Accepted Manuscript Continuous synthesis of lactulose in an enzymatic membrane reactor reduces lactulose secondary hydrolysis Azis Boing Sitanggang, Anja Drews, Matthias Kraume PII: DOI: Reference:
S0960-8524(14)00842-6 http://dx.doi.org/10.1016/j.biortech.2014.05.124 BITE 13532
To appear in:
Bioresource Technology
Received Date: Revised Date: Accepted Date:
31 March 2014 30 May 2014 31 May 2014
Please cite this article as: Sitanggang, A.B., Drews, A., Kraume, M., Continuous synthesis of lactulose in an enzymatic membrane reactor reduces lactulose secondary hydrolysis, Bioresource Technology (2014), doi: http:// dx.doi.org/10.1016/j.biortech.2014.05.124
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1 2
Continuous synthesis of lactulose in an enzymatic membrane reactor reduces lactulose secondary hydrolysis
3
Azis Boing Sitanggang1,3*, Anja Drews2, Matthias Kraume1
4 5 6
1
Chair of Chemical and Process Engineering, Technische Universität Berlin. Ackerstraße 76, 13355 Berlin, Germany. 2
7 8 9 10
3
HTW Berlin - University of Applied Science, Engineering II, School of Life Science Engineering. Wilhelminenhofstraße 75A, 12459 Berlin, Germany.
Department of Food Science and Technology, Bogor Agricultural University. Raya Darmaga St, Kampus IPB Darmaga Bogor 16680, West Java, Indonesia.
11 12 13 14 15 16 17
*Corresponding Author:
18
Name : Azis Boing Sitanggang, MSc
19
Email :
[email protected] 20
Phone : +49-(030) 314-72693
21
Fax
22 23
Chair of Chemical and Process Engineering, Technische Universität Berlin, Ackerstraße 76, 13355 Berlin, Germany
: +49-(030) 314-72756
24 25 26 27 28 29 30 1
31
Abstract
32
Newly developed parallel small-scale enzymatic membrane reactors (EMRs) were used to
33
enhance the synthesis of lactulose using β-galactosidase. Under batch operation, the
34
productivity of lactulose decreased abruptly from 2.72 down to 0.04 mglactulose/(Uenzymeh) over
35
35 h of reaction. This was presumably caused by the action of β-galactosidase which
36
performed secondary hydrolysis upon the produced lactulose. The continuous operations of an
37
EMR system led to continuous removal of lactulose in the reactors restricting lactulose
38
degradation caused by secondary hydrolysis. Therefore, continuous lactulose syntheses in the
39
EMRs yielded significantly higher specific productivities under “steady state” conditions.
40
Approximately 0.70 and 0.50 mglactulose/(Uenzymeh) for hydraulic residence times of 5 and 7 h
41
were reached, respectively. Continuous lactulose synthesis performed in an EMR system
42
conclusively can circumvent the drawbacks (e.g., secondary hydrolysis) of lactulose synthesis
43
encountered in batch operation. It is, therefore, beneficial in terms of enhanced lactulose
44
productivity and reduced enzyme consumption.
45
Keywords: Enzymatic membrane reactors (EMRs), lactulose, lactose, β-galactosidase,
46
transgalactosylation, process intensification.
47 48 49 50 51 52 53 54 2
55
1. Introduction
56
The potential of enzymes to catalyse bio-chemical reactions is enormous with some of the
57
economically interesting products of these reactions being their intermediates rather than their
58
end-products. For these reactions, the used enzymes must have at least two functions and it is,
59
therefore, plausible that similar enzymes used to catalyse the primary reactions are also
60
responsible for the degradation of the intermediates (Kirk et al., 2002). The syntheses of
61
lactulose and galactooligosaccharides (GOS) lead to interesting intermediate products.
62
Growing attention is recently paid to the production of lactulose and GOS as they are
63
considered to be the valuable substances from the group of di-/oligosaccharides highlighting
64
their prebiotic properties (Schuster-Wolff-Bühring et al., 2010). Lactulose is reported to have
65
a number of physiological effects on humans, such as enhancing colonic motility, enriching
66
the growth of probiotic bacteria, improving mineral absorptions, reducing the growth of
67
putrefactive bacteria and acting as laxative agent in the treatment of constipation (Schumann,
68
2002; Schuster-Wolff-Bühring et al., 2010; Seki et al., 2007).
69
There are two enzyme classes that can catalyse lactulose and GOS synthesis, such as
70
glycosyltransferases and glycosidases (Hancock et al., 2006; Mayer et al., 2004). These
71
enzymes have two main functions; (i) to hydrolyse lactose and (ii) transfer the galactosyl-
72
enzyme complex to galactosyl acceptors (van Rantwijk et al., 1999) (see Supplementary Data,
73
Fig. S1, modified from Shen et al., 2012). When the goal of the reaction is to hydrolyse
74
lactose using glycosyltransferases or glycosidases, water molecules act as galactosyl
75
acceptors and eventually glucose and galactose are produced as the end-products. However,
76
when other substances are used as galactosyl acceptors (i.e., fructose, alcohols etc.), a
77
transgalactosylation process occurs halfway during the reactions yielding lactulose, GOS and
78
other galactosides. These reactions are kinetically controlled and, therefore, undergo
79
secondary hydrolysis by the actions of the same enzymes (van Rantwijk et al., 1999; Wang et
80
al., 2012). 3
81
Many papers report on enzyme-catalysed lactulose and GOS syntheses (Hua et al., 2013,
82
2010; Mayer et al., 2004). Most of them were carried out in batch operations with typical
83
reactor working volumes of 1-500 mL (Hua et al., 2013; Martínez-Villaluenga et al., 2008;
84
Mayer et al., 2004). Syntheses of these transgalactosylated products under batch operation
85
normally utilise higher amounts of enzymes as the reactions have to be stopped after 8 h to
86
avoid secondary hydrolysis taking place on these substances. Therefore, summarising the
87
overall procedures (start-up and end activities) of lactulose and GOS syntheses it can be stated
88
that batch production is presumably infeasible from an economic point of view. Many
89
attempts have been made to increase the productivity of lactulose and GOS productions either
90
in batch or continuous processes, including (i) the application of organic solvents as reaction
91
media to reduce the amount of water in the reaction system and consequently hampering the
92
hydrolysis reaction rate of the produced transgalactosylated products, (ii) the application of
93
more than one enzyme during the synthesis, and (iii) other operational strategies (Hua et al.,
94
2010; Schuster-Wolff-Bühring et al., 2010).
95
For an alternative operational strategy, membrane reactors can be used to continuously
96
remove the transgalactosylated products during the reaction. Foda and Lopez-Leiva (2000)
97
reported continuous production of oligosaccharides (OS) from whey using a UF membrane
98
cell operated in cross-flow filtration mode. Within their studies, a constant flux operation was
99
performed by means of two peristaltic pumps. Due to the insufficient control design of the
100
reactor system, the reactor could only be operated for less than 3 h at a flux value of 15
101
L/(m²h). Chockchaisawasdee et al. (2005) produced GOS using β-galactosidase from
102
Kluyveromyces lactis in UF membrane bioreactors. The reactors could only be operated at
103
constant transmembrane pressure (TMP) and, therefore, the permeate flux declined over the
104
reaction course as cake layer deposition or fouling occurred on top of the UF membranes.
105
Thereby, the hydraulic retention time (HRT) increased during the process which led to rising
106
concentrations of GOS inside the reactor over time. As a consequence, secondary hydrolysis 4
107
of the produced GOS occurred faster. Considering these conditions, a proper control design
108
for a constant flux (i.e., HRT) operation during GOS and other transgalactosylation products
109
is of importance to restrict the impact of secondary hydrolysis taking place.
110
Mayer et al. (2010) synthesised lactulose in an EMR and in a packed-bed reactor (PBR)
111
continuously using β-glycosidase from Pyrococcus furious. The lactulose production in the
112
EMR was unsuccessful because the half-life of the enzyme was less than a day and further
113
investigations were not reported. On the other hand, the synthesis of lactulose in a PBR was
114
successful using the immobilised β-glycosidase. It is, however, assumed to be quite tedious as
115
the reported preparation of the enzyme immobilisation took more than 24 h.
116
Considering the economic value of the positive physiological properties of lactulose, there is a
117
need to develop a proper reactor and control system for lactulose synthesis to avoid secondary
118
hydrolysis occurring on the produced lactulose during the reaction. Hence, this study aimed to
119
further develop a screening and characterisation system based on parallel small-scale EMRs
120
(Lyagin et al., 2010) which allows continuous production of lactulose at a constant flux
121
operation with a short HRT. The two reactors in parallel make it possible to obtain duplicates
122
and reproducible results in a shorter space of time. Developing a control design for the reactor
123
system was firstly carried out to enable a constant flux operation (thus constant HRT) for a
124
long period of reaction. Moreover, continuous lactulose syntheses were conducted at different
125
HRT values and their productivities were compared with the one from the synthesis of
126
lactulose in a batch process.
127 128 129 130
2. Material and methods 2.1 Material 2.1.1 Reactor configuration
131
Initially, two parallel EMRs were built for batch and continuous operation purposes. The
132
scheme of the developed EMR system is shown in Fig. 1. An EMR consisted of a pressure-
133
stable glass container and a body (holder) which was modified from the XFUF-047 dead-end 5
134
test cell, Merck Millipore Darmstadt, Germany. These modifications were necessary to insert
135
the temperature, pH and substrate probes inside the reactor (see Fig. 1). The maximum
136
working volume of the reactor was 90 mL with a flat-sheet PES membrane (effective surface
137
area = 12.38 x 10-4 m²) placed at the bottom of the reactor. Piping (PTFE ID = 0.8 mm; OD =
138
1.6 mm) and other small parts for fittings were received from VWR International GmbH, and
139
Th. Geyer Berlin GmbH, Germany. UF membranes made of polyethersulfone (PES) with a
140
molecular weight cut-off (MWCO) of 10 kDa were obtained from Microdyn-Nadir GmbH,
141
Germany. The pressure-stable pH electrode (glass, Pmax = 6 bar, Tmax = 130 oC, 8 x 120 mm)
142
was purchased from ProSense BV, The Netherlands. For permeate measurement, a Kern EW
143
620-3NM precision balance was employed (Kern & Sohn GmbH, Germany). In addition, the
144
proportional pressure regulator MPPE (Festo AG & Co. KG, Germany) was used for
145
continuous operation.
146
Data acquisition (DAq) was realised as a combination of several National Instruments (NI)
147
modules, such as (i) NI 9201 16-channel, 100 kS/s/ch, 16-Bit, ±10 V analogue input module;
148
(ii) NI 9870 4-port, RS232 serial interface module and (iii) NI 9264 16-channel analogue
149
output module. These modules were mounted on a cRIO-9076 integrated 400 MHz real-time
150
controller and LX45 FPGA chassis system produced by National Instruments, Germany. This
151
typical chassis can fully control up to 12 parallel reactors, enabling the feasibility of this
152
system as a rapid screening and characterisation system. The software Laboratory Virtual
153
Instrument Engineering Workbench (LabVIEW) Professional, version 2012 was employed to
154
control the reactors and save all data.
155
2.1.2 Chemicals
156
The enzyme β-galactosidase from Kluyveromyces lactis (EC Number 232-864-1, G3665),
157
acetonitrile (271004), 2-Nitrophenyl β-D-galactopyranoside (ONPG, 73660), 2-Nitrophenol
158
(ONP, 19702), lactulose (61360), D-fructose (F0127), lactose (17814) were purchased from
6
159
Sigma-Aldrich, Germany. All other chemicals were analytical grade, obtained from Merck-
160
Millipore and from VWR International GmbH, Germany.
161 162
2.2 Control design and strategy for constant flux operation
163
Permeate mass was quantified by means of a precision balance and converted into volume for
164
calculating its flow rate. Furthermore, the flow rate served as an actuator for controlling flux
165
or HRT. With regards to the controller output (CO; pressure, bar), and process variable (PV;
166
flux, L/(m²h)), the filtration process showed a non-periodic response that follows a PT1T0
167
model as process dynamics (Schwarze, 1962) and has been reported elsewhere (Lyagin et al.,
168
2010). These process dynamics were generally recognised as first order plus dead time
169
(FOPDT) behaviour. To control such process dynamics, a proportional-integral-derivative
170
(PID) controller was implemented in the EMR system. Moreover, PID parameters were tuned
171
according to the method described by Kuhn (1995).
172
The flux (and thus also HRT) was feed-back controlled in the continuous process (see Fig. 1).
173
When the setting value (SV) of flux was inserted into the LabVIEW program, the program
174
automatically sent an analogue input commanding the proportional pressure regulator (3) to
175
open its valve allowing a certain amount of gas to be released from the N2 bottle (1). As the
176
substrate tank (4) was pressurised, substrate solution was fed into the reactor (5). Since the
177
reactor was completely filled with liquid, the additional volume from the substrate tank
178
consequently drove the same amount of liquid out of the reactor via the membrane which was
179
then collected by the precision balance (8). In the LabVIEW program this permeate mass was
180
converted to its volume (using the corresponding density of the solution) and eventually into
181
the real flux value (process variable, PV) using the membrane effective surface area and the
182
time of filtration. This loop was carried out within one second (1 s). Furthermore, the
183
difference between PV and SV was used by the PID controller to give the subsequent
7
184
command to the proportional pressure regulator again (3). This cycle was repeated during the
185
whole reaction process, giving the possibility to control a flux and thus HRT precisely.
186 187
2.3 Enzyme filtration
188
In order to study the membrane performance, enzyme filtration was conducted continuously
189
at an HRT of 3 h for 25 h. The significant difference between the membrane MWCO (i.e.,
190
10 kDa) and the enzyme’s molecular weight (i.e., β-galactosidase ~ 465 kDa (Appel et al.,
191
1965; Juers et al., 2000)) should give a complete rejection of the enzyme molecules. By
192
quantifying the protein concentration in the permeate (see section 2.6.2), the membrane
193
rejection was found to be 95 %. As reported by Lozano et al. (2014), a PES membrane with
194
10 kDa MWCO could also reject 99 % of a cellulase complex (celluclast 1.5L®) and 98 % of
195
cellobiase (novozym 188).
196 197
2.4 Operational procedure for lactulose synthesis
198
Lactulose was synthesised in the developed EMR system run respectively in batch and
199
continuous mode, using a bi-substrate (lactose and fructose) and K. lactis β-galactosidase. In
200
contrast to enzymes from the group of glycosyltransferases that usually need activated
201
substrates and cofactors, β-galactosidase (group of glycosidases) is generally cofactor
202
independent (Mayer et al., 2004). In addition, it is more relevant for industrial applications
203
(e.g., for hydrolysis of lactose) as it is commercially available and relatively inexpensive
204
(Mayer et al., 2004; Perini et al., 2013).
205
Under batch operation, a series of reactions was carried out without enzyme removal from the
206
reactor. A 1.0 mL sample was taken from the permeate side and compensated by the same
207
amount (i.e., 1.0 mL) of fresh substrate conveyed into the reactor at 0.5; 2: 3; 5; 8; 10 and 12
208
h. In order to enable this sampling procedure, a pressure (i.e., 0.5 bar) was applied in the 8
209
substrate tank manually. Though an exchange between inlet and permeate took place in the
210
system, the operation was still considered as a batch reaction as it had a very long HRT.
211
The effects of several parameters, such as total sugar concentration (g/L), enzyme
212
concentration (U) and molar ratio of lactose to fructose (mL/mF) on the lactulose production
213
were investigated in batch operation. For continuous operation, the concentration of lactulose
214
was monitored for different HRT values, initially 5 and 7 h. In both types of operation (batch
215
and continuous), the reactors were fully filled with a working volume of 90 mL, 50 mM
216
buffer phosphate pH 6.8 as the reaction medium, stirred at 200 rpm and incubated at a
217
constant temperature of 40 ± 1 °C.
218
The productivities of batch and continuous processes (at different HRT values) were
219
compared in terms of specific productivity calculated as follows:
220
(1)
221
where Ci = lactulose concentration at a certain sampling time ti, Vi = permeate volume
222
collected until ti (for batch mode Vi = 90 mL), ti = sampling time.
223 224 225
2.6 Analysis 2.6.1 Measurement of the enzyme activity
226
ONPG was used as a substrate to determine the activity of β-galactosidase. The procedure
227
was similar to the method reported by Hua et al. (2010) with several minor modifications. The
228
enzyme β-galactosidase generally hydrolyses ONPG to yield ONP. The color of the ONP
229
solution is yellow and can be monitored at 420 nm using a spectrophotometer (Specord ® 210
230
Plus – Analytik Jena AG, Germany). One unit (1 U) of enzyme activity is defined as the
231
amount of enzyme required for liberating the equivalent 1 µmol ONP per minute at 30 oC and
232
pH 6.8.
233 234 9
235
2.6.2 Determination of mono-, disaccharides and protein concentration
236
For mono- and disaccharide analysis, an HPLC was used, equipped with a Vertex Plus 250 x
237
4.6 mm Eurospher II 100-3 NH2 column (Knauer GmbH, Germany), a WellChrom K-2300 RI
238
detector and a K.1001 pump combined with an electric valve drive. The evaluation of the
239
resulting chromatograms was done by the software Eurochrom 2000. The mobile phase was a
240
mixture of acetonitrile and water (75:25) with an isocratic gradient, pumped at a constant flow
241
rate of 1.0 mL/min. Column temperature was set to 30 oC. For achieving a proper resolution,
242
samples were diluted with ultra-pure water for four times and with the mobile phase at a ratio
243
of 1:1 prior to injection. Finally 20 µL of the diluted sample were thoroughly injected into the
244
HPLC. The retention time was 30 min.
245
The enzyme concentration (expressed as protein concentration) in the permeate during the
246
enzyme filtration test was determined using the linearised Bradford method (Zor and Selinger,
247
1996). The protein concentrations (0-0.4 g/L) and their absorbance ratios at 590 nm and
248
450 nm were plotted to give a linear line.
249 250 251
3. Results and discussion 3.1 Flux stability
252
Tuning PID controller parameters (P, I and D) is still one of the most highlighted topics in
253
control design and it sometimes becomes the bottleneck of PID controller’s wider applications
254
(Skogestad, 2003). Tuning PID controllers basically depends on the process dynamics of the
255
plant. Therefore, systematic procedures to find suitable values for those parameters are
256
required (Skogestad, 2003). There are many strategies that have been introduced for tuning
257
PID parameters and have been extensively summarised in several works by different authors
258
(Aidan, 2009; Madhuranthakam et al., 2008; Syrcos and Kookos, 2005). Out of those, a
259
strategy introduced by Kuhn (1995) was selected within this study which also had been used
260
elsewhere (Lyagin et al., 2010). This strategy is based on an open-loop approach which is 10
261
considered not to be as time-consuming as the closed-loop one, because there is no need to
262
wait for several periods of oscillations during several trial-and-error attempts. Within this
263
study, the values for PID parameters were 0.0075, 0.360 min, and 0.087 min, respectively for
264
P, I and D. This setting was tested in the filtration of lactose and fructose with a total sugar
265
concentration of 500 g/L in a serial flux stepping operation. As can be seen in Fig. 2, when
266
the SV shifted, the CO changed accordingly to bring the PV as close as possible to the SV.
267
The CO was not jerky and consequently the PV responses were stable without producing any
268
overshoot response. The error between SV and PV was evaluated to be less than 5 %. This
269
result was comparable to the study of tuning PI/D parameters in a closed-loop method
270
reported by Skogestad (2003) and Shamsuzzoha and Skogestad (2010).
271 272 273 274
3.2 Transgalactosylation towards lactulose 3.2.1 Effects of the total sugar concentration (g/L), enzyme concentration (U) and molar ratio of lactose to fructose (mL/mF)
275
The investigation of the effects of several parameters (sugar concentration (g/L), enzyme
276
concentration (U) and molar ratio of lactose to fructose (mL/mF)) on the productions of
277
lactulose was done in batch processes. As can be seen in Fig. 3, a higher sugar concentration
278
constantly yielded a higher lactulose production. The maximum concentrations of lactulose
279
obtained were 6.69; 8.24 and 11.0 g/L for sugar concentrations of 300, 400 and 500 g/L,
280
respectively. A smaller amount of lactulose for a lower sugar concentration (i.e., 300 g/L) was
281
caused by a lower availability of the galactosyl-enzyme complex which was necessary to react
282
with fructose. In relation to these results, the yields of lactulose also increased slightly from
283
0.03 to 0.031 and 0.034 glactulose/glactose for sugar concentrations of 300, 400 and 500 g/L,
284
respectively. Lee et al. (2004) also reported an increased level of lactulose concentration at a
285
higher concentration of the substrate (a mixture of lactose and fructose) at a constant molar
286
ratio of lactose to fructose (mL/mF).
11
287
The small amount of enzyme (i.e., 100 U) was assumed to be inhibited in the highly
288
concentrated substrate since higher lactulose production was obtained when the enzyme
289
concentration was brought up to 300 U (Fig. 4a). However, at level of 500 U of enzyme
290
concentration, the obtained lactulose concentration was not much altered. Similar results have
291
been reported by Kim and Oh (2012) where the concentration of lactulose increased by the
292
increase of the enzyme load and eventually reached the maximum lactulose concentration at
293
optimum enzyme load (150 U/mL). In Fig. 4a, lactulose concentrations were up to 14.1 and
294
13.76 g/L for 300 and 500 U, respectively. Within this study, an enzyme concentration of 300
295
U was considered to be the nearly optimum enzyme concentration. When the enzyme
296
concentration was 500 U, the lactulose concentration reduced as compared to that of 300 U.
297
Excessive amount of enzyme was unfavorable since it led to the acceleration of both primary
298
and secondary hydrolysis (see Fig. 4a) and eventually resulted in a lower lactulose
299
concentration and higher lactose consumption.
300
It is worth mentioning that higher enzyme concentrations performed the transgalactosylation
301
faster. In the course of the reaction at 100 U of enzyme concentration, the maximum lactose
302
concentration (i.e., 11.0 g/L) was reached after 23 h (data not shown) whereas for 300 and
303
500 U the optimum reaction times were 5 and 3 h, respectively. Secondary hydrolysis
304
generally occurred as soon as lactulose was produced. This was due to secondary hydrolysis
305
being a kinetically controlled reaction. Moreover, secondary hydrolysis appeared more
306
visible, after the maximum concentration of lactulose has been obtained. As an example, at
307
300 U and after 5 h reaction, the lactulose concentration gradually decreased to 13.55, 12.36
308
and 11.60 g/L for 8, 10 and 12 h, respectively. Highlighting on this issue which typically
309
occurs in batch operation, a strategy for continuous removal of lactulose is of importance.
310
This withdrawal consequently needs a sufficient control design to control flux precisely
311
during the reaction. Based on lactose consumption (see Fig. 4a), it could be stated that the
312
influence of β-galactosidase on the primary hydrolysis (i.e., that of lactose) was not affected 12
313
by the lactulose secondary hydrolysis. Lactose consumption increased over the reaction time
314
to 95-98 % within 12 h. Formation of by-products, which were confirmed as GOS by Shen et
315
al. (2012), was also observed during the reaction. This was indicated by the increased number
316
of peaks that appeared in the HPLC chromatograms (data not shown). Interestingly,
317
secondary hydrolysis also occurred on those GOS. This was reflected by the reduction of the
318
peak areas of the GOS in the HPLC chromatograms after the maximum concentrations of
319
transgalactosylated products (lactulose, GOS) reached (data not shown).
320
Increasing the mL/mF ratio was found to decrease the produced lactulose concentration (Fig.
321
4b). When the mL/mF ratio was brought to 0.5, a lactulose concentration of 16.70 g/L was
322
obtained, whereas for a ratio of 2.0 and 1.0, lactulose concentrations were 8.64 g/L and 14.10
323
g/L, respectively, at 5 h reaction time. At a lower mL/mF ratio (i.e., 0.25), lactulose
324
concentration was enhanced to 17.91 g/L at 8 h. The yield of lactulose concentration reported
325
by Lee et al. (2004) was about 5 % (based on initial lactose concentration) or equal to19.6 g/L
326
using 10.4 g/L permeabilised cells of Kluyveromyces lactis at an optimum mL/mF ratio of 1.0.
327
Mayer et al. (2004) reported a maximum lactulose concentration of 16 g/L using β-
328
glycosidase from P. furiosus (CelB) at an optimum mL/mF ratio of 1/15. Apart from the types
329
of enzymes used during the synthesis of lactulose, lactulose concentration is strongly
330
dependent on mL/mF ratio (Guerrero et al., 2011). A large amount of fructose seems to be
331
advantageous for lactulose synthesis, as this composition allows a higher probability for
332
fructose to react with the galactosyl-enzyme complex (Guerrero et al., 2011; Hua et al., 2010).
333
It eventually can overshoot the transgalactosylation, resulting in a higher lactulose production.
334
However, it must be noted that higher amounts of fructose can also inhibit the catalytic
335
activity of the enzyme (i.e., β-glycosidase) as reported by Mayer et al. (2004). For further
336
studies in continuous lactulose syntheses, 500 of g/L sugar concentration, an enzyme
337
concentration of 300 U and an mL/mF ratio of 0.5 were used in the reactions.
13
338
3.2.2 Continuous syntheses of lactulose at different HRTs
339
Continuous syntheses of lactulose at two different HRT values were conducted for 35 h. As
340
can be seen in Fig. 5a, the reaction reached highest concentrations at 10 and 12 h,
341
respectively, at HRTs of 5 and 7 h. The maximum lactulose concentrations obtained for HRTs
342
of 5 and 7 h were almost similar, approximately 12 g/L. Nevertheless, it must be noted that
343
lactulose concentration at longer HRT (7 h) was slightly higher (~ 2 %) than at the shorter one
344
(5 h).
345
In batch operation (conditions: [sugar] = 500 g/L, [E] = 300 U, mL/mF ratio = 0.5), the
346
maximum concentration of lactulose was 16.70 g/L at 5 h. However, lactulose concentration
347
reduced remarkably down to 4.87 g/L over 35 h (Fig. 5a). This was presumably caused by the
348
action of β-galactosidase which performed secondary hydrolysis during the batch process.
349
Besides the sharp reduction of lactulose concentration, the primary hydrolysis also took place
350
rapidly in the batch operation leading to a lower level of the remaining lactose in the reactor
351
(down to 2.0 % after 35 h reaction, see also section 3.2.1). In comparison, the remaining
352
lactose concentrations in the reactor during continuous operations at different HRTs were
353
higher than in batch operation. The remaining lactose concentrations at 5 and 7 h HRT were
354
11.70 and 10.60 %, respectively.
355
Although at nearly constant lactulose concentration the remaining lactose concentrations at
356
the two HRT values were not significantly different, Pspec at 5 h HRT was vividly higher than
357
at 7 h, since with a shorter HRT a higher amount of permeate was produced over the time.
358
The values were 0.70 and 0.50 mglactulose/(Uenzymeh), respectively, for HRTs of 5 and 7 h over
359
35 h (Fig. 5b). Pspec in the batch process at 0.5 h was four times higher than the maximum Pspec
360
in continuous operation at HRT of 5 h. This might lead to the conclusion that batch operation
361
was superior to continuous ones and, therefore, that economically batch would be preferred.
362
However, when the unproductive time for start-up and end activities (preparation, filling,
363
heating, cleaning etc.) which was evaluated to be around 3.5 h, was considered in every cycle 14
364
of batch operation, its maximum Pspec was nearly similar to that continuous operation at 5 h
365
HRT, around 0.70 mglactulose/(Uenzymeh) (see the curve of normalised batch in Fig. 5b). The
366
advantage of continuous operation at 5 h HRT was even more, as this maximum value of Pspec
367
(0.70 mglactulose/(Uenzymeh)) remained nearly constant for 35 h and it was not encountered in
368
batch operation.
369
The selectivity values obtained from batch operation and continuous operations at different
370
HRT values (5 and 7 h) were also evaluated (Fig. 5c). Selectivity is defined as the ratio
371
between the amount of produced lactulose (mol) and consumed lactose (mol). The maximum
372
selectivity achieved by batch operation was slightly higher than the continuous ones, up to
373
0.07 mollactulose/molcons. lactose at 5 h. For continuous syntheses (5 and 7 h HRT), the selectivity
374
values were almost similar, approximately 0.055 mollactulose/molcons. lactose at 10 h. As
375
mentioned previously (see also Fig. 5a), the remaining lactose concentrations for both
376
continuous syntheses remained nearly constant after 12 h (thus lactose consumptions were
377
constant), whereas for batch operation the remaining lactose concentration reduced
378
remarkably till 1 % within 35 h. Hereby, the syntheses of lactulose under 5 and 7 h HRT were
379
also propitious because the reaction selectivity values only slightly decayed (2-2.5 % of
380
reduction) during 35 h reaction whilst the selectivity in batch operation reduced drastically
381
down to 0.020 mollactulose/molcons. lactose (72 % of reduction).
382
It is obvious that continuous synthesis of lactulose acquired better production performance
383
than the batch process in terms of Pspec. This was due to the capability of the continuous
384
process to maintain “steady state” conditions with only a slight reduction of the lactulose
385
concentrations (app. 3-4 % of reduction, see Fig. 5a) over 35 h of reaction course and the fact
386
that the produced lactulose was continuously withdrawn from the reactors insofar the
387
reaction. The reduction of lactulose concentration was presumably caused by enzyme
388
inactivation. Factors that might contribute to the inactivation were thermal inactivation, shear-
389
induced inactivation and adsorption of enzyme molecules on top of the membrane (Ganesh et 15
390
al., 2000; Thomas and Geer, 2011). Conclusively, continuous synthesis of lactulose in an
391
EMR system can circumvent secondary hydrolysis which is the main obstacle in batch
392
operation and thus maintain the selectivity of the reaction.
393
To our knowledge, there is only one published work about continuous synthesis of lactulose
394
in an EMR using β-glycosidase from Pyrococcus furiosus (CelB). Mayer et al. (2010)
395
reported the synthesis of lactulose to be stable for less than one day in an EMR. Within their
396
studies, a less stable continuous lactulose production in an EMR was thought to be influenced
397
by the small mL/mF ratio (i.e., 1/15) and an unspecific enzyme adsorption onto the membrane
398
at initial operation of the reactor. Though in their previous studies (Mayer et al., 2004) such
399
an optimum mL/mF ratio of 1/15 had been chosen, this ratio was still considered to have an
400
influence on the reduction of β-glycosidase activity in a long period of operation (continuous
401
process). Within this study, after 35 h, the synthesis of lactulose using this EMR system was
402
still at steady-state. Further studies of lactulose syntheses are still on-going using the
403
developed EMR system for a long-term operation (i.e., ~100 h).
404
Besides the enhanced productivity of lactulose and the reduced enzyme consumption in
405
continuous lactulose synthesis, there is still a major issue that has to be pointed out when
406
operating an EMR system. As a consequence of the constant flux operation, the TMP
407
increases over the reaction time due to the cake deposition or fouling. The increase of TMP
408
was higher at an HRT of 5 h than at 7 h (Fig. 6). Higher permeate flux resulting from a shorter
409
HRT leads to a higher rate of material (foulant) to be deposited onto the PES membrane.
410
Similar results have been reported elsewhere, and enzyme molecules are considered to
411
contribute to such fouling phenomena (Luo et al., 2013). From Fig. 6, it is, however, seen that
412
the differences between initial and final TMP for both HRT values were not so significant.
413
Therefore, a longer continuous lactulose synthesis using this EMR system is highly possible
414
regarding the operational time span of the membrane.
415 16
416 417
3.2.3 Performances of EMRs – precision of parallel reactors
418
For an efficient and fast process design, parallel reactors are normally used. The difference
419
between the performances of the reactors in the EMR system was therefore to be considered.
420
For both HRTs (5 and 7 h), the similarity of the reactors (evaluating those two reactors)
421
during continuous lactulose productions was at a high level, about 95 %. As mentioned
422
earlier, the PID controller embedded within this EMR system could monitor up to 12 parallel
423
reactors. The developed EMR system (in this study) holds prospective possibilities to be
424
expanded (increasing the number of the reactors) without affecting the precision of each
425
reactor. Such a parallel system can circumvent the well-known bottleneck during bioprocess
426
development, namely the step from batch-wise lab-scale studies to continuous full-scale
427
operation. Such long-term continuous studies are needed to initially evaluate optimum
428
reaction conditions prior to scaling up the EMR system.
429 430
4. Conclusions
431
A newly developed EMR system supported by a suitable control design was used to enhance
432
lactulose synthesis. Under batch operation, lactulose production was significantly dependent
433
on mL/mF ratio. In continuous operation, a shorter HRT (5 h) yielded a higher specific
434
productivity compared to the longer one (7 h). The secondary hydrolysis could be
435
circumvented during continuous operation, indicated by nearly constant lactulose
436
concentration lasting for up to 35 h. Optimisation of reaction conditions, such as HRTs,
437
agitation speeds, ionic strength of the media, etc. can possibly further enhance continuous
438
synthesis of lactulose using this developed EMR system.
439 440 17
441 442
Acknowledgement
443
German Academic Exchange Service (DAAD) is gratefully acknowledged for providing a
444
PhD scholarship to the first author. In addition, the authors would like to thank Gabriele
445
Hedicke and Bernd Schmidt for the technical assistances given during this study.
446 447
References
448 449
1. Aidan, O., 2009. Handbook of PI and PID controller tuning rules. Imperial College Press; 3 edition.
450 451
2. Appel, S.H., Alpers, D.H., Tomkins, G.M., 1965. Multiple molecular forms of βgalactosidase. J. Mol. Biol. 11, 12–22.
452 453 454 455
3. Chockchaisawasdee, S., Athanasopoulos, V.I., Niranjan, K., Rastall, R. a, 2005. Synthesis of galacto-oligosaccharide from lactose using beta-galactosidase from Kluyveromyces lactis: Studies on batch and continuous UF membrane-fitted bioreactors. Biotechnol. Bioeng. 89, 434–443.
456 457
4. Foda, M.I., Lopez-Leiva, M., 2000. Continuous production of oligosaccharides from whey using a membrane reactor. Process Biochem. 35, 581–587.
458 459
5. Ganesh, K., Joshi, J.B., Sawant, S.B., 2000. Cellulase deactivation in a stirred reactor. Biochem. Eng. J. 4, 137–141.
460 461 462
6. Guerrero, C., Vera, C., Plou, F., Illanes, A., 2011. Influence of reaction conditions on the selectivity of the synthesis of lactulose with microbial β-galactosidases. J. Mol. Catal. B Enzym. 72, 206–212.
463 464
7. Hancock, S.M., Vaughan, M.D., Withers, S.G., 2006. Engineering of glycosidases and glycosyltransferases. Curr. Opin. Chem. Biol. 10, 509–519.
465 466 467
8. Hua, X., Yang, R., Shen, Q., Ye, F., Zhang, W., Zhao, W., 2013. Production of 1-lactulose and lactulose using commercial β-galactosidase from Kluyveromyces lactis in the presence of fructose. Food Chem. 137, 1–7.
468 469
9. Hua, X., Yang, R., Zhang, W., Fei, Y., Jin, Z., Jiang, B., 2010. Dual-enzymatic synthesis of lactulose in organic-aqueous two-phase media. Food Res. Int. 43, 716–722.
470 471 472 473
10. Juers, D.H., Jacobson, R.H., Wigley, D., Zhang, X.J., Huber, R.E., Tronrud, D.E., Matthews, B.W., 2000. High resolution refinement of beta-galactosidase in a new crystal form reveals multiple metal-binding sites and provides a structural basis for alphacomplementation. Protein Sci. 9, 1685–1699. 18
474 475 476
11. Kim, Y.-S., Oh, D.-K., 2012. Lactulose production from lactose as a single substrate by a thermostable cellobiose 2-epimerase from Caldicellulosiruptor saccharolyticus. Bioresour. Technol. 104, 668–672.
477 478
12. Kirk, O., Borchert, T.V., Fuglsang, C.C., 2002. Industrial enzyme applications. Curr. Opin. Biotechnol. 13, 345–351.
479 480
13. Kuhn, U. 1995. Eine praxisnahe Einstellregel für PID-Regler: Die T-Summen-Regel (In German). Automatisierungstechnische Praxis, 5: 10–16.
481 482
14. Lee, Y.-J., Kim, C.S., Oh, D.-K., 2004. Lactulose production by beta-galactosidase in permeabilized cells of Kluyveromyces lactis. Appl. Microbiol. Biotechnol. 64, 787–793.
483 484 485
15. Lozano, P., Bernal, B., Jara, A.G., Belleville, M.-P., 2014. Enzymatic membrane reactor for full saccharification of ionic liquid-pretreated microcrystalline cellulose. Bioresour. Technol. 151, 159–165.
486 487
16. Luo, J., Meyer, A.S., Jonsson, G., Pinelo, M., 2013. Fouling-induced enzyme immobilization for membrane reactors. Bioresour. Technol. 147, 260–268.
488 489 490
17. Lyagin, E., Drews, A., Bhattacharya, S., Ansorge-Schumacher, M.B., Kraume, M., 2010. Continuous screening system for inhibited enzyme catalysis: a membrane reactor approach. Biotechnol. J. 5, 813–821.
491 492 493
18. Madhuranthakam, C.R., Elkamel, a., Budman, H., 2008. Optimal tuning of PID controllers for FOPTD, SOPTD and SOPTD with lead processes. Chem. Eng. Process. Process Intensif. 47, 251–264.
494 495 496
19. Martínez-Villaluenga, C., Cardelle-Cobas, A., Olano, A., Corzo, N., Villamiel, M., Jimeno, M.L., 2008. Enzymatic synthesis and identification of two trisaccharides produced from lactulose by transgalactosylation. J. Agric. Food Chem. 56, 557–563.
497 498 499
20. Mayer, J., Conrad, J., Klaiber, I., Lutz-Wahl, S., Beifuss, U., Fischer, L., 2004. Enzymatic production and complete nuclear magnetic resonance assignment of the sugar lactulose. J. Agric. Food Chem. 52, 6983–6990.
500 501 502
21. Mayer, J., Kranz, B., Fischer, L., 2010. Continuous production of lactulose by immobilized thermostable β-glycosidase from Pyrococcus furiosus. J. Biotechnol. 145, 387–393.
503 504 505
22. Perini, B.L.B., Souza, H.C.M., Kelbert, M., Giannini, P., 2013. Production of β Galactosidase from Cheese Whey Using Kluyveromyces marxianus CBS 6556. Chem. Eng. Trans. 32, 991–996.
506 507
23. Schumann, C., 2002. Medical, nutritional and technological properties of lactulose. An update. Eur. J. Nutr. 41 Suppl 1, 117–125.
508 509
24. Schuster-Wolff-Bühring, R., Fischer, L., Hinrichs, J., 2010. Production and physiological action of the disaccharide lactulose. Int. Dairy J. 20, 731–741.
19
510 511 512
25. Schwarze, G. 1962. Bestimmung der regelungstechnischen Kennwerte von P-Gliedern aus der Übergangsfunktion ohne Wendetangentenkonstruktion (In German). Zeitschrift messen, steuern, regeln, 5: 447–449.
513 514 515
26. Seki, N., Hamano, H., Iiyama, Y., Asano, Y., Kokubo, S., Yamauchi, K., Tamura, Y., Uenishi, K., Kudou, H., 2007. Effect of lactulose on calcium and magnesium absorption: a study using stable isotopes in adult men. J. Nutr. Sci. Vitaminol. (Tokyo). 53, 5–12.
516 517
27. Shamsuzzoha, M., Skogestad, S., 2010. The setpoint overshoot method: A simple and fast closed-loop approach for PID tuning. J. Process Control 20, 1220–1234.
518 519 520 521
28. Shen, Q., Yang, R., Hua, X., Ye, F., Wang, H., Zhao, W., Wang, K., 2012. Enzymatic synthesis and identification of oligosaccharides obtained by transgalactosylation of lactose in the presence of fructose using β-galactosidase from Kluyveromyces lactis. Food Chem. 135, 1547–1554.
522 523
29. Skogestad, S., 2003. Simple analytic rules for model reduction and PID controller tuning. J. Process Control 13, 291–309.
524 525
30. Syrcos, G., Kookos, I.K., 2005. PID controller tuning using mathematical programming. Chem. Eng. Process. Process Intensif. 44, 41–49.
526 527
31. Thomas, C.R., Geer, D., 2011. Effects of shear on proteins in solution. Biotechnol. Lett. 33, 443–456.
528 529
32. Van Rantwijk, F., Woudenberg-van Oosterom, M., Sheldon, R.A., 1999. Glycosidasecatalysed synthesis of alkyl glycosides. J. Mol. Catal. B Enzym. 6, 511–532.
530 531 532 533
33. Wang, K., Lu, Y., Liang, W.Q., Wang, S. Di, Jiang, Y., Huang, R., Liu, Y.H., 2012. Enzymatic synthesis of galacto-oligosaccharides in an organic-aqueous biphasic system by a novel β-galactosidase from a metagenomic library. J. Agric. Food Chem. 60, 3940– 3946.
534 535
34. Zor, T., Selinger, Z., 1996. Linearization of the Bradford protein assay increases its sensitivity: theoretical and experimental studies. Anal. Biochem. 236, 302–308.
536 537 538 539 540 541 542 543 20
544
List of Figures
545 546 547 548 549
Fig. 1. Schematic of the developed EMR system; (1) N2 bottle, (2) pressure reducer, (3) proportional pressure regulator, (4) substrate tank, (5) reactor, (6) flat-sheet PES membrane, (7) heating system, (8) precision balance. Q = quality parameter, pH. Dashed lines (- - -) indicate control lines.
550 551 552
Fig. 2. Process variables with their corresponding pressures as controlled by PID controller in a series of setting values (30, 60, 50 and 75 L/(m2h)). Solution used was a mixture between lactose and fructose at level of 500 g/L.
553 554 555
Fig. 3. Effect of total sugar concentration on the production of lactulose using β-galactosidase from Kluyveromyces lactis, [E] = 100 U, mL/mF ratio = 1.0, phosphate buffer pH 6.8, stirred at 200 rpm, 40 °C. (a) [sugar concentration] = 300, (b) 400, (c) 500 g/L.
556 557 558 559 560 561 562 563
Fig. 4. Effects of (a) enzyme concentration [E] (mL/mF ratio = 1.0) and (b) mL/mF ratio ([E] = 300 U) on the productions of lactulose using β-galactosidase from Kluyveromyces lactis, [sugar concentration] = 500 g/L, phosphate buffer pH 6.8, stirred at 200 rpm, 40 °C.
564 565
Fig. 6. Flux profiles and TMP evolutions during continuous syntheses of lactulose at different HRTs.
Fig.5. (a) Profiles of lactulose concentrations, (b) Pspec. and (c) selectivity values of batch and continuous lactulose productions using β-galactosidase from Kluyveromyces lactis, [sugar concentration] = 500 g/L, [E] = 300 U, mL/mF ratio = 0.5, phosphate buffer pH 6.8, stirred at 200 rpm, 40 °C.
566
21
567 568
Continuous synthesis of lactulose in an enzymatic membrane reactor reduces lactulose secondary hydrolysis
569
Azis Boing Sitanggang1,3*, Anja Drews2, Matthias Kraume1
570 571 572
1
Chair of Chemical and Process Engineering, Technische Universität Berlin. Ackerstraße 76, 13355 Berlin, Germany. 2
573 574 575 576
3
HTW Berlin - University of Applied Science, Engineering II, School of Life Science Engineering. Wilhelminenhofstraße 75A, 12459 Berlin, Germany.
Department of Food Science and Technology, Bogor Agricultural University. Raya Darmaga St, Kampus IPB Darmaga Bogor 16680, West Java, Indonesia.
577 578 579 580 581 582 583
*Corresponding Author:
584
Name : Azis Boing Sitanggang, MSc
585
Email :
[email protected] 586
Phone : +49-(030) 314-72693
587
Fax
588 589
Chair of Chemical and Process Engineering, Technische Universität Berlin, Ackerstraße 76, 13355 Berlin, Germany
: +49-(030) 314-72756
590 591 592 593 594 595 596 22
597
Highlights
598 599
Continuous operation was carried out at precisely controlled flux (i.e., HRT).
600
Secondary hydrolysis decreased by removal of lactulose during reaction.
601
Enhanced productivity of lactulose synthesis was achieved in continuous operation.
602
An efficient reduction of enzyme consumption in continuous synthesis was realised.
603
Continuous lactulose production for ~ 100 h is amenable.
604 605
23