications in diverse fields. Despite recent advances, few coating materials simultaneously achieve robustness and substrate independence while still retaining the capacity for genetically encodable functionalities. Here, we report Escherichia coli biofilm-inspired protein nanofiber coatings that simultaneously exhibit substrate independence, resistance to organic solvents, and programmable functionalities. The intrinsic surface adherence of CsgA amyloid proteins, along with a benign solution-based fabrication approach, facilitates forming nanofiber coatings on virtually any surface with varied compositions, sizes, shapes, and structures. In addition, the typical amyloid structures endow the nanofiber coatings with outstanding robustness. On the basis of their genetically engineerable functionality, our nanofiber coatings can also seamlessly participate in functionalization processes, including gold enhancement, diverse protein conjugations, and DNA binding, thus enabling a variety of proof-of-concept applications, including electronic devices, enzyme immobilization, and microfluidic bacterial sensors. We envision that our coatings can drive advances in electronics, biocatalysis, particle engineering, and biomedicine.


Surface modification of materials is an essential aspect of engineering and technology fields including electronics, biomedicine, catalysis, textiles, and industrial equipment (16). The application of diverse coatings is one of the major methods through which either surface properties of a substrate are changed or completely new properties to a finished product are imparted. Some advanced coating materials that have recently been developed include polyelectrolytes, proteins, polydopamine, and polyphenols (2, 714); however, certain limitations have prevented the widespread adoption and practical use of these materials. For example, although polydopamine and polyphenol coatings are substrate independent, both coating types are unstable in certain application environments: Polydopamine coatings suffer from easy detachment in polar solvents, whereas polyphenol coatings exhibit pH-dependent disassembly (7, 15).

Protein-based coating materials (e.g., bovine serum albumin, hydrophobins, and mussel foot proteins) have attracted considerable attention because of their outstanding biocompatibility, biodegradability, and environmental friendliness (11, 12, 16, 17). Amyloid proteins are particularly appealing as a potential source of bioinspired coatings, as their characteristic β-sheet structures exhibit high tolerance toward high temperature, organic solvents, and harsh pH conditions (18, 19). Recent work demonstrated that phase transition lysozyme (PTL), an amyloid protein coating material, could coat the surface of virtually any substrates and have outstanding robustness; however, it is notable that the applications reported for PTL to date have mainly exploited its intrinsic chemical properties (i.e., the aforementioned β-sheet structures) rather than its potentially genetically engineerable functionalities (10, 20, 21).

In nature, bacteria use biofilms to robustly coat an enormous number of surfaces, and these coatings promote cellular survival in harsh environments (22, 23). Fundamental studies have revealed that biofilms produced by Escherichia coli contain amyloid nanofibers, which are self-assembled by secreted monomers of the CsgA protein (the major protein component within the biofilms); these nanofibers provide mechanical strength and structural integrity to biofilms (Fig. 1A) (2426). In addition, a molecular dynamics study recently suggested that CsgA, owing to its unique protein sequence and structural features, should strongly adhere to both polar and nonpolar surfaces (27). For practical applications, multiple studies have shown that genetically engineered CsgA fusion proteins can be used as underwater adhesives, nanoparticle (NP) assembly scaffolds, patternable materials, biomimetic mineralization, and medical hydrogels (2832). In light of their intrinsic adherence toward diverse substrates as well as the fact that a variety of functional peptides and protein domains could be rationally inserted in the CsgA protein through a modular genetic strategy without disrupting their self-assembly into β-sheet structures, we rationalized that engineered CsgA fusion proteins could be used as a coating platform to endow materials with diverse functionalities. Conceivably, such genetically engineered CsgA-based coatings would likely achieve precise performance for myriad applications, likely far surpassing the scope of existing protein coating materials such as PTL and bovine serum albumin. However, exploiting the genetically programmable functionality of CsgA amyloid proteins as a coating material have not been widely explored.

Fig. 1 E. coli biofilm-inspired protein nanofiber coatings and corresponding proof-of-concept applications including electronic devices, enzyme immobilization, and microfluidic sensor.

(A) Illustration of natural E. coli biofilms, in which self-assembled CsgA nanofibers constitute the major protein component. (B) Modular genetic design of genetically engineered CsgA proteins enabled by rationally fusing desired fusion domains at the C terminus of CsgA. (C) Illustrations of producing diverse protein coatings via a solution-based fabrication approach for various applications based on genetically engineered functionalities such as electronic devices, enzyme immobilization, and microfluidic sensor (from top to bottom).

Here, we report a proteinaceous coating material platform based on genetically programmable CsgA fusion amyloid nanofibers. We successfully used a simple, aqueous solution–based fabrication method based on the amyloid protein self-assembly to generate thin-film materials that can conformably coat substrates with highly diverse compositions (e.g., polymeric, metal oxide, inorganic, and metal) and varied shapes (flat, round, pyramid, the interior of a microfluidic device, and even irregular or asymmetric structures). We demonstrate that these coating materials can be further decorated with various molecules and nano-objects such as fluorescent proteins, enzymes, DNA probes, and NPs. The robust coating materials maintained their integrity and functionality, even after exposure to various common organic solvents such as acetone and hexane or after high-temperature challenge. Last, we exploited the process simplicity, flexibility, and functional customization of our coating materials in proof-of-concept demonstrations for electronic devices including a touch switch and a pressure sensor, immobilized multienzyme systems for bioconversion production applications, as well as a hybrid amyloid/DNAzyme microfluidic sensor (Fig. 1, B and C). We anticipate that our genetically engineered CsgA coating materials, which are substrate independent, ultrastable, and afforded precisely with tailor-made and tunable functionality, will find broad application in electronics, biocatalysis, particle engineering, and biomedicine.


Functional characterization, environmental tolerance, and substrate universality of CsgAHis-tag protein coatings

Leveraging a modular genetic design, we constructed four genetically engineered CsgA variants: CsgAHis-tag, CsgASpyTag, CsgASnoopTag, and CsgADNA binding domain (DBD) (Fig. 1B). We expressed our engineered CsgA proteins as inclusion bodies using E. coli BL21(DE3) as a host and purified the proteins following a typical guanidine denaturation protocol for amyloid proteins (28, 30); this approach markedly reduced batch-to-batch variation and impurities. To produce coating materials, we dissolved the purified proteins in an aqueous solution and directly immersed diverse substrates into this protein solution overnight. We first conducted detailed characterization to confirm the coating-forming ability of the CsgA fusion proteins. We chose plates made of unmodified poly(tetrafluoroethylene) (PTFE)—a classical adhesion-resistant material—as the test substrate. After immersion of the substrate in fresh-made CsgAHis-tag monomer (His-tag fused at the C terminal of the CsgA protein) solution overnight, water contact angle tests showed that the contact angle of CsgAHis-tag nanofiber–coated PTFE was 72.7 ± 2.7°, whereas that of bare PTFE was 110.2 ± 3.2° (Fig. 2A). To test the coating effect, we first incubated the bare and coated PTFE in the presence of solution containing nickel–nitrilotriacetic acid (Ni-NTA)–decorated red-emitting quantum dots (QDs) (allowing thorough interactions between Ni-NTA–decorated QDs and CsgAHis-tag nanofibers) and subjected them to copious amount of water to remove nonspecific binding (33).

Fig. 2 Characterization, environmental tolerance, and substrate universality of CsgAHis-tag protein coatings.

(A) Top: Digital images and water contact angles (inset) of bare and CsgAHis-tag–coated PTFE; bottom: digital images of bare and coated PTFE substrates after incubation with QD solution and illumination under UV light. Photo credit: Yingfeng Li, ShanghaiTech University. (B) AFM height image of CsgAHis-tag–coated PTFE. (C) XPS spectra of bare and CsgAHis-tag–coated PTFE, CPS representing counts per second. (D) Schematic showing stability tests consisting of a water contact angle test and a QD binding test. (E) Water contact angle comparison of CsgAHis-tag coatings on PTFE substrates after organic solvent exposure. (F) Digital image of challenged CsgAHis-tag–coated PTFE substrates after incubation with QD solution and illumination under UV light. Photo credit: Yingfeng Li, ShanghaiTech University. (G) Water contact angles of bare and CsgAHis-tag–coated diverse polymer substrates. (H) Water contact angles of bare and CsgAHis-tag–coated various inorganic substrates.

The coated sample displayed bright and uniform red fluorescence under ultraviolet (UV) illumination, whereas the bare PTFE sample showed almost no fluorescence (Fig. 2A). This vast difference in fluorescence intensity was also verified quantitatively through photoluminescence spectroscopy (fig. S1A). Moreover, as revealed by atomic force microscopy (AFM) imaging, CsgAHis-tag nanofiber coatings were formed on the PTFE substrate (Fig. 2B and fig. S1B). X-ray photoelectron spectroscopy (XPS) was also performed to further analyze the surface composition after nanofiber coating, revealing newly appeared N 1s and O 1s peaks at ∼399 and 531 eV, respectively, thereby confirming the coating of CsgAHis-tag proteins on the PTFE substrate (Fig. 2C). Collectively, these results validate the nanofiber coating–forming ability of the genetically engineered CsgA proteins.

To demonstrate the stability of CsgAHis-tag nanofiber coatings in organic solvents, we conducted two kinds of tests: contact angle and QD binding (Fig. 2D). We first measured the contact angles of coated PTFE substrates before and after contact with common organic solvents including hexane, acetone, and dimethyl sulfoxide (DMSO). After immersion in these solvents for 24 hours, the contact angles of the substrates underwent almost no changes, indicating that our coatings had outstanding chemical endurance in these harsh solvents (Fig. 2E). Furthermore, digital images showed that CsgAHis-tag–coated PTFE substrates anchored with Ni-NTA QDs still displayed red fluorescence after contact with the aforementioned common organic solvents, again highlighting the organic solvent tolerance of our nanofiber coatings (Fig. 2F). The CsgAHis-tag proteins also have outstanding environmental tolerance even after long-term exposure to both acidic and basic aqueous solutions as described in a previous study (30).

We next assessed the thermal stability of CsgAHis-tag nanofiber coatings. To such ends, we first used NanoDSF (differential scanning fluorimetry) to determine melting temperatures of proteins using their intrinsic fluorescence change during a programmed temperature gradient increase (34). The fluorescence intensity change of a protein sample is directly correlated to the structural change (e.g., unfolding) of the protein over the heating process. Briefly, our NanoDSF analysis of CsgAHis-tag nanofibers and control bovine serum albumin proteins in solution revealed that whereas the serum albumin proteins began to unfold at ~65°C, the CsgAHis-tag nanofibers had impressive thermal stability, as indicated by the steady fluorescence intensity even at 95°C (fig. S2A). Moreover, the attenuated total reflection–Fourier transform infrared (ATR-FTIR) spectrum of the challenged CsgAHis-tag nanofiber sample showed that the typical β-sheet structures (absorption peak at ~1625 cm−1) were still retained in the nanofiber structures after heating in a 90°C oven for 24 hours (fig. S2B). In addition, water contact angle analysis and QD binding test indicated that CsgAHis-tag nanofibers were still completely coated over on the PTFE substrates even after challenge at 90°C for 24 hours (fig. S2C). These data thus reveal that our CsgAHis-tag protein coatings have outstanding thermal stability.

Biodegradability under appropriate protease conditions is considered as one of the attractive material attributes for protein-based coatings (17). To assess whether our CsgAHis-tag protein coatings have such on-demand biodegradability, we chose two enzymes, trypsin from bovine pancreas and fungal protease from Aspergillus oryzae (protease AO), in our studies. Thioflavin T (ThT; an amyloid specific dye) assay was used to monitor the digestion process of CsgAHis-tag nanofibers. As illustrated in fig. S2 (D and E), the decreasing fluorescence intensities indicate the gradual disappearance of the β-sheet structures over time, suggesting the structural instability of CsgAHis-tag nanofibers under trypsin or protease AO digestion conditions. We next challenged the stability of CsgAHis-tag nanofiber coatings by incubating the CsgAHis-tag nanofiber–coated PTFE plate in the two enzyme solutions (trypsin, 2.5 mg/ml; fungal protease, ≥55 U/g) for 24 hours and assessed the morphological and physicochemical properties with scanning electron microscopy (SEM) and water contact angle analysis, respectively. SEM images showed that very little amount of nanofibers was found on the substrate surface and water contact angle analysis revealed that the enzyme-treated substrates restored their hydrophobicity after nanofiber coating digestions (fig. S2, F to H). These data convincingly demonstrate that our CsgAHis-tag nanofiber coatings can be degraded in the presence of proteases. Collectively, our coating materials have strong environmental robustness while retaining their on-demand biodegradability, and thus can broaden the application scope of existing protein-based coating materials.

To establish that our CsgAHis-tag nanofiber coatings can be applied to other substrates, we coated several typical material substrates, including common organic polymers [polydimethylsiloxane (PDMS), polypropylene (PP), polystyrene (PS), and polyethylene terephthalate (PET)] and inorganics [indium tin oxide (ITO), Si, Au, stainless steel 304, fluorine-doped tin oxide (FTO), and glass]. Our results from water contact angle analysis revealed that CsgAHis-tag nanofibers were successfully coated on each of these substrates (Fig. 2, G and H). These applications convincingly demonstrate the substrate-independent nature of the genetically engineered CsgA protein coatings.

The apparently very broad substrate scope for our coatings raises interesting questions about the molecular interactions that occur between nanofibers and substrates. Previous molecular simulation research has demonstrated that the unique structural features as well as its unique amino acid sequence and diversity of the CsgA protein enable its strong adhesion capacity for both polar and nonpolar substrates (27). Therefore, on the basis of the above contact angle test results, we speculated that the hydrophobic residues within the CsgA protein such as alanine, proline, and valine could provide adhesion to hydrophobic surfaces such as PTFE and PDMS through hydrophobic interactions; that aromatic amino acids such as tyrosine, phenylalanine, and histidine may contribute to adhesion to PS and PET surfaces through π-π stacking interactions; and that charged and polar amino acids such as arginine, lysine, and glutamine could form strong interactions with oxides through electrostatic interactions (35).

Having illustrated the coating formation capacity as well as their basic physicochemical properties of genetically engineered CsgA coatings, we next focused on establishing proof of concept for multiple programmable functions for the CsgA fusion protein coatings.

CsgAHis-tag–Au conductive coatings and electronic devices

Flexible and wearable electronics play critical roles in our daily lives, and the introduction of metal NP–based conductive coatings within such devices is definitely a key step (36, 37). Existing conventional top-down approaches to obtain metal NP coatings often require high temperature and sometimes suffer from low interfacial adhesion (36, 37). Gold enhancement is a promising bottom-up process for fabricating Au-based conductive coatings (38, 39). However, this process preliminarily requires the ability to anchor Au NPs to the targeted substrates (38, 39). Such NPs can then be used to heterogeneously catalyze further Au deposition and form NP-structured coatings in an aqueous AuCl4 and hydroxylamine solution. In the previous section, we confirmed that CsgAHis-tag coatings could anchor Ni-NTA–capped QDs on substrates. Transmission electron microscopy (TEM) images confirmed that CsgAHis-tag nanofibers could firmly bind Ni-NTA–capped Au NPs (fig. S3A). We thus reasoned that our Au NP–bound CsgA nanofiber coatings could theoretically lead to a gold enhancement process on the surface of a substrate, potentially forming Au coatings consisting of closely packed NPs.

To test the feasibility of our concept, we first incubated a CsgAHis-tag–coated three-dimensional (3D)–printed pyramid with Ni-NTA–capped Au NPs. After assembly for 30 min, we transferred this pyramid into a “gold enhancement” solution (AuCl4 and hydroxylamine), allowing chemical reduction (Fig. 3A). Photographic images showed that the surface color of the pyramid was successfully changed from pristine white to typical tan (Fig. 3B). The above experimental results thus showed the feasibility of our fabrication process. The simple Au coating technique could be easily applied to various substrates, including polyimide (PI), PDMS, PET, PTFE, and PP, highlighting the substrate independence and conformability features of our nanofiber coatings (fig. S3B).

Fig. 3 Construction and characterization of CsgAHis-tag–Au conductive coatings and electronic devices.

(A) Schematic showing the fabrication of Au coatings based on CsgAHis-tag coatings. (B) Digital images of pristine and Au-coated CsgAHis-tag–modified 3D printed pyramids. Photo credit: Yingfeng Li, ShanghaiTech University. (C) Digital (left) and SEM (right) images of an Au interdigital electrode fabricated by a CsgAHis-tag coating–enabled gold enhancement process assisted by a patterned waterproof sticker. Photo credit: Yingfeng Li, ShanghaiTech University. (D) XPS spectrum of the Au interdigital electrode. (E) Capacitance change of the Au interdigital electrode with different distances between the electrode and a finger; the inset digital images indicate different distances. Photo credit: Yingfeng Li, ShanghaiTech University. (F) Digital images of the Au interdigital electrode as the sensing element in a touch switch. Photo credit: Yingfeng Li, ShanghaiTech University. (G) Digital (left) and SEM (right) images of pristine (top) and Au-coated textiles (bottom). Photo credit: Yingfeng Li, ShanghaiTech University. (H) Schematic diagram of a pressure sensor fabricated by Au-coated textiles along with an Au interdigital electrode (inset) and the corresponding current variation (∆I/I0) under different pressures. (I) Current variation as a function of time at two pressures (the inset digital images indicate the two different types of pressure applied). Photo credit: Yingfeng Li, ShanghaiTech University.

Having demonstrated the feasibility of conformable Au coating technique using the CsgAHis-tag protein as functional coating proteins, we next explored the fabrication of diverse electronic devices with increasingly complex functionalities. We first generated patterned Au coatings by first fabricating CsgAHis-tag coatings with commercially available patterned waterproof stickers, then incubating the substrates in an Au NP solution followed by an Au enhancement process (see the Supplementary Materials). Accordingly, we fabricated an interdigital electrode consisting of patterned Au coatings on a PDMS substrate that conformably stuck to the outer surface of a 50-ml centrifugation tube (Fig. 3C). As expected, SEM and AFM images indicated that the coating was composed of NPs, and further XPS analysis confirmed the appearance of Au element on the surface (Fig. 3, C and D, and fig. S3C). To demonstrate the potential application of this interdigital electrode, we carefully tested the capacitance change of the electrode when a finger approached and then moved away from the electrode. As illustrated in Fig. 3E, as a finger gradually began touching the electrode, the capacitance correspondingly decreased. Likewise, when the finger was removed, the capacitance was restored to the original value.

This behavior is attributed to the higher dielectric constant of the human body as compared to air: a higher dielectric constant reflects lower capacitance. In this way, such an electrode could be used as the sensing unit of a touch switch (40). We therefore linked this electrode to a circuit including a power source, a commercially available signal processing chip, and a light-emitting diode (LED). As shown in Fig. 3F, when no finger was in contact with the electrode, the LED was off; however, when a finger touched the electrode, the circuit was connected and the LED was on.

To assess the mechanical stability of the conductive Au coatings, we applied an abrasion test for our CsgAHis-tag–enabled Au conductive coatings following a previous approach for coating structures (37, 41). Specifically, we first attached a soft PET fabric on the Au-coated PET plate, followed by placing a 2-kg counterweight on the fabric. We then moved the fabric against the conductive surface of PET plates. As illustrated in fig. S4A, the sheet resistance had almost no change (~23 ohms/sq) even after 500 cycles of abrasion. In addition, although SEM images showed the abrasion traces on the surfaces, the morphology of the conductive layers consisting of highly packed irregular Au NPs remained unchanged (fig. S4, B to D). These findings highlight the mechanical robustness of our conductive coatings on the PET plates. Because CsgAHis-tag coatings are vulnerable to enzymatic digestions, we next used trypsin and protease AO to challenge the Au conductive coatings. The sheet resistance and the microstructures of conductive coatings had negligible changes after incubation with the enzyme solutions for 24 hours, indicating the strong resistance of Au coatings to proteolytic digestion (fig. S5, A and B). It is likely that the extremely compact Au coatings above the nanofiber coatings could hinder the direct contact of enzymes with CsgAHis-tag nanofiber coatings and thus protected the nanofiber layers from enzymatic digestion.

Motivated by the impressive durability of CsgAHis-tag nanofiber–enabled Au conductive coatings, we next turned to explore more exciting applications based on such coatings. We first coated PET textiles with CsgAHis-tag nanofibers and then fabricated Au-coated conductive textiles (fig. S6A). Photographic and SEM images indicated the vast differences between textiles in apparent color and micromorphology after the formation of Au coatings (Fig. 3G). Furthermore, energy-dispersive spectroscopy (EDS) result implied the uniform distribution of Au on the textile surface, and electron backscatter diffraction (EBSD) analysis showed that the in situ generated Au NPs were closely anchored on the entire PET textile (fig. S6, B and C). We next constructed a pressure sensor based on our Au-coated PET textiles (Fig. 3H, inset). Briefly, we first used the Au-coated PET textile to cover the aforementioned PDMS-based Au interdigital electrode and sealed it with 3M VHB tape. The constructed pressure sensor worked as designed following a specific working principle as follows: When a certain pressure that led to the compression of the hierarchical porous textile was applied, the contact area between the textile and electrode was increased, so the contact electric current increased correspondingly under a constant voltage. When the external pressure was removed, the textile recovered from the deformation because of its inert elasticity, and the current returned to the initial state (42). The large surface area and sufficient surface roughness of the Au-coated textile, as revealed by SEM and EBSD images (Fig. 3H and fig. S6C), reliably reflect the changes in contact resistance resulting from an external stimulus.

We next carefully conducted several critical tests on the prepared pressure sensor. The sensitivity of the pressure sensor is defined as S = δ(ΔI/I0) /δP, where ΔI is the relative current change, I0 is the current without external pressure, and P is the applied pressure (42). In the range from 1.25 to 17.50 kPa, the relation between the change in current and the applied pressure was linear, and the sensitivity S was 8.3 kPa−1 (Fig. 3H). Figure 3I shows two representative current profiles (∆I/I0) under two different pressures (5 kPa and finger press). After 300 cycles of bending (1-cm bending radius) or 500 cycles of repeated 5-kPa presses, the values of ΔI/I0 under various external pressures had negligible changes, emphasizing the stable performance of the pressure sensor (tables S1 and S2). In general, our pressure sensor has high sensitivity (8.3 kPa−1), mechanical flexibility (300 bends), and cycle stability (500 cycles).

Programmable amyloid coatings for tunable fluorescent materials and enzymatic bioconversion systems

Functional protein-immobilized particles have a broad spectrum of applications in biosensor, biocatalysis, and drug delivery (4345). However, existing approaches for protein-based conjugation of microparticles are largely based on nonspecific interactions (e.g., electrostatic interactions in enzyme immobilization on silica) (46). Accordingly, these systems typically lack specificity and functional tunability. Note that CsgA is a genetically engineerable protein, so it can be appended with a variety of functional tags. We next explored the functional flexibility of CsgA coatings for diverse applications ranging from fluorescent coating materials to enzymatic immobilization on spherical particles for optimized bioconversion reactions. To this end, we first developed CsgASpyTag (SpyTag fused at the C terminus of CsgA)/CsgASnoopTag (SnoopTag fused at the C terminus of CsgA)–coated SiO2 microparticles as a platform to enable easy and flexible conjugation reaction systems (Fig. 4A). SpyTag and SnoopTag can covalently conjugate with their partners, SpyCatcher and SnoopCatcher, respectively (47, 48). Therefore, our CsgASpyTag/CsgASnoopTag coatings should be suitable for ligation of corresponding SpyCatcher- and SnoopCatcher-fused proteins.

Fig. 4 Programmable amyloid coatings for tunable fluorescent materials and enzymatic bioconversion systems.

(A) Illustration of CsgASpyTag/CsgASnoopTag (1:1, weight ratio)–coated microparticles. (B) SEM images of a CsgASpyTag/CsgASnoopTag-coated SiO2 microparticle. (C) Schematic showing fluorescent proteins conjugated on CsgASpyTag/CsgASnoopTag nanofiber (top) and fluorescence microscopy images of corresponding fluorescent protein–conjugated CsgASpyTag/CsgASnoopTag-coated microparticles. (D) Schematic showing the immobilization of LDHSpyCatcher and GOXSnoopCatcher on a CsgASpyTag/CsgASnoopTag-coated microparticle. (E) Illustration of a dual-enzyme reaction system enabled by LDHSpyCatcher and GOXSnoopCatcher co-conjugated microparticles. (F) Conversion ratio of l-tert-leucine in two different microparticle systems (LDHSpyCatcher and GOXSnoopCatcher co-conjugated together on CsgASpyTag/CsgASnoopTag coatings versus LDHSpyCatcher-conjugated CsgASpyTag coatings along with GOXSnoopCatcher-conjugated CsgASnoopTag coatings) during a 3-hour reaction period. (G) Conversion ratio of l-tert-leucine in the CsgASpyTag/CsgASnoopTag coating system over five cycles of 3-hour reactions.

SEM images showed that the SiO2 microparticle surface was successfully covered with CsgASpyTag/CsgASnoopTag nanofibers (Fig. 4B). Furthermore, the fluorescence spectra revealed that, compared to pristine particles, CsgASpyTag/CsgASnoopTag-coated microparticles exhibited an obvious enhancement in fluorescence intensity at ∼480 nm induced by the specific interaction between ThT molecules and β-sheet structures (fig. S7A). ATR-FTIR analysis of CsgASpyTag/CsgASnoopTag-coated microparticles showed an obvious absorption peak at ~1625 cm−1 corresponding to a β-sheet structure (fig. S7B) (49). In addition, XPS analysis of CsgASpyTag/CsgASnoopTag-coated microparticles revealed characteristic peaks of amide bonds originating from the coated proteins (fig. S7C). All the above results highlighted that the surface of SiO2 microparticles could be modified by our CsgASpyTag/CsgASnoopTag nanofiber coatings. Subsequent fluorescence microscopy images showed that these nanofiber-coated SiO2 microparticles displayed uniform bright red, green, and merged yellow fluorescence, confirming that SpyCatcher-fused mCherry (mCherrySpyCatcher) and SnoopCatcher-fused GFP (GFPSnoopCatcher) were successfully conjugated on the particle surfaces (Fig. 4C). Note that the microspheres stacking to each other displayed heterogeneous fluorescence strength in the image, which was likely due to their different focal planes under the fluorescence microscopy. Collectively, these results illustrate an alternative way of using nanofiber-coated microparticles to realize diverse applications.

We next applied a similar strategy to achieve multienzyme immobilization coupling with coenzyme regeneration. To this end, we first constructed SpyCatcher domain–fused leucine dehydrogenase (LDH; EC1.4.1.9; “LDHSpyCatcher”) and SnoopCatcher domain–fused glucose oxidase (GOX; EC1.1.3.4; “GOXSnoopCatcher”) and coimmobilized on the CsgASpyTag/CsgASnoopTag-coated SiO2 microparticle (Fig. 4D). In this proof-of-concept reaction system, trimethylpyruvic (TMP) acid was converted into the high-value chemical l-tert-leucine by LDH from the soil bacterium Lysinibacillus sphaericus, a reaction that requires NADH [reduced form of nicotinamide adenine dinucleotide (NAD+)] as a coenzyme. Moreover, GOX from Bacillus subtilis can regenerate NADH by oxidizing low-value glucose into gluconic acid (Fig. 4E) Therefore, these two enzymes could assemble into an NADH-recycling system (Fig. 4E). We chose LDHSpyCatcher and GOXSnoopCatcher conjugated onto CsgASpyTag- and CsgASnoopTag-coated microparticles, respectively, as a control group. We used high-performance liquid chromatography (HPLC) to analyze the conversion ratio of l-tert-leucine.

As shown in Fig. 4F, in the first 3-hour reaction, the conversion ratio of l-tert-leucine in the CsgASpyTag/CsgASnoopTag coating system was about 50%, whereas there was only 30% conversion in the control system. We speculate that the substantial disparity may lie in substrate channeling (50). That is, in the CsgASpyTag/CsgASnoopTag coating system, the generated NADH could be immediately consumed by adjacent LDHSpyCatcher on the same particle surface. However, in the control system, the produced NADH would not be used until it arrived at the surface of LDHSpyCatcher-conjugated particles, thereby resulting in a slower reaction rate.

To demonstrate the recyclable use of these immobilized enzymes, we recollected the enzyme-conjugated CsgASpyTag/CsgASnoopTag-coated microparticles via simple centrifugation. We then transferred these particles into a new reaction solution and again assessed the conversion ratio of l-tert-leucine. We found that the ratio did not significantly change over a series of five reaction cycles of 3 hours each (Fig. 4G). These experimental results demonstrate that our genetically engineered protein coatings are highly suitable for biocatalytic applications.

Functional CsgADBD coating–enabled microfluidic bacterial sensors

RNA-cleaving fluorogenic DNAzyme (RFD) is a well-established technology for detecting bacteria, and the ability to immobilize RFD probes on material surfaces such as the interiors of microfluidic devices is highly demanded because it could enable substantial improvements in the efficiency and speed of detection (5153). Our genetically engineered CsgA fusion coatings represent a potentially alternative approach. We produced CsgADBD proteins with a C-terminally fused DNA-binding domain (DBD) originally from Vibrio fischeri (fig. S8A) (54). We aimed to use this tailored protein to modify the surface of a microfluidic channel and bind E. coli–specific RFD probes. We expected that upon interaction with target molecule(s) present in the supernatants of E. coli bacteria, these bound RFD probes would be converted into an active state that can catalyze the cleavage of the fluorogenic substrate, thereby producing a detectable fluorescent signal on the interiors of the microfluidic channel (Fig. 5A) (52).

Fig. 5 Functional CsgADBD coating–enabled microfluidic bacterial sensors.

(A) Schematic diagram of a DNAzyme-bound CsgADBD-coated microfluidic sensor device and an illustration of the DNAzyme detection mechanism. (B) Digital image of the microfluidic device. Photo credit: Yingfeng Li, ShanghaiTech University. (C) Fluorescence intensity of RFD-functionalized CsgADBD- and CsgAHis-tag–coated interiors of microfluidic channels upon exposure to supernatants from E. coli cultures of various cell densities. (D) 3D image of the RFD-functionalized CsgADBD coatings activated by E. coli culture (OD600 = 1) supernatants on the microfluidic channel.

To demonstrate the feasibility of our general design, we first incubated CsgADBD nanofibers with RFD probes. Agarose gel electrophoresis analysis indicated that CsgADBD nanofibers were able to bind these probes (fig. S8B). A standard PDMS microfluidic device was used for this experiment (Fig. 5B). We first coated the interior of a microfluidic channel and then conducted a Ni-NTA–capped QD binding test (the His-tag used for purification of CsgADBD protein can also be used to bind these QDs). The fluorescence microscopy image indicated that the channel interiors were homogeneously modified by CsgADBD proteins (fig. S8C). We next tested the detection performance by injecting a filtered supernatant from an E. coli culture into the channel and found that the CsgADBD-coated channel generated a strong fluorescent signal, whereas a control channel with a CsgA coating did not (Fig. 5C). Moreover, the fluorescence intensity increased linearly with the number of E. coli cells present in the samples [measured as OD600 (optical density at 600 nm); Fig. 5C]. In addition, 3D reconstructed images from fluorescence microscopy further confirmed that the resulting fluorescence was on the channel surface (Fig. 5D). These results establish proof of concept for the use of our genetically engineered protein coatings in diagnostic devices to monitor specific infectious pathogens.


In summary, we demonstrate that genetically engineered CsgA fusion proteins can be used as a functional coating system. These coatings have substrate universality, ultrastability, and genetically programmable functions. We also confirm that genetically engineered CsgA fusion protein nanofibers can modify various substrates with different compositions, sizes, shapes, and structures and show that these coatings exhibit outstanding chemical robustness. Moreover, these protein coatings offer flexible genetically programmable functionalization (e.g., NP anchoring, protein conjugation, and DNA binding). By combining the coatings with various fabrication processes, we established multiple proof-of-concept applications, including touch switching, pressure sensing, enzyme immobilization, and microfluidic sensors for bacterial detection. Given these unique coating features and the development of protein conjugation technologies, our genetically engineered CsgA fusion protein nanofiber coatings should serve as a versatile surface functionalization platform for electronics, biocatalysis, textiles, biomedicine, and other application areas.


Plasmid construction

All genes were synthesized by GENEWIZ and then amplified by polymerase chain reaction. The DNA fragment was cloned into pet-22b vectors (Nde I and Xho I sites) using one-step isothermal Gibson assembly. All constructs were sequence-verified by GENEWIZ.

Protein expression and purification

For CsgAHis-tag, CsgASpyTag, CsgASnoopTag, or CsgADBD protein, the corresponding plasmid was transformed into BL21(DE3) E. coli competent cell. The bacterial seed was grown for 16 hours at 37°C in shaking flasks (220 rpm/min) containing 20 ml of LB medium supplemented with carbenicillin (50 μg/ml). The culture was then added into 1 liter of LB and grown to OD600 ~1.0. Protein expression was induced with 0.5 mM isopropyl-β-D-thiogalactopyranoside (IPTG) at 37°C for 45 min. Cells were collected by centrifugation for 10 min at 4000g at 4°C. The cell pellet was then lysed in 50 ml of GdnHCl [8 M, 300 mM NaCl, 50 mM K2HPO4/KH2PO4 (pH 8)] for 12 hours at room temperature. Supernatants of the lysates were collected at 12,000g for 30 min before loading in a His-Select Ni-NTA column. The column was washed with KPI [300 mM NaCl, 50 mM K2HPO4/KH2PO4 (pH 8)] buffer and 40 mM imidazole KPI buffer and then eluted with 300 mM imidazole KPI buffer.

For mCherrySpyCatcher, GFPSnoopCatcher, LDHSpyCatcher, or GOXSnoopCatcher protein, the corresponding plasmid was transformed into BL21(DE3) E. coli competent cell. Cell seeds were cultured for 16 hours at 37°C in LB broth containing carbenicillin (50 μg/ml). The culture solution was then added into 1 liter of LB and grown to OD600 ~0.6. Protein expression was induced with 0.5 mM IPTG for 12 hours at 16°C. Cells were collected by centrifugation for 10 min at 4000g at 4°C. The collected cell pellets were then resuspended in KPI solution (50 ml) containing lysozyme (1 mg/ml) and incubated on ice for 30 min before ultrasound disruption. The purification follows the same procedure used for purification of the genetically engineered CsgA proteins. The purified proteins were stored at 4°C for later use.

Coating fabrication

To enable coating formation, given substrates (plates, pyramids, or textiles) were directly immersed in fresh eluted CsgAHis-tag monomer (1 mg/ml) solution. After 16 hours of incubation at room temperature (~25°C), proteins could form nanofiber coatings on substrates. The coated substrates were then washed by deionized H2O and dried by clean N2 and finally stored in a desiccative cabinet (~25°C) for further use.

To coat microparticles with functional proteins, 1 ml of CsgASpyTag, CsgASnoopTag, or CsgASpyTag/CsgASnoopTag (1:1, weight ratio) monomer solution (1 mg/ml) was added into 2-ml tube containing 100 μl of SiO2 aqueous solution (25 mg/ml). After 16 hours of incubation at room temperature (~25°C), microparticles were collected by centrifugation for 5 min at 1000g and washed by deionized H2O followed by further centrifugation. This process was repeated for three times to remove the loosely bound proteins. The coated microparticles were then stored in a 4°C refrigerator for further use.

PDMS channel was first fabricated by replica molding of a glass model and then pressed on the surface of a clean glass slide. To coat the PDMS microfluidic device channel, fresh eluted CsgADBD monomer solution was directly injected into the channel using a syringe and incubated for 16 hours at room temperature (~25°C). The microfluidic channel was then washed by deionized H2O through injection. The microfluidic device was stored in the refrigerator (4°C) for further use.

QD binding test

Synthesis of Ni-NTA–capped QDs was performed following a previous report (33). To ensure thorough QD binding on protein-coated flat substrates, the substrates were immersed in the aqueous QD solution (ca. 500 nmol/ml) at room temperature (~25°C) and incubated for 30 min. The substrates were then washed by deionized H2O and dried by high-pressure N2 for further characterization. To ensure QD binding in a microfluidic device, QD solution was injected into the channel using a 1-ml syringe. After incubation for 30 min at room temperature (~25°C), the channel was washed by deionized H2O for further characterization.

Stability test

For the stability test of CsgAHis-tag coatings in organic solvents, 30 ml of acetone, hexane, or DMSO was poured into a 9-cm glass culture dish containing the CsgAHis-tag–coated PTFE substrates. After challenge at room temperature (~25°C) for 24 hours, the PTFE substrates were washed by deionized H2O and dried by high pressure N2 for further characterization. For the high temperature challenge, CsgAHis-tag–coated PTFE substrates were directly placed in an oven (90°C) for 24 hours and then taken out for further characterization.

Enzyme digestion test

CsgAHis-tag–coated PTFE substrates or Au-coated PET substrates were placed in 9-cm culture dishes containing 30 ml of solution of trypsin (2.5 mg/ml) from bovin pancreas or fungal protease (≥55 U/g) from A. oryzae (protease AO). After incubation at 37°C for 24 hours, substrates were washed by deionized H2O and dried by high-pressure N2 for further characterization. For ThT assay, 100 μl of enzyme solution (trypsin, 2.5 mg/ml or fungal protease, ≥55 U/g) was added into the 96-well microplate containing 100 μl of CsgAHis-tag nanofiber protein solution (0.5 mg/ml). ThT was then added to a concentration of 20 μM. Fluorescence was measured every 0.5 min after shaking 5 s with a BioTek Synergy H1 microplate reader (excitation at 438 nm, emission at 495 nm, and cutoff at 475 nm) at 37°C.

Gold enhancement process

Preparation of Ni-NTA–capped Au NPs was based on a previous report (33). To perform a gold enhancement process, CsgAHis-tag nanofiber–coated pyramid or textile substrates were first immersed into Ni-NTA–capped Au NP solution. After incubation at room temperature (~25°C) for 30 min, the substrates were washed with deionized H2O and dried by high-pressure N2. The substrates were then transferred into a 50-ml gold enhancement solution containing AuCl4 (50 mg/ml) and hydroxylamine (100 mg/ml). After reaction for 10 min at room temperature (~25°C), substrates were washed by deionized H2O and dried by high-pressure N2.

To prepare patterned Au coatings including interdigital electrode, substrates were first covered by waterproof stickers followed by producing patterned CsgAHis-tag nanofiber coatings through protein solution incubation. The patterned CsgAHis-tag nanofiber coatings were then bound with Ni-NTA–capped Au NPs, followed by a standard gold enhancement procedure described above. After drying, the stickers were carefully peeled off using a tweezer to produce the patterned CsgAHis-tag nanofiber–enabled Au coatings.

Abrasion test

Bare PET fabric was attached on the Au conductive coatings formed on a PET plate, followed by placing a 2-kg counterweight on the fabric. The abrasion test was achieved by moving the bare PET fabric. Sheet resistance of PET-based conductive coatings was measured with a four-probe ohmmeter (HPS 2523).

Electronic device test

The capacitance of the interdigital electrode was measured with an LCR (inductance, capacitance, and resistance) meter (HG2817A) at a voltage of 1 V and a frequency of 100 kHz at room temperature (~25°C). To fabricate the pressure sensor, Au-coated PET textile was covered on the PDMS-based Au interdigital electrode. Then, the textile and bottom Au electrode were sealed with a 3M VHB tape. Functional performances of the pressure sensor including current change under different pressures were assessed with an electrochemical work station (CHI 660E) at room temperature (~25°C).

Protein conjugation

For fluorescent protein conjugation, 1 ml of mCherrySpyCatcher/GFPSnoopCather (1:1, weight ration) aqueous solution (1 mg/ml) was added into a 2-ml tube containing the CsgASpyTag/CsgASnoopTag-coated SiO2 microparticles. After incubation for 1 hour at room temperature (~25°C), fluorescent protein–conjugated microparticles were collected by centrifugation for 5 min at 1000g and washed by KPI solution followed by further centrifugation. This process was repeated for three times to remove those unreacted loosely bound fluorescent proteins.

For enzyme immobilization, 1 ml of LDHSpyCatcher, GOXSnoopCather, or LDHSpyCatcher/GOXSnoopCather (1:1, weight ration) aqueous solution (1 mg/ml) was added into a 2-ml tube containing the CsgASpyTag-, CsgASnoopTag-, or CsgASpyTag/CsgASnoopTag-coated SiO2 microparticles, respectively. After incubation for 1 hour at room temperature (~25°C), the enzyme-immobilized microparticles were collected by centrifugation for 5 min at 1000g and washed by 100 mM phosphate buffer followed by further centrifugation. This process was repeated for three times to remove those unreacted loosely bound enzymatic proteins.

Enzyme activity test

The enzyme-immobilized microparticles were resuspended in 100 μl of 100 mM phosphate buffer, 50 μl of LDHSpyCatcher immobilized microparticles, and 50 μl of GOXSnoopCatcher immobilized microparticles added into the reaction solution, or 100 μl of LDHSpyCatcher/GOXSnoopCatcher immobilized microparticles was directly pipetted into the reaction solution. The reaction mixture containing 50 mM glucose, 0.1 mM NAD+, 50 mM ammonium chloride, 50 mM TMP acid, and 100 mM phosphate buffer (pH 8.0) was fixed with a final volume of 1 ml. The reaction was then conducted at 37°C under continuous shaking in a microplate reader.

To analyze the yield of l-tert-leucine, a 20-μl sample was filtered with a 220-nm syringe filter and analyzed by reversed-phase HPLC using a 1200 Series chromatograph and ZORBAX SB-C18 column (4.6 mm × 150 mm, 5 μm) at 35°C. The mobile phase composed of 2 mM CuSO4 was set with a flow rate of 1.0 ml/min. Quantitative analysis of the l-tert-leucine was monitored with a UV spectra detector at 210 nm (55).

The yield of l-tert-leucine was determined using the following equationη=Practical concentration of L-tert-leucineTheoretical concentration of L-tert-leucine (50 mM)×100%

For recyclable usage of the enzymes, the LDHSpyCatcher/GOXSnoopCatcher immobilized microparticles were collected by centrifugation for 5 min at 1000g after each round of reaction. The microparticles were then resuspended in 100 μl of 100 mM phosphate buffer and pipetted into a new reaction solution for another new round of reaction. The yield of l-tert-leucine in the new reaction system was determined following the same equation.

Microfluidic sensor fabrication and test

The synthesis of RFD probes was based on a protocol described in a previously published study (52). The microfluidic channel was then homogeneously coated with CsgADBD proteins following a typical fabrication protocol described in the coating fabrication process.

To ensure thorough binding of RFD probes onto the protein coatings on the microfluidic channel, DNAzyme in 100 mM tris-HCl (pH 8.0) and 0.2 mM EDTA binding buffer was injected into the microfluidic channel. After incubation for 2 hours at room temperature (~25°C), the channel was washed with 1 ml of injected 100 mM tris-HCl to remove loosely bound RFD probes.

E. coli K12 (MG1655) cell culture with different cell densities (OD600) was injected into the channels after filtration using a 220-nm PTFE filter. The channel was monitored by fluorescence microscopy, and the relative fluorescence intensity was calculated using the imaging software of the fluorescence microscopy.

AFM imaging

Samples were tested with Asylum MFP-3D-Bio using the tapping mode with AC160TS-R3 cantilevers (Olympus, k ≈ 26 N/m, ν ≈ 300 kHz). The data are presented in Fig. 2B and figs. S1B, S2C, and S8A.

Contact angle test

The water contact angle of samples was tested with a contact angle goniometer (SL200KS). The substrate was placed on the stage, and 1-μl droplet of water was dropped onto the surface of the substrate. The data are presented in Fig. 2 (A, F, G, and H) and fig. S2 (C, G, and H).

X-ray photoelectron spectroscopy

XPS spectrum was obtained with Thermo Fisher Scientific ESCALAB 250 Xi. The data are presented in Figs. 2C and 3D and fig. S7C.


NanoDSF curve was obtained with NanoTemper Prometheus NT.48. The data are presented in fig. S2A.

SEM imaging

Samples were coated with Au for 30 s with an SBC-12 sputter coater. SEM images including EBSD and EDS images were acquired with JEOL 7800 Prime or JSM-6010. The data are presented in Figs. 3 (C and G) and 4B and figs. S2 (F to H), S4 (B to D), S5B, and S6 (A to C).

TEM imaging

TEM images were obtained on an FEI T12 transmission electron microscope operated at 120-kV accelerating voltage. The data are presented in fig. S3A.

ATR-FTIR spectroscopy

Protein-coated microparticles, bare microparticles, or protein nanofibers were put on the ATR crystal directly. Spectra were recorded from 1700 to 1600 cm−1 using a nominal resolution of 2 cm−1 with Spectrum Two (PerkinElmer). The data are presented in figs. S2B and S7B.

Fluorescence microscopy imaging

Fluorescence imaging was performed on an Olympus IX83, Leica DMi8, or LSM 710 fluorescence microscope. Cy5 channel of Leica DMi8 was used to image RFD. The data are presented in Figs. 4C and 5D and fig. S8C.

Fluorescence spectroscopy

Photoluminescence spectra were collected using HORIBA FL-3 with excitation at 350 nm. The data are presented in figs. S1A and S7A.


Supplementary material for this article is available at

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: We thank X. Wang for AFM training. AFM characterization was executed at the Analytical Instrumentation Center (AIC), and SEM and TEM characterization were performed at the Electron Microscopy Center (EMC) at School of Physical Science and Technology (SPST), ShanghaiTech University. Funding: This work was partially sponsored by the Commission for Science and Technology of Shanghai Municipality (grant no. 17JC1403900), the Joint Funds of the National Natural Science Foundation of China (Key Program No. U1932204), and the National Science and Technology Major Project of the Ministry of Science and Technology of China (grant no. 2018YFA0902804). C.Z. also acknowledges start-up funding support from ShanghaiTech University and 1000 Youth Talents Program, granted by the Chinese Central Government. Author contributions: C.Z. conceived the concept and directed the research. C.Z., Y.L., and K.L. designed and conducted the experiments and data analysis. X.W. synthesized QDs and performed TEM. M.C. participated in coating fabrication process. P.G. and J.Z. fabricated microfluidic devices. F.Q. participated in protein purification. C.Z., Y.L., and K.L. wrote the manuscript with help from all authors. Competing interests: The authors have filed a provisional patent based on this work with the China Intellectual Property Office (PCT/CN2018/085988). The authors declare no other competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.
View Abstract

Stay Connected to Science Advances

Navigate This Article