Professional Documents
Culture Documents
Please cite this article as: Kirchmayr, M.R., Segura-García, L.E., Lappe-Oliveras, P., Moreno-Terrazas,
R., Rosa, M.d.l., Mathis, A.G., Impact of environmental conditions and process modifications on
microbial diversity, fermentation efficiency and chemical profile during the fermentation of mezcal in
Oaxaca, LWT - Food Science and Technology (2017), doi: 10.1016/j.lwt.2016.12.052.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Title
PT
1
Unidad de Biotecnología Industrial, Centro de Investigación y Asistencia en Tecnología y
Diseño del Estado de Jalisco, A.C.
RI
2
Instituto de Biología, Universidad Nacional Autónoma de México.
SC
3
Departmento de Ingeniería y Ciencias Químicas, Universidad Iberoamericana.
Abstract
U
AN
Mezcal, a distilled alcoholic beverage derived from cooked agave plants, is manufactured
traditionally on a small scale where raw materials and production practices differ noticeably
between production regions, localities and even factories resulting in a set of highly distinctive
M
products. The present study aimed to draw a comprehensive picture of the mezcal production in
Oaxaca State, relating the microbial consortia with the chemical compounds detected during
mezcal fermentation, and to elucidate the impact of process modifications made in order to
D
combining traditional and molecular techniques, detecting up to 21 different yeast species and
27 different bacterial species during a single process. Fermentation kinetics and volatile
compound generation varied strongly between the studied processes. A highly diverse
EP
microbiota associated to these traditional alcoholic fermentations was detected wherein several
yeast species and numerous bacterial groups built up the microbial consortia. Modifications
introduced between two production seasons showed a clear impact on microbial diversity. The
C
Keywords
1 1. Introduction
2 Fermented alcoholic beverages and distilled spirits derived from agave plants are produced in most
3 parts of Mexico since prehispanic times. Tequila, the most famous representative, is produced with
4 just one agave species (Agave tequilana Weber var. azul) and frequently on an industrial level
PT
5 (Cedeño, 2003). Other distilled spirits such as Bacanora, Raicilla, Sotol and Mezcal are produced
RI
6 traditionally and on small scale (reviewed in Lappe-Oliveras et al., 2008). In the specific case of
7 Mezcal, different states of the country or parts of them belong to the denomination of origin
SC
8 “Mezcal” (States of Oaxaca, Guerrero, San Luis Potosi, Zacatecas, Durango and parts of
9 Tamaulipas, Michoacán, Guanajuato and recently Puebla). Both raw materials and production
U
10 practices differ noticeably between production regions, localities and even factories resulting in a
AN
11 set of highly distinctive products. In general, the manufacturing process involves six stages: harvest
12 of agave plants and removal of leafs, cooking of agave core or stem, milling or crushing of cooked
M
13 core, fermentation of agave must, distillation of fermented must, and optional aging (Figure 1).
D
14 In most regions, the production of mezcal is a secondary activity complementing other agricultural
15 activities carried out throughout the year. Often, no process control is applied and, as a result of a
TE
16 lack of measurements and awareness, the processes are commonly quite inefficient and product
17 quality varies enormously between batches. The demand of mezcal is increasing and little by little,
EP
18 distillers are making efforts to standardize the production processes by introducing common
19 practices such as the use of commercial or autochthonous yeast starter cultures and defined
C
20 fermentation setups. However, the impact of environmental factors such as localization of the
AC
21 distilleries, ambient temperature and the alteration of traditional practices have not been examined
22 even though pronounced changes in microbiota, fermentation kinetics and the chemical composition
24 Reports of the complex composition of different mezcals revealed that numerous aroma compounds
25 derive from raw materials while others are generated during the different production stages
1
ACCEPTED MANUSCRIPT
26 (Lachenmeier et al., 2006; Mancilla-Margalli and Lopez, 2002; Molina-Guerrero et al., 2007). The
27 extract of cooked agave is the fermentable substrate and its low assimilable nitrogen content,
28 fructose as the main fermentable carbohydrate and the presence of several inhibitory compounds
29 make this a challenging matrix for microbial growth (Arrizon and Gschaedler, 2002; Flores-Berrios
PT
30 et al., 2005). In spite of its importance few scientific studies were conducted on the identification
RI
32 Minakata et al. (2008) identified the yeast and bacteria present in the mezcal fermentation from
SC
33 Agave salmiana (mezcal from the state of San Luis Potosi). These authors showed that bacteria,
34 mainly lactic acid bacteria and Zymomonas mobilis, dominate microbial diversity present in this
U
35 mezcal fermentation. Only three yeasts species were identified: Clavispora lusitaniae, Pichia
39 lusitaniae, Torulaspora delbrueckii, Candida ethanolica and Saccharomyces exiguus were detected.
D
40 In another region of mezcal production, the state of Durango, Paez-Lerma et al. (2013) pointed out
TE
43 the different fermentations studied only S. cerevisiae and T. delbrueckii were found. This research
44 demonstrates that the S. cerevisiae strains obtained in Durango State are phylogenetically
C
45 independent from the strains isolated in other regions of Latin America and Europe. Finally,
AC
46 Narváez-Zapata et al. (2010) studied the bacteria present in fermentation of agave in the state of
48 composti, Lb. parabuchneri, and Lb. plantarum was described. No work has been published on the
49 microbiota present in the fermentations in the state of Oaxaca, and this despite this state is the
50 leading producer of mezcal of Mexico. Currently more than 80% of the certified production is
51 generated in Oaxaca.
2
ACCEPTED MANUSCRIPT
52 The present study aimed to draw a comprehensive picture of the mezcal production in Oaxaca State
53 of Mexico relating the microbial consortia with the chemical compounds detected during mezcal
54 fermentation, and to elucidate the impact of process modifications made in order to improve
55 fermentation performance.
PT
56 2. Material and methods
RI
57 2.1 Mezcal production and sample collection
SC
58 The characterization of the mezcal production processes was carried out in two distilleries in the
59 state of Oaxaca, Mexico. Distillery A, located in Las Margaritas, a small mountain settlement, in the
U
60 Municipality of San Pedro Totolápan (16° 62’ N, 96° 04’ W, 780 masl), annual mean temperature
AN
61 of 24.3 °C, with April and May being the warmest months with an average temperature of 36 °C;
62 Distillery B, in the town of Matatlán, Municipality of Santiago Matatlán, (16° 52’ 30’’ N, 96° 23’
M
63 44’’ W, 1740 masl), annual mean temperature of 20 °C, and the highest varies from 28 °C to 35 °C
65 In distillery A in 2008, after cooking and crushing of the agave stems, 800 Kg of agave pulp and
TE
66 fiber were poured into a 1200 L vat and left to sit for 24 hours, then 400 L of water at
67 approximately 25 °C were added and 24 hours later the vat content was stirred. In 2009, the same
EP
68 preparation steps were carried out however 600 Kg of agave pulp and fiber and 500 L of water was
69 used. Additionally, approximately 185g of ammonium sulfate was supplemented before water
C
70 addition. In distillery B in 2008, after cooking and crushing of the agave stems, 500 Kg of agave
AC
71 pulp and fiber were poured into a 1000 L vat and left to sit for several hours. Then, 200 L and 240 L
72 of warm water (40 ºC) were added in two stages with 24 hours difference and then the vat content
73 was stirred. In 2009, only 390 Kg of agave pulp and fiber were poured into the vat and left to sit for
74 several hours. Together with 500 L of warm water (38 ºC) 10 L of a wild inoculum were added and
75 the fermentation vat was covered with a plastic tarp. After 24 hours, the vat content was stirred and
3
ACCEPTED MANUSCRIPT
76 kept covered during fermentation. The wild inoculum was prepared 24 hours prior to fermentation
77 with 1 L of fermented agave must from a previous fermentation and 10 L of fresh agave juice,
78 containing both bacteria (107 CFU/mL) and yeasts (105-106 CFU/mL). Aeration was accomplished
79 with an air pump and temperature was maintained above 25 ºC with a light bulb.
PT
80 Two 15 mL agave must samples were collected in early April 2008 and 2009 at different times of
RI
81 fermentation, for microbial and chemical analyses. 13 and 14 samples were taken in 2008, and 11
82 and 12 in 2009 from distilleries A and B, respectively. After sampling, microbial analyses were
SC
83 performed in situ as described below; samples for chemical analyses were frozen immediately and
87 aliquots were spread plated in duplicate onto WL Nutrient agar (WLA) (Fluka, Buchs, Switzerland)
88 supplemented with chloramphenicol (100 µg/mL, Sigma-Aldrich, Steinheim, Germany) for yeast
D
89 count, or with cycloheximide (100 µg/mL, Sigma-Aldrich) for acidophilic bacteria; and onto
TE
90 Zymomonas agar (yeast extract 3 g/L, malt extract 3 g/L, casein peptone 5 g/L, glucose 20 g/L,
91 cycloheximide 0.01 %, ethanol 3 %, agar 15 g/L), incubated anaerobically for Zymomonas (Obire,
EP
92 2005). Plates were incubated in situ at environmental temperature for 7 days during which they
93 were checked daily for population and morphology of colonies. Colonies were selected based on
C
94 morphology characteristics and isolates were purified by sub culture for identification. For each
AC
97 Yeast isolates were identified by examination for their morpho-physiological properties (Kurtzman
98 et al., 2011) and by sequencing of the D1/D2 domain of the LSU 26S rRNA gene. Procedure for
99 DNA extraction and PCR amplification (primers NL1 (5´ -GCA TAT CAA TAA GCG GAG GAA
4
ACCEPTED MANUSCRIPT
100 AAG- 3´) and NL4 (5´ -GGT CCG TGT TTC AAG ACG G- 3’) (Invitrogen, Carlsbad, CA, USA))
102 All sequences were obtained by Macrogen USA Corp. (Rockville, MD, USA); consensus sequences
103 were generated with the CLC Main Workbench 5.5 (CLC Bio, Sweden) and aligned in GenBank
PT
104 database using the BLASTn program (Altschul et al., 1997).Yeast sequences with a 99-100 %
RI
105 similarity were considered co-specific (Supplementary Table S1).
SC
107 All bacteria isolates were preliminary characterized by macro and micro-morphological features
U
108 and physiological tests (Holt et al., 2000). The identification of the different bacterial groups was
AN
109 performed as follows: LAB according to the methodology of Kozaki et al. (1992) and Duan et al.
110 (2008); AAB following the procedures of Schüller et al. (2000); Zymomonas using API 50 CHE and
M
111 API 20E kits (bioMériux, Marcy I´Etoile, France) and the additional tests recommended by Coton
112 et al. (2006); Endospore forming bacilli with API 50 CHB (bioMériux) and complementary test
D
113 recommended by Ludwig et al. (2009); Gram positive, catalase positive cocci with API Staph
TE
114 (bioMériux); Gram negative, oxidase negative bacilli with API 20E; and Gram negative, oxidase
116 For PCR amplification of the referred E. coli 16S rRNA gene nucleotide sequence, cells were
117 directly collected from a fresh bacterial colony and suspended in 25 µL of PCR reaction mix
C
118 (Master mix, Qiagen, Germany) containing 1 µL of each eubacterial universal primers, P27F (5´-
AC
119 AGA GTT TGA TCC TGG CTC AG-3’) and P1495r (5´-CTA CGG CTA CCT TGT TAC GA-3´)
120 (Baggi et al., 2004) provided by IBT UNAM, México. PCR reaction was performed according to
121 the methodology of Esteve-Zarzoso et al. (1999) adapted for bacteria. PCR program began with an
122 initial denaturing step at 95 °C for 25min followed by 35 cycles of denaturation (1 min at 94 °C),
123 annealing (2 min at 55 °C) and elongation (2 min at 72 °C), and a final extension step for 10 min at
5
ACCEPTED MANUSCRIPT
124 72 °C. Sequences were obtained by the Laboratorio de Biología Molecular de la Biodiversidad y de
125 la Salud, Instituto de Biología UNAM (Mexico); alignment and edition were done with BioEdit
126 v7.0.5 program (Hall, 2007). Edited sequences were aligned in the GenBank database using
127 BLASTN program (Altschul et al., 1997). Bacteria sequences with a 99-100 % similarity were
PT
128 considered co-specific (Supplementary Table S2).
RI
129 2.6 Direct reducing sugars (DRS) determination
130 DRS concentration was determined by the 3,5-dinitrosalicylic acid (DNS) method (Miller, 1959).
SC
131 Sample absorbance was evaluated in a microplate reader Model 680XR (Biorad Laboratories Inc.,
132 CA, USA) at 540 nm; sugar concentration was calculated by comparison with a glucose calibration
135 Volatile compound profiles were obtained using a Head Space Sampler (SHS Model 7694 E,
136 Hewlett Packard, Agilent Technologies, Palo Alto, CA, USA) coupled to gas chromatograph (GC
D
137 Hewlett Packard 6890, Agilent Technologies, Palo Alto, CA, USA) with flame ionization detector
TE
138 (FID). Substances were separated under conditions as described before (Arellano et al., 2012).
139 Compound identification was done according to their retention time, comparing with the retention
EP
140 time obtained with ethanol, methanol, ethyl acetate, ethyl lactate, acetaldehyde, isobutanol,
141 propanol and amyl alcohols purchased from Sigma-Aldrich Canada (Oakville, ON, Canada), and
C
144 Fermentation yield was calculated by relating the amount of ethanol produced with the reducing
145 sugars consumed (initial sugar concentration minus residual sugars). Due to the heterogeneity of the
146 agave must, the initial concentration of sugar was estimated based on the amount of agave used in
6
ACCEPTED MANUSCRIPT
147 each fermentation (approximately 20 % w/w of cooked agave stems are DRS, unpublished results).
148 Based on the quantity of sugar consumed and considering the theoretical maximum yield of
149 conversion of sugars to ethanol (51 %) the amount of alcohol capable of being produced was
150 determined (Arrizon and Gschaedler, 2002). With the data of the concentration of ethanol obtained
PT
151 in each fermentation, the alcohol efficiency was calculated, i.e. the percentage of sugars that were
RI
153 3. Results and Discussion
SC
154 The four mezcal elaboration processes studied in 2008 and 2009 in distilleries A and B were similar
155 (Figure 1). In both distilleries, the raw material used was juice, pulp and bagasse from cooked A.
156
U
angustifolia cores, which was generally spontaneously fermented, excepting fermentation in
AN
157 distillery B in 2009. However, several differences were found in the setup of the fermentations,
158 which affected fermentation kinetics, microbial diversity as well as volatile composition of the
M
159 fermented musts. The evolution of some parameters measured during the fermentations carried out
D
160 in the two distilleries in 2008 and 2009 are shown in Figure 2.
TE
162 In distillery A, in the fermentation carried out in 2008, the initial concentration of fermentable
EP
163 sugars was about 20 g/L, which increased to 153.5 g/L after 24 hours due to a slow delivery of these
164 from the agave bagasse to the liquid medium. Thereafter, the sugars were utilized in accordance
C
165 with microbial growth and the production of ethanol. The agave extracts gave initial yeast
AC
166 populations of less than 102 CFU/mL, which quickly grew to about 2x107 CFU/mL before dying off
167 (Figure 2). In 2009, the initial concentration of fermentable sugars was higher (51.9 g/L) increasing
168 to 120 g/L after 12 hours of fermentation. Higher levels of sugars were found in the 2008
169 fermentation because more cooked agave and less water was utilized than in 2009, leading to a
170 higher ethanol concentration for that year. The agave extracts gave initial yeast populations of about
7
ACCEPTED MANUSCRIPT
171 107 CFU/mL, which quickly grew to about 108 CFU/mL after 12 hrs. The exogenous nitrogen
172 source gave rise to the high initial yeast population in the 2009 fermentation, which grew about one
173 log higher that in the 2008 process. As stated in section 2.1, the agave bagasse in poured into the
174 fermentation vats 24 hours before water addition, which explains this finding. Accordingly, few
PT
175 bacteria were detected in the fresh agave extracts for the 2008 fermentation but high initial
176 population (108 CFU/mL) occurred in the 2009 fermentation. For both fermentations, temperature
RI
177 was not controlled but it started at ambient of 26-27 ºC in the moment of water addition, increasing
SC
178 to 40 ºC before decreasing to 36-38 ºC at the end of both fermentations.
179 In distillery B, the initial concentration of fermentable sugars was also higher in the 2008
U
180 fermentation than in 2009 (159.7 g/L and 126.6 g/L) reaching its maximum of 272.5 g/L and 198.7
AN
181 g/L after 48 hours and 18 hours, respectively. As in distillery A, the higher amount of agave and the
182 lower volume of water utilized for vat formulation gave rise to these differences. The agave extracts
M
183 gave a low initial yeast population (<102 CFU/mL) in 2008, increasing to about 1x107 CFU/mL
184 after 24 hrs. The inoculum preparation carried out in 2009 prior to vat formulation, led to a higher
D
185 initial yeast population (9x105 CFU/mL) growing quickly to 1.84x107 CFU/mL. As conditions for
TE
186 inoculum preparation did not exclusively promote growth of yeast, also a higher bacterial
187 population was observed at the beginning of the 2009 fermentation. For both fermentation in
EP
188 distillery B, warm water was utilized for vat formulation, explaining similar initial temperatures
189 (34-34.5 ºC), dropping less in 2009 fermentation due to the higher metabolic activity of the
C
191 The duration of the fermentation processes studied in both distilleries was clearly different, being
192 approximately 50 % longer in distillery B. The temperature changes inside the fermentation vats
193 were clearly different. The higher ambient temperature (25-40 °C) combined with a higher
194 microbial load attributed to a shorter process in distillery A. Although the consumption rates of
195 carbohydrates and the final ethanol concentration were higher in distillery A, the calculated
8
ACCEPTED MANUSCRIPT
196 alcoholic fermentation efficiencies were considerably low (38.9 % and 34.9 %). A high residual
197 sugar concentration (62 g/L) was detected at the end of the 2008 fermentation, representing a great
198 loss in terms of ethanol production. Despite a lower residual DRS concentration, the alcoholic
199 fermentation efficiency calculated for the fermentation carried out in distillery A did not improve. A
PT
200 probable reason could be the use of great part of the carbohydrates for growth and not for ethanol
201 production. In distillery B, the consumption of sugars generally was slower. However, due to less
RI
202 residual sugars a similar efficiency (38.7 %) as in distillery A was achieved in 2008. In 2009, from
SC
203 less cooked agave more alcohol was produced which caused a substantial increase in efficiency of
U
205 3.2. Microbial diversity
AN
206 553 fungal isolates were obtained from the four processes which were identified as 33 different
207 yeast species in 19 genera (Table 2); this is the first report of yeast micobiota associated to mezcal
M
208 fermentations in Oaxaca State and reflects a more diverse microbiota than those reported in
D
209 processes carried out in other production regions (reviewed in Lappe et al., 2008). Several of these
210 genera (Table 2) had not been previously isolated from fermentations carried out with cooked agave
TE
211 (Escalante-Minakata et al., 2008; Lachance, 1995; Paez-Lerma et al., 2013; Verdugo-Valdéz et al.,
212 2011). The most frequently isolated yeast species from each of the fermentations was S. cerevisiae,
EP
213 representing a 43.8 and 53.1 % of the total yeast isolates in distillery A (2008, 71 isolates; 2009, 93
214 isolates), and 35.2 and 62.8 % of the isolates in distillery B (2008, 43 isolates; 2009 59 isolates),
C
215 respectively. This observation is concordant with previous studies on fermentations of cooked
AC
216 agave in other production regions where the dominance of this yeast has been observed (Lachance,
218 After S. cerevisiae, the next most prevalent species were Kluyveromyces marxianus,
220 (distillery A) and T. delbrueckii and Z. bisporus (distillery B), suggesting a constant participation in
9
ACCEPTED MANUSCRIPT
221 these fermentations (Table 2). Specifically, K. marxianus and T. delbrueckii seem to join S.
222 cerevisiae along alcoholic fermentations as they were also identified in processes in San Luis Potosi
223 (Verdugo-Valdéz et al., 2011), Durango (Paez-Lerma et al., 2013), Michoacán (Perez et al., 2013)
224 as well as in tequila production (Lachance, 1995). The other 28 identified yeast species were
PT
225 isolated only sporadically at low numbers making it impossible to state their contribution to
226 alcoholic fermentations with cooked agave. Meanwhile the number of yeasts in distillery A
RI
227 maintained similar between both years of sampling, a change of identified taxa was observed in
SC
228 2009. It seems that the addition of an exogenous nitrogen source did not affect the diversity
229 (Shannon Indices, Table 2). In contrast, in distillery B, both the number and the taxa of yeast
U
230 species changed, which may be due to the addition of the wild inoculum. The number of S.
231 cerevisiae isolates increased 28 %, probably because its dominance in the wild inoculum, which
AN
232 explains are more pronounced drop in diversity index (Shannon Indices, Table 2).
M
233 527 bacterial strains were isolated from the four monitored fermentations and were identified as 36
234 species belonging to 19 different genera (Table 3). Nevertheless, five groups of isolates could not
D
235 be classified at species level. The highest number of isolates corresponded to spore-forming bacteria
TE
236 making up to 50 % of the bacteria identified in 2009 in distillery B (Table 3). The presence of this
237 group can be explained by the contact of the agave stems after the cooking step with the soil and the
EP
238 stone mill and in general by the “open-air” design of the distilleries. However, the impact of these
239 bacteria on the alcoholic fermentation processes remains unknown. After spore-forming bacteria,
C
240 the next most prevalent prokaryotic group was lactic acid bacteria, representing up to 48.9 % of the
AC
241 isolates in 2008 in distillery B, with Lactobacillus brevis and Lb. casei being the most prevalent
242 species. In concordance with the studies of Escalante-Minakata et al. (2008) and Narváez-Zapata et
243 al. (2010), the presence of several LAB species as for example Lb. plantarum confirms that
245 mobilis in all processes confirms its strong association to alcoholic fermentations carried out with
10
ACCEPTED MANUSCRIPT
246 agave as raw material (Escalante et al., 2012). In fact, this bacterium was isolated and described for
247 the first time from pulque almost 100 years ago (Lindner, 1924; Lindner, 1928). In the present
248 study, for the first time, the presence of acetic acid bacteria (AAB) is reported for mezcal
249 fermentation, being most abundant in 2008 in distillery A; probably due to a prolonged duration of
PT
250 the fermentation. AAB have been reported for the fermentation of tequila and pulque, however, not
251 the same species were identified (Escalante et al., 2012; Lachance, 1995). The remaining bacterial
RI
252 isolates, which were found occasionally, could have been brought into the fermentation through the
SC
253 environment or by biological vectors.
254 The prokaryotic diversity reported in the present study is comparable between both distilleries in
U
255 2008 (Shannon Indices, Table 3) and to previous reports from other mezcal production regions,
AN
256 where matching bacterial groups have been reported (Escalante-Minakata et al., 2008; Narváez-
257 Zapata et al., 2010). The different groups of bacteria identified in the present work confirm that
M
258 fermentations carried out in the production of mezcal are complex, where several microbial groups
259 intervene resulting in acidic and alcoholic fermentation processes. Previous studies of the
D
260 spontaneous fermentations of agave sap for the elaboration of pulque described the presence of the
TE
261 same bacterial groups carrying out in parallel an acidic, alcoholic and viscid fermentation
262 (Escalante et al., 2008, 2012). In both distilleries, a decrease in the diversity of bacterial groups was
EP
263 observed when comparing processes sampled in 2008 and 2009, being more pronounced in
264 distillery A (Shannon Indices, Table 3). The addition of exogenous nitrogen (distillery A) showed a
C
265 slightly greater impact on diversity as the employment of the starter culture (distillery B), reducing
AC
266 the total number of species more than 60 %. This observation is supported by the production of
267 ethanol by yeasts (10.5 g/L at the beginning of fermentation, Table 1) and by higher growth rates of
268 specific yeast and bacterial populations before water addition. The prevalence of the different
269 bacteria during the whole fermentation may be explained in all processes by the constant
11
ACCEPTED MANUSCRIPT
270 availability of carbohydrates as well as by the relative low ethanol concentration, which probably
PT
273 The volatile compounds measured during the fermentation were ethyl lactate, ethyl acetate,
274 methanol, acetaldehyde, isobutanol, propanol and amyl alcohols. The Mexican official regulation
RI
275 for mezcal established concentration ranges for some of the mayor volatile compounds important
276 for the aroma profile of the final product, i.e. ethanol, methanol or higher alcohols.
SC
277 The concentration of esters as well as the relation between the principal two, ethyl lactate and ethyl
U
278 acetate, varied between the four fermentations, reaching in almost every case its maximum at the
AN
279 end of the process (Figure 3). During 2008, the concentrations of ethyl lactate were 48 mg/L and
280 200 mg/L, in distillery A and B, respectively. In the subsequent year, in none of the processes this
M
281 compound was detected. Regarding ethyl acetate it was found to be produced more during the
282 second year of sampling reaching values of 15.1 to 21.6 mg/L. Similar values were reported by in
D
283 fermented agave must from San Luis Potosí (Verdugo-Valdéz et al., 2011); probably these are
TE
284 within the range of concentrations of ethyl acetate produced during the traditional fermentations of
285 mezcal.
EP
286 LABs and AABs present during the fermentation of mezcal play a major role in the generation of
287 organic acids (i.e. lactic and acetic acid) which are substrates transformed to esters by yeast
C
288 enzymes during fermentation or by chemical processes during distillation. The decrease in diversity
AC
289 and number of isolates of LAB and the increase of AAB isolates in processes of 2009 with respect
290 to the year before, may explain the changes observed in the concentration of ethyl lactate and ethyl
291 acetate (Table 2). In order to confirm this observation further studies have to be conducted and the
292 conversion of organic acids into esters need to be evaluated in controlled conditions.
12
ACCEPTED MANUSCRIPT
293 Methanol is an alcohol that derives from the cooked agave stems; since during the cooking process
294 the methoxylated pectins present in the raw material are hydrolyzed and transformed to this
295 compound. Methanol as well as sugars are initially attached to the must fibers and pulp and its
296 release depends on the addition of warm water to the must and the stirring during the first stages of
PT
297 fermentation; the maximum concentration of methanol varied from 46.9 to 140.2 mg/L and between
298 45 to 77.4 mg/L at the end of the processes. These values are important because the NOM states a
RI
299 concentration of 100 mg/L of methanol as maximum allowed in the distillate, so a proper
SC
300 distillation must be performed to fulfill the regulation.
301 The acetaldehyde concentration also augments during the initial stage of the process due to the
U
302 increase in yeast counts. As it is an intermediate, its presence is thought to be related to an
AN
303 inefficient conversion to ethanol catalyzed by alcohol dehydrogenase produced by the
304 microorganism. In 2008, in distillery A, acetaldehyde and residual sugar concentrations at the end
M
305 of the process were high indicating an incomplete fermentation process. In the three other
306 monitored fermentations, the final level of acetaldehyde was low. In the fermented musts of San
D
307 Luis Potosí concentrations of 14.24 and 3.04 mg/L were detected at the end of the processes
TE
308 (Verdugo-Valdéz et al., 2011), the latter value is similar to those determined in distillery B in 2008
309 and in both processes in 2009; it can be considered that one of the fermentations was not efficient to
EP
311 The production of higher alcohols as propanol, isobutanol and amyl alcohols are associated to the
C
312 production of ethanol (Pretorius, 2000). The concentration of ethanol and higher alcohols showed a
AC
313 similar behavior. The most pronounced change in the composition of this volatile group was
314 observed in distillery A (Figure 3); in 2009, the production of propanol and isobutanol was higher
315 while the concentration of amyl alcohols was lower. This change is associated to the use of an
316 additional nitrogen source as it was observed before in other agave must fermentations (Arrizon and
13
ACCEPTED MANUSCRIPT
318 4. Conclusions
319 In the present study, a highly diverse microbiota associated to traditional alcoholic fermentations
320 carried out for the production of mezcal is reported. Not only several yeast species but also
321 numerous bacterial groups build up the microbial consortia associated to these processes. The
PT
322 present work is the first report integrating the identification of both yeast and bacteria with the
RI
323 generation of major volatile compounds produced during the fermentation of cooked agave. In spite
324 of the similarities found between the production processes carried out in both distilleries, the
SC
325 dissimilar set up of the fermentations and the environmental conditions, i.e. the ambient
326 temperature, led to clear differences in the kinetics and the outcome of the fermentations.
327
U
Modifications introduced in both distilleries showed a clear impact on microbial diversity,
AN
328 highlighting the need of prior isolation and preservation of microbial species. The employment of
329 wild inoculums in the fermentation of cooked agave juice may be a good practice in order to
M
330 increase alcoholic fermentation efficiency, nevertheless prominent changes in the volatile
D
331 compound profile would be expected. In order to explain and predict changes in the chemical
332 composition of the fermented musts, however, concrete studies with controlled mixed fermentations
TE
334 5. Acknowledgements
335 The authors thank the Mexican Council of Science and Technology (CONACYT) for the financial
C
336 support through student grants and the SEP-CONACYT grant 24556. Besides, the authors thank Dr.
AC
337 Graham Fleet for his critical observations made on the manuscript.
338 6. References
14
ACCEPTED MANUSCRIPT
339 Altschul, S.F., Madden, T.L., Schäffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J., 1997.
340 Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic
342 Arellano, M., Gschaedler, A., Alcazar, M., 2012. Major Volatile Compounds Analysis Produced
PT
343 from mezcal Fermentation Using Gas Chromatography Equipped Headspace (GC–HS). In: Salih B
RI
344 (Ed.), Gas Chromatography in Plant Science, Wine Technology, Toxicology and Some Specific
SC
346 Arrizon, J., Gschaedler, A., 2002. Increasing fermentation efficiency at high sugar concentrations
347 by supplementing an additional source of nitrogen during the exponential phase of the Tequila
348
U
fermentation process. Can. J. Microbiol. 48, 965-970.
AN
349 Arrizon, J., Gschaedler, A., 2007. Effects of the addition of different nitrogen sources in the Tequila
M
350 fermentation process at high sugar concentration. J. Appl. Microbiol. 102, 1123-1131.
351 Baggi, G., Cavalca, L., Francia, P., Zangrossi, M., 2004. Chlorophenol removal from soil
D
352 suspentions: effects of a specialised microbial inoculum and degradable analogue. Biodegradation
TE
354 Cedeño, M., 2003. Tequila production from agave: historical influences and contemporary process.
355 In: Jacques KA, Lyons TP and Kelsall DR (Eds.), The alcohol textbook, 4th edn. Nottingham
C
357 Coton, M., Laplace, J. M., Auffray, Y., Coton, E., 2006. Polyphasic study of Zymomonas mobilis
358 strains revealing the existence of a novel subspecies Z. mobilis subsp. francensis subsp. nov.,
359 isolated from French cider. Int. J. Syst. Evol. Microbiol. 56, 121-125.
360 Duan, Y., Tan, Z., Wang, Y., Li, Z., Qin, G., Huo, Y., 2008. Identification and characterization of
361 lactic acid bacteria isolated form Tibetan Qula cheese. J. Gral. App. Microbiol. 54, 51-60.
15
ACCEPTED MANUSCRIPT
362 Escalante, A., Giles-Gómez, M., Esquivel, G., Matus Acuña, V., Moreno Terrazas, R., López-
363 Munguía, A., Lappe-Oliveras, P., 2012. Pulque fermentation. In: Hui, Y.H., Özgül Evranuz, E.
364 (Eds.), Handbook of Plant-Based Fermented Food and Beverage Technology. CRC Press, Boca
PT
366 Escalante, A., Giles-Gómez, M., Hernández, G., Cordova-Aguilar, M.S., López-Munguía, A.,
RI
367 Gosset, G., Bolivar, F., 2008. Analysis of the bacterial community during the fermentation of
368 pulque, a traditional Mexican alcoholic beverage, using a polyphasic approach. Int. J. Food
SC
369 Microbiol. 124, 126-134.
370 Escalante-Minakata, P., Blaschek, H.P., Barba De La Rosa, A.P., Santos, L., De Leon-Rodriguez,
371
U
A., 2008. Identification of yeast and bacteria involved in the Mezcal fermentation of Agave
AN
372 salmiana. Lett. Appl. Microbiol. 46, 626-630.
M
373 Esteve-Zarzoso, B., Belloch, C., Uruburu, F., Querol, A., 1999. Identification of yeast by RFLP
374 analysis of the 5.8S rRNA gene and the two ribosomal internal transcribed spacers. Int. J. Syst.
D
376 Flores-Berrios, E.P., Alba González, J.F., Arrizon Gaviño, J.P., Romano, P., Capece, A.,
377 Gschaedler-Mathis, A., 2005. The uses of AFLP for detecting DNA polymorphism, genotype
EP
378 identification and genetic diversity between yeasts isolated from Mexican agave-distilled beverages
379 and from grape musts. Lett. Appl. Microbiol. 41, 147-52.
C
AC
380 Hall, T., 2007. BioEdit. Biological Sequence Alignment Editor for Win95/98/NT/2K/XP. Ibis
382 Holt, J.G., Krieg, N.R., Sneath, P.H.A., Staley, J.T., Williams, S.T., 2000. Bergey’s Manual of
383 Determinative Bacteriology, ninth ed. Lippincott, Williams & Wilkins, Philadelphia.
16
ACCEPTED MANUSCRIPT
384 Kozaki, M., Uchimura, T., Okada, S., 1992. Experimental Manual of Lactic Acid Bacteria.
386 Kurtzman, C.P., Fell, J.W., Boekhout, T., 2011. The Yeasts. A taxonomic study, Elsevier,
387 Amsterdam.
PT
388 Kurtzman, C.P., Robnett, C.J., 1998. Identification and phylogeny of ascomycetous yeasts from
RI
389 analysis of nuclear large subunit (26S) ribosomal DNA partial sequences. Antonie van
SC
391 Lachance, M.A., 1995. Yeast communities in a natural tequila fermentation. Antonie Van
U
392 Leeuwenhoek 68, 151-60.
AN
393 Lachenmeier, D.W., Sohnius, E.M., Attig, R., Lopez, M.G., 2006. Quantification of selected
394 volatile constituents and anions in Mexican Agave spirits (Tequila, Mezcal, Sotol, Bacanora). J.
M
396 Lappe-Oliveras, P., Moreno-Terrazas, R., Arrizon-Gaviño, J., Herrera-Suaréz, T., García-Mendoza,
TE
397 A., Gschaedler-Mathis, A., 2008. Yeasts associated with the production of Mexican alcoholic
398 nondistilled and distilled Agave beverages. FEMS Yeast Res. 8, 1037-1052.
EP
399 Lindner, P., 1924. La importancia práctica y científica del estudio del pulque. Revista Mexicana de
401 Lindner, P., 1928. Gärungsstudien über Pulque in Mexiko. Ber. Westpreuss. Bot. Zool. Ver. 50,
AC
402 253-255.
403 Ludwig, W., Schleifer, K., Whitman, W., 2009. Class I. Bacilli class. nov. In: Volume 3: The
404 Firmicutes, Bergey’s Manual of Systematic Bacteriology second ed. Springer, Dordrecht.
405 Mancilla-Margalli, N., López, M.G., 2002. Generation of Maillard compounds from inulin during
406 the thermal processing of Agave tequilana Weber var. azul. J. Agric. Food Chem. 50, 806-812.
17
ACCEPTED MANUSCRIPT
407 Miller, G.L., 1959. Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal.
410 Jiménez-Islas, H., Cárdenas-Manríquez, M., Rico-Martinez, R., 2007. Compuestos volátiles en el
PT
411 mezcal. Revista Mexicana de Ingeniería Química 6, 41-50.
RI
412 Narváez-Zapata, J.A., Rojas-Herrera, R.A., Rodríguez-Luna, I.C., Larralde-Corona, C.P., 2010.
413 Culture-independent analysis of lactic acid bacteria diversity associated with Mezcal fermentation.
SC
414 Curr. Microbiol. 61, 444-450.
U
415 Obire, O., 2005. Activity of Zymomonas species in palm-sap obtained from three areas in Edo State,
AN
416 Nigeria. Journal of Applied Sciences & Environmental Management. 9, 25-30.
417 Paez-Lerma, J.B., Arias-García, A., Rutiaga-Quiñones, O.M., Barrio, E., Soto-Cruz, N.O., 2013.
M
418 Yeasts isolated from the alcoholic fermentation of Agave duranguensis during mezcal production.
420 Pérez, E., González-Hernández, J.C., Chávez-Parga, M.C., Cortés-Penagos, C., 2013.
421 Caracterización fermentativa de levaduras productoras de etanol a partir de jugo de Agave cupreata
EP
423 Pretorius, I.S., 2000. Tailoring wine yeast for the new millennium: novel approaches to the ancient
C
425 Schüller, G., Hertel, C., Hammes, W., 2000. Gluconacetobacter entanii sp. nov., isolated from
426 submerged high acid industrial vinegar fermentations. Int. J. Syst. Evol. Microbiol. 50, 2013-2020.
427 Verdugo-Valdéz, A., Segura-Garcia, L., Kirchmayr, M., Ramírez-Rodríguez, P., González-
428 Esquinca, A., Coria, R., Gschaedler-Mathis, A., 2011. Yeast communities associated with artisanal
429 Mezcal fermentations from Agave salmiana. Antonie van Leeuwenhoek 100, 497-506.
18
ACCEPTED MANUSCRIPT
430 Tables
Distillery A Distillery B
Parameters
2008 2009 2008 2009
Vat volume (L) 1200 1200 1000 1000
Yeasts Initial <1.0E+02 1.76E+07 <1.0E+02 9.00E+05
PT
(CFU/mL) Maximum 2.84E+07 1.2 E+08 1.23E+07 1.84E+07
Final 1.00E+05 3.50E+04 4.50E+05 2.10E+06
RI
Bacteria Initial <1.0E+02 4.64E+09 1.51E+04 5.45E+06
(CFU/mL) Maximum 1.02E+07 4.64E+09 1.03E+07 1.90E+08
Final 8.01E+05 5.15E+08 1.11E+06 7.72E+07
SC
Temperature Initial 27 26 34 34.5
(°C) Maximum 40 40 34 34.5
Final 38 36 26.2 32
U
DRS Initial 20.1 51.9 159.7 126.6
AN
(g/L) Maximum 153.5 120.0 272.5 198.7
Residual 62.1 6.8 12.7 24.3
Ethanol Initial 0.3 10.5 0.1 0.4
M
19
ACCEPTED MANUSCRIPT
PT
432 *theoretical value, see section 2.8.
RI
U SC
AN
M
D
TE
C EP
AC
20
ACCEPTED MANUSCRIPT
Factory A Factory B
Yeast species
2008 2009 2008 2009
Candida apicola 5 (3.1%) 2 (1.1%) 2 (1.6%)
Candida boidinii 3 (2.5%) 1 (1.1%)
Candida parapsilosis 6 (3.7%)
Candida rugosa 2 (1.1%) 1 (1.1%)
PT
Candida zemplinina 2 (1.1%)
Citeromyces matritensis 1 (0.6%) 2 (1.6%)
Clavispora lusitaniae 2 (1.1%) 1 (1.1%)
RI
Cryptococcus albidus 1 (0.6%) 1 (0.8%) 2 (2.1%)
Cryptococcus uniguttulatus 1 (1.1%)
Debaryomyces hansenii 1 (0.6%) 2 (1.6%)
SC
Dekkera anomala 1 (0.6%) 5 (5.3%)
Hanseniaspora guilliermondii 1 (0.6%) 2 (1.1%) 2 (1.6%)
Hanseniaspora osmophila 1 (0.8%) 1 (1.1%)
U
Kluyveromyces lactis var. drosophilarum 6 (3.4%)
Kluyveromyces marxianus 3 (1.9%) 13 (7.4%) 5 (4.1%) 7 (7.4%)
AN
Meyerozyma guilliermondii 7 (4.3%) 5 (2.9%) 6 (4.9%)
Pichia fermentans 2 (1.1%)
Pichia kluyveri 1 (0.6%) 1 (0.8%)
M
21
ACCEPTED MANUSCRIPT
Factory A Factory B
Bacteria species
2008 2009 2008 2009
Lactic acid bacteria
Lactibacillus brevis 2 (15.4%) 14 (14.9%) 1 (0.7%)
Lactobacillus casei 8 (3.0%) 11 (11.7%) 3 (2.0%)
Lactobacillus ghanensis 1 (1.1%) 1 (0.7%)
PT
Lactobacillus harbinensis 4 (1.5%)
Lactobacillus hilgardii 1 (7.7%)
Lactobacillus paracasei 1 (0.4%) 7 (7.4%)
RI
Lactobacillus plantarum 2 (0.7%)
Lactobacillus spp. 30 (11.2%) 1 (7.7%) 11 (11.7%) 16 (10.5%)
Leuconostoc spp. 5 (1.9%) 1 (1.1%) 1 (0.7%)
SC
Oenococcus oeni 1 (0.4%)
Weisella spp. 1 (0.4%) 1 (1.1%)
Number of isolates 52 (19.4%) 4 (30.8%) 46 (48.9%) 22 (14.5%)
U
Number of species 8 3 7 5
AN
Acetic acid bacteria
Acetobacter pasteurianus 2 (0.7%) 1 (1.1%) 4 (2.6%)
Acetobacter tropicalis 4 (1.5%) 2 (15.4%) 2 (2.1%) 2 (1.3%)
Acetobacter senegalensis 9 (3.4%) 1 (7.7%) 2 (2.1%) 10 (6.6%)
M
Zymomonas
Zymomonas mobilis 27 (10.1%) 2 (15.4%) 11 (11.7%) 4 (2.6%)
Number of isolates 27 (10.1%) 2 (15.4%) 11 (11.7%) 4 (2.6%)
EP
Number of species 1 1 1 1
Aerobic sporulating bacteria
Bacillus amyloliquefaciens 11 (4.1%) 5 (5.3%) 3 (2.0%)
C
22
ACCEPTED MANUSCRIPT
PT
Rhodococcus erythropolis 4 (4.3%)
Rhodococcus qingshengii 1 (1.1%)
Number of isolates 1 (7.7%) 5 (5.3%) 1 (0.7%)
RI
Number of species 1 2 1
Gram positive, catalase positive, cocci
Kocuria rhizophila 31 (11.6%) 2 (2.1%) 14 (9.2%)
SC
Kocuria spp. 2 (2.1%)
Number of isolates 31 (11.6%) 4 (4.3%) 14 (9.2%)
Number of species 1 2 1
U
Enterobacteriaceae
AN
Enterobacter aerogenes 1 (1.1%) 1 (0.7%)
Enterobacter cloacae 10 (3.7%) 16 (10.5%)
Number of isolates 10 (3.7%) 1 (1.1%) 17 (11.2%)
Number of species 1 1 2
M
23
ACCEPTED MANUSCRIPT
PT
440 and Temperature ( ) during traditional fermentation of Mezcal in distilleries A in (A)
441 2008 and (B) 2009 and distillery B in (C) 2008 and (D) 2009.
RI
442 Figure 3: Production of volatile compounds during traditional fermentation of Mezcal in
SC
443 distillery A in (A) and 2008 (B) 2009 and distillery B in (C) 2008 and (D) 2009: ethyl
24
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
Highlights
PT
RI
U SC
AN
M
D
TE
C EP
AC