SHORT COMMUNICATION

Feeding strategies for the enhanced production of a-arbutinin the fed-batch fermentation of Xanthomonas maltophilia BT-112

Chunqiao Liu • Peng Zhang • Shurong Zhang •

Tao Xu • Fang Wang • Li Deng

Received: 17 December 2012 / Accepted: 17 May 2013

� Springer-Verlag Berlin Heidelberg 2013

Abstract To develop a cost-effective method for the

enhanced production of a-arbutin using Xanthomonas

maltophilia BT-112 as a biocatalyst, different fed-batch

strategies such as constant feed rate fed-batch, constant

hydroquinone (HQ) concentration fed-batch, exponential

fed-batch and DO-control pulse fed-batch (DPFB) on

a-arbutin production were investigated. The research

results indicated that DPFB was an effective method for

a-arbutin production. When fermentation with DO-control

pulse feeding strategy to feed HQ and yeast extract was

applied, the maximum concentrations of a-arbutin and cell

dry weight were 61.7 and 4.21 g/L, respectively. The

a-arbutin production was 394 % higher than that of the

control (batch culture) and the molar conversion yield of

a-arbutin reached 94.5 % based on the amount of HQ

supplied (240 mM). Therefore, the results in this work

provide an efficient and easily controlled method for

industrial-scale production of a-arbutin.

Keywords a-Arbutin � Fed-batch fermentation �Xanthomonas maltophilia � Hydroquinone � DO-control

pulse fed-batch

Introduction

a-Arbutin, a glycosylated hydroquinone (HQ), is com-

mercially used in the cosmetic industry. It has inhibitory

function against tyrosinase, a critical enzyme for generat-

ing pigments, which leads to the prevention of melanin

formation, resulting in a whitening effect on the skin [1]. It

was reported that the human tyrosinase inhibition of

a-arbutin was much more effective than its isomer, b-arbu-

tin, and the whitening effect of a-arbutin was more than 10

times higher than b-arbutin [2]. Worldwide demand for

a-arbutin is increasing year by year, however whether

a-arbutin can compete with b-arbutin is exclusively depen-

dent on its cost in commercial production.

Although b-arbutin is found in various plants including

bearberry, wheat, and pear, a-arbutin is mainly produced

by enzymatic synthesis. To date, many attempts have been

made to synthesize a-arbutin. In general, two approaches

were employed. One was to use whole cells such as

Bacillus subtilis [3] and Xanthomonas campestris [4], and

the other was to employ carbohydrate-active enzymes,

including sucrose phosphorylase and dextransucrase from

Leuconostoc mesenteroides [5, 6], a-glucosidase from

X. campestris and Saccharomyces cerevisiae [7, 8], and

amylosucrase from Deinococcus geothermalis [9]. How-

ever, there are still several issues that need to be addressed

to produce a-arbutin within the targeted cost, such as

enhancing the HQ tolerance of biocatalyst and lowering the

costs of synthesis processes. Therefore, it is very desirable

to establish a simple alternative method for a-arbutin

synthesis with high efficiency.

Fed-batch culture is a batch culture fed continuously or

sequentially with substrate without the removal of fer-

mentation broth, which is generally superior to batch and

continuous processing, and is especially beneficial when

Chunqiao Liu, Peng Zhang contributed equally to this work.

C. Liu (&) � P. Zhang � S. Zhang � T. Xu � F. Wang � L. Deng

Beijing Bioprocess Key Laboratory, College of Life Science

and Technology, Beijing University of Chemical Technology,

Beijing 100029, China

e-mail: liuchunqiao010@yahoo.com.cn

L. Deng

Amoy-BUCT Industrial of Bio-technovation Institute,

Amoy 361022, China

123

Bioprocess Biosyst Eng

DOI 10.1007/s00449-013-0980-9

changing substrate concentrations affect the biomass and

production of the desired product [10]. HQ, a glucosyl

acceptor of a-arbutin, induces apoptosis in vivo by

changing the cellular redox status by reducing the cellular

thiol level and increasing the cellular reactive oxygen

species level [11]. When the concentration of HQ in a

medium exceeds a certain value, the cells will undergo

apoptosis. Since the high concentration of HQ during cul-

tivation process was unfavorable for the growth of cells,

accurate regulation of the HQ concentration in microor-

ganism fermentation is necessary to prevent accumulation

of HQ to a toxic level. However, to our knowledge, until

now the research of fed-batch fermentation for high con-

centration and efficient production of a-arbutin is not

addressed in any publication.

In our previous experiments, Xanthomonas maltophilia

BT-112 was employed to perform a bioconversion reac-

tion (a-arbutin synthesis) using sucrose as a glucosyl

donor and HQ as an acceptor (scheme 1) [12].To relieve

the shock of HQ on the cell for high concentration of

a-arbutin by X. maltophilia BT-112, an accurate control

of HQ supply is necessary. The influences on biomass and

a-arbutin content from different fed-batch strategies

including constant feed rate fed-batch (CRFB), constant

HQ concentration fed-batch (CHFB), exponential fed-

batch (EFB) and DO-control pulse fed-batch (DPFB) were

studied in this paper.

Materials and methods

Microorganism and culture medium

The X. maltophilia BT-112 strain used in all our experi-

ments was screened by our lab [13]. It was maintained in a

medium consisting (in g/L) 10 sucrose, 10 peptone, 5 yeast

extract, 0.5 MgSO4, 1 K2HPO4, 1 KH2PO4, 2 NaCl and 15

agars. The medium for cell growth or inoculum preparation

contained the following (in g/L): 20 sucrose, 5 peptone, 3

yeast extract, 0.5 MgSO4, 1 K2HPO4, 1 KH2PO4 and 2

NaCl. The pH was adjusted to seven prior to sterilization at

121 �C for 20 min. In batch and fed-batch fermentations,

the medium was the same as that for cell growth with

additional HQ and yeast extract feeding.

Fermentation conditions and methods

All experiments were carried out in a 5 L jar fermenter

(Shanghai Baoxing Bioengineering Equipment Ltd.,

Shanghai, China) with an initial broth volume of 2 L at

30 �C, and the agitation speed was set at 300 rpm (revo-

lutions per min) with airflow rate of 1.0 vvm (volume of air

per volume of culture and per min) to ensure complete

mixing of the fermentation broth. The inoculum was

incubated at 30 �C with 150 rpm for 15 h on a shaking

incubator (Taicang City Experimental Equipment Factory,

Suzhou, China) before inoculation into the 5 L fermenter

with 10 % inoculum volume. The dissolved oxygen (DO)

was measured with an autoclavable O2 sensor (Mettler-

Toledo Process Analytical, Inc., Greifensee, Switzerland).

Substrates, 250 mL sucrose (4.8 M) and 250 mL HQ

(2.4 M), were added according to detailed arrangement of

each experiment after 12 h fermentation. During the pro-

cess, the concentrations of cells and a-arbutin were mea-

sured in the culture medium.

Batch fermentation: 250 mL sucrose solution (sucrose

concentration in the culture medium was 480 mM) and

250 mL HQ solution (HQ concentration in the culture

medium was 240 mM) were pumped into the fermenter at

once and lasted for 72 h. In the course of the entire fer-

mentation, nothing was added.

All fed-batch fermentations were initiated as a batch

culture with 250 mL sucrose solution (sucrose concentra-

tion in the culture medium was 480 mM), and the feeding

substrate was pumped into the fermenter using a computer

coupled peristaltic pump. In CRFB fermentation, 250 mL

HQ solution was pumped into the fermenter at a feeding

rate of 10 mL/h. In CHFB fermentation, the residual HQ

concentration in the culture medium was maintained in

40 mM by feeding HQ solution. In DPFB fermentation,

feeding of HQ solution commenced (10 mL/h) when the

DO decreased to 10 % air saturation and continued until

the DO increased to 30 % air saturation. When fermenta-

tion with DO-control pulse feeding strategy began to feed

Scheme 1 Synthesis of a-arbutin from hydroquinone and sucrose by X. maltophilia BT-112

Bioprocess Biosyst Eng

123

HQ and yeast extract, 250 mL HQ solution with 6.25 g

yeast extract (2.5 g/L in the original culture medium) were

pumped into the fermenter by DPFB method.

In EFB fermentation, the nutrient feeding rate can be

determined by Eq. (1), which is derived from a mass bal-

ance with the assumption of a constant cell yield on sub-

strate and constant maintenance coefficient throughout the

fermentation [14, 15]; thus

Fm tð Þ ¼ lYX=S

þ m

� �x t0ð ÞV t0ð ÞX0elðt�t0Þ ð1Þ

where Fm(t) is the mass flow rate of substrate at time

t (g/h), l is the specific growth rate (h-1), YX/S is the

theoretical cell yield on substrate [g dry-cell-weight

(DCW)/g], m is the specific maintenance coefficient (g/g

DCW h), t0 is the time at which feeding is started, and x(t0)

and V(t0) are cell concentration (g DCW/L) and culture

volume (L) at t0, respectively. The feeding HQ was fed into

reactor using a computer coupled peristaltic pump by the

feeding rate determined by Eq. (1).

Analytical assays

Biomass concentration was determined by DCW. Five

milliliters of fermentation broth was centrifuged at

4,200 rpm for 20 min. The pellet was washed twice with

5 mL of distilled water and dried at 105 �C to constant

weight.

a-Arbutin and HQ were determined by a HPLC

system (Shimadzu LC-10ATvp, Kyoto, Japan) with a

reversed phase C18 column (250 9 4.6 mm, 5 lm,

DiamodsilTM). The mobile phase was composed of

methanol aqueous solution at a volumetric ratio of 5:95,

and run at a flow rate of 1.0 mL/min. Absorbance

detection wavelength was set at 280 nm. Each sample

was filtered through 0.45 lm micro-membrane, and a

10 lL of the resulting filtrate was loaded into the HPLC

system for a single run. Each run of culture experiments

and analysis was replicated thrice. The working cali-

bration curve on a-arbutin and HQ standard solution

showed good linearity over the range of 0.5–10.0 and

0–5.0 g/L, respectively. The regression line for a-arbu-

tin and HQ were Yarbutin = 1,000,000Xarbutin ? 59,668

(R2 = 0.9996) and YHQ = 2,000,000XHQ ? 67,778

(R2 = 0.9993) respectively, where Yarbutin (YHQ) and

Xarbutin (XHQ) are the peak area and the concentration of

a-arbutin (HQ) (g/L), respectively.

The conversion yield of a-arbutin is given by:

g %ð Þ ¼ M1 �M2

M1

� 100 % ð2Þ

where M1 is the mole of HQ before transformation (mol)

and M2 is the remaining mole of HQ after the conversion

(mol).

Results and discussion

From the results of different fermentation methods

(Table 1), the batch culture had the lowest biomass, con-

version and a-arbutin production than others. This could

have resulted from the toxicity of high HQ concentration in

the batch culture condition [11].Using fed-batch fermen-

tation to maintain HQ at a low level in fermentation broth,

the inhibitory effect of HQ on a-arbutin production was

avoided and the process efficiency was greatly enhanced.

Comparing the data shown in Table 1, the CRBF, EFB

and CHFB strategies were not satisfactory too. The DPFB

strategy had better results in a-arbutin concentration

(55.2 g/L), while the DO-control pulse feeding HQ and

yeast extract had the best a-arbutin outcome (61.7 g/L).

According to the results of CRBF, EFB and CHFB strat-

egies (details not shown in Table 1), CRFB strategy was

easy to operate, however, it could not combine very well

with the cell growth online. To avoid the problem of CRFB

fermentation, EFB strategy was taken into consideration.

The disadvantage of using EFB strategy was that it was

very difficult to control the process in time, because cell

growth would deviate from its original growth orbit when

HQ was added. In CHFB, since the HQ concentration was

very difficult to determine online, the problem of this fed-

batch lies in the lag of HQ adjustment.

To address these problems in CFFB, EFB and CHFB

strategies, an important parameter in microbial fermenta-

tion, DO concentration, which can accurately and timely

reflect the actual cell growth, was considered as an adjusted

control parameter in the X. maltophilia BT-112 fermenta-

tion. The process of a-arbutin production by DPFB

Table 1 Comparison of different fed-batch strategies

Biomass

(g/L)

Conversion

(%)

a-Arbutin

(g/L)

Batch culture 2.18 19.2 12.5

Constant feed rate fed-batch 3.54 54.3 35.4

Constant HQ concentration

fed-batch

3.39 62.1 40.5

Exponential fed-batch 3.68 71.8 46.9

DO-control pulse fed-batch Ia 3.78 84.5 55.2

DO-control pulse fed-batch IIb 4.21 94.5 61.7

a a-Arbutin production by DO-control pulse feeding HQb a-Arbutin production by DO-control pulse feeding HQ and yeast

extract

Bioprocess Biosyst Eng

123

fermentation was investigated and the results are shown in

Fig. 1. In the early phase of fermentation (0–32 h), the

feeding of HQ will cause DO increase for a short interval

because of its toxicity to cells (cell growth was repressed).

When HQ feeding finished (32 h), cells continued to grow

with the bioconversion of HQ and DO remained relatively

constant at 23 % (32–56 h). After 56 h of fermentation,

increase in biomass and a-arbutin content slows down and

the DO began to rise At the end of fermentation, the

maximal biomass and concentration of a-arbutin were 3.78

and 55.2 g/L, respectively. The molar conversion yield of

a-arbutin based on the amount of HQ supplied reached

84.5 %. In DPFB fermentation, the feeding of HQ was an

appropriate value and the conversion yield of a-arbutin was

enhanced.

Yeast extract which work as a supplier of vitamins,

growth factors can affect microbial growth and its bio-

logical activity. A number of studies show that higher cell

concentration could be obtained with adding yeast extract

in microbial fermentation [16–18]. To improve the cell

density thereby increasing the bioconversion of HQ, the

yeast extract (yeast extract concentration in the culture

medium was 2.5 g/L) and HQ solution were pumped into

the fermenter by DPFB method (results are shown in

Fig. 2). According to Fig. 2, during the course of fed-

batch fermentation, the DCW increased rapidly and the

maximum cell growth (4.21 g/L) was obtained after 60 h

of fermentation. This clearly shows that controlling

nutrient (yeast extract) concentration in an optimal range

is an efficient way of cultivating cells to higher concen-

tration. At the end of fermentation, the final concentration

of a-arbutin reaches 61.7 g/L with a molar conversion

yield of 94.5 % based on the amount of HQ supplied

(240 mM). In DO-control pulse feeding, both HQ and

yeast extract were added to an appropriate amount based

on cell growth. Comparing with the traditional batch

culture (Table 1), the yield of a-arbutin and maximal cell

dry weight enhanced 394 and 93 %, respectively. Addi-

tionally, the fed-batch method used in the present work

was simple and easily controlled. Thus, this method is an

efficient way of cultivating cells and producing a-arbutin

to higher concentration.

Fig. 1 Effects of DO-control pulse fed-batch fermentation. The

fermentation was carried out in a 5 L jar fermenter with an initial

broth volume of 2 L at 30 �C with 300 rpm for 72 h

Fig. 2 Effects of DO-control pulse feeding hydroquinone and yeast

extract. The fermentation was carried out in a 5 L jar fermenter with

an initial broth volume of 2 L at 30 �C with 300 rpm for 72 h

Table 2 Comparison of a-arbutin synthesis by different biocatalysts

Biocatalyst Donor HQ

(mM)

D:HQa

(mol)

Conversion

(%)

Production

(g/L)

Productivity

(g/L/h)

Reference

Sucrose phospholylase (L. mesenteroides) Sucrose 18 5:1 46.5 2.3 0.15 Kitao and Sekine [5]

Dextransucrase (L. mesenteroides) Sucrose 450 1:2 0.4 0.5 0.08 Seo et al. [6]

a-Glucosidase (X. campestris) Maltose 45 27:1 55.6 6.8 0.19 Sato et al. [7]

a-Glucosidase (S. cerevisiae) Maltose 9 167:1 13 0.4 0.02 Prodanovic et al. [8]

Amylosucrase (D. geothermalis) Sucrose 23.6 10:1 90 5.8 0.24 Seo et al. [9]

B. subtilis strain X-23 G5b 9 1:2 24.8 0.6 0.04 Nishimura et al. [3]

X. campestris WU-9701 cells Maltose 45 27:1 93 11.4 0.32 Kurosu et al. [4]

X. maltophilia strain BT-112 Sucrose 240 2:1 94.5 61.7 0.86 This study

a The molar ratio of donor: hydroquinoneb G5 is maltopentaose

Bioprocess Biosyst Eng

123

Previously, many approaches have been employed to

synthesize a-arbutin. Among the data reported by several

researchers (Table 2) [3–9], the highest production of

a-arbutin to date was obtained by a Japanese research group

[4] who used lyophilized X. campestris WU-9701 cells as a

biocatalyst with 45 mM HQ and 1.2 M maltose. The pro-

duction of a-arbutin in their research was 11.4 g/L which is

lower than our result (61.7 g/L). Furthermore, the molar

conversion yield of a-arbutin reached 94.5 % based on the

amount of HQ supplied (240 mM) by using DO-control

pulse feeding HQ and yeast extract strategy fermentation.

Therefore, the present investigation is a promising process

and feasible for industrial production of a-arbutin.

Conclusions

This study shows that fed-batch fermentation has the

potential to provide a cost-effective and efficient method

for a-arbutin production. DPFB was selected as the most

suitable strategy for transglycosylation of HQ. Using DO-

control pulse feeding HQ and yeast extract strategy fer-

mentation, the maximal HQ tolerance of cells and con-

centration of a-arbutin were 240 mM and 61.7 g/L,

respectively. The a-arbutin production was 394 % higher

than that of the control (batch culture), indicating a

potential for reducing cost. Furthermore, this method was

feasible and easily controlled. Such a fed-batch strategy is

very promising from an industrial perspective for a-arbutin

production and would provide reference for other similar

reaction systems.

Acknowledgments This project has been funded by the Key Pro-

jects in the National Science & Technology Pillar Program during the

12th 5 years Plan Period (2011BAD22B04), the National Natural

Science Foundation of China (21246005), and the National Basic

Research Program of China (973 program) (2013CB733600).

References

1. Funayama M, Arakawa H, Yamamoto R (1995) Effects of a- and

b-arbutin on activity of tyrosinases from mushroom and mouse

melanoma. Biosci Biotech Biochem 59:143–144

2. Kazuhisa S, Takahisa N, Koji N (2004) Inhibitory effects of

a-arbutin on melanin synthesis in cultured human melanoma cells

and a three-dimensional human skin model. Biol Pharm Bull

27:510–514

3. Nishimura T, Kometani T, Takii H, Terada Y, Okada S (1994)

Purification and some properties of a-amylase from Bacillus

subtilis X-23 that glucosylates phenolic compounds such as

hydroquinone. J Ferment Bioeng 78:31–36

4. Kurosu J, Sato T, Yoshida K, Tsugane T, Shimura S, Kirimura K,

Kino K, Usami S (2002) Enzymatic synthesis of a-arbutin by a-

anomer-selective glycosylation of hydroquinone using lyophi-

lized cells of Xanthomonas campestris Wu-9701. J Biosci Bioeng

93:328–330

5. Kitao S, Sekine H (1994) a-D-Glucosyl transfer to phenolic

compounds by sucrose phosphorylase from Leuconostoc mesen-teroides and production of a-arbutin. Biosci Biotech Biochem

58:38–42

6. Seo ES, Kang J, Lee JH, Kin GE, Kim GJ, Kim D (2009) Syn-

thesis and characterization of hydroquinone glucoside using

Leuconostoc mesenteroides dextransucrase. Enzyme Microb

Technol 45:355–360

7. Sato T, Hasegawa N, Saito J, Umezawa S, Honda Y, Kino K,

Kirimura K (2012) Purification, characterization, and gene

identification of an a-glucosyl transfer enzyme, a novel type a-

glucosidase from Xanthomonas campestris WU-9701. J Mol

Catal B Enzym 80:20–27

8. Prodanovic R, Milosavic N, Sladic D, Zlatovic M, Bozic B,

Velickovic TC, Vujcic Z (2005) Transglucosylation of hydro-

quinone catalysed by a-glucosidase from baker’s yeast. J Mol

Catal B Enzym 35:142–146

9. Seo DH, Jung JH, Ha SJ, Cho HK, Jung DH, Kim TP, Baek NI,

Yoo SH, Park CS (2012) High-yield enzymatic bioconversion of

hydroquinone to a-arbutin, a powerful skin lightening agent, by

amylosucrase. Appl Microbiol Biotechnol 94:1189–1197

10. Lee J, Lee SY, Park S, Middelberg APJ (1999) Control of fed-

batch fermentations. Biotechnol Adv 17:29–48

11. Shen DX, Shi X, Wang Y, Fu JL, Zhou ZC (2003) Inhibitory

effect of thioredoxin on cytotoxicity of hydroquinone. Chin J

Pharmacol Toxicol 17:55–60

12. Liu CQ, Zhang SR, Zhang P (2006) Biocatalytic synthesis of a-

arbutin by Xanthomonas BT-112. Chin J Catal 27:361–364

13. Liu CQ, Zhang SR, Zhang P (2006) Mutation breeding of Xan-thomonas maltophilia for the synthesis of a-arbutin. J Beijing

Univ Chem Technol 33:13–16

14. Yang XM, Xu L, Epstein L (1992) Production of recombinant

human interferon-a by Escherichia coli using a computer-con-

trolled cultivation process. J Biotechnol 23:291–301

15. Nor ZM, Tamer MI, Scharer JM, Moo-Young M, Jervis EJ (2001)

Automated fed-batch culture of Kluyveromyces fragilis based on

a novel method for on-line estimation of cell specific growth rate.

Biochem Eng J 9:221–231

16. Zhang XL, Xia ZJ, Liu Y, Zhao BX, Zheng XG, Xiong ZH

(2001) Effects of nutrition and its feeding strategy on the

expression of a-amylase in fed-batch cultivation of recombinant

yeast. Chem Res Chinese Universities 15:144–148

17. Liu XL, Xiao CM, Xu YT, Qi W, Du LX (2001) High-density

cultivation of Bacillus coagulans using fed-batch method. Bio-

technology 11:20–23

18. Yu EK, Saddler JN (1983) Fed-batch approach to production of

2,3-butanediol by Klebsiella pneumoniae grown on high substrate

concentrations. Appl Environ Microbiol 46:630–635

Bioprocess Biosyst Eng

123