Contribution of reutericyclin production to the stable persistence of

Lactobacillus reuteri in an industrial sourdough fermentation

Michael G. Ganzle*, Rudi F. Vogel

TU Munchen, Lehrstuhl fur Technische Mikrobiologie, Weihenstephaner Steig 16, 85350 Freising, Germany

Received 31 July 2001; received in revised form 26 February 2002; accepted 3 March 2002

Abstract

Reutericyclin is a small molecular weight antibiotic produced by the sourdough isolate Lactobacillus reuteri LTH2584. This

strain was isolated from an industrial sourdough, SER, in 1988. To determine whether reutericyclin formation contributes to the

stable persistence of L. reuteri in sourdough, evaluations were made on whether reutericyclin-producing strains were among L.

reuteri isolates from the SER sourdough obtained in 1994 and 1998. These strains were characterised on species and strain level

by physiological tests and randomly amplified polymorphic DNA (RAPD) and restriction fragment length polymorphism

(RFLP) patterns. Reutericyclin production in dough was evaluated by two methods, a bioassay and HPLC. Throughout 10 years

of continuous propagation, reutericyclin-producing L. reuteri strains were present in SER sourdough. All isolates exhibited

similar physiological properties and molecular typing revealed closely related patterns. Two isolates obtained in 1994 and 1998

were identical. Reutericyclin produced in situ by L. reuteri was active in dough against reutericyclin-sensitive L.

sanfranciscensis. The reutericyclin concentration in dough fermented with L. reuteri was 5 mg kg� 1. The results indicate that

reutericyclin production contributed to the stable persistence of L. reuteri in sourdough. Because reutericyclin is produced in

active concentrations during sourdough fermentations, it is a suitable candidate for use as natural preservative.

D 2002 Elsevier Science B.V. All rights reserved.

Keywords: Sourdough; Reutericyclin; Lactobacillus reuteri

1. Introduction

Sourdough fermentation is employed in bread

production because of the high sensory quality of

sourdough bread. Sourdough fermentation contributes

to the characteristic flavour of bread (Hansen and

Hansen, 1996), improves bread texture, and delays

bread staling and microbial spoilage of bread (Armero

and Collar, 1996, 1998; Rosenquist and Hansen,

1998; Corsetti et al., 2000).

The sourdough microflora is composed of stable

associations of lactobacilli and yeasts. The single most

important microorganism prevailing in sourdough is

Lactobacillus sanfranciscensis. Depending on the

fermentation conditions, other species such as L.

pontis, L. panis, L. frumenti, or L. reuteri occur in

relevant cell counts (for review, see Hammes and

Ganzle, 1997; Vogel et al., 1999). The most prominent

metabolic activity of microorganisms in sourdough is

the production of acid and carbon dioxide. Gas

production is required for leavening of the dough

0168-1605/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved.

PII: S0168 -1605 (02 )00146 -0

* Corresponding author. Tel.: +49-8161-713959; fax: +49-

8161-713327.

E-mail address: michael.gaenzle@bl.tum.de (M.G. Ganzle).

www.elsevier.com/locate/ijfoodmicro

International Journal of Food Microbiology 80 (2002) 31–45

unless baker’s yeast is added. Dough acidification is a

prerequisite for rye baking to inhibit the flour a-

amylase. Acidification furthermore activates cereal

phytases, resulting in almost complete phytate hydrol-

ysis during sourdough fermentation (Fretzdorff and

Brummer, 1992; Tangkongchitr et al., 1982).

Enhanced proteolysis is observed in sourdough or in

sterile, acidic doughs, compared to doughs fermented

with a straight dough process (Kratochvil and Holas,

1984; Thiele et al., 2002). Sourdough fermentation

promotes a solubilisation of pentosans at the dough

stage and thus enhances water binding of dough

(Martinez-Anaya and Devesa, 2000). Other metabolic

activities of sourdough lactobacilli which are of

importance for bread quality are their proteolytic

activity (Gobbetti et al., 1996), the conversion of

arginine to ornithine enhancing the roasty flavour of

bread (Ograbek et al., 1999; Thiele et al., 2002), and

the production of exopolysaccharides in dough (Kor-

akli et al., 2001), which potentially affects bread

texture and staling.

A reproducible and controlled composition and

activity of the sourdough microflora is paramount to

achieve a constant quality of sourdough bread. In

bakery practice, sourdough is usually sustained by

repeated inoculation. Anecdotal reports exist of sour-

doughs maintained over several centuries and contin-

uous use of a sourdough over seven decades has been

documented (Bocker et al., 1990). Despite annual

changes in raw materials, seasonal changes in temper-

ature as well as ample opportunity for contamination

from either the raw material or the bakery environ-

ment, the sourdough microflora often is remarkably

stable. Monitoring of the microflora of two industrial

sourdoughs over a period of 7 months revealed only

minor shifts in the composition of lactobacilli (Rosen-

quist and Hansen, 2000). In BRS, a sourdough starter

propagated according to traditional procedures, the

composition remained stable on strain level of over a

period of at least two decades (Spicher and Schroder,

1978; Bocker et al., 1990; Ganzle et al., 1998).

Several factors account for the dominance of

sourdough lactobacilli. First, their carbohydrate

metabolism is highly adapted to the main substrates

in dough, maltose and fructose. Utilisation of maltose

via maltose phosphorylase and the pentose phophate

shunt with fructose as co-substrate results in a higher

energy yield than homofermentative maltose degrada-

tion (Stolz et al., 1995, 1996). Second, the growth

requirements of L. sanfranciscensis with respect to

temperature and pH match the conditions encountered

during sourdough fermentation (Ganzle et al., 1998).

Third, antimicrobial compounds may contribute to the

stable persistence of lactobacilli in sourdough fermen-

tations (Corsetti et al., 1996; Olsen et al., 1995).

A small molecular weight antibiotic, reutericyclin,

produced by the sourdough isolate L. reuteri

LTH2584 was purified and characterised (Holtzel et

al., 2000). Reutericyclin is active against a broad

range of Gram-positive bacteria in concentrations of

less than 1 mg l � 1 and the inhibitory spectrum

includes those lactic acid bacteria relevant in sour-

dough fermentations (Ganzle et al., 2000). Based on

the conditions favouring reutericyclin production in

laboratory media, Ganzle et al. (2000) suggested that

reutericyclin may be produced in active concentra-

tions in dough. L. reuteri LTH2584 was isolated in

1988 from SER, an in house rye sourdough prepared

for the production of a commercially available baking

aid (Bocker et al., 1995; Bocker, personal communi-

cation). The microflora of SER sourdough was moni-

tored over a period of 10 years. Over these 10 years of

continuous propagation, corresponding to about

50000 generations of microbial growth, considerable

shifts were observed concerning the composition of

the dough microflora. However, relevant cell counts

of L. reuteri were observed at each isolation time. It

was the aim of this study to determine whether

reutericyclin production is relevant for dough ecology.

It was therefore determined (i) whether reutericyclin

producing strains are among the SER isolates obtained

in 1994 and 1998, (ii) whether those isolates are

identical to L. reuteri LTH2584 on species or strain

level, and (iii) whether reutericyclin is produced in

active concentrations in dough.

2. Materials and methods

2.1. Microorganisms and media

L. reuteri st rains LTH2584, TMW1.106,

TMW1.112 and TMW1.656 were isolated in 1988,

1994, or 1998 from SER sourdough. The flora compo-

sition of SER sourdough in 1988, 1994, and 1998 is

summarised in Table 1. L. reuteri TMW1.614 is a

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–4532

sorghum sourdough isolate (Hamad et al., 1997) and

L. reuteri TMW1.693 is the DSM 20016T type strain.

L. sanfranciscensis LTH2581 is a reutericyclin-sensi-

tive sourdough isolate (Bocker et al., 1990; Ganzle et

al., 2000). Strains were grown in mMRS4 (Stolz et al.,

1993) containing the following per liter: 10 g tryptone,

5 g meat extract, 5 g yeast extract, 10 g maltose, 5 g

fructose, 5 g glucose, 2.6 g KH2PO4, 4 g K2HPO4�3H2O, 3 g diammonium citrate, 3 g NH4CL, 0.5 g

cysteine�HCL, 1 g Tween 80, 0.2 mg MgSO4�7H2O, 0.05 g MnSO4�H2O, and 0.5 Ag each of

cobalamine, folic acid, niacin, panthotheic acid, pyr-

idoxal, and thiamine. Buffered mMRS4 contained 20

g l�1 maltose, 10 g l�1 each of glucose and fructose, 5

g sodium acetate� 3H2O, and other compounds as in

mMRS4. The modified API 50 CH medium used to

determine the sugar fermentation patterns was com-

posed as mMRS4 but did not contain maltose, glu-

cose, or fructose, and additionally contained 170 mg

l � 1 bromcresol purple. Solid media for enumeration

of cell counts additionally contained 15 g agar agar

per liter. L. reuteri strains were incubated at 37 jC and

L. sanfranciscensis was incubated at 30 jC.

2.2. Inhibitory activity of L. reuteri culture super-

natants, dough extracts, and minimum inhibitory con-

centration of reutericyclin

Antimicrobial activity was determined by using a

critical dilution assay on microtiter plates essentially

as previously described (Ganzle et al., 2000). Twofold

serial dilutions of the analyte in mMRS4 were inocu-

lated with the indicator strain L. sanfranciscensis to a

cell count of about 107 cfu ml � 1 and incubated at 30

jC for 16 h. In experiments where L. reuteri

LTH2584, TMW1.106, TMW1.112, or TMW1.656

were used as indicator strains, plates were incubated

at 37 jC. Growth of the indicator strain was judged by

measuring the optical density at 595 nm (OD595) and

the amount of analyte causing 50% growth inhibition

was defined as d50. In experiments where dough

extracts were used as analyte, it was verified by pH

measurements on the microtiter plates that growth

inhibition was not attributable to variations of pH.

The inhibitory activity was calculated as 1/d50 and

expressed as arbitrary units (AU) per milliliter. In

samples with a known reutericyclin concentration, the

reutericyclin concentration causing 50% growth

inhibition was defined as minimum inhibitory con-

centration (MIC). Results are shown as meansF stan-

standard deviation of triplicate experiments.

2.3. Purification of reutericyclin from cultures of L.

reuteri

(i) Analytical purification of reutericyclin from

cultures of L. reuteri LTH2584, TMW1.106,

TMW1.112, and TMW1.656. Purification of reuter-

icyclin was essentially performed as described previ-

ously (Ganzle et al., 2000). In short, cells from 250 ml

overnight cultures of L. reuteri strains in buffered

mMRS4 were harvested by centrifugation, washed

once with 50 mM phosphate buffer, pH 2.5, and

reutericyclin was extracted from the cells by stirring

the cells for 1 h in 100 ml 50 mM phosphate buffer,

pH 6.5 containing 30% (w/w) isopropanol (cell

extract). Cells were removed by centrifugation, and

NaCl was added to the supernatant until saturation.

The organic phase was removed, evaporated to dry-

Table 1

Microflora of SER sourdough at isolations of 1988, 1994, and 1998

Strain cell counts (cfu� 109/g)

1988a 1994b 1998c

L. reuteri

LTH2584d1.5 L. spec.

TMW1.104e1.0 L. frumenti 0.3

L. spec.

LTH3568e1.5 L. pontis

TMW1.107

1.0 L. amylovorus 1.3

L. mucosae 0.3 L. reuteri

TMW1.106

0.6 L. pontis 0.3

LTH3566d L. pontis

TMW1.108

0.3 L. reuteri

TMW1.656

0.2

L. reuteri

LTH3569

0.3 L. reuteri

TMW1.112

0.1

two other strains < 0.1 L. spec.

TMW1.102e0.1

six other

strains

< 0.1

a The results of the isolation performed in 1988 was published

in Bocker et al. (1995), and in Bocker (personal communication).b Muller, 1995.c Muller et al., 2000, 2001. The authors did not provide cell

counts for individual strains; only one strain of L. reuteri,

TMW1.656, was isolated from this batch of dough.d L. reuteri LTH2584 and L. mucosae LTH3566 produced

reutericyclin and an unknown inhibitory compound, respectively

(Ganzle et al., 1995, 2000).e L. spec. indicates species of Lactobacillus unknown at the

time of isolation.

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–45 33

ness, and solids resuspended in 2 ml isopropanol/

water (8:2). Reutericyclin was recovered from the

organic phase of this concentrated cell extract (1

ml). Concentrated cell extracts were analysed by

HPLC using a 250� 4.6 mm Luna 5A C18 column

(Phenomenex, Torrance, USA) coupled to an UV

detector. Samples were eluted with a solvent gradient

at a flow of 1 ml min� 1 (0 min: 50% acetonitrile,

50% H2O, 0.1% trifluoroacetic acid; 30 min: 100%

acetonitrile, 0.1% trifluoroacetic acid).

(ii) Preparative purification of reutericyclin from

cultures of L. reuteri LTH2584. Concentrated cell

extract from 2 l culture of L. reuteri LTH2584 was

prepared as described above and additionally purified

by gel filtration and hydrophobic interaction chroma-

tography as previously described (Ganzle et al.,

2000). This reutericyclin stock solution was stored

in isopropanol/water (8:2) at � 20 jC and used for all

experiments where reutericyclin was used as internal

or external standard. The concentration of this stock

solution was determined by HPLC using synthetic,

epimeric reutericyclin (EMC, Tubingen, Germany) as

calibration standard.

2.4. Determination of sugar fermentation patterns

Sugar fermentation patterns of L. reuteri strains

were determined by using the API 50 CH kit (Bio-

Merieux, Marcy l’Etoile, France) according to the

instructions of the manufacturer. Strains were incu-

bated at 37 jC for 48 h in modified API 50 CH

medium before the test strips were evaluated.

2.5. Randomly amplified polymorphic DNA (RAPD)

and restriction fragment length polymorphism (RFLP)

patterns

Chromosomal DNA from strains of L. reuteri was

prepared according to Lewinton et al. (1987). RAPD

patterns were generated according to Muller et al.

(2001) with the oligonucleotide primer M13V (5V-GTT TTC CCA GTC ACG AC-3V). The reaction

volume was 50 Al and contained the following: 5 Al10� reaction buffer, 3.5 mMMgCl2, 200 nM each of

the four deoxynucleotides, 1.5 U Taq polymerase (all

reagents from Amersham Pharmacia, Uppsala, Swe-

den), 20 pmol of the primer M13V, and 0.5 Algenomic DNA. The amplification reaction was carried

out in TopYield strips (Nunc, Denmark) using a

Mastercycler Gradient (Eppendorf, Germany) thermo-

cycler with the following cycling programme: 94 jC,3 min; 40 jC, 5 min; 72 jC, 5 min (2 cycles); 94 jC,1 min; 60 jC, 2 min; 72 jC, 3 min (32 cycles). PCR

fragments were electrophoretically separated on a 1%

agarose gel in 0.5� TBE buffer. DNAwas visualised

by staining with ethidium bromide and the gels were

documented with a digital camera and E.A.S.Y. soft-

ware (both Herolab, Griesheim, Germany). RAPD-

PCR was performed in duplicate.

RFLP patterns were made with EcoRI or PstI

digested chromosomal DNA. Chromosomal DNA

from L. reuteri strains was digested at 37 jC overnight

with 240 Uml�1 EcoRI (TaKaRa, Shiga, Japan) or 200

Uml�1PstI (Promega,Madison, USA). DigestedDNA

was separated by TBE agarose gel electrophoresis and

blotted on a nylon membrane (Hybond N+, Amersham

Pharmacia) by capillary blotting according to the

instructions of the manufacturer. A probe targeting

the 3Vhalf of the gene coding for 16S RNA in L. reuteri

was amplified by PCR with L. reuteri DSM20016T

chromosomal DNA as template and primers 616V

(5VAGA GTT TGATYM TGG CTC AG-3V, 3Vtermi-

nus of the primer located at position 31, Escherichia

coli numbering convention, Brosius et al., 1978) and

609RII (5VACTACY VGG GTATCTAAK CC-3V, 3Vterminus of the primer located at position 786, E. coli

numbering convention). The PCR product was labelled

and hybridised to the EcoRI and PstI digested DNA

using AlkPhos direct labelling and detection kit (Amer-

sham Pharmacia) according to the instructions of the

manufacturer.

2.6. Production of reutericyclin in wheat sourdoughs

2.6.1. Dough fermentation with L. reuteri and L.

sanfranciscensis

Wheat sourdoughs were prepared with 10 g white

wheat flour, 0.2 g NaCl and 9 g sterile tap water.

Sourdoughs were inoculated with L. reuteri LTH2584

or L. sanfranciscensis LTH2581. Cells from an over-

night culture of these strains in mMRS4 were washed

twice in sterile tap water, resuspended in water to a

cell count of about 2� 108 cfu ml� 1 and 1 ml of cell

suspension was added to the dough. Doughs were

mixed to homogeneity with a spatula and incubated at

34 jC. Growth and acid production were monitored

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–4534

by determination of viable cell counts and dough pH.

To distinguish between L. reuteri and L. sanfrancis-

censis in co-cultures, appropriate dilutions were plated

in quadruplicate and two plates each were incubated

for 24 h at 42 jC and for 24 h at 28 jC. After theseincubation times, large white colonies formed by L.

reuteri and no growth of L. sanfranciscensis was

observed at 42 jC, whereas at 28 jC, transparent

pinpoint colonies were formed by L. reuteri and large

white colonies were formed by L. sanfranciscensis.

The following fermentations were carried out: (i) L.

reuteri or L. sanfranciscensis in single culture. (ii) L.

reuteri and L. sanfranciscensis in co-culture. (iii) L.

sanfranciscensis in single culture with the addition of

reutericyclin stock solution to a concentration of 0,

1.8, 9.4, and 37.5 mg kg�1. Representative results of

three independent experiments are shown.

2.6.2. Extraction of reutericyclin from wheat sour-

doughs

Wheat sourdough was prepared as described

above, inoculated with L. reuteri LTH2584 or L.

sanfranciscensis as a negative control, and incubated

for 24 h at 34 jC. For extraction of reutericyclin, 8 g

of isopropanol were added to 40 g of dough, 12 g 50

mM histidine buffer (pH 5.85) and 5 M NaOH to

adjust the dough pH to 5.9F 0.05. Doughs were

extracted by stirring for 1 h at room temperature and

solids were removed by centrifugation at 15000� g

for 30 min at 0 jC and filtration through a 45-Amfilter. To the dough extracts were added 2 ml iso-

propanol and NaCl to saturation, the organic phase

was removed and evaporated to dryness in a rotary

evaporator. Solids were resuspended in isopropanol/

water (8:2) and the organic phase was collected to

obtain 0.300 ml concentrated dough extract for each

sample. Concentrated dough extracts and intermedi-

ates of the dough extraction protocol were analysed

by determination of their antimicrobial activity. To

obtain a biological assay specific for reutericyclin,

the inhibitory activity of dough extracts towards L.

sanfranciscensis LTH2581 (reutericyclin-sensitive)

was compared to the inhibitory activity towards L.

reuteri LTH2584 (reutericyclin-resistant). Reutericy-

clin inhibitory activities were converted to reuter-

icyclin concentrations by determination of the MIC

of reutericyclin stock solution on the same microtiter

plate. HPLC analysis of concentrated dough extracts

was carried out as described above but with a modi-

fied solvent gradient. Samples were eluted with a

solvent gradient at a flow of 1 ml min�1 (0 min:

75% acetonitrile, 25% H2O, 0.1% trifluoroacetic acid;

30 min: 100% acetonitrile, 0.1% trifluoroacetic acid).

Synthetic, epimeric reutericyclin was used as calibra-

tion standard. For each dough extraction, an extraction

was carried out in parallel which received 80 Agreutericyclin (concentration of 2 mg kg�1) as internal

standard just prior to the extraction step. The amounts

of reutericyclin recovered from L. sanfranciscensis

doughs (negative control) were used to determine the

recovery factor of the extraction procedure. Dough

fermentations were carried out in two independent

experiments and results are shown as meansF stan-

standard deviation.

2.7. Absorption of reutericyclin to gluten and starch

Reutericyclin was dissolved to a concentration of

240 mg l�1 in 50 mM histidine buffer, pH 5.85,

containing 24% (w/w) isopropanol. To this solution

were added 20 g l�1 soluble starch (Merck, Darm-

stadt, Germany) or 20 g l�1 gluten from wheat

(Sigma, Deisenhofen, Germany) and the suspensions

were incubated in a rotary shaker for 1 h. Solids were

removed by centrifugation and the reutericyclin con-

centration in the supernatant was determined by

HPLC. The experiment was carried out in triplicate.

3. Results

3.1. Reutericyclin production by L. reuteri strains

isolated from SER sourdough

The microflora of SER sourdough, a sourdough

propagated by repeated inoculation for at least 20

years, was monitored in 1989, 1994, and 1998 (Table

1). Although major shifts in the microflora are appa-

rent between these isolations, L. reuteri strains were

isolated each time. Strain L. reuteri LTH2584 is

known to produce reutericyclin and it was evaluated

whether reutericyclin-producing strains were still

present in dough in 1994 and 1998. Fig. 1 shows

the antimicrobial activity of four of these L. reuteri

isolates towards L. reuteri LTH2584 and L. sanfran-

ciscensis LTHB2581. The culture supernatant of all

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–45 35

four strains was strongly inhibitory to L. sanfrancis-

censis, whereas growth of the reutericyclin-resistant L.

reuteri LTH2584 was not inhibited, indicating that

reutericyclin may be the inhibitory compound. It was

furthermore observed that the putative reutericyclin

producing SER isolates had a high resistance to

purified reutericyclin (Fig. 1). Strain TMW1.106 with

the lowest reutericyclin-resistance of the L. reuteri

SER isolates still exhibited a reutericyclin MIC of

26.9 mg l�1, exceeding that of L. sanfranciscensis by

a factor of 100. Other L. reuteri isolates and most

other lactic acid bacteria are inhibited by reutericyclin

concentrations ranging from 0.1 to 1 mg l�1 (Ganzle

et al., 2000).

To verify that the inhibitory activity of L. reuteri

strains isolated from SER sourdough is indeed attrib-

utable to reutericyclin, it was purified from cultures of

these strains according to previously established

methods (Ganzle et al., 2000). The purification

involved extraction of cells with 30% propanol (cell

extract), solvent extraction of this cell extraction, and

separation of reutericyclin from remaining contami-

nants by reversed phase chromatography. The yield of

the purification protocol for the four strains is shown

in Table 2. The identification of the active compounds

as reutericyclin was based on the HPLC retention

times (Fig. 2), the ratio of UV adsorbances at the

wavelengths of 210, 280, and 305 nm (data not

shown), and ratio of UV adsorbance to inhibitory

activity (MIC, Table 2). All relevant properties were

Fig. 1. Antimicrobial activity towards L. reuteri LTH2584 (reutericyclin-resistant) and L. sanfranciscensis LTH2581 (reutericyclin-sensitive),

and reutericyclin resistance of L. reuteri strains LTH2584, TMW1.106, TMW1.112, TMW1.656, and L. sanfranciscensis LTH2581. Panel A:

inhibitory activity of neutralised culture supernatants of L. reuteri strains and L. sanfranciscensis LTH2581 towards L. sanfranciscensis

LTH2581 (black bars) and towards L. reuteri LTH2584 (white bars). Panel B: minimal inhibitory concentration of reutericyclin towards L.

reuteri strains and L. sanfranciscensis LTH2581. Results indicate meansF standard deviation of three independent experiments.

Table 2

Purification of reutericyclin from cultures of L. reuteri strains

LTH2584, TMW1.106, TMW1.112 and TMW1.656

L. reuteri strain

activity recovered

during purification

LTH2584 TMW1.106 TMW1.112 TMW1.656

Cell extract

(AU� 10�3)

15.2 8.1 9.5 26.1

Concentrated

cell extract

(AU� 10�3)

6.8 3.8 6.4 25.6

Purification

yield (%)

44% 47% 67% 98%

Purification

yield (mg)a2.9 1.0 2.0 8.7

MIC reutericyclin

(mg l�1)b0.18 0.20 0.25 0.19

a The reutericyclin concentration in concentrated cell extract

was determined by HPLC.b The minimum inhibitory concentration was determined using

L. sanfranciscensis LTH2581 as indicator strain.

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–4536

in agreement with that of synthetic reutericyclin

(chemical properties) and reutericyclin from strain

LTH2584. Table 2 furthermore shows that the yield

of the purification protocol was 40% or higher for all

strains. Taking into account losses during liquid han-

dling and the accuracy of the assay for determination

of antimicrobial activity, this result indicates that

reutericyclin is the major inhibitory compound of

these strains as previously shown for L. reuteri

LTH2584 (Ganzle et al., 2000). Differences were

observed with respect to the amounts of reutericyclin

produced by various strains. Strain LTH2584 and

TMW1.656 were more effective reutericyclin pro-

ducers than TMW1.106 and TMW1.112, however,

because the cultures from which reutericyclin was

isolated were not characterised with respect to growth

kinetics and cell density, this may reflect experimental

error.

3.2. Differentiation of L. reuteri SER isolates by

physiological properties and molecular typing

Differentiation of four L. reuteri isolates from SER

on strain level was performed to determine whether

these isolates are different strains dominating the SER

fermentation at different times between 1988 and

1998, or whether these isolates are clones of one

single strain persisting over 10 years. The strain

TMW1.112 could be distinguished from the three

other isolates on the basis of the colony morphology

on mMRS4 agar. Sugar fermentation patterns were

determined with the API CH50 system. Strains

LTH2584, TMW1.106 and TMW1.656 utilised

ribose, xylose, galactose, glucose, maltose, lactose,

melibiose, and sucrose. Strain TMW1.112 utilised all

of these sugars and additionally gluconate. L. reuteri

TMW1.112 was thus differentiated from the other

isolates on the basis of physiological properties.

RAPD patterns and RFLP patterns were generated

to further distinguish strains on molecular level. The

SER isolates were compared to one other sourdough

isolate and the L. reuteri type strain. The RAPD

patterns are shown in Fig. 3. The RAPD patterns

consisted of up to 10 main fragments sized between

0.3 and 4 kbp. Only three fragments were common for

all L. reuteri strains and SER isolates were clearly

distinguished on the basis of RAPD patterns from

the DSM type strain and the sourdough isolate

TMW1.614. Strains TMW1.106 and TMW1.656

exhibited identical RAPD patterns. Reproducible dif-

ferences were observed between these two strains and

the patterns of LTH2584 and TMW1.112. However, all

main fragments amplified from DNA of strains

TMW1.656 or TMW1.106 were also amplified from

DNA of strain LTH2584, and 8 of these 10 fragments

were obtained with DNA from strain TMW1.112. The

RFLP patterns of L. reuteri strains obtained with PstI

digested DNA are shown in Fig. 4. Digestion of

chromosomal DNA with PstI and hybridisation with

a 16S rDNA targeting probe produced up to five

distinct bands ranging from 1 kbp to about 5 kbp.

Virtually identical patterns were obtained with DNA

from strains TMW1.656 and TMW1.106 and all other

strains exhibited different patterns. No single band was

observed in all L. reuteri strains. However, two bands

were common to all SER isolates and four bands were

common in strains LTH2584 and TMW1.106/

Fig. 2. Chromatograms of concentrated cell extracts prepared from

L. reuteri LTH2584, TMW1.106, TMW1.112 and TMW1.656 on a

reversed-phase C18 HPLC column. Synthetic reutericyclin eluted at

35 min; chromatograms were offset by 22 AU.

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–45 37

Fig. 4. RFLP fingerprints of L. reuteri chromosomal DNA digested with PstI: lane 1, LTH2584; lane 2, TMW1.106; lane 3, TMW1.112; lane 4,

TMW1.656; lane 5, TMW1.614; lane 6, TMW1.693.

Fig. 3. RAPD patterns of L. reuteri strains. Lanes from left to right: lane 1, TMW1.693; lane 2, TMW1.614; lane 3, TMW1.656; lane 4,

TMW1.112; lane 5, TMW1.106; lane 6, molecular weight marker; lane 7, LTH2584. The result is representative of two independent

experiments.

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–4538

TMW1.656. The PstI RFLP thus supported the results

of the RAPD patterns. EcoRI digestion and hybrid-

isation did not provide as much differentiation as PstI

digestion and yielded only two bands of about 4–6 kbp

for all strains (data not shown). The patterns of strains

TMW1.656 and 1.106 were identical but different from

the other strains. Because these two strains were iso-

lated at different times from a continuously propagated

sourdough, it can be concluded that these are two

isolates of a single strain.

3.3. Production of reutericyclin in sourdough

3.3.1. Dough fermentation with L. reuteri and L.

sanfranciscensis

Reutericyclin formation in sourdough was first

evaluated by sourdough fermentations with reuter-

icyclin-producing L. reuteri in co-culture with the

reutericyclin-sensitive L. sanfranciscensis. In single

culture, each strain grew to cell counts of greater 109

cfu ml�1 within 10 h, corresponding to a drop in

dough pH from 6.5 to 3.5. In doughs inoculated with

L. reuteri and L. sanfranciscensis in co-culture, both

strains initiated exponential growth in the first 6 h of

fermentation. However, L. sanfranciscensis ceased to

grow after 6 h and cell counts were reduced by one

log cycle within the next 24 h. Because the acid

production and the corresponding pH drop in L.

reuteri and L. sanfranciscensis sourdoughs was com-

parable and thus might not account for the decrease of

L. sanfranciscensis cell counts, this growth arrest and

inactivation are attributable to reutericyclin.

To estimate the MIC of reutericyclin in sourdough,

reutericyclin was added to dough inoculated with L.

sanfranciscensis to concentrations ranging from 0 to

37.5 mg kg�1 and growth of the indicator strain and

the pH were monitored (Fig. 6). Addition of 1.8 mg

kg�1 reutericyclin resulted in a slight but significant

growth inhibition, and reutericyclin in concentrations

of 9.4 and 37.5 mg kg�1 reduced L. sanfranciscensis

cell counts by 0.2 and 1 log cycle, respectively,

already at the onset of fermentation. In these two

doughs, L. sanfranciscensis reached cell counts of less

than 108 cfu mg�1, and the dough pH did not drop

below 4.0. Using the same criterion for MIC deter-

mination as it was used for the MIC in liquid media—

50% growth inhibition after 16 h of growth—the

reutericyclin MIC in dough can be estimated to range

Fig. 5. Growth and acid production of L. reuteri LTH2584 and L. sanfranciscensis in single culture and in co-culture during sourdough

fermentation. Shown are the cell counts of L. reuteri (.) and L. sanfranciscensis (z) in single culture (open symbols) and in co-culture (closed

symbols). Grey symbols indicate the pH of doughs inoculated with L. reuteri ( ), L. sanfranciscensis ( ) or both strains ( ). The result is

representative of three independent experiments.

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–45 39

from 2 to 10 mg kg�1, exceeding the MIC in mMRS4

by a factor of more than 10. Based on comparison of

the L. sanfranciscensis growth curves presented in

Figs. 5 and 6, it can be estimated that L. reuteri-

produced reutericyclin in dough in a concentration

ranging from 1 to 10 mg kg�1.

Fig. 6. Growth and acid production of L. sanfranciscensis in wheat sourdoughs in the presence or absence of reutericyclin. Doughs were

inoculated with 5� 106 cfu ml�1 L. sanfranciscensis and reutericyclin was added to a concentration of 0 (.), 1.8 mg kg�1 (n), 9.4 mg kg�1

(E), and 37.5 mg kg�1 (z) at the onset of fermentation. Closed symbols indicate cell counts, open symbols indicate the pH of the dough. Data

are representative of two independent experiments.

Fig. 7. Chromatograms of reutericyclin extracted from doughs fermented with L. sanfranciscensis and L. reuteri on a reversed-phase C18 HPLC

column. Each dough was extracted twice, with and without addition of reutericyclin (RTC) as internal standard during dough extraction.

Synthetic reutericyclin eluted at 32.5 min and chromatograms were offset by 4 AU.

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–4540

3.3.2. Extraction of reutericyclin from wheat sour-

doughs

In order to obtain direct evidence for reutericyclin

production in sourdough fermentations, reutericyclin

was extracted from doughs fermented with L. reuteri.

L. sanfranciscensis sourdoughs were used as negative

control. Reutericyclin extraction was carried out from

each dough with or without reutericyclin addition as

internal standard. Extraction with aqueous solvents or

butanol failed to recover reutericyclin from the

doughs (data not shown). Reutericyclin could be

extracted at neutral pH with histidine buffer contain-

ing 40% isopropanol. Reutericyclin in dough extracts

was identified and quantified with a HPLC assay as

well as the bioassay for antimicrobial activity, using

reutericyclin-resistant L. reuteri and reutericyclin-sen-

sitive L. sanfranciscensis as indicator strains. Fig. 7

shows the chromatograms of the four extracts and

Table 3 shows the reutericyclin concentrations. No

inhibitory activity was extracted from doughs fer-

mented with L. sanfranciscensis. Dough extracts from

either L. sanfranciscensis or L. reuteri sourdoughs

exhibited no inhibitory activity towards L. reuteri

LTH2584 (data not shown), indicating that antimicro-

bial activity recovered from L. reuteri doughs is

attributable to reutericyclin. The reutericyclin concen-

trations determined by HPLC and the bioassay dif-

fered significantly. The reutericyclin concentration in

L. sanfranciscensis dough with internal standard

added to a level of 2 mg kg�1 just prior to extraction

was estimated as 0.2F 0.1 and 0.5F 0.03 mg kg�1

with the bioassay and HPLC, respectively. This dis-

crepancy indicates that only 25% of the reutericyclin

present in dough was recovered with the extraction

procedure. Additionally, compounds were extracted

which interfere with reutericyclin inhibitory activity.

Using the L. sanfranciscensis doughs with internal

standard as reference, reutericyclin concentrations in

L. reuteri doughs were corrected for the reutericyclin

recovery of the extraction procedure and a concen-

tration of 5.2F 1.5 and 4.3F 1.2 mg kg�1 was

determined with the HPLC and bioassay, respectively.

Correspondingly, a concentration of 6.2 mg kg�1 was

determined for L. reuteri doughs which additionally

received 2 mg kg�1 reutericyclin as internal standard

prior to dough extraction. This reutericyclin concen-

tration is furthermore consistent with the growth

inhibition of L. sanfranciscensis in dough by exter-

nally added reutericyclin (Fig. 6).

3.4. Absorption of reutericyclin to gluten

To provide a rationale for the high MIC of reuter-

icyclin in wheat doughs, and for the poor recovery

during the extraction procedure, it was evaluated

whether reutericyclin specifically binds to dough

components. Gluten and starch were suspended in a

240 mg l�1 reutericyclin solution in His/Propanol

buffer. After incubation for 1 h, undissolved starch

and gluten were removed by centrifugation and the

reutericyclin concentration was determined by HPLC.

The reutericyclin concentration in buffer determined

by HPLC was 236F 5 mg l�1 in the presence or

absence of starch, arguing against reutericyclin bind-

ing to starch. However, addition of gluten to the

solution reduced the reutericyclin concentration to

166F 7 mg l�1, indicating substantial binding of

reutericyclin to the suspended gluten polymer. It was

furthermore observed that a part of the gluten was

solubilised by the His/POH buffer, which may

account for the reduced inhibitory activity of reuter-

icyclin in dough extracts.

4. Discussion

This study provided evidence for production of the

antibiotic reutericyclin in active concentrations in

Table 3

Reutericyclin concentration in wheat sourdoughs estimated by the

biological assay and HPLC determination

Reutericyclin (mg/kg)

HPLC assay Bioassay

Control 0F 0 0F 0

Control + internal standarda 0.5F 0.03 0.2F 0.1

L. reuteri 1.1F 0.1 0.3F 0.1

L. reuteri + internal standarda 1.4F 0.2 0.5F 0.1

Concentrations corrected for recovery of internal standardb

L. reuteri 5.2F 1.5 4.3F 1.2

L. reuteri + internal standarda 6.3F 0.4 6.2F 2.2

a Reutericyclin (2 mg kg�1) was added to the doughs prior to

the extraction procedure.b The recovery of reutericyclin from L. sanfranciscensis doughs

with internal standard were used as reference.

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–45 41

sourdough fermentation. The results furthermore pro-

vided circumstantial evidence that reutericyclin pro-

duction provides a competitive advantage to the

producer strains and contributed to the stable persis-

tence of L. reuteri in an industrial sourdough fermen-

tation over a period of 10 years.

Reutericyclin production in dough to was shown in

this work by a challenge test demonstrating growth

inhibition of a reutericyclin-sensitive indicator strain

in sourdough, and by HPLC after dough extraction.

The growth inhibition of L. sanfranciscensis by L.

reuteri observed in this work is mainly attributable to

reutericyclin formation. Antimicrobial compounds

other than reutericyclin, e.g. reuterin (Axelsson et

al., 1989) or bacteriocins, do not contribute to the

inhibitory activity of L. reuteri LTH2584 (Ganzle et

al., 2000). Based on numerous studies on the growth

of L. sanfranciscensis LTH2581 in sourdough and

laboratory media (Bocker et al., 1990; Ganzle et al.,

1998; Thiele et al., 2002; Korakli et al., 2001) it can

be excluded that the observed levels of organic acids

produced by L. reuteri cause inactivation of L. san-

franciscensis LTH2581 in sourdough. Previous

attempts to demonstrate the formation of inhibitors

by lactic acid bacteria during food fermentations have

relied solely on bioassays because of the lack of

suitable extraction methods (Goff et al., 1996; Ryan

et al., 1996). The amount of reutericyclin produced in

dough was in the same range as that produced in

laboratory media containing Tween 80 (Ganzle et al.,

2000; this study). Addition of emulsifiers such as

Tween 80 to mMRS4 left the amount of reutericyclin

associated with the producer cells did not change but

enhanced the amount dissolved in the medium com-

pared to media with other sources of fatty acids

(Ganzle et al., 2000). The inhibitory concentration

of reutericyclin in dough was greater by a factor of

about 10 compared to mMRS4, indicating that only a

part of reutericyclin produced in dough is biologically

active. Adsorption of reutericyclin to gluten and

possibly other dough components are likely to

account for this effect. There is no indication of

appreciable reutericyclin degradation by microbial or

cereal enzymes during dough fermentation. Reuter-

icyclin formation by L. reuteri follows a primary

metabolite kinetics (Ganzle et al., 2000) but was

extracted in relevant concentrations from late station-

ary cultures of L. reuteri in dough.

In the challenge test presented here, L. reuteri fully

inhibited growth of L. sanfranciscensis when grown

in co-culture in sourdoughs. In contrast, in the prac-

tical situation of the SER fermentation, several other

strains of lactobacilli were isolated in high cell counts

from the same dough as reutericyclin-producing L.

reuteri strains. It was previously shown that those

strains isolated with L. reuteri LTH2584 from SER in

1988 have a two- to fourfold higher tolerance to

reutericyclin than L. sanfranciscensis strains and most

other lactobacilli (Ganzle et al., 2000). This finding

indicates that tolerance to reutericyclin may evolve

upon repeated exposure and highlights that factors

other than production of antimicrobials—adaptation

to substrates available in dough, and growth require-

ments with respect to pH, temperature, and ionic

strength—are of prime importance in sourdough

ecology.

Reutericyclin production during food fermentation

as well as suitable means for its quantification in

foods are prerequisite for the utilisation of this com-

pound as natural preservative. A possible application

of reutericyclin in food is the inhibition of rope-

forming bacilli in bread. Bread spoilage by rope-

forming bacilli remains a problem in bread production

(Rocken and Spicher, 1993; Rosenkvist and Hansen,

1995) and bacteriocins of lactic acid bacteria were

ineffective against these spoilage organisms in bread

(Rosenquist and Hansen, 1998). Rope-forming strains

of Bacillus subtilis are as sensitive to reutericyclin as

L. sanfranciscensis (Ganzle et al., 2000) and reuter-

icyclin concentrations in dough as they were observed

in this work could be expected to inhibit growth of B.

subtilis in bread. It remains unknown, however,

whether reutericyclin remains active after baking of

the dough. During bread baking, the crumb is heated

to 98 jC for about 1 h and data on the heat stability of

reutericyclin is lacking.

The SER L. reuteri isolates were characterised on

species level at the time of isolation (Bocker et al.,

1995; Muller et al., 2001). To determine the identity

or diversity of the isolates on strain level, a poly-

phasic approach was used including physiological

properties, RAPD and RFLP patterns. RAPD patterns

differentiate isolates of lactic acid bacteria below

species level. Screening of large strain collections of

L. sakei and L. curvatus (Berthier and Ehrlich, 1999)

as well as L. sanfranciscensis (Zapparoli et al., 1998)

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–4542

revealed that different RAPD patterns were generally

obtained from different isolates although not all

strains were discriminated by their RAPD patterns.

The primers used in this work for strain differentia-

tion by RAPD patterns discriminated on subspecies

level strains of lactobacilli isolated from duck intes-

tines and cereal fermentations (Kurzak et al., 1998;

Muller et al., 2001). Ribotyping of lactobacilli pro-

vides discrimination of lactobacilli including L. reu-

teri, on strain level (Barney et al., 2001; Stahl et al.,

1994; Zhong et al., 1998; Rodtong and Tannock,

1993).

The strain TMW1.112 was discriminated from

other isolates based on the criteria sugar fermentation

patterns and colony morphology as well as the RAPD

and RFLP patterns. The strain LTH 2584 was dis-

criminated from the three other strains by means of

molecular typing. The isolates TMW1.106 and

TMW1.656 were found to be identical in all methods.

The strain TMW1.656 produced more reutericyclin

upon incubation mMRS4 than strain TMW1.106,

however, experiments under more standardised con-

ditions are required to consider this result as criterion

for strain differentiation. Because 4 years or about

20 000 generations of bacterial growth elapsed

between the isolation of strains TMW1.106 and

TMW1.656, this result provides evidence for the

stable persistence of a single strain over a long period

of continuous sourdough propagation. The differen-

ces in RAPD and RFLP patterns between strains

TMW1.106/TMW1.656 and LTH2584 may indicate

the presence of unrelated strains at the various iso-

lation times. However, monitoring of growth of

defined strains of lactobacilli in gnotobiotic mice

demonstrated that RFLP patterns could change within

22 month (Rodtong and Tannock, 1993). Because

strains TMW1.106/TMW656 and LTH2584 were vir-

tually identical with respect to physiological proper-

ties, and exhibited highly similar RAPD and RFLP

patterns, we favour the interpretation that these differ-

ences reflect minor rearrangements on the chromo-

some after a few thousand generations in the same

strain rather than the presence of a new strain.

Independent of the question of diversity or identity

of strain LTH2584, TMW1.106, and TMW1.656, it

was shown that reutericyclin-producing strains per-

sisted in a traditionally prepared sourdough over 10

years or 50000 generations of microbial growth. It

can be thus concluded that reutericyclin production

contributed to the stable persistence of L. reuteri

strains in dough.

Acknowledgements

Monika Hadek is acknowledged for excellent

technical assistance, Martin Muller and Matthias

Ehrman for helpful discussions and support during

the work, and their careful revision of the manuscript.

References

Armero, E., Collar, C., 1996. Antistaling additives, flour type and

sourdough process effects on functionality of wheat doughs. J.

Food Sci. 61, 299–303.

Armero, E., Collar, C., 1998. Crumb firming kinetics of wheat

breads with anti-staling additives. J. Cereal. Sci. 28, 165–174.

Axelsson, L.T., Chung, T.C., Dobrogosz, W.J., Lindgren, S.E.,

1989. Production of a bread spectrum antimicrobial substance

by Lactobacillus reuteri. Microb. Ecol. Health Dis. 2, 131–

136.

Barney, M., Volgyi, A., Navarro, A., Ryder, D., 2001. Riboprinting

and 16S rRNA gene sequencing for identification of brewery

Pediococcus isolates. Appl. Environ. Microbiol. 67, 553–560.

Berthier, F., Ehrlich, S.D., 1999. Genetic diversity within Lactoba-

cillus sakei and Lactobacillus curvatus and design of PCR pri-

mers for its detection using randomly amplified polymorphic

DNA. Int. J. Syst. Bacteriol. 49, 977–1007.

Bocker, G., Ernst Bocker GmbH & Co KG., Minden., personal

communication.

Bocker, G., Vogel, R.F., Hammes, W.P., 1990. Lactobacillus san-

francisco als stabiles Element in einem Reinzucht–Sauerteig–

Praparat. Getreide Mehl Brot 44, 269–274.

Bocker, G., Stolz, P., Hammes, W.P., 1995. Neue Erkenntnisse zum

Okosystem Sauerteig und zur Physiologie der sauerteigsypi-

schen Stamme Lactobacillus sanfrancisco und Lactobacillus

pontis. Getreide Mehl Brot 49, 370–374.

Brosius, J., Palmer, J.L., Kennedy, J.F., Noller, H.F., 1978. Complete

nucleotide sequence of a 16S ribosomal RNA gene from Escher-

ichia coli. Proc. Natl. Acad. Sci. U. S. A. 75, 4801–4805.

Corsetti, A., Gobbetti, M., Smacchi, E., 1996. Antibacterial activity

of sourdough lactic acid bacteria: isolation of a bacteriocin-like

inhibitory substance from Lactobacillus sanfrancisco C57. Food

Microbiol. 13, 447–456.

Corsetti, A., Gobbetti, M., de Marco, B., Balestrieri, F., Paoletti, F.,

Russi, L., Rossi, J., 2000. Combined effect of sourdough lactic

acid bacteria and additives on bread firmness and staling. J.

Agric. Food Chem. 48, 3044–3051.

Fretzdorff, B., Brummer, J.-M., 1992. Reduction of phytic acid

during breadmaking of whole-meal bread. Cereal Chem. 69,

266–270.

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–45 43

Ganzle, M.G., Hertel, C., Hammes, W.P., 1995. Antimicrobial ac-

tivity in lactobacilli from sourdough. Beijerinck Centennial. Mi-

crobiol Physiology and Gene Regulation: Emerging Principles

and Applications. 10–14.12.1995. Delft Univ. Press, Delft, pp.

380–381.

Ganzle, M.G., Ehmann, M., Hammes, W.P., 1998. Modelling of

growth of Lactobacillus sanfranciscensis and Candida milleri

in response to process parameters of the sourdough fermenta-

tion. Appl. Environ. Microbiol. 64, 2616–2623.

Ganzle, M.G., Hoeltzel, A., Walter, J., Jung, G., Hammes, W.P.,

2000. Characterization of reutericyclin produced by Lactobacil-

lus reuteri LTH2584. Appl. Environ. Microbiol. 66, 4325–

4333.

Gobbetti, M., Smacchi, E., Corsetti, A., 1996. The proteolytic sys-

tem of Lactobacillus sanfrancisco CB1: purification and char-

acterization of a proteinase, a dipeptidase, and an amino-

peptidase. Appl. Environ. Microbiol. 62, 3220–3226.

Goff, J.H., Bhunia, A.K., Johnson, M.G., 1996. Complete inhibition

of low levels of Listeria monocytogenes on refrigerated chicken

meat with pediocin AcH bound to heat-killed Pediococcus acid-

ilactici cells. J. Food Prot. 59, 1187–1192.

Hamad, S.H., Dieng, M.C., Ehrmann, M.A., Vogel, R.F., 1997.

Characterization of the bacterial flora of Sudanese sorghum

flour and sorghum sourdough. J. Appl. Microbiol. 83, 764–

770.

Hammes, W.P., Ganzle, M.G., 1997. Sourdough breads and related

products. In: Wood, B.J.B. (Ed.), Microbiology of Fermented

Food. Chapman & Hall, London, pp. 199–216.

Hansen, A., Hansen, B., 1996. Flavour of sourdough wheat bread

crumb. Z. Lebensm.-Unters. -Forsch. 202, 244–249.

Holtzel, A., Ganzle, M.G., Nicholson, G.J., Hammes, W.P., Jung,

G., 2000. The first low-molecular-weight antibiotic from lactic

acid bacteria: reutericyclin, a new tetramic acid. Angew. Chem.,

Int. Ed. 39, 2766–2768.

Korakli, M., Rossmann, A., Ganzle, M.G., Vogel, R.F., 2001. Su-

crose metabolism and exopolysaccharide production in wheat

and rye sourdoughs by Lactobacillus sanfranciscensis. J. Agric.

Food. Chem. 49, 5194–5200.

Kratochvil, J., Holas, J., 1984. Untersuchungen uber proteolytische

Vorgange in Roggensauerteig. Getreide Mehl Brot 38, 330–

332.

Kurzak, P., Ehrmann, M.A., Vogel, R.F., 1998. Diversity of lactic

acid bacteria associated with ducks. Syst. Appl. Microbiol. 21,

588–592.

Lewinton, J., Greenaway, S.D., Spillane, B.J., 1987. Rapid small

scale preparations of bacterial genomic DNA suitable for clon-

ing and hybridization analysis. Lett. Appl. Microbiol. 5, 51–53.

Martinez-Anaya, M.A., Devesa, A., 2000. Influence of enzymes in

sourdough wheat breadmaking. Changes in pentosans. Food Sci.

Technol., Int. 6, 109–116.

Muller, M., 1995. Indentifizierung von Laktobazillen aus einem

Roggensauerteig. Diploma thesis at the TU Munchen, Lehrstuhl

fur Technische Mikrobiologie (Prof. R.F. Vogel). Unpublished.

Muller, M.R.A., Ehrmann, M.A., Vogel, R.F., 2000. Lactobacillus

frumenti sp. nov., a new lactic acid bacterium isolated from rye-

bran fermentations with a long fermentation period. Int. J. Syst.

Evol. Microbiol. 50, 2127–2133.

Muller, M.R.A., Wolfrum, G., Stolz, P., Ehrmann, M.A., Vogel,

R.F., 2001. Monitoring the growth of Lactobacillus species dur-

ing a rye flour fermentation. Food Microbiol. 18, 217–227.

Ograbek, D., Brandt, M.J., Hammes, W.P., 1999. Arginine metab-

olism of Lactobacillus sanfranciscensis and L. pontis in sour-

dough. Poster, Presented at the XVII ICC Conference, Valencia,

June 1999.

Olsen, A., Halm, M., Jakobsen, M., 1995. The antimicrobial activity

of lactic acid bacteria from fermented maize (kenkey) and their

interaction during fermentation. J. Appl. Bacteriol. 79, 506–

512.

Rocken, W., Spicher, G., 1993. Fadenziehende Bakterien—Vorkom-

men, Bedeutung, Gegenmaßnahmen. Getreide Mehl Brot 47,

30–35.

Rodtong, S., Tannock, G.W., 1993. Differentiation of Lactobacillus

strains by ribotyping. Appl. Environ. Microbiol. 59, 3480–

3484.

Rosenkvist, H., Hansen, A., 1995. Contamination profiles and char-

acterisation of Bacillus species in wheat bread and raw materials

for bread production. Int. J. Food Microbiol. 26, 353–363.

Rosenquist, H., Hansen, A., 1998. The antimicrobial effect of or-

ganic acids, sour dough and nisin against Bacillus subtilis and B.

licheniformis isolated from wheat bread. J. Appl. Microbiol. 85,

621–631.

Rosenquist, H., Hansen, A., 2000. The microbial stability of two

bakery sourdoughs made from conventionally and organically

grown rye. Food Microbiol. 17, 241–250.

Ryan, M.P., Rea, M.C., Hill, C., Ross, R.P., 1996. An application in

cheddar cheese manufacture for a strain of Lactococcus lactis

producing a novel bread-spectrum bacteriocin, lacticin 3147.

Appl. Environ. Microbiol. 62, 612–619.

Spicher, G., Schroder, R., 1978. Die Mikroflora des Sauerteiges: IV.

Mitteilung: Untersuchungen uber die Art der in ‘‘Reinzucht-

sauern’’ anzutreffenden stabchenformigen Milchsaurebakterien

(Genus Lactobacillus Beijerinck). Z. Lebensm.-Unters. -Forsch.

167, 342–354.

Stahl, M., Pettersson, B., Molin, G., Uhlen, M., Arhne, S., 1994.

Restriction fragment length polymorphism of Lactobacillus reu-

teri and Lactobacillus fermentum, originating from intestinal

mucosa, based on 16S rRNA genes. Syst. Appl. Microbiol.

17, 108–115.

Stolz, P., Bocker, G., Vogel, R.F., Hammes, W.P., 1993. Utilisation

of maltose and glucose by lactobacilli isolated from sourdough.

FEMS Microbiol. Lett. 109, 237–242.

Stolz, P., Vogel, R.F., Hammes, W.P., 1995. Utilization of electron

acceptors by lactobacilli isolated from sourdough: I. Lactobacil-

lus sanfrancisco. Z. Lebensm.-Unters. -Forsch. 201, 91–96.

Stolz, P., Hammes, W.P., Vogel, R.F., 1996. Maltose-phosphorylase

and hexokinase activity in lactobacilli from traditionally pre-

pared sourdoughs. Adv. Food Sci. 18, 1–6.

Tangkongchitr, U., Seib, P.A., Hoseney, R.C., 1982. Phytic acid: III.

Two barriers to the loss of phytate during breadmaking. Cereal

Chem. 59, 216–221.

Thiele, C., Ganzle, M.G., Vogel, R.F., 2002. Contribution of sour-

dough lactobacilli, yeast, and cereal enzymes to the generation

of amino acids in dough relevant for bread flavor. Cereal Chem.

79, 45–51.

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–4544

Vogel, R.F., Knorr, R., Muller, M.R.A., Steudel, U., Ganzle, M.G.,

Ehrmann, M.A., 1999. Non-dairy lactic fermentations: the ce-

real world. Antonie van Leeuwenhoek 76, 403–411.

Zapparoli, G., Torriani, S., Dellaglio, F., 1998. Differentiation of

Lactobacillus sanfranciscensis strains by randomly amplified

polymorphic DNA and pulsed-field gel electrophoresis. FEMS

Microbiol. Lett. 166, 325–332.

Zhong, W., Millsap, K., Bialkowska-Hobrzanska, H., Reid, G.,

1998. Differentiation of Lactobacillus species by molecular typ-

ing. Appl. Environ. Microbiol. 64, 2418–2423.

M.G. Ganzle, R.F. Vogel / International Journal of Food Microbiology 80 (2002) 31–45 45