Presence of sourdough lactic acid bacteria in commercial total mixed ration silage as revealed by denaturing gradient gel electrophoresis analysis


Naoki Nishino, Department of Animal Science, Graduate School of Natural Science and Technology, Okayama University, Okayama 700-8530, Japan. E-mail:


Aims:  To characterize the bacterial communities in commercial total mixed ration (TMR) silage, which is known to have a long bunk life after silo opening.

Methods and Results:  Samples were collected from four factories that produce TMR silage according to their own recipes. Three factories were sampled three times at 1-month intervals during the summer to characterize the differences between factories; one factory was sampled 12 times, three samples each during the summer, autumn, winter and spring, to determine seasonal changes. Bacterial communities were determined by culture-independent denaturing gradient gel electrophoresis. All silages contained lactic acid as the predominant acid, and the contents appeared stable regardless of factories and product seasons. Acetic acid and 1-propanol contents were different between factories and indicated seasonal changes, with increases in warm seasons compared to cool seasons. Both differences and similarities existed among the bacterial communities from each factory and product season. Lactobacillus parabuchneri was found in the products from three of four factories. Various sourdough lactic acid bacteria (LAB) were identified in commercial TMR silage; Lactobacillus panis, Lactobacillus hammesii, Lactobacillus mindensis, Lactobacillus pontis, Lactobacillus frumenti and Lactobacillus farciminis were detected in many products. Moreover, changes owing to product season were distinctive, and Lact. pontis and Lact. frumenti became detectable in summer products.

Conclusion:  Sourdough LAB are involved in the ensiling of commercial TMR silage. Silage bacterial communities vary more by season than by factory. The LAB species Lact. parabuchneri was detected in the TMR silage but may not be essential to the product’s long bunk life after silo opening.

Significance and Impact of the Study:  Commercial TMR silage resembles sourdough with respect to bacterial communities and long shelf life. The roles of sourdough LAB in the ensiling process and aerobic stability are worth examining.


Many by-products from the food and beverage industries are used to produce animal feed. Most of these by-products have high moisture contents and lack fermentable substrates for ensiling without treatment. Desirable preservation can be achieved if sugars or sugar-rich materials are added at ensiling, but this practice may yet be insufficient to balance the nutritional components required for food-producing animals.

Feeding value can be upgraded to a targeted composition if many ingredients are used to create a total mixed ration (TMR). The production of TMR silage has been proven useful in Japan to incorporate various by-products into rations. In general, more than ten ingredients are mixed to prepare TMR silage. Wet brewers grains (BG) and soybean curd residue (SC) are often used as the main ingredients and are blended with lower proportions of dry feeds such as grass hay, legume hay, cracked maize, wheat bran and beet pulp. Molasses is usually added to fortify fermentable sugars. The silage product is commercialized principally in the form of a portable (300–400 kg) bag silo. Most manufacturers do not have sufficient storage area for large quantities of silage, so they prefer short-term storage to facilitate product shipments.

Although aerobic deterioration occurs within 2 days when BG and SC are ensiled alone, deterioration can be delayed for as long as 7 days when the by-product is stored as TMR silage (Wang and Nishino 2008). In previous studies, we isolated Lactobacillus buchneri as the predominant species of lactic acid bacteria (LAB) in BG-containing TMR silage and confirmed that inoculation with this species could suppress the spoilage of grass and maize silages (Nishino and Touno 2005; Wang and Nishino 2009). The presence of Lact. buchneri in BG-containing TMR silage was also verified by denaturing gradient gel electrophoresis (DGGE) analysis (Wang and Nishino 2009). However, product stability was nearly the same in SC-containing TMR silage without DGGE-detectable Lact. buchneri (Wang and Nishino 2008).

Because our previous studies were conducted in laboratory silos with BG and SC from specific factories, results might have been biased owing to the controlled silo size, factory, region and local environment. If Lact. buchneri was indigenous to the factories or local ingredients, the association with aerobic stability would not occur in TMR silage produced in other regions. In this study, therefore, we collected commercial TMR silage from different factories in Japan and evaluated the fermentation products and bacterial communities. Unlike typical crop silages, TMR silages are produced during all seasons; hence, storage properties may differ owing to varying storage temperatures. A total of 21 samples were collected from four factories, and the differences between factories and product seasons were characterized. Sourdough LAB were identified without any expectations, because they have never been isolated from crop silages.

Materials and methods


Samples were collected from four factories located in central (factories A and B) and southwest (factories C and D) parts of Japan. The samples from factories A, B and C were collected between June and August 2008, with three different product lots sampled at 1-month intervals. The samples from factory D were collected between January and November 2009, with four different product seasons sampled three times each during 1 week. A core sample (about 1 kg) was taken from a bag silo (300–400 kg), because our preliminary study indicated that DGGE profiles appeared the same among outer-top, inner-top, outer-bottom and inner-bottom samples.

Wet BG, soy sauce cake and maize gluten feed were main by-products used in the factories A, B and C, and dried distillers grain with soluble was used instead of soy sauce cake in the factory D. Total numbers of ingredients were 17, 12, 12 and 19 for the factories A, B, C and D, respectively. Storage periods varied from 12 to 55 days, and samples were shipped frozen to our laboratory. For factories A, B and C, mean daily minimum and maximum temperatures during the periods of silage production were 11 and 26°C, 17 and 32°C, and 19 and 33°C, respectively. For factory D, mean daily minimum and maximum temperatures were −2 and 10°C, 5 and 25°C, 20 and 30°C, and 5 and 22°C, respectively and correspond to samples collected in January–February (winter), April–May (spring), July–August (summer) and October–November (autumn), respectively.

Chemical analyses

Dry matter (DM) contents were determined by oven drying at 60°C for 48 h. Silage pH, lactic acid, short-chain fatty acids and alcohols were determined from water extracts. Lactic acid, acetic acid and ethanol contents were determined by an ion-exclusion polymeric HPLC method with refractive index detection (López-Tamames et al. 1996). A portion of the water extracts was passed through a 0·20-μm filter and 10 μl was injected into an ICSep COREGEL-87H column (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan) containing a cation exchange polymer in the hydrogen ionic form. The mobile phase was 0·004 mol l−1 sulphuric acid at a flow rate of 0·6 ml min−1 at 60°C. Because the refractive index detection was not sensitive enough to determine small amounts of alcoholic products, 1-propanol and 1,2-propanediol contents were analysed using gas chromatography (Parvin and Nishino 2009).

Denaturing gradient gel electrophoresis

DGGE was performed as previously described (Parvin and Nishino 2010). In brief, PCR was used to amplify a variable (V3) region of the bacterial 16S rRNA gene using the forward primer GC357f (5′-CGCCCGCCGCGCGCGGCGGGCGGGGCGGGGGCACGGGGGGCCTACGGGAGGCAGCAG-3′) and the reverse primer 517r (5′–ATTACCGCGGCTGCTGG–3′). The GC-clamp PCR products were separated according to their sequences with a DCode Universal Mutation Detection System (Bio-Rad Laboratories, Inc., Tokyo, Japan). The samples were applied directly onto 10% (w/v) polyacrylamide gels prepared in a denaturing gradient from 25 to 50% using urea and formamide (7 mol l−1 urea and 40% [v/v] formamide as 100% denaturants).

Cloning and sequencing of DGGE bands

Selected bands were excised from the DGGE gels, and each band was placed in 10 μl of sterilized water at 4°C overnight to allow DNA diffusion. Extracted DNA was amplified by PCR using 357f (without GC-clamp) and 517r primers, and the reaction products were purified using a commercial cleanup kit (GeneClean kit; Qbiogene, Carlsbad, CA, USA). Purified PCR products were cloned into the pTAC-1 vector, and the resulting plasmids were transformed into Escherichia coli DH5α competent cells (DynaExpress TA cloning kit; BioDynamics Laboratory Inc., Tokyo, Japan). Colony PCR was performed to confirm the inserts using primers for the vector (M13f and M13r). The sequencing reaction was carried out using a BigDye® Terminator v3.1 Cycle Sequencing kit (Applied Biosystems Inc., Foster City, CA, USA), and DNA sequences were analysed using an ABI PRISM® 310 sequencer (Applied Biosystems Inc.). BLAST searches of the GenBank database were performed to determine the closest relatives of partial 16S rRNA gene sequences.

Statistical analysis

Data were subjected to one-way analysis of variance, and differences owing to factories (2008 samples) and product seasons (2009 samples) were tested by Tukey’s multiple comparison. Analyses were conducted using jmp software (ver. 7; SAS Institute, Tokyo, Japan).


Effect of factories (2008 samples)

Mean pH values from three lots were 4·62, 4·20 and 3·99 in the A, B and C products, respectively, and pH was higher in the A products than in the other products (Table 1). Lactic acid contents varied from 11·8 to 16·4 g kg−1, with no difference found between the products. Acetic acid (3·96–6·95 g kg−1) and 1-propanol (1·04–3·30 g kg−1) contents were highest and lowest in the C and B products, respectively. Ethanol (1·47–2·61 g kg−1) and 1,2-propanediol (0·06–0·83 g kg−1) contents were not different between factories. Propionic and butyric acids were not detected in any products. Although the ratio of lactic acid to acetic acid was more than double in the A and B products compared to the C products, the differences did not reach significance owing to large variations within the A and B products.

Table 1.   Differences owing to product factories in fermentation products of commercial total mixed ration silage (2008 samples)
  1. Means for three separate samplings. Values in the same row with different superscript letters are significantly different (P < 0·05).

Dry matter (g kg−1)611a563b543b6·92
Lactic acid (g kg−1)16·413·811·82·44
Acetic acid (g kg−1)5·02b3·96b6·95a0·40
Ethanol (g kg−1)2·611·472·380·44
1-Propanol (g kg−1)1·32ab1·04b3·30a0·69
1,2-Propanediol (g kg−1)0·340·830·060·50
Lactic to acetic acid ratio3·463·461·700·35

Several common bacteria were detected in the A, B and C products (Fig. 1). Weissella paramesenteroides (bands 1 and 7), Lactobacillus mindensis (band 10), Lactobacillus panis (band 12), Lactobacillus parabuchneri (band 15) and Curtobacterium citreum (band 17) were found in at least one lot from each factory. Lactobacillus vaccinostercus (band 4) and Pediococcus damnosus (band 14) were found in the A and B products but were not detectable in the C products. The bands indicative of Lactobacillus fermentum (band 5), Lactobacillus hammesii (band 6) and Streptococcus spp. (band 11) were present only in the A products, and the band indicative of Lactobacillus frumenti (band 19) was found only in the B products. Lactobacillus acetotolerans (band 9) and Lact. panis (band 16) were detectable in the B and C products, and Pediococcus acidilactici (band 13) was found in the A and C products.

Figure 1.

 Bacterial communities in commercial total mixed ration silage collected from three factories with three separate samplings (2008 samples). (1) Weissella paramesenteroides; (2) W. paramesenteroides; (3) Lactobacillus plantarum; (4) Lactobacillus vaccinostercus; (5) Lactobacillus fermentum; (6) Lactobacillus hammesii; (7) W. paramesenteroides; (8) Lact. plantarum; (9) Lactobacillus acetotolerans; (10) Lactobacillus mindensis; (11) Streptococcus spp.; (12) Lactobacillus panis; (13) Pediococcus acidilactici; (14) Pediococcus damnosus; (15) Lactobacillus parabuchneri; (16) Lact. panis; (17) Curtobacterium citreum; (18) Lactobacillus oris and (19) Lactobacillus frumenti.

Effect of seasons (2009 samples)

The pH value (4·68) of winter products was higher than the pH values (c. 4·4) of products from the other seasons (Table 2). Although lactic acid contents (16·0–20·8 g kg−1) were stable regardless of product season, acetic acid (5·46–13·5 g kg−1) and 1-propanol (0·00–8·84 g kg−1) contents were higher in TMR silage produced under warm and hot conditions. The changes in ethanol contents (3·59–12·2 g kg−1) were the opposite owing to the product season; ethanol contents were highest in the winter products and lowest in the summer products. The 1,2-propanediol contents (0·13–1·36 g kg−1) were not affected by product season, and propionic and butyric acids were not detected in any products. Differences were found in the lactic acid to acetic acid ratio (1·32–2·95), with the highest and lowest values observed in winter and summer products, respectively.

Table 2.   Differences owing to product seasons in fermentation products of commercial total mixed ration silage (2009 samples)
 Winter (January–February)Spring (April–May)Summer (July–August)Autumn (October–November)SE
  1. Means for three separate samplings. Values in the same row with different superscript letters are significantly different (P < 0·05).

Dry matter (g kg−1)5535625515738·14
Lactic acid (g kg−1)16·020·617·820·81·62
Acetic acid (g kg−1)5·46c10·8ab13·5a7·50bc0·87
Ethanol (g kg−1)12·2a9·97a3·59b9·44a1·07
1-Propanol (g kg−1)0·00c4·89b8·84a1·50c0·64
1,2-Propanediol (g kg−1)0·131·360·190·820·34
Lactic to acetic acid raito2·95a1·98bc1·32c2·75ab0·22

Weissella paramesenteroides (band 2) and Lactobacillus brevis (band 6) were detected in all samples regardless of product season (Fig. 2). The band indicative of Lactobacillus plantarum (band 3) was clearly found in winter, spring and autumn products but appeared eliminated in two of three summer products. The band indicative of Lactobacillus pobuzihi (band 11) was also faint in summer products compared to products from the other seasons. Bands indicative of Lact. fermentum (band 4) and Pediococcus pentosaceus (band 8) were found in spring products, whereas those of Lactobacillus farciminis (band 9) were distinctive in winter products compared to products from the other seasons. Lactobacillus panis (band 10) was detected in all spring and summer products and in parts of winter and autumn products. Lactobacillus pontis (band 12) was found only in spring and summer products, and Lactobacillus vaginalis (band 13) and Lact. frumenti (band 14) were present only in summer products.

Figure 2.

 Bacterial communities in commercial total mixed ration silage collected from one factory during four seasons with three separate samplings each (2009 samples). (1) Weissella paramesenteroides; (2) W. paramesenteroides; (3) Lactobacillus plantarum; (4) Lactobacillus fermentum; (5) Lactobacillus brevis; (6) Lact. brevis; (7) Lact. brevis; (8) Pediococcus pentosaceus; (9) Lactobacillus farciminis; (10) Lactobacillus panis; (11) Lactobacillus pobuzihi; (12) Lactobacillus pontis; (13) Lactobacillus vaginalis and (14) Lactobacillus frumenti.


The production TMR silage has been practised in Japan to utilize wet by-products as ruminant feeds. Farmers have recognized that these products are highly resistant to aerobic deterioration, but an explanation for this property remains unclear. Because the silages were frozen and shipped immediately after sampling, no data were available on the stability in the presence of air. However, the silages could be regarded as stable as usual, because farmers and silage manufacturers did not see any distinctive properties on the products that were sampled. For DNA-based microbial community analysis, freezing is considered acceptable to the storage of environmental samples (Lauber et al. 2010).

Our previous studies using laboratory silos and wet by-products produced from specific factories demonstrated that Lact. buchneri may be involved in the ensiling process and aerobic stability (Wang and Nishino 2008) despite the fact that it was not found in the material used in the TMR mixture. During the progress of ensiling fermentation, Lact. buchneri becomes detectable by both plate culture and DGGE analysis, and small amounts of 1,2-propanediol also become evident (Wang and Nishino 2009). The DGGE profiles were similar between 14- and 56-days TMR silages (Wang and Nishino 2009); therefore, the variation of storage periods (from 12 to 55 days) may have little influence on the analyses of bacterial community in this study.

Our previous work showed that Lact. buchneri may occur more often in BG-containing silage than in SC-containing silage (Wang and Nishino 2008), but this LAB species was not confirmed in any samples in this study. Nevertheless, Lact. parabuchneri, a LAB species similar to Lact. buchneri in terms of spoilage inhibition and 1,2-propanediol production (Oude Elferink et al. 2001), was detected in 2008 samples regardless of their originating factories. Moreover, the detection of small amounts of 1-propanol, a 1,2-propanediol metabolite produced by Lactobacillus diolivorans (Krooneman et al. 2002), suggested that Lact. buchneri and Lact. parabuchneri had been activated in the TMR silages. Therefore, this study supports our findings that Lact. buchneri may be involved in the stability of TMR silage, although it remained undetectable by DGGE analysis.

Many heterofermentative LAB were identified in the TMR silages. W. paramesenteroides, Lact. parabuchneri and Lact. panis were detected in many of the 2008 samples, and W. paramesenteroides, Lact. brevis and Lact. panis were frequently detected in the 2009 samples. Weissella paramesenteroides can be found in crop silages (Cai et al. 1998), and its inability to inhibit aerobic deterioration has been demonstrated previously (Zhang et al. 2000). Lactobacillus brevis was also ineffective in suppressing spoilage after silo opening (Wang and Nishino 2009). Lactobacillus panis has never been isolated from crop silages; thus, no information is available on the involvement of this LAB species in the ensiling fermentation process and aerobic spoilage. Among other heterofermentative LAB found in this study (Lactobacillus vaccinostercus, Lact. fermentum, Lact. hammesii, Lactobacillus oris, Lact. frumenti and Lactobacillus vaginalis), only Lact. fermentum is usually found in silage. However, Lact. fermentum has been shown to have no inhibitory activity against aerobic deterioration (Wang and Nishino 2009).

Lactobacillus panis, Lact. hammesii, Lact. mindensis, Lact. frumenti, Lact. farciminis and Lact. pontis are typical sourdough LAB (Meroth et al. 2003; Corsetti and Settanni 2007; De Vuyst et al. 2009). Lactobacillus plantarum, Lact. fermentum and Lact. brevis are also often isolated from sourdough, but these LAB can be found in many fermented foods including silage. Commercial TMR silage resembles sourdough in terms of bacterial communities, and both fermented foods are known to have a long shelf life. Sourdough LAB have never been isolated from crop silages, and it is difficult to explain why TMR silage can harbour such LAB species. The presence of sourdough LAB was common to all four factories examined in this study. The use of cereal ingredients and the relatively low DM contents (500–600 g kg−1) in the TMR silage may account for the findings. Large proportions of wet BG (a residue obtained after extracting wort from mashed malted barley grains), maize and wheat bran are often used to produce TMR silage, and the DM contents of sourdough (Gül et al. 2005) are similar to those of TMR silage. The appearance of Lact. frumenti and Lact. pontis in spring and summer products (2009 samples) agreed with the findings of Meroth et al. (2003), who demonstrated that although Lactobacillus sanfranciscensis can dominate sourdough prepared at 25°C, Lact. panis, Lact. pontis and Lact. frumenti may become predominant in sourdough prepared at 30 and 40°C. The lack of Lact. panis in winter products (2009 samples) can thus be regarded as acceptable. However, Lact. sanfranciscensis, a representative LAB species in sourdough (Gobbetti 1998), was not detected even in the winter products.

The involvement of community changes in sourdough LAB with seasonal changes in fermentation products remains unknown. This study suggested that although lactic acid production may be stable regardless of storage temperature, ethanol production can increase and acetic acid and 1-propanol production will decrease if the TMR silage is produced at a low storage temperature. However, in a model experiment of sourdough production using Lact. sanfranciscensis, Lact. plantarum and Saccharomyces cerevisiae as a starter, ethanol and lactic acid contents increased and acetic acid content decreased when the temperature was raised from 25 to 35°C (Gobbetti et al. 1995). The present results are different from the model study; yeast activity, assumed by ethanol production, appeared to diminish in the TMR silage at warm rather than cool ambient temperatures, although the reason for this observation is unclear. The detection of Lact. panis, Lact. pontis and Lact. frumenti in warm seasons seemed to be associated with increases in the antifungal acetic acid in the TMR silage; however, the activities of inhibiting yeasts and producing 1-propanol in LAB have not been confirmed. Interestingly, the bands indicative of Lact. pontis and Lact. frumenti became distinctive, and the band indicative of Lact. plantarum diminished in summer products. If activations increased in the heterofermentative LAB and decreased in Lact. plantarum in the hot season compared to other seasons, the acetic acid content could increase and ethanol production would decrease as observed in the 2009 samples. The high acetic acid content in summer products may be advantageous for spoilage inhibition. Further experiments are needed to clarify the effect of storage temperature and the activity of sourdough LAB in the ensiling fermentation process.

Lactobacillus acetotolerans was detected in the 2008 samples from the B and C products. Lact. acetotolerans is a homofermentative LAB species and can tolerate a high acetic acid environment at a concentration of 40–50 g l−1 at pH 3·5 (Entani et al. 1986). Although Lact. acetotolerans is usually found in vinegar, it can also be isolated from sourdough (Gül et al. 2005). Because the acetic acid contents in the samples that detected Lact. acetotolerans (B2 and C1) were similar to other 2008 samples, the reason Lact. acetotolerans was found in the TMR silage was unclear. Lactobacillus pobuzihi is also a homofermentative LAB species commonly found in the 2009 samples. Lactobacillus pobuzihi was isolated from pobuzihi, a Taiwanese fermented cummingcordia (Chen et al. 2010), but no information is available on its role in food and feed preservation.

All products examined in this study contained wet BG as an ingredient. This might account for the presence of the LAB species Pediococcus damnosus in the TMR silages, because it was isolated from a brewery environment (Sakamoto and Konings 2003). It may be worthwhile to determine LAB species of each major ingredient in the commercial products, although indigenous LAB do not necessarily dominate the bacterial community of TMR silage (Wang and Nishino 2008).

In conclusion, although Lact. parabuchneri is detected in commercial TMR silages, it may not be essential to the product’s long bunk life after silo opening. Mixed activities of sourdough LAB may collectively contribute to resistance to aerobic deterioration. The reasons why sourdough and commercial TMR silage bacterial communities resemble each other is unknown, and the roles of sourdough LAB in the ensiling process and aerobic stability are worth examining. Because of the suggested role of a fungal association in the prolonged shelf life of sourdough, additional studies of the fungal communities of TMR silage are also worthwhile.