Send Orders for Reprints to Reprints@benthamscience.ae Application of Pretreatment, Bioaugmentation and Biostimulation for Fermentative Hydrogen Production from Maize Silage

Bacteria produce hydrogen during anaerobic dark digestion of carbon rich natural resources including renewable cellulosic materials. The purpose of this work was to study the impact of maize silage pretreatment with Trichoderma fungi, bioaugmentation with defined bacterial inocula and/ or biostimulation with humic acids and an additional inorganic nitrogen source on the fermentative hydrogen production in laboratory batch assay. Experiments were carried out with and without Trichoderma asperellum pretreated silage. The selected bacterial inocula consisted of Clostridium, Enterobacter and Tissierella species, with or without Bacillus mycoides. Headspace gas composition, the amount of dry particulate matter, chemical oxygen demand and concentration of volatile fatty acids in liquid were determined. Bacterial communities were studied with fluorescence in situ hibridization. The predominant cultivable microbial species were isolated and identified. The study demonstrated a significant increase of hydrogen production from maize silage by indigenous bacteria after pretreatment with Trichoderma in comparison with silage untreated with Trichoderma. From tested factors, pretreatment, biostimulation with additional nutrients (ammonium nitrate and/ or humic acids) and bioaugmentation with defined bacterial inocula, pretreatment demonstrated significant improvement of hydrogen production from maize silage. Thereby, aerobic treatment with Trichoderma could be recommended for the pretreatment of silage for the purpose of fermentative production of hydrogen.


INTRODUCTION
Hydrogen is expected to be a widely used energy carrier in the future.Bacteria produce hydrogen using multiple biochemical reactions.Anaerobic digestion is the most studied form of these transformations, when microorganisms use carbon rich natural resources including renewable cellulosic materials.Anaerobic digestion of complex organic matter consists of four stages: hydrolysis, acidogenesis, acetogenesis and methanogenesis.Hydrogen can be produced from acidogenic fermentation and consumed in hydrogenotrophic methanogenesis.
The common biomass categories used in Europe are animal manure and slurry, agricultural residues and byproducts, digestible organic wastes from food and agro industries, the organic fraction of municipal waste and catering waste, sewage sludge and dedicated energy crops [1].Plant materials contain cellulose and lignin.The majority of these macromolecules can be gradually cleaved in aerobic conditions by fungal and bacterial interaction [2].
Before anaerobic digestion, biological pretreatment can be performed by applying either microorganisms or enzymes to the lignocellulose material [3].The studies using microorganisms comprise mainly those using white-and soft-rot fungi [4] that degrade the lignin fraction as lignin is a barrier to enzymatic hydrolysis of cellulose and also cause unproductive binding of cellulases.The rate of biological treatment is much lower than for most other pretreatment methods and that has for many years been considered to be a factor that prevents industrial use.However, in recent years interest has increased and new studies have been performed involving other types of microorganisms, e.g.brown rot fungi that give a degradation of hemicelluloses and cellulose [5].
Trichoderma species are known as cellulolytic fungi.They contain enzyme complexendoglucanases, exoglucanases or cellobiohydrolases, and β-glucosidases [2].However different species, for example T. reesei, do not secrete enough β-glucosidase to ensure effective conversion of biomass to low-molecular carbohydrates [6].Also other plant components such as starch and water-soluble carbohydrates can be used for hydrogen production.For example, feed corn contains 31% starch and 22-25% watersoluble carbohydrates [7] which are used with no difficulty by bacteria in fermentation processes.
To improve biodegradation and H 2 -producing capability, a biostimulation approach has been suggested.Experiments with supplementation of fermentation mixtures with cane molasses [8], waste potato starch and NH 4 NO 3 [9] were provided.Studies of nutrient formulations are still necessary for amelioration of hydrogen production.
Galore microbial species, belonging to mesophilic genera Enterobacter, Bacillus, Clostridium [10], Escherichia, Klebsiella, Paenibacillus, Pantoea etc. [11] are reported to produce hydrogen through dark fermentation.Enriching H 2producing pure and mixed microbial cultures from anaerobic substrates, mostly sludge are described in several studies [12][13][14].Usually microbial cultures are used for biodegradation experiments with defined and relatively simple model substrates, for example starch [15], xylose [16] or sewage sludge [8].Nevertheless, microbiologists are still seeking ideal cultures.In our work, bioaugmentation or inoculation of native lignocellulosic substrate with preestablished H 2 -producing and H 2 -production supporting microorganism association was attempted.
The purpose of this study was to evaluate the influence of maize silage pretreatment with Trichoderma asperellum and bioaugmentation with defined bacterial inocula on the hydrogen production in laboratory dark fermentative batch assay.Biostimulation with humic acids and additional inorganic nitrogen source was also analyzed.

Substrate and Pretreatment of Substrate
Maize silage was obtained from BUFPI Ltd. (Latvia).Pretreatment was carried out with fungus Trichoderma asperellum MSCL (Microbial Strain Collection of Latvia) 309.One hundred grams of silage were mixed with 2.5 ml of suspension of T. asperellum in a concentration of 10 9 colony-forming units (cfu) per ml, and solid state fermentation was carried out for three weeks at room temperature.

Batch Experiments
Hydrogen production was studied at 5 g untreated silage (0.76 g dry mass) or 5 g with Trichoderma asperellum pretreated silage (1.00 g dry mass) and sterile water or supplements in total volume of 30 ml in 50 ml serum bottles (Supelco Analytical, USA) at 37 °C for 46 days.Bottles were sealed with butyl rubber stoppers (Gotlands Gummifabrik, Sweden) and aluminium crimps (Supelco Analytical, USA), and flushed with argon (AGA Ltd., Latvia).
The following 11 variants were created: Cuntreated silage; Huntreated silage (C) supplemented with 0. The presented value in this study was the average value of the duplication.

Methods of Chemical Analyses
Evolved gases from the headspace of the bottles in the amount of 10 cm 3 were collected with a graduated, gas-tight syringe from the test system and tested in the massspectrometer RGAPro-100 (Setaram Instrumentation, France).The RGAPro is a residual gas analyzer designed to work with the PCTPro-2000 or any other gas process for real-time compositional analysis of gas sorption and reaction processes.The system uses a 100 atomic mass units (amu) quadrupole mass spectrometer with better than 1 amu resolution.The unit for mass spectra graphic on x-ass is m/z, where m is amu of measured molecule and z is the charge.The data from mass spectrometer were analyzed by RGA 3.0 Software for SR Residual Gas Analysers program [17].
The liquid samples taken from the bottles were centrifuged at 5000 rpm for 10 min, and then the supernatants were filtered through 0.45 m cellulose acetate membranes and mixed with acetonitril (SAFC, USA) in ratio 1:1.The concentration of the volatile fatty acids (VFAs) was determined using modular UPLC system Agilent 1290 Infinity series (Agilent Technologies, Germany).HPLC column with a dimension of 4.6 x 150 mm (Atlantis T3, 3 µm, Waters, Ireland) was placed in line from the LC system and attached directly to the 1290 Infinity DAD detector (Agilent Technologies, Germany).All samples were scanned over an UV range of 200-400 nm.A flow rate of 0.4 ml/min was maintained for all sample analyses.Elution solvents consist of 1 mM sulfuric acid and acetonitril in 12 minutes gradient mode.Acetonitril (HPLC grade) was supplied by LiChrosolv (Merck KGaA, Germany) and sulphuric acid (99.999%) was supplied by Sigma-Aldrich (St. Louis, USA).The water used was purified with a Milli-Q water purification system from Millipore (Bedford, USA).Data were obtained and processed with MassHunter system (Agilent Technologies, USA).Chemical Oxygen Demand (COD) was estimated in supernatants after centrifugation of liquids at 10000 rpm for 5 min.using RD 125 Thermoreactor (Lovibond, Germany) and MD200 Photometer (Lovibond, Germany).
The pH values were measured using a pH meter AD 1405 (Adrona, Latvia).

Isolation and Identification of Culturable Microorganisms
Bacteria were isolated in agar plates with R2A medium (SIFIN, Germany) at a temperature of 37 C.Serial decimal dilutions of samples were incubated aerobically or anaerobically (GasPak Anaerobe Pouch, Becton & Dickinson, USA).The cultivation was repeated until a single morphology was detected by microscopic observation.Isolated pure cultures were identified biochemically with BBL Crystal ID kits for Gram-Positive bacteria, Enteric/ Nonfermenter bacteria and Anaerobe (Becton & Dickinson, USA).

Fluorescence in situ Hybridization (FISH)
The basic method described by Pernthaler et al. [18] and Fuchs et al. [19] was used.The samples were air-dried, fixed in 4% para-formaldehyde for 1.5-2 hours.After drying, the samples were dehydrated in ethanol (50, 80, and 96%).The following 16S rRNA-targeted oligonucleotide probes were used: LGC353B, GAM42a and Chis150 (Table 1).The formamide concentration for the hybridization (2 h, 46 °C) and NaCl concentration for the washing (20 min, 48 °C) are listed in Table 1.The probes, marked at their 5'-end with Cy3, were purchased from Eurofins MWG Operon (Germany).Epifluorescent microscope DM 2000 (Leica Microsystems, Germany) was used for examination of hybridization results.

SEM
Air-dried samples were treated with solution of 96% ethanol and diethyl ether in the ratio 1:1 and analysed using scanning electron microscope Mira/LMU (Tescan, Czech Republic).

Production of Gases
In fermentation variants with untreated silage 1-2% of hydrogen gas was produced but release of hydrogen in the amount of 15-21% from headspace gases was established in all variants with silage pretreated with fungus Trichoderma asperellum (Fig. 1).Low H 2 production (≤ 0.3%) was observed in bioaugmentation variants with six bacteria (TB) but the variant with five bacteria (TB-B) showed ten times higher production after 7 and 28 days of incubation.Maximum hydrogen production was observed after 28 days of incubation in all of the variants with the exception of TB.Delayed and prolonged release of hydrogen was detected in TN and TNB, where pretreated silage was supplemented with inorganic nitrogen.
The proportion of carbon dioxide reached 40-43% from the volume of headspace gases and the ratio of hydrogen and carbon dioxide reached 1:2.Dynamic of CO 2 production (Fig. 2) resembled dynamic of H 2 production.Pearson`s correlation coefficients r=0.9509 (p <0.00001) and r=0.8931 (p <0.00021), on the 7 th and 28 th day respectively, indicated a strong positive relationship between the values of CO 2 and H 2 .
Concentration of methane was below 0.0% from gases in all cases.

Biodegradation of Silage
SEM analyses demonstrated the effect of fungal pretreatment on the structure of maize silage.Fig. (3A  shows silage without pretreatment and there we can see relatively little affected plant surface whereas Fig. (3B) shows pretreated silage with partially degraded cell walls.
Loss of dry mass during fermentation was established.Fig. (4) shows the proportion of consumed amount of dry mass of silage in comparison with initial input in all variants.Fungal pretreatment increased the loss of dry mass during fermentation for an average 13%.Moreover, the loss during fungal pretreatment was 24% and this parameter was not taken into account in the following calculations.Humic acid supplement significantly (p<0.05)increased the consumption of silage in untreated variant H but bacterial inoculums decreased the consumption in variants B and TNB in comparison with variants C and T (p <0.05).

Chemical Changes During Fermentation
To find out the amount of remaining dissolved organic matter at the end of the experiment, COD was estimated.Obtained results (Fig. 5) show that variants without silage pretreatment contained on average 30% higher concentration   HPLC analyses provided information on the concentration of important organic acidsformic acid, acetic acid, lactic acid, propionic acid and butyric acidin liquid at the end of fermentation.Data summarized in Table 3 show that pretreatment of silage resulted in decrease of the total concentration of organic acids for 10.5-36.9% in comparison with pretreated silage in all comparable variants.The proportion of formic acid varied from 3.0% in B to 5.9% in TB-B.Concentration of acetic acid was low and anywhere exceeded 0.2%.Proportion of lactic acid and propionic acid  was in the range from 20.8% to 30.6% and from 33.4% to 40.4% respectively, in variants with untreated silage and from 0.0% to 15.7% and from 2.8% to 42.4% respectively, in variants with pretreated silage.Butyric acid was the predominant acid in variant H with a proportion of 42.3% and in all pretreated variants with exception of TB.Moderate negative correlation r =-0.7138 (p =0.0136) was found between concentration of total volatile fatty acids (Table 3) and pH (Table 2).Moderate positive correlation r =0.7116 (p =0.0141) was found between concentration of total volatile fatty acids (Table 3) and COD (Fig. 5).Volatile fatty acids constituted from 30.1% (variant TNB) to 83.7% (variant T) from the COD remaining in the liquid after fermentation.

Detected Microorganisms
FISH analyses of fermented liquids were provided with three oligonucleotide probes and indicated the presence of Clostridium clusters I and II including C. beijerinckii in all pretreated variants without TB, presence and dominance of Bacillus spp. in five variants (C, T, TB, TB-B and TNB), and absence of Gamma-proteobacteria (Table 4).
From one to four predominant species of bacteria from every variant were isolated and identified (Table 4).From six bioaugmentation variants, Bacillus mycoides, one of the inoculated bacteria, was isolated in one case, from TB.Other Lactobacillus johnsonii was a sole isolated lactic acid bacterium and it was found in three variants (C, HB and TB-B).Eight variants contained Propionibacterium spp.and six variants, all belonging to pretreated silage, contained Clostridium beijerinckii.We failed to isolate any fungus from fermented silage despite Trichoderma asperellum conidia in SEM (Fig. 3C) in pretreated silage.

DISCUSSION
Renewable cellulosic materials are known as a sustainable source for biofuel production.This study demostrated the feasibility of biohydrogen production by dark fermentation of maize silage at temperature of 37 C.The provided 46 day long experiments showed that indigenous bacteria induce loss of dry mass from 10% to 42% (Fig. 4) and produce a little amount of carbon dioxide (Fig. 2) and almost no hydrogen (Fig. 1) from native, untreated silage.At the same time, the decrease of pH from 6.55 to 4.40-4.21and the accumulation of fatty acids, in particular propionic acid and butyric acid, was observed (Table 3).Predominant Propionibacterium species producing propionic acid were isolated from variants H, B and HB (Table 4).Khanal et al. [23] found that the specific hydrogen production rate is the highest for the pH range of 5.5-5.7.When pH drops to 4.5 or below, the clostridial populations reach the stationary growth phase and the reactions shift from a hydrogen/ acid production phase to a solvent production phase [24].In addition, the optimum pH for methanogenesis is between pH 6.5 and pH 8.0 [25] therefore we suppose that methanogenesis was inhibited in our experiments due to low pH values.Obtained data allow assume that microbial activities became exhausted on day 37.The remaining COD indicated on the unutilized soluble compounds including volatile fatty acids, mainly butyric acid and propionic acid, which were not be able to serve for substrates in particular conditions.Silage pretreatment with fungus Trichoderma asperellum for three weeks at room temperature strengthened fermentation and releasing of hydrogen and carbon dioxide what indicates fermentative activities [26].The amount of consumed dry mass exceeded 45% in four pretreated variants THB, T, TB and TB-B.However, the difference of hydrogen production was high among these variants.No variant of bioaugmentation (TB, TB-B) or bioaugmentation plus biostimulation (THB) exceeded control variant T in relation to hydrogen production as well carbon dioxide production.In the case of active fermentation (TB-B, TN, THB, T, and TH), the volume of hydrogen reached 29-33% of the total gas volume.Variant T also demonstrated the least remaining COD in the solution after fermentation (Fig. 5).One of the predominat bacteria was Clostridium beijerinckii in pretreated variants with the exception of TB which demonstrated a low level of hydrogen production.It is known that C. beijerinckii belongs to hydrogen-producing organisms [26][27][28].Fungi Trichoderma asperellum inoculated in silage at the beginning of aerobic pretreatment procedure had lost their viability after 46 days of anaerobic fermentation.
To enhance hydrogen gas production, Kim et al. [29] reported the addition of nitrate (KNO 3 1 g/l) for methanogenesis inhibition.Our efforts to promote hydrogen production with an additional nitrogen source, ammonium nitrate, did not lead to the expected results.Either nitrogen was in sufficient quantities in the silage, or there was too much concentration of additional nitrate, 5 g/l.For comparison, Wang and Jin [9] have described process optimization of biohydrogen production from molasses by using 1.2 g/l NH 4 NO 3 .However, we observed metabolic changes in variants with additional nitrogen, i.e., decrease of the proportion of lactic acid to zero, decrease of propionic acid to 2.8-8.0%, as well as an increase of formic acid to 5.8-6.7% and in particular, an increase of butyric acid to 85-91%.An excess nitrogen has been found to promote the production of higher organic acids and ethanol, resulting in lowering hydrogen production efficiency [30].
The addition of humic acids in silage reduced the release of hydrogen in all the studied variants and decreased (p <0.05) the proportion of the consumed amount of dry mass in variants without pretreatment (Fig. 4).It is known that diversity of bacteria can utilize humic substances as electron donors for anaerobic respiration [31], compete with methanogenesis [32] and suppress hydrogenotrophic methanogenesis [33].As confirmed in our study, an abundance of humic acids adversely affects the production of hydrogen.Some aerobic Bacillus species and facultative anaerobic Enterobacteriaceae are known as hydrogen producers and anaerobic environment creators [34].Chang et al. [35] suggested contribution of Bacillus and Clostridium coculture for development of industrialized bio-fuels and biohydrogen producing systems from biomass.Our attempt to make an effective bacterial community for hydrogen production from maize silage did not achieve the expected aim.Bioaugmentation of silage with defined bacterial inocula consisting of anaerobic (Clostridium butyricum, C. paraputrificum, Tissierella praeacuta) and facultative anaerobic bacteria (Enterobacter asburiae, E. cloacae) with or without aerobic Bacillus mycoides did not increase hydrogen (Fig. 1) and CO 2 production (Fig. 2) and amount of consumed dry mass (Fig. 4).Although inoculated bacteria failed to become predominant, with the exception of Bacillus mycoides in TB, their effect was unfavorable for fermentation processes.In particular, adverse effects were shown by B. mycoides.If we compare concentration of organic acids at the end of fermentation, variant TB showed greater amount of total fatty acids than TB-B.Exactly, concentration of lactic acid and propionic acid was increased in the presence of B. mycoides.To our best knowledge, there are no investigations about anaerobic metabolism of this species but it is known that its close relatives can grow anaerobically by respiration with nitrate as a terminal electron acceptor and produce lactate, acetate and 2,3butanediol as the major anaerobic fermentation products [36].Screening of 203 Bacillus natural isolates provided by Beric et al. [37] demonstrated that 127 strains exhibit antimicrobial activity.Accordingly, special attention should be paid to the selection of appropriate Bacillus species in future experiments.
Acetate is the most important intermediate in anaerobic environments but acetate constituted only 0.0-0.2% of total volatile fatty acids at the end of fermentation (Table 3).We assume that acetate has been utilized.The ability to degrade acetate is widely spread among methanogenic, sulfatereducing and nitrate-reducing microorganisms [38].Methane was not found in our analyses therefore we suppose that our fermentation liquids contained active acetate-utilizing and nitrate-reducing bacteria.Elefsiniotis et al. [39] investigated the ability of naturally-produced volatile fatty acids act as a carbon source for the removal of nitrate and found that the denitrifier population had a preference for acetic acid, and butyric acid and propionic acid were consumed only after acetate concentrations began to decline.According to investigation of S. F. Magram [40], formic acid is not a good carbon source for denitrification.

CONCLUSION
The present study was an attempt to improve the biohydrogen production using various nutrient and seed formulation methods on silage.From tested factors, pretreatment, biostimulation with additional nutrients (ammonium nitrate and/ or humic acids) and bioaugmentation with defined bacterial inocula, pretreatment demonstrated significant improvement of hydrogen production from maize silage.Thereby, aerobic pretreatment with Trichoderma asperellum MSCL 309 and possibly other species and strains of fungi from genus Trichoderma could be recommended for the pretreatment of silage for the purpose of fermentative production of hydrogen.

Fig. 5 )
variants with pretreated silage.Weak negative correlation r =0.4691 (p =0.1456) was found between COD (and the proportion of consumed dry mass (Fig.4).Initial pH in the fermentationmedium was 0.11 lower after fungal pretreatment than without pretreatment.During anaerobic fermentation, pH values decreased in all variants (Table 2) especially in untreated silage where pH reached 4.21 in variant C. The pH values ranged from 5.45 in TNB to 4.74 in TH in the pretreated silage.