Self-Assembly of Plodia interpunctella Silk Particles: Mechanisms and Encapsulation Strategies

Abstract

Silk proteins represent a unique class of biopolymers produced by arthropods that can be leveraged to form a wide range of biomaterials, including particles at the nano- and micro-scale. Silk-derived particles are stable, biodegradable into amino acids, and can entrap and stabilize cargo for applications in healthcare, agriculture, and advanced materials. The utility of silk particles stems from their inherent protein folding, leading to crystalline domains that kinetically trap and stabilize cargoes. Organization and presence of these domains is dependent on synthesis methodology, protein composition (silk fibroin, silk sericin, silk composites), and encapsulated cargo molecules. Herein, we evaluate regenerated silk particles derived from Plodia interpunctella silk through mechanisms of phase separation and nanoprecipitation with and without small molecule encapsulants. Shifts in particle properties are expected to result from silk fiber source, mechanism of protein self-assembly between particle synthesis methods, and cargo molecule encapsulation. The morphology, crystallinity, size, dispersity, and zeta potential of particles derived from non-degummed P. interpunctella silk fibers exhibit distinct properties from degummed B. mori silk fibroin particles. P. interpunctella silk particles are capable of encapsulating small molecules of varying size, hydrophobicity, and charge with high efficiency. This study investigates silk-based biomaterials derived from native, non-degummed Plodia interpunctella silk fibers and demonstrates that protein composition and structural motifs of P. interpunctella silk fibroin influence particle formation and properties. Use of non-degummed silk fibers reduces the processing time required to regenerate silk proteins into biomaterials, offering a promising platform for expanding the functional space of silk-based biomaterials.

Graphical Abstract

1. Introduction

Biopolymer-based particle systems have established utility in healthcare, agriculture, and advanced materials applications due to their stability, ability to biodegrade into nontoxic byproducts, and capacity to entrap and stabilize cargo. Polymeric particles can be formed on the nano- to microscale, utilizing a variety of top-down or bottom-up approaches to obtain quick synthesis of stable particles for encapsulating active payloads or supporting gel and emulsion-based materials.^{1, 3, 4} Naturally derived polymeric particles represent a rising sustainable effort in particle synthesis, utilizing plant-based biopolymers such as cellulose and alginate or animal- and insect-derived proteins such as gelatin, chitosan, and silk to generate biodegradable and environmentally safe alternatives to synthetic materials.^{3, 5-8} Advantageous properties of biopolymer-based particles can be expanded through chemical modification or conjugation, cargo loading, or through composite polymer blends,^{3, 7-9} often manipulating the balance between structural and biologically or chemically active characteristics.

Silk proteins can be extracted from silk fibers with or without purification processes to generate aqueous silk biopolymer solutions that can be regenerated into various biomaterial formats.^{1, 10, 11} The utility of silk-based particles stems from their functionality and inherent crystallinity upon formation, supported by distinct differences in silk protein production, composition, and secondary structure.^12-18 Silk-based particles can be synthesized through various methodologies that exploit the hydrophobic and charged properties of structural silk proteins, causing the biopolymers to aggregate and form small particles.^{1, 12, 14, 19} Semicrystalline, water insoluble micro- and nanoparticles can be generated through nanoprecipitation, liquid-liquid phase separation, emulsion, electrospraying, and other strategies that vary in crystalline forming mechanisms, dependent on solvent exposure, time, and temperature.^{1, 12, 14, 15, 19-21} Modification of silk particle dispersity and stability can be tuned through experimental parameters to further define the biological fate and interactions of the particles with the environment and entrapped cargo.^{13, 14, 19} Varying the composition or properties of silk proteins used to form silk particles, such as molecular weight, amino acid composition, or the inclusion of additional structural or functional proteins,^{12, 13, 19} is also effective in modifying the formation and corresponding physical properties of silk particles.

Silk micro- and nanoparticles are primarily derived from silk fibroin (SF) of the domesticated silkworm, Bombyx mori (B. mori, mulberry silkworm). In silkworms, SF comprises the structural core of silk fibers and is coated by an outer layer of functional proteins (sericins, seroins, mucins) with a wide range of functions (adhesion, barrier functions, antifungal, antimicrobial).^{2, 22-25} Silk fibroin is often composed of fibroin heavy chain (FibH), fibroin light chain (FibL), and the small glycoprotein P25 (Fibrohexamerin) in Lepidoptera, with the exception of the Saturniidae family.^26-28 B. mori FibH is a highly repetitive structural protein with extensive crystalline regions that dictate the stability, distribution of cargo, and degradation of particles and other biomaterials.^{1, 14, 19, 29} However, the use of a singular fibroin protein sequence limits the variety of particle properties achievable. Across many silk-producing species, the secondary structure and composition of functional and structural proteins that comprise silk fibers are highly diverse,^{2, 22, 30, 31} leading to distinctive properties that are tailored to specific applications as fibers and as regenerated biomaterials.^16-18 The utility of many fiber coating proteins in silk-based particle systems is often underexplored and limited, though it is largely focused on sericins, the primary class of proteins present in the outer coating.^{25, 26, 32-35} Sericin particles are often formed in combination with other structural proteins, as the hydrophobic collapse of sericin proteins is unstable due to their high water solubility, low degradation temperature, and pH sensitivity.^{25, 33, 35} However, the increased abundance of polar side chains offers high chemical reactivity for silk sericins to bind molecules and cargo or crosslink and co-polymerize with stable, structural polymers to extend their antimicrobial properties and ability to activate the mammalian immune system into biomaterial applications.^{25, 34-37} Alternative silk sources and proteins represent an underutilized natural resource to broaden the application space and improve the understanding of how differing protein structures translate to the structure and function of a material both for applications in healthcare as well as agricultural and advanced manufacturing fields.

This work evaluates the formation of polymeric particles derived from Plodia interpunctella (P. interpunctella, pantry moth) silk as an alternative silk-based particle system. P. interpunctella is a common agricultural pest in the Pyralidae family that produces silk throughout its lifecycle in addition to spinning a thin cocoon for pupation.^38-41 P. interpunctella silkworms can be easily reared in a laboratory setting with control over environmental parameters that optimize the production of wandering silk,^{38, 39} leading to the formation of silk sheets that are easily collected from rearing containers and proposed previously as a prospective silk source for biomaterial studies.^{38, 42, 43} Wandering silk is produced by silkworms in their late 4^th-5^th instar stages and is often more abundant, can be collected without sacrificing the insect or inducing life cycle changes, and requires less time to collect from rearing containers compared to P. interpunctella cocoon silk, which is often entangled with remaining food and frass.^{38, 42, 44} Recent work characterizing P. interpunctella wandering silk fibers and silk sheets have evaluated the structural, physical, and bioactive properties of non-degummed and degummed fibers in comparison to other lepidopteran silks.⁴² Degummed P. interpunctella silk fibers exhibit shifts in morphology and mechanical properties compared to native fibers, corresponding to the removal of the outer fiber coating and heat treatment to silk fibroin proteins, in addition to a mass loss between 45-55%.^{38, 42} These findings suggest that P. interpunctella silk has a higher abundance of coating proteins than B. mori silk fibers (20-30% mass loss upon degumming) and is consistent with recent studies on a closely related pyralid species, Galleria mellonella (G. mellonella).²⁴ The study of P. interpunctella has largely been limited to its pest habits, lifecycle and development,^{39, 41, 45} and the production,^{38, 44} properties,⁴² or cell-interactions of native and degummed silk fibers and silk extracts.^{42, 43, 46} This study represents the first regenerated silk biomaterial generated from P. interpunctella silk fibers and the first regenerated biomaterial derived from a pyralid species, focusing on the formation of P. interpunctella silk microparticles.

2. Materials and Methods

2.1. FibH Secondary Structure Prediction

Secondary structures of P. interpunctella FibH (accessed from Kawahara et al.⁴⁷) and B. mori FibH (accession number: AF226688_1) proteins were predicted using MacVector’s Robson Garnier secondary structure algorithm based on amino acid sequence. Secondary structure predictions on the full length of P. interpunctella and B. mori FibH proteins are depicted in Supplemental Figure S1.

2.2. Silk Solution Preparation

2.2.1. P. interpunctella silk

Silk sheets were collected from P. interpunctella populations reared on a standardized wheat bran diet at 24 °C and 65 % relative humidity.^{38, 42} Non-degummed silk sheets were manually cleared of loose debris before solubilization in 9.3 M LiBr solution for 4 hours at 80 °C. Solutions were filtered with Miracloth (22-25 μm) and dialyzed against ultrapure water using 3.5 kDa MW dialysis membrane tubing for 48 hours to remove residual salt ions. Silk solutions were re-filtered with Miracloth to remove any insoluble silk or remaining impurities. The concentration (w/v %) of silk solutions were obtained by drying a known volume of silk solution and weighing the remaining dried silk.

2.2.2. B. mori SF

B. mori SF solution was prepared using previously described methods.¹ Briefly, 5 grams of cocoons were cut into dime-sized pieces and boiled in 0.02 M sodium carbonate (Sigma-Aldrich, USA) for 15 minutes, rinsed in ultrapure water to remove residual salt, and air-dried for 48 hours. Degummed SF mats were solubilized in 9.3 M LiBr solution for 4 hours at 60 °C. Solutions were dialyzed against ultrapure water using 3.5 kDa MW dialysis membrane tubing for 48 hours to remove residual salt ions. Aqueous SF solutions were centrifuged (4000 x g, 20 min, 4°C) to remove any insoluble silk or remaining impurities. The concentration (w/v %) of SF solutions were obtained by drying a known volume of silk solution and weighing the remaining dried solids.

2.3. Gel Electrophoresis

Molecular weight distributions of P. interpunctella silk and B. mori SF solutions were visualized with sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). All samples were run in triplicate and under reducing conditions through the addition of Bolt reducing agent. Solubilized silk was loaded (4 μg/well) into a NuPAGE 3-8% Tris-acetate gel (Invitrogen, Waltham, MA). A high molecular weight standard ladder was loaded into 2 lanes at a volume of 5 μL (HiMark Pre-stained Protein Standard, Invitrogen). Gels were loaded into a Mini gel tank and run with Bolt 1X MES SDS Running Buffer and 1 mL of Bolt antioxidant. Gels were run at 100 V for 10 minutes to allow proteins to enter the gel before increasing the voltage to 200 V until the dye front reached the bottom of the gel (~20 minutes). Gels were stained with a Colloidal Blue staining kit (LC6025, Invitrogen) and imaged on an Odyssey Fc (LI COR, Lincoln, NE) at 700 nm with a 30 second exposure time. MW distributions were obtained through the densiometric analysis along the length of each lane. Representative SDS-PAGE band profiles were generated by extracting grayscale intensity across the complete length of a well for each sample using ImageJ analysis software gel analysis tools (Supplemental Section 3).

2.4. Particle Fabrication

2.4.1. Phase Separation

Silk particles were formed through phase separation from polyvinyl alcohol (PVA) as described previously with modifications.^{1, 13, 14} Stock solutions of P. interpunctella silk and B. mori SF and PVA 5 % (w/v) were combined in for a total solution volume of 5 mL. Polymer concentration was held constant at 0.25 % (w/v) across all conditions. To modulate particle size, the weight ratio of silk:PVA in solution was varied between 1:4 or 1:8. The solution was then vortexed for 30 seconds before casting in a 100 x 15 mm petri dish and dried into a film overnight. Once dry, the film was dissolved in 20 mL of ultrapure water for 30 minutes while shaking. The solution was centrifuged (4000 g x 20 min, 4 °C). Pelleted particles were resuspended in 5 mL of ultrapure water and sonicated for 15 seconds at 15 % amplitude.

2.4.2. Nanoprecipitation

Silk particles were formed through nanoprecipitation as described previously with slight modifications.^{19, 20} P. interpunctella silk and B. mori SF polymer solution (0.25 % w/v) were added dropwise using a syringe pump to acetone at a drop length of 8 cm with constant stirring with a 25 mm x 8 mm stir bar. Solutions were prepared by adding 4 mL of silk polymer solution to 18 mL of acetone to maintain a concentration of >75 % (v/v) acetone. To modulate particle size, stir rate varied between 500 and 1000 rpm with a constant flow rate of 0.5 mL/min. Polymer concentration was held constant at 0.25 % (w/v) across all conditions. The nanoparticle suspension was allowed to evaporate for 12-24 hours before centrifuging (4000 g x 40 min, 4 °C). Pelleted particles were resuspended in 4 mL of ultrapure water and sonicated for 15 seconds at 15 % amplitude. Centrifuging and resuspension in water were repeated to rinse any remaining acetone from particle suspension.

2.5. Particle Loading

Loading of P. interpunctella silk particles was evaluated with curcumin, doxorubicin-HCl, and alcian blue as model cargo molecules of variable size, hydrophilicity, and charge. Stock solutions of cargo (1 mg/mL) were mixed with silk proteins in solution at a 1:100 mass ratio (molecule:silk) prior to phase separation or nanoprecipitation. Supernatants after pelleting particles were analyzed for residual cargo concentrations using UV-Vis spectrometry. Standard calibration curves were generated for analysis of encapsulation efficiency. Encapsulation efficiency was determined by dividing the mass of cargo encapsulated within particles by the initial mass loaded to aqueous silk solution (Equation 1).

$$Encapsulationefficiency({}^w∕_w\%)=\frac{massofcargoinparticles(mg)}{theoreticalmassofcargoadded(mg)}\times 100$$ (1)

2.6. Scanning Electron Microscopy (SEM)

Particles suspended in water were pipetted onto ZEISS/LEO SEM sample stubs with carbon conductive tape (Catalog No. 16202, Ted Pella, Inc., USA). Samples were dried overnight and sputter-coated with 12 nm of gold prior to imaging. Images were obtained on a Phenom Pure benchtop SEM using a backscattered electron detector with a standard sample holder and 5 kV acceleration voltage.

To supplement CT analysis of particle internal structure, histological cross sections of silk particles were imaged by SEM. Lyophilized silk particles were fixed in fixed in 10% phosphate buffered formalin (ThermoFisher Scientific, USA) for 4 hours prior to dehydration through a series of ethanol solutions of increasing concentrations before being transferred to xylene (ThermoFisher Scientific, USA). Dehydrated samples were left in a paraffin wax (ThermoFisher Scientific, USA) bath overnight before being embedded into wax molds for sectioning. Samples were sectioned at 5 μm thickness and mounted on SEM sample stubs. Samples were dried overnight and sputter-coated with 6 nm of gold.

2.7. Particle Size Distribution

The hydrodynamic diameter of silk particles in ultrapure water was measured on a Beckman Coulter LS 13 320 particle size analyzer. Particle formulations were pooled in triplicate (n=3) for analysis with refractive indices set at 1.37 for water and 1.54 for proteins. Average particle size was determined from number % distributions. To analysis particle dispersity, the span of particle size was calculated for each condition based on volume-distribution 10, 50, and 90% percentiles (Equation 2).

$$\text{Span}=\frac{{\mathrm{d}}_{\mathrm{V},90}-{\mathrm{d}}_{\mathrm{V},10}}{{\mathrm{d}}_{\mathrm{V},50}}$$ (2)

2.8. Zeta Potential

The zeta potential of silk particle formulations in ultrapure water was determined by dynamic light scattering at 25°C. Zeta potential was measured with a Universal Dip Cell Kit (ZEN1002) using a Zetasizer Nano-ZS Malvern Instrument at 25°C. Refractive indices were set at 1.33 for water and 1.60 for proteins.

2.9. Fourier Transform Infrared (FTIR) Spectroscopy

Particle formulations were pooled in triplicate (n=3) and lyophilized at −80°C and 0.300 mbar for 2 days before analysis. Lyophilized silk particles were analyzed using attenuated total reflectance (ATR) with a zinc selenide crystal on a Nicolet iS50 FTIR Spectrometer (ThermoFisher Scientific, USA) (UF Nanoscale Research Facility). Spectra were collected over 64 scans over a 4,000-650 cm⁻¹ wavenumber range at a resolution of 4 cm⁻¹. Background spectra were collected at the same conditions and subtracted from the sample spectra in OMNIC^™ Spectra Software before analysis with Origin data analysis software (OriginLab Corporation, USA). Deconvolution of protein secondary structure was performed according to Hu et al.⁴⁸ The amide I region (1,590-1,720 cm⁻¹) can be correlated to secondary structures as follows: 1,605–1,615 cm⁻¹ as side chains, 1,616-1,637 cm⁻¹ and 1,697–1,703 cm⁻¹ as β-sheet structures, 1,638–1,655 cm⁻¹ as random coils, 1,656–1,662 cm⁻¹ as α-helices, and 1,663–1,696 cm⁻¹ as turns.⁴⁸

2.10. Nano-Computed Tomography (nanoCT)

Silk particles were suspended in Lugol’s iodine contrast solution (Sigma-Aldrich, USA) for 2 hours and rinsed in ultrapure water to remove excess contrast solution. Stained particles were lyophilized at −80°C and 0.300 mbar for 2 days and then adhered to a carbon fiber rod for stability during scanning. Particles were scanned on a ZEISS Xradia 620 Versa X-ray microscope (XRM) (Carl Zeiss X-Ray Microscopy, Germany) with a 20X objective at 80 kV voltage, 10 W power, and 125 A current to achieve a pixel size of 0.3 μm. 3D volume files were analyzed and rendered with Volume Graphic’s VGStudio Max v2024.2 software suite. Grayscale value-based segmentation was utilized to isolate individual particles and measure particle diameter, volume, surface area, wall thickness, and grayscale intensity profiles.

2.11. Statistical Analysis

Experimental data are primarily expressed as average ± standard deviation with a minimum n=3. GraphPad Prism 10.2.1 (GraphPad Software Inc., USA) was used for statistical analysis. Analysis was performed with appropriate-size analysis of variance (ANOVA) with Tukey’s test post-hoc analysis. Statistical significance is reported as *p<0.05, **p<0.01, ***p<0.001, and ****p<0.0001 with key features shown in the figures and tables and further information provided in the supplemental information. Specifics on the statistical test, p values, and definition of n are all present in individual figure captions.

3. Results and Discussion

3.1. P. interpunctella and B. mori FibH proteins are distinct

Complete genome assemblies have recently been generated for P. interpunctella;^{47, 49} however, the proteins that comprise P. interpunctella silk fibers have not yet been reported outside of the FibH protein.^{47, 50} The P. interpunctella FibH protein possesses sequence irregularity, amino acid composition, and crystalline-forming motifs that are distinct from B. mori FibH (Figure 1) and other arthropod silks utilized in biomaterials,^{2, 42, 47, 51} suggesting the formation of secondary structures that yield differences in polymer function. The repetitive regions of pyralid FibH proteins are highly species-specific, varying in the length and arrangement of repeat sequences that determine the regularity of the protein sequence.^{2, 50, 52} P. interpunctella FibH possesses the general repeat form of 4 hydrophobic residues (I or V), 1-4 hydrophilic residues (E, D, N, Q, R), and an alanine- and sericin-rich crystalline motif followed by terminal repeat sections comprised of glycine-rich repeats (GX, GGX) that contain a high proportion of bulky residues (L, W, Y) (Figure 1BC). The terminal repeat section of P. interpunctella is highly variable, leading to a higher degree of sequence irregularity, or interruption of predicted crystalline β-sheet structures, as visualized in Figure 1A. Higher sequence irregularity has been previously correlated to decreased strength and rigidity in P. interpunctella silk fibers compared to B. mori and other lepidopteran silks,^{42, 50} as shorter, more irregular repeats could lead to decreased inter- and intramolecular interactions of crystalline regions. The differing crystalline content, sequence irregularity, and amino acid composition between P. interpunctella and B. mori FibH proteins is hypothesized to be the primary driver of physical property differences between P. interpunctella and B. mori silk fibers^{2, 42, 51} and regenerated biomaterials, in addition to the composition and function of other silk fiber proteins, discussed further in Supplemental Section 2.

Figure 1. Characteristics of B. mori (Bm) and P. interpunctella (Pi) FibH.

(A) Robson-Garnier predicted secondary structures in FibH proteins representative of the first 2000 amino acids in B. mori and P. interpunctella FibH. Predictions show that B. mori β-sheet region is uninterrupted and makes up a majority of the sequence compared to the P. interpunctella, suggesting a higher percentage of amino acids participate in β-sheet formation in B. mori compared to P. interpunctella. Complete Robson-Garnier predicted structures for the complete FibH protein of B. mori and P. interpunctella can be visualized in Supplemental Figure S1. (B) Amino acid composition of FibH proteins. Nonpolar amino acids F, M, W, and C were lower than 1% of protein composition and were excluded for visibility. (C) Repeat motif coverage of B. mori and P. interpunctella FibH adapted from Aikman et al.²

3.1.1. Molecular weight distribution of regenerated silk solution is process dependent

P. interpunctella silk can be solubilized in similar ways to B. mori SF (degummed 15 minutes)¹ (Figure 2A) and exhibits molecular weight distributions over the same range (Figure 2B). However, polymer length distribution and uniformity vary between the two regenerated silk solutions, corresponding to the treatment of native silk fibers prior to dissolution. P. interpunctella solution is derived from non-degummed silk fibers, resulting in distinct bands corresponding to FibH (~470 kDa) and FibL and P25 (~27-30 kDa) (Figure 2B). The band assigned to FibH is 13-14% larger than native P. interpunctella FibH protein based on amino acid sequence (413 kDa),^{47, 52} suggesting that FibH-FibL disulfide bonds and P25 interactions may still exist in solution. Samples are prepared under reducing conditions prior to SDS-PAGE analysis, which could disrupt FibH-FibL disulfide bonds, but are less likely to break non-covalent interactions with P25 or non-covalent interactions between SF and coating proteins.^{27, 28} In addition, high molecular weight proteins such as FibH may migrate differently than globular protein ladders,⁵³ influenced by high hydrophobicity or residual charge after denaturation that may add to discrepancy in band MW assignment. Despite some small discrepancies, these data suggest that the full length of the P. interpunctella FibH protein is maintained in the regenerated silk solution. This is expected given that the solubilization process does not contain a degumming step, where SF proteins are degraded due to hydrolysis at high temperatures.

Figure 2. Molecular weight distributions of regenerated silk solution.

(A) Schematic of methods used to form P. interpunctella silk solution. Methods are adapted from the extraction of B. mori SF with modifications.¹ (B) Molecular weight distributions of proteins in non-degummed P. interpunctella silk solution and 15-minute degummed B. mori SF solution determined by SDS-PAGE gel analysis. Position and molecular weight (kDa) of protein bands corresponding to a high molecular weight ladder are labeled in gray. The raw gel image can be visualized in Supplemental Figure S3.

As a non-degummed system, proteins present within the outer coating of P. interpunctella fibers, such as sericins or seroins, remain in solution, though they are not clearly visualized by SDS-PAGE. Absence of bands corresponding to outer coating proteins could result from their lower concentration respective to silk fibroin proteins or possible degradation of sericins and other coating proteins during dissolution.^{2, 54, 55} Thermal denaturation of P. interpunctella sericins is theorized to occur between 25-50 °C,⁴² lower than the temperature at which P. interpunctella silk is solubilized in LiBr (80 °C), leading to degradation of polymer length indicated by the light smears observed in gel analysis (Figure 2B, Supplemental Figure S3). The molecular weight distribution of non-degummed P. interpunctella silk solution is similar to a recent study on solubilized non-degummed B. mori silk.⁵⁴ Use of non-degummed silk in regenerated biomaterial formats could retain structures and interactions more similar to native silk proteins as the length of FibH is maintained, rather than degradation of silk fibroin proteins and removal of sericins that occurs during degumming. This is supported by the comparison of non-degummed B. mori silk and degummed B. mori SF for spinning artificial silk fibers, in which regenerated fibers from non-degummed silk solution more closely resembled native silk structure with mechanical properties that strongly outweighed regenerated degummed SF fibers.⁵⁴ Degummed B. mori SF solution consists of a broad distribution of SF fragments (Figure 2B, average MW ~250 kDa) resulting from degradation of SF proteins during degumming processes. This effect has been explored previously to modify the molecular weight and polymer interactions within regenerated SF particles, tuning resulting particle properties.^{12, 13, 19}

The inclusion of variable polymer length and composition are expected to influence the assembly of polymers into particles, modifying the interactions between proteins, particle size, charge, and the overall stability of particles in suspension. Furthermore, the inclusion of sericins and other coating proteins within P. interpunctella silk offers opportunities downstream for chemical modification and cross-linking, owing to their increased chemical reactivity.^{25, 36} With the inclusion of multiple silk protein classes that compose P. interpunctella silk solution, we discuss P. interpunctella regenerated biomaterials as a silk-based system in contrast to B. mori SF-derived materials. While previous studies of B. mori silk have identified that the combination of fibroin and sericin proteins can lead to unwanted immune responses,^{56, 57} the protein(s) in B. mori that induce mammalian inflammatory responses have not been clearly identified. Therefore, it is unclear whether the interaction or underlying protein(s) that promotes inflammatory responses is conserved in P. interpunctella. Initial studies on P. interpunctella silk fibers did not result significant differences in cell activity on non-degummed versus degummed fibers, in addition to the demonstrated utility of P. interpunctella silk extracts in antimicrobial and anticancer applications.^{43, 46, 58} Investigation is warranted on the purification of P. interpunctella SF to evaluate the immune response of non-degummed and degummed regenerated P. interpunctella biomaterials for human applications. Nonetheless, the utility of non-degummed silk particles can extend further than biomedicine, as the growing diversity of biopolymer-based particles finds utility across many fields. In addition to minimizing impacts to silk protein composition, the lack of degumming processes reduces processing time required to transform native silk fibers into a regenerated biomaterial. Use of non-degummed silk is less common in B. mori silk-derived biomaterials, yet it represents a unique, advantageous route with decreased protein treatments and higher yield of native silk protein properties and function. This work serves to evaluate the synthesis and properties of P. interpunctella particles as a baseline for prospective applications.

3.2. Protein size distribution and synthesis method govern particle assembly

The hydrophobic collapse of silk proteins during particle formation transforms aqueous polymers into water-insoluble, semicrystalline polymeric particles. In this work, we investigate property differences in particles formed through nanoprecipitation with acetone and liquid-liquid phase separation with polyvinyl alcohol (PVA) (Figure 3) as a function of insect silk fiber source. Differences in the speed of polymer assembly and β-sheet formation, in addition to polymer characteristics and composition, are expected to influence the properties and encapsulation of cargo within silk particles.^{14, 19, 20, 59-61} In nanoprecipitation, the rapid mixing of the silk polymer solution with the non-solvent (acetone) causes supersaturation of polymers within the system, generating a quick, almost instantaneous, aggregation of polymers from solution (Figure 3A).^{59, 60} Polymer aggregation is driven by interfacial surface tension differences and shear forces imposed on the system through stirring or agitation, often leading to smaller particle size distributions and dispersity at higher agitation rates.^{19, 59, 60} Alternatively, aggregation of polymers in phase separation occurs as water is removed from the system, a much slower precipitation period, which often allows for larger particles to grow and form a broader size distribution (Figure 3B).^{13, 14} Particle synthesis through nanoprecipitation and phase separation can be tuned through varying polymer concentration, molecular weight, ratio of silk:solvent, temperature, and imposed shear forces, previously employed to modulate the size and stability of B. mori SF particles.^{12-14, 19} To assess the feasibility to tune P. interpunctella silk particle formation, the stir rate during nanoprecipitation (500 vs. 1000 rpm) and weight ratio of silk:PVA during phase separation (1:4 vs. 1:8) is varied to modify imposed shear forces and silk biopolymer intermolecular interactions (dilution), respectively (Figure 3B).

Figure 3. Mechanisms and methodology of particle formation.

(A) Representation of silk protein self-assembly mechanisms for phase separation with polyvinyl alcohol (PVA) and nanoprecipitation with acetone. (B) Schematic of methods used to form particles. The weight ratio of silk:PVA (phase separation) and stir rate (nanoprecipitation) were used to tune particle size distributions.

3.2.1. Variable silk protein composition yields more disperse particles

P. interpunctella silk particles are first investigated alongside B. mori SF particles as a comparison to existing standards in silk-based particles and to investigate how protein composition and molecular weight distributions influence the organization of proteins within particles. Impacts of synthesis speed and protein composition can be visualized through SEM of particles formed through phase separation and nanoprecipitation (Figure 4A, Supplemental Figure S5). Phase separation resulted in visually disperse particles spanning a broad distribution of sizes in both P. interpunctella silk and B. mori SF particles. The median hydrodynamic diameter between P. interpunctella and B. mori phase separation particles were almost equivalent (2.4 μm and 2.35 μm, respectively; Supplemental Figure S6). However, the span of particle sizes was significantly greater in P. interpunctella silk (span of 2.8) compared to B. mori SF (span of 1.01, p=0.0003) (Figure 4BC). Substantial differences between particle size distributions were also observed between P. interpunctella and B. mori particles formed through nanoprecipitation, wherein P. interpunctella silk particles display a median diameter and span more than double that of B. mori SF particles (Figure 4C). P. interpunctella nanoprecipitation silk particles have a D50 value of 2.26 μm and size span of 3.04, which is significantly different to both values for B. mori SF particles, 280 nm (p<0.0001) and 1.02 (p=0.0001), respectively (Figure 4BC). These disparities in particle size distribution (further visualized in Supplemental Figure S6) are hypothesized to extend from the heterogeneous composition, conservation of native FibH protein length, and differing protein structures in P. interpunctella silk solution compared to B. mori SF solution (Figure 2). These silk polymer solution characteristics are expected to modify the hydrophobic collapse of silk proteins into particles, which is dependent on amino acid composition, protein structures, and molecular weight. Silk composition was more significant (p<0.01) than synthesis method for particle span, indicating that molecular weight distributions between P. interpunctella silk and B. mori SF largely influenced particle dispersity.

Figure 4. Influences of molecular weight and synthesis method on particle properties.

(A) Representative SEM images of particles formed with P. interpunctella (Pi) silk or B. mori (Bm) SF through phase separation (weight ratio 1:8) and nanoprecipitation (stir rate 500 rpm). Comparison of (B) median particle size D50, (C) diameter span, (D) zeta potential, and (E) β-sheet content for P. interpunctella and B. mori particles at each condition. Data is expressed at average ± standard deviation (n=3). Size distribution data is summarized in Supplemental Figure S6. Analyzed with 2-Way ANOVA with Tukey’s test post-hoc analysis. Statistical significance is reported as ***p<0.001 and ****p<0.0001.

The hydrophobic collapse of native FibH proteins present in P. interpunctella particles likely differs from SF fragments in B. mori particles, attributed primarily to molecular weight differences, but also to differing regularity of crystalline structures and the inclusion of supplemental silk proteins in P. interpunctella silk. Large proteins will collapse into larger particles than those formed from smaller molecular weight proteins under the same conditions. Additionally, P. interpunctella FibH is considered highly irregular in repeat structure, with varying arrangements and lengths of repeat sequences (Figure 1A).^{2, 42, 51} This irregularly, combined with the presence of hydrophilic proteins in P. interpunctella silk solution, could correspond to necessity for more P. interpunctella proteins to enable hydrophobic collapse, resulting in larger and more non-uniform particles. The assembly of silk proteins into P. interpunctella particles was confirmed by SDS-PAGE analysis of redissolved silk particles (Supplemental Figure S4). Bands assigned to FibH and FibL/P25 are observed in both phase separation and nanoprecipitation particles, in addition to the appearance of light smears (30-250 kDa) attributed to the degradation of sericins and other supplemental proteins. The intensity of coating proteins decreases in particles compared to solution (Supplemental Figure S4B), suggesting that some coating proteins did not collapse into silk particles and remained in solution (PVA or acetone), due in part to their more hydrophilic nature.^{25, 34, 35} While further analysis of protein composition could aid in this discussion, the lack of a fully annotated P. interpunctella genome or extensive studies of transcriptomics within the silk gland of P. interpunctella poses challenges for further assessments and data reconstruction from tools like proteomics. We fully anticipate that these particles contain some of the silk fiber coating proteins, which are hypothesized to be kinetically trapped within the β-sheets of the P. interpunctella silk fibroin. However, the contribution of these coating proteins to the hydrophobic collapse is hypothesized to be quite low, given their predicted hydrophilic properties.⁵²

The stability of silk particle suspensions is dependent on particle synthesis method (p<0.0001) and silk composition (p<0.001). P. interpunctella silk particles display high stability in phase separation (−16.8 mV) and nanoprecipitation (−18.9 mV), in contrast to B. mori SF particles that have a significant drop (p<0.0001) in zeta potential in phase separation (−7.3 mV) versus nanoprecipitation (−18.8 mV). Discrepancies in zeta potential between synthesis methods could be due to variation in mixing and solvent exposure, possibly generating a different organization or packing of proteins within particles and shifting surface charge distributions.¹² Though this effect is not completely understood, or yet explored in P. interpunctella silk particles, it has been previously theorized by Solomun et al for B. mori SF particles.¹² Manual versus microfluidic-assisted particle fabrication had substantial differences in zeta potential while silk composition was unchanged.¹² Exposure to PVA and vortex mixing during phase separation compared to acetone and high stir rates during nanoprecipitation could differentially affect the organization and density of proteins within and on the surface of particles, modifying the density of charges present and shifting the measured zeta potential (Figure 3B). Further differences between P. interpunctella silk particles and B. mori SF particles are expected considering the compositional differences between B. mori SF and P. interpunctella silk (Figure 2B). The zeta potential of P. interpunctella particles is larger than B. mori SF for both conditions, indicating greater stability in aqueous suspension at neutral pH (Figure 4D). Greater stability is hypothesized to primarily result from compositional differences in FibH proteins, in which P. interpunctella FibH has a lower isoelectric point (3.73) than B. mori FibH (4.07) and higher percentage of negatively charged amino acids (Glu, Asp) (Figure 1B). These protein characteristics point to P. interpunctella FibH having a more negative charge than B. mori at neutral pH, leading to higher zeta potential. It is also important to consider the contributions of outer coating proteins within P. interpunctella particles, as sericins are often negatively charged and would further decrease the surface charge of the protein molecule. Increased zeta potential and stability in suspension is largely beneficial in controlled release or drug delivery applications, preventing aggregation or clumping of particles before the cargo is transported to the target site.

3.2.2. Shifts in crystalline content and organization influence particle size and dispersity

β-sheet crystal formation in particles occurs as silk proteins aggregate and precipitate out of solution, originating from the semicrystalline structure of FibH proteins. The time period of β-sheet formation is expected to vary between phase separation and nanoprecipitation, due to removal of water from the protein backbone⁶² as water evaporates from the silk-PVA film and upon exposure to acetone and shear forces during nanoprecipitation (Figure 3A).^{14, 19, 20} Different speeds of β-sheet formation, in addition to varying FibH secondary structures and composition between P. interpunctella and B. mori, were equally significant in influencing the total crystalline content of silk particle formulations (p<0.05). P. interpunctella particles have high levels of crystallinity in both phase separation (51%) and nanoprecipitation (44%), similar to B. mori SF particles (37% and 51%, respectively) but with differing trends in overall crystalline content (Figure 4E).

B. mori SF particle crystallinity is consistent with literature,^{12, 13} with hypothesized trends that the rapid self-assembly of SF and exposure to acetone generates more crystalline particles than gradual phase separation. An opposite trend in crystallinity is observed in P. interpunctella silk, wherein phase separation resulted in more crystalline particles than nanoprecipitation (Figure 4E). While high agitation and organic solvents are highly influential in inducing β-sheet formation in B. mori SF materials, it is unclear if these hold the same control over P. interpunctella silk proteins. β-sheet crystalline structures are hypothesized to emanate from different amino acid repeat motifs and increased sequence irregularity in P. interpunctella FibH compared to B. mori FibH (Figure 1AC).^{2, 42, 51, 52} In addition, P. interpunctella FibH is predicted to have a lower overall crystalline content based on its amino acid sequence (Figure 1A), which has observed between native P. interpunctella and B. mori silk fibers.⁴² These structural protein variations could translate to the sensitivity or extent to which regenerated silk proteins would form ordered crystalline structures upon exposure to high shear and organic solvents. Furthermore, the length of native P. interpunctella FibH is maintained in a non-degummed silk particle system versus degummed B. mori SF (Figure 2). Limited time of β-sheet formation could impose different constraints, such as mobility in solution, on the assembly of large fibrillar proteins (400-470 kDa) than shorter fragments (~250 kDa) (Figure 2B). Longer time periods would allow larger FibH proteins to form favorable protein intermolecular interactions and arrange in more ordered, crystalline secondary structures than the almost instantaneous formation during nanoprecipitation that might promote more disorganized crystalline domains and lower intermolecular interactions. Similar trends of this hypothesis are seen in B. mori SF particles, wherein increasing average MW (decreased degumming times) corresponded to slight decreases in β-sheet content of particles.^{12, 13}

3.3. P. interpunctella silk particle dispersity can be tuned while maintaining stability

As a preliminary investigation to modulate intermolecular interactions and rates of β-sheet formation in P. interpunctella silk particles, the silk:PVA weight ratio and stir rate during nanoprecipitation were varied (Figure 5, Figure 3B). Particle morphology was consistent at each condition with a broad distribution of particle size, visualized by SEM (Figure 5A-D, Supplemental Figure S5). Phase separation particles formed at a 1:4 silk:PVA weight ratio have visibly increased abundance of larger, mid-range particles (5-10 μm) (Figure 5A) in contrast to 1:8 weight ratio particles that appear to exist primarily at small (<5 μm) and large scales (>10 μm) (Figure 5B). This trend was expected as the change in silk:PVA ratio should influence protein mobility in solution during particle formation, resulting in shifts in intermolecular interactions during mixing. Both nanoprecipitation formulations have visible small and large size range microparticles; however particles formed at 1000 rpm (Figure 5D) have an increased amount of particles <5 μm at the expense of large particles, observed at 500 rpm (Figure 5C). Thus, the increase in applied energy through an increased stir rate led to faster particle formation reducing the time for intermolecular interactions during particle formation. Overall, the decreased silk:PVA weight ratio (1:4) and lower stir rate (500 rpm) decrease constraints on particle assembly mechanisms (Figure 3A), reducing the surface area of interaction between silk proteins and PVA and applied shear forces and mixing time, respectively.

Figure 5. Influences of fabrication parameters on P. interpunctella particle properties.

Representative SEM images of particles formed with P. interpunctella silk through phase separation at a (A) 1:8 and (B) 1:4 silk:PVA weight ratio. Representative SEM images of particles formed with P. interpunctella silk through nanoprecipitation with a (C) 500 and (D) 1000 rpm stir rate. Comparison of (E) median particle size D50, (F) diameter span, (G) zeta potential, and (H) β-sheet content for P. interpunctella particles at each condition. Data is expressed at average ± standard deviation (n=3). Size distribution data is summarized in Supplemental Figure S6. Analyzed with 1-Way ANOVA with Tukey’s test post-hoc analysis. Statistical significance is reported as ***p<0.001 and ****p<0.0001.

Modification of particle synthesis in phase separation and nanoprecipitation led to shifts in particle size distributions while maintaining stability and crystallinity. The average hydrodynamic particle diameter did not largely shift between conditions, approximately ~2.5 μm (Figure 5E). However, the span of particle sizes significantly shifted between process modifications (Figure 5F). Decreasing the silk:PVA ratio from 1:8 to 1:4 resulted in an increase in particle span from 2.8 to 4.6 (p<0.0001), respectively. The increase in particle size dispersity can be attributed in part to the reduced exposed surface area of silk proteins to PVA at lower silk:PVA ratios and to changes in the potential for intermolecular interactions during the particle formation process. Both result in the formation of larger particles as silk proteins were less diluted during evaporation. Alternatively, doubling the stir rate during nanoprecipitation lowered particle span from 3 to 2.3 (p=0.0004), exhibiting a more uniform size distribution from increased shear forces imposed on the aqueous silk proteins upon addition to acetone.

As process modifications did not change the composition of particles or significantly shift particle diameter, zeta potential was also consistent across conditions (Figure 5G) since it is largely a function of protein charge and medium, which was also unchanged. There was no statistical significance in β-sheet content across particle formulations (Figure 5H). However, increasing the stir rate to 1000 rpm led to a similar average β-sheet content (44% to 42%). We hypothesized that quicker particle formation with faster stir rates could increase the rate of β-sheet formation in a manner that reduced intermolecular interactions and restrained the assembly of proteins into ordered or favorable orientations, leading to more disorganized and less crystalline particles. However, the difference between 500 rpm and 1000 rpm did not yield significant changes in β-sheet content as quantified by deconvolution of FTIR spectra.

Although the modifications of particle synthesis parameters shifted particle properties slightly, additional modifications should be explored to reduce particle dispersity and result in a uniform population for targeted applications. Uniformity of particle size leads to more predictable release profile and more consistent behavior, enhancing effectiveness in specific applications. Filtration could be utilized to separate synthesized particles into more uniform distributions; however, tuning synthesis methodology through energy input (i.e., sonication, agitation, temperature) or solvent choice can narrow particle dispersity without post-processing.^{14, 19} One of the most influential factors, however, is the molecular weight distribution of proteins prior to phase separation or nanoprecipitation. This could be explored through the use of degummed P. interpunctella silk to shift average MW, similar to previous studies of B. mori SF.^{12, 13, 19} Shifts in physical properties were previously observed between non-degummed and degummed P. interpunctella fibers,⁴² though the impact of degumming on molecular weight distribution of P. interpunctella silk proteins is currently under investigation. Alternatively, purification of the high molecular weight FibH band (Figure 2B) could decrease the protein dispersity and possibly translate to a more uniform micro-scale particle system. These compositional and process modifications are ultimately dependent on the internal organization and assembly of proteins within the particle.

3.4. Size-dependent variation in particle density and internal structure

P. interpunctella particles formed through phase separation and nanoprecipitation exhibit similar external morphology and properties (Figure 5), though it is unclear if the speed and mechanisms of particle assembly influence the internal structure of silk particles. To assess shifts in the organization and consistency within particles of different synthesis methods and of different sizes, nano-computed tomography (nanoCT) was employed to visualize a broad population of particles and identify trends in particle characteristics.

CT scans of lyophilized particles formed through phase separation (1:8 weight ratio, Figure 6A) and nanoprecipitation (500 rpm stir rate, Figure 6B) showed a similar spread of particle diameters to hydrodynamic diameter measurements (Figure 5E) and variations in internal structure, visualized in Supplemental Video 1. Particles were segmented based on grayscale intensity differences, wherein brighter areas represent more dense features while darker gray materials are less dense and have less polymer. Owing to the size dispersity of P. interpunctella particles, a broad range of diameters, surface area, and volumes were observed, offering the opportunity to evaluate variations in internal structures and polymer organization at multiple size-scales. Particles less than 7 μm have greater surface area relative to total volume (SA/V > 1) while increasing diameters (> 7 μm) decreased surface area relative to volume (Figure 6C). Higher SA/V ratio particles could have faster and more efficient transport of cargo into and out of the particle, while transport within low SA/V may be limited due to larger volumes of interfering polymer that hamper diffusion or particle degradation. Additionally, particles with decreasing SA/V ratios could display a higher surface charge density, leading to the larger zeta potential observed in B. mori SF nanoprecipitation nanoparticles versus phase separation microparticles (Figure 4D). These differences could translate to differing particle stability and release profiles of cargo based on diffusion limitations and size-dependent structure variations.

Figure 6. Internal structures of P. interpunctella silk particles.

3D rendering of segmented particles formed through (A) phase separation with a 1:8 weight ratio and (B) nanoprecipitation with a 500 rpm stir rate. (C) Surface area to volume ratios as a function of particle diameter (n=15). (D) Grayscale intensity line profiles through silk particles with diameters (i.) less than 5 μm, (ii.) between 5-10 μm, and (iii.) greater than 10 μm. Representative CT particle cross sections corresponding to each size division are pictured to the right of each plot for phase separation (outlined in blue) and nanoprecipitation (outlined in orange). Scale bars are 2 μm. (E) Percent of total particle volume that is composed of densely packed polymer as a function of particle diameter (n=15). (F) Representative SEM images of histological cross sections of phase separation (blue outline) and nanoprecipitation (orange outline) particles. Scale bars are 2 μm.

P. interpunctella particles exhibit similar size-dependent morphology and density variation throughout the particle structure (Figure 6D), independent of synthesis method. As particle size increases, the internal morphology shifts from uniform density or polymer organization to the development of core-shell-like structure, characterized by an inner sponge-like core surrounded by a denser outer wall. However, as particle size increases, the thickness of the shell is maintained around 2 μm for both nanoprecipitation and phase separation (Supplemental Figure S8), agreeing with previous studies on B. mori SF phase separation microparticles.¹³ Observation of an inner porous core with a denser outer wall was also visually confirmed through SEM of lyophilized particle cross sections (Figure 6F). Microparticles formed through phase separation and nanoprecipitation exhibit sponge-like internal structures characterized by irregular pores of varying size, shape, and density (Figure 6F). Smaller particles, less than 5 μm, exhibit relatively uniform density throughout the middle of the particle volume (Figure 6D,i).¹³ Compared to mid-range microparticles (5-10 um, Figure 6D,ii), the inner core of large microparticles (>10 um, Figure 6D,iii) comprises more of the particles’ total diameter. Maintenance of constant wall thickness with increasing particle diameter results in a decrease in percentage dense volume of the particle (Figure 6E), quantified through separate segmentation of the inner core versus total particle volume (Supplemental Figure S8).

These nanoCT and SEM investigations are the first to confirm prior hypotheses¹³ of silk particle internal structures. Implications of these morphological differences are important to encapsulation and release applications, as the diffusion or transport through dense areas would likely differ from a spongy-matrix core. The distribution of polymers within particles, in addition to the uniformity of these features across a broad range of diameters, better motivates the need to engineer a uniform size distribution as varying surface area-to-volume ratios and polymer density would influence the transport of cargo within particles and available surface areas for modifications or interactions with environmental cues.

3.5. Comparative analysis of small molecule encapsulation

Loading of P. interpunctella particles formed through phase separation (1:8 weight ratio) and nanoprecipitation (500 rpm stir rate) was evaluated at a 1:100 (molecule: silk) ratio by mass by adding curcumin stock to silk solution prior particle fabrication (Figure 3B). Curcumin loading did not significantly impact particle morphology (Figure 7AB) or size distribution (Supplemental Figure S9), which is likely due to its small size and low loading ratio relative to silk proteins. Preliminary investigations into other small molecules, alcian blue and doxorubicin-HCl, that vary in size, charge, and hydrophilicity also did not impact particle morphology and size distribution (Supplemental Figure S9). It is expected that the small molecules are distributed throughout the volume of the particle, entrapped within the dense wall and inner sponge-like core visualized through CT (Figure 6), without disrupting silk polymer interactions and causing morphological variations.

Figure 7. Curcumin encapsulation in P. interpunctella silk particles.

(A) Representative SEM and brightfield image of curcumin-loaded phase separation particles. (B) Representative SEM and brightfield image of curcumin-loaded nanoprecipitation particles. (C) Encapsulation efficiency of curcumin in silk particles. (D) Zeta potential determined by DLS. (E) β-sheet content determined by FTIR. Data is expressed at average ± standard deviation (n=3). Analyzed with unpaired parametric t-test. Statistical significance is reported as **p<0.01.

Encapsulation of curcumin in P. interpunctella particles was more efficient through nanoprecipitation than phase separation (Figure 7C). The encapsulation efficiency of nanoprecipitation was approximately 75%, almost double the efficiency during phase separation (~40%). As unloaded, or silk-only, particles formed through nanoprecipitation and phase separation had similar internal structures (Figure 6) and overall β-sheet content (Figure 5H), the discrepancy in encapsulation efficiency is attributed to the method of particle synthesis. Both synthesis methods utilize hydrophilic solvents (acetone, PVA) to induce the hydrophobic collapse of silk proteins into β-sheet structures.^{1, 13, 14, 19, 20} Curcumin is not expected to have large differences in affinity to one solvent versus the other, though it might partition into the non-solvent phases differently. Therefore, the decreased efficiency in encapsulation during phase separation versus nanoprecipitation is attributed to the speed of particle assembly and formation of β-sheet crystalline structures. Supersaturation of silk proteins and cargo during nanoprecipitation^{59, 60} causes an increase in localized concentrations, promoting faster aggregation and the likelihood of co-precipitation of silk and cargo, improving encapsulation efficiency. As phase separation occurs over a much longer time period than nanoprecipitation, it is possible that discrepancies in precipitation rates of silk proteins and cargo manifest in measured encapsulation efficiency. The precipitation of silk proteins likely occurs at a different rate than curcumin, leading to lower encapsulation efficiency as the two are not co-precipitated. Additionally, during drying of the silk-PVA film and further separation of silk particles from PVA,^{1, 13, 14} there are opportunities for diffusive loss of curcumin from the particles. Improving phase separation encapsulation efficiency could be explored by employing methods that delay polymer precipitation, such as polymer/PVA concentration or temperature, to promote co-precipitation of the aqueous phases. Alternatively, passive encapsulation of particles in suspension could improve consistency in encapsulation across the two methods, which have been successfully explored in B. mori particles.^{15, 20}

The stability and crystalline content of P. interpunctella particles was also maintained upon loading of curcumin. Curcumin has a net neutral charge at physiological pH and should not modify the surface charge of particles, which is supported by zeta potential values that are consistent with unloaded P. interpunctella particles (Figure 7D, Figure 5G, see Supplemental Table S2 for statistical comparisons). Encapsulation of charged molecules will likely cause some shifts in particle stability, increasing or decreasing zeta potential, respectively. Charged molecules may also have variable interactions with silk proteins based on charge differences that could influence encapsulation efficiency. This is observed upon encapsulation of alcian blue and doxorubicin-HCl, which are both positively charged (Supplemental Figure S9). Loading of alcian blue and dox-HCl resulted in an increase in zeta potential, showing a decrease in suspension stability (Supplemental Table S2). The β-sheet content of curcumin-loaded P. interpunctella particles is consistent with unloaded particles, maintaining around 50% crystalline content for both phase separation and nanoprecipitation (Figure 7E, Supplemental Table S3). Encapsulation of alcian blue and dox-HCl also did not shift crystalline content with P. interpunctella particles (Supplemental Figure S9). This suggests that self-assembly of P. interpunctella silk proteins into particles and formation of β-sheet structures are not largely influenced by the molecules and loading ratios investigated herein. The capacity to encapsulate small molecules of varying size, charge, and hydrophobicity, while maintaining particle properties consistent with unloaded particles, warrants further investigations into targeted applications of P. interpunctella silk particles as a biopolymer-based delivery system.

4. Conclusions

For the first time, we demonstrated the use of P. interpunctella silk solution to form silk microparticles, comparing P. interpunctella silk particles to B. mori silk fibroin particles. We validate that formulations and techniques developed for regenerated B. mori silk fibroin solutions can be modified and applied to P. interpunctella silk solutions, but that resulting particle properties are driven by differences in amino acid composition and secondary structures between the two species. Specifically, we use phase separation in PVA and nanoprecipitation in acetone to compare silk source and formulation parameters, demonstrating that the highest zeta potential was achieved with P. interpunctella silk particles formed via nanoprecipitation (~ −19 mV). Results show that particle properties depend on silk polymer chemical structure, composition, individual protein sequences, and molecular weight distributions and can be further modulated through fabrication methodologies.

Furthermore, we establish that degumming of the native silk fibers is not required to regenerate P. interpunctella silk fibers or form silk particles. This is critical because it allows us to maintain the native FibH structure and high molecular weight (~413 kDa) found in the fiber, skipping the traditional boiling step and resulting thermal degradation. We anticipate that this enables some conservation of the native protein-protein interactions found within the fiber during the particle formation process. Moreover, results also demonstrate the tunability of P. interpunctella silk particle formation, while also establishing its potential as a biopolymer-based encapsulation system relevant to controlled release applications. This opens the door for P. interpunctella to serve as an additional source of silk fibroin and silk proteins (e.g., sericins, seroins) for delivery applications in healthcare, agriculture, and in other uses of advanced materials.

Moving forward, laboratory rearing of P. interpunctella gives us an opportunity to use this silk particle platform to gain fundamental insight into the function of silk fiber coating proteins (e.g., sericins, seroins, mucins). The P. interpunctella particle system will offer opportunities to chemically crosslink sericins, incorporate genetically modified fiber proteins, and evaluate the unique impact individual coating proteins have on overall particle properties using synthetic biology tools. As we expand these results, we will focus on evaluating differences between non-degummed, degummed, and purified SF systems to understand how purification and separation methods of SF (and resulting shifts in MWD) influence particle properties and cell response towards biomedical applications.

Supplementary Material

Complete Robson-Garnier plots (Figure S1), silk fibroin proteins (Table S1), alignment of P. interpunctella candidate silk proteins (Figure S2), solution gel analysis (Figure S3), particle gel analysis (Figure S4), higher magnification SEM images (Figure S5), particle size parameters and distributions (Figure S6), particle segmentation and wall thickness (Figure S8), encapsulation of alcian blue and doxorubicin (Figure S9), summary of statistical comparisons for unloaded and loaded particles (Table S1, Table S2).

All raw data is available at this link:

https://www.dropbox.com/scl/fo/649v5c9vob3u5465c59nm/AMXUT2qWYAjNymS4GWGFglA?rlkey=hw3wu93nhnaawwyz9h5ysrpbw&st=1n24u4zb&dl=0