Amino acid coating enables micromotor operation in physiological conditions

Significance

In recent years, the use of self-electrophoretic micro/nanomotors in biomedicine has grown, prompting a focus on their stability in high-ionic-strength environments. Here, we propose a simple amino acid modification strategy that significantly boosts the ion tolerance of light-driven titanium dioxide motors, enabling functionality in biological media. This is achieved by increasing surface conductivity, allowing the motors to work in high-ionic environments and, combined with light and ultrasound, effectively induce tumor cell apoptosis. This approach opens possibilities for biomedical applications of self-electrophoretic motors.

Abstract

Physiological environment with high ionic strength will quench the propulsion of micro/nanomotors (MNMs) by suppressing electric double layers, especially for those motors based on electrolyte diffusiophoresis and electrophoresis. Herein, we demonstrate an efficient, general, and simple strategy to improve the ion tolerance of light-driven titanium dioxide (TiO₂) micromotors with amino acid surface modification. Compared to the bare TiO₂ counterpart, L-arginine (Arg)-treated TiO₂ micromotors display over 200 times higher ion tolerance, which is mainly attributed to the increased surface conductivity. This simple ion tolerance improvement strategy can also be applicable to other motors driven by self-electrophoresis. As TiO₂ is an efficient sonosensitizer, we combined the light-guiding ability with ultrasound to generate reactive oxygen species to effectively induce in situ tumor apoptosis. We envision that this simple amino acid surface modification can not only provide a solution for MNMs to tolerate the ionic environment but also open up opportunities for further biomedical and translational research of MNMs.

Micro/nanomotors (MNMs) are artificial miniature devices capable of converting chemical or external energies from the surrounding environment into mechanical motion. Due to their autonomous movement, MNMs have attracted significant research interest in various fields (1–8), such as environmental remediation and biomedicine. In terms of biomedical applications, MNMs are excellent candidates for the next generation of therapeutic micro- and nanocarriers, thanks to their efficient cargo towing and effective penetration capabilities. Collectively, the studies in the motor field surged in the past decades, leading to the successful design and fabrication of a variety of MNMs with multifunction of loading, transportation, and release of cargo (such as nucleic acids, proteins, and cells) (9–15). The unceasing development of MNMs has gradually made “nanorobots” in science fiction and movies a reality and is also gradually answering the bold conjecture of Nobel laureate Richard Feynman in 1959 in his famous speech “There is Plenty of Room at the Bottom” of “swallowing a surgeon.”

However, the complex biological environment is a long-standing challenge for the efficient actuation of MNMs, where the plasma protein biofouling, high viscosity, and high ionic strength all deteriorate the performance of MNMs (16, 17). Most crucially, the inability of MNMs to function in a highly saline solution is mainly attributed to an inherent property of the electrophoresis process, based on the fundamental theory of electrokinetic (16). Therefore, it is still challenging to utilize electrophoretic MNMs for biological applications. The driving force of electrophoretic MNMs originated from the Coulombic interaction between the charged surface of MNMs and the oppositely charged Debye layer. This interaction spans hundreds of nanometers in a low-concentration electrolyte solution (18, 19). However, the Debye layer on the surface of MNMs collapses significantly in high ionic conditions. And the counterions are drawn to the MNMs to screen the surface charge and result in a net decrease in ionic concentration in the diffuse layer. This hinders the build-up of ionic concentration gradients around the MNMs, thus limiting the motion capability of MNMs (18, 19). Therefore, improving the ion tolerance of MNMs has been the focus of numerous studies. Tang et al. (20) showed that the ion tolerance of self-electrophoretic motion is proportional to the surface conductivity of MNMs, which could be improved with ion-conductive polyelectrolyte coating. However, in situ polymerization of polymeric coating to the motor surface is complex and difficult to control, while the ion tolerance is still insufficient to support biomedical application. In addition to polyelectrolyte coating, porous materials with high photocatalytic activity can also overcome the bottleneck of ion tolerance. Metin Sitti et al. (21) synthesized poly(heptazine imide) (PHI) carbon nitride porous microparticles as light-driven micromotors, which maintains efficient mobility even in high ionic strength [up to 5 M sodium chloride (NaCl)] and biological media. They addressed the high ion tolerance as the result of an interplay between the optoionic effects and the microparticle’s structural and textural porosity, which allowed ion migration into and through the micromotors to overcome the ionic quenching behavior. However, the PHI structure and activity are not generally available, and a general, simple, and effective method for ion tolerance improvement is still highly desirable.

Here, using light-driven titanium dioxide (TiO₂) as a model of self-electrophoretic MNMs, we show that the Achilles heel of electrophoresis MNMs can be overcome with biocompatible zwitterion L-arginine (Arg) surface modification via simple electrostatic adsorption (Fig. 1). By significantly increasing the surface conductance with the absorbed Arg to suppress the collapse of Debye layers, the ion tolerance of TiO₂ micromotors has increased over 200 times. The efficacy of other amino acids, including glycine (Gly), glutamic acid (Glu), and phenylalanine (Phe), to the ion tolerance is tested and demonstrated as well. This simple ion tolerance improvement strategy can also be applicable to other motors driven by self-electrophoresis. To demonstrate the biomedical application of ion-tolerant MNMs, Arg-modified TiO₂ motors are utilized for the active sonodynamic treatment of tumor cells. Under ultraviolet (UV) light, TiO₂-Arg motors are steered toward 4T1 tumor cells, and most of the cells are killed when US stimulation is applied. With this simple, general, and effective approach for constructing MNMs with high ion tolerance just by surface modification of amino acids, the biomedical applications of self-electrophoretic motors are expected soon.

Fig. 1.

(A and B) Schematic of the TiO₂ micromotor without and with Arg modification. Arg was physically adsorbed on the surface of TiO₂. (C) TiO₂-Arg with ionic tolerance presented controllable navigation toward the targeted 4T1 cell under the guidance of UV light. Then ROS was released by US stimulation, resulting in tumor cell apoptosis.

Results

Fabrication and Characterization of TiO₂ Micromotors.

TiO₂ microparticles were first synthesized based on a sol–gel approach (Fig. 2A) (22). As shown in SI Appendix, Fig. S1, TiO₂ microparticles with spherical morphology were clearly observed by transmission electron microscopy (TEM). The resulting TiO₂ microparticles (0.83 mg/mL) were then added into Arg solution (0.58 mg/mL), and the layer of Arg was formed through electrostatic adsorption (23). The TEM image of TiO₂-Arg micromotors (SI Appendix, Fig. S2) shared a similar morphology to TiO₂ with an average hydrodynamic particle size of about 1,048 nm (Fig. 2B). Energy-dispersive X-ray (EDX) spectroscopy mapping analysis was further performed to verify the elemental composition of TiO₂ and TiO₂-Arg motors. EDX images illustrated the existence and the distribution of C, Ti, O, and N, respectively (Fig. 2C and SI Appendix, Fig. S3). The zeta potential of TiO₂ before and after Arg modification was then measured. After Arg adsorption, the zeta potential of TiO₂-Arg micromotors was further decreased from −35.7 ± 2.8 to −43.3 ± 3.4 mV (Fig. 2D), demonstrating the successful modification of Arg on the surface of motors. The chemical structure of TiO₂-Arg was further characterized by Fourier-transform infrared spectroscopy (FTIR) (SI Appendix, Fig. S4). In the 3,300 to 3,500 cm⁻¹ region, both TiO₂ (24) and TiO₂-Arg exhibited absorption peaks, primarily attributed to N-H stretching (from amino or guanidinium groups of Arg) and O-H stretching (from TiO₂ hydroxyl groups or adsorbed water). After Arg functionalization, two distinct peaks at 3,321 cm⁻¹ and 3,331 cm⁻¹ appeared in TiO₂-Arg, whereas bare TiO₂ only exhibited a broader peak at 3,250 cm⁻¹. This shift suggested the incorporation of N-H vibrations from Arg and its interaction with TiO₂ surface groups. In the 1,600 to 1,700 cm⁻¹ region, FTIR spectrum of bare TiO₂ showed a peak at 1,625 cm⁻¹, attributed to bending vibrations of surface hydroxyl (-OH) or adsorbed water. Upon Arg functionalization, this peak shifted slightly to 1,631 cm⁻¹, likely due to overlapping contributions from intrinsic absorption of TiO₂ and functional groups (C═O in the carboxyl group or C═N in the guanidinium group) of Arg, which interacted with TiO₂ surface via hydrogen bonding or electrostatic interactions. However, due to the strong intrinsic absorption of TiO₂ in this region, Arg-related signals were not easily distinguishable in FTIR spectra. To further confirm the functionalization of Arg, thermogravimetric (TG) analysis was conducted to quantify the mass ratio of Arg modification on TiO micromotors. TiO₂ micromotors only had one stage with thermal weight loss, which was caused by physically adsorbed water molecules and the chemisorbed-OH. In contrast, for TiO₂-Arg micromotors (Fig. 2E), two stages of thermal weight loss during the heating process were clearly observed. The thermal loss in the first stage was attributed to the same cause as that of bare TiO₂, namely, the loss of physically adsorbed water molecules and chemisorbed hydroxyl groups (∼20%). The second stage was attributed to the desorption of Arg. Therefore, the mass ratio of Arg modification in TiO₂-Arg micromotors was 1.6%, thus supporting the FTIR results and further verifying the successful functionalization of TiO₂ by Arg. Next, the photoelectric behaviors of micromotors were tested by UV-Vis diffuse reflectance spectra (DRS), and the results of TiO₂-Arg and TiO₂ micromotors are presented in Fig. 2F and SI Appendix, Fig. S5, respectively. Their absorptions were confined to the UV region. The absorption edge for TiO₂-Arg was located at 370.59 nm, essentially coinciding with that of TiO₂. According to the Kubelka–Munk equation (25), the energy band gap (Eg) of TiO₂-Arg was calculated to be approximately 3.02 eV, indicating that the obtained micromotors could be activated by UV light.

Fig. 2.

Fabrication and characterization of TiO₂-Arg micromotors. (A) Fabrication process of TiO₂-Arg micromotors. (B) Particle size of TiO₂ and TiO₂-Arg micromotors measured by dynamic light scattering (DLS). (C) EDX images illustrating the elemental distribution of Ti, O, C, and N in TiO₂-Arg micromotors. (Scale bar: 500 nm.) (D) The zeta potential of TiO₂ and TiO₂-Arg micromotors measured by DLS (mean ± SD, n = 3). (E) TG measurement of TiO₂-Arg micromotors. (F) UV-vis DRS of TiO₂-Arg. [The Inset shows the linear fitting applied to the near-linear region (335 to 355 nm) of the UV-Vis absorption curve].

Light-Actuated Propulsion and Adaptive Motion of TiO₂-Arg Micromotors.

Under UV illumination at an oblique incidence angle (θ = 24.5°), TiO₂-Arg micromotors exhibit asymmetric light absorption, forming distinct illuminated and shadowed hemispheres (Fig. 3A). This spatial asymmetry gives rise to uneven photocatalytic activity across the surface, enabling directional redox reactions. Upon photoexcitation, electrons in the valence band (VB) are promoted to the conduction band (CB), generating electron–hole pairs (26). Due to the higher photon flux on the illuminated side, more carriers are produced, resulting in a locally elevated quasi-Fermi level. Consequently, the photogenerated electrons in the CB migrate from the illuminated (higher-energy) side to the shadowed (lower-energy) side, minimizing internal energy gradients. This process is further assisted by the accumulation of holes (h⁺) on the illuminated side, which raises the local surface potential and establishes a potential gradient across the particle surface. This separation of charge carriers drives spatially distinct redox reactions: Oxidation occurs on the illuminated hemisphere, while reduction occurs on the electron-rich shadowed side. The resulting ion imbalance along with the localized electric field gives rise to self-electrophoresis. Experimental observations support this mechanism, as an increase in ionic strength significantly reduces propulsion velocity, consistent with electrostatic screening effects that suppress self-electrophoretic motion. Photocatalytic water splitting acts as the fuel source in this system, producing H₂ and O₂ via spatially separated redox reactions. Although the net reaction does not directly yield free protons, the redox reactions lead to local accumulation of H⁺ and OH⁻ near the particle–fluid interface, inducing interfacial ionic imbalance and short-range electric fields that drive ionic migration. A contribution from diffusiophoresis cannot be entirely excluded, as the reaction products on the TiO₂-Arg surface may create a concentration gradient capable of inducing propulsion. This contrasts with previously reported mechanisms for isotropic TiO₂ micromotors powered by hydrogen peroxide, where oxygen gradients dominate and self-diffusiophoresis is the main propulsion mode (22).

Fig. 3.

Propulsion of TiO₂-Arg micromotors in different fluids. (A) Schematic of TiO₂-Arg micromotors actuated by UV light (B) Representative light-driven trajectory of a TiO₂-Arg micromotor in H₂O. Color encodes time progression from 0 to 7.5 s. (C) The movement speed and the corresponding mean square displacement (MSD) of TiO₂ and TiO₂-Arg as a function of time lag (Δt) (mean ± SD, n = 20). (D) Average speeds of TiO₂-Arg micromotors under different UV intensities in PBS. (The Inset shows the corresponding dependence of the normalized speed to the intensity of UV light.) (mean ± SD, n = 20). (E) The trajectory of a TiO₂-Arg motor, with time color-coded (yellow to purple), moving toward 4T1 (i) and a red blood cell (RBC) (ii). All experiments in Fig. 3, except for Fig. 3D, were conducted under UV light at a wavelength of 365 nm with an intensity of 80 mW/cm².

Building on the mechanistic insights, we evaluated the propulsion behavior of bare TiO₂ and arginine-functionalized TiO₂ (TiO₂-Arg) micromotors in deionized water and phosphate-buffered saline (PBS) under UV illumination (365 nm, 80 mW/cm²) with a light incidence angle of 24.5°. In the absence of illumination, irregular Brownian motion was observed for TiO₂-Arg particles (Movie S1). Upon UV illumination, they displayed directional movement along nearly straight trajectories in H₂O, indicating effective light-driven actuation (Fig. 3B and Movie S2). Quantitative analysis revealed that the velocity of TiO₂-Arg micromotors reached 19.50 ± 1.84 μm/s (Fig. 3C), slightly higher than that of unmodified TiO₂ micromotors, which moved at 17.16 ± 2.93 μm/s under the same conditions (Movie S3). This improvement is attributed to the increased surface potential introduced by Arg modification, as confirmed by zeta potential measurements (Fig. 2C), which further enhances the electrophoretic driving force. In PBS (ionic strength ~165 mM), TiO₂-Arg micromotors exhibited sustained propulsion, whereas unmodified TiO₂ micromotors showed only minimal movement, comparable to Brownian motion (SI Appendix, Fig. S6). To further investigate light-responsive behavior under physiological conditions, we then examined the motion of TiO₂-Arg micromotors in PBS under varying light intensities (Fig. 3D). The normalized speed of TiO₂-Arg micromotors showed a positive correlation with the light intensity. When the intensity of the applied light increased from 50 to 200 mW/cm², the speed enhanced simultaneously (Movie S4). In contrast, bare TiO₂ showed negligible light responsiveness in PBS, even at the highest light intensity (200 mW/cm², Movie S5). To assess micromotor stability in more complex physiological fluids, we further examined their behavior in Roswell Park Memorial Institute-1640 (RPMI-1640) medium (ionic strength ~153 mM), which simulates the nutritional environment of cell growth. Despite the more complex liquid environment compared to PBS, TiO₂-Arg micromotors retained directional propulsion under illumination, with a speed of 3.12 ± 0.91 μm/s (SI Appendix, Fig. S7), demonstrating their ability to move even in biologically relevant environment.

Future biomedical applications normally require MNMs that can precisely target specific location in complicated biological media. In order to demonstrate the precise motion behavior, TiO₂-Arg micromotors were guided toward the target 4T1 cell in RPMI-1640 and RBC in PBS (Fig. 3E), respectively (Movies S6 and S7). By controlling the on/off state and direction of the light source, TiO₂-Arg micromotors navigated along a predesigned path toward the target cells. The direction of motion can be adjusted by changing the angle of incidence, allowing the micromotors to target specific cells. The change in light angle drives the micromotors toward the direction of higher photocatalytic activity. However, the speed of the micromotor decreased significantly as the motor gradually approached the cell, likely due to the electrostatic interactions between the negatively charged micromotors and cell surfaces. Under the light guidance, individual micromotor independently executed self-navigation or self-targeting in ionic conditions, as demonstrated above.

We then examined the collective behaviors (27) of the micromotors. To rule out alternative clustering mechanisms such as phototaxis or photothermal effects, all experiments were conducted under spatially uniform UV illumination (incident at 24.5°) with less than 0.3% variation in light intensity across the 1,360 μm × 1,350 μm field of view. No external chemical fuels were introduced. Following UV illumination, TiO₂-Arg micromotors exhibited directional motion toward the regions of high photocatalytic activity. In deionized water, rapid aggregation was observed within 2 to 3 min, and dense flocks were formed by 5 min (SI Appendix, Fig. S8 and Movie S8). In contrast, aggregation in PBS was slower and less pronounced, with only moderate clustering observed after 6 to 7.5 min (SI Appendix, Fig. S9 and Movie S9). These results suggest that electroosmotic effects may still contribute under high-ionic-strength conditions; however, their influence is notably attenuated compared to low-ionic-strength environments. Photothermal effects are unlikely to explain the observed medium-dependent trend, as they do not vary with ionic strength. Mechanistically, this behavior arises from an intrinsic photoinduced electroosmotic process: Surface hydroxyl groups (–OH) on the TiO₂-Arg surface dissociate under illumination, generating H⁺ and OH⁻ ions (28, 29). The faster diffusion of H⁺ establishes a local electric field, which induces attractive electroosmotic interactions between negatively charged micromotors, thereby promoting their aggregation. Photoelectrochemical asymmetry further amplifies this field, guiding micromotors collectively toward active regions. Such electroosmotic interactions may also enable cooperative behaviors like cargo transportation or environmental sensing (29).

To evaluate the effect of UV irradiation on the motion performance of TiO₂-Arg micromotors, the samples subjected to 10 min of UV exposure were stored overnight and then tested in PBS solution. The results showed that despite prolonged UV exposure and extended storage, the micromotors remained capable of effective motion in PBS, indicating that the enhancement effect of Arg modification persisted. Additionally, to further investigate the degradation behavior induced by UV irradiation, liquid chromatography–tandem mass spectrometry analysis was conducted to monitor the changes in surface-bound Arg after UV exposure. The results revealed that a substantial portion of Arg underwent photodegradation, with an overall degradation rate of approximately 72% (SI Appendix, Fig. S10). Despite this degradation, the experimental results confirmed that the remaining Arg modification was still sufficient to sustain the motion of micromotors in ionic environments, further validating its persistent impact on ionic tolerance.

Enhanced Ion Tolerance of Self-Electrophoretic Micromotors with Arg Modification.

To quantify the ion tolerance of micromotors, we introduced the concept of “Media Effective Ionic Strength” (EI₅₀) (20), defined as the ionic strength of solution that causes a 50% decrease in the migration speed of MNMs. In self-electrophoretic MNMs, the ion tolerance is determined by surface conductivity, as described by

$${\text{EI}}_{50}={\mathrm{K}}^{\mathrm{\sigma }}/\mathrm{a}{\mathrm{\Lambda }}_{\mathrm{m}},$$ [1]

where K^σ represents the surface conductivity, “a” represents the particle characteristic size, and Λ_m represents the electrolyte molar conductivity. The speed of TiO₂-Arg and TiO₂ micromotors was measured across a range of ionic strengths (Fig. 4 A and B), showing a clear decrease in velocity with increasing ionic strength. This trend was further analyzed by plotting the inverse speed (1/speed) against ionic strength (Fig. 4 C and D), which yielded a linear relationship and allowed for extraction of the EI₅₀ value. Consistent with this analysis, meansquare displacement results also confirmed reduced motility at higher ionic strength (SI Appendix, Fig. S11). The EI₅₀ of the bare TiO₂ micromotor was determined to be 0.064 ± 0.0023 mM (Fig. 4E and Movie S10), consistent with the previously reported EI₅₀ of all electrophoretic MNMs (less than 0.1 mM) (19). In contrast, the EI₅₀ of TiO₂-Arg micromotors exhibited a markedly improved EI₅₀ of 15.69 ± 1.84 mM (Movie S11), representing a 245-fold enhancement in ionic tolerance compared to their unmodified counterparts.

Fig. 4.

Ion tolerance enhancement for TiO₂-Arg micromotors. The speed of TiO₂-Arg (A) and TiO₂ (B) micromotors at varying ionic strengths. (C) Linear fit of the inverse speed versus ionic strength for TiO₂-Arg. (D) Linear fit of the inverse speed versus ionic strength for TiO₂. (E) EI₅₀ values of TiO₂-Arg and TiO₂. (F) Electro-osmosis flow (EOF) velocity field around TiO₂ (i) and TiO₂-Arg micromotors (ii) in DI water. EOF velocity field around TiO₂ (iii) and TiO₂-Arg micromotors (iv) in PBS. All speed-related data are presented as mean ± SD (n = 20). All experiments with speed were conducted under UV light with a wavelength of 365 nm with an intensity of 80 mW/cm².

To gain mechanistic insight into this enhancement, we examined the electrokinetic behavior of micromotors under varying ionic conditions. Ion migration within the electrical double layer (Debye layer) (16), which governs all electrokinetic phenomena, is a critical factor in ion tolerance. In this instance, the thickness of the Debye layer significantly affects the effectiveness of the electrophoresis and electrolyte diffusiophoresis. The variation in EI₅₀ is not solely governed by ionic concentration but may also reflect attenuation of surface electric fields and a reduction in electroosmotic slip velocity. Although the primary propulsion mechanism is self-electrophoresis, the associated electric field also induces electroosmotic flow along the particle surface, which contributes to effective motion. This hypothesis is supported by our zeta potential measurements: Both TiO₂ and TiO₂-Arg micromotors exhibited a continuous decline in surface potential across the 0 to 150 mM ionic strength range (SI Appendix, Fig. S12), indicating progressive electric double layer (EDL) compression. This monotonic decrease in surface charge potential likely compromises both the self-generated electric field and the resulting electroosmotic propulsion, thereby contributing to the observed EI₅₀ values. Briefly, when the Debye layer is compressed, the effective region for electric field generation becomes narrower, weakening the electroosmotic flow (EOF). EOF refers to the fluid motion induced by a local electric field acting on the charged double layer near the TiO₂ micromotor surface and plays a key role in driving micromotor motion. Conversely, high surface conductivity enables ample ionic current flow along the Debye layer, despite the high solution conductivity, thereby generating enough EOF for micromotor propulsion. Consequently, enhancing surface conductivity can mitigate the collapse of the Debye layer, which prevents the quenching of self-electrophoresis. Therefore, a modified electrokinetic model for self-electrophoretic motors was proposed to highlight the significance of surface conductance in locomotion. The surface resistance of TiO₂ and TiO₂-Arg was measured using the transfer length method shown in (SI Appendix, Fig. S13). The average surface conductivity (K^σ) of TiO₂ and TiO₂-Arg were 1.43 × 10⁻⁷ S and 1.26 × 10⁻⁵ S (enhanced by two orders of magnitude), respectively. Therefore, we attribute the effect of Arg modification on the ion tolerance to the higher surface conductivity, which prevented the quenching of electrophoresis due to the Debye layer compression.

To reveal the mechanism behind the ion tolerance enhancement of TiO₂ micromotors after Arg modification, numerical modeling of the EOF velocity field around the micromotor in a solution with different ionic strengths was conducted. Regardless of the degree of alteration, both TiO₂ and TiO₂-Arg micromotors exhibited similar tangential EOF in deionized (DI) water (Fig. 4 F, i and ii). However, when TiO₂ motors were suspended in PBS, the formed EOF was suppressed entirely as expected due to electrophoresis quenching (Fig. 4 F, iii). In contrast, for TiO₂-Arg micromotors, an efficient EOF was clearly observed around the motor despite the self-electrophoresis field being blocked partially (Fig. 4 F, iv), which was the primary contributor to the improvement of ion tolerance. To sum up, the numerical simulation of EOF velocity and surface conductivity demonstrated that the TiO₂-Arg micromotor had higher surface conductivity, which supported the efficient propulsion by suppressing the compression of “Debye layers.”

Ion Tolerance in TiO₂ Micromotors: Amino Acid Modifications and Effects of Size.

In this section, we examine the ion tolerance of TiO₂-Arg micromotors, focusing on the effects of particle size and amino acid surface modifications. Surface modifications are known to significantly enhance micromotor performance, particularly in terms of ion tolerance, which is crucial for their functionality in various environments. The ion tolerance provided by other amino acid modifications, including Gly, Phe, and Glu, was also investigated. TiO₂-Gly, TiO₂-Phe, and TiO₂-Glu micromotors were prepared and placed in both water and PBS for propulsion investigation. While the propulsion speeds were suppressed in PBS compared to water, the micromotors were still able to move effectively in PBS (Fig. 5A and Movies S12–S14). TiO₂-Gly micromotors exhibited the highest speed, with an average velocity of 3.85 ± 0.75 μm/s. Comparatively, the speed of TiO₂-Phe and TiO₂-Glu were 3.07 ± 0.65 μm/s and 2.89 ± 1.12 μm/s, respectively. The propulsion speeds of TiO₂-Gly, TiO₂-Phe, and TiO₂-Glu micromotors were also measured at different ionic strengths. Fig. 5B and SI Appendix, Fig. S14 show the dependence of the inverted speed of these micromotors on the ionic strength of the solution. Based on these measurements, we calculated the EI₅₀ values of TiO₂-Gly, TiO₂-Phe, and TiO₂-Glu, which were found to be 20.24 ± 1.28 mM (Movie S15), 16.86 ± 1.24 mM (Fig. 5C and Movie S16), and 13.77 ± 0.288 mM (Movie S17), respectively. Notably, TiO₂-Gly micromotors exhibited the highest EI₅₀ of 20.24 ± 1.28 mM, while TiO₂-Arg micromotors showed the fastest motion speed in PBS. This may be attributed to the more negative surface potential of TiO₂-Arg (SI Appendix, Fig. S15), as confirmed by our zeta potential measurements, which could enhance electrostatic interactions. Additionally, photocatalytic efficiency, assessed via methylene blue degradation, indicated that TiO₂-Arg exhibited slightly superior photocatalytic performance compared to TiO₂-Gly, which might contribute to its enhanced propulsion efficiency to some extent (SI Appendix, Fig. S16). Moreover, the surface conductivity of TiO₂-Gly, TiO₂-Glu, and TiO₂-Phe micromotors were measured using the same impedance technique. The results shown in SI Appendix, Fig. S17 indicated that the surface conductivity of these micromotors was also considerably higher than that of bare TiO₂ micromotors.

Fig. 5.

Ionic strength–dependent behavior of amino acid–modified micromotors. (A) Speeds of TiO₂-Gly, TiO₂-Phe, and TiO₂-Glu micromotors in water and PBS. (B) Linear fit of the inverse speed versus ionic strength for TiO₂-Gly. (C) EI₅₀ values of TiO₂-Gly, TiO₂-Phe, and TiO₂-Glu. Speeds of 580 nm (D) and 1.72 µm (E) TiO₂-Arg micromotors at varying ionic strengths. (F) EI₅₀ values of 580 nm and 1.72 µm TiO₂ and TiO₂-Arg micromotors. (G) Linear fits of inverse speed versus ionic strength for ZnO-Arg. (H) Linear fits of inverse speed versus ionic strength for SiO₂@CeO₂-Arg. (I) EI₅₀ values of ZnO, ZnO-Arg, SiO₂@CeO₂ and SiO₂@CeO₂-Arg. All quantitative data (speed and EI₅₀) are based on n = 20 and are presented as either box plots (median ± interquartile range) or bar charts with error bars indicating mean ± SD, as specified in each panel. All experiments with speed were conducted under UV light with a wavelength of 365 nm with an intensity of 80 mW/cm².

We also investigated the dependence of ion tolerance on the size of the micromotors (30–33). The sol–gel method was used to fabricate TiO₂-Arg motors with various diameters (see Materials and Methods for details). As shown in SI Appendix, Fig. S18, TiO₂-Arg motors with 580 nm and 1.72 μm were successfully synthesized. We tested the motion speeds of TiO₂-Arg micromotors of these two sizes, both before and after Arg modification, under different ionic strengths (Fig. 5 C and D and SI Appendix, Fig. S19). Linear fitting of the inverse speed and ionic strength allowed us to calculate the EI₅₀ values, as shown in Fig. 5E. The EI₅₀ of the smaller TiO₂-Arg motor (580 nm) was enhanced to 29.11 ± 1.19 mM, which was 4.4 times higher than that of the larger ones with a diameter of 1.72 μm (EI₅₀ = 6.56 ± 0.365 mM) (Movies S18–S21). The above-mentioned results showed that the ion tolerance of self-electrophoretic motors was closely related to the particle size, and further improvement in ion tolerance could be achieved by decreasing the dimension of MNMs as predicted in Eq. 1. Reducing particle size can decrease the contribution to the solution conductivity (Λ_m) (20), thereby enhancing the ion tolerance. Because particles at micro/nanoscales possess a higher specific surface area, allowing for increased surface interactions with ions in the solution. These interactions lead to ion adsorption and exchange, consequently reducing the concentration of free ions in the solution. Reducing particle size can weaken the interaction between each particle and the ions, thereby decreasing its impact on conductivity.

Extending Arg Modification to Enhance Ion Tolerance of ZnO and CeO₂ Micromotors.

To further validate the generality of our approach beyond TiO₂, we extended its application to two additional photocatalytic materials: ZnO and CeO₂, with the latter grown on SiO₂ to form a core-shell structure (SiO₂@CeO₂). ZnO particles with a diameter of 1.82 μm (SI Appendix, Fig. S20) were synthesized using a solvothermal method, while SiO₂@CeO₂ particles with a diameter of 460 nm (SI Appendix, Fig. S21) were synthesized using the sol–gel method followed by solvothermal approach (details in Materials and Methods). We tested the propulsion speeds of these two micromotors before and after Arg modification under different ionic strengths. By performing linear fitting of the inverse speed and ionic strength (Fig. 5 G and H and SI Appendix, Figs. S22 and S23), we calculated the EI₅₀ values, which are shown in Fig. 5I. For unmodified ZnO and SiO₂@CeO₂, the EI₅₀ values did not exceed 0.1 mM, consistent with theoretical expectations. However, after Arg modification, the EI₅₀ value for ZnO-Arg increased from 0.043 ± 0.0042 mM to 0.53 ± 0.13 mM. For SiO₂@CeO₂-Arg motors, the EI₅₀ increased from 0.054 ± 0.0040 mM to 11.78 ± 0.32 mM (SI Appendix, Table S1). These findings highlight the effectiveness of amino acid modification in enhancing ionic conductance within the EDL, confirming that this strategy is not limited to TiO₂ but can be broadly applied to other photocatalytic systems.

The Application of TiO₂-Arg Micromotors with Ionic Tolerance.

The precise navigation and clustered movement of MNMs enable them to effectively gather at the site of the disease, providing assurance for active disease treatment. TiO₂, functioning as an acoustic sensitizer, can generate reactive oxygen species (ROS) to eradicate tumors when exposed to ultrasound (US) with low intensity. Therefore, our TiO₂-Arg motors with enhanced ion tolerance hold great potential for antitumor therapy. Regulated by the on/off state and the irradiated angle of the light source, TiO₂-Arg motors could target the tumor cells in an ionic solution (RPMI-1640 medium), and ROS were then produced under US stimulation. To assess the sonodynamic efficacy of TiO₂-Arg micromotors, 1,3-diphenylisobenzofuran (DPBF) (34) and Singlet Oxygen Sensor Green (SOSG) (35) probes were used to measure the US-activated ROS generation. As illustrated in Fig. 6A, the characteristic absorption peak (426 nm) of DPBF was substantially reduced during US treatment (1.0 MHz, 50% duty cycle, 1.05 W/cm²). For comparison, the absorption of the DPBF+TiO₂-Arg group (without US treatment) decreased slowly, as shown in SI Appendix, Fig. S24. Moreover, the absorption variation relative to the initial value was calculated by recording the absorbance at various intervals during the experiment. Based on SI Appendix, Fig. S25, the concentration of DPBF in the DPBF+US+TiO₂-Arg group decreased to 65% in 7 min, whereas for the DPBF+TiO₂-Arg group, the decrease was marginal, and the DPBF concentration remained above 97%. This indicates that our TiO₂-Arg could generate ROS under US, and the ROS production was also increased with increasing US stimulation time. SOSG, a highly selective reagent for singlet oxygen (¹O₂), was further used for the detection of ¹O₂. As depicted in Fig. 6B, the emission intensity of SOSG increased along with US irradiation time, which also indicated the ¹O₂ generation capability of TiO₂-Arg.

Fig. 6.

Sonodynamic antitumor effect of TiO₂-Arg micromotors. (A) DPBF UV-vis absorption decay curves after addition of TiO₂-Arg aqueous solution followed by US irradiation. (B) SOSG fluorescence intensity after adding TiO₂-Arg aqueous solution with US irradiation. (C) 4T1 cells viability after treating with TiO₂-Arg based on the CCK-8 assay. (D) AM and PI staining of the cells after different treatments. (Scale bar, 50 μm). (E) ROS generation probed by DCFH-DA. (Scale bar, 50 μm). (F) Z-stack confocal laser scanning microscopy images of in vitro 3D tumor spheroids treated with Rhodamine B–labeled TiO₂ (a) or TiO₂-Arg (b) to evaluate drug penetration. Images at different depths from the surfaces of the spheroids are shown. (Scale bar, 100 μm.) (G) AM and PI staining of the 4T1/NIH3T3 tumor sphere after different treatments. (Scale bar, 50 μm). (H) L02 Cell viability after treatment with TiO₂-Arg micromotors with different concentrations for 24 h. (I) Hemolysis evaluation of TiO₂-Arg micromotors. (J) The effect of UV and catalytic activity of TiO₂-Arg micromotors on cellular viability. All quantitative data are shown as box plots representing the median and interquartile range, with individual data points displayed (n = 3).

Owing to the robust sonodynamic responsiveness of TiO₂-Arg, the sonodynamic antitumor effect of TiO₂-Arg micromotors was then investigated. As shown in Fig. 6C, without US treatment, the viability of 4T1 cells was still 97% even when the concentration of TiO₂-Arg was increased to 175 µg/mL, indicating the relative biosafety of the TiO₂-Arg nanomotor itself. When subjecting cells to US stimulation at a power density of 1.05 W/cm², there was a pronounced, concentration-dependent decrease in cell survival rate, and the 4T1 cell viability rate was only 20% with a concentration of 175 µg/mL. Moreover, live/dead (AM and PI) staining was also performed, visually showing the therapeutic effects by green fluorescence (live cells) and red fluorescence (dead cells). As shown in Fig. 6D, only green fluorescence signal was detected in the control group, and both green and red fluorescence signals were observed in TiO₂-Arg, US, and US+TiO₂-Arg groups. The US+TiO₂-Arg group exhibited a much stronger red signal than the TiO₂-Arg and US group, indicating that TiO₂-Arg micromotors combined with US stimulation had the strongest antitumor effect. The AM/PI cellular fluorescence result was fully consistent with the CCK-8 assay, further confirming the effect of sonodynamic therapy. To demonstrate the ROS generation in 4T1 cells, 2,7-dichloro-dihydro-fluorescein diacetate (DCFH-DA) was further used (34). Fig. 5E presented that 4T1 cells in control and TiO₂-Arg groups did not exhibit significant ROS fluorescence signal, demonstrating that the TiO₂-Arg nanomotor itself did not obviously generate ROS. In contrast, a bright ROS fluorescence signal was clearly observed when combined with US treatment, which indicated the highest ROS production.

Next, we used 4T1 cells and NIH/3T3 fibroblasts to construct three-dimensional (3D) multicellular tumor spheroids to analyze the penetration ability and antitumor efficacy of active TiO₂-Arg. To visualize the penetration of motors within tumor spheroids, we labeled TiO₂ and TiO₂-Arg micromotors with fluorescent dye by electrostatic adsorption of Rhodamine B. First, Rhodamine B–labeled TiO₂-Arg micromotors were guided to the tumor spheroids using UV light, followed by a coincubation period of 6 h. For the TiO₂ micromotor group, we followed the same steps, but their poor ion tolerance led to minimal movement, resulting in only a few TiO₂ particles aggregating at the tumor spheroids. As shown in Fig. 6F, the Rhodamine B fluorescent signals of the TiO₂-Arg nanomotor were not only distributed at the edge of the spheroids but also diffused deep inside the tumor, whereas most of Rhodamine B in the TiO₂ group was retained at the surface layer of the tumor spheroids. As shown in SI Appendix, Fig. S26, the TiO₂-Arg group showed significantly higher Rhodamine B content than the TiO₂ group. This is attributed to the effective motility of the nanomotor in the culture medium, facilitating its penetration into the interior of the tumor spheroids. Then, live-dead staining was employed to assess the antitumor properties of the micromotors at a 3D level. As shown in Fig. 6G, the US+TiO₂-Arg group exhibited a much stronger red signal than other groups. Consistent with the results of two-dimensional cell experiments, the TiO₂-Arg nanomotor showed significant tumor inhibition efficacy on 3D tumor spheroids as well.

Evaluation of Biocompatibility of Nanomotors.

Biocompatibility is a major challenge for artificial MNMs, which should be addressed before effective biomedical applications can be realized (20). Consequently, the cytotoxicity of TiO₂-Arg micromotors was evaluated using a normal liver cell line, L02. Fig. 6H demonstrated that cell viability remained unchanged after incubating with different concentrations of TiO₂-Arg micromotors (0 to 200 μg/mL) for 24 h. To further assess the safety of our system, hemolysis experiments (36) were also carried out to evaluate the biocompatibility of Arg-TiO₂. The hemolysis rate for the positive control (water) was 100%, whereas the rate of TiO₂-Arg groups with different concentrations was almost less than 1% (Fig. 6I), indicating that our TiO₂-Arg was indeed biocompatible. In practical antitumor applications, cell exposure to UV irradiation is inevitable when guiding TiO₂-Arg micromotors toward tumor cells. Therefore, the impact of UV phototoxicity and catalytic activity of TiO₂-Arg micromotors on cell viability was evaluated. As shown in Fig. 6J, illumination at 365 nm (80 mW/cm²) with TiO₂-Arg micromotors (175 μg/mL) on 4T1 cells with varying durations (0 to 30 min) did not lead to a significant variation in cell viability during 24 h incubation period. Given the efficient propulsion of TiO₂-Arg micromotors in ionic solutions, their potential applications in cells and tissues appear to be promising.

Discussion

This study introduces a simple and effective method to significantly enhance the ion tolerance of light-driven TiO₂ micromotors through surface modification with Arg. Compared to unmodified TiO₂ micromotors, the Arg-modified versions maintain stable motion in high-ionic-strength environments such as PBS and RPMI medium, with a more than 200-fold increase in ion tolerance. This development offers a solution to the challenges faced by electrophoretic micromotors in complex physiological conditions.

The improvement in ion tolerance is primarily attributed to the increased surface conductivity provided by Arg modification. This enhancement prevents the collapse of the Debye layer, which is critical for maintaining electrokinetic propulsion in high ionic environments. By preserving the integrity of the Debye layer, Arg-modified micromotors can continue to operate effectively in high ionic conditions, demonstrating significantly improved ion tolerance. Although amino acid modification enhances ion tolerance and biomedical performance, it is important to note that amino acids undergo degradation upon UV exposure. This degradation may affect their surface charge and electrophoretic performance under long-term or intense light conditions. Nevertheless, the enhancement from Arg modification remains effective after UV irradiation, as shown by the continued motion of the micromotors in PBS. Similar effects were also observed for ZnO and CeO₂ micromotors, where Arg modification enabled stable propulsion under high-ionic-strength conditions, further demonstrating the versatility of this approach.

In addition to improved ion tolerance, Arg-modified TiO₂ micromotors also show great potential in biomedical applications, particularly in sonodynamic therapy. Under US stimulation, these micromotors generate ROS, effectively inducing apoptosis in tumor cells. The ability of these micromotors to penetrate 3D tumor spheroids further underscores their potential for targeted cancer therapies. Compared to unmodified TiO₂, Arg-modified micromotors exhibit superior movement speed and penetration capability, enhancing their therapeutic effects.

While other amino acid modifications, such as glycine, glutamic acid, and phenylalanine, also improved the ion tolerance of TiO₂ micromotors, Arg provided the most significant advantages in terms of speed and electrophoretic performance. This indicates that Arg not only enhances ion tolerance but also optimizes the overall functionality of the micromotors, making them more efficient in active propulsion and targeted therapeutic delivery. Furthermore, the size of the micromotors was found to significantly influence their ion tolerance, with smaller micromotors (580 nm) showing a higher tolerance than larger ones (1.72 µm). This size-dependent effect underscores the importance of geometric optimization in micromotor design. Mechanistically, smaller micromotors possess a higher surface-area-to-volume ratio, which enhances the contribution of surface conduction relative to bulk conduction. Additionally, when the particle size approaches the Debye length, the EDL becomes more prominent in electrokinetic interactions, facilitating more efficient electroosmotic slip. These factors collectively help preserve local electric field gradients essential for sustained propulsion. Our observations are consistent with prior reports on size-dependent phoretic behavior in polyelectrolyte-coated systems and align with classical electrokinetic theory.

In conclusion, Arg surface modification offers a simple yet highly effective solution to the long-standing challenge of poor ion tolerance in electrophoretic micromotors. This advancement opens possibilities for the use of micromotors in biomedical applications, particularly in drug delivery and cancer treatment, where precise, active, and targeted motion in complex physiological environments is essential. Future research can focus on further optimizing micromotor size and modification to expand their applicability.

Materials and Methods

Materials.

Tetrabutyl titanate (TBOT), titanium isopropoxide, NaCl, ethanol, methanol, ethylene glycol, L-arginine (Arg), and isopropyl alcohol were purchased from Aladdin (Shanghai, China). PBS was purchased from Sigma. HCl, H₂SO₄, H₂O₂, and acetone were bought from Guangzhou Chemical Reagent Factory. Formic acid was obtained from Macklin (Shanghai, China).

Fabrication of TiO₂-Based MNMs.

TiO₂ particles of various sizes (∼580 nm, 1 μm, and 1.72 μm) (22, 37, 38) were synthesized via modified sol–gel methods using TBOT in ethanol or ethylene glycol under different ionic conditions and aging times. Surface modification with L-arginine (TiO₂-Arg) was performed via electrostatic adsorption by incubating TiO₂ in Arg-containing PBS (pH 7.4) for 48 h, followed by washing and drying. Rhodamine B–labeled TiO₂ and TiO₂-Arg were obtained by mixing with dye solution (1 mg/mL) and stirring in the dark for 48 h.

Fabrication of ZnO and ZnO-Arg.

ZnO microspheres were prepared through a solvothermal method by reacting Zn(NO₃)₂·6H₂O with polyvinylpyrrolidone in ethylene glycol, followed by calcination at 500 °C. ZnO-Arg was obtained by incubating ZnO with Arg in Tris-HCl buffer under stirring for 48 h (39–41).

Fabrication of SiO₂@CeO and SiO₂@CeO-Arg.

SiO₂ spheres were first synthesized using the Stöber method. A CeO₂ shell was formed by hydrothermal deposition of Ce(NO₃)₃ and urea on SiO₂ templates at 160 °C, producing SiO₂@CeO₂ particles. Surface modification with Arg was achieved by dispersing the particles in aqueous Arg solution and stirring for 48 h (42).

Supplementary Material

Appendix 01 (PDF)

Movie S1.

Brownian motion of TiO₂-Arg micromotors in H₂O.

Movie S2.

Motion of TiO₂-Arg micromotors in H₂O with 80 mW/cm₂ of light.

Movie S3.

Motion TiO₂ micromotors in H₂O with 80 mW/cm₂ of light.

Movie S4.

Motion of TiO₂-Arg micromotors in PBS with different intensity of light.

Movie S5.

Motion of TiO₂ micromotors in PBS with 200 mW/cm₂ of light.

Movie S6.

TiO₂-Arg micromotors targeting the 4T1 cell with 80 mW/cm₂ of light.

Movie S7.

TiO₂-Arg micromotors targeting the RBC with 80 mW/cm₂ of light.

Movie S8.

TiO₂-Arg micromotors clustering behavior in H₂O with 80 mW/cm₂ of light.

Movie S9.

TiO₂-Arg micromotors clustering behavior in PBS with 80 mW/cm₂ of light.

Movie S10.

Motion of TiO₂ micromotors under different ionic strengths with 80 mW/cm₂ of light.

Movie S11.

Motion of TiO₂-Arg micromotors under different ionic strengths with 80 mW/cm₂ of light.

Movie S12.

Motion of TiO₂-Gly micromotors in PBS with 80 mW/cm₂ of light.

Movie S13.

Motion of TiO₂-Phe micromotors in PBS with 80 mW/cm₂ of light.

Movie S14.

Motion of TiO₂-Glu micromotors in PBS with 80 mW/cm₂ of light.

Movie S15.

Motion of TiO₂-Gly micromotors under different ionic strengths with 80 mW/cm₂ of light.

Movie S16.

Motion of TiO₂-Phe micromotors under different ionic strengths with 80 mW/cm₂ of light.

Movie S17.

Motion of TiO₂-Glu micromotors under different ionic strengths with 80 mW/cm₂ of light.

Movie S18.

Motion of TiO₂ (0.58 μm) micromotors under different ionic strengths with 80 mW/cm₂ of light.

Movie S19.

Motion of TiO₂-Arg (0.58 μm) micromotors under different ionic strengths with 80 mW/cm₂ of light.

Movie S20.

Motion of TiO₂ (1.72 μm) micromotors under different ionic strengths with 80 mW/cm₂ of light.

Movie S21.

Motion of TiO₂-Arg (1.72 μm) micromotors under different ionic strengths with 80 mW/cm₂ of light.

Data, Materials, and Software Availability

All study data are included in the article and/or supporting information.