Screening of white‐rot fungi for bioprocessing of wheat straw into ruminant feed

In this study, the biological variation for improvement of the nutritive value of wheat straw by 12 Ceriporiopsis subvermispora, 10 Pleurotus eryngii and 10 Lentinula edodes strains was assessed. Screening of the best performing strains within each species was made based on the in vitro degradability of fungal‐treated wheat straw.


Introduction
Exploring alternative feed ingredients for ruminant nutrition is important, not only for a sustainable animal production, but also to further ensure human food security. Agricultural biomass, such as wheat straw, is an attractive choice for this purpose. The use and challenges in the processing of this highly lignified biomass into a more digestible animal feed has been addressed in many studies (Shrivastava et al. 2014;Van Kuijk et al. 2015a). Different pretreatment methods have been used, including physical, chemical and biological methods (Sarnklong et al. 2010;Isroi et al. 2011). The goal of any pretreatment method is to break the lignin barrier and make cellulose and hemicellulose more accessible to fermentation in the rumen (Sarnklong et al. 2010). Each method, however, has a varying effect on the composition and structure of cellulose, hemicellulose and lignin (Mosier et al. 2005;Hendriks and Zeeman 2009). Biological methods, including the use of fungi have attracted the interest of many researchers due to an increasing global demand for eco-friendly approaches (Dinis et al. 2009;Shrivastava et al. 2011;Tuyen et al. 2012;Sharma and Arora 2014).
Lignin has a negative impact on the degradation of fibre in ruminants (Moore and Jung 2001). Degradation of lignin is a chemically difficult process since these polyphenolic compounds have a dense three-dimensional, highly branched structure that is linked with hemicellulose (Kubicek 2013). White-rot fungi are the only organisms capable of degrading lignin aerobically to CO 2 and water (Hatakka and Hammel 2010). This unique ability makes them favourable as candidates for the bioprocessing of highly lignified agricultural biomass into ruminant feed. However, not all fungal species exhibit a preferential degradation of lignin, i.e. selectively degrade lignin while utilizing low amounts of the structural carbohydrates. Tuyen et al. (2012) showed that out of 11 fungal species tested, only 468 Ceriporiopsis subvermispora (Pilat) Gilb. & Ryvarden 1985, Lentinula edodes (Berkeley) Pegler 1975 and Pleurotus eryngii var. eryngii (DC.) Qu el 1872, showed a high lignin to cellulose loss ratio, which increased the in vitro degradability in rumen fluid of the treated wheat straw by 20-60% after 7 weeks of solid-state incubation.
Bioprocessing of agricultural biomass, using white-rot fungi, is a complex process which is influenced by many factors, such as fungal species and strain, substrate and culture conditions (Van Kuijk et al. 2015a). Several studies worked on the optimization of the process, such as exploring the best fungi-substrate combinations (Tuyen et al. 2013), varying particle sizes of the substrate (Van Kuijk et al. 2016) and supplementing additives such as manganese and linoleic acid to the substrate (Van Kuijk et al. 2015b). However, little work has been done on selecting the best fungal strain to maximize the valorization of biomass for ruminant use. Fungi are diverse in their lignocellulose deconstruction mechanisms (Cragg et al. 2015), which emphasize the need to not only investigate and screen for the best performing fungal species, but also different strains within a particular species. Most studies selected fungal strains based on their enzyme activity (Mata and Savoie 1998;Cavallazzi et al. 2004). However, correlating qualitative enzyme activities with the selectivity of a fungus was difficult (Sharma and Arora 2010). Recently, we showed differences between two strains of C. subvermispora in improving the in vitro degradability of wheat straw, as well as in their growth and lignin-degrading enzyme characteristics (Nayan et al. 2017). The ultimate goal of the ongoing research is to study the variation among fungal strains, which can be a valuable input for future breeding activities to expand the potential of fungi in the bioprocessing of agricultural biomass into ruminant feed. Therefore, the degradability of the substrate, as determined by the in vitro gas production (IVGP)  was used as the main screening criterion.
The aim of the present study was to assess the biological variation among different strains within three species of white-rot fungi -C. subvermispora, P. eryngii and L. edodes, with respect to the formation of fungal biomass in time and the effect on IVGP. Three of the best strains for each species were selected based on IVGP and further compared for differences in the gas production kinetics and other fermentation characteristics; as well as changes in the mass balance of the wheat straw.

Preparation of fungal strains and spawn
Different strains of three white-rot fungi species: C. subvermispora (CS; 12 strains), P. eryngii (PE; 10 strains) and L. edodes (LE; 10 strains) (Table 1) from the collection of the Plant Breeding Group, Wageningen University & Research, the Netherlands, were used. All strains were cultured on malt extract agar (malt extract 20 g l À1 , KH 2 PO 4 0Á5 g l À1 , MgSO 4 Á7H 2 O 0Á5 g l À1 and Ca (NO 3 ) 2 Á4H 2 O 0Á5 g l À1 with a pH of 5Á4) and incubated at 24°C until the mycelia colonized most of the agar surface. The spawn of each strain was prepared by placing a piece of the agar culture (1Á5 9 2Á0 cm) into a container with sterilized (20 min, 121°C) sorghum grains and incubated at 24°C for 5 weeks.

Substrate preparation, fungi inoculation and sampling
A bale of organic wheat straw (Triticum aestivum L.) (300 kg; DM, dry matter content 880Á4 g kg À1 ) was purchased from a local farmer in the Netherlands and used for all incubations. Due to the limited capacity of the incubation chamber, treatments were conducted in three independent but adjacent periods with only one species (with different strains) for each period. Straw was chopped into c. 3 cm pieces and soaked in water for 3 days at room temperature. After draining the excess of water for 5 h, the moisture content of each batch of straw was recorded. The wheat straw was then distributed into 185 9 185 9 78 mm micropropagation containers (Combiness, Nevele, Belgium), which is equipped with a depth-filtration strip on the lid, and the amount of wet straw in each container was adjusted to contain 90Á2 AE 0Á3 g of DM for all batches.
All containers were autoclaved at 121°C for 1 h and allowed to cool overnight at room temperature. Pre-and postautoclaved wheat straw samples were taken for comparing the DM content. The handling and processing of the wheat straw was maintained in the same order throughout the experiment. The autoclaved straw was aseptically inoculated with the previously prepared spawn for each strain/species at 10% of the dry weight of the straw. Containers with inoculated straw and autoclaved straw (control) were incubated aerobically in triplicate at 24°C for 7 weeks in a climate-controlled chamber. Weekly samples were weighed, thoroughly mixed and 5 g of fresh sample was taken for pH measurements. The remaining sample was freeze-dried and ground over a 1 mm sieve using a cross beater mill (100AN; Peppink, Olst, the Netherlands).

Ergosterol estimation
Fungal biomass was estimated with an ergosterol assay (Niemenmaa et al. 2008). In brief, 200 AE 10 mg sample was weighed into a glass tube and after extraction with 10% (1 : 9) KOH/methanol solution for 10 min, the tube was saponified at 80°C for 60 min. After cooling to room temperature, 1 ml distilled water and 2 ml hexane was added before each tube was thoroughly shaken. The hexane layer was subsequently collected and the extraction steps were repeated. The pooled hexane layer was dried in a vacuum evaporation system (Rapidvap, Kansas, MO). The extracted ergosterol was re-dissolved in methanol and the solution was filtered into a high performance liquid chromatography (HPLC) vial for Waters HPLC-PDA analysis (Alliance HPLC system, Milford, MA). A reversed phase C18 column (250 9 4Á6 mm, Phenomex aqua 5 lm) was used in the HPLC instrument and the liquid phase was 90% methanol and 10% (1 : 1) 2-propanol/hexane. Cholecalciferol (vitamin D 3 ) was used as an internal standard. The ergosterol peak was detected at 280 nm.

In vitro gas production
All samples were subjected to IVGP, expressed as ml g À1 organic matter (OM) , where 0Á5 AE 0Á01 g of sample was incubated in 60 ml of buffered rumen fluid at 39°C. The incubation was conducted in 250-ml bottles (Schott, Mainz, Germany) for 72 h and the gas production was registered automatically. At the end of the incubation, rumen fluid samples from each bottle (600 ll) were taken for the determination of volatile fatty acids (VFA) and ammonia-N (NH 3 -N). The remaining content in the bottles was filtered through a crucible and dried in an oven at 103°C to determine the DM content (ISO 6496, 1999) and the ash content in a furnace at 550°C for 3 h (ISO 5984, 2002). The OM digestibility was calculated as a percentage of digested OM from the preincubated sample.
Only samples of selected fungal strains/species (three best performing strains from each species) were analysed for VFA and NH 3 -N. Individual VFA concentrations were separated by gas chromatography using hydrogen as carrier gas, while reaction of ammonia with phenol and hypochlorite was measured spectrophotometrically at 623 nm for NH 3 -N determinations. The kinetic parameters were determined by fitting the gas production data to a biphasic model (Groot et al. 1996), representing fermentation of the soluble (phase 1; 0-4 h) and the nonsoluble fractions (phase 2; 4-48 h) according to Tuyen et al. (2012). Only parameters of phase 2 (A 2 , B 2 , C 2 , t Rm2 , R m2 ) are presented here. A 2 is the asymptotic gas production (ml g À1 OM) of phase 2; B is the half time of the maximum gas production (h); C is a parameter to determine the steepness of the curve; t Rm is the time of the maximum fractional rate of substrate degradation (h); and R m the maximum fractional rate of substrate degradation (h À1 ).

Chemical analysis
All freeze-dried samples were analysed for DM and ash content. Only selected fungal strains/species samples were analysed for nitrogen and fibre content. Nitrogen content was determined by the Kjeldahl method (ISO 5983, 2005) and crude protein was calculated as N 9 6Á25. Neutral detergent fibre (NDF) was determined using a heat-stable amylase (thermamyl) and alcalase, using the standard procedures of Van Soest et al. (1991). Acid detergent fibre (ADF) and acid detergent lignin (ADL) were also determined using the standard method (Van Soest et al. 1991). Hemicellulose was calculated as the difference between NDF and ADF, while cellulose was calculated as the difference between ADF and ADL. Absolute amount (g) of each component was calculated from the remaining amount (g) of freeze-dried sample which was corrected for the DM content.

Statistics
Data were analysed per species using the general linear model in SAS 9.3 (SAS Institute Inc., Cary, NC, USA), followed by a post hoc multiple comparison using least significance differences. The statistical model was as follows: where Υ ij = response variable, l = overall mean, ST i = the effect of strain i, s j = effect of week j, (ST 9 s) ij = interaction of strain i and week j, and e ij = residual error with a mean of 0 and variance r 2 . For assessment and comparison of selected strains across species, nested analysis of variance without time interaction was conducted using the following statistical model: where Υ ijk = response variable ijk, l = overall mean, SP i = the effect of species i, ST j(i) = the effect of strain j nested within species i, s k(i j) = effect of week k, and e ijk = residual error with a mean of 0 and variance r 2 . SP i was considered a fixed effect, ST j(i) and s k(i j) as random effects. The CONTRAST statement was also used in SAS to allow comparison among different species group. Pearson Product-Moment Correlation (r) coefficients were determined among the measured variables. Probability values below 5% were considered significant. The ergosterol data were fitted to a polynomial regression models. A linear model was fitted first to the ergosterol data of a strain where after a quadratic term was added. If the quadratic term was significant, a cubic term was added to the model. The model with the highest significant (P < 0Á05) coefficient was taken as the best fit model. The rate of ergosterol changes over time (d Erg ) was calculated by substituting week 7 (x 2 = 7) and week 0 (x 1 = 0) into respective polynomial functions, f (x), and dividing the value by 7.

Results
Fungal growth and changes in the pH of substrate Figure 1 shows the pattern of change in ergosterol content in time for all fungal strains within each species. In general, the ergosterol concentrations in fungal-treated straw showed a curvilinear pattern for all strains, with a lag phase up to 3 weeks for L. edodes. The ergosterol data of each strain were fitted to a polynomial regression model to describe the growth pattern and to estimate the changes in ergosterol (d Erg ) over time ( Table 2). Most of C. subvermispora strains showed quadratic and linear growth patterns. All P. eryngii and L. edodes strains (except LE1 and LE9) mostly followed a cubic function. Four C. subvermispora strains (CS3, CS7, CS4 and CS11), showed a significantly higher (P < 0Á001) d Erg compared to other CS strains and CS1 had the lowest d Erg . Among P. eryngii strains, PE5 had the highest d Erg . High d Erg values among L. edodes strains were observed for LE10, LE1, LE2 and LE3. Two L. edodes strains (LE4 and LE5), showed irregular changes with the ergosterol contents of 271Á4 and 449Á4 lg g À1 , respectively, despite no apparent mycelium being observed after 7 weeks. These data were excluded from the analysis. Typical changes in pH for all fungal strains/species are shown in Fig. 2. Ceriporiopsis subvermispora strains significantly (P < 0Á001) decreased the pH of wheat straw from 5Á00 to 3Á52, while L. edodes strains (excluding LE4 and LE5) decreased (P < 0Á001) the pH to 3Á95. For P. eryngii strains, there were significant (P < 0Á001) increases in pH up to week 3, before it significantly decreased to pH 4Á46 at week 7. After 7 weeks of colonization, the pH of wheat straw treated with LE4 and LE5 increased to pH 6Á87 and 7Á17, respectively.
In vitro gas production IVGP was used to select the best performing fungal strains for each species. Overall, 17 out of 32 fungal strains tested, significantly (P < 0Á05) improved the IVGP of wheat straw after 7 weeks of colonization (Fig. 3). Seven strains of C. subvermispora significantly (P < 0Á05) increased the IVGP of the wheat straw, with CS1, CS6, CS12, CS8 and CS5 being highly significant (P < 0Á001) compared to untreated straw. Fermentation of CS1-treated wheat straw produced most gas (313Á2 ml g À1 OM), followed by CS6 and CS12 (293Á6 and 284Á2 ml g À1 OM, respectively). Among P. eryngii strains, three strains -PE6, PE2 and PE3 significantly (P < 0Á05) increased the IVGP of the wheat straw. PE6 had the highest IVGP (263Á0 ml g À1 OM), followed by PE2 and PE3 with a mean IVGP of 257Á9 and 252Á4 ml g À1 OM, respectively. Seven strains of L. edodes significantly (P < 0Á05) increased the IVGP of the straw where five of them were highly significant (P < 0Á001). LE8 showed the highest IVGP (297Á7 ml g À1 OM), followed by LE7 (289Á4 ml g À1 OM) and LE10 (287Á3 ml g À1 OM). LE4 and LE5, for which irregular changes in growth and pH were observed, showed a significantly (P < 0Á05) lower IVGP compared to the untreated straw.

471
Three strains from each species with the highest IVGP values were selected for further assessment and comparison across species. The selected strains (in order of IVGP values) were CS1, CS6 and CS12 for C. subvermispora; PE6, PE2 and PE3 for P. eryngii; and LE8, LE7 and LE10 for L. edodes.

Gas production kinetics and fermentation characteristics
The kinetic parameters of the second (nonsoluble) phase of the gas production curves (4-48 h) for the selected fungal strains are summarized in Table 3. Compared to the untreated wheat straw, all fungal-treated wheat straw showed significantly (P < 0Á01) higher A 2 values (except PE2 and PE3) and significantly (P < 0Á001) lower B 2 values, indicating a higher and faster fermentation of the substrate. High A 2 and low B 2 were accompanied by significantly (P < 0Á01) lower t Rm2 and higher R m2 values for all fungal-treated straw samples, compared to the control. Among different species, C. subvermispora strains showed a significant contrast (P < 0Á001) to the P. eryngii strains in all kinetic parameters. Other contrast wise comparisons among different fungal species showed significant differences (P < 0Á05), except for A 2 values for C. subvermispora vs L. edodes strains, and R m2 for P. eryngii vs L. edodes strains. The increases in IVGP correlated with the increases in in vitro OM digestibility (r = 0Á81; P < 0Á001).
All fungal-treated wheat straw showed a decrease in acetate (except for CS12) during fermentation in rumen fluid, but an increase in propionate production (Table 3). This observation was also reflected by the changes in acetate/propionate ratio with the lowest values observed for L. edodes strains, followed by P. eryngii and C. subvermispora strains. All species were significantly (P < 0Á001) different to each other in their acetate/propionate ratio. There was a significant (P < 0Á001) decrease in NH 3 -N production for C. subvermispora (except CS12) and L. edodes strains. The changes observed in all P. eryngii strains, however, were not significant. L. edodes strains showed the lowest NH 3 -N, followed by C. subvermispora and P. eryngii strains. Contrast wise, strains were significantly (P < 0Á001) different from each other.

Mass balance
The control, untreated wheat straw contained 213Á8 g kg À1 of DM and 33Á3 g kg À1 ash (on DM basis) with an absolute OM amount of 87Á2 AE 0Á3 g. All mass balance data were presented as g per 100 g of this starting OM. The mass balance data of the selected fungal species/strains are summarized in Table 4. All fungal pretreatments decreased the amount of OM of the wheat straw by 1Á6-7Á8%, with the decreases in CS6 and CS12-treated straw were not significant. The fungal pretreatment resulted in 6Á3 to 30Á1% increase in the amount of crude protein. All fungi significantly (P < 0Á001) degraded ADL and hemicellulose in the wheat straw. The largest decrease in the amount of ADL and hemicellulose was observed in CS1-treated straw with 52Á2 and 49Á7% lower ADL and hemicellulose, respectively, compared to the control. The absolute amount of cellulose, however, did not change significantly, except for CS1. Although CS1-treated straw showed a decrease in cellulose, it contained the highest 'enrichment' of total carbohydrates, i.e. highest carbohydrate to lignin ratio (C/L) in the remaining material. Wheat straw treated with C. subvermispora strains (except CS12) showed an overall high C/L, followed by L. edodes and P. eryngii strains. The total mass which was not identified ranged from 7Á8-24Á4% from the total amount of OM.

Discussion
Values with different superscripts within column of the same species are significantly (P < 0Á05) different. Erg week 7 : Ergosterol amounts after 7 weeks of colonization; d Erg : Calculated rate of ergosterol changes (lg g À1 per week) over 7 weeks using respective regression functions. LE4 and LE5 are omitted from the table due to irregular changes. *Regression coefficients of the polynomial functions with intercept b 0 with respective coefficient of determination (R 2 ) and root mean square error (RMSE).
Zakowska 2009), however a longer lag phase in our study contradicts the assumption made for these liquid culturebased models that growth should mainly occur at the beginning of the colonization. Fungi undergo a lag phase as a physiological preparation for an exponential phase, which depends on inoculum and available nutrients (Griffin 1996). A longer lag phase was observed for CS1 and CS2, compared to a previous study which used the same strains grown under similar conditions (Nayan et al. 2017). This was expected as the present study used 2Á6 times more substrate (fresh weight basis) compared to the latter study. The present study shows that L. edodes strains had a longer lag time compared to C. subvermispora and P. eryngii strains. For P. eryngii strains, there was an indication of a decrease in fungal biomass after week 5. Since ergosterol is relatively stable in substrate (Mille-Lindblom et al. 2004), this observation indicates that the fungi recycled mycelium, probably due to a decline in their ability to obtain nutrients from the substrate. A continuing decrease in the pH of the substrate (Fig. 2) signals the generation of metabolites that lower the pH, indicating a persistent formation of new biomass at the expense of old biomass (Falconer et al. 2005). Meanwhile, it is difficult to provide an explanation for the irregular changes observed in the ergosterol contents of LE4-and LE5-treated straw. A visual observation of these samples provided no indication for the presence of green molds (a sign of contamination). An increase in pH to near neutral and slightly alkaline for LE4 and LE5, respectively, might indicate that these fungi did not grow well on wheat straw or a probable bacterial contamination due to poor fungal growth. A decrease in substrate pH during the fungal colonization is a characteristic of lignin-degrading enzyme secretions (Liers et al. 2011), in addition to the production of organic acids (Hofrichter et al. 1999).
Degradability of treated wheat straw in rumen fluid, as measured using the IVGP technique , was used here (Fig. 3) as a screening criterion for the best performing fungal strains. This accurate time-related technique has been widely used in feed evaluation research, including in assessing the degradability of fungal-treated agricultural biomass (Tuyen et al. 2013;Van Kuijk et al. 2015a). Other important parameters have also been used as screening criteria such as selective delignification which resulted in a high C/L, growth characteristics and enzyme activities (Li et al. 2008;Chang et al. 2012;Cruz Ram ırez et al. 2012). Our choice of using the IVGP for strains screening is based on the fact that it provides a good indication for the improvement in the nutritive value of wheat straw in a convenient and robust manner. Moreover, an increase in IVGP has been correlated with an increase of C/L in the substrate (Karunanandaa and Varga 1996a;Tuyen et al. 2012). A selective fungus is desirable for bioprocessing of wheat straw into ruminant feed.
CS1, PE2 and LE7 were the same fungal strains as used in previous studies (Tuyen et al. 2012;Van Kuijk et al. 2015a;Nayan et al. 2017). The IVGP of CS1-treated wheat straw in the present study (313Á2 ml g À1 OM) was slightly higher than previously reported (300-305 ml g À1 OM). The IVGP of PE2-treated straw (257Á9 ml g À1 OM) was also within the range of previously reported for wheat straw treated with this strain (250-260 ml g À1 OM). The IVGP of LE7-treated wheat straw reported here (289Á4 ml g À1 OM) was slightly higher than that of Tuyen et al. (2012) (270-280 ml g À1 OM). Van Kuijk et al. (2015a) reported an IVGP value of 311Á3 ml g À1 OM when wheat straw was treated with this strain for 12 weeks. All selected IVGP data were subjected to a biphasic model (Groot et al. 1996) to assess the kinetic parameters of the gas production curves. This curve-fitting tool has been shown to be useful in comparing the gas production profiles of wheat straw treated with two strains of C. subvermispora, although the total IVGP values of both fungi were not significantly different (Nayan et al. 2017). We present only parameters of the second phase as it is related to the fermentation of structural carbohydrates (Groot et al. 1996;Cone et al. 1997). Results show that wheat straw treated with C. subvermispora strains had an overall better gas production profilehigher A 2 and lower B 2 , followed by L. edodes and P. eryngii strains. Fermentation of a low quality, highly fibrous feed results in an increase in acetate production, while an increase in rapidly fermentable carbohydrates will result in an increase in propionate production (Van Houtert 1993). Other factors that can also affect the acetate/propionate ratio, including substrate composition and availability, as well as microbial species present (Dijkstra 1994). An increase in IVGP indicates that the fungal pretreatment modifies the complexity of the cell wall and transforms the wheat straw into a more fermentable fibre source. The effect of a change from complex to easily digestible fibre sources (by means of chemical pretreatment) on the VFA production has been demonstrated by Griffith et al. (2016). They reported a similar decrease in acetate and acetate/propionate ratio, as well as an increase in propionate molar proportion. Similar results were also reported for rice straw treated with Cyathus stercoreus Total in vitro gas production (IVGP) of wheat straw treated with different fungal strains (within species) after 7 weeks of colonization. The IVGP values were sorted ascendingly within each species. Dashed-line marks the gas production of the control (untreated straw) and error bars indicate the SD. Asterisks show significant differences from the control at ***P < 0Á001, **P < 0Á01 and *P < 0Á05.
475 Table 3 The effects of wheat straw treated with different fungal strains/species on the in vitro organic matter (IVOM) digestibility, kinetic parameters of phase 2 of the gas production, volatile fatty acids (VFA) and ammonia production after 7 weeks Values with different superscripts within column are significantly (P < 0Á05) different. RMSE: root mean square error; A 2 : asymptotic gas production (ml g À1 OM) at phase 2; B 2 : half time of the maximum gas production (h); C 2 : parameters determine curvature of the graph; t Rm2 : time of the maximum fractional rate of substrate degradation (h); R m2 : maximum fractional rate of substrate degradation (h À1 ); HAc: acetic acid; HPr; propionic acid; HBu: butyric acid; TVFA: total volatile fatty acids including iso-acids; NH 3 -N: ammonia-N.
Values with different superscripts within column are significantly (P < 0Á05) different. RMSE, root mean square error; OM, organic matter; Cell, cellulose; Hcell, hemicellulose; ADL, acid detergent lignin; CP, crude protein; N.I., amount of nonidentified OM (not measured); C/L, total carbohydrate (cellulose + hemicellulose) to lignin ratio. (Karunanandaa and Varga 1996b). However, Karunanandaa and Varga (1996a) reported an increased acetate and acetate/propionate ratio in rice straw treated with Phanerochaete chrysosporium. A likely explanation is that this fungus is considered to be less selective and reduces the amount of carbohydrates considerably (Tuyen et al. 2012). A lower NH 3 -N was also reported for fungal-treated rice straw, compared to a control (Karunanandaa and Varga 1996b). During the in vitro incubation, a decrease in NH 3 -N concentration in the rumen fluid was followed by a concurrent increase in microbial N (Cone et al. 1997), suggesting that the fungal-treated straw may lead to a more efficient microbial N conversion in the rumen.
The selectivity of fungi in degrading lignin, as indicated by the 'carbohydrate enrichment' or C/L in the remaining material, was significantly correlated (r = 0Á64, P < 0Á001) to the IVGP values. Although the chemical analysis was not carried out for all 32 strains involved in this study, these correlations indicate the potency of the IVGP technique as a tool in screening the best performing fungal strains. One of the limitations of using the Van Soest et al. (1991) method to analyse the cell wall composition of the fungal-treated material is that, it does not differentiate cellulose in the substrate from chitin in the fungal biomass (cellulose-like structure with N-acetylglucosamine monomers). This is because the hydrolysis of chitin polymers requires a stronger concentrated acid (Einbu and V arum 2008). This may lead to an overestimation of the cell wall composition, which can explain the slight increase in cellulose observed for CS6-and PE6-treated wheat straw. Nonetheless, in the context of fungal pretreatment of lignocellulosic biomass, a selective fungus should be defined for its ability to degrade lignin and enrich the biomass with high amounts of available carbohydrates in the end products, i.e. high C/L in the remaining materials. Meanwhile, the increases of CP in fungal-treated straw may erroneously suggest that these fungi could fix atmospheric N, which is not true. The relative enrichment of CP is due to degradation of the cell wall components, leading to a concentration of N content (Van Kuijk et al. 2015c).
The present study also provides fundamental information for breeding purposes. For example, CS1 is one of the strains in the C. subvermispora collection that has no clamp connection and, therefore, is a monokaryon (A.S.M. Sonnenberg, pers. observ., 2017). Tello et al. (2001) showed that this monokaryon produced less biomass in liquid cultures and showed a higher lignindegrading enzyme activity compared to dikaryons. In our study, the low growth rate in the monokaryotic CS1 resulted in a high C/L, which corresponds to the higher IVGP of the treated wheat straw. Considering all these features, it might be worthwhile to recover the constituent nucleus of all dikaryotic strains and evaluate all recovered nuclei as monokaryons in future breeding activities.
Results clearly show the variation among strains within species in terms of growth patterns, fibre degradation capabilities, as well as improving the in vitro digestibility of wheat straw in rumen fluid. Nevertheless, there are some limitations in this screening study. White-rot fungi are known to vary in their performance on different lignocellulosic biomasses (Tuyen et al. 2013). Hence, the selected strains in this study will not necessarily be applicable for other agricultural biomasses. The wheat straw used in the study was heterogeneous in nature, which consisted of different parts of the plant. No prior separation of leaves and stalks was carried out, which may lead to variation in the chemical composition of the straw used. Nevertheless, thorough mixing was carefully done for each sub-batch of the straw to avoid biases. The variation in the wheat straw used, i.e. different cultivars and maturity stages with different culture conditions, may also influence the outcome of future studies using the selected fungal strains. In addition, under various treatment and environmental conditions, these fungi can differ in their enzyme activities and also the production of diverse secondary metabolites, which may affect the subsequent enrichment of the biomass and its degradability by the rumen microbes. This study presents a promising step towards capitalizing on the unique abilities of whiterot fungi to bio-transform high lignin-containing agricultural biomasses into digestible animal feed. Moreover, this finding could be further broadened into various applications, such as in the production of biofuels and industrial bioremediation.
In conclusion, there is a large biological variation among fungal strains within the same species for their capacity to degrade lignin in wheat straw. Ceriporiopsis subvermispora strains, especially strain 1, showed an overall high potential in improving the nutritive value of wheat straw, followed by L. edodes and P. eryngii strains. The outcome of this study underlines the importance to also assess different strains of the same species and select the best strain for bioprocessing of biomass into ruminant feed. The variation among fungal strains offers opportunities for breeders to map genomic regions that explain the difference in extent and selectivity of lignocellulose degradation. This knowledge can be used to stack favourable alleles and generate superior strains. the project "More Meat and Milk from Straw" which is sponsored by DEKA, ForFarmers and the Victam Foundation. The authors would like to acknowledge the scholarship provided by the Ministry of Higher Education of Malaysia and the Universiti Putra Malaysia. All funding bodies had no involvement in planning and conduct of the research. The authors also acknowledge Mrs. Saskia van Laar, Mrs. Xuan Huong van der Schans-Le and Ms. Laura Berns for their technical assistance.