Process Biochemistry 41 (2006) 293–298
Medium optimization of carbon and nitrogen sources for the
production of eucalyptene A and xyloketal A from Xylaria sp. 2508
using response surface methodology
Zeng Xiaobo a,*, Wang haiying b, He Linyu c, Lin Yongcheng c, Li Zhongtao a
College of Pharmacy, Wuhan University, 430072 Wuhan, PR China
College of Life Science, South-center University for Nationalities, 430074 Wuhan, PR China
School of Chemistry & Chemical Engineering, Sun Yat-sen University, 510275 Guangzhou, PR China
Received 30 March 2005; received in revised form 6 June 2005; accepted 7 June 2005
To optimize the production of novel ketal compounds and allenic ethers from a mangrove fungus Xylaria sp. 2508, optimization of the
culture medium was carried out using a two-step approach. A quick identiﬁcation of a suitable source of carbon and nitrogen was obtained by a
simple screening experiment, which was followed by an application of response surface methodology (RSM) for further optimization design.
A second order model was found to ﬁt the eucalyptene A production. Response surface analysis revealed that the optimum values of the tested
variables for the production of eucalyptene A were: glucose 2.86% (w/v), peptone 0.78% (w/v) and yeast extract 0.35% (w/v). A production of
4.89 Â 10À7 g mlÀ1 for eucalyptene A, which was close to the predicted response, was observed in veriﬁcation experiment. The production of
xyloketal A was 2.20 Â 10À7 g mlÀ1. In comparison to their production of basal medium, 1.53- and 1.32-fold increase had been obtained,
respectively. The 3-D response surface curves implied that high yeast extract concentration repressed the growth of Xylaria sp.
# 2005 Elsevier Ltd. All rights reserved.
Keywords: Medium optimization; Ketal compounds; Allenic ethers; Xylaria sp.; Response surface methodology
1. Introduction fungus. The xyloketals represent a series of novel ketal
compounds having a close biogenetic relationship. Xyloalle-
After almost 40 years of investigation, the number of nolide A has a novel cyclic peptide containing an allenic ether
compounds isolated from various marine organisms has of a N-( p-hydroxycinnamoyl) amide. Together with xyloal-
virtually soared and exceeds 10,000 , with hundreds of lenolide A there were two other allenic ethers found from the
new compounds still being discovered every year [2,3]. In spent culture medium . One of them, a known compound
2001 the number of new compounds and new stereo- eucalyptene A, exhibited antifungal activity. This was
chemical structures reported is 683 . Some of these revealed by means of bioautography on Cladosporium
marine nature products hold great potential for the cladosporioides for amounts as low as 50 mg . In
development of new and much needed drugs. the primary bioassay, xyloketal A inhibited acetylcholine
The mangrove fungus Xylaria sp. 2508, which was esterase at 1.5 Â 10À6 mol lÀ1 and xyloallenolide A depress-
collected from seeds of an angiosperm tree and identiﬁed as ed the growth of fungus [5,6]. The chemical conﬁgurations
Xylaria species (Ascomycota), was found to produce rich of xyloketal A, xyloallenolide A and eucalyptene A are
secondary metabolites. Xyloketals A–E  and xyloalleno- shown in Fig. 1. Further biological evaluation of them is in
lide A  were isolated from the spent culture medium of this progress.
Considering the potential bioactivities of the secondary
* Corresponding author. Tel.: +86 27 6875 2339/139 7168 2309;
products of the fungus Xylaria sp. 2508 we have attempted to
fax: +86 27 6875 4629. study and maximize the production of them. Preliminary
E-mail address: email@example.com (Z. Xiaobo). experiments indicated that eucalyptene A had a yield about 10
1359-5113/$ – see front matter # 2005 Elsevier Ltd. All rights reserved.
294 Z. Xiaobo et al. / Process Biochemistry 41 (2006) 293–298
Fig. 1. Chemical conﬁguration of xyloketal A, xyloallenolide A and eucalyptene A.
times of xyloallenolide A. It was also found that among and repeated twice at least. The mean values and the
xyloketals A–E, xyloketal A has the maximum outcome. standard deviation were calculated from the data obtained
Thus, in this experiment, we chose eucalyptene A and with these trials. The standard deviation in results was within
xyloketal A as targets to determine the production of ketal 10%.
compounds and allenic ethers from Xylaria sp. 2508,
respectively. 2.2. Micro-organism
The ‘one-at-a-time-approach’ is an operation frequently
used in biotechnology to obtain high yields of the desired The strain of the fungus Xylaria sp. (no. 2508) was
metabolic products in a microbial system. This method is not isolated from seeds of a mangrove angiosperm tree in Hong
only time consuming, but disregards the complex interac- Kong. It was stored at the College of Life Science, Sun Yat-
tions among various physicochemical parameters [8,9]. To sen University, Guangzhou, China.
overcome this difﬁculty, experimental factorial design and
response surface methodology (RSM) can be employed to 2.3. Inoculum preparation
optimize the medium components. The response surface
methodology, which includes factorial design and regression Starter cultures were maintained on cornmeal seawater
analysis, can build models to evaluate the effective factors agar (glucose 0.010 g mlÀ1, cornmeal 0.010 g mlÀ1, pep-
and study their interaction and select optimum conditions in tone 0.001 g mlÀ1 and agar 0.015 g mlÀ1, dissolved in fresh
a limited number of experiments [10–12]. Here, we report seawater). The agar with mycelium was transferred
statistical optimization of culture medium for enhanced aseptically to a 250 ml Erlenmeyer ﬂask containing
production of ketal compounds and allenic ethers from 100 ml of basal liquid medium (glucose 0.010 g mlÀ1,
Xylaria sp. 2508. peptone 0.002 g mlÀ1, yeast extract 0.001 g mlÀ1, NaCl
0.030 g mlÀ1). The ﬂask was incubated at 30 8C on a rotary
shaker (150 rpm) for 5–7 days. The mycelium was
2. Materials and methods aseptically transferred to a 1 l ﬂask containing 300 ml of
liquid medium for production experiments.
2.1. Chemicals and experimental statistics
2.4. Analytical methods
Isopropyl alcohol for HPLC was purchased from Merck
(Darmstadt, Germany). All other chemicals used were of Growth of Xylaria sp. 2508 was determined by weighting
analytical grade commercially available in China. All the the mycelia which was centrifuged (3000 Â g 10 min, at
experiments were carried out independently in triplicates ambient temperature) and washed twice with distilled water.
Z. Xiaobo et al. / Process Biochemistry 41 (2006) 293–298 295
Xyloketal A and eucalyptene A were analyzed by HPLC. levels (À1, 0, +1). For glucose, the corresponding actual
Culture medium from one ﬂask was centrifuged and the values of three coded levels were 1, 3 and 5%. For peptone
volume of supernatant was metered to 300 ml. Then 3 ml of they were 0.2, 1, 1.5% and for yeast extract they were
this liquor, 3 ml methanol and 50 ml isopropyl alcohol 0.2, 0.5, 1%. A series of 15 experiments including three
solution (4.187 Â 10À4 g mlÀ1, dissolved in methanol) were center points were performed. The minimum and maximum
mixed and centrifuged (5000 Â g, at ambient temperature, range of independent variables investigated and the full
5 min) for detection. HPLC (Agilent 1100 system) was experimental plan with respect to their coded forms are listed
carried out with a C-18 column (3.9 mm Â 300 mm, in Table 1.
Nova-Pak, Waters) and 80% MeOH (v/v, methanol/water) Cultures were stably incubated at 30 8C for 25 days. A
as the mobile phase at the ﬂow rate of 0.5 ml minÀ1. multiple regression analysis of the data was carried out to
To determinate the absorbance, the UV–vis detector was set obtain empirical models that deﬁne response (Y1 and Y2) in
to 220 nm. Isopropyl alcohol was used as an internal terms of the independent variables. For a three-factor
standard substance. Its relative correction factors to system, the following second-order polynomial equation
xyloketal A and eucalyptene A were 0.5288À1 and was then applied to the data by the multiple regression
1.3609À1, respectively. procedure:
2 2 2
2.5. Optimization of xyloketal A and eucalyptene A Y ¼ a0 þ a1 X1 þ a2 X2 þ a3 X3 þ a11 X1 þ a22 X2 þ a33 X3
production from Xylaria sp. 2508 þ a12 X1 X2 þ a13 X1 X3 þ a23 X2 X3 (1)
2.5.1. Selection of most suitable carbon and nitrogen with Y, being the predicted response; a0, intercept; a1, a2, a3,
sources by one-variable-at-a-time approach linear coefﬁcients; a11, a22, a33, squared coefﬁcients; a12,
For selection of the best source of carbon and nitrogen for a13, a23, interaction coefﬁcients.
xyloketal A and eucalyptene A production, various simple Using the above model, a SAS software package was
and complex carbon sources (glucose, sugar, lactose, used to analyze the variance and to obtain the optimum
maltose, glycerol, dextrin, starch, glyceryl monostearate) concentration of the medium components. The response
and inorganic and complex nitrogen sources (sodium nitrate, surface graphs were plotted to understand the effect of the
ammonium sulphate, urea, peptone, yeast extract, casein) selected variables individually and in combination and to
were used as substitute individually for carbon or nitrogen determine their optimum levels for maximal xyloketal A and
source in the basal liquid medium. In one test, one of the eucalyptene A production. All the experiments were carried
source of carbon (glucose) and nitrogen (peptone and yeast out independently in triplicates and the average values were
extract) in the basal liquid medium was replaced by matters presented.
mentioned above with a ﬁnal concentration of 2% (w/v) for
carbon source and 0.2% (w/v) for nitrogen source, while
other ingredients kept original concentrations. Before 3. Results
autoclaving, the pH value of the medium was adjusted
to 8.0 by adding 1 mol lÀ1 NaOH or 1 mol lÀ1 HCl. 3.1. Selection of most suitable carbon and nitrogen
Erlenmeyer ﬂasks (1000 ml) containing 300 ml of the sources by one-variable-at-a-time approach
culture medium were inoculated with 1% (w/v) fresh
mycelium. Yields of xyloketal A and eucalyptene A were Among the various carbon sources studied, Xylaria sp.
determined after 25-day incubation at 30 8C under stationary 2508 produced more eucalyptene A and xyloketal A in the
culture. presence of glucose and sucrose. This is shown on Table 2. It
is also shown that complex organic nitrogen sources, viz.,
2.5.2. Experimental design and optimization by RSM peptone, yeast extract and casein are observed to induce the
Based on the results obtained in the previous experiments, production eucalyptene A and xyloketal A while simple
glucose, peptone and yeast extract were found to be inorganic sources support no production of them. In medium
determining variables in xyloketal A and eucalyptene A containing simple inorganic nitrogen sources and some
production. Hence, these variables were selected for carbon sources, Xylaria sp. 2508 could scarcely grow. In all
optimizing a higher lever of production of xyloketal A and carbon sources being tested, Xylaria sp. 2508 could only
eucalyptene A. For this purpose, the response surface yield eucalyptene A and xyloketal A in the medium
methodology by using a 3 Â 3 factorial central composite containing glucose and sucrose. Considering that the sucrose
design (CCD) proposed by Box and Behnken  was molecule is composed of glucose and fructose and sucrose
adopted. Three independent variables were glucose (w/v %, supported more mycelium growth but didn’t induce more
X1), peptone (w/v %, X2) and yeast extract (w/v %, X3), and the yield of product than glucose (Table 2), glucose was selected
dependent variables were production of eucalyptene A as the source of carbon. And peptone and yeast extract were
(10À7 g mlÀ1, Y1) and xyloketal A (10À7 g mlÀ1, Y2). Each selected for the source of nitrogen in the following
independent variable was studied at three different coded experiment.
296 Z. Xiaobo et al. / Process Biochemistry 41 (2006) 293–298
Experimental designs used in RSM studies and observed responses
Test number X1 (glucose) X2 (peptone) X3 (yeast extract) Mean observed responses
Eucalyptene A (10À7 g mlÀ1) Xyloketal A (10À7 g mlÀ1) Mycelium weight (g)
1 À1 À1 0 2.610 0.97 16.40
2 À1 0 À1 2.047 1.02 16.98
3 À1 0 +1 0.121 n.d. 15.96
4 À1 +1 0 1.121 n.d. 16.22
5 0 À1 À1 3.058 1.39 20.92
6 0 À1 +1 0.254 0.13 15.02
7 0 +1 À1 2.610 1.16 23.74
8 0 +1 +1 0.356 n.d. 14.08
9 +1 À1 0 0.506 0.27 18.12
10 +1 0 À1 2.847 0.92 23.96
11 +1 0 +1 3.210 n.d. 11.24
12 +1 +1 0 1.015 n.d. 14.54
13 0 0 0 4.738 1.33 19.81
14 0 0 0 4.310 1.27 16.76
15 0 0 0 4.201 1.10 18.69
‘n.d.’ indicates that no xyloketal A was detected.
To ﬁnd the optimum rang of their concentrations for the cannot be got by response surface methodology. For
production of eucalyptene A and xyloketal A, Xylaria sp. eucalyptene A, the regression equation was obtained by
2508 was cultured in the basal liquid medium with varying applying the response surface methodology.
one parameter at a time and keeping the others constant. The coefﬁcients of the regression equation were
Data obtained (not reported here) was used for design of calculated using the software SAS. The eucalyptene A
response surface methodology. production (Y1) can be expressed in terms of the following
3.2. Optimization by response surface methodology
Y1 ¼ 0:772357 þ 1:250454X1 þ 5:448226X2 þ 0:144725X3
The observed responses of RSM experiments for 2 2 2
À 0:360308X1 À 4:967917X2 À 4:798811X3
studying the effects of three independent variables, viz.,
þ 0:656866X1 X2 þ 0:856881X1 X3
glucose, peptone and yeast extract, on production of
eucalyptene A and xyloketal A are presented together with þ 1:052876X2 X3 (2)
the mycelium weight in Table 1. Because xyloketal A cannot
be detected in several experiments where concentrations of where Y1 is the eucalyptene A production; X1, glucose
nitrogen sources are higher relatively, its regression equation concentration; X2, peptone concentration; X3, yeast extract
concentration. The results of the second order response
Table 2 surface model ﬁtting in the form of analysis of variable
Effect of various carbon and nitrogen sources on eucalyptene A and (ANOVA) are given in Table 3. A good ﬁt of the regression
xyloketal A production by Xylaria sp. 2508 model was checked by the R2 and R value, signiﬁcance of
Nutrient source Eucalyptene A Xyloketal A Mycelium ‘‘total regress F-value’’ (5.36) and non-signiﬁcance of ‘‘lack
(10À7 g mlÀ1) (10À7 g mlÀ1) weight (g) of ﬁt F-value’’ (13.87).
Carbon sources The 3-D response surface curves were then plotted to
Glucose 3.81 1.64 15.79 explain the interaction of medium components and the
Sucrose 3.30 1.81 20.61 optimum concentration of each component required for the
Lactose n.d. n.d. 9.37
Maltose n.d. n.d. 7.92
production of eucalyptene A (Fig. 2). In response surface
Glycerol n.d. n.d. 3.36 curves two factors varied at a time when the other factor
Dextrin n.d. n.d. 16.58 maintained at a ﬁxed level (zero level). As showing in graph
Starch n.d. n.d. 4.04 A, B and C in Fig. 2, the eucalyptene A production is
Glyceryl monostearate n.d. n.d. 10.89 inﬂuenced by all three independent variables. The maximum
Nitrogen sources point was achieved when glucose, peptone, and yeast extract
Sodium nitrate n.d. n.d. 3.67 concentrations were near their zero level.
Ammonium sulphate n.d. n.d. 4.05
Urea n.d. n.d. 4.13
Casein 1.61 1.05 15.62 3.3. Validation of the experimental model
Peptone 3.28 1.79 24.76
Yeast extract 3.42 1.54 25.62 The regression equation of eucalyptene A production was
‘n.d.’ indicates that no eucalyptene A or xyloketal A was detected. solved by using SAS software. The optimal values of each
Z. Xiaobo et al. / Process Biochemistry 41 (2006) 293–298 297
Table 3 10À7 g mlÀ1) increase in production of eucalyptene A was
Analysis of variance for the regression model of medium optimization of achieved compared with that in the basal medium
eucalyptene A production obtained from the experimental results
(3.19 Â 10À7 g mlÀ1). In veriﬁcation experiment, a 1.32-
Sources Degree of Sum of Mean F-value
fold increase of xyloketal A production (2.20 Â 10À7
freedom squares square
g mlÀ1) was obtained in comparison to the basal medium
(1.67 Â 10À7 g mlÀ1). The result indicated that the opti-
Linear 3 7.3840 2.4613 3.50
Quadratic 3 21.1858 7.0619 10.03a mized medium favors the production of xyloketal A.
Crossproduct 3 5.3850 1.7950 2.55
Total regress 9 33.9549 3.7728 5.36a
Residual 4. Discussion
Lack of ﬁt 3 3.3583 1.1194 13.87
Pure error 2 0.1614 0.0807 This work has demonstrated the use of a multifactorial
Total error 5 3.5198 0.7040
statistical approach for determining the conditions that lead
R 0.9519 to the optimum yield of secondary metabolites production.
Signiﬁcant at 95% level. Central composite design maximizes the amount of
information that can be obtained, while considering the
interaction of independent variables and limiting the numbers
test component in coded units are as follows: X1 = of individual experiments required. The response surface
À0.067275, X2 = À0.114614 and X3 = À0.610045. Their methodology, a smaller and less time consuming experi-
actual values were: 2.86% (w/v) glucose, 0.78% (w/v) mental design, could generally satisfy the optimization of
peptone and 0.35% (w/v) yeast extract. The model predicted many microbial processes [8,10].
that the production of eucalyptene A could reach Though the RSM experiments was not designed for
4.70 Â 10À7 g mlÀ1 by using the above optimized concen- Xylaria sp. 2508 growth, the 3-D response surface curves
trations of the variables. Veriﬁcation experiments were were plotted to understand the interaction of the medium
accomplished by using the optimized conditions and gave a components and their effect on mycelium growth. Graph D
production of 4.89 Â 10À7 g mlÀ1 of eucalyptene A, which in Fig. 2 shows that high concentration of yeast extract
was closer to the predicted response of 4.70 Â 10À7 g mlÀ1. repressed the growth of Xylaria sp. At all levels (+1, 0, À1) of
This result therefore corroborated the predicted values, yeast extract concentration, a change of peptone concentra-
and the effectiveness of the model. A 1.53-fold (4.89 Â tion cannot signiﬁcantly affect growth. Since Xylaria sp.
Fig. 2. Response surface curve of eucalyptene A production and mycelium weight of Xylaria sp. 2508. Graph A, B, C and D showed the effect of yeast extract,
peptone and glucose concentration and their interaction. The 3-D response surface curves were plotted with two independent variables at a time, the other factor
was maintained at the zero level.
298 Z. Xiaobo et al. / Process Biochemistry 41 (2006) 293–298
2508 was isolated from seeds of a mangrove angiosperm (863 Program), no. 2001AA624010. The authors thank Judy
tree, it might be a symbiote of this tree and might absorb Van Bruggen from Canada for the English reviewing.
some nutrients, especially those that couldn’t be synthesized
by itself, from the tree tissue. Yeast extract contains complex
nutrients such as vitamin, nucleic acid, lipid and other References
substances, which might be necessary for Xylaria sp. 2508.
Proper yeast extract concentration is necessary for growth  Proksch P, Edrada RA, Ebel R. Drugs from the seas—current status
and microbiological implications. Appl Microbiol Biotechnol 2002;
and secondary metabolites of Xylaria sp. 2508.
As with many other experiments have mentioned,  Faulkner DJ. Marine pharmacology. Antonie Van Leeuwenhoek
insufﬁcient oxygen supply is one of the most frequently 2000;77:135–45.
occurring problem associated with the application of  Faulkner DJ. Marine natural products. Nat Prod Rep 2002;19:1–48.
Erlenmeyer ﬂask cultivation. In microbial cultures, the  Blunt JW, Copp BR, Munro MG, Northcote PT, Prinsep MR. Marine
typical reaction behaviors during oxygen limitation may be natural products. Nat Prod Rep 2003;20:1–48.
 Lin YC, Wu XY, Feng S, Jiang GC, Luo JH, Zhou SN, Vrijmoed LLP,
the slowing down of the whole metabolism of the ¨
Jones EBG, Krohn K, Steingrover K, Zsila F. Five unique compounds:
microorganisms and the microorganisms sensitively react- xyloketals from mangrove fungus Xylaria sp. from the South China
ing to the oxygen supply with their product metabolism Sea coast. J Org Chem 2001;66:6252–6.
[13,14]. In our experiment we encountered this problem.  Lin YC, Wu XY, Feng S, Jiang GC, Zhou SN, Vrijmoed LLP. Jones
The preliminary experiment indicated that the Xylaria sp. EBG. A novel N-cinnamoylcyclopeptide containing an allenic ether
from the fungus Xylaria sp. (strain c2508) from the South China Sea.
2508 mycelium grew slowly under stationary culture and the Tetrahedron Lett 2001;42:449–51.
maximum production of eucalyptene A and xyloketal A  Arnone A, Cardillo R, Nasini G, de Pava OV. Two cinnamic allenic
could be detected between the 23rd and 25th day. Earlier ethers from the fungus Clitocybe eucalyptorum. Phytochemistry
studies [5,6] have shown that in batch cultivation with 1993;32:1279–81.
 Elibol M. Optimization of medium composition for actinorhodin
aeration system (300-l bioreactor containing 170 l of GYT
production by Streptomyces coelicolor A3(2) with response surface
medium) the incubation time ended at 80 h because of methodology. Process Biochem 2004;39:1057–62.
exhaustion of carbon sources. This was a much quicker  Abdel-Fattah YR, Saeed HM, Gohar YM, El-Baz MA. Improved
process than a stationary culture in Erlenmeyer ﬂask. In this production of Pseudomonas aeruginosa uricase by optimization of
work, low concentrations of nutrients have caused low process parameters through statistical experimental designs. Process
production of eucalyptene A. On the other hand, higher Biochem 2005;40:1707–14.
 Chauhan B, Gupta R. Application of statistical experimental design for
concentrations may cause inhibition of secondary products optimization of alkaline protease production from Bacillus sp. RGR-
synthesis. This fact was also observed in other experiments 14. Process Biochem 2004;39:2115–22.
on carbon and nitrogen repression effects [10,15]. Thus,  Box GEP, Behnken DW. Some new three level design for the study of
aeration and nutrients supply are the key problems for the quantitative variables. Technometrics 1960;2:455–75.
 Tanyildizi MS, Ozer D D, Elibol M. Optimization of a-amylase
following studies of Xylaria sp. 2508 cultivation. However,
production by Bacillus sp. using response surface methodology.
controlled fed-batch cultivation and chemostat cultures have Process Biochem 2005;40:2291–6.
so far been used successfully for improving metabolites ¨
 Buchs J. Introduction to advantages and problems of shaken cultures.
production from a number of microorganisms for a long Biochem Eng J 2001;7:91–8.
period of time [16–18]. The optimum culture medium  Kamzolova SV, Shishkanova NV, Morgunov IG, Finogenova TV.
Oxygen requirements for growth and citric acid production of Yarro-
obtained in this experiment gives a basis for further study
wia lipolytica. FEMS Yeast Res 2003;3:217–22.
with batch or fed-batch cultivation in a bioreactor for large-  Giesecke UE, Bierbaum G, Rudde H, Spohn U, Wandrey C. Produc-
scale production of target secondary metabolites from tion of alkaline protease with Bacillus licheniformis in a controlled
Xylaria sp. 2508. fed-batch process. Appl Microbiol Biotechnol 1991; 35:720–4.
 Lee JK, Song JY, Kim SY. Controlling substrate concentration in fed-
batch Candida magnoliae culture increases mannitol production.
Biotechnol Prog 2003;19:768–75.
 Sarkar D, Modak JM. Optimisation of fed-batch bioreactors using
genetic algorithms. Chem Eng Sci 2003;58:2283–96.
 Singh J, Vohra RM, Sahoo DK. Enhanced production of alkaline
This work was supported by the National High proteases by Bacillus sphaericus using fed-batch culture. Process
Technology Research and Development Program of China Biochem 2004;39:1093–101.