See discussions, stats, and author profiles for this publication at: [Link]

net/publication/230613200

Dissolving Feather Keratin Using Sodium Sulfide for Bio-Polymer

Applications

Article in Journal of Environmental Polymer Degradation · December 2011

DOI: 10.1007/s10924-011-0365-6

CITATIONS READS

94 3,353

3 authors:

Andrew J Poole Russell E Lyons

The Commonwealth Scientific and Industrial Research Organisation The University of Queensland
25 PUBLICATIONS 1,270 CITATIONS 118 PUBLICATIONS 3,522 CITATIONS

SEE PROFILE SEE PROFILE

Jeffrey S Church
The Commonwealth Scientific and Industrial Research Organisation
131 PUBLICATIONS 4,227 CITATIONS

SEE PROFILE

All content following this page was uploaded by Russell E Lyons on 28 May 2014.

The user has requested enhancement of the downloaded file.

J Polym Environ (2011) 19:995–1004
DOI 10.1007/s10924-011-0365-6

ORIGINAL PAPER

Dissolving Feather Keratin Using Sodium Sulfide

for Bio-Polymer Applications
Andrew J. Poole • Russell E. Lyons •

Jeffrey S. Church

Published online: 27 September 2011

Ó Her Majesty the Queen in Right of Australia 2011

Abstract Feather keratin has been widely studied for use oriented composite material [8, 9] of microfibrils within an
as a bio-based material. In this paper, we dissolve feather amorphous matrix [10], containing b-sheet crystallites and
keratin using industrial sodium sulfide to investigate the is highly crosslinked by virtue of 7 mol% cysteine [11].
yield, dissolved keratin characteristics, and properties of Primary chains are 10.2–10.4 kDa linear molecules con-
regenerated products to assess the potential of using sodium taining a high number of hydrophobic residues [11, 12].
sulfide as a means of converting waste feathers into a bio- Feather keratin has mechanical strength similar to wool,
polymer. Optimal conditions appeared to require short and is likely the most abundant form of hard keratin in
incubation times in order to give maximum strength in the nature [7, 13]. An estimated 5 million tonnes of feathers
regenerated product. This limits the yield to approximately are produced annually from a reliable production pipeline
55%. Air-dried films and acid-precipitated samples are all as a by-product of chicken meat farming [14], making it a
readily re-crosslinked, suggesting the re-crosslinking pro- dependable bio-polymer feedstock. Despite its properties,
cess is robust. Minimizing exposure to the highly alkaline feather keratin is usually disposed of as waste or rendered
conditions appears favorable to final product strength into a low-grade feed stock [7].
through minimizing alkaline chain damage. The b-sheet One appealing end-use is to use the keratin as a natural
structure of the parent keratin is largely maintained. The monomer for use in eco-composites and bio-plastics [15].
regenerated keratin was shown to have potentially attractive For this, the challenge is to develop a technique for dis-
physical properties for use as a bio-polymer. solving the keratin to produce solutions that are capable of
re-crosslinking, that have minimal degradation of the pri-
Keywords Feather keratin
Dissolution
mary protein chains, and that is industrially scalable. A
Sodium sulfide
Keratin films
Regenerated protein
wide number of techniques are available for dissolving
Mechanical properties hard keratin, often using reduction or oxidation reactions
[16], and more recently, ionic liquids [17–19]. The use of
2-mercaptoethanol as a supplier of thiols is known to
Introduction cleave the disulfide bonds without damaging the primary
chain [20].
Feather keratin has been widely studied for use as a bio- In an optimization of the 2-mercaptoethanol method,
based material [1–7]. It is a tough, high modulus, axially Schrooyen et al. [21–23] used a tenfold molar excess of
2-mercaptoethanol over the disulfide bond concentration in
conjunction with 8 M urea, 3 mM EDTA, 200 mM Tris,
A. J. Poole
J. S. Church (&)
CSIRO Materials Science and Engineering, PO Box 21, pH 9, and 40 °C to achieve a yield up to 75% in 30 min.
Belmont, VIC 3216, Australia The urea acted to disrupt hydrogen bonding and swell the
e-mail: [Link]@[Link] protein chains, giving better thiol ion penetration and an
increased reaction rate. Reagents were removed by dialysis
R. E. Lyons
CSIRO Livestock Industries, 306 Carmody Rd, St. Lucia, and re-crosslinking was prevented by the addition of
QLD 4067, Australia sodium dodecyl sulfate (SDS) to the extracted keratins,

123
996 J Polym Environ (2011) 19:995–1004

with 100–200% additions (on weight of extracted keratin) feather keratin using industrial grade sodium sulfide to
giving SDS-keratin complexes comparable in size to the investigate the yield, dissolved keratin characteristics, and
keratin primary chain [22]. The dissolved keratin was properties of regenerated products to assess the potential of
shown to readily re-crosslink in air to form solution cast using sodium sulfide as a means of converting waste
films. Other workers added SDS to the dissolution mixture feathers into a bio-polymer.
when dissolving feather or wool keratin [24–27] at
approximately 85% (on weight of native keratin) to
accelerate the extraction rate [24], although Schrooyen
Experimental
et al. [22] found no improvement from adding 140% SDS
(on weight of native keratin).
Materials
While the 2-mercaptoethanol method is a benchmark for
good yield and undamaged keratin, the high cost of
Feather Dissolution
2-mercaptoethanol means it is not industrially viable. A
potentially simpler and much cheaper scheme using a
Feathers were collected fresh and still wet from a local
single industrial chemical is to use sodium sulfide [28].
chicken slaughterhouse. They were washed in a horizontal
However, the reaction of sulfide with keratin is complex.
drum fitted with tines and rotating at 10 rpm. The washing
The sodium sulfide reacts in water (Scheme 1) to form
process consisted of rinsing in water at 60 °C, washing in
inorganic thiol (hydrosulfide) and hydroxyl ions [29],
0.4% aqueous Baymol, a non-ionic surfactant (Bayer,
giving the reaction mixture a high pH.
Australia), at 60 °C for 2 h, and then rinsing again at
The hydrosulfide anion reduces the protein disulfide
60 °C. After draining, the feathers were sealed in cotton
bond according to reaction Scheme 2. A mixture of the
bags, spun to remove excess water using a commercial
protonated and deprotonated forms of both of these highly
washing machine, and then dried in a commercial tumble
reactive amino acid residues, cysteine and perthiocysteine,
drier at 70 °C for 3 h. The dried feathers were milled
would be present. The effect of the high pH is twofold.
using a Wiley mill with 3.175 mm diameter die-plate
It acts to disrupt hydrogen bonding and disaggregate
holes.
the protein similar to urea in the 2-mercaptoethanol
The milled feather was digested by placing 50 g feather
scheme. In addition, the hydroxide ions can reduce the
in a 2 L conical flask with 500 mL of Na2S solution
disulfide bonds forming dehydroalanine by b-elimination
(technical grade, Ajax), filling the vessel headspace with
(Scheme 3) [30].
N2 gas, and incubating at 130 rpm and 30 °C in a shaker
The perthiocysteine residues produced by reaction
incubator; Na2S concentration and incubation time were
Schemes 2 and 3 decompose to form cysteine and sulphur.
experimental variables. The mixture was then centrifuged
Dehydroalanine residues are highly reactive and readily
at 1,0009 g for 10 min to isolate the supernatant from the
form cross-links with the amino acid side chains of cys-
un-dissolved solids. Additional dissolution experiments
teine and lysine to form lanthionine (Scheme 4) and ly-
carried out included the use of sodium dodecyl sulfate
sinoalanine (Scheme 5). These reactions have the potential
(approx. 95% purity, Sigma) and urea (analytical grade,
to provide a mechanism for cross-linking the proteins and
Ajax). A digestion run was also carried out in D2O
thus potentially improving the physical properties of the
(99.9 atom%, Aldrich). A reaction was carried out mim-
end product. However, the strong reducing conditions
icking the digestion process in which elemental sulfur
generated by sodium sulfide have the potential to damage
(precipitated, May & Baker Ltd., Dagenham, England) was
the protein backbone [31], although Jones and Meecham
used instead of feather keratin.
[28] supplied some evidence that they managed to cleave
the crosslinks without causing substantial damage to the
protein chain. Protein Films
A feather keratin dissolution scheme based on sodium
sulfide could show promise for industrial application pro- Films were cast from the supernatant in 140 mm diameter
vided that sufficient yield can be obtained without alkali- plastic Petri dishes that had been coated with a silicone
damage to the primary chain. In this paper, we dissolve based release agent. A heavy coating of the release agent
was applied to the surface of the Petri dish, buffed with
lint-free tissue until there was no visible trace of agent,
washed with tap water, and buffed dry. A 40 mL aliquot of
Na2S + H2O 2Na+ + HS- + OH-
keratin solution was added to the Petri dish, which was
Scheme 1 The formation of hydrosulfide and hydroxide ion from placed on a level surface inside a fume hood and dried
sodium sulfide under a constant stream of fresh air at room temperature.

123
J Polym Environ (2011) 19:995–1004 997

R-CH2-S-S-CH2-R + HS- R-CH2-SH + R-CH2-S-S-

Scheme 2 The reduction of disulfide bonds in cystine by hydrosulfide ion

C O NH C O NH

HC CH 2 S S CH 2 CH + OH - C CH 2 + H2N (CH2 ) 4 CH

NH O C NH O C

cystine
dehydroalanine lysine

C O NH C O NH
C CH 2 -
+ S S CH 2 CH + H 2O HC CH 2 NH (CH2 ) 4 CH
NH O C
NH O C

dehydroalanine perthiocysteine
lysinoalanine
Scheme 3 The formation of dehydroalanine by reduction of cystine
Scheme 5 The formation of lysinoalanine by addition of lysine to
dehydroalanine

C O NH Characterization

C CH 2 + HS CH 2 CH
Protein Analysis
NH O C
Protein concentration in the supernatant was determined
spectrophotometrically at 280 nm (Shimadzu UV-2550)
dehydroalanine cysteine
for samples diluted 1:100 with ultrapure water. A standard
curve of absorbance as a function of protein concentration
C O NH was determined gravimetrically by acid precipitating,
water rinsing, and then drying samples of known absor-
HC CH 2 S CH 2 CH bance overnight at 105 °C. It was found that an absorbance
of 0.1 at 280 nm equated to 8.74 g L-1 of protein.
NH O C
SDS-PAGE samples were incubated at 95 °C for 10 min
under reducing conditions (5% (v/v) 2-mercaptoethanol)
lanthionine and analyzed on 10% (w/v) Bis–Tris gels (Invitrogen).
Scheme 4 The formation of lanthionine by addition of cysteine to
Gels were visualized with Coomassie Brilliant Blue solu-
dehydroalanine tion (0.1% (w/v) and scanned using a Typhoon 9410
scanner (Amersham Biosciences). Data was processed
using ImageQuant image analysis software, version 5.2
When dry, films were water insoluble. Residual Na2S was (Molecular Dynamics, Amersham Biosciences) to deter-
readily removed by soaking for 2 h in either tap water or in mine molecular weights of bands and their relative con-
1% HCl, as shown by testing for thiol with lead-acetate centrations within each sample.
paper. Caution must be taken as lowering the pH has the
potential to release H2S gas. For films cast with SDS, the Mechanical Testing
residual detergent was removed by soaking overnight in
60% aqueous acetone saturated with KCl, adapted from the Dog bone shaped test coupons, 4 mm wide in the test
method of Lundgren [32]. Moisture content was measured region, were cut using a knife press (Brooks Pneumatic
from the wet state using standard conditions of 65% rh, Tool & Co, Melbourne, Vic) from films that were softened
21 °C. by soaking in water for 30 min. Ten coupons were cut from

123
998 J Polym Environ (2011) 19:995–1004

each film; 5 were tested in the conditioned state (65% rh, SEM Analysis
21 °C) and 5 were tested wet. For wet testing, the coupons
were soaked for 1 h in ultrapure water at 21 °C and tested Secondary electron images were obtained at low voltage
within 2 min of removal. Values reported are averages using a Schottky emission variable pressure scanning
from duplicate films, although breakages due to brittleness electron microscope (Hitachi, Japan). Prior to imaging the
reduced the number of coupons for testing in some samples were coated with 4 nm of chromium using a
instances. Coupons were stretched to breaking using an Dynavac Xenosput 2000.
Instron 5500 (Instron, Switzerland) fitted with a 100 N load
cell. The gauge length was 19 mm and the rate of extension
was 5 mm min-1. Film thickness was measured using Results and Discussion
digital calipers, taking an average of 10 readings from
along the length of the sample. Tensile strength, Young’s Feather Keratin Dissolution
modulus and extension at break were calculated using
Series IX Automated Materials Testing System software The cleaned and milled feather keratin was dissolved in
(Instron, version 8.10.00) and are reported as the means aqueous sodium sulfide solutions ranging in concentration
from the ten coupons tested. Statistical analysis (ANOVA from 0 to 50 g L-1. As the results in Fig. 1 show, the
and multiple comparison testing carried out at the 95% reaction rate in the first 30 min was rapid and dependent on
level of confidence) was carried out using Matlab (version the Na2S concentration. After the first 1 h of incubation,
R2010a, Mathworks) and the Statistics Toolbox. there was no difference between the 10 and 50 g L-1
treatments, both concentrations giving a maximum dis-
Raman Analysis solved keratin yield of 62% after 24 h. This value was only
increased 3% by extending the incubation to 48 h (result
Raman spectra were obtained at a resolution of 4 cm-1 not shown). This yield is lower than the 2-mercaptoethanol
using a RFS-100 FT-Raman spectrometer (Bruker) equip- scheme [21] of around 75%, but is achieved with a sig-
ped with an Nd:YAG laser (Adlas) operating at 1,064 nm nificantly lower level of thiol, there being no advantage in
and a liquid nitrogen cooled Germanium diode detector. going over a 3.5 fold excess—10 g L-1 Na2S providing
Raman data acquisition was performed using OPUS soft- 128 mmol thiol compared to the 36 mmol disulfide groups
ware v3.1 (Bruker). All data was collected using 180° found in feathers [21]. By comparison, the 2-mercap-
backscatter geometry. Aqueous samples were held in a toethanol scheme used a tenfold excess.
5 mm path length quartz cuvette with a mirrored backing As previous workers had found the addition of urea and
while precipitated and dried solids were packed into a 2 mm SDS to be beneficial [24], this was evaluated with disso-
cavity cell. Spectra were obtained by co-adding 512 scans lution solutions containing 10 g L-1 Na2S, 9 M urea and
collected using a laser power of 750 mW. If there were no 10 g L-1 SDS. The results in Fig. 2 show that urea
time restraints on data collection due to incubation, samples increased both the initial rate and the final yield. When
were analyzed in triplicate and averaged to produce the final added with urea, SDS had little effect after initially
spectra used for analysis. A Blackman–Harris 4-term apo- increasing the reaction rate, which is in agreement with
dization function was used. All spectral data manipulation Schrooyen’s [22, 23] findings that SDS gave no advantage
was carried out using Grams AI software v7.02. after 1 h incubation time. When added without urea the
For Raman analysis of the reaction mixtures, aliquots
were taken from the digestion vessel and spun down to 70 Na 2 S (g/L)
remove solids and thus stop the reaction. The clear 10 & 50
60
supernatant was decanted and used for the analysis. For
Raman analysis of solids precipitated from solubilized 50
7.5
Yield (%)

protein solutions the supernatant described above was 40

5
termed Sample A. This supernatant was centrifuged again 30
at higher force (10,0009 g for 10 min) and the highly 1
20
clarified supernatant was termed Sample B and the pellet
was termed Sample C. Aliquots of the Sample A super- 10
0
natant were freeze-dried after shell-freezing. Additional 0
0 5 10 15 20 25
aliquots were acid precipitated at pH 4 by drop-wise
Solubilisation Time (h)
addition of 1% HCl with vigorous stirring. Again, caution
must be taken as lowering the pH has the potential to Fig. 1 Solubilization of feather keratin as a function of time for
release H2S gas. different Na2S concentrations at 30 °C

123
J Polym Environ (2011) 19:995–1004 999

65
Urea+SDS
60 Urea

55 SDS
Yield (%)

Control
50

45

40

35
0 2 4 6 8
Incubation Time (h)

Fig. 2 Effect of added SDS and urea on protein dissolution using

10 g/L sodium sulfide

reaction proceeded more quickly until 3 h incubation time,

after which SDS had no effect on either rate or yield. The
initial differences in yield % observed between the
10 g L-1 dissolution shown in Fig. 1 and the control dis-
solution shown in Fig. 2 reflects the inherent variability in
the extraction process. Fig. 3 SDS-PAGE gel of dissolved keratin as a function of
All solutions containing Na2S in the first two experi- dissolution time using 10 g/L Na2S. Lanes are marked with the
ments were found to have a pH of 14; the high pH dis- dissolution time in hours; lane M is the molecular weight marker
standard. Numbers on the right represent distinct protein bands as
rupting hydrogen bonding to improve thiol access to the detected by ImageQuant image analysis software
interior of the protein. The improved yield with urea shows
access can be further improved beyond that achieved with subtraction spectra obtained for the reaction are shown in
high pH alone. Fig. 4.
The extent of damage caused by the high pH was While there were no significant changes in the frequency
assessed by analyzing the molecular weight fractions of the of the features, significant spectral intensity changes are
dissolved keratin as a function of reaction time using SDS- observed in the amide I (1,668 cm-1), S–H stretching
PAGE. The results in Fig. 3 show that seven distinct pro- (2,574 cm-1) and S–S stretching (449 cm-1) regions. In
tein bands were detected with little difference between the representation used in Fig. 4, bands that become more
times from 0.5 to 6 h. Longer treatment times show greater negative with time represent species that are decreasing in
smearing of the gel, suggesting chain damage could be concentration while those that increase in intensity with
occurring. Gel quantification shows 68–70% of material is time are increasing in concentration in the solution. A
present as the 10 kDa fraction consistently across the time
range investigated, and 15–19% is in the dimer form. By
24 h, the primary and dimer chains represent 98–99% of
the material. 1003
While the heavy 10 kDa bands could have obscured fine
Amide III
differences, the consistency in molecular mass across the
Relative Intensity

Amide I 449
shorter times and a lack of fragments below the 10 kDa
334
primary chain suggests the primary chain has remained
relatively intact at the shorter treatment times.
Raman spectra were obtained of aqueous keratin dis- 2574 24
solved in 1% Na2S as a function of incubation time with 6
4
treatments ranging from 0.5 to 29 h. The signal to noise
0.5
ratio of these spectra, while sufficient for analysis, was
limited as data collection time was restricted by the fre- 2750 2250 1750 1250 750 250
-1
quency that the reaction mixture was sampled. The spectral Wavenumbers (cm )
features attributable to water and the initial amount of dis-
Fig. 4 The 2,750–250 cm-1 region of the Raman subtraction spectra
solved Na2S were removed by subtraction of the spectrum obtained from the reaction mixture of milled feathers and 10 g/L
obtained from the initial 1% Na2S solution. The Raman Na2S at incubation times ranging from 0.5 to 24 h

123
1000 J Polym Environ (2011) 19:995–1004

graph depicting these changes as a function of dissolution As shown in Figs. 4 and 5, a spectral feature at
time is shown in Fig. 5. 449 cm-1 appears and grows linearly in intensity as a
The amide I band intensity increases rapidly within the function of dissolution time. Based on its frequency, strong
first hour of reaction, with no significant increases occurring intensity and sharpness, this band is likely associated with a
after that time. This trend is consistent with that determined S–S stretching vibration. Weaker features at 992 and
by UV analysis of the protein solutions (Fig. 1) and thus 334 cm-1 also appears to be growing in with increased
validates the method of Raman analysis. The amide I band dissolution time. Bands at the same frequencies were
peak maximum is observed at about 1,668 cm-1 with observed when the digestion was carried out in D2O
shoulders present at 1,660 and 1,680 cm-1. The frequency (results not shown) suggesting that the vibrations are not
of 1,668 cm-1 suggests that the protein has remained associated with protonated species.
largely in the b-sheet conformation [33]. The spectra of While the S–S stretching vibration of cystine is observed
solubilized proteins have amide III band components at near 512 cm-1, the corresponding S–SO3 mode of cys-
1,250 and 1,238 cm-1, corresponding to random coil and teine-S-sulfonate is observed at 407 cm-1 [35]. The S–S
b-sheet protein conformation, respectively [33]. The fre- and S–SO3 modes of cysteine-S-thiosulfonate are observed
quency overlap of b-sheet and random coil components in at 509 and 407 cm-1, respectively [36]. The possibility that
the amide I and III regions makes an accurate analysis very the 449 cm-1 Raman line is associated with –S–SH or –S–
difficult. It appears however that both b-sheet and random S–SH species was tested by bubbling air through the dis-
coil protein conformation is present at all dissolution times. solved protein solution. The S–H stretching vibration of the
The thiol S–H stretching vibration is observed at inorganic thiol anion observed at 2,574 cm-1 was found to
2,574 cm-1. The concentration of this species drops off decrease in intensity and disappeared but no changes were
with time in a somewhat linear fashion, no longer being detected in the 449 cm-1 band intensity.
detectable after 24 h. From comparison to the spectra It is possible that the 449 cm-1 band is associated with
obtained from the 1% sodium sulfide solution (not shown), an inorganic species such as the thiosulfate ion which has
the observed frequency can clearly be associated with that strong S–S stretching modes at 452 and 433 cm-1 in the
of the inorganic ion. For an aqueous solution of cysteine solid [36]. For thiosulfate, the A1 symmetric SO3 stretch-
this vibration is observed at 2,578 cm-1. Under the con- ing vibration is observed at 1,016 and 991 cm-1 and the
ditions of the dissolution, this group is expected to be degenerate (E mode) S–S–O deformation is observed at
ionized and thus not expected to be observed. 343 cm-1 [36]. Both of these Raman modes are moderate
Disulfide bond cleavage is a key part of keratin disso- in intensity. The frequencies and intensities of these
lution. The extent of bond cleavage can be monitored using vibrational modes are consistent with those observed in the
Raman spectroscopy as the S–S stretching vibration asso- keratin dissolution time series spectra. As the digestions are
ciated with cystine is observed as a well defined band carried out under N2, the formation of oxidized sulfur
complex near 512 cm-1 in the spectrum of the undigested species would however be limited by the level of O2
feathers [34]. In the spectra obtained from the keratin dissolved in the water.
protein solutions no clear evidence of this disulfide vibra- Elemental sulfur, S8, exhibits a strong Raman band at
tion is observed suggesting that a large portion of the 473 cm-1 with additional strong bands present at 217 and
disulfide bonds in the soluble protein have been cleaved. 152 cm-1 (spectrum not shown). While elemental sulfur is
not soluble in water, polysulfide salts including sodium
tetrasulfide are very soluble. Polysulfides are formed by the
100
[Protein, amide I] reaction of elemental sulfur and sodium sulfide, both of
% Change Peak Height

which are present during keratin dissolution. The com-

75 pound Na2S4, which has been shown to exist in solution as
[-S-S- 449cm-1]
a mixture of S42-, S52- and S2- species, exhibits a strong
50 sharp Raman line near 446 cm-1 [37]. This line has been
assigned to the stretching vibration of the central S–S
25 bonds. The stretching modes of the external S–S bonds are
[-SH 2574cm-1] observed as a weak shoulder at 482 cm-1. Considering the
0 weak nature of this band and the high S/N of our spectra in
0 5 10 15 20 25 30 this region it is not likely that this shoulder would be
Incubation Time (h) observed. When elemental sulfur was treated with sodium
sulfide solution for 24 h, the Raman spectrum obtained
Fig. 5 Graph of the % change in amide I, S–H stretching, S–S
stretching peak intensities as a function of incubation time for the from the clear supernatant exhibited a strong feature
solubilization reaction at 450 cm-1 as well as weak features at 997, 493 and

123
J Polym Environ (2011) 19:995–1004 1001

338 cm-1. These spectral features are in excellent agree- drop in peak area of the order of 25% can be estimated. A
ment with those observed in the spectra obtained from the very notable and related difference observed in the spec-
dissolution time series, confirming the presence of trum obtained from the pellet is the presence of bands
polysulfides. associated with elemental sulfur. No evidence of oxidized
sulfur species can be detected in any of the Raman spectra
Analysis of Isolated Protein Fractions collected from the samples in this series. The inorganic
thiol anion band (region not shown) and the 449 cm-1
In order to better understand the solubilization process, band, both of which are easily detected in spectra obtained
Raman spectra were obtained from protein solids obtained from the solubilized protein solutions, are no longer pres-
by acid-precipitating solutions after centrifuging the reac- ent. This suggests that they are either being destroyed or
tion mixture at 1,0009 g or 10,0009 g (Samples A and B, removed during the precipitation and clean-up procedures.
respectively) as well as the pellet (Sample C) of finely The major bands observed in spectra obtained from
divided material obtained by re-centrifuging Sample A at similar samples prepared from a 6 h feather keratin
10,0009 g. Spectra obtained from the dried proteins iso- digestion were found to be in good agreement with those
lated from a sample taken after 0.5 h of dissolution are observed for the corresponding 0.5 h samples. Significant
presented in Fig. 6. For comparison purposes the spectrum differences were however observed in the bands associated
from the milled feather keratin is also shown. with functional groups involving sulfur atoms. The peak
In general, the spectra obtained from the two precipi- area of the disulfide bond stretching vibration has further
tated samples, A and B are very similar to that obtained decreased in the precipitated samples and there is a sig-
from the milled feather sample (MF). Any spectral differ- nificant increase in the amount of elemental sulfur present
ences can be considered to be within the noise levels of the in the pellet.
data. Significant changes however are evident when these The presence of sulfur in the spectra obtained from
spectra are compared to the spectrum obtained from the the pellets suggests that dehydroalanine is being formed
pellet, Sample C. The amide I and III regions of the spectra according to reaction Scheme 1 and is immediately being
obtained from the precipitated fractions (A and B) are in destroyed in the formation of cross-links (Schemes 3 and
very good agreement with those observed for the milled 4). No spectroscopic evidence of the formation of these
feather suggesting that the protein conformation is largely cross-links, increased secondary amine functionality in the
b-sheet. case of lysinoalanine and increased C–S bonding (643 and
Unlike in the solution spectra (Fig. 4), the presence of 621 cm-1) for lanthionine formation, is detected as the
protein disulfide bonds are clearly observed in the spectra change in band intensity would be small compared to that
obtained from the precipitated solids suggesting that these of similar functionality that is already present in the pro-
bonds have reformed. Based on disulfide peak areas, there tein. The formation of these crosslinks, even in a minor
is a 10–15% drop in bond density for the precipitated amount, is however know to markedly decrease the solu-
samples compared to the milled feather. For the pellet, a bility of fibrous proteins [30]. Considering the drop in
number of disulfide bonds present in this fraction, the
formation of lysinoalanine and lanthionine crosslinks are
S highly probable.
Amide I
S
Relative Intensity

Feather Keratin Films

C -S-S- Mechanical properties such as tensile strength are important

B when considering a material for advanced applications. To
this end, the effect of dissolution time was investigated.
A
Keratin solutions were cast into films and air-dried to spon-
MF
taneously re-crosslink. The resultant films were pale yellow
and optically clear. An SEM image obtained from the surface
1700 1450 1200 950 700 450 200
of a typical film is shown as Fig. 7. The surface is homoge-
Wavenumbers (cm-1)
neous with a roughness on the 0.1 lm level. The small white
Fig. 6 The 1,800–200 cm-1 region of the Raman spectra obtained sub-micron specks are debris from cutting the film.
from dried solids precipitated from a 0.5 h dissolution (10 g/L Na2S) The moisture content of the films after conditioning was
after spinning for 10 min at 91,000 g (Sample A) and 910,000
approximately 13%. As reported elsewhere [34], Raman
g (Sample B) and the solid pellet obtained after 910,000 g (Sample
C). The spectrum obtained from milled feather keratin is shown as analysis showed the films were isotropic with b-sheet
trace MF conformation and a disulfide bond density similar to the

123
1002 J Polym Environ (2011) 19:995–1004

digestion (95% confidence limit) or 1 h digestion with SDS

(85% confidence limit). Measured in the wet state, the
Young’s moduli of films from the 0.5 and 1 h digestions
and the 1 h digestion with SDS were not statistically dif-
ferent. These samples did, however, have a statistically
higher Young’s modulus than films cast from the 6 and
24 h digestions. No differences in mechanical properties
were found between films before and after soaking to
remove residual sulfide.
The film produced from feathers digested for 1 h had
superior mechanical properties to the best film reported by
Schrooyen et al. [22] (Film 2-ME, Table 1). This was an
optimized film that had 52% of the cysteine residues cap-
ped using iodoacetic acid and contained 0.15 g glycerol/g
of keratin as plasticizer. Moore et al. [39] produced films
Fig. 7 An SEM image obtained from the surface of a typical film cast without capped cysteine residues using an extraction
from a 1 h dissolution using 10 g/L Na2S and soaked to remove
residual sulfide method based on those of Schrooyen et al. [21] and
Yamauchi et al. [24]. Their method incorporated 85% SDS
parent material, similar to the Raman findings for the (on weight of raw keratin) in the reaction mixture, a 1 h
precipitated protein material presented above. extraction time, and dialysis against distilled water to
In general, physical testing showed films with optimal remove reagents (Film 2-ME ? SDS, Table 1). The Na2S
mechanical properties were formed from short incubation films produced in the current study and those of Schrooyen
times (Table 1). et al. [22] were superior to the films of Moore et al. [39].
The Young’s modulus of the conditioned films from the The film strength (conditioned state) of the 1 h digested
0.5 and 1 h digestions are not statistically different. How- feathers is similar to that of industrial polyester resin [40–
ever, film cast from the 1 h digestion has significantly 42], but is weaker than native feather keratin, particularly
higher Young’s modulus than films from either the 6 h in the wet state, despite retaining b-sheet conformation and
re-forming crosslinks to a similar density as the parent.
This probably reflects the isotropic nature of the cast
Table 1 Physical testing results of keratin films material compared to the parent which has longitudinally
aligned microfibrils. Polymer drawing is known to increase
Sample Young’s Tensile Extension at
modulus (MPa) strength break (%) strength [43], demonstrating that the elastic modulus of
(MPa) isotropic polyester is increased from 2 to just under 17 by
Conditioned/ Conditioned/ Conditioned/ drawing, which reduces chain entanglement while
wet wet wet increasing chain alignment and crystallinity [44]. Stretch-
Films 0.5 h 1,241(182)/ 46(9)/8(2) 7(3)/8(3) ing of the films cast from the 1 h digested keratin was
121(39) shown to marginally increase alignment in the stretch
Films 1.0 h 1,568(271)/ 61(7)/6(2) 7(2)/10(3) direction [34], demonstrating alignment was possible.
95(46)
Films 6 h 990(499)/ 37(21)/3(1) 5(2)/17(4) General Discussion
23(10)
Films 24 h NR/17(8) NR/2(1) NR/15(3) Optimal conditions appeared to require short incubation
Films 1.0 h with 1,225(230)/ 48(8)/5(1) 5(1)/15(6) times in order to give maximum strength in the regenerated
SDS 73(43)
product. This limits the yield to approximately 55%. Even
Feather [38] 2,580/1,470 130/106 10.4/16.3
if incubation time is extended, yield is not significantly
Films 2-ME 1,344/– 30/– 3/–
[22]
improved. It is not known what factors prevented maxi-
mum dissolution from progressing substantially further.
Films 10/– 17/– 2/–
2-ME ? SDS Other workers also report maximum dissolution well below
[39] 100% for feather [21, 23] and wool [24, 31]. Using Na2S to
Polyester resin 4,000/– 45–85/– 2/– solublise wool, Happey and Wormell [31] were limited to a
[40–42] yield of 65% despite using a much higher concentration
SD are given in the brackets, 2-ME 2-mercaptoethanol, NR no result and long time frame (250 g L-1 Na2S, 24 h at 25 °C). In
due to brittleness and fragility of samples, – no result reported contrast to feather keratin, wool proteins are heterogeneous

123
J Polym Environ (2011) 19:995–1004 1003

with a generally higher molecular weight (10–55 kDa [45], could be readily regenerated through air-oxidation. The
higher cysteine content (10.5 mol% for whole merino fiber quality of dissolved keratin was controlled via the disso-
[46] with some sulfur-rich components containing lution time; with shorter times producing stronger regen-
12–41 mol% [47]). These factors could be expected to erated products. At these reaction times, alkaline chain
cause some components to dissolve faster than others. damage did not appear to be a significant problem.
Feather keratin protein is not only homogenous, but the Regenerated keratin films had superior mechanical prop-
precipitated dissolved protein and native keratins appear erties to films from a 2-mercaptoethanol digestion while
similar in structure based on the Raman analysis (Fig. 6), using a simpler scheme without requirement for dialysis.
and an excess of inorganic thiol was still present even after The films demonstrated that the dissolved keratin had
the dissolution had appreciably slowed (Fig. 5). A con- potentially attractive properties for use as a bio-polymer.
tributing factor to limited dissolution may be the difficulty
of the reductant gaining access to the interior of some Acknowledgments This work was funded by CSIRO Materials
Science and Engineering. The technical assistance of Ms. Andrea
larger keratin particles, as it was noticed the reaction Woodhead, Ms. Lisa O’Brien, Mr. Colin Veitch and Ms. Jacinta
mixture often contained undissolved particles of quill. Poole is greatly appreciated.
Air-dried films and acid-precipitated samples are all
readily re-crosslinked, suggesting the re-crosslinking pro-
cess is robust. Raman analysis revealed disulfide levels
were only 10–15% lower than the native material for short References
incubation (0.5 h) acid-precipitated samples, but samples
1. Barone J (2009) J Polym Environ 17:143–151
with longer incubation (6 h) had decreasing disulfide 2. Barone J, Schmidt W (2005) Compos Sci Technol 65:173–181
content. Analysis of the finely divided particulate matter 3. Bullions T, Hoffman D, Gillespie R, Price-O’Brien J, Loos A
isolated from the reaction mixture (Sample C) showed 25% (2006) Compos Sci Technol 66:102–114
4. Dweib MA, Hu B, O’Donnell A, Shenton HW, Wool RP (2004)
lower disulfide content than the bulk material as well as the
Compos Struct 63:147–157
presence of elemental sulfur, with the disulfide level further 5. Hong CK, Wool RP (2005) J Appl Polym Sci 95:1524–1538
decreasing and elemental sulfur level increasing with 6. Martı́nez-Hernández A, Velasco-Santos C, de Icaza M, Castaño
longer incubation. Considering its low solubility and low V (2005) Polymer 46:8233–8238
7. Reddy N, Yang Y (2007) J Polym Environ 15:81–87
disulfide crosslink content it can be postulated that this
8. Bonser RHC (1996) J Zool (Lond) 239:477–484
finely divided material is the result of re-crosslinking 9. Cameron GJ, Weiss TJ, Bonser RHC (2003) J Struct Biol 143:
occurring via a dehydroalanine route to form lanthionine 118–123
and/or lysinoalanine during the incubation. These cross- 10. Filshie BK, Rogers GE (1962) J Cell Biol 13:1–12
11. Arai K, Takahashi R, Yokote Y, Akahane Y (1983) Eur J Bio-
linking reactions are not prevented by the presence of
chem 132:501–507
reducing agent or high pH. 12. Fraser R, MacRae T, Parry D, Suzuki E (1971) Polymer 12:35–56
The formation of these agglomerates is likely to be 13. Onifade A, Al-Sane N, Al-Musallam A, Al-Zarban S (1998)
detrimental to the properties of a regenerated product as Bioresour Technol 66:1–11
14. Poole AJ, Church JS, Huson MG (2009) Biomacromolecules 10:
they were water-insoluble and so will not likely be cross-
1–9
linked into the larger structure. Within the constraints of 15. Pillai CKS (2010) Des Monomers Polym 13:87–121
the SDS-PAGE method, alkaline chain damage did not 16. Crewther WG, Fraser RDB, Lennox FG, Lindley H (1965) The
appear to occur within the first 6 h of the incubation as chemistry of keratins. In: Anfinsen CB Jr, Anson ML, Edsall JT,
Richards FM (eds) Advances in protein chemistry, vol 20.
there was no significant fragmentation of the primary
Academic Press, New York, p 191
chains. The loss in physical properties observed over this 17. Xie H, Li S, Zhang S (2005) Green Chem 7:606–608
incubation period (Table 1) can therefore not likely be 18. Sun P, Liu Z-T, Liu Z-W (2009) J Hazard Mater 170:786–790
attributable to the degradation of the primary chains. An 19. Zhao L, Tang YX, Zhao RF, Mao WK, Chen S, Hua J (2010)
Wool Text J 38:1–5
increased level of agglomerates in the films could however
20. Anfinsen CB, Haber E (1961) J Biol Chem 236:1361–1363
be one possible explanation for this drop-off. Therefore, 21. Schrooyen P, Dijkstra P, Oberthür R, Bantjes A, Feijen J (2000)
minimizing exposure to the highly alkaline conditions J Agric Food Chem 48:4326–4344
appears favorable to final product strength. 22. Schrooyen P, Dijkstra P, Oberthür R, Bantjes A, Feijen J (2001)
J Agric Food Chem 49:221–230
23. Schrooyen P, Dijkstra P, Oberthür R, Bantjes A, Feijen J (2001)
J Colloid Interface Sci 240:30–39
Conclusions 24. Yamauchi K, Yamauchi A, Kusunoki T, Kohda A, Konishi Y
(1996) J Biomed Mater Res Part A 31:439–444
25. Tanabe T, Okitsu N, Yamauchi K (2004) Mater Sci Eng C
Sodium sulfide dissolved feather keratin in a simple
24:441–446
scheme under ambient conditions, producing dissolved 26. Martelli S, Moore G, Paes S, Gandolfo C, Laurindo B (2006)
keratin that largely retains a b-sheet conformation and LWT—Food Sci Technol 39:292–301

123
1004 J Polym Environ (2011) 19:995–1004

27. Martelli SM, Moore GRP, Laurindo JB (2006) J Polym Environ 40. Acar M, Harper JF (2000) Comput Struct 76:105–114
14:215–222 41. EastCoastFibreglassSupplies, Guide to glass reinforced plastics,
28. Jones C, Mecham D (1943) Arch Biochem 2:209–223 [Link]
29. Sienko M, Plane R (1961) Chemistry. McGraw-Hill Book aspx. Accessed 16 Nov 2010
Company, New York 42. EcoTek, EcoTek product selection guide, [Link]
30. Maclaren J, Milligan B (1981) In: Wool science: the chemical com/images/uploads/lit_ecoteck_greenresins.pdf. Accessed 28
reactivity of the wool fibre. Science Press, Marrickville, Nov 2010
pp 99–101 43. Göschel U (2002) Structure development and mechanical
31. Happey F, Wormell R (1949) J Text Inst 40:T855–T869 behavior during uniaxial drawing of PET. In: Fakirov S (ed)
32. Lundgren H (1945) Text Res J 15:335–353 Handbook of thermoplastic polyesters, vol 1. Wiley, Weinheim,
33. Frushour BG, Koenig JL (1975) Raman spectroscopy of proteins. p 289
In: Clark RSH, Hester RE (eds) Advances in infrared and Raman 44. Gupta VB, Bashir Z (2002) PET fibers, films and bottles. In:
spectroscopy, vol 1. Heyden, New York, p 35 Fakirov S (ed) Handbook of thermoplastic polyesters. Wiley,
34. Church JS, Poole AJ, Woodhead AL (2010) Vib Spectrosc 53: Weinheim, p 317
107–111 45. Gillespie JM (1990) The Proteins of Hair and Other Hard
35. Church JS, Millington KR (1996) Biospectroscopy 2:249–258 a-Keratins. In: Goldman RD, Steinert PM (eds) Cellular and
36. Church JS, Evans DJ (2008) Spectrochim Acta Part A 69: molecular biology of intermediate filaments. Plenum Press, New
256–262 York, pp 95–128
37. Janz GJ, Downey JR, Roduner E, Wasilczyk GJ, Coutts JW, 46. Bradbury JH, Chapman G , King NLR (1967) Chemical com-
Eluard A (1976) Inorg Chem 15:1759–1763 position of the histological components of wool. In: Crewther
38. Taylor AM, Bonser RHC, Farrent JW (2004) J Mater Sci 39: WG (ed) Symposium on Fibrous Proteins Australia 1967. But-
939–942 terworths, Sydney, pp 368–372
39. Moore GRP, Martelli SM, Gandolfo C, do Amaral Sobral PJ, 47. Plowman JE (2003) J Chromatogr B 787:63–76
Laurindo JB (2006) Food Hydrocoll 20:975–982

123

View publication stats