Biochemical Engineering Journal 40 (2008) 460–464

Production of hyaluronic acid by repeated batch fermentation

Wei-Chih Huang a , Shu-Jen Chen b , Teh-Liang Chen a,∗
aDepartment of Chemical Engineering, National Cheng Kung University, Tainan 701, Taiwan
b Department of Chemical and Material Engineering, National Kaohsiung University of Applied Sciences,
Kaohsiung 807, Taiwan
Received 26 September 2007; received in revised form 9 December 2007; accepted 11 January 2008

Abstract
The production of hyaluronic acid (HA) by Streptococcus zooepidemicus with repeated batch fermentation has been investigated. It was found
that, with conventional operation, both maximal specific growth rate (μm ) and specific HA productivity (YP/X ) decreased with increasing seed
volume, suggesting that there exist some inhibitors in the broth. The removal of liquid in the seed was first attempted by installing nonwoven
fabrics (NWF) in the fermentor to retain some of the cells when draining the broth. However, this resulted in a loss of HA productivity, which in
turn was attributed to the growth of a sticky, non-HA-producing mutant on the NWF. Using an external cartridge filter to partially retain the cells,
followed by back-washing the filter with fresh medium for seeding, μm and YP/X could be maintained successfully at their batch levels during the
repeated cycles. In an operation that seeded 31% cell, the volumetric production rate of the repeated batch culture (0.59 g HA L−1 h−1 ) was found
to be 2.5-fold of the batch culture (0.24 g HA L−1 h−1 ).
© 2008 Elsevier B.V. All rights reserved.

Keywords: Fermentation; Batch processing; Submerged culture; Biosynthesis; Hyaluronic acid; Streptococcus zooepidemicus

1. Introduction high-cell-density fermentation or through a high-yield strain.

Hyaluronic acid (HA) is a mucopolysaccharide consisting
of alternating N-acetyl-d-glucosamine and d-glucuronic acid,
which has received great interest in the medical and cosmetic
markets [1–3]. In recent years, HA from microbial fermenta-
tion, rather than extraction from animal sources, is receiving
increased attention for avoidance of cross-species viral infection.
HA fermentations have been mostly by Streptococci spp.,
where HA is a capsular biopolymer shedding to the medium [4].
To aid the competence of the fermentation process, the develop-
ment of an economical process for mass production is necessary.
In the literature, most reports on HA fermentation have been
based on batch culture [5–9]. A major drawback of batch culture
is a long turnaround time, which greatly decreases the volumet-
ric production rate, and this in turn results in a high fixed cost
per unit product.
There are two approaches that could be employed to increase
the volumetric production rate. One is to increase HA concentra-
tion in the fermentor, which can be accomplished either through
∗ Corresponding author. Tel.: +886 6 2757575x62660; fax: +886 6 2344496.
E-mail address: t62660@mail.ncku.edu.tw (T.-L. Chen).
However, we found that, as the HA concentration higher than
4 g/L, the broth became too viscous to achieve efficient agitation
and aeration; as a result, the benefit of high HA concentration is
worn down by the defect of low efficiency of HA synthesis. The
other approach is to skip the turnaround phase by using contin-
uous culture. Continuous culture could also offer two benefits.
First, cell growth can be maintained at the exponential phase, so
that the excretion of cell wall proteins at the stationary phase [4]
can be avoided. Second, the cells can be controlled to grow at a
lower specific growth rate, which might result in HA of higher
molecular weight [5]. However, continuous culture has its inher-
ent defect—low efficiency of substrate utilization; and this is the
primary reason why continuous culture is seldom used in a com-
mercial process. Another adverse factor concerned is that the
efficiency of HA production would decrease during prolonged
operation [10,11].
Accordingly, repeated batch culture seems a promising mode
for HA fermentation, because it skips the turnaround time and
the lag phase. Unfortunately, no report on such subject has been
published to date. The aim of this work was to explore the
problems encountered during HA production by repeated batch
culture, to rationalize the problems, and to propose a feasible
operation strategy.

1369-703X/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.bej.2008.01.021
 W.-C. Huang et al. / Biochemical Engineering Journal 40 (2008) 460–464 461

2. Materials and methods

S. zooepidemicus ATCC 39920 was used as the HA producer.

The cells were maintained at −30 ◦ C in 50% (v/v) glycerol. The
isolation of a pure culture was achieved by streaking onto TSB
agar plates, which contained (per liter) 17 g of tryptone, 3 g of
soytone, 2.5 g of glucose, 2.5 g of K2 HPO4 , 5 g of NaCl, and
15 g of agar. Precultures were prepared in a 500-mL shaking
flask, with 100 mL of the TSB medium, at 37 ◦ C for 12 h.
HA production was performed in a 3 L fermentor (MDL-
300, B. E. Marubishi, Japan), with a working volume of 2 L.
The medium comprised of (per liter) 20 g of glucose, 10 g of
yeast extract, 1.7 g of tryptone, 0.3 g of soytone, 2 g of NaCl,
2.5 g of K2 HPO4 and 1.5 g of MgSO4 ·7H2 O. The fermentor was Fig. 1. HA production by repeated batch culture, using 5% broth as the seed.
inoculated with 5% (v/v) preculture, and operated at 37 ◦ C and Symbols: (䊉) cell density and () HA concentration.
pH 7.0 (with 5N NaOH). The aeration rate was 1 vvm, and the
concentration of dissolved oxygen (DO) was controlled above estimated to be 0.59 h−1 ; and the specific productivity (YP/X ),
10% air saturation by mixing the inlet air with oxygen when defined as the ratio of total increment of HA to total increment
necessary. Agitation was achieved with a gate impeller similar of cell mass, was 0.72 g HA/g cell. μm together with YP/X are
to that of Hiruta et al. [12], and an agitator speed of 400 rpm was good indicators for assessing the efficiency of the repeated batch
used. culture.
In repeated batch culture, the end of each cycle was deter- The repeated batch culture was first performed using a seed of
mined when there was a rise in DO level; it is a sign that the cells 5% broth. As can be seen in Fig. 1, both cell and HA concentra-
cease to grow. The medium concentration was adjusted accord- tions could be maintained as in the batch culture for at least seven
ing to the amount of cell in the seed, such that the repeated cycles repeated cycles. However, μm and YP/X are reduced to 0.51 h−1
were supposedly to achieve the same cell concentration as the and 0.62 g HA/g cell, respectively. To increase the volumetric
batch level. The nonwoven fabric (NWF) used to retain the cells production rate, volumes of the seed were further increased to
consists of a polypropylene core and a polyethylene surface. 10, 20, 30 and 40% of the broth, respectively. Fig. 2 depicts the
The fabric has a unit weight of 60 g/m2 and a specific gravity profiles of using 40% broth as the seed. Unfortunately, although
of 0.90. To equip the NWF in the fermentor, three pieces of the cell concentration could be maintained, HA production was
NWF (4 cm × 17 cm) were tied around the three baffles, unless reduced gradually to 1.3 g/L. The influence of seed volume in
otherwise noted. The cartridge filter used to retain the cells out- the repeated batch culture is summarized in Fig. 3, which shows
side the fermentor had a diameter of 3 cm and a length of 30 cm, both μm and YP/X decreased with increasing seed volume. It thus
where the NWF was used as the filter medium. After drain- suggests that in the broth there exist some inhibitors that hinder
ing the broth, the retained cells were released by back-washing cell growth and HA synthesis.
with fresh medium to seed the fermentor. The back-wash pro- It might also be possible that the inhibitors hinder cell growth,
cedure was ended with air-purge to avoid microbial growth in and a smaller specific growth rate results in a smaller specific HA
the cartridge. The amount of cells in the seed was adjusted by productivity; in other words, the inhibitors might not function
the amount of the NWF inside the filter as well as the number directly on HA synthesis. We therefore performed a fed-batch
of the cartridges. culture that was operated at a specific growth rate of 0.30 h−1 ,
Cell concentration was measured from optical density (OD)
of the broth at 660 nm using a spectrophotometer. Owing to a
change in cell morphology after entry into the stationary phase
[13], the correlations of OD with dry cell weight (DCW) were
DCW (g/L) = 0.399 × OD − 0.003 for the exponential growth
phase, and DCW (g/L) = 0.456 × OD − 0.012 for the station-
ary phase. HA concentration was determined by the carbazole
method [14], in which the optical density was measured at
525 nm and d-glucuronic acid was used as the standard.

3. Results and discussion

The preliminary results of batch culture showed that the cells

entered the stationary phase at 9 h, with a cell concentration
of 3.2 g/L and a HA concentration of 2.3 g/L; in addition, the
formation of HA was a growth-associated matter (data can also Fig. 2. HA production by repeated batch culture, using 40% broth as the seed.
be seen in Fig. 1). The maximal specific growth rate (μm ) was Symbols: (䊉) cell density and () HA concentration.
462 W.-C. Huang et al. / Biochemical Engineering Journal 40 (2008) 460–464
Fig. 3. Variations of μm and YP/X during the repeated batch culture, here showing
μm and YP/X decrease with increasing volume of seed. The repeated cycle of 0
denotes the initial batch culture. Seed volume: (䊉) 5%, () 10%, () 20%, ()
30%, and () 40% (v/v).
the lowest rate found in Fig. 3. However, the resulted YP/X was
still 0.72 g HA/g cell (fermentation profiles not shown). This
assumes that the inhibitors function on both cell growth and HA
synthesis.
With the assumption about the inhibitors, it is likely that both
μm and YP/X can be maintained during the repeated cycles if the
seed contains no liquid. We then installed NWF in the fermentor
to retain some of the cells when draining the broth. The retained
cells were then released and served as the seed for the next cycle.
The fermentation profiles thus obtained are shown in Fig. 4,
where the seed was estimated to be equivalent to a 3% inocu-
lum. It can be seen that cell concentration (of the broth) could
be maintained for two repeated cycles; thereafter, it decreased
gradually. The decrease in cell concentration was found to be
attributed to that more and more cells grew on the NWF. It
Fig. 4. HA production by repeated batch culture with NWF equipped in the
fermentor. Symbols: (䊉) cell density and () HA concentration.
Fig. 5. Comparison of batch fermentations between the non-mucoid mutant (䊉,
) and the normal strain (, ). Symbols: (䊉, ) cell density and (, ) HA
concentration.
is worth noting that the NWF functioned merely to block the
cells from draining in the first three cycles, yet surface growth
occurred in the succeeding cycles. On the other hand, although
the HA level could also be maintained for two repeated cycles,
it decreased sharply after the third cycle, and eventually, HA
production ceased at the sixth cycle. Nevertheless, the cells still
grew at a μm of 0.60 h−1 throughout the repeated cycles.
From Fig. 4, an important clue to the HA fermentation can
be obtained: the cells grown on the NWF, and their descen-
dants, must be a non-HA-producing mutant. To confirm this
suspicion, cells on the NWF were isolated on an agar plate; and
as expected, they appeared as small, rough colonies (diameter
of 0.5–0.8 mm, non-mucoid) rather than large, smooth colonies
(diameter of 3 mm, mucoid, normal strain) [10]. The percent-
ages of the mutant in the broth at the end of each cycle were
found to be 0, 0, 0, 0, 33, 79 and 100% for repeated cycles of 0
(the initial batch culture), 1, 2, 3, 4, 5 and 6, respectively. The
tendency of increase in this percentage was in accordance with
the reduction in the HA production. It is interesting to note that
the mutant appeared after 32 h (the third repeated cycle), not
from the beginning. This is similar to a previous report [10],
where showed the occurrence of non-HA-producing cells was
beyond 27 h in a chemostat culture.
Furthermore, a comparison between the non-mucoid mutant
and the normal strain was performed with batch fermentation,
as shown in Fig. 5. It can be seen that the extent of cell growth
as well as μm of both strains were quite similar, though the
mutant had a shorter lag phase; nevertheless, the mutant showed
no ability of HA production. Once again, the results of Fig. 5
are in agreement with the phenomena found in Fig. 4. It is worth
noting that the microscopic observation of both strains showed
the same cell morphology—they appeared as strands of 6–8 cells
during the exponential growth phase. A similar result was also
reported for group A streptococci [15].
To avoid growth of the non-mucoid mutant on the NWF, we
tied the NWF to the agitator, instead of the baffles. However,
similar results as in Fig. 4 were obtained (data not shown); the
 W.-C. Huang et al. / Biochemical Engineering Journal 40 (2008) 460–464 463

Fig. 6. HA production by repeated batch culture with a cartridge filter, where

the amount of cell seeded was 3%. Symbols: (䊉) cell density and () HA
concentration.

mutant is so sticky that it can cling onto the NWF, irrespective

of the higher shear force.
The above observations, together with previous works
[6,10,13,16], suggest the following hypothesis for HA fermen-
tation. S. zooepidemicus is an aerotolerant anaerobe; aeration Fig. 7. Variations of μm and YP/X during the repeated batch culture, here showing
has no effect on cell growth, but it markedly enhances HA μm and YP/X can be maintained at their batch levels if the seed contains no liquid.
Amount of cell seeded: (䊉) 3%, () 6.7%, () 16%, and () 31%.
production. During the exponential growth phase, the cocoid
cells associate into strand by capsule for themselves to be
shielded from oxygen. After entry into the stationary phase, tence of inhibitors in the broth. On the contrary, if we seed the
some inhibitors that hinder cell growth and HA formation would fermentor with cells only, the beneficial effect of high seeding
appear in the broth, and this limits the repeated batch culture to ratio on the volumetric productivity can be manifest. As can
using a seed of small volume. In continuous fermentations, a be seen in Fig. 8, using a seed of 31% cell, a productivity of
sticky non-mucoid mutant, which has the advantage of a longer 0.59 g HA L−1 h−1 could be obtained, which was 2.5-fold of the
retention time, would eventually become the prevailing species, batch culture.
and thus HA production ceases. The use of NWF, which provides
a tremendous surface area for the sticky cells to cling onto, is an
enticement for the prevalence of the non-HA-producing cells.
According to the hypothesis, HA production by repeated
batch culture can be performed as follows: partially block the
cells with an external filter when draining the broth, followed
by back-washing the filter with fresh medium to seed the fer-
mentor. Fig. 6 shows the repeated batch culture with a cartridge
filter, where the amount of cell seeded was estimated to be 3%.
It can be seen that during the repeated cycles, not only cell and
HA concentrations but also μm and YP/X could be maintained
at their batch levels. Similar results were also obtained when
seeding larger amounts of cell, as shown in Fig. 7; the strategy
to operate the repeated batch culture was thus quite success-
ful. It is perhaps worth mentioning that the external filter has an
added benefit—it helps to remove the sticky, non-HA-producing
mutant from the seed, if they would occur.
As mentioned above, increasing the volumetric production
rate, rather than increasing the product concentration, is the
focus of HA fermentation. Fig. 8 shows the comparison of HA
productivities in batch culture and in repeated batch cultures.
The batch culture had a HA productivity of 0.24 g HA L−1 h−1 ,
not counting the turnaround time. With conventional repeated Fig. 8. Comparison of volumetric productivities of HA between batch culture
(repeated cycle of 0) and repeated batch culture. In the upper panel (seed con-
batch culture (seed containing liquid), the productivity would taining liquid), the seed volumes were 5% (䊉), 10% (), 20% (), 30% (),
not increase with increasing seed volume; the beneficial effect and 40% (). In the lower panel (seed containing no liquid), the amounts of cell
of using large seed volume is worn down owing to the exis- inoculated were 3% (䊉), 6.7% (), 16% (), and 31% ().
464 W.-C. Huang et al. / Biochemical Engineering Journal 40 (2008) 460–464
4. Conclusions
HA from microbial fermentation would become the pre-
dominant source in the medical and cosmetic markets. The
improvement in volumetric productivity is crucial to HA fermen-
tation, and repeated batch culture seems to be the most feasible
approach. The problem encountered in conventional repeated
batch fermentation of HA by S. zooepidemicus is the existence
of some inhibitors in the broth, which wears down the benefit of
using large seed volume. Equipping a filtration medium (NWF,
in this work) inside the fermentor to retain the cells for serving
as the seed would lead to a more serious problem: a sticky, non-
HA-producing strain would grow on the solid surface and soon
becomes the prevailing species. With an external cartridge filter
to retain the cells when draining the broth, the repeated batch
culture can be employed successfully for HA production.
Acknowledgement
This study was supported by research grant NSC96-2221-E-
006-271 provided by the National Science Council of Taiwan.
References
[1] I.F. Radaeva, G.A. Kostina, A.V. Zmievskii, Hyaluronic acid: biological
role, structure, synthesis, isolation, purification, and applications, Appl.
Biochem. Microbiol. 33 (1997) 111–115.
[2] L. Lapčik Jr., L. Lapčik, S. De Smedt, J. Demeester, P. Chabreček, Hyaluro-
nan: preparation, structure, properties, and applications, Chem. Rev. 8
(1998) 2663–2684.
[3] B. Fong Chong, L.M. Blank, R. Mclaughlin, L.K. Nielsen, Micro-
bial hyaluronic acid production, Appl. Microbiol. Biotechnol. 66 (2005)
341–351.
[4] A. Mausolf, J. Jungmann, H. Robenek, P. Prehm, Shedding of hyaluronate
synthase from streptococci, Biochem. J. 267 (1990) 191–196.
[5] D.C. Armstrong, M.R. Johns, Culture conditions affect the molecular
weight properties of hyaluronic acid produced by Streptococcus zooepi-
demicus, Appl. Environ. Microbiol. 63 (1997) 2759–2764.
[6] W.C. Huang, S.J. Chen, T.L. Chen, The role of dissolved oxygen and
function of agitation in hyaluronic acid fermentation, Biochem. Eng. J.
32 (2006) 239–243.
[7] M.R. Johns, L.T. Goh, A. Oeggerli, Effect of pH, agitation and aeration on
hyaluronic acid production by Streptococcus zooepidemicus, Biotechnol.
Lett. 16 (1994) 507–512.
[8] J.H. Kim, S.J. Yoo, D.K. Oh, Y.G. Kweon, D.W. Park, C.H. Lee, G.H.
Gil, Selection of a Streptococcus equi mutant and optimization of culture
conditions for the production of high molecular weight hyaluronic acid,
Enzyme Microb. Technol. 19 (1996) 440–445.
[9] I. van de Rijn, R.E. Kessler, Growth characteristics of group A strep-
tococci in a new chemically defined medium, Infect. Immun. 27 (1980)
444–448.
[10] L.M. Blank, R.L. McLaughlin, L.K. Nielsen, Stable production of
hyaluronic acid in Streptococcus zooepidemicus chemostats operated at
high dilution rate, Biotechnol. Bioeng. 90 (2005) 685–693.
[11] D.C. Ellwood, C.G.T. Evans, G.M. Dunn, N. McInnes, R.G. Yeo, K.J.
Smith, Production of hyaluronic acid, US Patent 5,411,874 (1995).
[12] O. Hiruta, K. Yamamura, H. Takebe, T. Futamura, K. Iinuma, H. Tanaka,
Application of Maxblend fermentor for microbial process, J. Ferment.
Bioeng. 83 (1997) 79–86.
[13] W.C. Huang, S.J. Chen, T.L. Chen, Modeling the microbial pro-
duction of hyaluronic acid, J. Chin. Inst. Chem. Engrs. 38 (2007)
355–359.
[14] T. Bitter, H.M. Muir, A modified uronic acid carbazole reaction, Anal.
Biochem. 4 (1962) 330–334.
[15] B.A.B. Leonard, M. Woischnik, A. Podbielski, Production of stabilized
virulence factor-negative variants by group A streptococci during stationary
phase, Infect. Immun. 66 (1998) 3841–3847.
[16] P.P. Cleary, A. Larkin, Hyaluronic acid capsule: strategy for oxy-
gen resistance in group A Streptococci, J. Bacteriol. 140 (1979)
1090–1097.