Effect of inoculants and storage temperature on the microbial, chemical and mycotoxin composition of corn silage

Objective To evaluate the effect of lactic acid bacteria and storage temperature on the microbial, chemical and mycotoxin composition of corn silage. Methods Corn was harvested at 32.8% dry matter, and chopped to 1 to 2 cm. The chopped material was subjected to three treatments: i) control (distilled water); ii) 1×106 colony forming units (cfu)/g of Lactobacillus plantarum; iii) 1×106 cfu/g of Pediococcus pentosaceus. Treatments in triplicate were ensiled for 55 d at 20°C, 28°C, and 37°C in 1-L polythene jars following packing to a density of approximately 800 kg/m3 of fresh matter, respectively. At silo opening, microbial populations, fermentation characteristics, nutritive value and mycotoxins of corn silage were determined. Results L. plantarum significantly increased yeast number, water soluble carbohydrates, nitrate and deoxynivalenol content, and significantly decreased the ammonia N value in corn silage compared with the control (p<0.05). P. pentosaceus significantly increased lactic acid bacteria and yeast number and content of deoxynivalenol, nivalenol, T-2 toxin and zearalenone, while decreasing mold population and content of nitrate and 3-acetyl-deoxynivalneol in corn silage when stored at 20°C compared to the control (p<0.05). Storage temperature had a significant effect on deoxynivalenol, nivalenol, ochratoxin A, and zearalenone level in corn silage (p<0.05). Conclusion Lactobacillus plantarum and Pediococcus pentosaceus did not decrease the contents of mycotoxins or nitrate in corn silage stored at three temperatures.


INTRODUCTION
Corn is widely utilized as a ensiling material throughout the world, and corn plants in the field can be contaminated with fungi and mycotoxin formation may occur under unfavor able conditions, such as high temperature and drought stress. These fungi and mycotoxins produced can survive the ensiling process and cause animal health problems. Mycotoxins are low molecular weight secondary metabolites formed mainly by Aspergillus, Penicillium, and Fusarium species [1]. The most common mycotoxins in livestock feeds are aflatoxin B 1 (AFB 1 ), ochratoxin A (OTA), zearalenone (ZEA) and trichothecenes. Mycotoxin contami nation not only results in reduced animal feed intake, reproduction and feed conversion efficacy [2], but also may cause carcinogenesis, teratogenesis and immune system suppres sion as a result of chronic toxicity even at low levels [3,4]. Moreover, the presence of multiple mycotoxins is of particular concern due to potential synergistic effects on livestock exposed to moldy silage [3]. Another issue is that mycotoxins show higher resistance than mycelia to feedstuff processing and storage [5]. Thus, it is important to identify the methods for degrading or transforming mycotoxins in silage. Physical and chemical approaches, such as an addition of ammonia and adsorbents [4,6], may be dan gerous to applicators. In addition, some adsorbents even bind minerals and vitamins as well as mycotoxins, reducing feed quality [7].
A few studies have indicated that mycotoxins can be de graded or transformed by some microbes. Rumen microflora can degrade and inactivate mycotoxins [8], and intestinal microbes were shown to convert ZEA to αzearalenol and an unknown metabolite [9]. Some reduction of mycotoxins pro duced in the field was attributed to lactic acid bacteria (LAB) [10]; an in vitro study reported that binding of (deoxynivalenol) DON and ZEA is the major mode of action for LAB [11]. Microbial activity during silage fermentation caused the break down of mycotoxin ZEA [12], and Lactobacilli and Pediococcus species were reported to be able to transform some mycotoxins [13]. Mold normally grows at 10°C to 40°C, and Aspergillus and Penicillium species can grow at higher temperatures than Fusarium species [14]. However, mold growth under the op timum temperature range is not necessary for mycotoxins production. A comparison of the optimum temperature for mycotoxins formation by Aspergillus, Penicillium and Fusarium species [15], defined the temperature range of mycotoxins production for A. flavus (AFB 1 ), P. verrucosum (OTA) and Fusarium species (toxin T2, DON, nivalenol [NIV] and ZEA) to be 12°C to 40°C, 4°C to 20°C, and 24°C to 26°C, respec tively. To date, few studies have evaluated the effect of LAB on the stability of mycotoxins in corn silage stored at differ ent temperatures.
Thus, the objective of this study was to evaluate the effect of inoculants and storage temperature on the microbial pop ulations, fermentation characteristics, nutritive value and mycotoxins of corn silage.

Chemical and microbial analyses
At silo opening, corn silage mass was mixed manually prior to sampling. A subsample of 20 g was weighed into a blender jar, diluted with 180 mL of distilled water, and homogenized with a juicer for 2 min. Extracted solution was filtered through four layers of cheesecloth and one layer of qualitative filter paper, and analyzed for pH value using an electrode (PHS3C, INESA, Shanghai, China). Then, 2 mL of filtrate was centri fuged at 10,000 g at 4°C for 5 min and reserved for fermentation acids and ammonia N (NH 3 N) analysis. Lactic, acetic, pro pionic, and butyric acids were determined by high performance liquid chromatograph (HPLC, SHIMADZU10A, Kyoto, Japan). The HPLC system consisted of a Shimadzu system controller (SCL10A), and a Shodex Rspak KC811 SDVB gel column (300 mm×8 mm) with a column temperature of 50°C. The mobile phase was 3 mmol HClO 4 running at 1 mL/min, and the injection volume was 5 μL. A UV detector (SPD10A) was used for detection at 210 nm. NH 3 N was determined by the phenol method.
A second subsample of 200 g from the ensiled forage or silage was dried in a forceddraft oven at 65°C for at least 48 h to determine DM. Neutral detergent fiber and acid deter gent fiber were determined according to Van Soest et al [17]. Water soluble carbohydrates (WSC) were determined using the anthrone method. Nitrogen was analyzed according to Kjeldahl method. Crude protein was calculated by multiply ing 6.25 with N content. Ether extract (EE) was determined by petroleum ether extraction using filter bag technology. A XT4 filter bag was filled with 1 g of sample, sealed with a capper, and extracted in a X15I FAT EXTRACTOR (ANKOM Technology Corp., Macedon, NY, USA) with petroleum ether (analytical reagent grade, 30°C to 60°C) for 1 h. Petroleum ether was purchased from Sinopharm Chemical Reagent Co., Ltd (Beijing, China). Nitrate was determined by the salicylic acid method.
The third subsample of 20 g from silage was put into a sterile triangular flask, suspended in 180 mL of sterile solu tion (1 g of peptone and 9 g of sodium chloride per liter), and homogenized in a laboratory oscillator at a low speed for 30 min. A series of dilute solutions were prepared for microbial count. The LAB were measured on MRS agar by incubating plates at 37°C for 2 d under anaerobic conditions. Coliform bacteria (CB) were estimated following growth on violet red bile agar incubated at 37°C for 3 d. Yeast and mold were de termined on rose bengal medium after incubation at 28°C for 3 d and 5 to 7 d, respectively, with yeast and mold numbers counted separately according to their macromorphological characteristics. Bacteria and fungi were counted on the plates that yield 30 to 400 cfu and 1 to 100 cfu, respectively. All media used were obtained from Beijing Aoboxing Biotech Co., Ltd (Beijing, China).
Extraction procedure and clean-up: For silage extracting, the QuEChERS extraction method was used [18,19], where 5 g of DM silage was weighed into a 50 mL centrifuge tube, soaked in 10 mL of 2% formic acid for 30 min, and oscillated on a rotary shaker for 30 min at 240 rpm after an addition of 10 mL of acetonitrile. Subsequently, 4 g of MgSO 4 and 1 g of NaCl were added, and the tube was capped immediately (a brief hand shaking was done immediately after an addition of salts to prevent agglomeration). The slurry was immedi ately oscillated for 30 s with a vortex mixer and centrifuged at 10,000 g for 5 min at 4°C, with an aim to induce phase separation and mycotoxins partitioning. Removal of residual water and cleanup were performed simultaneously by using a rapid procedure called dispersive solidphase extraction (dispersiveSPE), where 2 mL of acetonitrile extract was mixed with 300 mg of MgSO 4 and 100 mg of C 18 endcapped silica sorbent (Agilent Technologies, Santa Clara, CA, USA) in a 15 mL centrifuge tube, and centrifuged at 10,000 g for 1 min at 4°C. A portion of the final solution (1 mL) was filtered to a sample bottle (Agilent Technologies, USA) by a 0.22 μm syringe filter, and stored at -20°C until further analysis.
Instrumental conditions: Instrumental conditions were set according to the Agilent Mass Spectrometry (MS) Technology Products Solutions solution package (mycotoxin, 12906470 Parameters, Agilent Technologies, USA). An Agilent infinity 1290 ultrahigh performance liquid chromatography (UHPLC) system (Agilent Technologies, USA) was utilized throughout the study. This instrument consisted of a UHPLC pump with a builtin microdegasser, an infinity autosampler with back flush function, and a temperature control compartment. Chromatographic separation was achieved using an Agilent Eclipse Plus column (1.8 μm, 100 mm×2.1 mm, Agilent Tech nologies, USA) at a flow rate of 300 μL/min, with a mobile phase consisting of solvent A and solvent B, where solvent A was water solution containing 1% acetic acid and 5 mM ammo nium acetate, and solvent B was methanol. The chromatogram of 8 mycotoxins mixed standard solution and gradient elu tion program are shown in Figure 1 and Table 1, respectively. The injection volume of each sample was 2 μL and column temperature was 40°C.
Analytes were detected using an Agilent 6470 triple quadru pole MS equipped with a Jet Stream electrospray ionization probe. MS was operated in dynamic multireaction monitor by monitoring two transitions (one quantifier and the other qualifier) for each analyte, with individual dwell time. MS pa rameters were as follows: gas temperature 300°C; sheath gas temperature 350°C; gas flow 7 L/min, sheath gas flow 11 L/min, and capillary 3,500 V, positive. The acquisition was performed in positive polarity, and the optimized MS conditions are outlined in Table 2. Linearity was established by injecting increasing concentrations (triplicates) of working solution (0.5, 1, 5, 10, 20, 50, and 100 ng/mL). Standard curves were linear in the range studied, showing correlation coefficients of >0.999. Quantification and detection limits were determined by spiked samples based on signaltonoise ratios of 10:1 for quantification and 3:1 for detection limit.

Statistical analysis
All microbial counts (LAB, yeast, mold, and CB) were trans formed to log10 units on fresh weight basis. Data of the fermentation characteristics, microbial populations, nutritive value and mycotoxin content in silage were analyzed by the Factorial Design model of SAS (9.1 version, SAS Institute Cary, NC, USA). The model used for the analysis was: Y = μ + treatment + storage temperature + treatment × storage temperature + ε, where Y = observation, μ = general mean, treatment = effect of CK or LP or PP, storage temperature = effect of 20°C or 28°C or 37°C, treatment×storage temperature = interaction between treatment and storage temperature, ε = residual error. Bon grouping contrasts were utilized to com pare means among four levels of one factor with the other factor immobilized. The significant difference was declared at p<0.05.

Chemical, microbial and mycotoxin composition of corn
Chemical and microbial composition and mycotoxin content of ensiling plant are presented in Table 3 and 4, respectively. Nitrate, DON, and ZEA content of corn was 383 mg/kg, 163 and 30.6 ng/g accordingly. Table 5 shows the microbial counts and fermentation char acteristics of corn silage treated with inoculants and storage temperature. LAB count in PPtreated silage silage stored at 20°C was higher than CK and LP (p<0.05), and their count          Table 5, the pH value in any silage was below 4.0, ranging from 3.69 to 3.85. The pH value in PPtreated silage conserved at 20°C and 28°C was higher than CK and LP (p<0.05). With the addition of LP or no additive, silage conserved at 37°C had a higher pH level compared to 20°C and 28°C (p<0.05). The application of PP not only resulted in a higher pH level in silage conserved at 37°C in compari son with 20°C and 28°C (p<0.05), but also caused a higher value of pH in silage conserved at 20°C compared to 28°C (p<0.05). NH 3 N content in any silage increased significant ly as conservation temperature increased from 20°C to 37°C (p<0.05), irrespective of treatment effect. NH 3 N content in LPtreated silage stored at any temperature was lower than CK (p<0.05), whereas PPtreated silage stored at 28°C had a lower level of NH 3 N in comparison with CK (p<0.05). The addition of LP lowered NH 3 N level in silage stored at 20°C and 37°C compared to PP (p<0.05). Without any additive, NH 3 N value in silage ensiled at 28°C and 37°C was higher compared to 20°C (p<0.05). With the use of LP, NH 3 N value in silage ensiled at 37°C was higher compared to 20°C and 28°C (p<0.05), and its content in silage ensiled at 28°C was higher than 20°C (p<0.05). The application of PP led to a higher content of NH 3 N in silage ensiled at 37°C compared with 20°C and 28°C (p<0.05). PA content in LPtreated silage ensiled at 20°C was higher than 28°C and 37°C (p<0.05), whereas there were not significant differences on PA content for all the other contrasts (p>0.05). Table 6 presents the effect of inoculants on the nutritive value of corn silage stored at three temperatures. DM content in CKtreated silage stored at 37°C was lower in comparison with LP and PP (p<0.05), and there were not significant dif ferences on DM content for all the other contrasts (p>0.05). CP level in any silage decreased significantly with storage tem perature increasing from 20°C to 37°C (p<0.05), irrespective of additive effect. CP level in CKtreated silage conserved at 28°C was lower compared to LP and PP (p<0.05), whereas its content in LPtreated silage conserved at 37°C was higher than CK and PP (p<0.05). With no additive, silage conserved at 20°C had a higher content of CP compared with 28°C and 37°C (p<0.05). The application of LP induced a higher value of CP in silage conserved at 20°C in contrast with 37°C (p< 0.05), and PPtreated silage conserved at 37°C had a lower value of CP than 20 and 28°C (p<0.05). WSC content in LP treated silage ensiled at any temperature was higher compared to CK and PP (p<0.05), whereas its content in PPtreated silage ensiled at 37°C was higher than CK (p<0.05). With the addi tion of PP or no additive, silage ensiled at 37°C had a higher level of WSC in comparison with 20°C and 28°C (p<0.05), and WSC level in silage ensiled at 20°C was higher in con trast with 28°C (p<0.05). The application of LP resulted in a higher value of WSC in silage ensiled at 37°C compared to 20°C and 28°C (p<0.05), whereas there were no differences on WSC value for all the remaining comparisons (p>0.05). Nitrate content in silage, ranging from 61.1 to 91.4 mg/kg, was lower compared with 383 mg/kg in corn. At 20°C, nitrate level in LPtreated silage was higher compared to CK and PP (p<0.05), and its level in silage treated with CK was higher than PP (p<0.05). LPtreated silage stored at 28°C had a higher level of nitrate than CK and PP (p<0.05). At 37°C, nitrate level in silage treated with LP was higher compared to CK and PP (p<0.05), and its level in silage treated with CK was lower in comparison with PP (p<0.05). The use of LP led to a higher value of nitrate in silage stored at 20°C compared to 28°C (p<0.05), whereas PPtreated silage stored at 37°C had a higher value of nitrate than 20°C and 28°C (p<0.05), and the application of PP caused a lower value of nitrate in silage stored at 20°C in contrast with 28°C (p<0.05).

Mycotoxin content of corn silage
Mycotoxin content of corn silage treated with LAB and storage temperature is stated in Table 7. NIV content in PPtreated silage stored at 20°C was higher than CK and LP (p<0.05), whereas its content in PPtreated silage stored at 28°C and 37°C was lower compared to CK and LP (p<0.05). NIV level in LPtreated silage was higher and lower in comparison with CK (p<0.05), at 28°C and 37°C, respectively. Without any ad ditive, NIV value in silage stored at 37°C was higher compared to 20°C and 28°C (p<0.05), and its value in silage stored at 20°C was lower than 28°C (p<0.05). The use of LP resulted in a lower value of NIV in silage stored at 20°C in contrast

DISCUSSION
In the present experiment (Table 5), LPtreated corn silage stored at any temperature had a lower level of NH3N and a higher number of yeast in comparison with CK, which is in agreement with Filya et al [20]. Without any additive, LAB number in corn silage stored at 37°C was lower than 20°C and 28°C, and a study found the similar phenomenon [21]. Ac cording to Table 6, the application of LP resulted in a higher content of WSC in corn silage stored at any temperature com pared to CK, it was reported that its concentration in corn silage treated with LP was lower than CK [20]. EE content decreased from 2.69% in corn to 2.09% to 2.29% of corn si lage, perhaps microbial activity during fermentation degraded some EE. Nitrate level in corn was 383 mg/kg, and decreased to 61.1 to 91.4 mg/kg after ensiling. Both corn and silage were safe in terms of nitrate content, because their concentrations were below the limit level (1,000 mg/kg). A study showed that nitrate contents in corn were 374 and 433 mg/kg, and reduced to 18.5 to 111 mg/kg of corn silages [22]. The eight mycotoxins in this experiment can be divided into Aspergillus and Fusarium toxins, with AFB 1 and OTA belonging to Aspergillus toxins, and Fusarium toxins consist ing of the remaining six toxins. As can be seen from Table 4 and 7, AFB 1 existed in pre and postfermented corn samples, and its content was higher in corn (30.2 ng/g) compared with corn silage (23.4 to 24.5 ng/g). This result indicated that AFB 1 had a slow degradation during fermentation, and agrees with Kalac and Woodford [23]. However, AFB 1 level was higher in corn silage when compared with prefermented samples [24], suggesting that A. flavus and A. parasiticus activity was en hanced during storage. Silage contamination with AFB 1 was also found in Argentina [25], Egypt [26] and France [27]. These phenomena above show that AFB 1 infection may be influenced by field environment where forage crop grow, and is usually concerned with tropical or subtropical regions with a higher temperature. Treatment and storage temperature did not have an effect on AFB 1 level in corn silage at this ex periment, however. OTA is a common contaminant of corn in temperate regions, and its level in corn was 34.4 ng/g at this experiment. Conservation temperature has a significant effect on OTA content in corn silage, and the application of LP and PP decreased its level in corn silage stored at 37°C compared to CK.
Most information for Fusarium toxins is for DON in corn silage. DON content in CKtreated corn silage stored at any temperature was lower than that of corn (Tables 4, 7), and two workers found a significant decrease in DON content after ensiling [28], perhaps because Fusarium species belong to field molds and their activity was suppressed by low oxygen and pH environment of silage [29]. In this experiment, storage temperature had a remarkable effect on DON level, and its level was not reduced in corn silage treated with LP and PP, whereas one study demonstrated that storage temperature did not affect DON stability [28]. Furthermore, their results showed that ensiling with low DM and prolonging storage time led to mycotoxin degradation of corn silage. DON level was also influenced by silage distribution in a silo, proved by two studies. One study showed that a higher content of DON was detected on the top of corn silage [27], and a lower con centration of it was found at the bottom of corn silage by the same workers [27]. On the whole, ZEA level kept unchanged in corn silage in this study, and this result was in accordance with Gonzales Pereyra et al [25].