The impact of degumming conditions on the properties of silk films for biomedical applications

Abstract

The degumming process to remove sericin decreases silk fiber strength; however, the impact of degumming on the mechanical properties of regenerated silk biomaterials has not been established. This study investigated the effect of degumming temperature, time, alkaline component and alkaline concentration on the mechanical properties of silk fibroin films. Sericin removal was estimated using weight loss; 10 samples with 12.2–29.4% weight loss were then further characterized in terms of fiber mechanical properties, fiber surface morphology, molecular weight distribution and film tensile strength. A negative correlation was found between weight loss and fiber tensile strength. This loss of fiber strength under harsher degumming conditions had a direct impact on the tensile strength of regenerated films. Mild degumming conditions (weight loss of 12.2%) led to higher film strength (8.9 MPa), whereas aggressive degumming conditions (with 29.4% weight loss) resulted in significantly weaker films (4.3 MPa). The presence of some residual sericin, after mild degumming, is likely to affect the mechanical properties of the regenerated silk films. These results will assist in the development of materials with mechanical and biocompatibility properties tuned to specific biomedical applications.

Keywords

Materials materials properties processing fabrication

The use of silk as a biomaterial has received increasing attention over recent years due to the material’s desirable tensile strength, biocompatibility and the ability to tune the rate of biodegradation to match the application.^1–3 Silkworm cocoons represent a versatile starting point for the manufacture of biomaterials since they can be processed into a large number of different forms, including non-woven mats, sponges, gels, powders or thin films.^2,4,5 Raw silk is composed of two fibers coated in a glue-like protein called sericin.⁶ The fiber is composed of fibroin, a protein comprising two subunits, a heavy chain (∼350 kDa) and a light chain (∼25 kDa).⁷ In order to manufacture biomaterials from silk, sericin is first removed in a process known as degumming. Unlike fibroin, sericin is water soluble at high temperatures so degumming most commonly involves boiling silk in water, sometimes with either salt or detergent before rinsing to remove the dissolved sericin.⁸ These degummed fibers can then be used to make non-woven mats or powders. Alternatively, the fibroin can be dissolved, most commonly using a chaotropic salt such as lithium bromide; the liquid silk solution is then used as the starting point to make regenerated silk materials, such as hydrogels, sponges or films.^2,9

The most commonly used method of degumming involves boiling the cocoons in a solution containing a wetting agent and an alkali. This method is favored because it removes high levels of sericin in a short time.¹⁰ Temperatures lower than 80℃ are not sufficient to remove sericin, even with extended degumming time,¹¹ while temperatures of 100℃ and higher can damage the fiber surface.¹¹ It is therefore considered that degumming is optimal somewhere in the temperature range of 90–100℃. The wetting agent is added to reduce the surface tension between the cocoons and water, which greatly improves the efficiency of degumming.¹¹ Commonly used agents include surfactants such as sodium dodecyl sulfate (SDS) and, more traditionally, olive oil (Marseille) soap.¹⁰ An alkaline solution is added to hydrolyze the sericin, aiding in its removal.^10,11

Although fibroin fibers are left largely intact after degumming, alkaline hydrolysis is a non-specific process. As a result, alkaline degumming always results in a certain degree of damage to the primary and higher order protein structures, most commonly within the amorphous (non β-sheet) regions.¹¹ In addition, the severe conditions can cause weakening of the non-covalent structures (hydrogen bonding and Van der Waal forces), causing a breakdown in the hierarchical structure of the fiber.¹² This damage results in a reduction in the average molecular weight and increase in dispersity.^13,14 As a result, the mechanical properties of fibroin, such as Young’s modulus and yield strength, are reduced¹⁴ and, in extreme cases, fibrillation becomes visible on the fiber surface.¹⁵ For this reason, many studies have sought to optimize conditions to remove sericin effectively while retaining as much of the intrinsic physical and mechanical properties of native fibroin as possible.^8,11,16–18

In addition to affecting fiber mechanical properties, degumming conditions can have a significant impact on the properties of silk biomaterials. Sericin removal is vital for biomaterials, because when combined with fibroin, it appears to be associated with an increased inflammatory response.¹⁹ There is also some evidence to suggest that the extent of degumming can have a negative impact on cell proliferation and viability of adherent cells.²⁰ Efficient removal of sericin is therefore a requirement to develop a biocompatible silk-based biomaterial. The average molecular weight of silk fibroin in biomaterials has also been associated with a higher degradation rate in vivo,²¹ so choosing conditions that minimize degradation may be useful in applications such as load bearing tissue where good mechanical strength and slow degradation during healing are desirable. In addition to these considerations, any material intended for use within the human body must be rigorously tested in order to meet the requirements of regulatory bodies. The production of biomaterials for biomedical applications must be carried out under controlled conditions to ensure consistency. Part of this process involves developing production methods (such as degumming) that are reproducible and well characterized.²⁰ In the case of degumming there is a need to measure and control residual sericin considering the possible immunological impact that it can have on clinical outcomes.

There has been a large volume of published work in recent years on the effect of silk processing on various properties of silk fibers. For example, the link between degumming severity and fiber damage (assessed by fiber morphology and molecular weight distribution) has been investigated,²² as has the effect of degumming on silk fiber tensile properties.^11,12,16 However, there has been relatively little attention paid to the possible link between degumming conditions and the mechanical properties of regenerated silk materials, such as silk films. Given the large number of possible variations in the degumming process (choice of alkali, temperature, degumming time, etc.), it is vital to investigate each factor systematically in order to determine its importance to the properties of degummed fibers as well as films cast from regenerated silk. This study therefore investigated the impact of temperature, alkaline agent, concentration and degumming time on the properties of degummed fibers as well as regenerated silk films. This work continues our development of silk films as a potential biomaterial to repair perforations of the tympanic membrane.^23–25 We attempted to tune the degumming process to minimize fibroin degradation and so maximize film mechanical strength. In order to find the optimal conditions, a range of degumming conditions were evaluated to determine percentage weight loss and visual appearance of surface and cross-sectional images of the samples. From the weight loss and scanning electron microscopy (SEM) results, a subset of conditions were chosen for further analysis including fiber tensile testing, assessment of fiber degradation and, ultimately, the tensile strength of films produced from these samples.

Materials and methods

Material preparation

Raw cocoons from multivoltine Bombyx mori silkworms were purchased from production centers in Northeast India. All experiments were conducted with cocoons from a single batch. Cocoons were cut into four pieces and washed thoroughly with warm (approximately 40℃) tap water to remove any foreign material. The pieces were then rinsed at least four times with room temperature deionized water (dH₂O) and dried overnight at 40℃.

Degumming

Degumming was performed in a rotary dying machine (Ahiba Nuance Top Speed); samples were degummed in aluminum pots with a material (g) to liquor (mL) ratio of 1:50. All degumming solutions contained a final concentration of 1 g/L of pure olive oil soap as a wetting agent (The Natural Olive Oil Soap Factory, Cowaramup, Australia). Degumming solutions were preheated to 40℃ for 1 h in a fan forced oven before being added to the aluminum pots. Cocoons were then soaked in soap solution for 15 min at 40℃ to wet the cocoons before the alkaline solution (sodium carbonate or sodium bicarbonate) was added at three concentrations, as described in Table 1. Each degumming cycle was 30 min at either 90℃ or 98℃ (see Table 1).

Table 1.

Summary of the degumming treatments and variables. Each treatment was performed in triplicate

Base type: sodium carbonate			Base type: sodium bicarbonate
Temperature (℃)	Number of times degummed	Base concentration (g/L)	Temperature (℃)	Number of times degummed	Base concentration (g/L)
90	1	0.5	90	1	0.5
		1.0			1.0
		2.0			2.0
	2	0.5		2	0.5
		1.0			1.0
		2.0			2.0
	3	0.5		3	0.5
		1.0			1.0
		2.0			2.0
98	1	0.5	98	1	0.5
		1.0			1.0
		2.0			2.0
	2	0.5		2	0.5
		1.0			1.0
		2.0			2.0
	3	0.5		3	0.5
		1.0			1.0
		2.0			2.0

Weight loss

Degumming performance was initially measured by determining the percentage weight loss of samples after degumming. Cocoon pieces (1 g weight) were dried overnight at 90℃, and weighed again to obtain a pre-degumming weight. Samples were placed on the scales for 2 min to allow the reading to stabilize before the exact weight was recorded. The parameters tested are summarized in Table 1.

Each degumming condition was performed on triplicate 1 g samples. After degumming, samples were dried again overnight and re-weighed; weight loss was defined as

Weightloss (%) = (1 - \frac{mass of degummed fibers}{mass of cocoons}) \times 100

Based on the results of the weight loss experiments, a number of degumming conditions were chosen for further analysis. Specifically, different conditions that produced similar weight loss results were chosen in order to see if the tensile properties of degummed silk were different. Treatment groups with the lowest weight loss and highest weight loss were also included as controls to show the effect of under-degumming and over-degumming on mechanical properties of the degummed silk. As a reference, the conditions used in previous work by our group were included; these conditions were 30 min degummed at 98℃ with 2 g/L sodium carbonate and 1 g/L Marseille soap.

Weight loss statistical analysis

The mean weight loss results from all treatment groups were compared using both a four-way analysis of variance (ANOVA) and a multiple regression using the software package statistiXL (version 1.8; www.statistixl.com). ANOVA was used to compare mean weight losses over the whole data set, while the multiple regression was used to compare each of the four independent variables (base type, base concentration, degumming temperature and number of times degummed). Means were deemed statistically different at p < 0.05.

Fiber cross-sections

The sericin content of raw silk fibers was independently determined visually in order to validate the weight loss experiments. Raw fibers were obtained by carefully peeling them off the cocoon. To obtain the silk fiber cross-sections, 20–25 fibers were embedded in resin (Spurr replacement kit; TAAB Laboratories; Berkshire, England) and then cured at 60℃ for 48 h. These fiber-embedded resin blocks were then cut perpendicular to the fiber axis using an ultramicrotome EMUC6 (Leica Microsystems, Wetzlar, Germany). The cut faces of the resin sections were gold coated and viewed under SEM at 3 kV accelerated voltage and a working distance of 6–7 mm. Sericin content was calculated by subtracting the cross-sectional area of the fibroin core from the cross-sectional area of the whole fiber as described below:

total area = area of entire fiber \times (brin 1 + brin 2 + sericin layer)

A total of 20 fibers from a raw cocoon sample were measured using ImageJ, version 1.47 (Wayne Rasband, National Institute of Health, USA).

Fiber morphology

Fiber samples from each of the treatments chosen from the weight loss experiments were gold-coated using a SCD 050 sputter-coater (Bal-Tec, Leica Microsystems) and visualized using a field emission scanning electron microscope (FE-SEM, Zeiss Supra 55VP or Zeiss Leo 1530). Images were acquired at a working distance of 5–7 mm and accelerating voltage of 3–5 kV.

Fiber tensile properties

For tensile measurements, degumming was performed using only the peeled outer layer of each cocoon, since silk fiber diameter and tenacity change along the length of the filament from the outer to the inner part of the cocoon.^26,27 One gram samples of these outer layer fibers were degummed according to the 10 conditions chosen from the weight loss experiments. Degumming was carried out as described above. Once dried, 100 fibers from each sample were pulled out of the fiber mass carefully so as to minimize pre-stretching. Fibers for testing were conditioned at 20℃ ± 2℃ and 65% ± 2% relative humidity for at least 48 h prior to tensile testing. The fibers were loaded onto a rack and the tensile strength measured using a Favimat automated tensile tester using 100 mg tension weights (Textechno, Mönchengladbach, Germany). Fibers were tested with a gauge length of 25 mm and an extension rate of 15 mm/min under standard conditions (20℃ ± 2℃ and 65% ± 2% relative humidity). The load cell of the instrument used was 210 cN. The linear density of each sample was calculated by the machine and used to standardize the breaking force for each fiber using a vibration method. Using this method, the resonance frequency is measured at a constant gauge length of 25 mm and appropriate pre-tension automatically adjusted by the instrument for maximum resonance. Linear density is calculated according to the following formula:

T_{t} = \frac{F_{V}}{(4 \times f^{2} \times L^{2})}

where L is the test section length.

The cross-sectional area was calculated from the measured linear density and known density of 1.35 g/cc of silk for each specimen to calculate the stress in MPa.

Film tensile strength

Each of the treatments chosen after the initial weight loss experiments was dissolved to make regenerated films, and the strength of these films was assessed using uniaxial tensile testing. Briefly, 1.25 g of degummed fibers from each sample was dissolved in 9.3 M lithium bromide at a material:liquor ratio of 1:7.5. Samples were dissolved at 60℃ for 5 h and 500 rpm shaking using a Thermomixer C (Eppendorf, Hamburg, Germany), then dialyzed against dH₂O at 4℃ for 3 days with six water changes. After dialysis, samples were centrifuged at 7000g at 4℃ for 15 min to remove any debris. The concentration of each sample was calculated by drying four aliquots of 500 µL at 60℃ for 3 hours and weighing the resulting films on a four-decimal-place balance. The concentration was determined to be the mean of the four films. Each sample of silk solution was then diluted to 3% w/v and 3.41 mL of this solution was pipetted into 55 mm Petri dishes to achieve a final film thickness of 30 µm. This volume was determined by measuring the actual thickness of films cast from different volumes of 3% solution; thickness measurements were made using SEM images of the film cross-sections (results not shown). Once dry, the films were removed and the thickness of each was measured by taking eight measurements across the film using a three-decimal-place digital micrometer (Kinchrome, Melbourne, Australia). The films were cut into 5 mm wide strips and soaked overnight in 70% ethanol to induce β-sheet formation. The cut strips were rinsed and transferred to dH₂O at least an hour before testing.

Strips were then wet tensile tested to break using a model 5967 tester (Instron, Norwood, MA, USA) equipped with a 5 N load cell and a custom-built, temperature-controlled water bath. Samples were submerged in dH₂O heated to 37℃ to mimic biological conditions; each sample was submerged and allowed to equilibrate to this temperature for at least 30 s. Tensile testing was conducted using a gauge length of 15 mm, a pre-load of 1 cN and an extension rate of 150 mm/min, which was chosen in order to break samples within approximately 20 s. The average thickness of each film and the standardized width of each strip were then used to calculate the cross-sectional area, which was used subsequently in the calculations for the tensile properties.

Fibroin molecular weight distribution

The degradation of fibroin during degumming was assessed using SDS polyacrylamide gel electrophoresis (SDS-PAGE). Degummed fiber samples for electrophoresis were dissolved using a similar method to that used to make the films described above. Degummed fibers (500 mg) from each sample were dissolved using the same 1:7.5 fiber:lithium bromide ratio. Samples for electrophoresis were dialyzed at 4℃ against dH₂O for 24 h with six water changes with mixing. The concentration of each sample was calculated using a protein assay kit based on the Bradford method (Bio-Rad, Hercules, CA, USA). Electrophoresis was carried out using a Mini Protean system (Bio-Rad, Hercules, CA, USA), and run with a modified Laemmli²⁸ method. Samples were suspended in running buffer but no reducing agent was added and the samples were not denatured. This modification was made in order to avoid causing the fibroin protein to gel and precipitate out of solution. Samples were run on a 4–20% Tris-glycine gradient gel (NuSep, Bogart, GA, USA) at 250 V for 45 min and visualized using Coomassie Brilliant Blue R (Sigma-Aldrich, St. Louis, MO, USA). Each lane contained 40 µg of protein. The running buffer consisted of 0.025 mol/L Tris, 0.192 mol/L glycine and 0.1% (w/v) SDS at a pH of 8.3.

Results

Weight loss

When investigating the effect of base type (sodium carbonate or sodium bicarbonate), base concentration (0.5 g/L, 1 g/L or 2 g/L), degumming temperature (90℃ and 98℃) and degumming time (1, 2 or 3 × 30 min cycles), all factors were found to have a significant effect on weight loss when compared using a four-way ANOVA (P < 0.05, Figure 1). Overall, the lowest weight loss value (12.2%) corresponded to degumming with 0.5% sodium bicarbonate, degummed once at 90℃, while the highest weight loss (29.4%) was found using 2 g/L sodium carbonate, degummed three times at 98℃ (Figure 1).

Figure 1.

Mean weight loss for samples degummed using different conditions. Results are presented for degumming at 90℃ ((a) and (c)) and 98℃ ((b) and (d)), as well as degumming with sodium carbonate ((a) and (b)) and sodium bicarbonate ((c) and (d)).

When comparing individual factors using a multiple regression, sodium carbonate was more efficient at degumming in terms of weight loss than sodium bicarbonate, with a higher minimum weight loss (24.7% for sodium carbonate versus 11.3% for sodium bicarbonate). That is, when all other factors were equal, weight loss for all samples degummed with sodium carbonate (Figures 1(a) and (b)) was higher on average than that of the equivalent samples degummed with sodium bicarbonate (Figures 1(c) and (d)). For each base, increasing concentration generally resulted in higher weight loss (P = 0.08). However, the effect of base concentration was less pronounced for sodium carbonate than for sodium bicarbonate and less pronounced at 98℃ compared with 90℃. Temperature had a significant effect on weight loss (P = 0.000); all samples degummed at 98℃ resulted in higher weight loss than the equivalent samples degummed at 90℃. Finally, the number of times degummed (and so degumming time) significantly affected weight loss (P = 0.000), with a strong positive correlation between the number of times degummed and the average percentage weight loss. Again, the impact of the number of times degummed was less pronounced at 98℃ than at 90℃ and less pronounced for the sodium carbonate samples compared with those degummed with sodium bicarbonate (Figure 1). The results suggest that at 98℃ and/or degumming with sodium carbonate, increasing the concentration and time of treatment have a negligible affect on weight loss. On the other hand, at 90℃ and/or using sodium bicarbonate, increasing concentration and time of treatment can significantly influence degumming performance. Within the range of parameters used, we also found that there was high weight loss, with less variation between treatments when either sodium carbonate was used or degumming was performed up to three times, irrespective of other conditions.

Based on the weight loss results, 10 treatments were chosen for further analysis (Table 2). The sampling strategy included treatments that gave the lowest and highest weight loss, and intermediate range treatments with similar weight loss but different degumming conditions, in order to establish whether the same level of sericin removal could be achieved with milder conditions in order to reduce degumming costs and ensure less damage to the intrinsic properties of silk.

Table 2.

Summary of the treatments chosen based on weight loss results for further analysis

Treatment group	Degumming conditions
Treatment group	Base concentration (g/L)	Base type	Degumming temperature (℃)	Number of times degummed
1	0.5	Sodium carbonate	90	3
2	0.5	Sodium carbonate	98	3
3	0.5	Sodium bicarbonate	90	3
4	0.5	Sodium bicarbonate	98	3
5	2	Sodium carbonate	90	3
6	2	Sodium bicarbonate	90	3
7	2	Sodium carbonate	90	1
8	2	Sodium bicarbonate	90	1
9	2	Sodium carbonate	98	1
10	2	Sodium carbonate	98	3

Validation of sericin content by measuring cross-sectional area

Weight loss is often correlated with the loss of sericin; however, during base degumming, it is possible that weight loss also comes from hydrolysis of silk fibroin, which is not desired. Therefore, the removal of sericin can be monitored by imaging the fiber surface and cross-section as well as monitoring tensile properties of fibers. Figure 2 shows a SEM image of two silk fiber brins encased in a thick layer of sericin obtained from the outer layer of a silk cocoon. The cross-sections of raw silk fibers were used to calculate sericin content based on area. These measurements were used to validate the weight loss data. Area calculations from 20 such cross-sections suggest an average sericin content of 42.2 ± 3.6% by volume. It is known that the sericin gradually reduces from the outer towards the inner part of the shell, and we reported 37.4 ± 2.7% of sericin content based on weight loss in the outer part of silk cocoons.²⁹ Therefore, the weight loss data follows closely with the sericin content in silk fibers, considering expected lower density of sericin compared to silk fiber.

Figure 2.

Cross-section of a raw (un-degummed) silk fiber showing the two brins (B1 and B2) encased within a layer of sericin (S). Calculations to determine the percentage sericin based on surface area can be found in the text.

Fiber surface morphology

SEM imaging of the degummed fibers revealed the progressive removal of the sericin layer as the weight loss of the samples increased. Raw cocoon samples showed the classic morphology of two fiber brins encased in a thick layer of sericin (Figure 3(a)). Even under the mildest conditions with the least weight loss (12.2%), the fibroin brins had separated so that the fibers were a tangle of individual fibers. However, this treatment showed a significant proportion of sericin still present on the fiber surface (Figure 3(b)); the sericin appeared to have bubbled and cracked during the degumming (Figure 3(b), arrows) but did not lift off completely. At 25.7% weight loss (Figure 3(c)) most of the sericin appeared to have been removed, revealing the characteristic striated surface of the fibroin fiber. However, some sericin deposits still remained across the surface. At 27.7% and 27.9% weight loss (Figures 3(d) and (e)), no sericin was visible; however, some damage to the fibroin surface was evident (Figure 3(d), arrow). The treatment with the highest recorded weight loss (29.4%) showed no sericin, but extensive fibroin fibrillation was evident (Figure 3(f), arrow).

Figure 3.

Scanning electron micrographs of raw (a) and degummed (b)–(f) fibers with increasing weight loss: (a) raw fiber; (b) 12.2% weight loss, arrows indicate bubbling in sericin coating; (c) 25.7% weight loss; (d) 27.7% weight loss, arrow indicates damage to the fibroin surface; (e) 27.9% weight loss; (f) 29.4% weight loss, arrow indicates fibroin fibrillation. F: fibroin; S: sericin.

Fiber tensile properties

When plotted against weight loss, the fiber tenacity of the treatments chosen for testing loosely followed a negative linear correlation (R²= 0.63, Figure 4), suggesting that as more sericin is removed (i.e. higher weight loss), mechanical strength of the degummed silk decreases proportionally. As expected, the treatment with the highest mechanical strength (Treatment 3; 603.8 MPa) also showed the lowest weight loss (25.7%), while the treatment with the lowest strength (Treatment 10, 466.0 MPa) had the highest weight loss (29.37%). Treatment group 8, which showed a weight loss of 12.2%, could not be tested for mechanical properties since the degummed fibers were strongly bound together by the residual sericin, making it impossible to remove individual fibers without damaging them. Treatments 1, 4 and 9 were found to have a fiber strength that was above the line of best fit (Figure 4, represented by triangle symbols), so these were nominated as candidates for the optimal degumming conditions.

Figure 4.

Correlation between fiber tensile strength and weight loss from the 10 degummed samples chosen for further analysis. Number indicates treatment group (see Table 2). R² value denotes the coefficient of determination for the line of best fit.

Assessment of fibroin degradation by SDS-PAGE

The same 10 degummed fibroin treatment groups visualized by SEM and tested for tensile properties were also dissolved in lithium bromide and visualized using SDS-PAGE. It is expected that silk fibroins with high alanine and glycine contents in heavy chains should show some migration (smearing) on SDS-PAGE gels compared with normal globular proteins.¹³ Although all treatments showed signs of smearing, some samples showed a higher proportion of lower molecular weight banding than others, suggesting these samples were more degraded (Figure 5). For example, when comparing fibers degummed with the same sodium carbonate concentration (0.5 g/L) and the same number of times degummed, the sample degummed at 90℃ showed less smearing than the equivalent sample degummed at 98℃ (Figure 5, lanes 1 and 2). The level of degradation correlated closely with the fiber tenacity (Table 3), with the three treatments that gave the darkest smearing (Figure 5, lanes 2, 5 and 10) showing the lowest fiber tenacity.

Figure 5.

Sodium dodecyl sulfate polyacrylamide gel electrophoresis analysis of the degummed samples. Numbered lanes 1–10 represent the 10 degumming treatment groups described in Table 2.

Table 3.

Fiber tensile strength of degummed samples (mean ± SD)

Treatment group	Tensile strength (MPa)
1	582 ± 59
2	554 ± 69
3	604 ± 58
4	578 ± 70
5	544 ± 54
6	559 ± 98
7	573 ± 56
8	N/A
9	570 ± 71
10	466 ± 70

NB: treatment group 8 could not be tested since the degummed fibers were strongly bound by residual sericin, making them impossible to extract without damage. Refer to Table 2 for testing conditions for the 10 treatment groups

Film tensile properties

Wet tensile testing of films created from each of the 10 treatments revealed some key differences in mechanical properties. The 10 treatments could be broken into two statistical groups based on tensile strength (Table 4). One group of seven treatments showed a mean tensile strength in the range of 4–6 MPa, while the remaining three treatments (1, 3 and 8) were substantially higher, in the range of 8–10 MPa (P < 0.05), although Treatment 3 was statistically similar to Treatment 9, indicating some overlap between the two groups. There was no apparent relationship between tensile strength and modulus; the modulus of the treatment groups ranged from 43.5 MPa for Treatment 8 through to 67.4 MPa for Treatment 1 (Table 4). The elongation of most treatments was in the region of 100–150%, with 1 and 8 showing the highest elongation (88.6% and 265.8%, respectively). Treatment 10 showed by far the lowest elongation at only 22.2% (Table 4).

Table 4.

Tensile properties of silk films (mean ± SD)

Treatment group	Tensile strength (MPa)	Young's modulus (MPa)	Strain (%)	Sample size (n)
1	10.8 ± 2.9	64.6 ± 17.2	353.3 ± 93.8	20
2	6.1 ± 0.7	53.2 ± 15	132.7 ± 110.2	18
3	8.2 ± 2.3	55 ± 26.2	150.9 ± 103.4	16
4	5.3 ± 0.8	63.3 ± 17.4	88.6 ± 75.6	21
5	5.6 ± 1.4	48.5 ± 16.7	97.5 ± 51.4	16
6	5.3 ± 1	46.3 ± 9	101.4 ± 74.9	13
7	5.7 ± 1.7	58.7 ± 17.2	98.6 ± 38.4	14
8	8.9 ± 2.1	43.5 ± 18.5	265.8 ± 65.9	13
9	6.3 ± 1.4	47.3 ± 12.5	140.1 ± 50.8	18
10	4.3 ± 0.9	47.1 ± 13.4	22.1 ± 6.4	14

Refer to Table 2 for testing conditions for the 10 treatment groups

Film tensile strength did show a positive linear correlation with elongation (R²= 0.93, graph not shown). The optimal treatments identified in fiber testing showed mixed results; Treatment 1, which showed promising fiber strength, also showed excellent film properties, with the highest tensile strength and modulus of any treatment. Treatments 4 and 9, however, showed average tensile strength and elongation, although Treatment 4 showed the second highest modulus of all groups.

Discussion

This study established for the first time a link between degumming conditions, fiber appearance and properties and the mechanical properties of regenerated silk films. Through screening initially by weight loss and then by fiber tensile strength, SEM and SDS-PAGE, three “optimal” treatment groups (Treatments 1, 4 and 9) were found to have a good compromise between sufficient sericin removal with minimal fiber degradation and relatively high tensile strength. Of these groups, Treatment 1 (3 × degummed with 0.5 g/L sodium carbonate at 90℃) gave films with high tensile properties and excellent elongation. However, the weight loss of this sample at 27.3% was slightly lower than the other two treatments. Given that sericin removal is considered to be vital for biomedical applications, a slightly higher weight loss with complete sericin removal but a slight loss of strength is considered to be optimal. For this reason the conditions that met these requirements was once degummed using 2 g/L sodium carbonate at 98℃ (Treatment 9). This sample showed sufficient degumming (27.7% weight loss) with higher than average fiber and film mechanical properties.

Statistical analyses of the weight loss results indicate that all of the factors tested showed a significant effect on weight loss. Of all of the factors, base concentration (from 0.5 to 2 g/L) had the least impact, suggesting that changing the concentration is not a priority when optimizing conditions. On the other hand, degumming temperature, base type and number of times degummed had a greater impact. It has been previously suggested that degumming duration is the primary factor influencing fiber tensile properties, with a steady deterioration in tensile strength, modulus and strain as degumming duration increases.¹¹ However, the results presented here suggest that it is possible to obtain superior tensile strength at longer durations by reducing the degumming temperature, base concentration and/or base type. For example, sample 1 (0.5 g/L Na2CO3, 3 × degummed at 90℃) showed the highest film tensile strength, modulus and elongation, despite being degummed three times. The effect of temperature on degumming efficiency is well established; optimization for commercial scale degumming at 98℃ often allows degumming for as little as 15 min.¹⁰ Weight loss studies also reveal an interaction between temperature and time: at lower temperatures, degumming time had more of an impact on sericin removal than at higher temperatures. This should therefore be taken into account when choosing appropriate degumming conditions.

The combination of the weight loss matrix and fiber imaging used in this study proved to be a powerful screening method. Using this method, it was determined that samples with weight loss of 12–25% showed residual sericin (using SEM), while the sample with 29.4% weight loss showed extensive fibrillation, indicating that the extra weight loss above 28% was achieved through the loss of some fibroin. Samples with a weight loss of between 27% and 28% showed optimal degumming, with no visible sericin and minimal damage to the fibroin surface. It must be noted, however, that even these samples showed some surface cracking, indicating that there is not a point at which all sericin is removed without any fibroin degradation. This is due to the non-specific nature of alkaline hydrolysis as a degumming method.¹¹ As sericin is progressively removed and the surface of the fibroin brins is exposed, the sodium carbonate will inevitably begin hydrolyzing the fibroin.¹⁵ This observation that fiber damage begins before the complete removal of sericin confirms a similar result in a previous study.^16,22

The fiber strength and SDS-PAGE results also revealed that it was possible to manipulate degumming conditions to achieve a similar level of sericin removal (as indicated by similar weight loss), but less fibroin degradation and higher fiber tensile strength. For example, Samples 4 and 5 both showed identical weight loss (27.7%), but Sample 5 with the more aggressive base at higher concentration showed greater smearing after SDS-PAGE, with the 25 kDa light chain band almost completely absent. The light chain of this sample was clearly visible. The fiber tensile properties of these two samples reflected this observation, with Sample 5 having a lower fiber tensile strength than Sample 4, even with the same weight loss. Overall, the optimal samples (Treatments 1, 4 and 9) were found to have good sericin removal (above 27% weight loss) but maintained a tensile strength above the line of best fit. The tensile strength of these three samples was all higher than 560 MPa, which is remarkable given that raw, un-degummed fibers have a strength of around 600 MPa.³⁰ Only one other sample (Sample 3) was able to achieve a higher tensile strength (603.8 MPa), but at the cost of less efficient sericin removal (with a weight loss of 25.7%).

The treatment-induced differences in fiber mechanical properties appeared to carry through to the films, with lower degumming weight loss corresponding with films with better mechanical properties. Of the three optimal treatments with good fiber properties, films from Treatment 1 showed high strength and modulus. However, while Treatments 4 and 9 showed excellent fiber properties, the strength of the films was lower at 5.29 and 6.29 MPa, respectively. Films from Treatments 3 and 8 (not identified as optimal conditions) also showed high tensile strength (above 8 MPa). It is assumed that the fibers from this sample would similarly show high strength, but this sample could not be fiber tested since the high residual sericin content made extracting fibers from the degummed mass impossible. Interestingly, the three samples with significantly higher strength than all other samples (Samples 1, 3 and 8) were all degummed at 90℃, indicating that film strength may be more dependent on degumming temperature than fiber strength. Similarly, Sample 10, which showed fibrillation and poor fiber mechanical properties gave the lowest film tensile strength (4.3 MPa) and extremely low extension (22.1%). There was, however, no clear overall correlation between the fiber mechanical properties and the film mechanical properties. It is therefore possible that there are other factors affecting the mechanical properties of the films than degumming alone. Of all of the films tested, those from Sample 1 had by far the highest mechanical strength (10.8 MPa) and exceptional elongation (353.3%), even when compared with samples of lower weight loss. The exact reason for this result is unclear. This sample showed good weight loss (27.3%) but slightly lower than Samples 2, 4, 5, 9 and 10, which all had a weight loss of 27.7% or higher. It is possible that this sample may therefore have a small amount of residual sericin present compared with the higher weight loss samples, which were presumably completely degummed. It is therefore possible that this small amount of residual sericin plays a part in the mechanical properties of the regenerated films. A recent study on regenerated silk fibers found that small amounts of residual sericin^17,22 resulted in higher tenacity and elongation in regenerated silk fibers.¹⁷ The high extension of this sample may result from the sericin within the film acting both as a plasticizer and a bridge with hydrogen bonding. A similar plasticizing role has been ascribed to glycerol to explain an increase in extension that was observed when dry testing silk/glycerol films.³¹ While present, this effect seems to be highly dependent on degumming conditions, as samples with both higher and lower weight loss showed lower tensile strength and much lower extension. The presence of an additional factor, such as the plasticizing effect of sericin, may help to explain the lack of a clear trend between the mechanical properties of the silk fibers and regenerated fibroin films.

The determination of the exact sericin content of silk is complicated, as the proportion of sericin to fibroin varies between silkworm species and also from the inner to the outer cocoon layers. Despite the variety of degumming conditions used in this study, weight loss mostly remained around 27–28%. This might indicate that this is the point at which all sericin is removed. These results are in agreement with previous weight loss studies, which estimate B. mori sericin content for silk ranging from 25% to 30%.^15,22,32,33 In order to validate the weight loss results, the sericin content of B. mori silk was independently calculated by measuring the proportion of total area taken up by the sericin layer when viewing a cross-section of raw cocoon fibers. The average sericin content (volume based) determined using this method was 42.2%. Although significantly higher than the weight loss results, the fibers used for cross-sections were all taken from the outer-most layer of the raw cocoon, which is known to have proportionally more sericin than the inner cocoon layers.³³ The 42.2% determined here is similar to weight loss experiments with the outer layer only, which determined an average weight loss of 37.4%.²⁹ An accurate sericin content would therefore require the averaging of cross-sectional areas taken from all cocoon layers, making it unfeasible as a validation method for weight loss studies. It is therefore suggested that this method be used in conjunction with fiber surface imaging to qualitatively monitor sericin removal.

Conclusions

By systematically investigating four degumming factors and screening the resulting samples, it was found that good degumming (as determined by weight loss and SEM) could be achieved with minimal fiber degradation (as determined by SEM and SDS-PAGE) and minimal loss of fiber strength (as determined by tensile testing). This study also demonstrated for the first time that degradation caused by degumming has a significant impact on the mechanical properties of regenerated silk films. Degumming conditions of 30 min at 98℃ using 2 g/L sodium carbonate are proposed as suitable for biomedical applications where mechanical strength is desirable. These conditions were found to offer a good balance of efficient sericin removal with minimal loss of strength due to degradation. The strength of regenerated silk films may be affected by the presence of residual sericin, which acts as a plasticizer and a bridge with hydrogen bonding. Sericin removal was favored over strength to maximize biocompatibility.

Footnotes

Funding

This work was supported by an Australian Research Council Linkage Grant (LP110200547).

Acknowledgment

Thank you to Dr Luke Henderson for the useful discussions.

References

Altman

Diaz

Jakuba

. Silk-based biomaterials. Biomaterials 2003; 24: 401–416.

Vepari

Kaplan

. Silk as a biomaterial. Prog Polym Sci 2007; 32: 991–1007.

Kundu

Rajkhowa

Kundu

. Silk fibroin biomaterials for tissue regenerations. Adv Drug Del Rev 2013; 65: 457–470.

Rajkhowa

Wang

. Ultra-fine silk powder preparation through rotary and ball milling. Powder Technol 2008; 185: 87–95.

Rajkhowa

Wang

Kanwar

. Fabrication of ultrafine powder from eri silk through attritor and jet milling. Powder Technol 2009; 191: 155–63.

Chen

Porter

Vollrath

. Morphology and structure of silkworm cocoons. Mater Sci Eng C 2012; 32: 772–778.

Jin

Kaplan

. Mechanism of silk processing in insects and spiders. Nature 2003; 424: 1057–61.

Pérez-Rigueiro

Elices

Llorca

. Effect of degumming on the tensile properties of silkworm (Bombyx mori) silk fiber. J Appl Polym Sci 2002; 84: 1431–1437.

Duki

Forys

. Designing silk-silk protein alloy materials for biomedical applications. J Vis Exp 2014, pp. e50891–e50891.

10.

Gulrajani

. Degumming of silk. Rev Prog Color Relat Top 1992; 22: 79–89.

11.

Teh

TKH

Toh

S-L

Goh

JCH

. Optimization of the silk scaffold sericin removal process for retention of silk fibroin protein structure and mechanical properties. Biomed Mater 2010; 5: 035008–035008.

12.

Wang

H-Y

Zhang

Y-Q

. Effect of regeneration of liquid silk fibroin on its structure and characterization. Soft Matter 2013; 9: 138–145.

13.

Rajkhowa

Wang

Kanwar

. Molecular weight and secondary structure change in eri silk during alkali degumming and powdering. J Appl Polym Sci 2011; 119: 1339–1347.

14.

Yamada

Nakao

Takasu

. Preparation of undegraded native molecular fibroin solution from silkworm cocoons. Mat Sci Eng C-Bio S 2001; 14: 41–46.

15.

Freddi

Mossotti

Innocenti

. Degumming of silk fabric with several proteases. J Biotechnol 2003; 106: 101–112.

16.

Jiang

Liu

Wang

. Tensile behavior and morphology of differently degummed silkworm (Bombyx mori) cocoon silk fibres. Mater Lett 2006; 60: 919–925.

17.

Kim

. The effect of residual silk sericin on the structure and mechanical property of regenerated silk filament. Int J Biol Macromol 2007; 41: 346–353.

18.

Pérez-Rigueiro

Elices

Llorca

. Tensile properties of silkworm silk obtained by forced silking. J Appl Polym Sci 2001; 82: 1928–1935.

19.

Panilaitis

Altman

Chen

. Macrophage responses to silk. Biomaterials 2003; 24: 3079–3085.

20.

Wray

Gallego

. Effect of processing on silk-based biomaterials: reproducibility and biocompatibility. J Biomed Mater Res B 2011; 99B: 89–101.

21.

Cao

Wang

. Biodegradation of silk biomaterials. Int J Mol Sci 2009; 10: 1514–1524.

22.

Yoon

. Effect of degumming condition on the solution properties and electrospinnablity of regenerated silk solution. Int J Biol Macromol 2013; 55: 161–168.

23.

Levin

Redmond

Rajkhowa

. Preliminary results of the application of a silk fibroin scaffold to otology. Otolaryng Head Neck 2010; 142. Supplement 1: S33–S35.

24.

Rajkhowa

Levin

Redmond

. Structure and properties of biomedical films prepared from aqueous and acidic silk fibroin solutions. J Biomed Mater Res A 2011; 97A: 37–45.

25.

Levin

Redmond

Rajkhowa

. Utilising silk fibroin membranes as scaffolds for the growth of tympanic membrane keratinocytes, and application to myringoplasty surgery. J Laryngol Otol 2013; 127 Supplement 1: S13–20.

26.

Zhao

H-P

Feng

X-Q

S-W

. Mechanical properties of silkworm cocoons. Polymer 2005; 46: 9192–9201.

27.

Zhao

H-P

Feng

X-Q

Cui

W-Z

. Mechanical properties of silkworm cocoon pelades. Eng Fract Mech 2007; 74: 1953–1962.

28.

Laemmli

. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970; 227: 680–685.

29.

Kaur

Rajkhowa

Tsuzuki

. Photoprotection by Silk Cocoons. Biomacromolecules 2013; 14: 3660–3667.

30.

Omenetto

Kaplan

. New opportunities for an ancient material. Science 2010; 329: 528–531.

31.

Wang

. Insoluble and flexible silk films containing glycerol. Biomacromolecules 2009; 11: 143–150.

32.

Gupta

Agrawal

Chaudhary

. Cleaner process for extraction of sericin using infrared. J Cleaner Prod 2013; 52: 488–494.

33.

Mondal

Trivedy

Kumar

. The silk proteins, sericin and fibroin in silkworm, Bombyx mori Linn. – a review. Caspian J Env Sci 2007; 5: 63–76.