Enhanced enzymatic hydrolysis of langostino shell chitin with mixtures of enzymes from bacterial and fungal sources

Article abstract of DOI:10.1016/S0008-6215(03)00269-6

A combination of enzyme preparations from Trichoderma atroviride and Serratia marcescens was able to completely degrade high concentrations (100 g/L) of chitin from langostino crab shells to N-acetylglucosamine (78%), glucosamine (2%), and chitobiose (10%). The result was achieved at 32°C in 12 days with no pre-treatment (size reduction or swelling) of the substrate and without removal of the inhibitory end-products from the mixture. Enzymatic degradation of three forms of chitin by Serratia/Trichoderma and Streptomyces/Trichoderma blends was carried out according to a simplex-lattice mixture design. Fitted polynomial models indicated that there was synergy between prokaryotic and fungal enzymes for both hydrolysis of crab chitin and reduction of turbidity of colloidal chitin (primarily endo-type activity). Prokaryotic/fungal enzymes were not synergistic in degrading chitosan. Enzymes from prokaryotic sources had much lower activity against chitosan than enzymes from T. atroviride.

Full text of DOI:10.1016/S0008-6215(03)00269-6

Carbohydrate Research 338 (2003) 1823ꢀ1833
/
www.elsevier.com/locate/carres
Enhanced enzymatic hydrolysis of langostino shell chitin with
mixtures of enzymes from bacterial and fungal sources
Bruno Giuliano Garisto Donzelli,a,* Gary Ostroff,b Gary E. Harmana
a Department of Horticultural Sciences, Department of Plant Pathology, Cornell University, Geneva, NY 14456, USA
b Biopolymer Engineering, Inc., Eagan, MN 55121, USA
Received 12 August 2002; accepted 10 May 2003
Abstract
A combination of enzyme preparations from Trichoderma atroviride and Serratia marcescens was able to completely degrade high
concentrations (100 g/L) of chitin from langostino crab shells to N-acetylglucosamine (78%), glucosamine (2%), and chitobiose
(10%). The result was achieved at 32 8C in 12 days with no pre-treatment (size reduction or swelling) of the substrate and without
removal of the inhibitory end-products from the mixture. Enzymatic degradation of three forms of chitin by Serratia/Trichoderma
and Streptomyces/Trichoderma blends was carried out according to a simplex-lattice mixture design. Fitted polynomial models
indicated that there was synergy between prokaryotic and fungal enzymes for both hydrolysis of crab chitin and reduction of
turbidity of colloidal chitin (primarily endo-type activity). Prokaryotic/fungal enzymes were not synergistic in degrading chitosan.
Enzymes from prokaryotic sources had much lower activity against chitosan than enzymes from T. atroviride.
# 2003 Elsevier Ltd. All rights reserved.
Keywords: N-Acetylglucosamine; Trichoderma; Serratia; Streptomyces; Hydrolysis; Chitin
1. Introduction
high temperatures and results in production of toxic
wastes. Moreover, the extreme conditions used in the
process may result in unwanted modifications to the
hydrolysis products.
Conversely, enzymatic degradation of crustacean
shells is environmentally friendly but is more complex
since it involves both production of the enzyme and the
digestion of the substrate. More importantly, enzymatic
processes have been plagued by low yields and have
resulted in incomplete conversion of chitin into its
monomer.6,7
N-Acetylglucosamine and glucosamine are amino su-
gars having therapeutic potential for the treatment of a
variety of diseases, including arthritis,1,2 inflammatory
bowel disease,3 and general inflammatory damage.4
Commercial production of these amino sugars currently
relies upon acid hydrolysis of de-proteinized and de-
mineralized crustacean shells.5 Acid hydrolysis is rela-
tively efficient but involves strong acids (4ꢀ8 M HCl) at
/
However, there has been substantial recent progress in
studies of chitinases from both prokaryotic and eukar-
yotic sources.8ꢀ10 Of these, the most widely examined
enzyme sources for commercial chitin degradation are
strains of Serratia marcescens.8 However, its enzyme
preparations do not completely convert chitin to
GlcNAc but instead accumulate a blend of both
monomers and chito-oligomers (mainly chitobiose).11
The phenomenon is particularly evident when high
Abbreviations: GlcNAc, N-acetyl-
D-glucosamine; GlcN,
D
glucosamine; MUA, 4-methylumbelliferyl N-acetyl-b-
-
glucosaminide; MUB, 4-methylumbelliferyl-N,N?-diacetyl-b-
chitobioside; nahase, N-acetylhexosaminidase (EC 3.2.1.52);
endo, Endochitinase (EC 3.2.1.14).
* Corresponding author. Present address: Boyce Thompson
Institute for Plant Research Tower Rd, Ithaca NY 14853-1801,
USA. Fax: ꢁ1-607-2542958.
/
E-mail address: bdd1@cornell.edu (B. Giuliano Garisto
Donzelli).
concentrations of chitin (5ꢀ10% w/v) are digested.
/
0008-6215/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0008-6215(03)00269-6

1824
B. Giuliano Garisto Donzelli et al. / Carbohydrate Research 338 (2003) 1823ꢀ1833
/
Chitinolytic enzymes from fungi in the genus Tricho-
derma have also been extensively studied.12 Unlike their
bacterial counterparts, Trichoderma chitinolytic pre-
parations have a high ratio of exochitinase to endochi-
tinase activity and release almost exclusively monomeric
GlcNAc from chitin. Insofar as we are aware, enzymes
from these fungi have not been examined for their
abilities for commercial or pilot-scale production of
GlcNAc from crustacean chitin, although enzymes from
T. reesei are widely used for commercial hydrolysis of
cellulose and other plant polysaccharides.
This study describes the efficacy and the synergy of
mixtures of enzymes from Trichoderma atroviride, S.
marcescens and Streptomyces albidoflavus in hydrolyz-
ing several forms of chitin, including native chitin from
langostino crab shells, colloidal chitin, and chitosan.
2. Results
2.1. Hydrolysis of langostino crab chitin
Fig. 1. Time-course of N-acetylglucosamine release from
langostino chitin during incubation with S. marcescens and
T. atroviride enzyme blends. (A) Stirred flasks, 32 8C in, 350
rpm; (B) shaken flasks, 37 8C, 220 rpm. Serratia/Trichoderma
filtrate percent ratios: (-"-) 100:0; (-j-) 75:25; (-'-) 50:50; (-
Enzyme preparations from the three organisms differed
markedly in their endochitinase and nahase levels as
measured by methylumbelliferyl substrates. The activity
of the preparation from S. marcescens was several-fold
higher than that from S. albidoflavus or T. atroviride.
Further, the ratios of endochitinase to nahase activity
from preparations from the various microbes differed by
about an order of magnitude with the highest propor-
tions of exochitinase activity in the enzymes from T.
atroviride (Table 1).
m-) 25:75; (-ꢃ-) 0:100. Fifty grams per litre of chitin were
/
initially present in all the digestions; additional 50 g/L were
added at the time indicated by the arrows.
ratio of 75:25 Serratia/Trichoderma (Fig. 1). This
mixture provided about 100 g/L of GlcNAc (measured
As a first step, we examined the release of GlcNAc
from chitin by enzyme blends from the most active
sources, namely, from S. marcescens, and from the
fungus T. atroviride (Fig. 1).
The lowest rate of release with single preparations was
with the enzymes from T. atroviride. Enzymes from S.
marcescens released GlcNAc rapidly through day 2 and
then the rate slowed. After the second addition of chitin
at day 7 the rate of GlcNAc release again increased. The
mixtures of S. marcescens and T. atroviride enzyme
preparations were more effective than any single enzyme
preparation, with the greatest release occurring at a
by Elsonꢀ
added.
/
Morgan assay) from 100 g/L of total chitin
These results largely were confirmed by HPLC
analyses of the reaction mixtures. The 75:25 mixture
released per liter 78 g of GlcNAc, 2 g of GlcN and 10 g
of chitobiose at the termination of the test (Table 2).
Unblended S. marcescens yielded 47 g/L of GlcNAc,
and no detectable GlcN, while 37% of the total detected
amino sugars was chitobiose. Larger oligomers were
never detected, even though the technique was capable
of resolving up to the 7-mer. Digestions carried out with
T. atroviride enzymes alone released primarily GlcNAc
Table 1
Endochitinase (Endo) and N-acetylhexosaminidase (Nahase) activities of chitinolytic enzyme preparations used in this work
Enzyme source
Endo activity (U) a
Nahase activity (U) b
Ratio E/N
Serratia marcescens
Streptomyces albidoflavus
Trichoderma atroviride
822
25
193
23
4.3
1.1
17
39
0.43
a
Uꢂ
/
release of 1 nmol of methylumbelliferone/s mL from 4-methylumbelliferyl-
D
-N,N?-diacetyl-b-chitobioside.
-glucosaminide.
b
Uꢂ
/
release of 1 nmol of methylumbelliferone/s mL from 4-methylumbelliferyl N-acetyl-b-
D

B. Giuliano Garisto Donzelli et al. / Carbohydrate Research 338 (2003) 1823ꢀ
/
1833
1825
Table 2
HPLC quantification of the hydrolysis products released from chitin digestion (stirring flasks) with enzyme mixtures from different
organisms
Blend composition
GlcNAc (g/L)
GlcN (g/L)
GlcNAc2 (g/L)
GlcNAc3 (g/L)
Glucose (g/L)
Total (g/L)
SM 100%
SM:TA 75%:25%
TA 100%
47.10 (62.8%)
77.90 (86.0%)
26.90 (95.5%)
B
/
0.004 (0.0%)
2.17 (2.4%)
0.95 (3.4%)
27.90 (37.2%)
10.50 (11.6%)
0.32 (1.1%)
B
/
0.020 (0.0%)
0.020 (0.0%)
0.020 (0.0%)
B
/
0.005 (0.0%)
0.005 (0.0%)
0.005 (0.0%)
75.00
90.57
28.17
B
/
B
/
B
/
B
/
Analysis was performed by high-performance anion-exchange chromatography with pulsed amperometric detection on a Dionex
DX 500 apparatus. Abbreviations are: SM, Serratia marcescens; TA, Trichoderma atroviride; GlcNAc, N-acetylglucosamine; GlcN,
glucosamine; GlcNAc2, chitobiose; GlcNAc3, chitotriose.
(96% of the total detectable amino sugars); of the
remainder, GlcN was more abundant than chitobiose.
These results were confirmed in other independent
replications. Thin layer chromatography (TLC) analyses
of the reaction mixtures confirmed the results of the
HPLC analysis (data not shown).
prokaryotic/Trichoderma mixtures as compared to any
enzyme used singly (Fig. 2). The interaction terms of the
models explain up to 35% (Serratia/Trichoderma) and
48% (Streptomyces/Trichoderma) of the overall GlcNAc
released (Table 4) with the best theoretical blend being
55:45 Serratia/Trichoderma and 45:55 Streptomyces/
Trichoderma.
Product profiles from these partial digestions were
analyzed by TLC (data not shown) and were consistent
with product composition of the complete hydrolysis.
Neither single nor blended chitinase preparations
yielded chito-oligomers higher than chitotriose. The
presence of Trichoderma chitinases in the mixtures
increased GlcNAc yield relative to Serratia and Strep-
tomyces enzymes used singly.
In order to describe better the Serratia/Trichoderma
system, we repeated the experiment using a two compo-
nent, five-level simplex-lattice mixture design (Table 3).
We also included a second set of digestions with a
different prokaryotic/fungal mixture (Streptomyces/Tri-
choderma). GlcNAc released at 72 h was used to fit
quadratic equations. The models were statistically sig-
nificant (Pꢂ
/
0.0002; adjusted R2ꢀ
0.95) and pointed to
/
the presence of a strong positive interaction in both the
Table 3
Runs and corresponding results of the simplex-lattice mixture designs presented in this paper: GlcNAc release from langostino crab
chitin; turbidity reduction of colloidal chitin; reducing sugar release from chitosan
Run
#
SM
(%)
TA
(%)
GlcNAc from langostino chitin Colloidal chitin turbidity red. Reducing sugars from chitosan (GlcN
(g/L)
(%)
g/L)
1
2
3
4
5
6
7
8
100
100
75
50
50
25
0
0
0
20.9
21.7
30.7
28.9
30.5
27.2
18.2
17.8
63
63
66
66
70
58
25
18
0.034
0.015
0.078
0.135
0.126
0.197
0.252
0.264
25
50
50
75
100
100
0
Run SA (%) TA
#
GlcNAc from langostino chitin Colloidal chitin turbidity red. Reducing sugars from chitosan (GlcN
g/L)
(%)
(g/L)
(%)
1
2
3
4
5
6
7
8
100
100
75
50
50
25
0
0
0
12.3
12.4
25.8
30.1
28.9
27.2
18.3
18.5
12
11
21
23
23
30
25
23
0.158
0.182
0.180
0.231
0.221
0.254
0.271
0.274
25
50
50
75
100
100
0
SM, S. marcescens; SA, S. albidoflavus; TA, T. atroviride.

1826
B. Giuliano Garisto Donzelli et al. / Carbohydrate Research 338 (2003) 1823ꢀ1833
/
between different enzyme sources (Fig. 5). TLC analysis
of the digestion products showed that Trichoderma
chitinolytic preparations accumulate both GlcN and a
range of oligomers, while prokaryotic preps released
mainly the dimer and oligomers larger than the tetramer
(Fig. 6).
2.4. Both Trichoderma and prokaryotic chitinase
preparations are differently inhibited by hexosamines
We hypothesized that a possible explanation for the
prokaryotic/Trichoderma synergism could be related to
lower enzyme inhibition by hexosamines as compared to
the unblended preparations.13 To test this hypothesis,
we measured the decrease in hydrolytic activity against
MUA or MUB of prokaryotic/Trichoderma enzymes
blended according to a simplex-lattice mixture design in
the presence of high concentrations of either GlcNAc or
GlcN. GlcNAc or GlcN was added to give a range of
about five- to 1400-fold excesses compared to the
substrates. The highest concentrations of GlcNAc were
severely inhibitory to nahase activity for all of the
enzymes and enzyme mixtures, but the enzyme mixtures
from T. atroviride or S. albidoflavus were somewhat less
inhibited than those from S. marcescens. Conversely,
the presence of GlcN was more inhibitory to the
enzymes from T. atroviride than from the bacterial
sources. Mixtures of prokaryotic/Trichoderma enzymes
were not significantly interactive in degrading MUA in
the presence of GlcNAc and GlcN compared with single
enzyme sources (Fig. 7).
Fig. 2. Plot of the fitted models estimated for NAG release
from langostino chitin after 72 h of incubation with Serratia/
Trichoderma (m) and Streptomyces/Trichoderma (') mix-
tures. Corresponding equations are also reported. GlcNAc72h
(g/L): GlcNAc released after 72 h of digestion; SA, SM, and
TA represent, respectively the percentages of Streptomyces,
Serratia, and Trichoderma enzyme in the mixtures.
2.2. Prokaryotic and fungal enzyme preps are synergistic
in degrading colloidal chitin
The activity of any of the enzyme mixtures against
MUB was inhibited less by either GlcNAc or GlcN than
was activity against MUA. Generally, the bacterial
sources were less affected by GlcNAc or GlcN than
the enzymes from T. atroviride. Activity of the enzymes
from S. albidoflavus against MUB was increased at
concentrations of GlcNAc above 5 g/L. However, unlike
activity against MUA, blends displayed significant
interaction in the hydrolysis of MUB. Serratia/Tricho-
derma mixtures were less inhibited by GlcN than were
the Streptomyces/Trichoderma blends (Fig. 8).
The ability of Serratia/Trichoderma and Streptomyces/
Trichoderma to hydrolyze colloidal chitin was also
modeled using a similar simplex-lattice mixture design
(Table 3 and Fig. 3). Both Serratia and Streptomyces
enzymatic preparations were synergistic with Tricho-
derma enzymes in reducing turbidity of colloidal chitin
suspensions (significant at Pꢂ
/
0.0002; adjusted R2ꢀ
/
0.94). The interaction between enzyme preparations
was responsible for about 45 and 26% of the total
turbidity reduction for Serratia/Trichoderma and Strep-
tomyces/Trichoderma, respectively (Table 4). The only
reaction product from colloidal chitin detected with
TLC with any enzyme or mixtures of enzymes was
GlcNAc (Fig. 4).
3. Discussion
Chitin, one of the most abundant organic polymers on
Earth, is a heterogeneous polysaccharide with different
crystalline forms (a, b and g) and characterized by
variable levels and patterns of acetylation. In nature,
degradation of all these forms of chitin is carried out
efficiently by the combined activity of the enzymatic
systems of a multitude of microorganisms.
Heretofore, complete enzymatic degradation of com-
plex native chitin has not been reported. Instead, studies
of enzymatic hydrolysis of chitin exploited single
2.3. Trichoderma enzymes are more active on chitosan
than their prokaryotic counterparts
When chitosan was used in another simplex-lattice
design experiment, the quantity of reducing sugars
released by enzymes from T. atroviride was 10- and
1.6-fold higher than Serratia or Streptomyces enzymes,
respectively (significant at Pꢂ
/
0.0001; adjusted R2ꢀ
/
0.94), but the model gave no evidence for synergy

B. Giuliano Garisto Donzelli et al. / Carbohydrate Research 338 (2003) 1823ꢀ
/
1833
1827
Table 4
Weight of both direct (D) and interaction (I) components to the overall hydrolytic activity predicted by the models describing the
release of GlcNAc from langostino shell chitin and the reduction of turbidity in colloidal chitin suspensions for Serratia/
Trichoderma (A) and Streptomyces/Trichoderma (B) mixtures
(A) Serratia/Trichoderma
Mixture composition (%)
Langostino shell (GlcNAc release)
Colloidal chitin (Turbidity Red.)
Serratia
Trichoderma
D
I
D
I
100
90
80
70
60
50
40
30
20
10
0
0
100
85
75
69
66
65
65
68
73
83
100
0
100
93
85
76
68
62
57
55
56
64
100
0
10
20
30
40
50
60
70
80
90
100
15
25
31
34
35
35
32
27
17
0
7
15
24
32
38
43
45
44
36
0
(B) Streptomyces/Trichoderma
Streptomyces
Trichoderma
D
I
D
I
100
90
80
70
60
50
40
30
20
10
0
0
100
72
60
54
52
52
54
58
65
77
100
0
100
85
78
75
74
75
77
80
85
91
100
0
10
20
30
40
50
60
70
80
90
100
28
40
46
48
48
46
42
35
23
0
15
22
25
26
25
23
20
15
9
0
microbial sources and were effective only on easily
degradable forms of the polymer such as b-chitin7 or
on relatively low concentrations of a-chitin.7,14 Typi-
cally, the substrate was either physically (size reduction,
heat-swelling) or chemically pre-treated in order to
increase enzyme action.6,14,15 In other cases, the effi-
ciency of the process was increased by continuously
removing end-products.16
In this work, we chose to examine synergistic mixtures
of enzymes that more closely approximate those that
occur in nature. However, we used organisms whose
complexes of chitinolytic enzymes are well studied,
produce a variety of activity types and possess enzymes
with diverse abilities to interact with the substrate. T.
atroviride strain P1 produces CHIT42, which is an
endochitinase;17 CHIT73, which is an exochitinase,18
and CHIT38, which is a chitobiosidase (which requires
at the least the trimer for activity),12 as well as several
larger molecular weight exochitinases. S. marcescens
produces at least three chitinolytic enzymes, and again
the enzymes and genes are well known.19,20 The
chitinolytic system of S. albidoflavus is not as well
characterized as the other two, but multiple chitinolytic
enzymes are known to be produced and the enzyme
complexes from closely related species have been
thoroughly investigated.9,21
The primary goal of the current research was to
discover enzyme mixtures that were capable of relatively
complete release of hexosamines even from complex
chitin, such as the langostino shells used in this study.
We hypothesized: (a) that blends of enzymes with
different types of activities were more likely to be
effective than single enzymes; and (b) that highly diverse
mixtures were more likely to be effective than enzymes
from a single source. These hypotheses were proven
correct by the current study. We view as unlikely the
discovery of any single enzyme, or even enzymes from a
single organism, that will provide complete degradation
of crustacean chitin.
We demonstrated that it is possible to significantly
increase enzyme-mediated hexosamine production by
exploiting synergism among enzyme preparations ob-

1828
B. Giuliano Garisto Donzelli et al. / Carbohydrate Research 338 (2003) 1823ꢀ1833
/
The origin of synergism showed by both Serratia/
Trichoderma and Streptomyces/Trichoderma mixtures is
unknown but it is likely to derive from the differential
ability of the single preparation in: (a) accessing chitin
crystalline structure; (b) removing impurities such as
proteins and lipids;22 and (c) providing non-chitinolytic
proteins with activity against chitin.15 The enzymes
produced by these organisms, and the genes that encode
them, are dissimilar.23,24,17,18 This difference is especially
pronounced between prokaryotic and eukaryotic
sources.25 Synergism could be due to an increased
efficiency of the blends in accessing/de-crystallizing the
polymer, a feature considered of extreme importance for
the hydrolysis of ligno-cellulosic substrates.26 However,
in our experiments with suspensions of colloidal chitin,
there was also synergy between bacterial and fungal-
source enzymes. Rapid decreases in turbidity are asso-
ciated primarily with endo-type enzymes, indicating that
enzymes of this class from different sources are syner-
gistic in reduction of chain length in the amorphous
substrate. Synergy may also have been enhanced by the
activity of the enzymes from T. atroviride on deacety-
lated regions of the crystalline polymer, since low levels
of glucosamine in chitin molecules might block activity
of prokaryotic enzymes with low chitosanase activity.
Similarly to other Trichoderma isolates, T. atroviride
strain P1 displays high chitosanase activity.27,28
Fig. 3. Plot of the fitted model estimated for turbidity
reduction of colloidal chitin from langostino chitin by
Serratia/Trichoderma (m) and Streptomyces/Trichoderma
(') mixtures. Corresponding equations are also reported.
CC turbidity red (%): percent reduction of turbidity in chitin
suspension; SA, SM, and TA represent, respectively the
percentages of Streptomyces, Serratia, and Trichoderma
enzymes in the mixtures.
tained from taxonomically distant sources. The process
described is simple, since it involves the incubation of
de-proteinized and de-mineralized crab shells in the
presence of an appropriate enzyme mixture with con-
tinuous and vigorous stirring. At the same time, the
process is efficient because it allows the complete
saccharification of langostino crab shells.
Synergism shown by bacterial/fungal mixtures could
not be explained by a relief of hexosamine inhibition.
Only Serratia/Trichoderma blends performed better
Fig. 4. TLC separation of the hydrolysis products from colloidal chitin after incubation with either Serratia/Trichoderma or
Streptomyces/Trichoderma enzyme preparation blends. For each sample Trichoderma component is the complement to the
percentage reported in the figure. Reactions were concentrated 5.5ꢃ in Speedvac and 3 mL/lane were loaded. Oligo: N-
/
acetylglucosamine (50 mg), chitobiose (25 mg), chitotriose (25 mg) and chitotetraose (25 mg). GlcN: glucosamine (50 mg). The image
was digitally enhanced by uniformly replacing the background color (dark yellow) with white color.

B. Giuliano Garisto Donzelli et al. / Carbohydrate Research 338 (2003) 1823ꢀ
/
1833
1829
against MUB in the presence of GlcN, compared to the
unblended preparations.
The hydrolytic process illustrated in this paper can be
further improved. A possible strategy is expanding the
range of organisms contributing to the chitinolytic
blend. Synergism among hydrolases of different origin
seems a general phenomenon and the test of mixtures
including two or more chitinolytic preparations from a
wider array of sources is likely to provide blends with
better performances.29,30
4. Experimental
4.1. Chitinolytic enzyme production
Chitinolytic enzymes from S. marcescens QMB1466
(ATCC 990) used for this work was a prototype
commercial preparation provided by Biopolymer En-
gineering (Eagan, MN).
Culture filtrates from T. atroviride strain P1 (ATCC
74058) were prepared by inoculating 250 mL Erlen-
meyer flasks containing 100 mL of medium with conidia
Fig. 5. Plot of the fitted models estimated for the release of
reducing sugars (expressed as GlcN) from chitosan by
Serratia/Trichoderma (m) and Streptomyces/Trichoderma
(') culture filtrate blends. Corresponding equations are also
reported. RSGlcN (g/L): reducing sugars released after 4 h of
digestion and expressed as GlcN. SA, SM, and TA represent,
respectively, the percentage of Streptomyces, Serratia, and
Trichoderma enzymes in the mixture.
(5ꢃ
106/mL final concentration) harvested from colo-
/
nies grown on potato dextrose agar (Difco Laboratories,
Detroit, MI). The medium contained 15 g/L KH2PO4, 2
g/L (NH4)2HPO4, 1.6 g/L MgSO4, 0.6 g/L CaCl2, 2.5
mg/L FeSO4×
H2O, 1.0 mg/L ZnSO4×
CuSO4×5H2O, 0.013 mg/L (NH4)6Mo7O24×
/
7H2O, 20 mg/L FeCl3, 1.6 mg/L MnSO4×
7H2O, 1 mg/L CoCl2, 0.04 mg/L
4H2O, 1 g/L
/
/
/
/
glucose, 5 g/L lactose, 8 g/L chitin from crab shells
(Sigma Chemical Co., St. Louis, MO) and 1 g/L
Fig. 6. TLC separation of the hydrolysis products from chitosan after incubation with either Serratia/Trichoderma or Streptomyces/
Trichoderma enzyme preparation blends. For each sample, Trichoderma component is the complement to the percentage reported in
the figure. Reactions were concentrated 5.5ꢃ in Speedvac and 3 mL/lane were loaded. Oligo: N-acetylglucosamine (50 mg),
/
chitobiose (25 mg), chitotriose (25 mg), and chitotetraose (25 mg). GlcN: glucosamine (50 mg). The image was digitally enhanced by
uniformly replacing the background color (dark yellow) with white color.

1830
B. Giuliano Garisto Donzelli et al. / Carbohydrate Research 338 (2003) 1823ꢀ1833
/
Fig. 7. Effect of hexosamine concentration on nahase activity (MUA) of blends containing (as percentage) 0/100, 25/75, 50/50, 75/25
or 100/0 of either Serratia/Trichoderma (A and B) or Streptomyces/Trichoderma (C and D) enzymes. Each blend was incubated in
the presence of 25 (-m-), 10 (-"-), 5 (-'-), 1 (-j-), and 0.1 (-ꢃ
/
-) g/L of either GlcNAc (A and C) or GlcN (B and D). Activity levels
are expressed as the percent of the activity that a particular bend showed in the absence of hexosamine. The experiment was executed
following a simplex-lattice mixture design.
scleroglucan (Carbomer, Westborough, MA). The pH
was adjusted to 6.5. Flasks were incubated on a rotary
shaker (150 rpm) at 25 8C for 9 days.
Culture filtrates from S. albidoflavus strain NRRL B-
16746 were prepared as previously described;31 however,
the medium contained 2.5 g/L of casein (Fisher Scien-
tific, Fair Lawn, NJ) and the growth time was 6 days.
Both enzyme preparations produced for this study
were culture filtrates harvested by centrifugation and
filtration to remove the producing organism.
75 and 0/100. Digestions were incubated in 25 mL
Erlenmeyer flasks either at 32 8C on magnetic stirrers
(350 rpm) or at 37 8C on an orbital shaker (220 rpm).
After 7 days of incubation, the concentration of
langostino shells was brought to 100 g/L. Reactions
were incubated for 7 more days under the same
conditions. Samples were taken every 24 h.
4.3. Synergism in the hydrolysis of chitin substrates by
prokaryotic/Trichoderma mixtures
4.2. Hexosamine production from langostino crab chitin
The following assays were conducted on Serratia/
Trichoderma and Streptomyces/Trichoderma chitinase
blends according to a two-component, five-level sim-
plex-lattice mixture design that measured, in separate
experiments: (a) release of GlcNAc from langostino
shell chitin; (b) reduction of turbidity of colloidal chitin
suspensions; or (c) release of reducing sugars (expressed
as GlcN) from chitosan. Results were fitted by multiple
linear regression using the program Design-Expert 6.0
(Stat-Ease Inc., Minneapolis MN).
Deproteinized and demineralized langostino shell flakes
(ꢀ
/
90% chitin with a degree of acetylation ꢀ85%;
/
caustic solubles, 1%; moisture, 4.6%; ashes, 1%) were
provided by Biopolymer Engineering and used without
any further treatment. Particle size was variable (up to 3
mm). Enzymatic reactions contained 50 g shells/L in
presence of 100 mM NaOAc buffer, pH 5.0, and
included blends of S. marcescens and T. atroviride at
the following percentage ratios: 100/0; 75/25; 50/50; 25/

B. Giuliano Garisto Donzelli et al. / Carbohydrate Research 338 (2003) 1823ꢀ
/
1833
1831
Fig. 8. Effect of hexosamine concentration on hydrolytic activity against MUB of blends containing (as percentage) 0/100, 25/75,
50/50, 75/25 or 100/0 of either Serratia/Trichoderma (A and B) or Streptomyces/Trichoderma (C and D) enzymes. Each blend was
incubated in the presence of 25 (-m-), 10 (-"-), 5 (-'-), 1 (-j-), and 0.1 (-ꢃ
/
-) g/L of either GlcNAc (A and C) and GlcN (B and D).
Activity levels are expressed as the percent of the activity that a particular blend showed in the absence of hexosamine. The
experiment was executed following a simplex-lattice mixture design.
4.4. Release of GlcNAc from langostino shell chitin
Trichoderma enzyme blends diluted 1:10 were added to
the substrate and incubated for 3 h at 37 8C. Reduction
of turbidity was calculated as the percent decrease in
absorbance measured at 600 nm.
Fifty g/L of langostino shell fragments were incubated
with prokaryotic/Trichoderma mixtures. Digestions
were buffered at pH 5.0 with 100 mM NaOAc and
incubated at 32 8C on magnetic stirrers (350 rpm). After
72 h, solids were removed by centrifugation and
GlcNAc concentration was measured.
4.6. Release of reducing sugars from chitosan
Chitosan was prepared as previously described.33 One
volume of enzyme solutions or blends (500 mL) was
incubated with one volume of 5% chitosan (w/v)
suspended in 200 mM NaOAc, pH 5.0. Reactions were
incubated at 37 8C for 4 h on an orbital shaker (215
rpm). Solids were removed by centrifuging the digestion
4.5. Turbidity reduction of colloidal chitin suspensions
Colloidal chitin turbidity reduction, which results pri-
marily from endochitinase activity,32 was measured in
96-well microtiter plates using 50 mL of colloidal chitin
for 10 min at 14,000g. Supernatants were diluted 5ꢃ in
/
suspension (OD600
ꢂ/0.6) as the substrate. Fifty micro-
water and reducing sugar release was assayed using
GlcN as the standard.
liters of either Serratia/Trichoderma or Streptomyces/

1832
B. Giuliano Garisto Donzelli et al. / Carbohydrate Research 338 (2003) 1823ꢀ1833
/
2. Muller-Fassbender, H.; Bach, G. L.; Haase, W.; Rovati,
L. C.; Setnikar, I. Osteoarthritis Cartilage 1994, 2, 61ꢀ69.
3. Salvatore, S.; Heuschkel, R.; Tomlin, S.; Davies, S. E.;
4.7. Chitinase activity measurement
/
Exo- and endo-chitinase activities were measured by the
release of 4-methylumbelliferone from either 4-methy-
Edwards, S.; Walker-Smith, J. A.; French, I.; Murch, S. H.
Aliment Pharmacol. Ther. 2000, 14, 1567ꢀ
4. Kamel, M.; Hanafi, M.; Bassiouni, M. Clin. Exp. Rheu-
matol. 1991, 9, 17ꢀ21.
5. Ferrer, J.; Paez, G.; Marmol, Z.; Ramones, E.; Garcia, H.;
Forster, C. F. Bioresour. Technol. 1996, 57, 55ꢀ60.
6. Carroad, P. A.; Tom, R. A. J. Food Sci. 1978, 43, 1158ꢀ
/1579.
lumbelliferyl
N-acetyl-b-
D
-glucosaminide
(MUA,
Sigma) or 4-methylumbelliferyl-N,N?-diacetyl-b-chito-
bioside (MUB, Sigma), respectively. Both substrates
were used at 0.1 mM final concentration in the presence
of 50 mM NaOAc buffer at pH 5.0 and 1% BSA fraction
V (Fisher Scientific). One unit of nahase or endochiti-
nase activity was defined as the amount of enzyme
/
/
/
1161.
7. Pichyangkura, R.; Kudan, S.; Kuttiyawong, K.; Sukwat-
tanasinitt, M.; Aiba, S. Carbohydr. Res. 2002, 337, 557ꢀ
/
required to release 1 nmol of methylumbelliferoneꢃ
/
sꢄ1 from MUA or MUB in 1 mL reaction volume at
37 8C.
559.
8. Watanabe, T.; Kimura, K.; Sumiya, T.; Nikaidou, N.;
Suzuki, K.; Suzuki, M.; Taiyoji, M.; Ferrer, S.; Regue, M.
J. Bacteriol. 1997, 179, 7111ꢀ7117.
/
9. Saito, A.; Fujii, T.; Yoneyama, T.; Redenbach, M.; Ohno,
T.; Watanabe, T.; Miyashita, K. Biosci. Biotechnol.
4.8. Chitinase inhibition activity by GlcNAc and GlcN
Biochem. 1999, 63, 708ꢀ
10. Felse, P. A.; Panda, T. Bioprocess. Eng. 2000, 23, 127ꢀ
/
710.
In order to investigate whether prokaryotic/Tricho-
derma enzyme blends had reduced sensitivity to GlcNAc
and GlcN compared to unblended preparations, we
measured percentage variation in the chitinolytic activ-
ity against both MUA and MUB in the presence of 0,
0.1, 1, 5, 10, 25 g/L of either GlcNAc or GlcN. Assays
were carried out following a two component, five-level
simplex-lattice mixture design.
/
134.
11. Aloise, P.A.; Lumme, M.; Haynes, C.A., in: R.A.A.
Muzzarelli (Ed.), Chitin Enzymology, vol. 2, Atec Edi-
zioni, Grottammare (Italy), 1996.
12. Lorito, M. In Trichoderma and Gliocladium; Harman, G.
E.; Kubicek, C. P., Eds.; Vol. 2; Taylor and Francis:
London, 1998.
13. Letzel, M. C.; Synstad, B.; Eijsink, V. G. H.; Peter-
Katalinic, J.; Peter, M. G. Adv. Chitin Sci. 2000, 4, 545ꢀ
/
4.9. Amino sugar detection and measurement
552.
14. Sashiwa, H.; Fujishima, S.; Yamano, N.; Kawasaki, N.;
Nakayama, A.; Muraki, E.; Hiraga, K.; Oda, K.; Aiba, S.
GlcNAc release was measured using the ElsonꢀMorgan
/
assay.34 Reducing sugar release was measured using
the Somogyi assay.35 TLC was carried out as described
previously using pre-coated silica gel plates (Sigma,
cat. # Z122785).36 After separation, plates were air
Carbohydr. Res. 2002, 337, 761ꢀ
15. Sukwattanasinitt, M.; Zhu, H.; Sashiwa, H.; Aiba, S.
Carbohydr. Res. 2002, 337, 133ꢀ137.
/763.
/
16. Haynes, C.A.; Aloise, P.; Creagh, A.L. U.S. Patent
5,998,173, 1999.
dried for 1 h, spayed with anilineꢀdiphenylamine
/
17. Hayes, C. K.; Klemsdal, S.; Lorito, M.; Di Pietro, A.;
Peterbauer, C.; Nakas, J. P.; Tronsmo, A.; Harman, G. E.
reagent (Sigma, cat. # A8142) and baked at 140 8C for
4 min. In addition, sugars were quantitated at Econo-
tech (British Columbia, Canada) by high-performance
anion-exchange chromatography with pulsed ampero-
metric detection on a Dionex DX 500 apparatus as
described.37
Gene 1994, 138, 143ꢀ
18. Peterbauer, C. K.; Lorito, M.; Hayes, C. K.; Harman, G.
E.; Kubicek, C. P. Curr. Genet. 1996, 30, 325ꢀ331.
/148.
/
19. van Aalten, D. M.; Synstad, B.; Brurberg, M. B.; Hough,
E.; Riise, B. W.; Eijsink, V. G.; Wierenga, R. K. Proc.
Natl. Acad. Sci. USA 2000, 97, 5842ꢀ5847.
/
20. Suzuki, K.; Sugawara, N.; Suzuki, M.; Uchiyama, T.;
Katouno, F.; Nikaidou, N.; Watanabe, T. Biosci. Biotech-
Acknowledgements
nol. Biochem. 2002, 66, 1075ꢀ1083.
/
21. Hain, T.; Ward-Rainey, N.; Kroppenstedt, R. M.; Stack-
ebrandt, E.; Rainey, F. A. Int. J. Syst. Bacteriol. 1997, 47,
The assistance of Kristen Ondik is gratefully acknowl-
edged as is the advice of Karl Siebert on modeling. The
research at Cornell University was supported in part by
a grant from Biopolymer Engineering.
202ꢀ
22. Synowiecki, J.; Quawi Al Khateeb, N. A. A. Food Chem.
2000, 68, 147ꢀ152.
23. Fuchs, R. L.; McPherson, S. A.; Drahos, D. J. Appl.
Environ. Microbiol. 1986, 51, 504ꢀ509.
/
206.
/
/
24. Harpster, M. H.; Dunsmuir, P. Nucleic Acids Res. 1989,
17, 5395.
References
25. Gongora, C.E. Ph.D. Thesis, Cornell University, 1999.
26. Mansfield, S. D.; Mooney, C.; Saddler, J. N. Biotechnol.
1. Hochberg, M. C.; Dougados, M. Best Pract. Res. Clin.
Rheumatol. 2001, 15, 583ꢀ
/593.
Prog. 1999, 15, 804ꢀ816.
/

B. Giuliano Garisto Donzelli et al. / Carbohydrate Research 338 (2003) 1823ꢀ
/
1833
1833
27. Inglis, G. D.; Kawchuk, L. M. Can. J. Microbiol. 2002,
48, 60ꢀ70.
28. Nogawa, M.; Takahashi, H.; Kashiwagi, A.; Ohshima, K.;
32. Tronsmo, A.; Harman, G. E. Anal. Biochem. 1993, 208,
74ꢀ79.
33. No, H. K.; Cho, Y. I.; Kim, H. R.; Meyers, S. P. J. Agric.
Food Chem. 2000, 48, 2625ꢀ2627.
34. Chaplin, M. F. In Carbohydrate Analysis*
/
/
Okada, H.; Morikawa, Y. Appl. Environ. Microbiol. 1998,
895.
/
64, 890ꢀ
/
/
A Practical
29. Sheehan, J.; Himmel, M. Biotechnol. Prog. 1999, 15, 817ꢀ
/
Approach; Chaplin, M. F.; Kennedy, J. F., Eds.; IRL
827.
30. Tarantili, P. A.; Koullas, D. P.; Christakopoulos, P.;
Kekos, D.; Koukios, E. G.; Macris, B. J. Biomass
Press: Oxford, 1986.
35. Somogyi, M. J. Biol. Chem. 1952, 195, 19ꢀ23.
36. Tanaka, T.; Fujiwara, S.; Nishikori, S.; Fukui, T.; Takagi,
/
Bioenergy 1996, 10, 213ꢀ219.
/
M.; Imanaka, T. Appl. Environ. Microbiol. 1999, 65,
31. Broadway, R. M.; Williams, D. L.; Kain, W. C.; Harman,
G. E.; Lorito, M.; Labeda, D. P. Lett. Appl. Microbiol.
5338ꢀ
/
5344.
37. Dionex, Technical Note 40. 1998.
1995, 20, 271ꢀ276.
/