J Surfact Deterg

DOI 10.1007/s11743-015-1668-8

ORIGINAL ARTICLE

Extraction and Fermentation-Based Purification of Saponins

from Sapindus mukorossi Gaertn.
Wu Heng • Zhang Ling • Wang Na •
Guo Youzhi • Weng Zhen • Sun Zhiyong •

Xu Deping • Xie Yunfei • Yao Weirong

Received: 24 March 2014 / Accepted: 2 January 2015

Ó AOCS 2015

Abstract In the present study, the extraction and purifi- area of Taiwan as well as in forests below 1,000 m in
cation of saponins from Sapindus mukorossi Gaertn. were China, India, and Japan [1]. The fruit of S. mukorossi
examined for effective utilization of the saponin resource. Gaertn. (Sapindus), better known as soapnuts in the tropical
Saponins were extracted from S. mukorossi Gaertn. using and sub-tropical regions of Asia [2], is generally used as a
water. The conditions of the water extraction process, in- commercial cleanser, and has been shown to possess
cluding extraction temperature, extraction time, number of medicinal properties, including antidermatophytic, anti-
times of extraction, and solvent-material ratio were opti- tussive, and antihelmintic activities [3]. The extract of the
mized. The yield of total Sapindus saponins (TSS) from the soapnut pericarps is widely used for hair and body cleaning
pericarp was 33.41 % and its purity in the extract was purposes in traditional Taiwanese culture, and is also em-
45.71 %. The saponin solution was further concentrated to ployed as an expectorant in folk medicine for cough relief,
1/6–1/7 of its original volume, and dried yeast BV818 that detoxification, and defervescence [4].
adapted to the concentrated Sapindus saponins solution Although many methods are currently available to ex-
(SW) was screened. The activation conditions, inoculum tract and purify saponins, including water extraction [5, 6],
amount, fermentation temperature, and fermentation period ethanol precipitation [7, 8], flocculation [9, 10], butanol
were optimized. By using the dried yeast under optimized extraction [11], use of macroporous resins [12], ultrafil-
conditions, the purity was increased to 75.50 %. The yield of tration [13], and foam separation [14, 15], they still possess
the byproduct ethanol was 5.33 % (w/v), while the content of many disadvantages. For example, in water extraction,
TSS in the final product decreased from 18.29 to 15.30 % although the extraction rate is higher, there will often be a
(w/v). These results could contribute to the development of number of monosaccharides, oligosaccharides, amino
industrial-scale production of Sapindus saponins. acids, proteins, mucus quality are extracted as well, which
decrease the purity of the saponins; in ethanol precipita-
Keywords Sapindus mukorossi Gaertn.  Saponins  tion, there was alcohol and energy consumption when re-
Extraction  Concentrate  Fermentation  Purification covering the solvent, easy to absorb moisture, higher
requirements for equipment and production conditions,
high cost and long processing; in flocculation, it is hard to
Introduction obtain a reasonable flocculant; in butanol extraction, there
is solvent consumption, the purity and yield are not high;
The Chinese soapberry tree (Sapindus mukorossi Gaertn.) with macroporous resin, ethanol recovery is not complete,
is a small deciduous tree, widely distributed in the lowland it is a complex process, high reagent and energy con-
sumption; in ultrafiltration, the film is easily blocked and
difficult to clean; in foam separation, it is inefficient, and
W. Heng  Z. Ling  W. Na  G. Youzhi  W. Zhen  prone to back mixing phenomena, difficult industrial pro-
S. Zhiyong  X. Deping  X. Yunfei  Y. Weirong (&)
duction, and could not be used for commercial production
School of Food Science and Technology, Jiangnan University,
1800 Lihu Avenue, Wuxi 214122, Jiangsu, China of saponins. Hence, a rapid and simple method for the
e-mail: yaoweirongcn@jiangnan.edu.cn separation of total Sapindus saponins (TSS) is needed.

123
 J Surfact Deterg

TSS can dissolve in water or ethanol, and, according to (98 % purity) was purchased from Sigma (Sigma Co. Ltd.,
the similarity principle of compatibility, water and ethanol St. Louis, MO, USA).
can be used as solvents for extraction. Although the yield
of TSS with ethanol extraction is slightly higher than that Extraction of TSS
with water, because of the high cost of ethanol extraction,
recovery needs, increasing energy consumption, and the Water, being inexpensive and safe, was used as the sol-
risk of flammability, water was selected for extraction in vent. A total of 100 g of pretreated pericarp of S.
this study. mukorossi powder (dry ground to 20–40 mesh) and a
Yeast fermentation to purify the Sapindus saponins after certain volume of water were added to a 1,000-mL flask
water extraction is a new method that has been reported and allowed to macerate under specific conditions. The
only in a few studies. Fermentation of crude Ficaria resulting extract was filtered through a gauze (Purchased
saponin by different fungal strains allows the elimination from a local pharmacy), and the residue obtained was
of free sugars [16] to obtain further purification of the main extracted twice with a solvent-material ratio of 1:2–1:3
saponin. Jin Yueqing, Zheng Jingcheng, and Gui Xiaohua for 1–2 h using the same procedure described earlier. The
[5] reported that tea saponin prepared by the purification factors such as extraction time, extraction temperature,
the techniques of water-extraction and fermentation is number of times of extraction, and solvent-material ratios
purer than that purified by general water-extraction. This were employed in single factor test, and the results are
combination technique also increases the extraction rate of shown in Fig. 1. The extract obtained was further
tea saponin, compared with ethanol or n-butanol extraction. analyzed.
Thus, in the present study, the water extraction process and WTSS
separation of Sapindus saponins from the pericarp of S. Yield % ¼  100 % ð1Þ
WPericarp
mukorossi by yeast fermentation were examined.
Fermentation can remove impurities such as sugars and WTSS ¼ CTSS  VExtract ð2Þ
proteins from Sapindus saponins, making them suitable for where Yield % represented the yield of total Sapindus
use in the chemical and pharmaceutical industries. Fur- saponins (TSS), WTSS (mg) was the weight of TSS in the
thermore, the ethanol formed as a byproduct of fermenta- extract and WPericarp (g) was the weight of S. mukorossi
tion can be processed to obtain ethanol fuel, an additional pericarp, respectively. CTSS (mg/mL) was the concentra-
energy source, which is environmentally friendly and tion of TSS in the extract obtained, and VExtract (mL) was
cheap. As a result, this technique has aroused great interest the volume of the extract obtained, respectively.
worldwide.
In the present study, different yeast strains were Preparation of the SW Product
screened to determine the strain that could consume the
maximum amount of sugars in the concentrated Sapindus The extract obtained as described in the previous section
saponins solution (SW) to produce ethanol and purify the was further concentrated 6–7-fold in the concentration of
TSS without consuming it. AnQi Grape dried yeast was TSS by a climbing film evaporator (50–120 kg/h,
found to be the best strain to purify the TSS, after its ac- Changshu Pharmaceutical Machinery Co., Ltd, China),
tivation medium, and inoculum amount, temperature and and then a much more concentrated product, called the
time of SW were optimized. This strain can utilize impu- SW product in this paper, was obtained for further
rities such as sugars and proteins of the water extraction fermentation.
process, consume minimum saponins, and produce max-
imum ethanol yield. Thus, fermentation using AnQi Grape
Determination of the Physicochemical Indices
dried yeast is natural, convenient, and efficient, and can be
of the Fermentation Liquid
used easily for the industrial-scale production of saponins.
The following physicochemical indices can track the
growth status of yeast, to obtain ethanol and purify it to a
Materials and Methods
good nutritional status.
Raw Materials 1. Content of soluble solids: The soluble solids (%) were
measured [17].
Samples of the pericarp of S. mukorossi (soapberry) were 2. Total sugars: The total sugars was determined by
obtained from the Fujian Yuanhua Forestry Bio-Technique employing the Lane-Eynon method [18].
Co. Ltd in Fujian province, China. AnQi yeast was pur- 3. Total saponins: The total saponins was determined
chased from the AnQi Company in China. Oleanolic acid according to the Vanillin–Sulfuric acid assay [19].

123
J Surfact Deterg

Fig. 1 Effect of extraction temperature, extraction time, solvent–material ratio, and number of times of extraction

Plotting of the standard curve: 10 mg/mL oleanolic Sample determination: The liquid sample contained TSS
acid was used as the standard Sapindus saponin was diluted several times and 100 lL of the diluted solu-
solution. To eight 20-mL dry and clean colorimetric tion was added into the colorimetric tube, and the same
tubes numbered from 1 to 8 containing 100, 90, 80, 70, procedure described for plotting of the standard curve was
60, 50, 40, and 20 lL of ultra-pure water, 0, 10, 20, 30, followed.
40, 50, 60, and 80 lL of 10 mg/mL standard Sapindus WTSS
saponin solution were added, respectively. The solu- Saponin purity % ¼  100 % ð3Þ
WSolublesolids
tions were evenly mixed using a vortex shaker and
heated in a water bath at 60 °C. Next, 0.4 mL of 5 % where Saponin purity % is the purity of the TSS in the
(wt%) vanillic aldehyde–glacial acetic acid solution liquid sample, WTSS(mg) is the weight of the TSS and
was added in combination with 1.4 mL of analytically WSoluble solids (mg) is the weight of soluble solids in
pure perchloric acid as the color development reagent the liquid sample, WSoluble solids (mg) = CSoluble solids 9
and shaken well. The tubes were heated in a water bath VLiquid sample.
at 70 °C for 20 min, and cooled to room temperature in WTSS
an ice water bath. Subsequently, 10 mL of glacial Saponin content % ¼  100 % ð4Þ
VLiquidsample
acetic acid were added as the solvent and mixed well.
The absorbance was determined at about 478 nm using where saponin content % is the concentration of the TSS in
an ultraviolet–visible spectrophotometer (Shimadzu the liquid sample, WTSS(mg) is the weight of the TSS and
Corporation, Japan). VLiquid sample (mL) is the volume of liquid sample.

123
 J Surfact Deterg

4. Ethanol content dry yeast; 3, High-activity dry yeast for rice wine; 4, Angel
grape wine dry yeast RV171; 5, Tianying high-sugar high-
Standard solution preparation A total of 3,000 mg/mL
activity dry yeast; 6, AnQi distiller’s yeast; 7, Angle high-
ethanol stock solution and 3,000 mg/mL n-propanol stock
temperature-resistant high-activity dry yeast; and 8, AnQi
solution (2.5 mL each) were transferred into a 100-mL
grape wine dry yeast BV818) respectively were added into
glass container and mixed well.
100 mL of SW preheated to 35 °C and diluted twofold.
Test solution preparation The fermentation liquid of
The yeasts were activated for 15–30 min in a water bath at
Sapindus saponins (SWF) and n-propanol stock solution
35 °C.
(2.5 mL each) were added into a 100-mL glass container.
The container was sealed with a rubber plug, shaken well,
Yeast Adaptability
and heated in a water bath at 60 °C for 30 min.
Chromatographic conditions A chromatographic column
A total of 0.5 mL of each activated yeast culture was
HP-INNOWAX- 1900 30 M, with a membrane thickness of
inoculated into 50 mL of SW. The container was sealed
0.25 lm, inner column diameter of 0.25 mm, outer column
with eight layers of gauze and incubated at 30 °C and
diameter of 0.32 mm, inlet temperature of 200 °C, detector
100 rpm. At 0, 6, 12, 18, 24, and 30 h, the SWF sample
temperature of 220 °C, programmed temperature rise of
was obtained and the cells were counted under the micro-
50 °C (kept constant for 10 min) at 5 °C/min to 180 °C
scope by using the blood counting chamber.
(kept constant for 5 min), N2 as carrier gas at a flow rate of
50 mL/min, H2 as combustion gas at a flow rate of 50 mL/
Fermentation Ability of the Yeast Strains
min, air as combustion assisting gas at a flow rate of 75 mL/
min, and a split ratio of 30:1, was employed.
A total of 2.5 mL of each activated yeast culture was
Determination of the K value of the standard solution A total
inoculated into 50 mL of SW and the container was sealed
of 1.5 mL of the gas above the standard solution was removed
and incubated at 30 °C. The total weights of the culture
with a syringe and immediately injected into the gas chro-
were measured at 0, 12, 24, 36, and 48 h, respectively, and
matograph. The peaks, which appeared at about 1.5 and 2.2 min
the weight lost through CO2 release (g) was determined.
for ethanol and n-propanol, respectively, were recorded.
The fermentation rate curve was plotted with the fermen-
The resulting values obtained from three measurements were
tation time (h) on the X-axis versus weight lost through
averaged, and the K value was calculated as follows:
CO2 release (g) on the Y-axis.
SEthanol in standard solution
K¼ ð5Þ
Snpropanol in standard solution Fermentation Method
where SEthanol in standard solution is the peak area of ethanol in
standard solution and Sn-propanol in standard solution is the peak A total of 2.5 mL of each activated yeast culture was
area of n-propanol in standard solution. inoculated into 50 mL of SW and the container was sealed
Determination of ethanol content in the samples A total and incubated at 30 °C. On day 4, the fermentation liquid
of 1.5 mL of the gas above the test solution was removed was sampled, and the yeast cells were removed by cen-
with a syringe and immediately injected into the gas trifugation at 8,000 rpm for 10 min. The fermentation
chromatograph. The resulting values obtained from three liquid obtained after centrifugation was analyzed for the
measurements were averaged. The ethanol content in the contents of TSS, total sugars, soluble solids, and ethanol.
sample was determined by using the following equation:
Optimization of the Yeast Activation Conditions
CEthanol content¼ 1  SEthanol ð6Þ
K Snpropanol CðEthanol content in standard solutionÞ

Activation Method
where SEthanol is the peak area of ethanol, Sn-propanol is the
peak area of n-propanol and CEthanol content in standard solution A total of 24 conical flasks were taken and numbered from
is the ethanol content in standard solution. 1 to 8, respectively, with three replicates for each ex-
periment. Subsequently, 50 mL of tap water, 2 % sugar
Screening of the Yeast Strains for the Fermentation solution, yeast extract peptone dextrose medium (YPD
of SW medium), SW, SW diluted twofold (1/2 SW; 20 % soluble
solids, 9.15 % saponin content, and 7.13 % total sugars),
Yeast Activation SW diluted fourfold (1/4 SW; 10 % soluble solids, 4.58 %
saponins, and 3.57 % total sugars), 1:1 mixture of 1/2 SW
A total of 10 g of different yeast strains numbered 1–8 (1, and 2 % sugar solution (SWS), and 1:1 mixture of 1/2 SW
AnQi super dry yeast; 2, Tianying low-sugar high-activity and YPD medium (SWY) were added into the flasks,

123
J Surfact Deterg

respectively, and preheated to 35 °C. Then, 5 g of the yeast Temperature

strains numbered 1–8 was added to the corresponding
conical flasks and shaken well. The yeasts were activated To 15 dry and clean 1,000-mL conical flasks each con-
for 15–30 min in a water bath at 35 °C. taining 800 mL of SW, 1.5 % dried yeasts were inoculated.
The flasks were sealed with a silicone plug, wrapped in
Fermentation Method kraft paper, and incubated at 20, 25, 30, 33, and 37 °C,
respectively (three replicates for each temperature). On day
A total of 2.5 mL of the activated yeast cultures was re- 4, the fermentation liquid was sampled and subjected to
spectively, inoculated into SW preheated to 30 °C. The centrifugation for 10 min at 8,000 rpm. The supernatant
container was sealed and incubated at 30 °C. obtained was analyzed for the contents of TSS, total sugars,
soluble solids, and ethanol.
Determination of the Germination Rate
Time
According to this method [20], the activated yeast culture
and the fermentation liquid were diluted to the appropriate
To 21 dry and clean 1,000-mL conical flasks each con-
concentration at 2, 6, and 10 h, respectively. Then, to 1 mL
taining 800 mL of SW, 1.5 % dried yeasts were inoculated.
of the diluted sample, a drop of 0.1 % methylene blue was
The flasks were sealed with a silicone plug, wrapped in
added. After 5 min, the sample was placed on a slide,
kraft paper, and incubated at 30 °C for 1, 2, 3, 4, 5, 7, and
covered with a cover slip, and observed under a microscope
10 days, respectively (three replicates for each fermenta-
to detect germination. The germination rate was calculated
tion time). Then, samples of the fermentation liquid were
as follows:
subjected to centrifugation at 8,000 rpm for 10 min, and
number of germinated cell the supernatant was analyzed for the contents of TSS, total
Germination rate % ¼
total yeast cell count sugars, soluble solids, and ethanol.
 100 % ð7Þ

Data Analysis
Yeast Cell Count
For all the experiments, 3–6 replicates were employed. The
The activated yeast culture and the fermentation liquid at 2, results were expressed as means ± standard deviations,
6, and 10 h were diluted to appropriate concentrations. and Origin 8.5 software was used. Two groups of data were
Then, about 100 lL of the sample was added to the blood analyzed by an independent sample t test using SPSS 17.0,
counting chamber and the cells were counted under a mi- while multiple groups of data were compared with Dun-
croscope. The yeast cell count was calculated as follows: can’s test. The data were marked by letters and P \ 0.05
A 1 was considered as being statistically significant.
Yeast cell count ¼  25  B ð8Þ
5 0:0001
where A is the yeast cell count in the five middle grids of
the blood counting chamber and B is the dilution degree of Results and Discussion
the yeast culture.
Single Factor Test for the Extraction of Sapindus
Optimization of the Yeast Fermentation Conditions Saponins

Inoculum Amount The main factors affecting TSS yield are the extraction
temperature, extraction time, solvent–material ratio, and
To 18 dry and clean 1,000-mL conical flasks each con- number of times of extraction, therefore they were in-
taining 800 mL of SW, 0.1, 0.2, 0.5, 1, 1.5, and 2 % dried cluded as factors in the single factor test. The results of
yeasts were inoculated, respectively (three replicates for the single factor test are shown in Fig. 1. An extraction
each inoculum amount). The flasks were sealed with a temperature of 40 °C, extraction time of 7 h, a solvent–
silicone plug, wrapped in kraft paper, and incubated at material ratio of 1:7, and extraction for three times gave
30 °C. On day 4, the fermentation liquid was sampled. The the best results. The water extraction process carried out
yeast cells were removed by centrifugation at 8,000 rpm for under these conditions yielded 33.41 % of TSS, with a
10 min and the supernatant obtained was analyzed for the purity of 45.71 %, and was used as a control for the
contents of TSS, total sugars, soluble solids, and ethanol. subsequent fermentation reaction.

123
 J Surfact Deterg

Screening of the Yeast Strains significant, when compared with the control (P \ 0.05).
However, a significant difference was noted with respect to
The adaptability, fermentation ability, and physicochem- sugar consumption. The ranking of the strains in the increasing
ical indices of the fermentation liquid exhibited by differ- order of the residual total sugars in the SWF at the end of
ent dried yeast strains were measured. fermentation was as follows: 8 \ 7 \ 5 \ 6 \ 2 \ 4 \
1 \ 3. Thus, the sugar consumption of strain 8 was the
Adaptability of the Yeast Strains highest, and the difference was significant, when compared
with the other strains, whereas strain 3 consumed the lowest
As shown in Fig. 2a, each yeast strain had an obvious amount of sugar. Furthermore, all the eight strains consumed
adaptation period in the first 6 h of fermentation. The yeast only a small amount of saponins, when compared with the
cell count per unit volume basically remained constant or control (P \ 0.05). Except for the significant difference in
became even slightly lower. A good adaptability was ex- saponin consumption between strain 7 and strains 8, 4, and 5,
hibited by the yeast strains 7 and 8 in SW. These two strains and between strain 5 and strains 2, 3, and 6 (P \ 0.05), there
quickly showed high growth after inoculation into SW and was no significant difference between the other strains
began fermentation. At 24 h, the yeast cell count of strains 7 (P \ 0.05). As there was little difference in the saponin con-
and 8 reached 4.575 9 108 and 4.8 9 108 cells/mL, re- sumption between the strains, the increase in the purity of
spectively. On the other hand, the adaptability of the other six saponins in the SWF at the end of the fermentation mainly
strains was low, with the yeast cell count at 24 h ranging depended on the sugar consumption by the strains. The
from 1.475 9 108 to 2.225 9 108 cells/mL (P \ 0.05). ranking of the strains in the decreasing order of the purity of
saponins was as follows: 8 [ 7 [ 5 [ 4 [ 6 [ 2 [ 1 [ 3.
Fermentation Ability of the Yeast Strains At the end of the fermentation, the highest quantity of
ethanol was produced by strain 8, with a yield of 2.7 %,
The fermentation ability of the yeast is an important indi- which was significantly higher than that produced by the
cator of its quality. The yeast strains can produce ethanol other seven strains. The ranking of the strains in the de-
under anoxic fermentation conditions. This is accompanied creasing order of ethanol production was as follows:
by the generation of CO2, which is released from the fer- 8 [ 7 [ 5 [ 6 [ 2 [ 4 [ 1 [ 3. Thus, based on the ex-
mented liquid, so that the weight of the whole fermentation perimental results with respect to the yeast strains’ adapt-
system decreases. The fermentation rate is determined ability to SW, fermentation ability, purification effect,
partially by the weight of the fermentation liquid. saponin consumption, sugar consumption, and ethanol-pro-
As shown in Fig. 2a, the yeast strains 7 and 8 immedi- ducing ability, AnQi grape wine dry yeast BV818 (strain 8)
ately entered the fermentation stage, and their fermentation was selected for fermentation-based purification of saponins.
ability was high. The weight loss resulting from CO2 re-
lease occurred mainly in the first 24 h, and the number of Strain Activation Conditions
yeast cells reached the highest level after 24 h. When
compared with strain 8, yeast strain 7 reached the fer- The rehydration and activation of the dried yeasts are in-
mentation stage more rapidly, but its fermentation ability fluenced by the type of the activation medium, temperature,
was weakened at the later stage. On the other hand, the and time period employed. It has been reported that the
yeast strains 5 and 6 entered the fermentation stage less activation of dried yeasts is optimal at an activation tem-
rapidly, and the general fermentation process was mild. perature of 38 °C and an activation time period of
The weight loss resulting from CO2 release mainly oc- 15–30 min. Hence, in the present study, only the effects of
curred in the first 36 h. In terms of the overall fermentation the type and concentration of the activation medium were
speed, both strains 7 and 8 exhibited a similar fermentation studied, and the results are shown in Table 1.
rate, while the rest of the yeast strains 1–6 was much lower, As can be seen from the results obtained, the purity of
considering of the rapid fermentation, short fermentation saponins achieved using the activation media, 2 % sucrose
period and ethanol-producing ability the strains 8 and 7 solution, 1/2 SW, and 1/4 SW was the highest, reaching
were much more suitable to be used as fermentation strains 58.47, 58.50, and 59.47 %, respectively. However, yeasts
in SW. activated with 2 % sucrose solution consumed higher
amount of saponins, reducing the saponin content to
Analysis of the Physicochemical Indices 13.05 %. It may be that when the sucrose used as the sole
of the Fermentation Liquid carbon source was not enough for the activation of the strain,
the strain has to get the carbon source from the saponin. Thus,
As shown in Fig. 2b, all eight yeast strains were able to con- when compared with the other activation media, use of 1/2
vert the sugars in SW to ethanol, and the difference was SW or 1/4 SW for the rehydration and activation of dried

123
J Surfact Deterg

Fig. 2 a Comparison of the growth and fermentation ability of yeast; 7, angle high-temperature-resistant high-activity dry yeast; and
different yeast strains in SW. b Strain selection. Note: 1, AnQi super 8, AnQi grape wine dry yeast BV818. SW contained 40.00 % soluble
dry yeast; 2, Tianying low-sugar high-activity dry yeast; 3, high- solids, 18.29 % TSS, and 14.25 % total sugars, and the saponin purity
activity dry yeast for rice wine; 4, AnQi grape wine dry yeast RV171; was 45.71 %
5, Tianying high-sugar, high-activity dry yeast; 6, AnQi distiller’s

yeasts resulted in lower saponin consumption, better saponin obviously decrease (P [ 0.05) and the ethanol yield also did
purification, and higher yield of ethanol. In addition, these not significantly increase (P \ 0.05). The purity of the
two rehydrating and activating media increased the adapt- saponins first increased and then stabilized. In particular, the
ability of the strains to the substrate. maximum purity of saponins was achieved when the
inoculum amount was 1.5 %. The inoculum amounts of 1,
Optimization of the Fermentation Conditions 1.5, and 2 % did not produce any obvious difference
(P [ 0.05). The reasons may be the following: (1) The solid
The fermentation conditions, including the inoculum content of SW was high, when the inoculation amount was
amount, temperature, and time, with AnQi grape wine low; (2) the dissolved oxygen and sugar in the substrate was
dried yeast BV818 (strain 8) in SW were optimized, and enough for yeast proliferation, but as the substrate concen-
the results are shown in Fig. 3. tration was too high, the permeability pressure distribution in
each yeast cell was too high; this in turn affected the
Inoculum Amount physiological processes in the yeast cells; (3) the amount of
biomass was small and consumed less sugar, protein and
When the amount of inoculum was increased to 1.5 or 2 %, other impurities; (4) when the amount of the inoculum was
the total sugar content in the fermentation liquid did not increased to a certain degree, the normal metabolism of

123
 J Surfact Deterg

W warm water, 2 % S 2 % sugar solution, SW concentrated Sapindus saponin solution, 1/2 SW SW diluted twofold (20 % soluble solids, 9.15 % saponin content, and 7.13 % total sugars), 1/4
55.97 ± 0.95bc

SW SW diluted fourfold (10 % soluble solids, 4.58 % saponins, and 3.57 % total sugars), SWS 1:1 mixture of SW diluted twofold and 2 % sugar solution SWY, 1:1 mixture of SW diluted
nutrients in the medium was not enough to sustain the yeast,

1.90 ± 0.17d

0.79 ± 0.13d

9.26 ± 0.38b
8.35 ± 0.32c

15.18 ± 0.88a

1.87 ± 0.03c
the aging of the yeast accelerated, the rate of sugar utilization
declined, the conversion rate of ethanol decreased, and a lot
SWY

of metabolic waste was generated, which caused a decrease

in the purity of the saponins. Thus, the optimal inoculum
amount was concluded to be 1.5 %.

cd
1.78 ± 0.13d

0.825 ± 0.04d

6.45 ± 0.33d

9.64 ± 0.77b
14.86 ± 0.73a

c
53.82 ± 1.86

1.86 ± 0.02
Temperature
SWS

The total sugar content was the lowest in SWF at 30 and

33 °C, corresponding to the highest purity of saponins and
ethanol yield. At 30 °C, the purity of saponins reached
b
4.04 ± 0.28a

3.56 ± 0.12a

10.55 ± 0.61a

15.32 ± 0.39a
a

8.45 ± 1.25c
59.47 ± 2.16

2.75 ± 0.05
74.31 %, the total sugar content decreased to 3.16 %, and the
ethanol yield was as high as 5.38 %. Although the saponin
1/4 SW

content significantly decreased, it was still within the in-

dustrially acceptable range. When the temperature was in-
creased, the yield of ethanol first increased, and then
decreased, reaching a maximum at 30 °C. The reasons for
3.92 ± 0.12ab

ab

b
3.31 ± 0.23a

11.15 ± 0.34a

15.5 ± 0.21a

8.62 ± 0.43c
58.50 ± 0.59

2.72 ± 0.09

this observation may be that the yeast strain was sensitive to

temperature when the temperature was low. Yeast fermen-
1/2 SW

tation with the 30 °C temperature produced ethanol faster,

However, when the temperature was too high, the yeast
fermentation enzyme activity slowed down, reducing the
The lowercase superscripts in the same row represents the significant difference between the data (P \ 0.05)
3.73 ± 0.22b

2.07 ± 0.15b

6.5 ± 0.30d

13.95 ± 0.22b
e

10.45 ± 0.91a
e
50.58 ± 0.67

1.12 ± 0.03

rate of fermentation to produce ethanol. Thus, it was con-

cluded that the optimal fermentation temperature was 30 °C.
SW

Time

After the start of the fermentation, the total sugar content

cd

9.47 ± 0.77b
2.52 ± 0.16c

1.62 ± 0.13c

8.05 ± 0.23c

14.92 ± 0.58a

c
54.41 ± 1.19

1.88 ± 0.08

obviously declined and reached the lowest level on day 4.

The purity of saponins first increased and then became
stabilized, reaching the highest value around day 3 or 4.
YPD

The ethanol content first increased and then decreased,

exhibiting the highest level on day 4, which subsequently
decreased slightly, and may have been converted to acetic
3.89 ± 0.17ab

ab
9.65 ± 1.17b

7.89 ± 1.22d
3.41 ± 0.12a

13.05 ± 0.52c

a
58.47 ± 1.94

2.90 ± 0.05

acid and acetaldehyde. Thus, it was concluded that the

optimal fermentation time period was 4 days.
2 %S

In summary, based on the experimental results, the optimal

conditions for the fermentation and purification of Sapindus
saponins using AnQi grape wine dry yeast BV818 were de-
de
Table 1 Comparison of the activation media

termined as follows: rehydration and activation of the dried

2.02 ± 0.13b

9.51 ± 1.07b
d
2.41 ± 0.21c

5.65 ± 0.31e

12.98 ± 0.96c
52.72 ± 1.35

1.73 ± 0.02

yeasts with 1/2 or 1/4 SW, inoculum amount of 1.5 %, tem-

Number of yeast cells (108cells/mL±SD)

perature of 30 °C, and a fermentation period of 4 days. Under

Fermentation budding rate (%±SD)

these conditions, the purity of saponins increased from 45.71

Activation budding rate (%±SD)
W

to 75.50 %, the ethanol yield was 5.33 % (w/v), and the

Fermentation results (%±SD)

saponin content decreased from 18.29 to 15.30 % (w/v),

twofold and YPD medium
Total sugar content (%)

which was within the industrially acceptable range.

Saponin content (%)
Saponin purity (%)

Ethanol yield (%)

Activation medium

Conclusions
10 h

10 h

In the present study, water extraction and microbial

fermentation were adopted to separate and purify TSS.

123
J Surfact Deterg

Fig. 3 Optimization of the fermentation conditions, including inoculum amount, temperature, and time

The best conditions for the water extraction process were The results of our study showed that, yeast fermentation
as follows: an extraction temperature of 40 °C, an ex- to purify the Sapindus saponins initially extracted by water,
traction time of 7 h, a solvent–material ratio of 1:7, and not only achieved the purpose of purification of the sapo-
three repeats of the extraction procedure. Under these nins, but also generated ethanol as a byproduct fuel. Thus,
conditions, the TSS yield from the pericarp was this process was particularly superior to others, therefore
33.41 %, with a purity of 45.71 %.The optimum acti- capable of industrial application.
vation and fermentation conditions for AnQi grape wine
dry yeast BV818 were as follows: pre-activation in SW Acknowledgments This article was supported by the key project of
transformation and industrialization of scientific and technological
with 10 or 20 % soluble solid content, temperature of achievement in Fujian Province (No. 2012N3008), the National
30 °C, inoculum amount of 1.5 %, and fermentation Natural Science Foundation of People’ Republic of China (No.
period of 4 days. Under these conditions, the purity of 21203076), Open Project of State Key Laboratory of Supramolecular
saponins increased from 45.71 to 75.50 %, the total Structure and Materials (sklssm 201328), Jiangsu province science
and technology support program for social development
sugar content significantly decreased from 14.25 to (BE2012631), the Fundamental Research Funds for the Central
3.02 %, the ethanol yield increased to 5.33 % (w/v), and Universities (JUSRP51309A), National Key Technology R&D Pro-
the Sapindus saponin content slightly decreased from gram in the 12th Five year Plan of China (No. 2012BAD36B02,
18.29 to 15.30 % (w/v). 2014BAD04B03) and the Priority Academic Program Development

123
 J Surfact Deterg

of Jiangsu Higher Education Institution (PAPD). The manuscript has 17. Cui Z (2012) The production of total Sapindus-saponins from
been edited by International Science Editing. Sapindus mukurossi Gaertn. Jiangnan University, Wuxi
18. University WlI (1998) Food analysis
19. Hiai S, Oura H, Nakajima T (1976) Color reaction of some
saposenins and saponins with vanillin and sulfuric acid. Planta
References
Med 29(02):116–122
20. Zhu Gejian TS (1994) Industrial microbiology experiment tech-
1. Huang T (ed) (2003) Editorial Committee of the Flora of Taiwan. nical manual. China Light Industry Press, Beijing
Department of Botany, Taipei
2. Nakayama K, Fujino H, Kasai R, Mitoma YYN, Tanaka O (1986)
Solubilizing properties of saponins from Sapindus mukurossi Wu Heng received her M. Pharm. degree from Jiangnan University in
Gaertn. Chem Pharm Bull 34(8):3279–3283 China. She is working at the Shanghai Institute for Food and Drug
3. Kazuo RWG (1996) Saponins used in food and agriculture Control (SIFDC), China.
4. Cy W (1977) Flora Yunnanica, vol 1. Science Press, Beijing
5. Yueqing J, Jingcheng Z, Xiaohua G (2008) Study on purification Zhang Ling is a doctoral student, majoring in food science, in the
of tea saponin by water extraction and fermentation. Cereal Food School of Food Science and Technology at Jiangnan University,
Ind 15:34–37 China.
6. Roy D, Kommalapati R, Mandava S, Valsaraj K, Constant W
(1997) Soil washing potential of a natural surfactant. Env Sci
Wang Na is a postgraduate, majoring in food engineering, in the
Technol 31(3):670–675
School of Food Science and Technology at Jiangnan University,
7. Houzeng R, Longhua G (2002) The extraction technology of
China.
Supindus-saponin. Jiangxi Sci 20(1):55–58
8. Jieru S, Kongchang C, Mingfang Z, Lei Z (2002) Study of surface
activity and application of Sapindus mukurossi. China Surfact Guo Youzhi is the general manager of the Fujian Yuanhua Forestry
Deterg Cosmetics 32(4):16–19 Bio-Technique Co. Ltd., China.
9. Shujuan L, Huarong T, Dangling J (2005) Study on extraction
and application of theasaponin in camellia cake. Cereal and Oils Weng Zhen is the chairman of the Fujian Yuanhua Forestry Bio-
1:13–15 Technique Co. Ltd., China. His company is focused on the planting
10. Rongxiang L, Quanfen Z, Chunhua X (1996) Progress and ten- and developing of Chinese soap berry trees.
dency of the researches on tapon in Bio-activity and its utiliza-
tion. Tea Sci 16(2):81–86 Sun Zhiyong is the vice-general manager of the Fujian Yuanhua
11. Zhaojun Y, Hongquan L, Qunfang L (2004) Extraction and pu- Forestry Bio-Technique Co. Ltd., China.
rification of tea saponin camellia seed. Feed Industry Magazine
25(12):11–13 Xu Deping is an associate professor in the School of Food Science
12. Fengyu W, Daofu Z (2007) China Patent No.: S. I. P. O. o. and Technology, Jiangnan University, China. He is now focusing on
T. P.R.C the extraction and application of natural products.
13. Fengyu W, Hui X, Jincheng Y (2008) Separation and purification
of Supindus-saponin by ultrafiltration. Membrane Sci Technol Xie Yunfei is an associate professor in the School of Food Science
28(2):85–88 and Technology, Jiangnan University in China. She is now focusing
14. Zhilong X, Daijia Z, Lingyun J (2001) Foam Separation of gin- on the extraction and application of natural products.
senosides of Panax ginseng. Chinese J Process Eng 1(3):289–292
15. Fengyu W, Jing Z, Hui X (2009) Purification of Sapindus-saponin Yao Weirong is a professor in the School of Food Science and
by foam separation. Chin Tradit Pat Med 31(7):1021–1024 Technology, Jiangnan University, China. She is now focusing on the
16. Pourrat H, Regerat F, Lamaison J, Pourrat A (1979) Purification developing and application of natural cleaning agents in the food
of main saponin from tuber of ficaire (Ficaria ranunculoides industry.
Moench) by a strain of Aspergillus niger (author’s translation).
Paper presented at the Annales Pharmaceutiques Françaises

123