Mass spectrometry for the characterization of brewing process

FRIN-06362; No of Pages 8 Food Research International xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Food Research International journal homepage: www.elsevier.com/locate/foodres

Mass spectrometry for the characterization of brewing process Adriana Fu Vivian, Caroline Tiemi Aoyagui, Diogo Noin de Oliveira, Rodrigo Ramos Catharino ⁎ INNOVARE Biomarkers Laboratory, School of Pharmaceutical Sciences, University of Campinas, Campinas, São Paulo 13083-877, Brazil

a r t i c l e

i n f o

Article history: Received 1 June 2016 Received in revised form 10 August 2016 Accepted 13 August 2016 Available online xxxx Keywords: Beer Brewing Processomics Mass spectrometry PLS-DA

a b s t r a c t Beer is a carbonated alcoholic beverage produced by fermenting ingredients containing starch, especially malted cereals, and other compounds such as water, hops and yeast. The process comprises five main steps: malting, mashing, boiling, fermentation and maturation. There has been growing interest in the subject, since there is increasing demand for beer quality aspects and beer is a ubiquitous alcoholic beverage in the world. This study is based on the manufacturing process of a Brazilian craft brewery, which is characterized by withdrawing samples during key production stages and using electrospray ionization (ESI) high-resolution mass spectrometry (HRMS), a selective and reliable technique used in the identification of substances in an expeditious and practical way. Multivariate data analysis, namely partial least squares discriminant analysis (PLS-DA) is used to define its markers. In both positive and negative modes of PLS-DA score plot, it is possible to notice differences between each stage. VIP score analysis pointed out markers coherent with the process, such as barley components ((+)-catechin), small peptide varieties, hop content (humulone), yeast metabolic compounds and, in maturation, flavoring compounds (caproic acid, glutaric acid and 2,3-butanediol). Besides that, it was possible to identify other important substances such as off-flavor precursors and other different trace compounds, according to the focus given. This is an attractive alternative for the control of food and beverage industry, allowing a quick assessment of process status before it is finished, preventing higher production costs, ensuring quality and helping the control of desirable features, as flavor, foam stability and drinkability. Covering different classes of compounds, this approach suggests a novel analytical strategy: “processomics”, aiming at understanding processes in detail, promoting control and being able to make improvements, © 2016 Published by Elsevier Ltd.

1. Introduction Beer is a carbonated alcoholic beverage produced by fermenting ingredients containing starch, mainly malted grains such as barley and wheat (Beltramelli, 2013). In addition to cereals, beer has water, hops and yeasts, and may include other ingredients such as fruits, herbs and brewing adjuncts (Sleiman & Venturini Filho, 2008). In general, beers are composed of carbohydrates, minerals, vitamins, polyphenols and amino acids (Quifer-Rada et al., 2015). The beer production process, regardless of the batch size and style, involves the steps of malting, mashing, wort production, filtration, boiling, fermentation, maturation and bottling. The considerable interest surrounding beer is justifiable: it is one of the most widespread alcoholic beverages and is the fifth most consumed beverage in the world, standing only behind tea, coffee, milk and soda (Fillaudeau, Blanpain-Avet, & Daufin, 2006). Despite its millenary tradition, popularity and varieties, the quantification and identification of beer components and ingredients still represent a challenge due to the high complexity of the beer matrix, lack of commercial standard, instability of ingredients and structural similarity of the ⁎ Corresponding author. E-mail address: [email protected] (R.R. Catharino).

constituents (Ivanova & Spiteller, 2014). The main classification used is according to fermentation process. There are two large groups, which may be divided into Ales and Lagers. Ale beers are also known as top fermenting because its fermentation is at the top of the tanks. The lagers are known as bottom fermenting, since fermentation takes place at the bottom of the tanks (Beltramelli, 2013). In the malting process, starts the production of hydrolytic enzymes during germination, such as enzymes that break down starch (α-amylase and β-amylase), the cell walls (β-glucanases) and proteins (carboxypeptidases and endopeptidases). Then, to stop the enzymatic activity, the grains are dried at above 85 °C, ranging up to more than 200 °C depending on the aimed malt flavor and color (Baxter & Hughes, 2001). The drying process also has the function of removing volatiles that may contribute to the emergence of off-flavors in the final product and help in grain conservation, since it eliminates all the moisture (Morado, 2009). The next step involves grinding the malted grains in a mill. The crushed malt goes under heating to gelatinize the starch and make it more susceptible to enzymatic attack. The wort is the result of the breakdown of amylose and amylopectin molecules into small sugars, as maltose and glucose. The wort is also rich in soluble proteins, peptides and amino acids, important elements in the stability and quality

http://dx.doi.org/10.1016/j.foodres.2016.08.008 0963-9969/© 2016 Published by Elsevier Ltd.

Please cite this article as: Vivian, A.F., et al., Mass spectrometry for the characterization of brewing process, Food Research International (2016), http://dx.doi.org/10.1016/j.foodres.2016.08.008

2

A.F. Vivian et al. / Food Research International xxx (2016) xxx–xxx

of beer (Baxter & Hughes, 2001). There is no exact definition of the intervals and temperature in wort production, which is a characteristic of each beer and varies according to the desired result. Nevertheless, four temperature ranges may be set for the process, based on enzymatic activation range, as follows (Morado, 2009): • 40–45 °C: Enzymatic activation. Beta-glucanase enzymes become active. • 50–55 °C: Proteolysis. • 60–72 °C: Action of alpha and beta-amylase. • 76–78 °C: Enzymatic inactivation.

Then, it is the moment to add hop pellets and bring to boiling temperature, causing the following effects: sterilization of the wort, enzyme inactivation, water evaporation (concentration of the wort) and coagulation of proteins. This process is also important for extraction of α- and β- hops acids, responsible for beer bitterness, flavor and antiseptic action (Bernotienë, Nivinshiene, Butkienë, & Mochkute, 2004). The next stage, the fermentation process, is the phase in which sugars are converted into alcohol. This phase depends on many cellular mechanisms, being influenced by other factors such as temperature, pH, dissolved oxygen, agitation, availability of sugars and other nutrients, the presence of contaminants and inoculation rate (Bokulich & Bamforth, 2013). During the process, the simplest sugars are fermented first, using glucose, fructose, sucrose, maltose and maltotriose, in that order (Palmer, 2006). When the stationary phase of fermentation starts, the majority of yeasts are removed and then maturation starts. The remaining yeast promotes two effects: higher carbon dioxide production and chemical removal of undesirable compounds such as diacetyl, acetaldehyde and acid sulphide (Yamauchi, Okamoto, Murayama, Kajino, & Noguchi, 1995). Moreover, this phase aims to initiate beer clarification (sedimentation of the yeast cells) and maintain the beer in the reduced state, preventing oxidation during storage. During maturation, higher alcohols and fatty acids do not change significantly. However, other esters intensify, consisting mainly: ethyl acetate, isoamyl acetate, ethyl caprylate and ethyl caproate (BENTO, CARVALHO, & Silva, 2007). After the maturation, the beer must be bottled. Before this process, the remaining yeast may be removed and the beer is then cooled and may be filtered. Mass spectrometry (MS) is a technique widely used in the identification of substances, with several ionization techniques. Great improvements have been made, especially on electrospray ionization (ESI), which caused the increased sensitivity of equipment and contributed to the application of MS in a number of different products, contributing to advances in quality control (QC)(Gillespie & Winger, 2011). The progress of ESI in recent years is due to technical versatility and specificity, so it can be applied to biomolecules and biomarkers processes found in complex matrices (Riccio et al., 2010). The simplicity for sample preparation and speed in obtaining fingerprints with several chemical information increases its applicability in several areas (Quifer-Rada et al., 2015). The sensory properties of food and drinks are key characteristics for customer acceptability and QC. However, in brewing, QC has a subjective evaluation by a combination of visual, flavor and taste perceptions that can lead to misunderstandings. Considering this scenario, a fast, objective, versatile and analytical method such as ESI-MS is desirable for quality and process controls (Riccio et al., 2010). Moreover, current trends in manufacturing processes rely on the search for variables to predict and clarify performances and events. When elements are not redundant, have a well-understood connection and data are not extensive, multiple linear regression is advisable. On the other hand, the PLS-DA (Partial Least Squares- Discriminant Analysis) algorithm is appropriate when there are multi-collinear data and much information. It is valuable for categorization purposes and biomarker election, presenting the results in a visual and graphical way, e.g. score plots (Brereton & Lloyd, 2014). In combination with mass

spectrometry fingerprinting, PLS-DA becomes a powerful tool for process description and control, identifying the main compounds of any process. In this trend, the present contribution focuses at demonstrating the ability of this group of techniques in pointing out markers to follow up processes based on quality and productivity aspects, as well as evidencing different compounds of interest such as off-flavors. 2. Materials and methods 2.1. Chemical and reagents Methanol, ammonium hydroxide and formic acid solutions were purchased from J. T. Baker (Xalostoc, Mexico) and used with no further purification. Deionized water was obtained with a Milli-Q system (Millipore, USA). 2.2. Samples Samples were obtained from a craft brewery, Cervejaria Nacional, São Paulo, Brazil. The India Pale Ale (IPA) style was the study focus. Samples were withdrawn from the end of the following stages: mash, final wort, boiling, fermentation and maturation. Experiments were carried out in five replicates. The beer sample has 60 IBUs, 7.5% ABV, and is contains Pilsen, Pale Ale and Cara Munich malt, in addition to Hallertauer Magnum, Chinook, Cascade and Sincoe hops. 2.3. Sample preparation The beer samples were stored at 4 °C before analysis for up to 24 h, protected from light exposure. Samples were prepared by placing 10 μL of beer sample in a 1 mL vial with 990 μL of methanol:water (1:1). For positive analysis, 10 μL of sample solution were added to 990 μL of methanol:water (1:1) and 1 μL of formic acid. For negative analysis, the same dilution was adopted, but with the addition of 1 μL of ammonium hydroxide instead of acid. All samples were filtered before analysis. 2.4. Equipment MS fingerprints were obtained using an LTQ-XL Orbitrap Discovery instrument (Thermo Scientific, California, USA). The injection pump uses a continuous flow of 10 μL.min−1 and a 500-μL Hamilton Gastight glass syringe (Hamilton, Nevada, USA). High-resolution mass spectra were obtained after a period of 30 s, in a scan range from 50 to 2000 m/z, in both positive and negative modes. 2.5. Data management Fingerprints obtained were processed using MetaboAnalyst 3.0 and tables of m/z values as functions of intensities were extracted from the bulk spectra. In order to identify chemical markers of beer production samples, score plots were obtained after PLS-DA of normalized data in combination with VIP score analysis considering the five main steps of beer production (mash (M1), wort (M2), boiling (M3), fermentation (M4) and maturation (M5)). To evaluate if there were differences considering the overall process a previous graphical analysis of score plots was made. For VIP score analysis, data were processed considering the process sequence, in other words, comparisons were made between M1–M2, M2–M3, M3–M4 and M4–M5, and signals were considered markers only if presenting a VIP score greater than 2. No data filtering and transformation were performed apart from range scaling. Signal identification was performed using Lipid Maps, FoodDB and Metlin databases considering a mass error less than 2 ppm.

Please cite this article as: Vivian, A.F., et al., Mass spectrometry for the characterization of brewing process, Food Research International (2016), http://dx.doi.org/10.1016/j.foodres.2016.08.008

A.F. Vivian et al. / Food Research International xxx (2016) xxx–xxx

3

381

100 95 90 85 80 75 325

70 65 60 258

55

360

50 45

116

40

543

35 30 25

130 145 156

20

219

15

296 294 309

175

10

229

85

522

474

382

458

326

544

383

M1

5 0 100

200

300

400

500

377

100

600

700

800

900

1000

800

900

1000

800

900

1000

800

900

1000

m/z

95 90

439

85 341

80 161

75 70 65 60 55

221

50 179

45 40

191

35 379 383

30 25 20

387

215 15

263 101

10

89

5

143 128

281

440 470

431 425

342

339 308

259

503 601 683

M2

539

474

0 100

200

300

400

500

377

100

600

700

600

700

600

700

m/z

95 90 85 80 75 70 65 60 55 50 45 40 35 379

30 25 20 15

161 431

215

10

503

383

341

163

5

539

M3

0 100

200

300

400

500 m/z

191

100 95 90 85

383

80 75 70 65 60 55 221 50 45 40 35 161 339

30 25 179

20

259

15 195

10 133

5

545

341

263 290

333

377

384 425 419

503

M4

439

0 100

200

300

400

500 m/z

223

100 95 90 149

85 80 75 70 65 60 55

283

50 45 40

209

35 30 25 20

153

15

189

357 343

267 224

10

981

297

M5

5 0 100

200

300

400

500

600

700

800

900

1000

m/z

Fig. 1. Representative mass spectra (fingerprints) of each beer production stage (M1 - mash; M2 - wort; M3 - boiling; M4 - fermentation; M5 - maturation.). Negative ion mode. It is possible to identify anionic species as [M + Cl]− adducts of maltose (m/z 377.06) and matotriose (m/z 379.05); and [M + H]− species for maltose (m/z 341.09), maltotriose (m/z 503.12) and glucose (m/z 179.04).

Please cite this article as: Vivian, A.F., et al., Mass spectrometry for the characterization of brewing process, Food Research International (2016), http://dx.doi.org/10.1016/j.foodres.2016.08.008

4

A.F. Vivian et al. / Food Research International xxx (2016) xxx–xxx

381

100 95 90 85 80 75 70 65 60 55 50 45 40

258

35 30 25

543

325

219

133

20

156 130

15

175

10

360

309

294

M1

382

205 221

5

383

474

544

0 100

200

300

400 381

100

500

600

700

800

900

1000

900

1000

900

1000

m/z

95 90 85 80

365

75 70 175

65 60 55 50 45 40 35

325

30

360

25

527

20

258

15

130 147

10

127

203

543

382

309 294

219

383

5

M2

474

0 100

200

300

381400

100

500

600

700

800

m/z

95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20

543

219

15

175

M3

10

383

5 0 100

200

300

400

500

600

700

800

m/z

175

100

258

95 90 85 80

325

75 70 65 60 55 50 45

309

40

381

35 30

543

25 156 145 127

20 15 10 5

85 97

294 266

337

182 205 244 234

M4

471 705

544

0 100

200

300

400

500

600

700

800

900

1000

900

1000

m/z

175

100 95 90 85 80 75 70 65 60 55

258

50 45 40

325

35 30 25

309

156

20

266

10

381

294

15 116 145

182 205

5

244

268

543

337

M5

527

0 100

200

300

400

500

600

700

800

m/z

Fig. 2. Fingerprints of each stage of beer production (M1 - mash; M2 - wort; M3 - boiling; M4 - fermentation; M5 - maturation). Positive ion mode. It is possible to identify cationic species corresponding to [M + K]+ and [M + Na]+ adducts of maltose (m/z 381.10), maltotriose (m/z 543.17) and glucose (m/z 219.04).

Please cite this article as: Vivian, A.F., et al., Mass spectrometry for the characterization of brewing process, Food Research International (2016), http://dx.doi.org/10.1016/j.foodres.2016.08.008

A.F. Vivian et al. / Food Research International xxx (2016) xxx–xxx

3. Results and discussion Fingerprints obtained are presented in Figs. 1 and 2. PLS-DA results are presented in Figs. 3 and 4, for negative (neg) and positive (pos) modes, respectively. In both cases, it is possible to notice statistical differences between each stage due to the graphic separation presented in the score plots. Results after VIP score analysis and components identification are summarized in Tables 1 and 2 (Jewison et al., 2011; Preedy, 2011; Yannai, 2012) and 2(Collin, Jerkovic, Bröhan, & Callemien, 2013; Enebo, Blomgren, & Johnsson, 1955; Ge, Usack, Spirito, & Angenent, 2015; Ng, Jung, Lee, & Oh, 2012; Preedy, 2011; Yannai, 2012). Spectra presented expected signals for beer sugars as cationic species, corresponding to [M + K]+ and [M + Na]+ adducts of maltose (m/z 381.10), maltotriose (m/z 543.17) and glucose (m/z 219.04); anionic species were observed as [M + Cl]− adducts of maltose (m/z 377.06) and matotriose (m/z 379.05); and [M + H]− species for maltose (m/z 341.09), maltotriose (m/z 503.12) and glucose (m/z 179.04)(Araújo et al., 2005). All of them appear mainly in M2-Wort and M3-Boiling spectra, with a lower intensity in M4-Fermentation. Despite not being a quantitative method, this variation of intensity is coherent to what is expected from the process, since fermentation transforms sugars into alcohol, and thus its concentration tends to reduce. In addition to sugars, the statistical analysis using PLS-DA process pointed out process markers, i.e., compounds that indicate whether the transformation is occurring based on sample differences. The score plots comparing all five stages (mash (M1), wort (M2) boiling (M3), fermentation (M4) and maturation (M5)), show that there is precision

5

between replicates, since they cluster among each other. As each sample clusters are far from each other, this illustrates that there are differences between the steps and, after VIP score analysis, it is possible to find relevant markers. Different peptides appear as markers of many phases, mainly in M2 (Wort). Considering that in this phase the proteolysis of malt content happens at 50–55 °C (Morado, 2009), it is coherent with the process, although only generic molecules were observed as markers, as they also appeared in M3-boiling, M4-Fermentation and M5-Maturation phases. Despite being generic, amino acids and peptides play important roles in beer quality and brewing process, especially as foam stabilizing agents, and flavor and haze development agents (combined with polyphenols) in beer. Wort amino acids are the major source of nitrogen for fermentation yeast, and are consumed according to their proprieties. Proline, for example, is the last to be consumed, and its presence in final beer is directly related to product stability (haze) and, due to its reaction with maltose (through the Maillard reaction mechanism), it may affect directly the color development in beer. Moreover, there is a direct correlation among each amino acid, fermentation temperature and the esters and higher alcohols formed, molecules that are essential to beer flavor development (Preedy, 2011). The undesirable taste of vicinal diketones, for instance, is result of valine and isoleucine intermediates biosynthesis during fermentation process. One method to control diacetyl production suggested is valine feedback inhibition control (Krogerus & Gibson, 2013). Beer phenols also contribute directly to beer flavor, color and haze, and may be found in hops, barley and malt. During mashing, malt flavonoids are dissolved in the wort, but after filtration, they tend to

Fig. 3. PLS-DA plots from negative mode spectra illustrate that each stage is different from each other and the replicates are precise between them. M1 - mash; M2 - wort; M3 - boiling; M4 fermentation; M5 - maturation.

Please cite this article as: Vivian, A.F., et al., Mass spectrometry for the characterization of brewing process, Food Research International (2016), http://dx.doi.org/10.1016/j.foodres.2016.08.008

6

A.F. Vivian et al. / Food Research International xxx (2016) xxx–xxx

Fig. 4. PLS-DA plots from positive mode spectra illustrate that each stage is different from each other and the replicates are precise between them. M1 - mash; M2 - wort; M3 - boiling; M4 fermentation; M5 - maturation.

disappear through binding to coagulated proteins, for example. Thus, (+)-catechin found in M2-Wort is a marker that is originated from malt grain, and it is the main monomeric unit found, with an important role since it can induce haze upon storage (Collin et al., 2013). In other studies, (−)-epicatechin, (−)-catechin gallate and (−)-epicatechin gallate and epigallocatechin were also detected (Yannai, 2012). Epigallocatechin metabolites were found as markers in M4-Fermentation phase, and it is suggested to be result of S. cerevisae metabolism and

action of its many O-methyltransferase (Niewmierzycka & Clarke, 1999). Other compounds that are important to beer flavor were found in hops and appear as markers of M3-Boiling step, such as humulone and flavone, expressive components of hop pellets (Bernotienë et al., 2004). Cohulupone, one marker of M5-Maturation, is minor hop constituent and an oxidation product of colupulone. During fermentation and maturation, it does not undergo any other transformation, possibly explaining why it appears as a marker in the final beer. It is responsible

Table 1 Markers found after PLS-DA and VIP score analysis- Negative ion mode. Stage (spectra)

Ion

m/z (δPPM)

[M +

503.1229 Peptide Cys-Leu-Met-Cys (0) 575.1398 Peptide Asp-Asp-Glu-Tyr (2)

Cl]− [M + Boiling (M3)

Cl]− [M + H]− [M +

H]− Fermentation [M + (M4) H]− [M + H]− Fermentation [M + (M4) H]− [M + H]−

Compound

Compound description

Remaining peptides from wort proteolysis. Yeast will consume according to amino acid proprieties (Preedy, 2011).

472.1725 Peptide Tyr-Glu-Tyr (2) 221.0683 Flavone (1)

Crystalline compound present in hops (Yannai, 2012)

545.1339 Peptide Asp-His-Cys-His (0)

Remaining peptides from wort proteolysis and not used in yeast metabolism.

443.1064 Peptide Cys-Cys-Gly-Tyr (0) 425.0946 Epigallocatechin-3-O-(4-hydroxybenzoate) Epigallocatechins are isolated from barley grains, hops and final beer (Yannai, 2012). It is a (2) metabolite of S. cerevisae (Jewison et al., 2011), by means of O-methyltransferase enzyme. 485.1147 Epigallocatechin (3) 3-O-(3,5-di-O-methylgallate)

Please cite this article as: Vivian, A.F., et al., Mass spectrometry for the characterization of brewing process, Food Research International (2016), http://dx.doi.org/10.1016/j.foodres.2016.08.008

A.F. Vivian et al. / Food Research International xxx (2016) xxx–xxx

7

Table 2 Markers found after PLS-DA and VIP score analysis- Positive ion mode. Stage (spectra)

Ion

m/z (δPPM)

Compound

[M +

366,1330 (1) 528,1945 (0)

Peptide Met-Thr-Asp Peptide Met-Phe-Gln-Cys

529,1963 (0)

Peptide Ser-Glu-Met-Tyr

528,1959 (1)

Peptide Asn-Arg-Asn-Cys

291.0795 (2)

(+)- Catechin

Phenol found in large amount in hops, can be found in small concentration in malt grain (10–100 ppm) and is responsible for astringency mouth feel (Collin et al., 2013).

133.0424 (2)

Glutaric Acid

Non-volatile organic acid (Enebo et al., 1955).

221,0807 (2)

Humulone

Main component of hops (Yannai, 2012).

381.1075 (1) 382.1101 (1)

Peptide Gly-Asp-Cys-Ser Peptide Cys-Glu-Met

366.1330 (1)

Peptide Met-Glu-Ser

435.1640 (0)

Peptide Pro-Pro-Pro-Ser

H]+ [M + H]+ [M + Wort (M2)

H]+ [M + H]+ [M + H]+ [M + H]+ [M + H]+ [M +

Boiling (M3)

H]+ [M + H]+ [M + H]+ [M + K +] [M + H]+ [M +

Maturation (M5)

H]+ [M + H]+ [M + H]+

117,0829 (1) 252.1201 (1)

Compound description

Peptides resulting from proteolysis.

Remaining peptides from wort. Yeast will consume according to amino acid proprieties (Preedy, 2011).

Caproic Acid

Result of yeast activity (Ge et al., 2015).

2,3- butanediol glucoside

Result of S. cerevisae metabolism (Ng et al., 2012)

318.1487 (2)

Peptide Cys-Val-Pro

Peptide remaining from raw materials.

319, 1825 (2)

Cohulupone

Cohulupone is a constituent of hops

* Arg: Arginine; Asn: Asparagine; Asp: Aspartic Acid; Cys: Cysteine; Met: Methionine; Gln: Glutamine; Glu: Glutamic acid; Gly: Glycine; His: Histidine; Phe: Phenylalanine; Ser: Serine; Thr: Threonine; Tyr: Tyrosine;

for the pleasant bitterness in beer, although it is less potent than isohumulone (Laws, 1968). Caproic acid, also direct linked to beer quality and sensorial aspects, was detected as maker in M5-Maturation. Depending on its concentration, it may be considered either an off-flavor or a desirable component in beer (Horák, Čulík, Jurková, Čejka, & Kellner, 2008). Detecting its quantity during maturation is a way to predict if the final product will be according to what is expected or not. Another important compound in maturation that is related to quality is 2,3-butanediol, result of conversion of diacetyl in acetoin through diacetyl reductase, then reduction of acetoin by butanediol dehydrogenase (Ng et al., 2012). Diacetyl is an undesirable off-flavor, and it is important to be converted to 2,3butanediol during fermentation and maturation, as the latter has a larger threshold detection. Other markers found were non-volatile organic acids, compounds present in wort and in larger concentration in the final beer, due to fermentation metabolism, and some more expressively in top-fermenting beers (Ales). One example is glutaric acid, which is the content in final beer that characterizes the fermentation procedure, specific from Ales (Enebo et al., 1955). Off-flavor signs are observed during the process. Despite not being considered a process change marker, they are remarkable to final quality aspects. The importance of flavor and its components is not only to give the final quality characteristics to beer, but is also a strong indicator that the process was successful (Morado, 2009). The anionic species [M + H−] found are p-methane-8-thiol (M1; m/z 185.1076), a precursor of catty urine flavor and ethyl mystrate (M3, m/z 255.2396) – an ester with vegetable oil flavor. It is important to highlight that they appear in early stages (M1 and M3) which allow to previous evaluate if the process is happening according to expected, without the necessity of waiting to taste the final product.

4. Conclusion Direct injection ESI-HRMS demonstrated to be suitable for a rapid, simple and robust fingerprinting approach of complex mixtures such as beer, as well as for pointing out the main chemical players in process. The technique is convenient for process monitoring, since it requires small sample size, simple handling and fast injection in a mass spectrometer. This is an attractive alternative for the control of food and beverages, particularly when associated with qualitative multivariate statistics. PLS-DA shows clear differences between the phases in the overall process and VIP score analysis allowed to identify the main markers of the process, malt and hop components, in addition to substances resulting from mashing, fermentation and maturation processes, such as humulone, glutaric acid, 2,3-butanediol and caproic acid. Besides markers, MS is capable of evidencing other important substances such as off-flavors precursors. Our approach is able to evidence whether the process is happening according to expected before it is finished, without subjective sensorial analysis, thereby suggesting a new area for mass spectrometry application, in a zoom-in ‘processomics’ approach to better understand any process, and thus control and improve it. The main contributions for developing further applications is the possibility of detection of off-flavors before obtaining the final product, prevention of higher production costs and assuring final quality, monitoring contamination and controlling desirable characteristics, such as presence of esters and bitter compounds, foam stability and, ultimately, drinkability. Author contributions A.F.V.: designed and conducted the mass spectrometry experiments, analyzed the results, wrote the paper and prepared the figures. C.T.A.:

Please cite this article as: Vivian, A.F., et al., Mass spectrometry for the characterization of brewing process, Food Research International (2016), http://dx.doi.org/10.1016/j.foodres.2016.08.008

8

A.F. Vivian et al. / Food Research International xxx (2016) xxx–xxx

conducted the mass spectrometry experiments and analyzed the results. D.N.O. and R.R.C.: idealized all experiments, analyzed the results and revised the paper. Acknowledgements Cervejaria Nacional supported this study. All authors have read and approved the final manuscript. We also thank FAPESP for the financial support that allowed this work to be performed (Grants # 2011/ 50400-4, 2014/00084-2 and 2015/06809-1). References Araújo, A. S., Da Rocha, L. L., Tomazela, D. M., Sawaya, A. C., Almeida, R. R., Catharino, R. R., & Eberlin, M. N. (2005). Electrospray ionization mass spectrometry fingerprinting of beer. Analyst, 130(6), 884–889. Baxter, E. D., & Hughes, P. S. (2001). Beer: Quality, safety and nutritional aspects. Royal Society of Chemistry. Beltramelli, M. (2013). Cervejas: Brejas & Birras. Cidade São Paulo: Editora Leya, Editora Tainã Bispo. Bento, C., Carvalho, G., & Silva, J. (2007). Elementos biotecnológicos fundamentais no processo cervejeiro: 3° parte–A maturação. São Paulo: Revista Analytica, 27. Bernotienë, G., Nivinshiene, O., Butkienë, R., & Mochkute, D. (2004). Chemical composition of essential oils of hops (Humulus lupulus L.) growing wild in Aukstaitija. Chemija, 15(2), 31–36. Bokulich, N. A., & Bamforth, C. W. (2013). The microbiology of malting and brewing. Microbiology and Molecular Biology Reviews, 77(2), 157–172. Brereton, R. G., & Lloyd, G. R. (2014). Partial least squares discriminant analysis: Taking the magic away. Journal of Chemometrics, 28(4), 213–225. Collin, S., Jerkovic, V., Bröhan, M., & Callemien, D. (2013). Polyphenols and beer quality. Natural products (pp. 2333–2359). Springer. Enebo, L., Blomgren, G., & Johnsson, E. (1955). Low molecular non-volatile organic acids in wort and beer. Journal of the Institute of Brewing, 61(5), 408–411. Fillaudeau, L., Blanpain-Avet, P., & Daufin, G. (2006). Water, wastewater and waste management in brewing industries. Journal of Cleaner Production, 14(5), 463–471. Ge, S., Usack, J. G., Spirito, C. M., & Angenent, L. T. (2015). Long-term n-caproic acid production from yeast-fermentation beer in an anaerobic bioreactor with continuous product extraction. Environmental Science & Technology, 49(13), 8012–8021.

Gillespie, T. A., & Winger, B. E. (2011). Mass spectrometry for small molecule pharmaceutical product development: A review. Mass Spectrometry Reviews, 30(3), 479–490. Horák, T., Čulík, J., Jurková, M., Čejka, P., & Kellner, V. (2008). Determination of free medium-chain fatty acids in beer by stir bar sorptive extraction. Journal of Chromatography A, 1196, 96–99. Ivanova, B., & Spiteller, M. (2014). Solid-state determination of hop bitter acids in beer by UV–MALDI–Orbitrap mass spectrometry. Journal of Food Measurement and Characterization, 8(4), 343–355. Jewison, T., Knox, C., Neveu, V., Djoumbou, Y., Guo, A. C., Lee, J., ... Sinelnikov, I. (2011). YMDB: The yeast metabolome database. Nucleic Acids Research (gkr916). Krogerus, K., & Gibson, B. R. (2013). Influence of valine and other amino acids on total diacetyl and 2, 3-pentanedione levels during fermentation of brewer's wort. Applied Microbiology and Biotechnology, 97(15), 6919–6930. Laws, D. (1968). Hop resins and beer flavour V. The significance of oxidized hop resins in brewing. Journal of the Institute of Brewing, 74(2), 178–182. Morado, R. (2009). Larousse de cerveja: Larousse do Brasil. Ng, C. Y., Jung, M. -y., Lee, J., & Oh, M. -K. (2012). Production of 2, 3-butanediol in Saccharomyces cerevisiae by in silico aided metabolic engineering. Microbial Cell Factories, 11(1), 1. Niewmierzycka, A., & Clarke, S. (1999). S-Adenosylmethionine-dependent methylation in Saccharomyces cerevisiae identification of a novel protein arginine methyltransferase. Journal of Biological Chemistry, 274(2), 814–824. Palmer, J. J. (2006). How to brew. EUA: Brewers Publications. Preedy, V. R. (2011). Beer in health and disease prevention. Academic Press. Quifer-Rada, P., Vallverdú-Queralt, A., Martínez-Huélamo, M., Chiva-Blanch, G., Jáuregui, O., Estruch, R., & Lamuela-Raventós, R. (2015). A comprehensive characterisation of beer polyphenols by high resolution mass spectrometry (LC–ESI-LTQ-Orbitrap-MS). Food Chemistry, 169, 336–343. Riccio, M. F., Saraiva, S. A., Marques, L. A., Alberici, R., Haddad, R., Moller, J. C., ... Catharino, R. R. (2010). Easy mass spectrometry for metabolomics and quality control of vegetable and animal fats. European Journal of Lipid Science and Technology, 112(4), 434–438. Sleiman, M., & Venturini Filho, W. (2008). Relação entre percentual de malte e preço em cervejas tipo pilsen. Revista Energia na Agricultura, 23(1), 98–108. Yamauchi, Y., Okamoto, T., Murayama, H., Kajino, K., & Noguchi, K. (1995). Rapid maturation of beer using an immobilized yeast bioreactor. 2. Balance of total diacetyl reduction and regeneration. Journal of Biotechnology, 38(2), 109–116. Yannai, S. (2012). Dictionary of food compounds with CD-ROM. CRC Press.

Please cite this article as: Vivian, A.F., et al., Mass spectrometry for the characterization of brewing process, Food Research International (2016), http://dx.doi.org/10.1016/j.foodres.2016.08.008