Introduction

During the last decade, global climate change has raised public awareness about human beings’ excessive reliance on the usage of petroleum for energy and chemical production. Biorefinery of plant biomass to bio-based chemical and energy, such as lactic acid and bioethanol, has been suggested as part of the solutions to achieving sustainable developments. (Srirangan et al. 2012; del Castillo-Romo et al. 2018). Nevertheless, despite the renewable characteristics, the increasing demands of the food crops and terrestrial plant biomass for biorefinery directly compete with the arable lands and agricultural resources (Havlík et al. 2011; Valentine et al. 2012). These kinds of conflicts will only be elevated over the time with the growth of global population. In order to avoid this dilemma, aquatic biomass such as algae has been defined as the third-generation feedstock for biorefinery (Hoevers 2011). Among all kinds of aquatic biomass, algae are commonly acknowledged for their potential as a renewable resource to generate a variety of chemicals and fuels (Jones and Mayfield 2012; Bikker et al.2016; Resdi et al. 2016).

Sea lettuce belongs to the family of green macroalgae Ulvaceae which commonly exist in the littoral zone of the coast (Zhu et al. 2016). The Ulva species have been traditionally used as animal feed, herbal medicine, and food ingredient in many human civilizations (Vázquez-Rodríguez and Amaya-Guerra 2016). Additionally, the Ulva species have been extensively evaluated for different types of applications related to environmental biotechnology such as bioremediation of wastewater, bioabsorption of toxic metal ions from water bodies (Henriques et al. 2017; Shaaban et al. 2017), and biological indicator of the aqueous environment (Farias et al. 2017). Because of their rapid nutrient uptake capabilities, promotion of biomass reproduction and wide ranging salinity tolerance, many investigations have suggested that Ulva lactuca could also be used as a new feedstock for bioenergy production (Bruhn et al. 2011; van der Wal et al. 2013; Chen et al. 2015a, b). Yet, research related to storage and supply logistics of the Ulva species or other macroalgal biomass is relatively limited in comparison with studies of biomass conversion technology. Storage of macroalgal biomass is a challenge because of its high water content, and freshly harvested macroalgal biomass can spoil rapidly if the storage is not properly implemented.

Conventional storage of macroalgal biomass is usually done by oven dry or sun dry (Foscarini and Prakash 1990; Mabeau and Fleurence 1993); in addition, storage at low temperature may also work in preserving fresh macroalgae biomass (Onodera et al. 2011). However, these storage approaches require either high energy consumption or permissible weather conditions, which might not suitable in terms of low cost and a consistent supply of feedstock for bio-based chemicals and fuels (Kadam et al. 2015; Franco et al. 2017). Ensilage is a wet storage approach for biomass preservation, which had been traditionally employed in the industrial animal agriculture. The main objective of ensilage is to prevent the loss of carbohydrates by creating a low-pH environment where the microbial activity is inhibited or decreased. Storage of biomass through the ensilage process has many advantages over dry storage including low risk of fire and less loss of dry matter (Oleskowicz-Popiel et al. 2011). Nevertheless, in comparison with the ensiling of lignocellulosic biomass, there is only some information available on the ensiling of macroalgae. An early study done by Black (1955) examined ensiling of brown seaweeds for animal feed and chemical processing purposes and concluded that the macroalgae supports lactic acid fermentation; however, pH values of the seaweed silages did not decline below the desired level. A similar investigation reported by Herrmann et al. (2015) also indicated that high buffering capacities, insufficient fermentable sugars and low initial counts of lactic acid bacteria (LAB) are the main challenges for ensiling macroalgae. Many technical reports and research studies recommend that LAB counts should be higher than 105 CFU/g biomass for silage preparation (Weinberg and Muck 1996; Muck 2008; Basso et al. 2012; Abdul Rahman et al. 2017). There is no strict required value of initial fermentable sugar content for silage preparation, because the sugar profile varies with different sources of biomass. For ensilage, the amount of fermentable sugar is usually measured as water-soluble carbohydrates (WSC). A report suggests that at least 7% WSC is sufficient for preparing wheat straw silage (Yang et al. 2006). Another report suggests that at least 2.5% WSC is required to obtain an acceptable silage quality of fresh material, while 2% WSC is adequate for inoculated silage (Pettersson and Lindgren 1990). In order to accelerate anaerobic fermentation and drop the pH level, LAB inoculum, cell degrading enzymes, and chemical additives are often added at the beginning of ensilage (Kung and Shaver 2001). A successful preparation of macroalgae silage as fish hatchery feeds was done by adding LAB and yeasts at the beginning of ensilage (Uchida et al. 2004). Ensilage can also be employed as a biological approach to pretreat biomass for further processing. For example, enzymatic digestibility of grape pomace and sugar beet pulp was improved by ensilage (Zheng et al. 2011a, b; 2012). Thus, the aim of this study was to apply ensilage as a green approach for simultaneous preservation and pretreatment of Ulva lactuca and to test the fermentability of Ulva silage for biorefinery purposes by lactic acid fermentation. In order to overcome the problems associated with low initial LAB and low soluble carbohydrates in the Ulva biomass, the effects of factors including addition of cell wall degrading enzyme and inoculation of LAB at the beginning of ensilage on the carbohydrate preservation and subsequent dilute acid hydrolysis were investigated and reported. Additionally, lactic acid fermentation was performed to evaluate the fermentability of Ulva silage.

Materials and methods

Biomass preparation

Biomass of U. lactuca was purchased from local residents who collected the fresh biomass from the coast near the Hoping Island (Keelung, Taiwan). The fresh biomass was used directly or stored in Ziploc® bags at − 20 °C until use.

Chemicals and enzymes

All chemicals used in the present study are analytical grade purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan) except the sulfuric acid used to prepare the mobile phase for high-performance liquid chromatography (HPLC) analysis is HPLC grade. Cellulase enzyme complex ACCELLERASE®1500 was given by DuPont™ Genencor® Science (Rochester, NY, USA) as a gift.

LAB cultivation and ensilage inoculum preparation

Lactobacillus plantarum purchased from Bioresource Collection and Research Center (BCRC# 10069, Hsinchu, Taiwan) was used to ensile the biomass of U. lactuca and to ferment acid hydrolysate of ensiled biomass. L. plantarum was maintained and precultured in MRS broth (HiMedia M369-500G, Midland Scientific, Inc. Omaha, NE, USA).The preparation of ensilage inoculum was done according to the reference published by Zheng et al. (2011a, b).

Ensilage of U. lactuca

A 22 experimental design plus a central point was performed to test the effects of ensilage additives, cellulase (10 CMCU/g dry biomass) and LAB (1 × 106 cfu/g dry biomass), on the carbohydrate recovery after ensilage and the yield of reducing sugar after dilute acid hydrolysis. Triplicates of five base experimental combinations (Table 1) were generated by using a statistical software Design Expert v8 (Stat-Ease, Inc. MN, USA) and carried out in random order. Ensilage of sea lettuce was conducted in 500-ml airtight screw-capped PP jars at 25 °C in an isothermal incubator. Before ensilage, the water content of Ulva lactuca was first determined, and then 59 g equivalent dry weight biomass was packed into the jars. The final moisture content was adjusted to 80% by adding distilled water which contains the designed amount of LAB and cellulase. The content in the jar was manually mixed by using a stir rod and then tightly screw caped. The ensilage period was 30 days and the pH of each silage fermentation was monitored using a potable pH meter (CLEAN PH30 Tester, Clean Instrument Co., Ltd, New Taipei City, Taiwan) by opening jar at storage periods of 0, 7, 15, 22 and 30 days in an anaerobic glove box. After ensilage, total solid, WSC, total carbohydrates, and the biomass composition were determined. The organic acids in silage effluent and cold water extracts of silage samples were analyzed by HPLC.

Table 1 Coded and actual value of experimental design factors
Full size table

Dilute acid hydrolysis of Ulva biomass

Dilute acid hydrolysis was carried out in 250-ml screw-capped laboratory bottles (DURAN®, GL 45, Wertheim/Main, Germany) with 100 ml total reaction volume. Ensiled and raw biomass of sea lettuce was hydrolyzed in the presence of a designated concentration of sulfuric acid at 15% (w/w) solid loading and 80 °C for designated period. Two milliliters of the well-mixed hydrolysate was neutralized with CaCO3, and then the liquid and solid in the hydrolysate were separated by centrifugation at 6000 ×g for 20 min. The amount of reducing sugar in the supernatant was determined by dinitrosalicylic acid (DNS) method. The content of monosaccharides, organic acids, and fermentation inhibitors in the hydrolysate was identified and quantified using HPLC.

Lactic acid fermentation of acid hydrolysate

The acid hydrolysate (4% H2SO4, 120 min) of silage with both additives was neutralized to pH 5 by slow addition of 6 N NaOH. The neutralized acid hydrolysate was further hydrolyzed with ACCELLERASE®1500 (0.2 ml/g silage, dry basis). The reducing sugar concentration was also determined by DNS method. Then, the enzymatic hydrolysate was used as solution to prepare MRS medium without addition of glucose and the final concentration was also adjusted to 2% (w/w) by addition of sterilized DI water. Overnight culture of LAB (BCRC 10069) was inoculated to 25 ml silage hydrolysate MRS medium in a 125-ml screw-capped Erlenmeyer flask at initial optical cell density at 600 nm (OD600) equal to 0.2. Two milliliters of samples was withdrawn at 0, 8, 16, 24, 48, and 72 h of fermentation in an anaerobic glove box. The optical cell density and the concentration of lactic acid, acetic acid, and monosaccharides were determined.

Analytical methods

The pH values were averaged from three measurements of each sample by using a pH meter. The solid content was determined according to the NREL Laboratory Analytical Protocol (LAP) “Determination of Total Solids in Biomass and Total Dissolved Solids in Liquid Process Samples” (Sluiter et al. 2008). For compositional carbohydrate analysis, silage and raw biomass were dried at 45 °C and powdered by bead beating (Mini-BeadBeater-1, Bio Spec Products Inc., Bartlesville, OK, USA) with 3 stainless beads (diameter 3.2 mm) in 2-ml capped stainless steel microvials (catalog # 2007, Bio Spec). The powdered biomass was hydrolyzed by two-stage sulfuric acid hydrolysis according to a published report (Chen et al. 2015a, b) modified from the NREL Laboratory Analytical Protocol (LAP) “Determination of Structural Carbohydrates and Lignin in Biomass” (Sluiter et al. 2012). The acid hydrolysate was first neutralized using CaCO3 followed by static precipitation at 4 °C overnight and centrifugation at a relative centrifugal force of 6000 ×g for 20 min. The content of monosaccharides, organic acids and fermentation inhibitors resulted from ensilage and acid hydrolysis was analyzed by HPLC equipped with a Rezex ROA-organic acid column and guard cartridge (Phenomenex Inc., Torrance, CA, USA), RID and UV detectors (λ = 210 nm).

Data analysis

All presented values are the average of three replicates, unless specified otherwise. Design Expert V8 was used to perform statistical analyses (Version 8.0; SAS Institute, Raleigh, NC). The statistical significance of treatment effects and mean comparisons were determined by analysis of variance (ANOVA) with significance level α = 0.05.

Results and discussion

Effect of additives on the pH variation over the ensiling period

A quick and effective front-end fermentation during early ensiling is essential to drop the pH to lower than 4. The low-pH environment can prevent microbial growth and reduces the loss of dry matter and nutrients (Zheng et al. 2011a, b). In order to circumvent the problems associated with ensiling macroalgae, including insufficient fermentable sugars and low initial LAB count (Herrmann et al. 2015), commercial cellulase complex and LAB inoculum were applied in this study as silage additives. Because the biomass of Ulva sp. contains cellulose as part of its carbohydrate structure (Trivedi et al. 2011), the aim of adding cellulase complex was to release some fermentable sugars from the Ulva biomass at the beginning of ensilage and the LAB inoculum could successively promote the lactic acid fermentation. The result (Fig. 1) showed that only the silage treated with both additives could reach the target pH (< 4) in 15 days from the beginning of ensilage. There are a limited number of successful cases that can be found on the ensilage of Ulva species (pH < 4); one report demonstrated that additions of certain types of cellulase and co-inoculation of LAB and yeast could be beneficial for preparation of fermented marine silage feed from Ulva reticulate (Felix and Pradeepa 2011). Here, both the addition of cellulase complex (p value < 0.0001) and LAB inoculum (p value < 0.0001) were required to drop the pH value of the silage based on the statistical analysis (Table 2). Among all experimental sets, only the silage received both the cellulase complex (10 CMCU/g dry biomass) and LAB inoculum (106 cfu/g dry biomass), and it reached a pH value lower than 4 at the end of ensilage period. The addition of cellulase (coded coefficient = −0.66) seems to be slightly more important than the inoculum of LAB on silage (coded coefficient = −0.27). For silages without the addition of the cellulase complex, the pH remained above 5.5 and high concentrations of volatile fatty acids (VFA) were observed (Table 3), with the sign of spoilage and growth of molds appearing after 30 days of ensiling. These results are similar to reports that examined the ensilage of macroalgae without supplement of silage additives (Herrmann et al. 2015) or with inoculum of LAB only (Cabrita et al. 2017). The formation of VFA could be attributed to the activity of clostridia and other unwanted microbes. Clostridia and mold growth is not favored for ensilage because they can consume lactic acid and produce VFA resulting in a pH rise that opposes the preservation mechanism of ensilage. The growth of unwanted clostridia, mold, and yeast can also degrade carbohydrates and protein into VFA, ammonia, and biogas, leading to substantial loss of nutrients and dry matter (Kung Jr 2001). Based on the pH value (Table 3) and solid recovery (Table 4), a good quality silage of Ulva lactuca could be prepared by addition of cellulase complex and LAB inoculum at the beginning of ensilage.

Fig. 1
figure 1

Changes of silage pH value during the ensiling period. (+, +) with cellulase and LAB inoculum, (+, −) with cellulase only, (−, +) with LAB inoculum only, and (− ,−) without silage additives

Full size image
Table 2 Summary of analysis of variance (ANOVA)
Full size table
Table 3 Final pH, organic acids, and VFA concentration after a 30-day ensilage
Full size table
Table 4 Solid recovery and biomass compositions of ensiled and fresh Ulva lactuca biomass
Full size table

Effect of ensilage conditions on the biomass composition

Table 4 shows that all preparations of silage without addition of the cellulase complex lost more than thirty percent of solid after 30 days of ensiling. The loss of dry matter is clearly caused by the degradation associated with unwanted microbial activity because the pH value did not drop enough to fulfill the preservation mechanism of ensilage. Since the solid lost during the ensiling period is influenced by many factors including biomass type, moisture content, and environmental conditions (Muck 1988), the data derived from one feedstock might not be totally appropriate for predicting other feedstocks. Nevertheless, the solid recovery could still be an indicator of silage quality (Kung Jr 2001). An early report published by FAO has suggested that ensilage could preserve more than 90% of the energy content from the harvested plant biomass (McCullough 1978). For some types of plant biomass such as sugar beet pulp, the loss of dry matter could be negligible if ensilage is effective (Kreuger et al. 2011; Zheng et al. 2011a, b; Liu et al. 2016). On the contrary, some studies have suggested that dry matter loss could be substantially higher (> 10%). In this study, a good quality of Ulva silage with around 92% solid recovery was obtained from the silage preparation with both cellulase complex (10 CMCU/g dry biomass) and LAB inoculum (106 cfu/g dry biomass).

All the silage preparations contained relatively less percentage of sugar composition than the raw biomass, except the fraction of rhamnose and arabinose (Table 4). Since the pH values of silages without both additives did not drop below pH 4, the decrease in sugar content might be attributed to the consumption and degradation caused by microbial activity. The degradation of carbohydrates commonly happens in land plants and might be also applicable to macroalgae, for example, the loss of starch content during the postharvest processing (Finger et al. 1999). The data also show that the content of WSC is relatively low in the fresh biomass. The increased content of WSC in all silage preparations might have resulted from the partial breakdown of the carbohydrate structure during the ensilage period, especially the silage prepared with addition of the cellulase complex. The partial breakdown of cell components could also support better ensilage by increasing the accessibility of the cell walls and cytoplasmic carbohydrates for LAB which has cell-surface enzyme complexes that can degrade and utilize oligo- or poly-saccharides (Siezen et al. 2006).

The lignin and ash content in the recovered solid increased for all prepared silage in comparison with fresh biomass (Table 4). Since the content of each composition is a relative value to the dry matter, the increase in lignin and ash might have resulted from the loss of other dry matters such as carbohydrates and protein.

Effect of ensilage on dilute acid hydrolysis of Ulva silage

Normally, pretreatment is employed to improve the efficiency of chemical or enzymatic hydrolysis. Several pretreatment methods, including chemical, physical, and biological approaches, have been developed to increase fermentable sugar yields from lignocellulosic biomass. Ensilage could also be employed as an approach for biological pretreatment of biomass (Richard et al. 1998). For example, reports published by Ambye-Jensen, Morten et al. suggest that ensilage could be used as a biological pretreatment for grass (Ambye-Jensen et al. 2013a, b) and decrease the required temperature in the hydrothermal pretreatment of wheat straw (Ambye-Jensen et al. 2013). Another benefit is that the application of ensilage can improve the enzymatic digestibility of sugar beet pulp (Zheng et al. 2011a, b). Although Ulva species do not contain a significant amount of lignin which might hinder the efficiency of hydrolysis, high temperature is usually required to release the fermentable sugars from Ulva biomass (Choi et al. 2012; Hamouda et al. 2016). Therefore, diluted acid hydrolysis at high solid loading (15% w/w) and low temperature (80 °C) was performed to examine the pretreatment effect of ensilage on Ulva lactuca biomass. Figure 2 shows that under all tested conditions the silage could produce more reducing sugar than fresh biomass after dilute acid hydrolysis. The best reducing sugar yield was around 155 mg/g dry biomass. Because the yield of reducing sugar was comparable to the data published in other reports (Table 5), the result suggests that ensilage could be employed as an integrated green approach to simultaneously preserve and pretreat Ulva biomass. The higher reducing sugar yield means that the ensilage did reflect the pretreatment effect and decreased the required temperature and acid concentration of diluted acid hydrolysis.

Fig. 2
figure 2

Reducing sugar yield from dilute acid hydrolysis of Ulva biomass at different conditions

Full size image
Table 5 Comparison of sugar yields reported by different hydrolysis approaches and conditions
Full size table

Fermentability of Ulva silage hydrolysate

In order to test the fermentability of Ulva silage, lactic acid fermentation was performed after enzymatic hydrolysis of the neutralized acid hydrolysate. The result indicated that the additional enzymatic hydrolysis could increase the reducing sugar yield from 155 to 198 mg/g dry biomass. The yield of reducing sugar was comparable to previous reports published by Trivedi et al. (2013) and Kim et al. (2011). The results also suggested that the hydrolysate could be used as a substitute of glucose in MRS medium and supported lactic acid fermentation (Fig. 3). During the fermentation period, glucose was completely consumed in 24 h and other sugars were only partially consumed. The yield of lactic acid was around 0.58 g per g of consumed reducing sugars. The results derived from this study suggest that the ensilage could be a useful approach to preserve and pretreat Ulva biomass for fermentable sugar production. Additionally, the fermentation by-product including unfermented sugars, lignin, and microbial biomass could also be a good substrate for anaerobic digestion. The integration of fermentation and anaerobic digestion allows for complete utilization of Ulva silage with a better energy balance (Alrefai et al. 2017).

Fig. 3
figure 3

Lactic acid fermentation profile of Ulva silage hydrolysate

Full size image

Conclusion

In this study, ensilage was successfully employed as a green method to simultaneously preserve and pretreat macroalgae Ulva lactuca by adding cellulase complex and LAB inoculum. The silage of Ulva lactuca retained more than ninety percent of solid and most of the carbohydrates at the end of ensiling period. The ensiled Ulva lactuca showed higher digestibility than fresh biomass by diluted acid hydrolysis at moderate temperature and high solid loading. Additionally, the hydrolysate of Ulva silage was also proven to be fermentable by LAB for lactic acid production. Since the successful preparation of Ulva silage is relatively limited, the results derived from this study could be useful for the future development of utilizing Ulva biomass as feedstock for bioproductions.