International Journal of Food Microbiology 215 (2015) 71–78
Contents lists available at ScienceDirect
International Journal of Food Microbiology
journal homepage: www.elsevier.com/locate/ijfoodmicro
The fungal community structure of barley malts from diverse
geographical regions correlates with malt quality parameters
Mandeep Kaur a,⁎, John P. Bowman a, Doug C. Stewart b, David E. Evans c
a
b
c
Tasmanian Institute of Agriculture, University of TAS, Hobart, Australia
Cargill, Adelaide, Australia
Tassie Beer Dr, Lindisfarne, TAS, Australia
a r t i c l e
i n f o
Article history:
Received 12 March 2015
Received in revised form 3 August 2015
Accepted 24 August 2015
Available online 28 August 2015
Keywords:
Barley malt
TRFLP
Microbial community structure
Geographical location
a b s t r a c t
Malt is a preferred base for fermentations that produce beer or whisky. Barley for malt is grown under diverse
environments in different geographical locations. Malt provides an ecological niche for a varied range of microorganisms with both positive and negative effects on its quality for brewing. Little information exists in the literature on the microbial community structure of Australian malt as well as broader global geographical differences
in the associated fungal and bacterial communities. The aims of the present study were to compare the bacterial
and fungal community structures of Australian commercial malt with its international counterparts originating
from different geographical regions using terminal restriction fragment length polymorphism (TRFLP) fingerprinting and clone library analyses of ribosomal RNA genes. Further, the relationship between malt associated
microbial communities and conventional malt quality parameters was also compared. Results showed that
differences in fungal communities of malts from different geographical location were more pronounced than bacterial communities. TRFLP analysis discriminated high quality commercial malts with low fungal loads from
malts deliberately infected with fungal inocula (Fusarium/Penicillium). Malt moisture, beta-amylase, α-amylase
and limit dextrinase contents showed significant correlations with fungal community structure. This investigation concluded that fungal community structure was more important to subsequent malt quality outcomes
than bacteria.
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
Barley malt is to beer as grapes are to wine (Goldammer, 2008). The
most extensive use for barley malt worldwide is as a source of fermentable sugars for alcoholic fermentations, primarily beer but also for whiskey production. Approximately 10% of the world barley crop is used,
after malting, for the production of beer. Malt forms the base material
for making wort, the liquid extract that is fermented into beer. Different
malt types are used to generate different characteristics in beer products
including flavor, color and mouthfeel (Bamforth and Barclay, 1993).
Barley for malt is grown in a diverse range of environments and geographic locations. These include sub arctic Scandinavia to near the equator, in the mountains of Ethiopia and in South America, from below sea
level near the Dead Sea to high altitudes in the Andes and the
Himalayas, from humid, temperate regions, such as western Europe to
dry land areas in parts of North America (N. America), Africa, and
Australia (Briggs, 1978; Hunter, 1962; Rasmusson, 1985). In Australia
Spring, two-rowed, barley is grown as a “winter” crop in semi arid, temperate and intermediate climatic regions resulting in dry maturation
⁎ Corresponding author at: Tasmanian Institute of Agriculture, PO Box 98, University of
Tasmania, Hobart, TAS 7001, Australia.
E-mail address: [email protected] (M. Kaur).
http://dx.doi.org/10.1016/j.ijfoodmicro.2015.08.019
0168-1605/© 2015 Elsevier B.V. All rights reserved.
and harvest conditions. These conditions usually result in dry barley
(b 13% moisture) for storage and for subsequent malting. Such conditions maintain the germinative vigor of the barley that is a prerequisite
for malting, and inhibit the growth of microbes during storage. Such advantages contribute to Australia's supply of around 32% of the world export malting barley trade (http://www.e-malt.com/). Australian grown
barley has a reputation for being “bright and clean”, which is suggestive
of a low microbial load and rarity of mycotoxins such as deoxynivalenol
and ochratoxin A (Kaur et al., 2009).
Barley grains, covered by a fibrous husk, are normally colonized by a
wide variety of bacteria, yeasts and filamentous fungi (Flannigan, 2003).
These mixed populations are difficult to control and elimination is neither possible nor desirable (Laitila, 2008) in a practical sense. Grain associated microbes have both positive and negative effects on grain
quality in the field, in storage, at various stages during the malting process, and on the quality of the resulting malt and beer (Flannigan, 2003;
Justé et al., 2011; Noots et al., 1999).
According to Flannigan (2003), barley provides an ecological niche
for a diverse range of microorganisms, but the microbiota of different
barleys is remarkably similar to each other, and to other cereals. Barley
microbiota at harvest has been found to comprise the same limited
number of species. Studies on fungi associated with South African
(S. African) barley malt reported that predominant species in
72
M. Kaur et al. / International Journal of Food Microbiology 215 (2015) 71–78
S. African malt were the same as those found elsewhere in the
world; however the counts of these fungal taxa especially Fusarium
and Penicillium species were significantly lower than those reported in
the Northern Hemisphere (Rabie and Lübben, 1993). This is not surprising, as the S. African barley growing environment is in many ways
similar to that of Australia, in that the grain maturation and harvest conditions are generally dry. The microbial community structure of barley
malt can be influenced by other factors including growing location
(Birgitte et al., 1996), climatic conditions (Backhouse and Burgess,
2002; Doohan et al., 2003; Krstanović et al., 2005), malting techniques
(Flannigan et al., 1982), and storage and handling environments (Hill
and Lacey, 1983; Laitila et al., 2003). In addition, the detection and
enumeration techniques used for analysis may yield results that lead
to different conclusions (Jarvis and Williams, 1987; Rabie et al., 1997).
That is the microbial culture conditions may bias the composition determined because some microbes are unculturable or require very specific
culture conditions.
A survey of the literature showed that there is a growing body of
data for malt associated microbiota derived from N. American,
European and S. African locations. There is no little information in this
aspect for Australian malt or the broader global geographical differences
in malt associated fungal and bacterial communities.
Most microbial studies of barley or malt samples have been done
using conventional culture-dependent methods, comparing quantitative
changes in microbial communities. These conventional microbial
culture-dependent methods are biased towards the selective enrichment
of fast growing microorganisms adapted to high substrate concentrations that could potentially represent only a minor fraction of the resident microbial community. On the other hand, culture-independent
methods are now commonly employed to assess microbial community
diversity and dynamics in food based ecosystems (Bokulich and Mills,
2012). PCR-DGGE (denaturing gradient gel electrophoresis) was used
to monitor bacterial community dynamics during the malting process
in Finland (Laitila et al., 2007).
The objectives of the present study were to compare the bacterial
and fungal community structures of commercial Australian barley
malt with its international counterparts originating from different geographical regions using terminal restriction fragment length polymorphism (TRFLP) fingerprinting and clone library analysis of ribosomal
RNA genes. TRFLP analysis is a method developed for rapid profiling of
complex microbial populations. Being a high throughput fingerprinting
technique, TRFLP analysis has been applied extensively to the analysis of
fungal and bacterial communities (Schütte et al., 2008). While TRFLP
shares problems inherent to any PCR-based method (Lueders and
Friedrich, 2003; Qiu et al., 2001), it has been shown to provide a facile
means to observe changes in microbial community structure on temporal and spatial scales by monitoring the gain or loss of specific fragments
from the profiles. When coupled with rRNA gene clone library assessment and sequencing, additional specific information on the composition of microbial communities can be obtained. The final objective was
to establish a relationship between malt associated microbial communities and routine physicochemical malt quality parameters used in
malting and brewing industry.
2. Materials and methods
2.1. Sample collection and preparation
A total of 34 Australian commercial malt samples were collected
from different malt houses representing malt produced from barley
grown in different cropping zones and included different commercial
varieties (Baudin, Gairdner, Grimmett, Schooner and Sloop). International commercial malts were investigated with the sample numbers
and source countries being shown in Table 1. The samples from
Finland were malted from barley artificially inoculated with either a
Fusarium sp. or a Penicillium sp. to produce standard malt samples
Table 1
Detail of barley malt samples used in this study.
Country of origin
No. of samples
Argentina
Australia
Belgium
Denmark
Finland
France
North America
(N. America⁎)
Russia
Slovakia
South Africa (S. Africa⁎)
5
34
2
3
4 (two standard DON and two standard OTA samples)
6
3 (includes one standard Fusarium head
blight infected malt)
7
3
7 (includes one standard gushing malt)
⁎ Abbreviated and used there on.
having specified concentrations of deoxynivalenol (DON: 12 and
32 mg/kg) and ochratoxin A (OTA: 126 and 1099 μg/kg) mycotoxins
for routine laboratory mycotoxin studies. One malt sample (out of
7) from S. Africa had gushing properties and was prepared by artificially
inoculating the malt with Fusarium culmorum. Additionally, one sample
from N. America malts was produced from barley that was known to be
infected with Fusarium head blight. These samples had been analyzed
for these characteristics and information was provided by the sample
providers. The reason for including these known fungal infected malt
samples in the study was to examine the discriminative ability of
TRFLP technique and further statistical analyses applied in this study.
Collected samples were stored in airtight bags at room temperature before grinding. Samples (40 g) were ground in a Cyclone Sample Mill
using a 0.1 mm screen (UDY Corporation, CO, USA) and stored immediately at −20 °C until used for DNA extraction. Cross contamination between samples was avoided by blowing high pressure dry air through
the grinding mill and collection container in between the samples, and
taking only the middle portion of the ground sample from the container
for analysis.
2.2. DNA extraction and rRNA gene PCR amplification
DNA was extracted from ground samples (0.1 g) in duplicate using
the FastDNA® Spin Kit for Soil (Q Biogene, CA, USA) according to the
manufacturer's instructions except that the samples were homogenized
with a Retsch MM300 bead beater (Retsch GmbH, Haan, Germany) at
30/s frequency for 4 min. DNA samples were independently amplified
using 16S rRNA gene primers 27F (5′-AGAGTTTGATCMTGGCTCAG-3′)
labeled with D3 WellRED fluorescent dye (Sigma-Aldrich) and primer
1492R (5′-TACGGYTACCTTGTTACGACTT-3′) labeled with WellRED™
fluorescent dye D4 for bacteria and 28S rRNA gene primers NL1 (5′GCATATCAATAAGCGGAGGAAAAG-3′) labeled with D3 WellRED
fluorescent dye and primer NL4 (5′-GGTCCGTGTTTCAAGACGG-3′) labeled with D4 WellRED fluorescent dye for fungi (Beckman Coulter,
Australia Pty Ltd., NSW, Australia). Each 60 μl reaction mixture
contained 30 μl of 2 × ImmoMix Red™, 22.5 μl of sterile water (Bioline,
NSW, Australia), 3 μl of each forward and reverse primers (10 pmol) and
1.5 μl DNA template. The PCR amplification program was as follows; 95
°C, 10 min; 30 cycles (35 cycles for fungi) of 94 °C, 1 min; 55 °C, 1 min;
72 °C, 2 min with a final extension of 72 °C, 10 min in a PTC 200 Peltier
Thermal Cycler (MJ Research, Waltham, USA). After checking amplicons
using agarose electrophoresis PCR products (55 μl) were purified using
the UltraClean™ PCR Clean-Up Kit (MoBio Laboratories, Inc., Carlsbad,
CA, USA) according to the manufacturer's instructions.
2.3. TRFLP analysis of bacterial and fungal communities
Aliquots of purified PCR products were digested individually with
HaeIII, MspI and RsaI (for bacteria) and HaeIII, HinfI and RsaI (for
fungi) (New England Biolabs Inc., Ipswich, MA, USA) according to the
M. Kaur et al. / International Journal of Food Microbiology 215 (2015) 71–78
manufacturer's instructions. Digested PCR samples were then desalted
and purified using a 96 well ethanol precipitation method (Beckman
Coulter CEQ8000 Handbook). Digested products were mixed with
30 μl sample loading solution and 0.25 μl of 600 bp DNA size standard
(Beckman Coulter) and analyzed using the Beckman Coulter CEQ8000
Genetic Analysis System utilizing the modified Frag-4 method. According to this method samples were injected at 2.0 kV for 30 s, run for
90 min at capillary temperature of 50 °C at 4.8 kV. The TRFs were analyzed using the Beckman Coulter CEQ8000 Genetic Analysis System
software (version 8.0). Duplicate TRF profiles were generated for each
sample. Further, based on relative area of peaks (relative abundance)
whereby a threshold was set to 5%, only TRFs present with a size
between 60 and 640 bp regions were used for further analysis.
Duplicate profiles for each sample were analyzed with TAlign (Smith
et al., 2005) where fragments were binned with a 1.0 base confidence
interval. Fragments not present in duplicate samples were removed to
generate a consensus profile containing only TRFs that occurred in
both replicate profiles. The results of all enzyme digests for each sample
(separate for bacteria and fungi) were pooled and analyzed with Primer
v6 software (Primer-E Ltd., Plymouth Marine laboratory, UK). A similarity matrix of relative abundance data was calculated utilizing the Bray–
Curtis coefficient (Bray and Curtis, 1957). This resemblance matrix was
then tested with permutational multivariate analysis of variance
(PERMANOVA) (Anderson and Gribble, 1998) using Primer v6 to test
for differences in malt sample microbial community structure with bacterial and fungal TRF profiles tested separately and when combined. Differences between malts from different countries were considered
significant at P b 0.01 (using 9999 permutations). Canonical analyses
of principal coordinate (CAP) plots (Anderson and Willis, 2003) were
used to build visual representations of microbial community differences
between the sample groups, plotted along simple axes that are best at
discriminating among groups. Pearson's correlation was used to determine TRFs (both bacterial and fungal) that correlated with the CAP
axes for country of origin of malt samples used to constrain the data.
Further, the bacterial and fungal combined TRF data, excluding six
standard fungi inoculated/infected malt samples mentioned previously
were employed to determine species richness, evenness and diversity
using Margalef (d), Pielou's evenness (J'), Shannon (H') and Simpson's
diversity (1-λ′) indices, using routines in Primer v6. Analysis of variance
on these indices was performed in SAS statistical software, version 9.1
(SAS Institute Inc., Cary, NC). Similarity percentage (SIMPER) analysis
was used to calculate average Bray–Curtis similarity within individual
groups (countries) and differences among groups (countries). Multivariate dispersion (MVDISP) indices were also calculated to examine within group (country of origin) heterogeneity. Sample sets with large
dispersion index values possess high sample to sample variability in
comparison to other sample sets (Clarke and Warwick, 2001; Rees
et al., 2004).
A subset of malts (number of samples from Argentina, Belgium,
Denmark, France, Russia and Slovakia, as shown in Table 1) of known
cultivars with analogous properties was provided by a single international brewer group, ensuring comparability with regard to malting
conditions that was further analyzed for aforementioned microbial
analyses and routine malt quality parameters. The common measures
of malt quality analysis, provided by the supplier, were carried out according to Analytica — EBC (1998) recommended methods; moisture
(method 4.2), fine grind extract (method 4.5.1), and wort color was
measured using a photometer (Lovibond EBC Colorpod, Tintometer
Ltd., Wiltshire, UK), Kolbach Index (KI, method 4.9.1), α−amylase
(Evans, 2008), beta-amylase (Evans, 2008), limit dextrinase (LD,
Evans, 2008) and total protein content (method 4.3.1). The relationship
of these malt quality parameters with fungal TRFs was investigated
using distance-based linear model (DISTLM) and distance-based redundancy analysis (dbRDA) (Anderson and Gribble, 1998). For DISTLM the
forward selection option and 9999 permutations were used to identify
variables that were significantly correlated with the observed fungal
73
community structure. The vectors on the dbRDA ordination diagram
show strength and direction of the relationship between individual
vector (routine malt quality parameter) and construction of the
constrained ordination plot.
2.4. Clone libraries construction, sequencing and tentative assigning of TRFs
to different microbial taxa
Fungal and bacterial clone libraries were generated from DNA extracted from different malt samples for which the differences were
significant with regard to bacterial and fungal community structures
based on TRFLP data analysis. Bacterial 16S rRNA gene and fungal 28S
rRNA gene PCR products (using aforementioned primers and PCR
conditions) were cloned in Escherichia coli using TOPO TA Cloning®
Kit (Invitrogen, Carlsbad, USA), following the manufacturer's instructions. The correct insert size in each clone was checked by vector targeted PCR (primer M13F and M13R) and agarose gel
electrophoresis. The PCR products were purified using UltraClean™
PCR Clean-Up Kit (MoBio Laboratories, Inc., Carlsbad, CA, USA). Thirty two clones per sample were randomly selected, sent to and sequenced for this work by the Macrogen, Korea with the BigDye
Terminator Ready Reaction mix sequencing reaction kit using the
vector specific T3 or T7 promoters as primers; with sequence reactions ran on an automated DNA sequencer (Applied BioSystems).
The sequences were edited using BioEdit v7.0.9.0 (Hall, 1999) and
compared to GenBank database using standard BLASTN comparisons. Non-redundant sequences have been deposited in the GenBank
database under the accession numbers JX005948–JX006033, and
HQ143267. These sequences were virtually digested using the same
restriction enzymes used to cleave the PCR products from malt
DNA. The lengths of these theoretical TRFs were calculated and fragments were tentatively assigned to different bacterial/fungal taxa.
3. Results
3.1. Similarity analysis of TRFs
In order to test for the impact of malt sample origin on community
structural variation, the fungal and bacterial TRFLP profiles generated
for each sample were compared by the PERMANOVA approach. The
main PERMANOVA results showed significant differences in malt associated bacterial (P = 0.003), fungal (P = 0.0001) communities and
when both bacterial and fungal data were combined (P = 0.0001)
(Table 2). Australian malts were significantly different from malts produced in Argentina, Finland, France, Russia and S. Africa in terms of fungal community profiles. Similar results were observed from CAP plot
(Fig. 1). The plot showed distinct grouping of malts by geographic location. The first canonical axis separated Australian malts from Finnish
and Russian malts, whereas, axis 2 separated these from Argentina,
France and S. Africa. Fungal communities of Finnish malts were also different from Argentinean, French and Russian malts. The CAP plot (Fig. 1)
also showed distinct grouping of inoculated Finnish malts (inoculated
for DON and OTA mycotoxins), from most other countries except for
the standard gushing malt sample from S. Africa, two N. American
malts including the one that was Fusarium head blight infected, and
Russian malts. The axis 2 separated Finnish (inoculated malts) and
Russian malts. Significant differences with regard to fungal community
were also observed between samples from Denmark–Russia, France–
Argentina, France–Russia, France–Slovakia, N. America–Russia, Russia–
Argentina, Russia–Slovakia, S. Africa–Argentina, S. Africa–France and S.
Africa–Russia. The only significant difference in malt bacterial community structure was observed between samples from Finland (inoculated
malts) and France (P = 0.0046). CAP analysis also confirmed these results (data not shown).
To further resolve which bacterial and fungal TRFs contributed to the
differences observed, Pearson's correlations between the variables
74
M. Kaur et al. / International Journal of Food Microbiology 215 (2015) 71–78
Table 2
PERMANOVA P values for comparison of all the barley malt samples for microbial TRFLP
peak data. Significant P values are in bold (P b 0.01).
Fungi
Bacteria
Combined
Main test
0.0001
0.003
0.0001
Key pair wise comparisons
Argentina–Slovakia
Australia–Argentina
Australia–Belgium
Australia–Denmark
Australia–Finland
Australia–France
Australia–N. America
Australia–Russia
Australia–S. Africa
Australia–Slovakia
Belgium–Argentina
Belgium–Russia
Belgium–Slovakia
Denmark–Argentina
Denmark–Belgium
Denmark–Russia
Denmark–Slovakia
Finland–Argentina
Finland–Belgium
Finland–Denmark
Finland–France
Finland–N. America
Finland–Russia
Finland, S. Africa
Finland–Slovakia
France–Argentina
France–Belgium
France–Denmark
France–Russia
France–Slovakia
N. America–Argentina
N. America–Belgium
N. America–Denmark
N. America–France
N. America–Russia
N. America–S. Africa
N. America–Slovakia
Russia–Argentina
Russia–Slovakia
S. Africa–Argentina
S. Africa–Belgium
S. Africa–Denmark
S. Africa–France
S. Africa–Russia
S. Africa–Slovakia
0.0194
0.0031
0.0604
0.0137
0.0001
0.0001
0.0145
0.0001
0.0005
0.0252
0.0508
0.0294
0.1011
0.0189
0.1022
0.0086
0.0957
0.0087
0.1321
0.0283
0.005
0.2029
0.0038
0.0687
0.0303
0.0036
0.1109
0.1419
0.0004
0.009
0.0186
0.3004
0.1016
0.0118
0.0082
0.094
0.2943
0.0011
0.009
0.0014
0.2503
0.0332
0.0007
0.0005
0.079
0.3321
0.4984
0.3758
0.046
0.0155
0.135
0.1411
0.1129
0.1108
0.1343
0.7129
0.0523
0.0956
0.1259
0.1034
0.0576
0.0959
0.022
0.0691
0.0258
0.0046
0.1735
0.0153
0.3899
0.0572
0.5226
0.2868
0.0239
0.1729
0.0116
0.2692
0.0954
0.1047
0.0224
0.057
0.2788
0.1059
0.2754
0.056
0.7297
0.3681
0.6083
0.0963
0.5606
0.4219
0.032
0.0249
0.1354
0.0175
0.0001
0.0019
0.0283
0.0001
0.0064
0.0453
0.1375
0.0263
0.1011
0.0376
0.0988
0.0084
0.0999
0.0076
0.1394
0.0288
0.004
0.2204
0.0033
0.1259
0.0281
0.0097
0.1406
0.0561
0.0008
0.0114
0.0368
0.1006
0.1007
0.0131
0.0161
0.1073
0.0995
0.0018
0.0083
0.0142
0.3879
0.2469
0.0016
0.0007
0.2037
(TRFs) and the first four CAP axes were determined. The fungal TRFs
with N 0.55 correlation are shown as vectors on the CAP plot (Fig. 1).
The fungal TRFs 148 HaeR, 203 HinfF, 229 RsaR, 210 HinfF, 144 HinfR,
459 HaeF and 156 HaeR were more positively correlated with samples
from France, Belgium and Denmark and to some extent with samples
from Argentina. Fragments 96 HinfF, 97 HaeR, 154 HaeR, 215 HinfF,
311RsaF, 220 HinfF, 382 HinfR, 317 HaeR, 95 HaeF, 435 HaeF, and 113
RsaR were correlated with Finland (DON and OTA inoculated malt
samples), one malt sample from S. Africa (standard gushing malt),
two N. American malts (one being Fusarium head blight infected) and
Russian malts (only axes 1 and 2 shown in Fig. 1). The TRFs 100 HaeR,
135 HaeR, 112 HaeF and 200 RsaR were positively correlated with
Australian malt samples (axes 1, 2 and 4). Bacterial 125 RsaF, 126 RsaR
and 123 RsaF TRFs were positively correlated with samples from
Finland (inoculated malts) and 80 HaeR, 128 HaeR and 457 RsaF with
studied French malts (plot not shown).
Some of these TRFs were tentatively identified from clone libraries
as species belonging to fungal genera, i.e. Alternaria, Candida,
Cladosporium, Cryptococcus, Filobasidium, Fusarium, Geotrichum,
Hanseniaspora, Holtermanniella, Issatchenkia, Pichia, Saccharomyces,
Sporobolomyces, and Trichosporon, and bacterial genera, i.e. Arthrobacter,
Brachybacterium, Curtobacterium, Kineococcus, Microbacterium, Erwinia
and Pseudomonas (Table 3).
With SIMPER analysis the average percent similarity and the
number of combined bacterial and fungal TRFs needed to explain
90% of this similarity within a group were analyzed. Average microbial community similarity within a group was the lowest in N.
American malts and highest in Russian malts (Table 4). Lower homogeneity and higher within group variability were also indicated by
higher dispersion indices for malts from N. America, Australia and
S. Africa as compared to Russia, Slovakia, Denmark, Belgium, France
and Argentina. Considering the large number of samples in the
Australian group, the average similarity within this group was comparable with other groups that had fewer sample numbers such as S.
Africa. N. America malts also had the lowest Shannon's and Simpson
diversity indices, Pielou's species evenness and Margalef richness.
Australian malts were comparable with those from France, S. Africa
and Slovakia in species richness, evenness and diversity. Samples
from Belgium had the highest species richness, evenness and diversity. However, the analyses of variance of diversity indices for different geographical locations showed non-significant differences
among themselves (P N 0.05). When different geographical malt
groups were compared between themselves using SIMPER analysis
the N. America malts showed the greatest dissimilarity from S.
Africa malts (67.5%), closely followed by the Australia (65.5%),
Belgium (65.5%) and France (65.1%).
3.2. Relation between malt quality parameters and microbial communities
A subset of malt samples from different countries supplied by one
single international brewer was studied for routine malt quality parameters (Table 5) and microbial TRFLP analysis to establish relation among
these two approaches of malt quality evaluation.
The main PERMANOVA results showed significant differences in
fungal (P = 0.0001) community profiles of these malts. Further, pair
wise comparison showed significant differences among samples from
France and Russia, France–Argentina, France–Slovakia, and Russia–
Argentina (P b 0.01). Similar were the results when PERMANOVA statistical routine was applied to the brewer's malt quality parameters. The
differences with regard to bacterial communities were non-significant
(P = 0.0276). Similar results were observed from CAP analysis (data
not shown).
To find which fungal TRFs contributed to these differences,
Pearson's correlation (N 0.55) was determined. The 368 HinfR, 379
HaeF, 229 RsaR, 205 HinfF and 148 HaeR TRFs were positively correlated with France, 498 RsaF, 97 HaeR, 113 RsaR, 154 HaeR and 96
HinfF with Russia and 200 RsaR, 69 HinfR, 90 HaeF and 140 RsaF
with malt samples from Argentina. Some of these TRFs were
tentatively identified as fungal species of Candida, Cladosporium,
Cryptococcus, Filobasidium, Fusarium, Geotrichum, Holtermanniella,
Issatchenkia and Sporobolomyces genera.
The dbRDA plot of fungal TRFs data (Fig. 2) showed relationship between these TRFs and malt quality parameters. Out of malt moisture
content, fine grind extract, wort color, total protein content, Kolbach
Index, beta-amylase, α-amylase and limit dextrinase, the moisture,
beta-amylase, α-amylase and limit dextrinase contents of the malts
showed significant correlations (P b 0.004) with TRFs (30.5% of total
variations explained by the first two axes). The beta-amylase activity
was positively correlated with the fungal TRF profiles of samples from
France, Belgium, Denmark and Argentina and α-amylase negatively correlated with samples from France, Belgium and Denmark. The αamylase was positively correlated with fungal TRF profiles of malt samples from Slovakia. Positive correlations were observed among moisture
content, limit dextrinase and Russian malt sample fungal community
structure.
M. Kaur et al. / International Journal of Food Microbiology 215 (2015) 71–78
75
Fig. 1. Canonical analysis of principal coordinates (CAP) of fungal TRFLP peak data from barley malt samples based on Bray Curtis similarity matrix and grouped by the geographical
location of samples. Legends with circle around them represent DON and OTA inoculated Finnish malts, a standard gushing malt sample from S. Africa, and a Fusarium head blight infected
N. American malt sample. Vector overlays indicate Pearson's correlations between the ordination axes and individual taxa (only taxa with N0.55 correlation are shown).
4. Discussion
As reported by Laitila et al. (2007) and Normander and Prosser
(2000), analysis of barley-associated fungal and bacterial communities
is complicated by the co-recovery of barley chloroplast and mitochondrial 16S rRNA genes. This problem was resolved here by excluding barley chloroplast or mitochondrial rRNA gene derived peaks from the TRF
data before performing statistical analyses. With this achieved, TRFLP
analysis was demonstrated to be a useful tool to monitor spatial microbial population diversity in malts once optimized. TRFLP analysis was
further complemented by cloning and sequencing to assist with the
identification of dominant fragments within different TRFLP profiles.
Table 3
Tentative identification of bacterial and fungal TRFs based upon clone libraries data.
TRF
Predicted genus
135 HaeR, 144 HinfF, 156 HaeR
112 HaeF, 156 HaeR, 317 HaeR, 379 HaeF
90 HaeF, 135 HaeR, 156 HaeR, 382 HinfR
69 HinfR, 96 HinfF, 140 RsaF, 144 HinfF
96 HinfF, 144 HinfF
97 HaeR, 215 HinfF, 220 HinfF
140 RsaF, 154 HaeR, 210 HinfF, 435 HaeF
156 HaeR
100 HaeR
90 HaeF, 215 HinfF, 379 HaeF
317 HaeR
459 HaeF
113 RsaR, 135 HaeR, 200 RsaR
144 HinfF
Fungi
Alternaria
Candida
Cladosporium
Cryptococcus
Filobasidium
Fusarium
Geotrichum
Hanseniaspora
Holtermanniella
Issatchenkia
Pichia
Saccharomyces
Sporobolomyces
Trichosporon
80 HaeR
128 HaeR
125 RsaF
Bacteria
Arthrobacter, Brachybacterium,
Curtobacterium, Kineococcus,
Microbacterium
Erwinia
Pseudomonas
The discriminative ability of the approach (molecular techniques and
further statistical analyses) used in this study was demonstrated as
seen from distinct groupings of artificially and naturally inoculated or
infected test malts (samples from Finland, S. Africa and N. America denoted in Fig. 1). Further, tentative identification of TRFs associated
with this group confirmed the presence of Fusarium spp. However,
none of the identified fragments correspond to Penicillium sp. as was expected for the ochratoxin A (OTA) inoculated malts. This might be because of the limited number of clone libraries constructed in this
study which might not be sufficient to capture the greater microbial diversity associated with studied malts. This might be the reason that
many TRFs, such as 148 HaeR, 203 HinfF, 229 RsaR, 144 HinfR, 311RsaF,
95 HaeF, 126 RsaR, 123 RsaF, 80 HaeR and 457 RsaF could not be identified in this study. Also fungal taxa likely share many TRFs and thus may
not be resolved by the procedure adopted in this study. One way to
overcome this is to use next generation microbial sequencing methods
such as pyrosequencing and Miseq which are becoming more routine
in microbial ecology because of their competitive pricing and higher
resolution. However, the higher sequencing error rates and shorter sequences obtained with these next generation sequencing techniques
are not directly comparable with already established Sanger sequencing
and TRFLP used here in this study. Even with new approaches it is often
difficult to identify microbial communities accurately to species level
especially for fungi with great inter-genome variability and where taxonomic organization is still largely underway and until recently permitted dual nomenclature for a genetically identical microorganism/
pleomorph thus creating ambiguity in molecular-based identifications.
Another factor that can cause ambiguous and/or false identification is
incorrectly annotated sequences uploaded to the sequencing databases.
It is known that up to 20% of fungal sequences in databases are incorrectly annotated at the species level (Nilsson et al., 2006; Yamamoto
et al., 2014). Further, unlike bacteria there are no direct bioinformatics
pipelines for high throughput fungal sequencing data analysis especially
for 28S rRNA gene marker thus requiring great bioinformatics input.
Overall, this study stands as a useful reference point on which future
76
M. Kaur et al. / International Journal of Food Microbiology 215 (2015) 71–78
Table 4
Similarity percentage (SIMPER) analysis, multivariate dispersion (MVDISP), average Margalef species richness (d), Pielou's evenness (J'), Shannon (H') and Simpson's diversity (1-λ′)
indices (figures in parentheses show the standard deviation observed within each group) calculated for each barley malt geographical group TRFLP peak data.
Country of origin
SIMPER
MVDISP
d
J'
H'
1-λ′
Argentina
Australia
Belgium
Denmark
France
N. America
Russia
Slovakia
S. Africa
47.6
39.5
52.7
54.8
52.4
32.5
57.4
54.3
40.5
0.52
1.07
0.17
0.14
0.22
1.69
0.11
0.12
1.01
15.30 (2.575)
16.3 (22.681)
19.42 (2.022)
15.49 (1.152)
16.98 (2.383)
12.04 (0.742)
15.23 (1.533)
16.72 (0.019)
16.71 (2.107)
0.94 (0.016)
0.95 (0.022)
0.96 (0.000)
0.94 (0.017)
0.96 (0.017)
0.91 (0.018)
0.94 (0.013)
0.94 (0.016)
0.95 (0.022)
4.35 (0.229)
4.41 (0.247)
4.67 (0.127)
4.35 (0.137)
4.50 (0.203)
3.99 (0.136)
4.32 (0.136)
4.44 (0.085)
4.45 (0.213)
0.98 (0.005)
0.99 (0.006)
0.99 (0.001)
0.98 (0.004)
0.99 (0.004)
0.97 (0.006)
0.98 (0.003)
0.99 (0.002)
0.99 (0.005)
studies using more modern and reliable molecular microbial population
ecology techniques may be based.
Differences in microbial diversity, richness and evenness were observed among Australian malt samples, which might be the result of
the wide distribution of barley cropping regions in different agro climatic zones within Australia. Backhouse and Burgess (2002) and Backhouse
et al. (2004) have reported spatial differences between regions within
Australia in the distribution of Fusarium spp. associated with cereals.
In addition, the barley malt assessed in this study was sourced from different varieties, which might have contributed to within group heterogeneity. Different barley head architecture and differential responses to
microbial taxa could also account for generating increased biodiversity.
Barley varietal differences have previously been found to significantly
influence Fusarium populations and associated mycotoxin production
(Perkowski et al., 2003). Despite the large number of criteria (cultivation conditions in the field, agricultural practices, the barley cultivars,
lodging, fertilization, soil, storage conditions after harvesting, and different malting processes) this study still found that geographical region is
clearly an influential factor affecting microbial community structure of
grains as malt samples from different geographical locations tend to
form distinct groups especially for fungi, thus suggesting that initial microbial community on the barley grain is very influential on what is
present through the ensuing supply/manufacturing process.
Statistical analysis of TRFLP data showed geographical differences in
fungal community structure of Australian malts compared to those from
Argentina, Finland, France, Russia and S. Africa. Geographical differences
in fungal communities were also observed among other groups. The
only significant difference in malt bacterial community structure was
observed in Finland and France. Bacterial diversity seemed to be more
consistent in groups than fungal community suggesting a lesser influence of geographical factors on bacterial species. This might be due to
less host specificity and more cosmopolitan distribution of bacteria
than fungi in terrestrial plant systems.
The differences observed here were the combined effect of both
qualitative (type of fungi) as well as quantitative (relative abundance
of different fungi) dissimilarities as witnessed from differences in number and relative abundance of TRFs in different samples. One can argue
that whether the differences observed here were real or a result of PCR
artifacts (Lueders and Friedrich, 2003; Qiu et al., 2001). Great care was
taken into account to minimize these effects by replicating DNA extraction and TRFLP analyses steps, concurrent running of PCR and TRFLP for
all samples, binning of TRF singletons, using three different restriction
enzymes for bacteria and fungi, different fluorescent dyes labeled forward as well as reverse primers and removing any chimeras in sequencing. Differences in fungal populations on S. African malts compared to
Northern Hemisphere malts have been reported by Rabie and Lübben
(1993). Ackermann (1998) also observed lower fungal counts in S.
African barley malts as compared to malts from other countries.
However, both of these studies emphasized that the predominant
fungal species were the same on malts regardless of their geographical
origin. In contrast, this study found regional differences in malt associated fungal communities, comprising both quantitative and qualitative
differences in fungal populations. These observed differences are of interest since fungi can contribute significantly to the overall quality of
malts (Laitila et al., 2007). Certain fungal species or strains are more deleterious in their effect on malt quality, in particular production of mycotoxins or causing process failures such as premature yeast flocculation
during beer fermentation and effect on beer quality such as overfoaming/gushing (Evans et al., 1999; Jestoi et al., 2004; Kaur et al.,
2012; Kosiak et al., 2004; Shokribousjein et al., 2011). Conversely,
some can be beneficial to malt quality e.g., the antagonistic yeast,
Wickerhamomyces anomalus can be used as a malting starter culture to
improve malt quality (Laitila et al., 2011). This spatial displacement of
detrimental microbial taxa due to competition from less harmful or beneficial taxa could contribute to the overall improved quality and safety
of some malts compared to others.
This investigation also noted the presence of a large diversity of
yeasts (indicated in fungal clone libraries and corresponding TRFs)
which might have contributed to the differences between malt groups.
In previous studies, comparisons among barley and malt samples were
mostly concentrated on culturable filamentous fungi and less attention
has been paid to yeasts and yeast-like fungi with exception of the extensive work done by Laitila et al. (2006). Yeasts are the second most abundant microbes after bacteria in viable counts in pre harvest barley
(Flannigan, 2003) and are reported to survive during storage and
malting (Clarke and Hill, 1981; Haikara et al., 1977; Flannigan et al.,
1982). Furthermore, malt associated yeasts and yeast-like fungi are
known for their potential to produce extracellular hydrolytic enzymes
with a potential contribution to the overall malt enzyme spectrum
(Laitila et al., 2011). Work done by Laitila et al. (2006) revealed that several yeasts, especially basidiomycetous species, were active producers
of α-amylase, β-glucanase, cellulase and endo xylanase. In our study
Table 5
Mean value of physicochemical properties of selected barley malts from a single brewer (figures in parentheses show the standard deviation observed within each group).
Country of
origin
Moisture
(%)
Fine grind extract
(% dwt)
Wort color
(°EBC)
Total protein
content (% dwt)
Kolbach
Index (%)
beta-amylase
(U/g)
α-amylase
(U/g)
Limit dextrinase
(U/kg)
Argentina
Belgium
Denmark
France
Russia
Slovakia
4.0 (0.4)
4.2 (0.2)
4.3 (0.2)
4.3 (0.2)
4.9 (0.3)
4.0 (0.3)
80.5 (0.9)
79.6 (0.1)
81.3 (0.6)
79.7 (1.0)
80.9 (0.5)
82.0 (0.6)
3.6 (0.6)
3.8 (0.4)
4.7 (0.6)
3.8 (0.4)
4.6 (0.2)
3.8 (0.3)
10.9 (0.6)
11.3 (0.0)
9.9 (0.3)
11.3 (0.5)
10.7 (0.3)
10.6 (0.2)
40.9 (4.3)
32.7 (0.0)
37.2 (2.2)
34.7 (3.0)
38.3 (1.3)
38.7 (1.5)
610.4 (98.9)
762.5 (81.3)
577.7 (104.5)
769.0 (187.6)
420.4 (58.6)
542.7 (95.0)
193.0 (32.6)
126.0 (14.1)
145.0 (19.3)
120.0 (37.5)
181.2 (36.7)
191.3 (19.5)
330.6 (98.9)
109.0 (9.9)
198.3 (37.4)
120.5 (35.3)
320.2 (89.4)
206.3 (1.2)
M. Kaur et al. / International Journal of Food Microbiology 215 (2015) 71–78
77
Fig. 2. dbRDA analysis of variation of TRFLP fungal community structure constrained at geographical location as explained by malt physicochemical properties. Vectors represent
correlations of variables with community structure along the first two dbRDA axes by Pearson's correlations.
we also found relationships between fungal TRFs and their corresponding putative fungal species with malt quality parameters including betaamylase, α-amylase and limit dextrinase contents.
The impact of Fusarium on barley malt quality is well documented in
the literature. Besides producing mycotoxins such as DON, Fusarium
spp. are known to reduce wort β-glucan and viscosity and increase soluble nitrogen, free amino nitrogen, and wort color as compared to wort
produced from control malt (Haikara, 1983; Schwarz, 2003; Schwarz
et al., 2002). These changes in wort characteristics are indicative of the
action of fungal enzymes. Extreme structural proteolytic, hemicellulolytic and starch deterioration of Fusarium infected grains coupled
with higher protease and β-glucanase activities, a greater proportion of
free amino and soluble nitrogen, and a lower β-glucan content was also
reported by Oliveira et al. (2012) and Nielsen et al. (2014). The higher
malt moisture content of studied Russian malts likely resulted in greater
fungal growth during storage and their positive correlation with limit
dextrinase along with the putative TRF identifying Fusarium spp. supported the views of earlier researchers. Wet barley harvest conditions
result in higher grain moisture content favoring the growth of Fusarium
(Sarlin et al., 2005) which is capable of producing extracellular dextranase (Simonson et al., 1975). Although correlations among putative fungal taxa and enzymatic activities of the malts were found in this study,
future efforts should focus on confirming this with further research to
better quantify the relative amounts of enzymes contributed by fungi,
bacteria and grain.
5. Conclusion
An investigation was undertaken by microbial DNA based fingerprinting of Australian malt made from barley grown in different regions.
This was benchmarked against malts from barley grown internationally
by using TRFLP and cloning and sequencing analyses. The TRFLP approach proved appropriate because it is comparatively rapid and cost efficient and profiles from a large number samples can be assessed. Both
qualitative and quantitative differences were observed in bacterial and
especially fungal communities associated with malts produced in different geographical regions. Further, relationships between malt associated microbial diversity and routine physicochemical malt quality
parameters used in malting and brewing industry emphasized the critical though often overlooked role these microbes play in the overall
quality and safety of barley malts. Our analyses also indicated that
fungi were more important in influencing malt quality outcomes than
bacteria.
Acknowledgments
This research was funded by the Australian Research Council and the
Cargill (LP0560329), Australia. We greatly acknowledge Adam
Smolenski (Central Science Laboratory — Research, University of
Tasmania, Australia) for his assistance in TRFLP analysis.
References
Ackermann, A., 1998. Mycoflora of South African barley and malt. J. Am. Soc. Brew. Chem.
56, 169–176.
Analytica — EBC, 1998. European Brewing Convention. Verlag Hans Carl GetränkeFachverlag, Nürnberg, Germany.
Anderson, M.J., Gribble, N.A., 1998. Partitioning the variation among spatial, temporal and
environmental components in a multivariate data set. Aust. J. Ecol. 23, 158–167.
Anderson, M.J., Willis, T.J., 2003. Canonical analysis of principal coordinates: a useful
method of constrained ordination for ecology. Ecology 84, 511–525.
Backhouse, D., Burgess, L.W., 2002. Climatic analysis of the distribution of Fusarium
graminearum, F. psudograminearum and F. culmorum on cereals in Australia. Australas.
Plant Pathol. 31, 321–327.
Backhouse, D., Abubakar, A.A., Burgess, L.W., Dennis, J.I., Hollaway, G.J., Wildermuth, G.B.,
Wallwork, H., Henry, F.J., 2004. Survey of Fusarium species associated with crown rot
of wheat and barley in eastern Australia. Australas. Plant Pathol. 33, 255–261.
Bamforth, C.W., Barclay, A.H.P., 1993. Malting technology and the uses of malt. In:
MacGregor, A.W., Bhatty, R.S. (Eds.), Barley: Chemistry and Technology. American
Association of Cereal Chemists Inc., St. Paul, Minnesota, pp. 297–354.
Birgitte, A., Thrane, U., Svendsen, A., Rasmussen, I.A., 1996. Associated field mycobiota on
malting barley. Can. J. Bot. 74, 854–858.
Bokulich, N.A., Mills, D.A., 2012. Next generation approaches to the microbial ecology of
food fermentations. BMB Rep. 45, 377–389.
Bray, J.R., Curtis, J.T., 1957. An ordination of the upland forest communities of Southern
Wisconsin. Ecol. Monogr. 27, 325–349.
Briggs, D.E., 1978. Barley. Chapman and Hall Ltd, London.
Clarke, J., Hill, S., 1981. Mycofloras of moist barley during sealed storage in farm and
laboratory silos. Trans. Br. Mycol. Soc. 77, 557–565.
Clarke, K.R., Warwick, R.M., 2001. Change in Marine Communities: An Approach to
Statistical Analysis and Interpretation. PRIMER-E, Plymouth, UK.
Doohan, F.M., Brennan, J., Cooke, B.M., 2003. Influence of climatic factors on Fusarium
species pathogenic to cereals. Eur. J. Plant Pathol. 109, 755–768.
Evans, D.E., 2008. A more cost- and labor-efficient assay for the combined measurement
of the diastatic power enzymes beta-amylase, α-amylase and limit dextrinase. J. Am.
Soc. Brew. Chem. 66, 215–222.
Evans, D.E., Sheehan, M.C., Stewart, D.C., 1999. The impact of malt derived proteins on
beer foam quality. Part II: the influence of malt foam-positive proteins and nonstarch polysaccharides on beer foam quality. J. Inst. Brew. 105, 171–177.
Flannigan, B., 2003. The microbiota of barley and malt. In: Priest, F.G., Campbell, I. (Eds.),
Brewing Microbiology. Kluwer Academic/Plenum Publishers, New York, pp. 113–180.
78
M. Kaur et al. / International Journal of Food Microbiology 215 (2015) 71–78
Flannigan, B., Okagbue, R.N., Khalid, R., Teoh, C.K., 1982. Mould flora of malt in production
and storage. Brew. Distil. Int. 12 (31–33), 37.
Goldammer, T., 2008. The Brewer's Handbook — The Complete Book to Brewing Beer.
Apex Publishers, Clifton, Virginia.
Haikara, A., 1983. Malt and beer from barley artificially contaminated with Fusarium in
the field. Proceedings of the European Brewing Convention 19, pp. 401–408.
Haikara, A., Mäkinen, V., Hakulinen, R., 1977. On the microflora of barley after harvesting,
during storage and in malting. Proceedings of the European Brewing Convention 16,
pp. 35–46.
Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis
program for Windows 95/98/NT. Nucleic Acids Symp. Ser. 41, 95–98.
Hill, R.A., Lacey, J., 1983. The micro flora of ripening barley grain and the effect of preharvest fungicide application. Ann. Appl. Biol. 102, 455–465.
Hunter, H., 1962. The science of malting barley production. In: Cook, A.H. (Ed.), Barley and
Malt Biology, Biochemistry, Technology. Academic Press, New York, pp. 25–44.
Jarvis, B., Williams, A.P., 1987. Methods for detecting fungi in foods and beverages. In:
Beuchat, L.R. (Ed.), Food and Beverage Mycology. Van Nostrand Reinhold, New
York, pp. 599–636.
Jestoi, M., Rokka, M., Yli-Mattila, T., Parikka, P., Rizzo, A., Peltonen, K., 2004. Presence and
concentrations of the Fusarium-related mycotoxins beauvericin, enniatins and
moniliformin in Finnish grain samples. Food Addit. Contam. 21, 794–802.
Justé, A., Malfliet, S., Lenaerts, M., de Cooman, L., Aerts, G., Willems, K.A., Lievens, B., 2011.
Microflora during malting of barley: overview and impact on malt quality. Brew. Sci.
64, 22–31.
Kaur, M., Sheehy, M., Stewart, D.C., Bowman, J.P., Davies, N.W.A., Evans, D.E., 2009. Improving the cost efficiency of quality assurance screening for mycotoxins in malting
barley. J. Am. Soc. Brew. Chem. 67, 95–98.
Kaur, M., Bowman, J.P., Stewart, D.C., Sheehy, M., Janusz, A., Speers, R.A., Koutoulis, A.,
Evans, D.E., 2012. TRFLP analysis reveals that fungi rather than bacteria are associated
with premature yeast flocculation in brewing. J. Ind. Microbiol. Biotechnol. 39,
1821–1832.
Kosiak, B., Torp, M., Skjerve, E., Andersen, B., 2004. Alternaria and Fusarium in Norwegian
grains of reduced quality — a matched pair sample study. Int. J. Food Microbiol. 93,
51–62.
Krstanović, V., Klapec, T., Velić, N., Milaković, Z., 2005. Contamination of malt barley and
wheat by Fusarium graminearum and Fusarium culmorum from the crop years
2001–2003 in eastern Croatia. Microbiol. Res. 160, 353–359.
Laitila, A., 2008. More good than bad — microbes in the maltings. Brewer Distil. Int. 8,
52–54.
Laitila, A., Wilhelmson, A., Mikkola, H., Kauppila, R., Kotaviita, E., Olkku, J., Haikara, A.,
Home, S., 2003. Importance of drying and sound storage conditions on the microbiological safety and viability of malting barley. Proceedings of the European Brewing
Convention 29, pp. 1235–1245.
Laitila, A., Wilhelmson, A., Kotaviita, E., Olkku, J., Home, S., Juvonen, R., 2006. Yeasts in an
industrial malting ecosystem. J. Ind. Microbiol. Biotechnol. 33, 953–966.
Laitila, A., Kotaviita, E., Peltola, P., Home, S., Wilhelmson, A., 2007. Indigenous microbial
community of barley greatly influences grain germination and malt quality. Brewer
Distil. Int. 113, 9–20.
Laitila, A., Sarlin, T., Raulio, M., Wilhelmson, A., Kotaviita, E., Huttunen, T., Juvonen, R.,
2011. Yeast in malting, with special emphasis on Wickerhamomyces anomalus (synonym Pichia anomala). Antonie Van Leeuwenhoek 99, 75–84.
Lueders, T., Friedrich, M.W., 2003. Evaluation of PCR amplification bias by terminal restriction fragment length polymorphism analysis of small-subunit rRNA and mcrA genes
by using defined template mixtures of methanogenic pure cultures and soil DNA extracts. Appl. Environ. Microbiol. 69, 320–326.
Nielsen, L.K., Cook, D.J., Edwards, S.G., Ray, R.V., 2014. The prevalence and impact of Fusarium head blight pathogens and mycotoxins on malting barley quality in UK. Int.
J. Food Microbiol. 179, 38–49.
Nilsson, R.H., Ryberg, M., Kristiansson, E., Abarenkov, K., Larsson, K.H., Kõljalg, U., 2006.
Taxonomic reliability of DNA sequences in public sequence databases: a fungal perspective. PLoS One 1, e59.
Noots, I., Delcour, J.A., Michiels, C.W., 1999. From field barley to malt: detection and specification of microbial activity for quality aspects. Crit. Rev. Microbiol. 25, 121–153.
Normander, B., Prosser, J.I., 2000. Bacterial origin and community composition in the barley phytosphere as a function of habitat and presowing conditions. Appl. Environ.
Microbiol. 66, 4372–4377.
Oliveira, P.M., Mauch, A., Jacob, F., Waters, D.M., Arendt, E.K., 2012. Fundamental study on
the influence of Fusarium infection on quality and ultrastructure of barley malt. Int.
J. Food Microbiol. 156, 32–43.
Perkowski, J., Kiecana, I., Kaczmarek, Z., 2003. Natural occurrence and distribution of Fusarium toxins in contaminated barley cultivars. Eur. J. Plant Pathol. 109, 331–339.
Qiu, X., Wu, L., Huang, H., McDonel, P.E., Palumbo, A.V., Tiedje, J.M., Zhou, J., 2001. Evaluation of PCR-generated chimeras, mutations, and heteroduplexes with 16S rRNA
gene-based cloning. Appl. Environ. Microbiol. 67, 880–887.
Rabie, C.J., Lübben, A., 1993. The mycoflora of South African barley and barley malt. Proceedings of the Scientific and Technical Convention — Institute of Brewing 4,
pp. 55–73.
Rabie, C.J., Lübben, A., Marais, G.J., Jansen van Vuuren, H., 1997. Enumeration of fungi in
barley. Int. J. Food Microbiol. 35, 117–127.
Rasmusson, D.C., 1985. Barley. American Society of Agronomy, Madison, Wisconsin.
Rees, G.N., Baldwin, D.S., Watson, G.O., Perryman, S., Nielsen, D.L., 2004. Ordination and
significance testing of microbial community composition derived from terminal restriction fragment length polymorphisms: application of multivariate statistics.
Antonie Van Leeuwenhoek 86, 339–347.
Sarlin, T., Laitila, A., Pekkarinen, A., Haikara, A., 2005. Effects of three Fusarium species on
the quality of barley and malt. J. Am. Soc. Brew. Chem. 63, 43–49.
Schütte, U.M.E., Abdo, Z., Bent, S.J., Shyu, C., Williams, C.J., Pierson, J.D., Forney, L.J., 2008.
Advances in the use of terminal restriction fragment length polymorphism (TRFLP) analysis of 16S rRNA genes to characterize microbial communities. Appl.
Microbiol. Biotechnol. 80, 365–380.
Schwarz, P.B., 2003. Impact of head blight on the malting and brewing quality of barley.
In: Leonard, K.J., Bushnell, W.R. (Eds.), Fusarium Head Blight of Wheat and Barley.
American Phytopathological Society Press, St. Paul, MN, pp. 395–419.
Schwarz, P.B., Jones, B.L., Steffenson, B.J., 2002. Enzymes associated with Fusarium infection of barley. J. Am. Soc. Brew. Chem. 60, 130–134.
Shokribousjein, Z., Sylvie, M., Deckers, S.M., Gebruers, K., Lorgouilloux, Y., Baggermanc, G.,
Verachtert, H., Jan, A., Delcour, J.A., Etiennee, P., Rock, J.M., Michiels, C., Derdelinckx,
G., 2011. Hydrophobins, beer foaming and gushing. Cerevisia 32, 85–101.
Simonson, L.G., Liberta, A.E., Richardson, A., 1975. Characterization of an extracellular dextranase from Fusarium moniliforme. Appl. Microbiol. 30, 855–861.
Smith, C.J., Danilowicz, B.S., Clear, A.K., Costello, F.J., Wilson, B., Meijer, W.G., 2005. TAlign,
a web-based tool for comparison of multiple terminal restriction fragment length
polymorphism profiles. FEMS Microbiol. Ecol. 54, 375–380.
Yamamoto, N., Dannemiller, K.C., Bibby, K., Peccia, J., 2014. Identification accuracy and
diversity reproducibility associated with internal transcribed space-based fungal
taxonomic library preparation. Environ. Microbiol. 16, 2764–2776.