Tyrosine monomer nanocrystal as a potent ice recrystallization inhibition activity

Abstract

Ice growth inhibition is crucial in cryotechnology, as uncontrolled recrystallization during the frozen state and freeze-thaw cycles causes irreversible damage to biological samples. Nanoscale materials that mimic antifreeze proteins and exhibit ice recrystallization inhibition (IRI) activity have been explored as cryoprotectants; however, the structural features that govern potent IRI activity remain unclear. This study investigated the effects of nanoparticle size and functionality on the IRI activity. Polystyrene nanoparticles (PSNPs, 30–1000 nm) were used as inert nanoscale models, and amino acid derivatives with phenyl groups with or without hydroxyl functionality, including L-phenylalanine monomers, pentamers of L-phenylalanine (Phe-5), L-tyrosine, and 3,4-dihydroxy-L-phenylalanine, were examined. Among these, we found that tyrosine monomer nanocrystals (TMNs) display exceptionally potent IRI activity under both extracellular and intracellular conditions, which is attributed to nanoscale structure formation, hydroxyl functionality, and high colloidal stability. TMNs enhance cell survival during cryopreservation, even at low dimethyl sulfoxide concentrations, whereas Phe-5 and other analogs show limited activity owing to aggregation or lack of hydroxyl groups. These results elucidate the key factors influencing IRI activity, including nanoscale assembly with high colloidal stability and the presence of a hydroxy functional group. Therefore, considering the biocompatibility of L-tyrosine, our study shows that TMNs are promising supplementary materials for cryobiology.

---

1. Introduction

Cryopreservation is a fundamental technology for the long-term storage of various biosamples including protein-based therapeutics, cells, tissues, and transplantable organs. It plays a critical role in cryobiology, biotechnology, and regenerative medicine.^1,2 This process involves storing biological samples at ultra-low temperatures, commonly at −80 °C or −196 °C, to preserve them for extended periods. However, during freezing and freeze-thaw cycles, ice formation and growth can cause irreversible damage by physically disrupting cell membranes and tissue structures.^3,4 Ice formation can start from a small ice nucleus, and ice growth mainly proceeds through Ostwald ripening,⁵ which is a thermodynamically driven process wherein smaller ice crystals dissolve and redeposit onto larger ice crystals to minimize the overall surface energy. The growth of large ice crystals can physically damage biological samples, thereby making the inhibition of ice growth and recrystallization major challenges in improving cryopreservation efficiency.

Several approaches have been proposed to control ice crystal growth, and these include vitrification, modulation of cooling and heating rates, and the development of antifreeze compounds.¹ Notably, substance-based approaches show significant potential as integrating natural materials – such as antifreeze proteins (AFPs) – as well as biocompatible small molecules or polymers, can prevent irreversible injury without changing the standard cryopreservation procedures.⁶ AFPs or antifreeze glycoproteins (AFGPs), which are known for their strong ice recrystallization inhibition (IRI) activity, have been suggested for this purpose. However, their use is limited by challenges such as difficulties in large-scale production, high extraction and purification costs,⁷ and cytotoxicity to mammalian cells in some cases, potentially leading to immunological issues.⁸

To overcome these limitations, researchers have focused on developing alternative materials that inhibit ice growth and exhibit IRI activity, including AFP-mimicking molecules,⁹ synthetic polymers such as poly(vinyl alcohol) (PVA),¹⁰ two-dimensional materials like graphene oxides,¹¹ and hydrogels.¹² Hydroxyl groups are important functional groups because they serve dual roles: enabling hydrogen bonding with ice surfaces or other molecules,^13,14 while also modulating molecular self-assembly.¹⁵ Recent studies have explored the use of amino acids, small molecules, and nanoscale materials, while highlighting that factors such as structure formation and surface functional groups play key roles in determining IRI activity.^16,17 A natural amino acid monomer (i.e., L-phenylalanine) has been shown to suppress ice crystal growth. Amphiphilicity was reported to be a critical structural feature of IRI activity, and para-amino substituents experimentally achieved decreased IRI activity due to decreased amphiphilicity.¹⁸ Additionally, monomers without intrinsic IRI activity can exhibit IRI effects when forming self-assembled structures.^19,20 Thus, the size of the self-assembled structures and the functional groups on their surfaces are important factors that influence IRI activity.^21,22 Despite these findings, there is limited insight into the dependence of IRI activity on nanoparticle size or functionality.

In this study, we systematically investigated the effect of polystyrene nanoparticle (PSNP) size on IRI activity. PSNPs have a highly hydrophobic surface because of the abundance of phenyl rings in their structure. Styrene has the same benzyl motif as L-phenylalanine, and it is possible to synthesize PSNPs with precise size control ranging from the nanoscale to the microscale. Several studies have reported that when molecules form nanoscale assemblies, IRI activity can be enhanced or newly observed.^19,20 For this reason, it serves as a useful model system for elucidating the relationship between IRI activity and structure size. For comparison, we also selected the L-phenylalanine monomer (Phe-1) and pentamer of L-phenylalanine (Phe-5). Although Phe-1 showed IRI activity as per a previously published paper,¹⁸ the origin of this IRI activity is not clear. Warren et al. only mentioned the possibility of a nanoscale structure in the solution with Phe-1. Phe-5 was selected because it can be used to produce nanoscale structures. We examined the possibility of nanoscale structure formation in a solution of Phe-1 or Phe-5 and compared the IRI activities of Phe-1 and Phe-5. We further investigated the impact of a hydroxy group in the phenyl group on the IRI activity using an L-tyrosine monomer, with 3,4-dihydroxy-L-phenylalanine (DOPA-1) as a potent cryoprotectant candidate for molecular or nanoscale materials (Fig. 1a and Fig. S1). IRI activity was quantitatively evaluated using a splat-cooling assay (Fig. 1b).

Fig. 1 (a) Concept for studying the effect of styrene monomers, nanosized polystyrene particles, and nano- or micro-sized materials with phenyl groups and varying numbers of hydroxyl groups on ice recrystallization inhibition (IRI) activity. (b) Experimental setup of the splat-cooling method to measure IRI activity.

2. Experimental

Sample preparations

Amicon® Ultra Filters were used for solvent exchange of the PSNP standard solution. PSNPs (30, 50, 60, and 100 nm) were centrifuged three times (9300 rcf, 30 min) and dispersed in a saline solution ([NaCl] = 50 mM). The 200 and 1000 nm PSNPs were centrifuged at 15 000 rcf for 30 min and 2300 rcf for 15 min (using an Eppendorf tube), respectively. The supernatant was removed and dispersed in saline; this process was repeated three times. The particle size and zeta potential were measured using a transmission electron microscope (H-7100; Hitachi, Tokyo, Japan) and a particle size analyzer (ELSZ-1000; Otsuka Electronics, Osaka, Japan).

Phe-5 (1.0 mg) was dissolved in HFIP (500 µL) for 30 min on a shaker and lyophilized using a freeze dryer (IlShinBioBase, South Korea). The freeze-dried powder was dispersed in distilled water (DW) (500 µL, sonicated for 1 min) and incubated at room temperature (25 °C) for 1 h to allow self-assembly, which was followed by solvent exchange with saline.

The self-assembled Phe-5 structures were disrupted using high-intensity ultrasound (VC505, Sonics and Materials Inc., Newton, CT, USA). Samples dispersed in DW (2.0 mg mL⁻¹) were sonicated at 35% amplitude for 60 min in the pulse mode (5 s on/5 s off), and the conical tubes were kept in an ice-water bath to prevent excessive heating.

Characterization of particles and nanocrystals

The morphology and size were examined using a transmission electron microscope (H-7100; Hitachi, Tokyo, Japan). Hydrodynamic size distribution and zeta potential were analyzed using a particle size and zeta potential analyzer (ELSZ-1000, Otsuka Electronics, Osaka, Japan). Powder X-ray diffraction (XRD) patterns were recorded using a SmartLab diffractometer (Rigaku Corporation, Tokyo, Japan), with CuKα radiation (λ = 1.5406 Å) at 40 kV and 40 mA over the 2θ range of 10–70°. Data were collected in the continuous scan mode (step size 0.02° and scan rate 2° min⁻¹) and processed using the PDXL software (Rigaku Corporation, Tokyo, Japan).

Measurement of ice recrystallization inhibition (IRI) activity

IRI activity and dynamic ice shaping (DIS) were measured using a nanoliter osmometer (Otago Osmometers, Dunedin, New Zealand) equipped with an optical microscope (BX 51, Olympus, Japan). The plate cooling method was used to assess IRI activity. Styrene, PSNPs (30, 50, 60, 100, 200, and 1,000 nm), Phe-5, and the amino acids (Phe-1, TMN, and DOPA-1) were dissolved in saline ([NaCl] = 50 mM) at various concentrations. Each sample (10 µL) was dropped onto circular cover glasses and cooled to −78 °C using liquid nitrogen. The ice grain size was monitored using a microscope equipped with a polarizer on the cryo-stage (Peltier device) while maintaining the temperature of the circular cover glass surface at −6 °C for 30 min. The 50 largest ice grains were selected and their areas were quantified using ImageJ software to obtain average values.

Monitoring the dynamic ice shaping (DIS) properties

A sample solution (3.0 µL) was injected on both sides of the stage hole (Φ 2.0 mm) and frozen at −78 °C using liquid nitrogen. The sample was then transferred to a cryo-stage (Peltier device) maintained at −6 °C and slowly heated (0.3 °C min⁻¹) until a few ice grains remained. The shape of a single ice grain was monitored by cooling at a rate of 0.01 °C s⁻¹.

Observing intracellular ice-grains using fluorescence cryo-microscopy

To examine the formation of intracellular ice grains, human oral squamous cell carcinoma (HSC-3) cells were sandwiched between two cover glasses with fresh medium and mounted on a cold stage (THMS600; Linkam, UK). The samples were cooled at 100 °C min⁻¹, held at −196 °C for 2 h, and then heated to −6 °C at 100 °C min⁻¹ to induce intracellular ice recrystallization. They were annealed at −6 °C for 30 min before thawing.

Cell culture

HSC-3 and RAW 264.7 (mouse macrophages) were cultured in DMEM supplemented with 10% v/v FBS and 1% v/v antibiotic in a humidified 37 °C incubator with 5% CO₂. At 90% confluence, the cells were detached using trypsin-EDTA (HSC-3) or a cell scraper (RAW 264.7) and subcultured at a 1 : 3 ratio in T75 flasks.

Cell viability

Cell viability was assessed using the 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) cell proliferation assay, which uses MTS as the main reagent. HSC-3 cells were seeded into 96-well plates (10 000 cells per well) and incubated in a humidified 37 °C, 5% CO₂ atmosphere. After 24 h, the medium was replaced with phenol red-free DMEM containing either the amino acids (Phe-1 and TMN) or the amino acid derivative (DOPA-1). The cells were incubated for 12 h and washed, and cell viability was measured at 490 nm using a plate reader (BioTek Inc., Winooski, USA).

Cryopreservation

To evaluate the effectiveness of cryoprotectants, cells were incubated with 0.45 mg mL⁻¹ of TMN or Phe-1 for 30 min for cellular uptake. HSC-3 or RAW264.7 cells were detached from T75 flasks using trypsin-EDTA or a cell scraper and collected in 15 mL conical tubes. After centrifugation, the supernatant was removed and replaced with a fresh medium. A mixture of dimethyl sulfoxide (DMSO) (5.0 µL) and TMN or Phe-1 (495 µL, 0.45 mg mL⁻¹ in fresh medium without FBS) was added to suspensions containing 50 × 10⁴ cells. The samples were immediately transferred to liquid nitrogen for rapid cryopreservation and stored for one week. After storage, the cells were thawed in a 37 °C water bath for 3 min, and cell viability was assessed using the MTS assay.

Statistical analysis

To calculate the mean largest ice grain size, the ten largest ice grains were selected per experiment, and the procedure was independently repeated five times (N = 50). Error bars represent the standard deviation.

3. Results and discussion

IRI activity of polystyrene nanoparticles

The PSNP standards were characterized by measuring the particle size and zeta potential. As shown in Fig. S2, PSNPs exhibited a spherical morphology with average diameters of 31 ± 3 nm (30 nm PSNPs), 51 ± 3 nm (50 nm PSNPs), 61 ± 4 nm (60 nm PSNPs), 100 ± 6 nm (100 nm PSNPs), 203 ± 5 nm (200 nm PSNPs), and 994 ± 15 nm (1000 nm PSNPs). After washing twice with DW to remove surfactants and redispersing in saline, the zeta potentials ranged from −22.8 ± 1.1 mV to −43.1 ± 1.9 mV.

The styrene monomer (M_w 104.15, ≥99%; 1.0 mg mL⁻¹) showed no IRI activity in either DW or 50 mM saline. In DW, the ice grain size (23 060 ± 1021 µm²) was comparable to that of DW alone (22 712 ± 1415 µm²). Similarly, in saline, the ice grain size (23 008 ± 1722 µm²) was nearly identical to that of the saline solution alone (23 801 ± 1176 µm²). These results indicated that the styrene monomer was IRI-inactive at 1.0 mg mL⁻¹, regardless of the solvent used (Fig. S3).

A previous study found that a polymerized material (spherical shape, 400 nm) with an IRI-inactive monomer, even without ice-binding domains (e.g., –OH groups), exhibited enhanced IRI activity.¹⁹ To assess whether nanoscale polymerized styrene exhibited enhanced IRI activity, we measured the ice-grain size of PSNPs (30–1000 nm) in DW (Fig. 2). At 10 µg mL⁻¹ in DW, 30 nm PSNPs showed the mean largest ice grain size (MLGS) at 26.2% of the control, thereby indicating possible IRI activity (Fig. 2a). At 100 µg mL⁻¹, smaller PSNPs consistently produced smaller ice grains. However, 1000 nm PSNPs showed no decrease in ice grain size at any tested concentration. This result is likely attributable to the substantially lower number of particles present at an equivalent mass concentration, due to the significantly higher mass per particle; for example, a 1.0 mg mL⁻¹ suspension of 30 nm PSNPs contains approximately 3.7 × 10⁴ times more particles than a suspension of 1000 nm PSNPs (SI).

Fig. 2 Effect of the PSNP size (30, 50, 60, 100, 200, and 1000 nm) on IRI activity. Measured mean largest ice grain size and photographs of PSNPs (a) in DW and (b) in saline ([NaCl] = 50 mM).

However, measuring the ice grain size under low-ionic-strength conditions (e.g., DW) can lead to the misinterpretation of IRI activity in real biological systems, as previously shown for PEG (M_w 3500–4500).²³

Therefore, we examined PEG and PVA (M_w 9000–10 000) as negative and positive controls, respectively, in DW and 50 mM saline. In DW, PEG (1.0 mg mL⁻¹) exhibited an MLGS of 46.9%, thereby suggesting a potential IRI activity, and PVA also demonstrated potential activity at only 10 µg mL⁻¹ (MLGS 17.7%) (Fig. S4a). However, PEG exhibited no IRI activity in 50 mM saline, whereas PVA remained highly active (MLGS 10.9% at 1.0 mg mL⁻¹), which is consistent with previously reported results (Fig. S4b). These results indicate that 50 mM saline is a suitable solvent for assessing the IRI activity of materials. Therefore, we measured the ice grain size in the presence of PSNPs (30–1000 nm) in 50 mM saline. However, PSNPs of all sizes showed no decrease in the MLGS values, which were the same as those of the saline control (Fig. 2b), thus demonstrating the absence of significant IRI activity in the PSNPs. These results indicated that the presence of nanoscale materials without functionality was insufficient to confer IRI activity.

IRI activity of Phe-1 and Phe-5

Phe-1 has been reported to show considerable IRI activity at 20 mM in 10 mM saline due to its amphiphilic structure.¹⁸ Warren et al. suggested that the nanoscale structure of Phe-1 may contribute to this high IRI activity. According to a previous study, an alanine pentamer formed nanoscale structures that enhanced IRI activity compared with the corresponding amino acid monomers.²⁰ Therefore, we examined whether Phe-1 and the phenylalanine pentamer (Phe-5) could form self-assembled structures and evaluated their IRI activity (Fig. 3a). Phe-1 and Phe-5 (1.0 mg mL⁻¹) were dissolved in HFIP, respectively, lyophilized, redispersed in DW, and incubated at room temperature for one day. However, no specific self-assembled structures could be observed from Phe-1 at 1.0 and 5.0 mg mL⁻¹ (Fig. 3a-i, ii), whereas Phe-5 showed wire-like self-assemblies with widths of 30–50 nm for thin ones and >100 nm for thicker ones (Fig. 3a-iii). These assemblies are likely driven by the extensive π–π stacking among the aromatic groups of Phe-5.²⁴ As shown in Fig. 3a iv, extensive wire-like structures were frequently aggregated into bundled forms due to the strong hydrophobicity of Phe-5.²⁵ To reduce the size of these large aggregates, the Phe-5 assembled solution (2.0 mg mL⁻¹) was probe-sonicated for 60 min to produce fragmented nanostructures (<500 nm). However, upon replacing DW with 50 mM saline, rapid aggregation was observed within minutes and visibly altered the solution, with the whole process being confirmed by photographs and TEM images (Fig. 3b).

Fig. 3 (a) TEM images of (i) Phe-1 (1.0 mg mL⁻¹), (ii) Phe-1 (5.0 mg mL⁻¹), (iii) Phe-5 (1.0 mg mL⁻¹), and (iv) Phe-5 (2.0 mg mL⁻¹). (b) TEM images of phenylalanine pentamer (Phe-5) solution (2.0 mg mL⁻¹) before and after probe sonication and solvent change. (c) Mean largest ice grain size of the Phe-1 and self-assembled Phe-5 structure in saline ([NaCl] = 50 mM, with or without sonication). (d) Photographs of ice grains in saline: (i) Phe-1 (1.0 mg mL⁻¹), (ii) Phe-1 (5.0 mg mL⁻¹), (iii) Phe-5 (0.1 mg mL⁻¹, no sonication), and (iv) Phe-5 (2.0 mg mL⁻¹, no sonication) solution. (v) Phe-5 (0.1 mg mL⁻¹, sonicated for 60 min) and (vi) Phe-5 (2.0 mg mL⁻¹, sonicated for 60 min).

Next, we measured the MLGS values for Phe-1, Phe-5, and sonicated Phe-5 to evaluate the effect of self-assembly on IRI activity. For Phe-1, increasing the concentration from 2.0 mg mL⁻¹ to 5.0 mg mL⁻¹ decreased the MLGS from 64.2% to 48.7%, indicating concentration-dependent enhancement of IRI activity even in the absence of nanoscale structures in the solution.¹⁹ In contrast, the self-assembled Phe-5 solution (2.0 mg mL⁻¹) showed an MLGS of 69.1%, which was similar to that of Phe-1. Sonicated Phe-5 showed slightly improved IRI activity at low concentration (0.1 mg mL⁻¹); however, no significant improvement in IRI activity was observed at higher concentration (2.0 mg mL⁻¹), as shown in Fig. 3c and d. This is due to the poor colloidal stability of Phe-5 in saline, which can decrease the effective number of available assemblies and the concentration of dispersed structures.^25,26 Additionally, the same number of carboxyl and amine groups of Phe-1 and Phe-5 can be another reason for the almost same IRI activity.²⁷

IRI activity for the number of hydroxyl groups in the phenyl group

Hydroxyl groups are known to enhance the IRI activity by enabling hydrogen bonding, promoting lattice matching with ice surfaces, and improving solubility by increasing polarity.¹⁴ To evaluate the effect of the number of hydroxyl groups in the phenyl group, we compared Phe-1 (non-hydroxylated), L-tyrosine (mono-hydroxylated), and DOPA-1 (di-hydroxylated) (Fig. 4) and first examined whether these amino acids could form self-assembled structures using TEM and dynamic light scattering (DLS) analysis (Fig. 4a–f). All powdered samples were dissolved in HFIP, lyophilized, and dispersed in DW. TEM revealed observable structures at low (0.1 mg mL⁻¹) and high (5.0 mg mL⁻¹) concentrations of Phe-1, L-tyrosine, and DOPA-1 (Fig. 4a–c). However, DLS analysis revealed no detectable nanoscale structures in the solution states of Phe-1 and DOPA-1 (Fig. 4d and f). For the TEM analysis, the samples were dried to form structures on the TEM grid. In contrast, the L-tyrosine solution showed a detectable intensity in the DLS analysis (Fig. 4e-ii). Therefore, it can be concluded that Phe-1 and DOPA-1 cannot produce nanostructures in DW, while L-tyrosine can produce nanoscale structures. It is established that L-tyrosine can be crystallized in the orthorhombic system (space group P2₁2₁2₁).²⁸ As shown in Fig. S5, the XRD patterns of the obtained L-tyrosine matched those of commercially purchased L-tyrosine powder, thereby confirming that there was no difference in the crystalline structure. We named the obtained nanostructure of L-tyrosine as an L-tyrosine monomer nanocrystal (TMN). The orthorhombic lattice supports a three-dimensional hydrogen-bonding network, and π–π stacking between aromatic rings further stabilizes assemblies.^29,30 The zeta potentials of TMNs and commercial L-tyrosine monomers were determined to be −14.71 mV and −19.96 mV, respectively. This indicates that hydroxyl (−OH) or carboxyl (−COOH) groups dominate the surface of nanocrystals. TEM and DLS analyses clearly showed the presence of nanoscale aggregates and micrometer-scale fibrous structures in TMNs and commercial L-tyrosine monomers (Fig. S6).

Fig. 4 Self-assembled nanostructures and IRI activity of Phe-1, TMNs, and DOPA-1. TEM images of (a) Phe-1 ((i): 0.1 mg mL⁻¹ and (ii): 5.0 mg mL⁻¹), (b) TMNs ((i) 0.1 mg mL⁻¹ and (ii) 5.0 mg mL⁻¹), and (c) DOPA-1 ((i) 0.1 mg mL⁻¹ and (ii) 5.0 mg mL⁻¹). Size distribution of (d) Phe-1 ((i) 0.1 mg mL⁻¹ and (ii) 5.0 mg mL⁻¹), (e) TMN ((i) 0.1 mg mL⁻¹ and (ii) 5.0 mg mL⁻¹), and (f) DOPA-1 ((i) 0.1 mg mL⁻¹ and (ii) 5.0 mg mL⁻¹). (g) Mean largest ice grain size (0.1–5.0 mg mL⁻¹) and photographs (0.1 and 5.0 mg mL⁻¹) of Phe-1, TMNs, and DOPA-1 in the saline ([NaCl] = 50 mM).

Fig. 4g shows the IRI activity of Phe-1, TMN, and DOPA-1 solutions. Phe-1 showed an IRI activity at concentrations of 3.0 and 5.0 mg mL⁻¹, while concentrations below 1.0 mg mL⁻¹ yielded ice grain sizes comparable to the saline control. In contrast, TMNs exhibited very strong IRI activity even at 0.1 mg mL⁻¹, thereby reducing the MLGS by 25.1% (approximately 75% ice grain size reduction). This potency suggests that hydrogen bonding provided by the hydroxyl group, together with the stability of the self-assembled structures, is central to suppressing ice recrystallization. The influence of TMN nanostructures on IRI activity was further assessed.^17,20 The TMN solutions were filtered through a 200 nm membrane to remove large aggregates. The L-tyrosine solubility in DW was 0.45 mg mL⁻¹; therefore, TMNs were prepared at 1.0 mg mL⁻¹, allowed to dissolve fully, and then filtered using a 200 nm pore membrane filter. At 1.0 mg mL⁻¹, TMNs exhibited comparable IRI activity before and after filtration (MLGS: 16.9% vs. 19.8%; Fig. S7a and c). Although the microscale structures were removed (Fig. S7b), the TMN solution showed significant IRI activity in the presence of nanostructures, which indicated the dominant role of the nanoscale structure in IRI activity.

DOPA-1 exhibited no detectable IRI activity within the tested range (0.1–5.0 mg mL⁻¹). Although its two hydroxyl groups should promote hydrogen bonding at the ice interface, intermolecular hydrogen bonding and autoxidation of DOPA-1 to dopaquinone may further suppress the IRI activity.³¹

Ions such as Na⁺, Cl⁻, and PO₄³⁻ in PBS or under other high-salinity conditions can disrupt interactions between IRI-active molecules and ice crystals.^17,25,26 Thus, the IRI activities of amino acid solutions were compared in PBS. Consequently, TMNs retained IRI activity at concentrations above 1.0 mg mL⁻¹ with MLGS values below 55% (Fig. S8).

DIS, which is often associated with ice-binding behavior, was examined for Phe-1, TMNs, and DOPA-1 at 1.0 mg mL⁻¹. Each sample exhibited isotropic crystal growth, thereby indicating that Phe-1, TMNs, and DOPA-1 did not bind to specific ice planes (Fig. S9).³² This is consistent with the fact that specific ice-binding is not required for IRI activity³³ and is absent in various IRI-active molecules, including surfactants, low-molecular-weight carbohydrates, and their derivatives.³⁴

Cryoprotective performance of TMNs and intracellular ice growth inhibition

DMSO is widely used as a cryoprotective agent because it suppresses ice formation and growth;³⁵ however, high concentrations of DMSO are cytotoxic and can cause adverse effects.³⁶ Replacing DMSO with an alternative is the most desirable method; however, this is yet to be feasible. Current strategies have focused on reducing DMSO content by supplementing IRI-active compounds.^37,38

We evaluated the cytotoxicity of TMNs to examine their potential as a cryoprotectant supplement in DMSO. Cells were treated with 0.1, 1.0, and 5.0 mg mL⁻¹ Phe-1, TMN, or DOPA-1 (Fig. S10). Although amino acids are generally biocompatible, their high concentrations can lead to cytotoxicity. Neither Phe-1 nor TMNs exhibited cytotoxicity up to 5.0 mg mL⁻¹. In contrast, DOPA-1 significantly reduced cell viability at 1.0 mg mL⁻¹. When dissolved at ≥1.0 mg mL⁻¹ in DW or saline, DOPA-1 solutions gradually turned brown over 14 days (Fig. S11) owing to melanin formation via autoxidation. This process proceeds through a series of oxidative transformations involving dopaquinone and leucodopachrome as intermediates and generates reactive oxygen species (ROS), including superoxide anions (O₂˙⁻), hydrogen peroxide (H₂O₂), and hydroxyl radicals (˙OH), which contribute to oxidative stress and cellular damage.^39,40 These results indicated that DOPA-1 is not a suitable candidate for cryoprotective applications.

Next, we assessed the cryoprotective potential of TMNs and Phe-1 in cell culture medium containing 1% DMSO. DMSO is typically used at concentrations of up to 10% (v/v) in cell culture media; however, 10% DMSO is too high to evaluate the cryoprotective potential of additional materials. As shown in Fig. 5a, reducing the DMSO concentration to 2.5% resulted in a cell viability of approximately 90%, which was comparable to that of 10% DMSO. Thus, concentrations below 2.5% were considered suitable for evaluating the contribution of the added compounds to cell viability. TMN nanostructures remained stable (<500 nm) in DMEM or DMEM + 1% DMSO, as confirmed by DLS and TEM (Fig. S12). We measured the cell viability in DMSO 1% in the presence of TMNs or Phe-1. HSC-3 cells were incubated with 0.45 mg mL⁻¹ of TMN or Phe-1 for 30 min to allow cellular uptake and then cryopreserved for 7 days in medium containing 1% DMSO and TMN or Phe-1. After rapid thawing at 37 °C, viability was assessed via MTS assay. TMNs increased cell viability by approximately 7.3% relative to that of the 1% DMSO control, whereas Phe-1 showed no improvement (Fig. 5a). In RAW264.7, TMNs increased the viability by 15.4%, which indicates that its cryoprotective effect is not limited to a specific cell type (Fig. 5b). To further evaluate the long-term cryoprotective efficacy of TMNs, we assessed post-thaw cell proliferation over 2 days. In HSC-3 cells, proliferation after 2 days reached 84.17% with TMNs + 1% DMSO compared to 61.4% with 1% DMSO alone. Similarly, in RAW264.7 cells, proliferation increased from 55.9% to 102.14% upon TMN addition (Fig. S13). These results demonstrate that TMNs not only enhance immediate post-thaw viability but also support sustained cell proliferation, indicating robust long-term cryoprotection beyond simple survival.

Fig. 5 Cryoprotective performance of TMNs and inhibition of intracellular ice growth. (a) Cell viability of HSC-3 cells after 1 week of cryopreservation following treatment with DMSO, DMSO + TMN, or DMSO + Phe-1. (b) Cell viability of RAW264.7 after 1 week of cryopreservation following treatment with DMSO, DMSO + TMN, or DMSO + Phe-1. (c) Temperature curve of the cryo-microscope stage during fluorescence imaging. Numbers and proportional distributions of intracellular ice crystals were determined at 60 and 120 s. Normalized ice fractions at 60 s or 120 s in (d) the nontreated cell (control), (e) 1% DMSO treated cells, (f) TMN solution (1% DMSO + 0.45 mg mL⁻¹ TMN) treated cells, and (g) Phe-1 solution (1% DMSO + 0.45 mg mL⁻¹ Phe-1) treated cells. Fluorescence images were acquired at 0 s, 60 s, 120 s, and 30 min (scale bar: 100 µm and inset scale bar: 10 µm).

Intracellular ice formation was monitored to elucidate the mechanism underlying cryoprotective efficacy of TMNs at ultralow temperatures.⁴¹ Since both extracellular and intracellular ice growth critically affect cell viability,⁴² we examined whether TMN uptake delays intracellular ice growth. Cells stained with a sodium indicator dye were mounted on coverslips, rapidly frozen to −196 °C and maintained at this temperature for 2 h. Intracellular ice growth was quantified at 60 s and 120 s post-freezing (Fig. 5c), followed by analysis of 200–300 ice grains per group. The stained molecules are released from the ice crystals, which caused the intracellular ice particles to appear as dark spots.

In untreated cells, the proportion of cells with an ice fraction below 0.4 decreased from 94% at 60 s to 63% at 120 s post-freezing, thereby indicating the growth of intracellular ice (Fig. 5d). A decrease in the proportion of cells with an intracellular ice fraction below 0.4 indicates an increase in intracellular ice crystal size, which proves that ice growth was not inhibited. In TMN-treated cells (0.45 mg mL⁻¹), this fraction decreased only slightly, from 94% to 90%, thereby indicating effective inhibition of ice growth (Fig. S14a).

In contrast, Phe-1-treated cells decreased from 82% to 62%, similar to untreated cells, which suggested that Phe-1 did not inhibit intracellular ice growth (Fig. S14b). Significant reductions in fluorescence intensity were observed in the samples treated with both amino acids after 30 min, which indicated the near absence of viable cells. Notably, rapid freezing and slow thawing protocols used to assess intracellular ice growth inhibition impose more severe stress on cells than typical cryopreservation protocols.⁴² Additionally, it has been confirmed that cryoprotective agents supplemented with minimal amounts of ice-recrystallization inhibitors do not provide adequate protection.⁴³ Therefore, when treated under the same conditions as the cryopreservation experiments, the TMN solution (0.45 mg mL⁻¹ TMN + 1% DMSO) exhibited ∼70% of cells with an ice fraction below 0.4 at 120 s, thereby indicating effective inhibition of ice crystal growth and formation of smaller ice crystals. Qualitative analysis showed that TMNs achieved the highest percentage of cells with low intracellular ice, followed by 1% DMSO (67%), the untreated control (63%), and Phe-1 + 1% DMSO (62%) (Fig. 5d–g). A numerical comparison between the TMN solution and 1% DMSO revealed that similar proportions of cells with ice crystal counts below 0.4 were 70 and 67%, respectively. Nevertheless, the ice crystal distribution map at 120 s demonstrated pronounced clustering in the lower left quadrant of the TMN solution, which indicated suppression of ice crystal growth despite the presence of intracellular ice. Additionally, the higher fluorescence intensity at 30 min in the TMN-treated samples supported increased cell survival, while corroborating the results of the MTS viability assay.

4. Conclusions

In this study, we investigated the factors affecting the IRI activity, with a focus on the monomers and self-assembled nanostructures of styrene and amino acids. Styrene monomers exhibited no IRI activity at any tested concentration, and PSNPs, despite forming nanoscale particles, showed no detectable IRI activity regardless of the particle size. This indicated that the hydrophobic domains alone were insufficient for IRI activity. Phe-1 exhibited no noticeable formation of self-assembled structures. Phe-5 produced an extensive aggregate with low IRI activity owing to its low colloidal stability. In contrast, L-tyrosine monomers produced TMNs very efficiently and demonstrated strong IRI activity even at low concentrations (0.1 mg mL⁻¹), which was attributed to the formation of self-assembled nanocrystals of tyrosine side chain via π–π and π–OH interactions.^15,30 And ice-binding materials (e.g., AF(G)P) can directly bind ice crystals, potentially generating needle-like structures that damage cells during cryopreservation. In contrast, non-ice-binding TMNs, which exhibit only IRI activity, are more suitable for cell cryopreservation applications.^25,44,45 Previous studies have suggested no significant differences in IRI activity between the L- and D-forms of amino acids and their mixtures, thus implying that D-tyrosine would likely exhibit pronounced IRI activity.¹⁸L-Pro₈ oligomers (Proline 8-mer) at 40 mg mL⁻¹ in 7.5% (v/v) DMSO + 10% (v/v) ethylene glycol (EG) achieved a mouse oocyte cryopreservation survival rate of 99.11%, an improvement from 95.93% in the control group (15% DMSO + 15% EG).⁴⁶ In comparison, atactic-PVA (a-PVA) at 1.0 mg mL⁻¹ yielded approximately 40% cell recovery for sheep RBC cryopreservation.⁴⁷ Notably, TMNs at a lower concentration of 0.5 mg mL⁻¹ in 1.0% DMSO (base DMEM) demonstrated superior cryopreservation efficacy, achieving 69.04% survival in RAW264.7 cells. When applied in cryopreservation, TMNs combined with DMSO improved cell survival, thereby effectively suppressing both intracellular and extracellular ice recrystallization at low DMSO concentrations and reducing cellular damage. Although TMNs cannot fully replace DMSO due to its limited solubility and the resulting constraints on usable concentrations, identifying TMNs as IRI-active materials remains important for the field of cell cryopreservation. Other studies have demonstrated cases where DMSO combined with IRI agents enhances cryopreservation efficacy.^37,48 As TMNs are a cluster of highly biocompatible L-tyrosine residues, it can be a promising supplemental ingredient for cryopreservation applications. Cryodamage arises from multiple factors acting synergistically, including ice crystal-induced membrane disruption, osmotic stress, and toxicity of cryoprotectant agents. Future studies should focus on synthesizing or screening compounds with structural features that enhance IRI activity, while ensuring colloidal stability, solubility at high concentrations, and biocompatibility. Such approaches are expected to enable a more effective control of ice recrystallization and improve cryopreservation efficiency. In conclusion, this study identified key factors contributing to IRI activity and proposed strategies for utilizing IRI-active agents as cryoprotectants, thus providing valuable insights into the development of more efficient cryopreservation techniques for diverse biological samples.

Author contributions

Yong Duk Kim: data acquisition, repetition, data analysis, investigation, methodology, and writing of the original draft. Yedam Lee: data acquisition, repetition, and analysis. Dong June Ahn: supervision, resources, writing, reviewing, and editing. Dong-Kwon Lim: conceptualization, supervision, resources, writing, reviewing, and editing.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting the findings of this study are available within the article and its supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6tb00077k.

Acknowledgements

This work was supported by Korea University, the KU-KIST Research Fund, and the National Research Foundation (NRF) of Korea (RS-2025-16070563).

Notes and references

1. T. Chang and G. Zhao, Adv. Sci., 2021, 8, 2002425.
2. J. Yang, L. Gao, M. Liu, X. Sui, Y. Zhu, C. Wen and L. Zhang, Trans. Tianjin Univ., 2020, 26, 409–423.
3. A. Fowler and M. Toner, Ann. N. Y. Acad. Sci., 2006, 1066, 119–135.
4. J. O. Karlsson, E. G. Cravalho, I. H. Borel Rinkes, R. G. Tompkins, M. L. Yarmush and M. Toner, Biophys. J., 1993, 65, 2524–2536.
5. P. W. Voorhees, J. Stat. Phys., 1985, 38, 231–252.
6. X. Qin, Z. Chen, L. Shen, H. Liu, X. Ouyang and G. Zhao, Nano-Micro Lett., 2023, 16, 3.
7. V. A. Tirado-Kulieva, W. R. Miranda-Zamora, E. Hernández-Martínez, L. R. Pantoja-Tirado, D. L. Bazán-Tantaleán and E. W. Camacho-Orbegoso, Heliyon, 2022, 8, e10973.
8. X. Zhang, H. Qi, J. Yang, X. Chen and L. Zhang, Ind. Eng. Chem. Res., 2023, 62, 12063–12072.
9. M. Leclère, B. K. Kwok, L. K. Wu, D. S. Allan and R. N. Ben, Bioconjugate Chem., 2011, 22, 1804–1810.
10. R. C. Deller, J. E. Pessin, M. Vatish, D. A. Mitchell and M. I. Gibson, Biomater. Sci., 2016, 4, 1079–1084.
11. H. Geng, X. Liu, G. Shi, G. Bai, J. Ma, J. Chen, Z. Wu, Y. Song, H. Fang and J. Wang, Angew. Chem., Int. Ed., 2017, 56, 997–1001.
12. Y. Xie, M. Wang, X. Liu, K. Zhou, Z. Wang, F. Xu, H. Zhou, H. Hu and B. Xu, J. Agric. Food Chem., 2024, 72, 21763–21771.
13. F. Bachtiger, T. R. Congdon, C. Stubbs, M. I. Gibson and G. C. Sosso, Nat. Commun., 2021, 12, 1323.
14. N. Jeon, I.-h Jeong, E. Cho, I. Choi, J. Lee, E. H. Han, H. J. Lee, P. C. W. Lee and E. Lee, JACS Au, 2023, 3, 154–164.
15. B. P. Partlow, M. Bagheri, J. L. Harden and D. L. Kaplan, Biomacromolecules, 2016, 17, 3570–3579.
16. R. Surís-Valls, T. P. Hogervorst, S. M. C. Schoenmakers, M. M. R. M. Hendrix, L. Milroy and I. K. Voets, Biomacromolecules, 2022, 23, 520–529.
17. Z. Ding, C. Wang, B. Zhou, M. Su, S. Yang, Y. Li, C. Qu and H. Liu, Nano Lett., 2022, 22, 5307–5315.
18. M. T. Warren, I. Galpin, M. Hasan, S. A. Hindmarsh, J. D. Padrnos, C. Edwards-Gayle, R. T. Mathers, D. J. Adams, G. C. Sosso and M. I. Gibson, Chem. Commun., 2022, 58, 7658–7661.
19. P. G. Georgiou, H. L. Marton, A. N. Baker, T. R. Congdon, T. F. Whale and M. I. Gibson, J. Am. Chem. Soc., 2021, 143, 7449–7461.
20. Y. D. Kim, W. H. Jung, D. J. Ahn and D.-K. Lim, Nano Lett., 2023, 23, 9500–9507.
21. P. G. Georgiou, I. Kontopoulou, T. R. Congdon and M. I. Gibson, Mater. Horiz., 2020, 7, 1883–1887.
22. J. Zhu, P. Zhu, Y. Zhu, Y. Ye, X. Sun, Y. Zhang, O. J. Rojas, P. Servati and F. Jiang, Carbohydr. Polym., 2024, 335, 122059.
23. C. I. Biggs, C. Stubbs, B. Graham, A. E. R. Fayter, M. Hasan and M. I. Gibson, Macromol. Biosci., 2019, 19, 1900082.
24. M. Reches and E. Gazit, Nano Lett., 2004, 4, 581–585.
25. T. Li, Y. Zhao, Q. Zhong and T. Wu, Biomacromolecules, 2019, 20, 1667–1674.
26. T. Li, M. Li, Q. Zhong and T. Wu, Carbohydr. Polym., 2020, 240, 116275.
27. M. T. Warren, I. Galpin, F. Bachtiger, M. I. Gibson and G. C. Sosso, J. Phys. Chem. Lett., 2022, 13, 2237–2244.
28. M. N. Frey, T. F. Koetzle, M. S. Lehmann and W. C. Hamilton, J. Chem. Phys., 1973, 58, 2547–2556.
29. C. Ménard-Moyon, V. Venkatesh, K. V. Krishna, F. Bonachera, S. Verma and A. Bianco, Chem. – Eur. J., 2015, 21, 11681–11686.
30. R. Y. Adhikari and J. J. Pujols, Nano Select, 2022, 3, 1314–1320.
31. Q. Guo, J. Chen, J. Wang, H. Zeng and J. Yu, Nanoscale, 2020, 12, 1307–1324.
32. Z. Wang, M. Li and T. Wu, J. Colloid Interface Sci., 2023, 629, 728–738.
33. R. Y. Tam, S. S. Ferreira, P. Czechura, J. L. Chaytor and R. N. Ben, J. Am. Chem. Soc., 2008, 130, 17494–17501.
34. C. J. Capicciotti, M. Leclère, F. A. Perras, D. L. Bryce, H. Paulin, J. Harden, Y. Liu and R. N. Ben, Chem. Sci., 2012, 3, 1408–1416.
35. J. E. Lovelock and M. W. H. Bishop, Nature, 1959, 183, 1394–1395.
36. Z. Shu, S. Heimfeld and D. Gao, Bone Marrow Transplant., 2014, 49, 469–476.
37. K. A. Murray, R. M. F. Tomás and M. I. Gibson, ACS Appl. Bio Mater., 2020, 3, 5627–5632.
38. D. Kaiser, N. M. Otto, O. McCallion, H. Hoffmann, G. Zarrinrad, M. Stein, C. Beier, I. Matz, M. Herschel, J. Hester, G. Moll, F. Issa, P. Reinke and A. Roemhild, Front. Cell Dev. Biol., 2021, 9, 750286.
39. A. N. Basma, E. J. Morris, W. J. Nicklas and H. M. Geller, J. Neurochem., 1995, 64, 825–832.
40. A. R. Soares, R. Marchiosi, C. Siqueira-Soares Rde, R. Barbosa de Lima, W. Dantas dos Santos and O. Ferrarese-Filho, Plant Signal Behav., 2014, 9, e28275.
41. T. K. Won, A. Shin, S. Y. Lee, B.-S. Kim and D. J. Ahn, Adv. Mater., 2025, 37, 2416304.
42. C. Lee, Y. Lee, W. H. Jung, T.-Y. Kim, T. Kim, D.-N. Kim and D. J. Ahn, Sci. Adv., 2022, 8, eadd0185.
43. Y. Sun, D. Maltseva, J. Liu, T. Hooker, V. Mailänder, H. Ramløv, A. L. DeVries, M. Bonn and K. Meister, Biomacromolecules, 2022, 23, 1214–1220.
44. T. Li, M. Li, V. P. Dia, S. Lenaghan, Q. Zhong and T. Wu, Int. J. Biol. Macromol., 2020, 165, 2378–2386.
45. M. Chow-Shi-Yée, J. G. Briard, M. Grondin, D. A. Averill-Bates, R. N. Ben and F. Ouellet, Protein Sci., 2016, 25, 974–986.
46. Q. Qin, L. Zhao, Z. Liu, T. Liu, J. Qu, X. Zhang, R. Li, L. Yan, J. Yan, S. Jin, J. Wang and J. Qiao, ACS Appl. Mater. Interfaces, 2020, 12, 18352–18362.
47. S. Jin, L. Yin, B. Kong, S. Wu, Z. He, H. Xue, Z. Liu, Q. Cheng, X. Zhou and J. Wang, Sci. China Chem., 2019, 62, 909–915.
48. Y. Wang, F. Zhang, X. Niu, G. Liu, D. Li, C. Xie, R. Yang and P. Wang, ACS Appl. Mater. Interfaces, 2026, 18, 13371–13383.

---