You are on page 1of 13

International Journal of Food Science and Technology 2016 1

Original article
Inferring the role of microorganisms in water kefir fermentations

Abigail Martınez-Torres, Sandra Gutierrez-Ambrocio, Pamela Heredia-del-Orbe, Lourdes Villa-Tanaca &


Cesar Hern
andez-Rodrıguez*
Departamento de Microbiologıa, Escuela Nacional de Ciencias Biol
ogicas, Instituto Politecnico Nacional, Prol. Carpio y Plan de Ayala, Col.
Casco de Santo Tom
as. Distrito Federal, CP 11340, Mexico, D.F., Mexico
(Received 5 August 2016; Accepted in revised form 25 October 2016)

Summary Water kefir is a slightly alcoholic, lactic and acetic beverage fermented by yeasts, lactic acid bacteria and
acetic acid bacteria that are associated with the polysaccharide of the water kefir grains. In this study, the
three main metabolic products of microorganisms were evaluated during a traditional 192-h water kefir
fermentation and also after inoculating the microorganisms in fresh medium or sterilised broth from dif-
ferent fermentation stages. The first process to occur was alcoholic fermentation, carried out in particular
by Saccharomyces cerevisiae. After 24 h, lactic and acetic acid accumulation was generated by Lactobacil-
lus hilgardii and Acetobacter tropicalis. By the end of fermentation, ethanol had been almost entirely con-
sumed and oxidised to acetic acid, possibly by a dissimilatory route of Acetobacter species. An original
hypothetical diagram is proposed for the carbon flux from sucrose, and the metabolic role of the main
yeasts and bacteria is assigned for the distinct stages of water kefir fermentation.
Keywords Acetobacter tropicalis, Lactobacillus hilgardii, Saccharomyces cerevisiae, water kefir fermentation.

1989; Bergmann et al., 2010; Waldherr et al., 2010;


Introduction
Gulitz et al., 2011; Miguel et al., 2011).
Water kefir (WK) (also known as sugary kefir or There have been some attempts to understand the
‘tepache de tibicos’) is a fermented beverage made with interactions of the microorganisms in WK fermenta-
panela, a sugar cane product mainly composed of 80– tions. In particular, two studies suggest that different
89% sucrose, 10% reducing sugars and about 0.4% symbiotic relationships can be established between
protein (Guerra & Mujica, 2010). Some fruits (e.g. figs yeasts and lactic acid fermenters isolated from different
or lemons) are usually added to the WK fermentation WK fermentations. Firstly, the interaction of a
broth to enhance flavour (Pidoux et al., 1988; Reiß, polysaccharide producer (Lactobacillus hilgardii) with
1990). Fermentation results in WK vinegar after a yeast (Saccharomyces florentinus) enhances the sur-
2 weeks. vival of bacteria in glucose yeast extract medium and
The starter culture to produce the beverage is made increases their growth and lactic acid production.
of amorphous, compact, translucent grains, named Meanwhile, bacteria affect the growth and ethanol
‘tibicos’ or WK grains (WKGs). These grains are production of the yeasts (Leroi & Pidoux, 1993). Sec-
mainly composed of water and a gelling polysaccha- ondly, a mutualistic relationship was also established
ride, in which different bacteria and yeasts are embed- in another report on the enhanced cell yields of the
ded. The particular microbiota depends on the two partners of a co-culture of Lactobacillus nagelii
location of sample collection along with the specific and Lactobacillus hordei with Zygotorulaspora flo-
procedure for preparing the beverage (Gulitz et al., rentina (formerly S. florentinus) (Stadie et al., 2013).
2011; Miguel et al., 2011). Nevertheless, Saccha- To our knowledge, only one study has explored the
romyces cerevisiae yeasts as well as certain species of microbial dynamics in a WK fermentation during
Lactobacillus and acetic acid bacteria have been 192 h, in that case using panela and figs as the sub-
repeatedly observed (Moinas et al., 1980; Pidoux, strate at an initial pH of 4.6. The resulting microbial
diversity in grains and supernatant included five Lacto-
bacillus species, Acetobacter lovaniensis/fabarum and
*Correspondent: Fax: +52 55 57 29 62 07; two yeast species (S. cerevisiae and Dekkera bruxellen-
e-mail: chdez38@hotmail.com sis). The entire sugar source was consumed after 96 h

doi:10.1111/ijfs.13312
© 2016 Institute of Food Science and Technology
2 Fermenting microorganism of water kefir A. Martınez-Torres et al.

of fermentation, and the first to be completely and the pressure 10 PSI. A sample of 3 mL was intro-
exhausted was sucrose at 24 h. The most representa- duced into a 20-mL Pyrex glass vial, which was sealed
tive metabolites were ethanol and lactic acid, although with a silicon/Teflon septum and an Al tap. Operating
some acetic acid appeared at 144 h (Laureys & De conditions for the HS-GC-FID are specified in Table 1.
Vuyst, 2014). Lactic acid was determined using HPLC, as previ-
The aim of the present study was to assign a role, ously described (Gezginc et al., 2015), on a Varian
during WK fermentation, to the main microbial guilds 920-LC HPLC apparatus and a C18 column. The
– yeasts, lactic acid bacteria and acetic acid bacteria. absorbance was detected at a wavelength of 214 nm
A general characterisation was formulated of the order and a temperature of 30 °C.
of primary metabolite accumulation as fermentation
progressed. Then, an evaluation was made of the
Microbial isolation
microbial consumption of carbon sources as well as
the production by specific microorganisms of meta- For microbial isolation, samples were taken every 6 h
bolic products in fresh medium and in sterilised broth during 72 h of fermentation. WKGs were strained and
from different stages of WK fermentation. In this way, 1 g was crushed with a pestle and suspended in 9 mL
it was possible to establish the main producers of of distilled water. Decimal dilutions were carried out
metabolites in the system. The current results represent until reaching 106 and then were inoculated on five
an advance in the understanding of microbial succes- different media: MRS medium (De Man et al., 1960),
sion and dynamics during WK fermentation. MRS modified medium with sucrose instead of glu-
cose, GYC medium [10% wt/vol dextrose, 1% wt/vol
yeast extract, 2% wt/vol calcium carbonate], YPD
Materials and methods
medium [2% wt/vol dextrose, 1% wt/vol yeast extract,
2% wt/vol peptone] and sucrose–casein peptone med-
Conditions and sampling time of WKGs
ium [5% wt/vol sucrose, 1% wt/vol casein peptone].
WKGs were obtained from a traditional Mexican cul- Colonies having different macroscopic and microscopic
ture and propagated in a 5% panela solution with tap morphologies were isolated and preserved in 25% glyc-
water containing 0.2–1.5 mg L1 sodium hypochlorite, erol at 72 °C. Similar decimal dilutions were per-
incubated at 20 °C for 96 h. This was repeated ten formed with the fermentation broth or supernatant of
times before utilising the grains for further analyses. each sample until reaching 103.

Physical and chemical properties of WK fermentation DNA extraction, amplification and sequencing of the 16S
rRNA bacterial gene and the ITS1-5.8S rRNA-ITS2
Fermentation was performed in triplicate using
region of yeasts
100 mL of 5% panela solution in flasks, inoculated
with 1 g of WKGs and incubated at 26 °C for 240 h. DNA extraction was performed according to the
Samples of grains and supernatant, obtained every Hoffman & Winston (1987). PCR amplifications of
24 h, were conserved at 4 °C to measure wet weight 16S rRNA bacterial genes were performed utilising a
and pH as well as to determine the concentration of
sugars and the quantity of metabolites. The weight Table 1 HS-GC-FID conditions
increase curve was adjusted with the Gompertz equa-
tion, a sigmoid function widely used to describe micro- Headspace parameters
bial growth (Zwietering et al., 1990).
Headspace temperature 80 °C
The sucrose, fructose and glucose concentrations
Thermostatic heating time 5 min
were determined by Sucrose, D-Fructose and D-Glu- Pressurisation time 0.2 min
cose assay procedures of the K-SUFRG 06/14 Kit Injection time 0.12 min
(Megazyme, Ireland) according to the manufacturer’s Withdrawal time 0.5 min
instructions. Absorbance was measured with a micro- Needle temperature 110°C
plate spectrophotometer at 540 nm (Multiskan GO, Transfer line temperature 115°C
Thermo scientific, USA). Sample volume 3 mL
Ethanol and acetic acid were estimated by headspace GC parameters
gas chromatography (HS-GC-FID) on a Perkin-Elmer Temperature/time 1 55 °C/7 min
AutoSystem XL Gas Chromatograph fitted with a Temperature (oven)/time 2 160 °C/0 min
Temperature/time 3 220 °C/2.5 min
flame ionisation detector (FID) and a Perkin Elmer HS
Heating speed 1 10 °C/min
40 XL Automatic Headspace Sampler, using a CAR- Heating speed 2 20 °C/min
BOWAX column (Polyethylene glycol, 30 m; 32 mm Injector and detector temperature 150 °C
i.d.; film thickness 0.5 lm). The carrier gas was nitrogen

International Journal of Food Science and Technology 2016 © 2016 Institute of Food Science and Technology
Fermenting microorganism of water kefir A. Martınez-Torres et al. 3

reaction mixture containing 100 ng of template DNA,


Growth assays and metabolite production/consumption on
1X Green Go TaqÒFlexi Buffer, 10 pM of dNTPs
fresh and conditioned WK media
mixture, 1X GoTaqÒFlexi DNA, 10 pM of each pri-
mer and 3.5 mM of MgCl2 in a final volume of 25 lL. Fresh medium (FM, corresponding to 5% panela solu-
The universal bacterial primers 8F (50 GCG GAT tion) and supernatants from four different fermentation
CCG CGG CGG CTG CAG AGT TTG ATC CTG times, consisting of 24 h (24FWK), 48 h (48FWK),
GCT 30 ) and 1492R (50 GGC TCG AGC GGC CGC 72 h (72FWK) and 96 h (96FWK), were centrifuged at
CCG GGT TAC CTT GTT ACG ACT T 30 ) were 18 516 g for 30 min and successively filtered through
used (Relman, 1993). The PCR procedure included an membranes of different pore sizes (0.8, 0.6 and 0.2 lm).
initial denaturation step at 94 °C for 5 min, followed These sterilised samples of FM and FWK were used as
by 35 cycles at 94 °C for 1 min, 55 °C for 1 min, culture media for previously isolated bacteria (Aceto-
72 °C for 2.5 min and a final extension step at 72 °C bacter orientalis, Acetobacter okinawensis, Acetobacter
for 10 min. For PCR amplification of the ITS1-5.8S tropicalis, Pseudoarthrobacter chlorophenolicus, Lacto-
rRNA-ITS2 region of yeasts, the aforementioned bacillus casei and L. hilgardii) and yeasts (S. cerevisiae,
reaction mixture and PCR conditions were utilised Candida californica and Pichia membranifaciens).
and the primers ITS1 (50 -TCCGTAGGTGAACC Microbial growth was determined by absorbance
TGCGG-30 ) and ITS 4 (50 -TCCTCCGCTTATTGAT (620 nm) every 24 h for a total of 96 h, except that the
ATGC-30 ) were included in the reaction mix (White growth of L. hilgardii was measured by wet weight.
et al., 1990). PCR products were purified using the Metabolite production was analysed by HS-GC-FID
ZymocleanTM Gel DNA Recovery Kit (ZYMO and HPLC at each fermentation stage. The relevant fer-
RESEARCH, USA) and sequenced by a DNA mentations in the process were confirmed in modified
sequencing service (Macrogen Inc., Korea). MBMM supplemented with each carbon source. The
production or consumption of ethanol, lactic acid and
acetic acid was determined after 96 h of incubation in
Phylogenetic analysis for microbial identification
each condition, using HS-GC-FID and HPLC as
Sequences were compared using BLAST (Altschul described previously. For controls, FM and FWK
et al., 1990) and taxonomically related species were (from each time) were employed without inoculation.
collected from GenBank at the National Center for
Biotechnology Information (NCBI; http://www.ncbi.
Results
nlm.nih.gov/Taxonomy/taxonomyhome.html/). All the
sequences were aligned and manually edited with SEA-
Isolation and identification of microorganisms
VIEW software (Galtier et al., 1996). The most appro-
priate model of nucleotide substitution was selected Microbial identification by the phylogenetic approach
with the jModelTest 2.0 program (Darriba et al., indicated the presence of seven bacteria, including Lac-
2012). Phylogenetic trees were constructed using tobacillus ghanensis, L. casei/paracasei, L. hilgardii,
MEGA6 software (Tamura et al., 2013; http:// P. chlorophenolicus (formerly Arthrobacter chlorophe-
www.megasoftware.net/), with 1000 bootstrap replica- nolicus), A. orientalis, A. tropicalis and A. okinawensis
tions to assess nodal support in the tree. All sequences (Fig. 1), as well as three yeasts, S. cerevisiae, C. cali-
were submitted to GenBank under accession numbers fornica and P. membranifaciens (Fig. 2).
KX214626-KX214656 for bacteria and KX279834- Lactobacillus spp., Acetobacter spp. and S. cerevisiae
KX279858 for yeasts populations were estimated at approximately
107–108 UFC g1 in grain. Meanwhile, L. casei/para-
casei, the three aforementioned Acetobacter species
Carbon assimilation profiles of bacteria and yeasts
and S. cerevisiae were detected in the supernatant with
The carbon assimilation profiles of ten bacterial and populations of approximately 105–106 UFC mL1 at
twenty-two yeast isolates were determined in 96-well 24 h of incubation. L. hilgardii (WKGPB13) was only
plates with 100 lL of modified MES-buffered minimal isolated from WKGs, in sucrose MRS medium, panela
medium (MBMM) (Rathnayake et al., 2013), using medium and sucrose–casein peptone medium, produc-
ammonium nitrate instead ammonium chloride and ing irregular translucent jelly colonies with a hard con-
each carbon source at 0.5% (glucose, fructose, sucrose, sistency (Fig. 3).
lactic acid, acetic acid and ethanol). Additionally, both
organic acids and ethanol were evaluated at 0.25% as
Characterisation of WK fermentation
sole carbon sources. The plates were incubated at
28 °C for 168 h and absorbance was measured each The increase in wet weight and pH was determined
24 h, with a microplate spectrophotometer at 620 nm during 216 h of fermentation using 5% panela medium
(Multiskan FC, Thermo scientific, USA). inoculated with 1 g of WKGs per 100 mL. Grain

© 2016 Institute of Food Science and Technology International Journal of Food Science and Technology 2016
4 Fermenting microorganism of water kefir A. Martınez-Torres et al.

Figure 1 Maximum-likelihood phylogenetic


tree of a partial sequence of the 16S rRNA
gene in microorganisms isolated from water
kefir. With the Jukes–Cantor substitution
model, the sequence of Bacteroides fragilis
(NR119164) was used as the external group.
The bar indicates the number of substitu-
tions per site. The numbers in the nodes rep-
resent the bootstrap value of 1000 replicates.
The numbers in brackets are the GenBank
accession numbers of each sequence.

biomass increased 2.7 times and the initial pH of 7 The main metabolites in WK were quantified and
decreased to 3.5 (Fig. 4a). The biomass of WKGs kept their production was determined during 192 h of fer-
accumulating even after a pH of 4 was reached in the mentation (Fig. 4b). Ethanol started to accumulate at
supernatant. the onset of fermentation and reached its maximum

International Journal of Food Science and Technology 2016 © 2016 Institute of Food Science and Technology
Fermenting microorganism of water kefir A. Martınez-Torres et al. 5

Saccharomyces cerevisiae (KC588952)


WK11B, WK49, WK50, WK26, WK57, WK4B, WK13C
Saccharomyces cerevisiae (DQ167469)
Saccharomyces cerevisiae (KC254075)
99 WK146, WK90A, WK33B, WK45, WK33D, WK17C, WK209
Saccharomyces cerevisiae (KC515364)
97 WK79, WK44, WK165, WK93A, WK110, WK1, WK49
Saccharomyces cerevisiae (CP006468)
Saccharomyces yakushimaensis (AB097400)
Candida californica (JX188105)
Candida californica (JX188102)
WK64, WK172, WK214
100Candida californica (JX188103)
WK205
71
Candida californica (DQ104728)
Candida ethanolica (FJ662418)
Pichia membranifaciens (AF270935)
Figure 2 Maximum-likelihood phylogenetic 96 WK199
tree of a partial sequence of the ITS1-5.8S Pichia membranifaciens (DQ198954)
rRNA-ITS2 gene in microorganisms isolated 99
Pichia membranifaciens (KP132516)
from water kefir. The Tamura 3 substitution 70
91 Pichia membranifaciens (DQ104715)
model was used, with the sequence of
Yarrowia lipolytica (DQ683010) as the Pichia manshurica (FM199959)
99
external group. The bar indicates the Pichia fermentans (KF646196)
number of substitutions per site. The num- Candida inconspicua (KP131717)
bers in the nodes represent the bootstrap
Yarrowia lipolytica (DQ683010)
value of 1000 replicates. The numbers in
brackets are the GenBank accession numbers
of each sequence. 0.05

(a) (b)

Figure 3 (a) Plate with L. hilgardii colonies


in sucrose–casein peptone solid medium. (b)
Image of a single colony of L. hilgardii in
sucrose–casein peptone solid medium.

concentration of 4.29 g L1 at 96 h, and then gradu- The consumption of sucrose and the presence of glu-
ally diminished from 144 h to 192 h when it reached cose and fructose in the fermentation of WK were
0.7 g L1. Lactic acid accumulation was detected as of quantified as well (Fig. 4c). Sucrose was gradually
24 h, reached 1.38 g L1 at 48 h and then stabilised assimilated from 46.61 to 4.76 g L1 during the fer-
between 1.29 and 1.79 g L1 (with no reduction mentation. The end of sucrose assimilation was not
observed during fermentation). The accumulation of associated with a decrease in pH, evidenced by the fact
acetic acid began at 24 h and rose to a maximum con- that a pH of 3.6 was detected as of 72 h. The initial
centration of 28.35 g L1 at 192 h. concentration of glucose and fructose in the panela

© 2016 Institute of Food Science and Technology International Journal of Food Science and Technology 2016
6 Fermenting microorganism of water kefir A. Martınez-Torres et al.

medium was 0.83 and 2.04 g L1, respectively. The


Carbon assimilation profiles of bacteria and yeasts
level of both monosaccharides increased slightly by
24 h, to 1.5 and 3.04 g L1, respectively. This may Carbon utilisation tests were performed with ten bacte-
have been due to the presence of extracellular invertase ria (six Acetobacter spp., three Lactobacillus spp. and
activity, as occurred with S. cerevisiae (data not P. chlorophenolicus) and twenty-two yeasts (nineteen
shown). After 48 h, the values of glucose and fructose S. cerevisiae, two C. californica and one P. mem-
oscillated from 0.0–0.74 to 0.0–0.54 g L1, respec- braneafaciens). All bacteria were able to use sucrose,
tively. fructose or glucose, but not acetic acid as a sole car-
bon source (Fig. 5a). Acetobacter spp. grew better in
lactic acid (Fig. S1). As expected, yeasts could also
grow with glucose, fructose or sucrose as the sole car-
bon source. The majority of the S. cerevisiae, P. mem-
branifaciens and C. californica strains grew in ethanol
and a relatively low lactic acid concentration (0.25%)
(Fig. 5b). Only P. membranifaciens grew in acetic acid
(Fig. 6).

Figure 4 (a) pH variations and kinetics of the biomass of water


kefir grains during 216 h of fermentation (the kinetics data were
adjusted with the Gompertz equation). (b) The concentration of
ethanol, lactic acid and acetic acid during 192 h of WK fermenta-
tion. (c) The concentration of sucrose, glucose and fructose during Figure 5 Carbon source utilization by (a) bacteria and (b) yeasts
192 h of WK fermentation. isolated from WK. Glucose, fructose and sucrose were at 0.5 w/v.

International Journal of Food Science and Technology 2016 © 2016 Institute of Food Science and Technology
Fermenting microorganism of water kefir A. Martınez-Torres et al. 7

(Fig. 8c). In some stages, it seemed that the produc-


Microbial role in WK fermentation
tion and consumption of lactic acid occurred simulta-
This assay was carried out to explore the environmen- neously, although this effect was not confirmed in
tal changes that occurred during fermentation when modified MBMM supplemented with glucose, fructose,
using each of the following microorganisms: L. casei, sucrose or lactic acid (data not shown).
L. hilgardii, A. tropicalis, A. okinawensis, A. orientalis, Acetic acid was generated in FM and 24FWK,
P. chlorophenolicus, S. cerevisiae, P. membranifaciens mainly by L. hilgardii. In the later stages, it was pro-
and C. californica. Five fermentation stages were duced mainly by A. tropicalis, and to a lesser extent by
assayed, including FM, 24FWK, 48FWK, 72FWK A. orientalis, A. okinawensis, L. casei and
and 96FWK. P. chlorophenolicus (Fig. 8d).
After inoculating microorganisms in FM, the growth
of yeasts and P. chlorophenolicus was significant and
Discussion
that of L. casei slight (Fig. 7a). However, none of the
acetic acid bacteria grew in this condition. Growth WK should not be confused with milk kefir (MK),
was noted in the different fermentation stages as fol- another fermented beverage that needs polysaccharide
lows: the yeasts and A. okinawensis at 24FWK grains having microbes embedded as starters. MK uses
(Fig. 7b), the yeasts at 48FWK (Fig. 7c), S. cerevisiae milk and its by-products as the substrate for fermenta-
and P. membranifaciens at 72FWK (Fig. 7d) and none tion to yield kefiran grains, a glucogalactan mainly
of the microorganisms at 96FWK (Fig. 7e). L. hilgar- synthetised by Lactobacillus kefiranofaciens (Wszolek
dii showed increased biomass in FM, 24FWK and et al., 2001 Piermaria et al., 2008; Magalh~aes et al.,
48FWK. 2010a, 2011; Zajsek et al., 2011; Rodrigues et al.,
At the early stages of fermentation, the majority of 2016). Meanwhile, WKGs employ sucrose as the sub-
microorganisms reduced the pH of the medium com- strate and produce an a 1–6 glucose dextran syn-
pared to the internal control (Fig. 8a). The decrease in thetised by L. hilgardii (Pidoux et al., 1990; Waldherr
pH was particularly evident with acetic acid bacteria, et al., 2010). Although both beverages harbour micro-
reaching pH <5.3. bial communities that include yeast, lactic acid bacte-
Saccharomyces cerevisiae was the main ethanol pro- ria and acetic acid bacteria, the relative abundance
ducer in all stages, particularly early in fermentation and composition of each of these are very different
(FM, 24FWK and 48FWK). Ethanol was also gener- (Gulitz et al., 2013; Zamberi et al., 2016). The fermen-
ated by L. hilgardii in the late stages (72FWK and tation and growth of WKGs have been reported in
96FWK), but not by C. californica or P. membranefa- milk (Hsieh et al., 2012), and the particular WK
ciens at any stage (8B). On the other hand, A. tropi- employed in the current study can ferment milk. How-
calis consumed ethanol and transformed it into acetic ever, no growth of WKGs was observed, mainly
acid during the late stages (72FWK and 96FWK), because L. hilgardii does not assimilate or produce
causing almost complete depletion (Fig. 8b and d). polysaccharides from lactose (data not shown).
Lactobacillus hilgardii induced the accumulation of a Some reports have explored the microbial complexity
moderate amount of lactic acid, mainly in FM, in WK (Moinas et al., 1980; Pidoux, 1989; Gulitz et al.,
24FWK and 48FWK, while L. casei and P. chlorophe- 2011, 2013; Miguel et al., 2011; Laureys & De Vuyst,
nolicus produced scarce quantities of the same 2014). In this study, seven bacteria (L. ghanensis, L. ca-
sei/paracasei, L. hilgardii, A. orientalis, A. tropicalis,
A. okinawensis and P. chlorophenolicus) and three
yeasts (S. cerevisiae, C. californica and P. membranifa-
ciens) were isolated and identified. To our knowledge,
this is the first report of A. okinawensis and
P. chlorophenolicus isolated in this environment.
Lactobacillus hilgardii has been widely reported in
WK studies (Pidoux, 1989; Magalh~aes et al., 2010b;
Gulitz et al., 2011). It was presently isolated and cul-
tured in medium containing only sucrose, producing
colonies very similar to the WKGs found in panela
medium. Hence, this bacterium is probably responsible
for inducing the formation of these grains. Indeed,
L. hilgardii, previously known as Bacterium vermi-
forme, was reported as the polysaccharide producer in
Figure 6 Kinetics of Pichia membranifaciens growing in different WK in 1892 (Ward, 1892). Another study (Pidoux,
carbon sources. 1989), almost 100 years later, reported a colonial

© 2016 Institute of Food Science and Technology International Journal of Food Science and Technology 2016
8 Fermenting microorganism of water kefir A. Martınez-Torres et al.

Figure 7 During 96 h of incubation, kinetics of the growth of bacteria and yeasts isolated from WK in the different fermentation stages: (a)
FM, (b) 24FWK, (c) 48FWK, (d) 72FWK and (e) 96FWK.

morphology of pure cultures of L. hilgardii similar to L. hordei, can produce polysaccharides and have been
that observed presently. It must be mentioned that related to WKG formation (Pidoux et al., 1988; Gulitz
there are other reports of the same species forming et al., 2011). Nonetheless, the current results are in
slimy or mucoid colonies in sucrose MRS medium agreement with previous reports in the sense that that
(Waldherr et al., 2010; Davidovic et al., 2015). More- L. hilgardii is the microbe mainly responsible for the
over, other WK inhabitants, such as Lactobacillus bre- production of polysaccharides to form the grains of
vis, Leuconostoc mesenteroides, L. nagelii and WK fermentation (Pidoux, 1989).

International Journal of Food Science and Technology 2016 © 2016 Institute of Food Science and Technology
Fermenting microorganism of water kefir A. Martınez-Torres et al. 9

Figure 8 (a) pH variation produced by microorganisms isolated from WK in each fermentation stage. Concentration of (b) ethanol, (c) lactic
acid and (d) acetic acid in different media with microorganisms isolated from WK.

Although about twenty Lactobacillus and ten Aceto- (Magalh~aes et al., 2010b; Gulitz et al., 2011, 2013;
bacter species have been reported in WK beverages, in Laureys & De Vuyst, 2014). It turns out that the short
many studies, these species were identified only through length of the sequences (200–300 nt) obtained with
the use of biochemical and physiological tests (Pidoux, these methods are not suitable for species-level identifi-
1989; Rubio et al., 1993). Due to the phenotypic plastic- cation by the phylogenetic approach and similitude.
ity of bacteria and the limitations inherent in commer- Additionally, only BLAST analyses have been per-
cial identification tests, species are frequently formed without taking into account any formal taxo-
misidentified, particularly those in the Lactobacillus nomic criterion. As no accession numbers were
genus. DNA hybridisation analyses have proven that provided, confirmation, taxonomical reclassification
some biochemically identified species were misidentified, and phylogenetic comparisons are not possible. More-
such as confusing L. hilgardii with L. brevis (Lonvaud- over, no phylogenetic trees were included to show
Funel et al., 1991). The usefulness of API 50 CH carbo- formally described species, much less several strains
hydrate fermentation strips and other phenotype identi- of the identified species. Given that no reliable
fication methods is greatly diminished by a high level of species-level identifications have been performed, the
phenotype variability (as in the case of the commensal variety of microorganisms in WK may not be as broad
Lactobacillus bacterium from the human vagina) and as supposed.
limited databases. Hence, 16S rRNA sequencing may be The grain biomass herein exhibited a 2.7-fold
preferable (Boyd et al., 2005). increase, which is different from other reports on the
Indeed, bacterial 16S rRNA or the 18S or 26S quantification of grain biomass. Nevertheless, these
rRNA partial gene sequences of yeasts obtained from results must be carefully compared because of differ-
PCR-DGGE amplicons and pyrosequencing have been ences in the grains and the media employed, as well as
used for molecular identification of WK microbiota in the initial pH and incubation temperature. In a

© 2016 Institute of Food Science and Technology International Journal of Food Science and Technology 2016
10 Fermenting microorganism of water kefir A. Martınez-Torres et al.

previous study using 6% panela medium inoculated vinegars, which contain from 10.8 to 111.7 g L1 (Li
with 3 g of grains per 100 mL, there was only a 1.7- et al., 2015).
fold increase in the wet weight of WKGs, although the The content of sugars was also measured in the pre-
decrease in pH was similar to that found presently sent study, finding that during fermentation, the level
(pH 3.4) (Rubio et al., 1993). Another fermentation of of sucrose diminished, although this sugar was not
kefir, performed with brown sugar and figs, started at completely consumed. In other reports, sucrose was
a pH of 4.8 and obtained only a small increase in the exhausted and grain biomass growth was associated
biomass of the grains, approximately 1.7 fold, verify- with disaccharide availability (Laureys & De Vuyst,
ing the relevance of pH in this process (Laureys & De 2014). Another study reported that the supply of
Vuyst, 2014). Furthermore, in the present study, the sucrose was almost exhausted within 24 h, the level
biomass kept accumulating even after a pH of 4 was dropping from approximately 47.5 to 1.2 g L1, while
reached in the supernatant, which confirms that an ini- the level of fructose rose above 20 g L1. In the cur-
tial alkaline pH leads to higher biomass yields rent contribution, neither glucose nor fructose under-
(Fig. S2). It is also possible that the polysaccharide went any marked increase, their values remaining
biomass of WKGs buffers the environment inside the below 0.74 and 0.54 g L1, respectively, for most of
biofilm and temporally protects L. hilgardii and the the process. According to previous reports, the three
other microorganisms from the acidic pH of the super- main sugars are totally consumed before 96 h
natant. Other bacterial biofilm polysaccharides gener- (Magalh~aes et al., 2010b; Laureys & De Vuyst, 2014).
ate an internal environment protected from external The carbon assimilation profiles of the microorgan-
acidic conditions, such as the biofilms of Lactobacillus isms evidenced that all bacteria were able to use
plantarum and Streptococcus mutans (Zhu et al., 2001; sucrose, glucose or fructose as a carbon source. Aceto-
Kubota et al., 2008). bacter spp. grew better in lactic acid, a common fea-
Quantification of the main metabolites in WK dur- ture in this genus because it can oxidise lactate and
ing 192 h demonstrated that the first fermentation was acetate to carbon dioxide and water (Lisdiyanti et al.,
alcoholic, as has been reported previously with similar 2000, 2001; Iino et al., 2012). The yeasts could utilise
WK fermentations (Reiß, 1990; Rubio et al., 1993; most of the carbon sources, except that only P. mem-
Magalh~ aes et al., 2010b; Laureys & De Vuyst, 2014). branifaciens was able to proliferate in the presence of
The maximum ethanol concentration was 4.29 g L1, acetic acid. The Pichia genus (particularly P. membran-
a moderate value considering the broad range ifaciens) is polyphyletic. This species is associated with
obtained in other studies (0.18–20.3 g L1) (Rubio the Pichia clade that includes some Issatchenkia spp.
et al., 1993; Laureys & De Vuyst, 2014). After 144 h, as well as Candida species, such as C. ethanolica and
the ethanol content gradually declined until reaching C. californica (Kurtzman et al., 2008). P. membranifa-
0.7 g L1, an effect not previously reported for any ciens is highly distributed in fermentations, as are
other WK fermentation, although it is known to occur other related species in the same clade, presenting a
in apple cider vinegar production and spontaneous predilection for ethanol and simple organic acids as
cocoa bean fermentation (De Vuyst & Weckx, 2016; carbon sources and having poor fermentative power

Stornik et al., 2016). (Suh et al., 2006).
The second fermentation was lactic, and the maxi- To study their metabolic capacity and possible role
mum concentration of this metabolite was 1.70 g L1, during WK production, microorganisms were inocu-
less than that previously reported (Laureys & De lated in sterilised broth taken from WK at different
Vuyst, 2014). Even during a fermentation of only fermentation stages. Overall, yeasts were able to grow
24 h, lactic acid reached a concentration up to in most of the fermentation stages, while P. chlorophe-
2.5 mg mL1 (Magalh~ aes et al., 2010b). On the other nolicus and L. casei grew in FM and A. okinawensis in
hand, no decrease in accumulated lactic acid was 24FWK. Although the other bacteria did not prolifer-
herein observed during fermentation, which is in agree- ate, they were metabolically active at the different
ment with previous reports on WK fermentation stages of fermentation. The pH was reduced by all
(Rubio et al., 1993; Laureys & De Vuyst, 2014). bacteria in FM (except P. chlorophenolicus), possibly
Acetic fermentation occurred as of 24 h, resulting in because even a very low amount of lactic acid or acetic
an acetic acid concentration above 28 g L1 by the acid, when excreted in a panela medium with any
end of fermentation. This represents the highest level buffering capacity, is capable of acidifying it. In the
reported for WK, in which concentrations have ranged other fermentation stages, no pH decrease was
between 1 and 7 g L1 (Rubio et al., 1993; Laureys & observed compared to the internal control.
De Vuyst, 2014). In Mexico, it is common for WK or Previous reports on WK have assumed that ethanol
‘tepache de tibicos’ to be fermented for 2–3 weeks to and acetic acid are afforded by any yeast or acetic acid
obtain homemade vinegar, a product with an acetic bacterium, respectively, present in the consortium
acid content similar to other commercial or traditional (Magalh~aes et al., 2010b; Laureys & De Vuyst, 2014).

International Journal of Food Science and Technology 2016 © 2016 Institute of Food Science and Technology
Fermenting microorganism of water kefir A. Martınez-Torres et al. 11

The current results are not in agreement with this The present identification of the most metabolically
assumption. Indeed, ethanol was mainly yielded by active species during WK fermentation has provided
S. cerevisiae, a traditional ethanol producing yeast further insights into the dynamics of this process,
related to the production of alcoholic beverages (Fay allowing for the proposal of a minimal and efficient
& Benavides, 2005). The two other yeasts identified by consortium for WK production, formed by L. hilgar-
the phylogenetic approach herein employed, C. califor- dii, S. cerevisiae and A. tropicalis. This simplified con-
nica and P. membranifaciens, were unable to produce sortium could be used in a new fermentation process
ethanol in any fermentation stage. for beverage development, focusing on the examina-
Acetic acid, on the other hand, was initially gener- tion of microbial interaction and expression patterns
ated by L. hilgardii, and later by Acetobacter spp., as well as in vitro evolution, among other experiments.
mostly A. tropicalis. The latter species consumed etha- In previous reports on WK fermentations, the
nol and oxidised it to acetic acid. After 96 h, other assignment of the fermentation process to particular
microorganisms also produced acetic acid, which yeasts and bacteria was based on previous knowledge
explains both the decrease in ethanol at 144 h and the of the fermentation abilities of each microbial group.
high concentration of acetic acid at 192 h of fermenta- To our knowledge, the current contribution demon-
tion. In a previous study, acetic acid bacteria appeared strates for the first time the production of the main
after 144 h and therefore were not an important part primary metabolites by specific microorganisms after
of the WK ecosystem (Laureys & De Vuyst, 2014). It inoculating these into fermentation broths from differ-
is evident that kefir beverages are very different ent stages (thus having distinct chemical and physico-
according to the particular region (even within the chemical compositions).
same country), as evidenced by a Brazilian report In conclusion, the present WK preparation resulted
(Miguel et al., 2011). in an initial increase (appearing as of 0 h) in alcohol,

Sucrose
(a) (b) 0h
Extracellular invertase

Intracellular invertase

pH 7
Glucansucrase

Biomass 1g
g L–1
–1
S. cerevisiae A. tropicalis L. hilgardii Sugars Metabolites gL
A. orientalis Sucrose 46.6 Ethanol 0
P. membranifaciens Glucose 0.83 Lactic acid 0
Fructose 2.04 Acetic acid 0

24 h
pH 4.97
Fructose Glucose Fructose Glucose Fructose Dextran Biomass 1.52 g
g L–1
–1
(Polysaccharide grain) Sugars Metabolites gL
Sucrose 36.06 Ethanol 1.62
Glycolytic enzymes

Glucose 1.5 Lactic acid 0.09


Fructose 3.04 Acetic acid 0.51

All species 48 h
pH 4.02
Piruvate
Biomass 2.34 g
Sugars g L–1 Metabolites gL
–1

Sucrose 17.6 Ethanol 4.11


dehydrogenase
dehydrogenase
decarboxylase

Glucose 0.7 Lactic acid 1.37


Pyruvate
Pyruvate
Pyruvate

Fructose 0.54 Acetic acid 2.97


Lactate dehydrogenase

72 h
Aldehyde pH 3.61
dehydrogenase Biomass 2.75 g
g L–1
–1
Acetaldehyde Acetyl CoA Sugars Metabolites gL
A. tropicalis Sucrose 15.56 Ethanol 4.2
CoA transferase
CoA transferase

L. hilgardii S. cerevisiae L. hilgardii


A. tropicalis A. orientalis A. tropicalis Glucose 0 Lactic acid 1.12
A. orientalis A. orientalis Fructose 0.24 Acetic acid 8.99
Alcohol dehydrogenase
Alcohol dehydrogenase

96 h
pH 3.59
Biomass 3.13 g
Acetyl-P
g L–1
–1
Sugars Metabolites gL
Acetate kinase

Acetate kinase

Sucrose 13.56 Ethanol 4.23


Glucose 0.17 Lactic acid 1.44
Fructose 0.31 Acetic acid 13.29

Lactic acid Ethanol Acetic acid

Figure 9 (a) Hypothetical flux of carbon during a WK fermentation from sucrose to the main metabolic products. The alcoholic, lactic acid
and acetic acid fermentations and production of acetic acid from ethanol are indicated. WK microorganisms are located in each relevant meta-
bolic pathway of sucrose hydrolysis and the various fermentations. The hypothetical enzymes in each reaction are shown next to the arrows.
(b) The substrate consumption, pH evolution, biomass accumulation and fermentation progress from 0 to 96 h.

© 2016 Institute of Food Science and Technology International Journal of Food Science and Technology 2016
12 Fermenting microorganism of water kefir A. Martınez-Torres et al.

this being most abundantly yielded by S. cerevisiae. Fay, J.C. & Benavides, J.A. (2005). Evidence for domesticated and
Lactic and acetic acid began to accumulate after 24 h, wild populations of Saccharomyces cerevisiae. PLoS Genetics, 1,
e5.
mainly produced by L. hilgardii (the producer of Galtier, N., Gouy, M. & Gautier, C. (1996). SEAVIEW and PHY-
polysaccharide in WKG) and A. tropicalis, respec- LO_WIN: two graphic tools for sequence alignment and molecular
tively. Finally, the fermentation process generated an phylogeny. Computer Applications in the Biosciences, 12, 543–548.
acetic acid accumulation and WK vinegar as of 192 h. Gezginc, Y., Topcal, F., Comertpay, S. & Akyol, I. (2015). Quanti-
tative analysis of the lactic acid and acetaldehyde produced by
By the end of fermentation, ethanol had almost been Streptococcus thermophilus and Lactobacillus bulgaricus strains iso-
entirely consumed and oxidised to acetic acid, possibly lated from traditional Turkish yogurts using HPLC. Journal of
by a dissimilatory route of Acetobacter spp. Based on Dairy Science, 98, 1426–1434.
the present results, we suggest an original hypothetical Guerra, M.J. & Mujica, M.V. (2010). Physical and chemical proper-
scheme of the carbon flow in WK production from ties of granulated cane sugar “panelas”. Food Science and Technol-
ogy (Campinas), 30, 250–257.
sucrose to metabolite products (Fig. 9a), and provide Gulitz, A., Stadie, J., Wenning, M., Ehrmann, M.A. & Vogel, R.F.
an image outlining the changes occurring in WK dur- (2011). The microbial diversity of water kefir. International Journal
ing 96 h of fermentation (Fig. 9b). of Food Microbiology, 151, 284–288.
Gulitz, A., Stadie, J., Ehrmann, M.A., Ludwig, W. & Vogel, R.F.
(2013). Comparative phylobiomic analysis of the bacterial commu-
Acknowledgments nity of water kefir by 16S rRNA gene amplicon sequencing and
ARDRA analysis. Journal of Applied Microbiology, 114, 1082–
We give thanks to Bruce Allan Larsen for reviewing 1091.
the English version of this manuscript and to Jesus Hoffman, C.S. & Winston, F.A. (1987). A ten-minute DNA prepara-
tion from yeast efficiently releases autonomous plasmids for trans-
Luna for photos of L. hilgardii. We are grateful for formation of Escherichia coli. Gene, 57, 267–272.
the support provided by the Central de Instru- Hsieh, H.H., Wang, S.Y., Chen, T.L., Huang, Y.L. & Chen, M.J.
mentacion de Espectroscopıa and the Departamento (2012). Effects of cow’s and goat’s milk as fermentation media on
de Graduados en Alimentos, both of the ENCB-IPN, the microbial ecology of sugary kefir grains. International Journal
and are beholden to Dr. Juan Carlos Villalobos Rocha of Food Microbiology, 157, 73–81.
Iino, T., Suzuki, R., Kosako, Y., Ohkuma, M., Komagata, K. &
for his critical review. AMT, SGA and PHO thank Uchimura, T. (2012). Acetobacter okinawensis sp. nov., Acetobacter
CONACyT and BEIFI-IPN for fellowships. This work papayae sp. nov., and Acetobacter persicus sp. nov.; novel acetic
was supported by the Secretarıa de Ciencia y Tec- acid bacteria isolated from stems of sugarcane, fruits, and a flower
nologıa del DF (PICSO12-136, ICYTDF/206/2012) in Japan. The Journal of General and Applied Microbiology, 58,
235–243.
and the Instituto Politecnico Nacional, SIP 20151102 Kubota, H., Senda, S., Nomura, N., Tokuda, H. & Uchiyama, H.
and SIP 20161850. CHHR and LVT are thankful for (2008). Biofilm formation by lactic acid bacteria and resistance to
COFFAA, SNI and EDI fellowships. environmental stress. Journal of Bioscience and Bioengineering, 106,
381–386.
Kurtzman, C.P., Robnett, C.J. & Basehoar-Powers, E. (2008). Phylo-
References genetic relationships among species of Pichia, Issatchenkia and
Williopsis determined from multigene sequence analysis, and the
Altschul, S., Gish, W., Miller, W., Myers, E. & Lipman, D. (1990). proposal of Barnettozyma gen. nov., Lindnera gen. nov. and Wick-
Basic local alignment search tool. Journal of Molecular Biology, erhamomyces gen. nov. FEMS Yeast Research, 8, 939–954.
215, 403–410. Laureys, D. & De Vuyst, L. (2014). Microbial species diversity, com-
Bergmann, R.S.D.O., Pereira, M.A., Veiga, S.M.O.M., Schneedorf, munity dynamics, and metabolite kinetics of water kefir fermenta-
J.M., Oliveira, N.D.M.S. & Fiorini, J.E. (2010). Microbial profile tion. Applied and Environmental Microbiology, 80, 2564–2572.
of a kefir sample preparations: grains in natural and lyophilized Leroi, F. & Pidoux, M. (1993). Characterization of interactions
and fermented suspension. Food Science and Technology (Camp- between Lactobacillus hilgardii and Saccharomyces florentinus iso-
inas), 30, 1022–1026. lated from sugary kefir grains. Journal of Applied Bacteriology, 74,
Boyd, M.A., Antonio, M.A. & Hillier, S.L. (2005). Comparison of 54–60.
API 50 CH strips to whole-chromosomal DNA probes for identifi- Li, S., Li, P., Feng, F. & Luo, L.X. (2015). Microbial diversity and
cation of Lactobacillus species. Journal of Clinical Microbiology, their roles in the vinegar fermentation process. Applied Microbiol-
43, 5309–5311. ogy and Biotechnology, 99, 4997–5024.
Darriba, D., Taboada, G.L., Doallo, R. & Posada, D. (2012). jMo- Lisdiyanti, P., Kawasaki, H., Seki, T., Yamada, Y., Uchimura, T. &
delTest 2: more models, new heuristics and parallel computing. Komagata, K. (2000). Systematic study of the genus Acetobacter
Nature Methods, 9, 772–772. with descriptions of Acetobacter indonesiensis sp. nov., Acetobacter
Davidovic, S.Z., Miljkovic, M.G., Antonovic, D.G., Rajilic- tropicalis sp. nov., Acetobacter orleanensis (Henneberg 1906) comb.
Stojanovic, M.D. & Dimitrijevic-Brankovic, S.I. (2015). Water nov., Acetobacter lovaniensis (Frateur 1950) comb. nov., and Aceto-
Kefir grain as a source of potent dextran producing lactic acid bac- bacter estunensis (Carr 1958) comb. nov. The Journal of General
teria. Chemical Industry, 69, 595–604. and Applied Microbiology, 46, 147–165.
De Man, J.C., Rogosa, D. & Sharpe, M.E. (1960). A medium for Lisdiyanti, P., Kawasaki, H., Seki, T., Yamada, Y., Uchimura, T. &
the cultivation of lactobacilli. Journal of Applied Bacteriology, 23, Komagata, K. (2001). Identification of Acetobacter strains isolated
130–135. from Indonesian sources, and proposals of Acetobacter syzygii sp.
De Vuyst, L. & Weckx, S. (2016). The cocoa bean fermentation pro- nov., Acetobacter cibinongensis sp. nov., and Acetobacter orientalis
cess: from ecosystem analysis to starter culture development. Jour- sp. nov. The Journal of General and Applied Microbiology, 47, 119–
nal of Applied Microbiology, 121, 5–17. 131.

International Journal of Food Science and Technology 2016 © 2016 Institute of Food Science and Technology
Fermenting microorganism of water kefir A. Martınez-Torres et al. 13

Lonvaud-Funel, A., Fremaux, C., Biteau, N. & Joyeux, A. (1991). Stadie, J., Gulitz, A., Ehrmann, M.A. & Vogel, R.F. (2013). Meta-
Speciation of lactic acid bacteria from wines by hybridization with bolic activity and symbiotic interactions of lactic acid bacteria and
DNA probes. Food Microbiology, 8, 215–222. yeasts isolated from water kefir. Food Microbiology, 35, 92–98.
Magalh~ aes, K.T., Pereira, M.A., Nicolau, A. et al. (2010a). Produc- 
Stornik, A., Skok, B. & Trcek, J. (2016). Comparison of cultivable
tion of fermented cheese whey-based beverage using kefir grains as acetic acid bacterial microbiota in organic and conventional apple
starter culture: Evaluation of morphological and microbial varia- cider vinegar. Food Technology and Biotechnology, 54, 113–119.
tions. Bioresource Technology, 101, 8843–8850. Suh, S.O., Blackwell, M., Kurtzman, C.P. & Lachance, M.A. (2006).
Magalh~ aes, K.T., Pereira, G.V.M., Dias, D.R. & Schwan, R.F. Phylogenetics of Saccharomycetales, the ascomycete yeasts.
(2010b). Microbial communities and chemical changes during fer- Mycologia, 98, 1006–1017.
mentation of sugary Brazilian kefir. World Journal of Microbiology Tamura, K., Stecher, G., Peterson, D., Filipski, A. & Kumar, S.
and Biotechnology, 33, 1–10. (2013). MEGA6: molecular evolutionary genetics analysis version
Magalh~ aes, K.T., Dragone, G., de Melo Pereira, G.V. et al. (2011). 6.0. Molecular Biology and Evolution, 30, 2725–2729.
Comparative study of the biochemical changes and volatile com- Waldherr, F.W., Doll, V.M., Meißner, D. & Vogel, R.F. (2010).
pound formations during the production of novel whey-based kefir Identification and characterization of a glucan-producing enzyme
beverages and traditional milk kefir. Food Chemistry, 126, 249–253. from Lactobacillus hilgardii TMW 1.828 involved in granule forma-
Miguel, M.G.D.C.P., Cardoso, P.G., Magalh~aes, K.T. & Schwan, tion of water kefir. Food Microbiology, 27, 672–678.
R.F. (2011). Profile of microbial communities present in tibico Ward, H.M. (1892). The gingerbeer plant and the organisms com-
(sugary kefir) grains from different Brazilian States. World Journal posing it; a contribution to the study of fermentation yeast and
of Microbiology and Biotechnology, 27, 1875–1884. bacteria. Philosophical Transactions of the Royal Society of London
Moinas, M., Horisberger, M. & Bauer, H. (1980). The structural B, 83, 125–197.
organization of the Tibi grain as revealed by light, scanning and White, T.J., Bruns, T., Lee, S. & Taylor, J.W. (1990). Amplification
transmission microscopy. Archives of Microbiology, 128, 157–161. and direct sequencing of fungal ribosomal RNA genes for phyloge-
Pidoux, M. (1989). The microbial flora of sugary kefir grain (the gin- netics. In: PCR Protocols: A Guide to Methods and Applications
gerbeer plant): biosynthesis of the grain from Lactobacillus hilgardii (edited by M.A. Innis, D.H. Gelfand, J.J. Sninsky & T.J. White).
producing a polysaccharide gel. MIRCEN Journal of Applied Pp. 315–322. New York: Academic Press.
Microbiology and Biotechnology, 5, 223–238. Wszolek, M., Tamime, A.Y., Muir, D.D. & Barclay, M.N.I. (2001).
Pidoux, M., Brillouet, J. & Quemener, B. (1988). Characterization of Properties of kefir made in Scotland and Poland using bovine,
the polysaccharides from a Lactobacillus brevis and from sugary caprine and ovine milk with different starter cultures. LWT-Food
kefir grains. Biotechnology Letters, 10, 415–420. Science and Technology, 34, 251–261.
Pidoux, M., De Ruiter, G.A., Brooker, B.E., Colquhoun, I.J. & Zajsek, K., Kolar, M. & Gorsek, A. (2011). Characterisation of the
Morris, V.J. (1990). Microscopic and chemical studies of a gelling exopolysaccharide kefiran produced by lactic acid bacteria
polysaccharide from Lactobacillus hilgardii. Carbohydrate Poly- entrapped within natural kefir grains. International Journal of
mers, 13, 351–362. Dairy Technology, 64, 544–548.
Piermaria, J.A., Mariano, L. & Abraham, A.G. (2008). Gelling prop- Zamberi, N.R., Mohamad, N.E., Yeap, S.K. et al. (2016). 16S
erties of kefiran, a food-grade polysaccharide obtained from kefir Metagenomic microbial composition analysis of kefir grain using
grain. Food Hydrocolloids, 22, 1520–1527. MEGAN and BaseSpace. Food Biotechnology, 30, 219–230.
Rathnayake, I.V.N., Megharaj, M., Krishnamurti, G.S.R., Bolan, Zhu, M., Takenaka, S., Sato, M. & Hoshino, E. (2001). Influence of
N.S. & Naidu, R. (2013). Heavy metal toxicity to bacteria—Are starvation and biofilm formation on acid resistance of Streptococ-
the existing growth media accurate enough to determine heavy cus mutans. Oral Microbiology and Immunology, 16, 24–27.
metal toxicity? Chemosphere, 90, 1195–1200. Zwietering, M. H., Jongenburger, I., Rombouts, F. M. & Van’t Riet,
Reiß, J. (1990). Metabolic activity of Tibi grains. Zeitschrift f€ ur K. (1990). Modeling of the bacterial growth curve. Applied and
Lebensmittel-Untersuchung und Forschung, 191, 462–465. Environmental Microbiology, 56, 1875–1881.
Relman, D.A. (1993). Universal bacterial 16S rRNA amplification
and sequencing. In: Diagnostic Molecular Microbiology: Principles
and Applications (edited by D.H. Persing, T.F. Smith, F.C. Ten- Supporting Information
over & T.J. White) Pp. 489–495. Washington, DC: American Soci-
ety for Microbiology. Additional Supporting Information may be found in
Rodrigues, K.L., Ara ujo, T.H., Schneedorf, J.M. et al. (2016). A the online version of this article:
novel beer fermented by kefir enhances anti-inflammatory and Figure S1. Kinetics of Acetobacter spp. growing in
anti-ulcerogenic activities found isolated in its constituents. Journal
of Functional Foods, 21, 58–69.
different carbon sources; A. tropicalis, A, okinawensis
Rubio, M.T., Lappe, P., Wacher, C. & Ulloa, M. (1993). Microbial and A. orientalis
and biochemical studies of the fermentation of sugary solutions Figure S2. Water kefir biomass production in differ-
inoculated with tibi grains. Revista Latinoamericana de Microbi- ent pH values.
ologıa, 35, 19–31.

© 2016 Institute of Food Science and Technology International Journal of Food Science and Technology 2016

You might also like