Effects of endophytic and saprophytic fungi on in vitro methanogenesis

Abstract

The predominance of pasture in the diet of ruminants in New Zealand presents specific challenges for mitigation of enteric methane emissions. However, there is preliminary evidence from both laboratory and animal-based studies that some fungi may suppress enteric methanogenesis. This study determined the antimethanogenic potential of a range of endophytic (Neotyphodium species) and saprophytic (Geotrichum, Monascus, Mortierella and Penicillium species) fungi. Candidate fungi were selected on the basis of their production of various classes of secondary compounds and screened in an in vitro batch-culture fermentation assay. No strains of endophyte in ryegrass or tall fescue suppressed methanogenesis compared with their endophyte-free controls. The supernatant fractions from three strains of Mortierella wolfii were the most promising fungi identified. They suppressed methanogenesis as effectively as an inhibitor of methanogenesis (bromoethane sulphonate at 30 µmoles/L), without also reducing overall fermentation. These strains should be investigated further by screening fresh cultures grown under contrasting conditions to establish some variation in putative bioactive compounds.

Keywords:

• in vitro methanogenesis
• perennial ryegrass
• tall fescue
• endophyte fungus
• saprophyte fungus
• Mortierella wolfii

Introduction

Feed type affects enteric methane production and forage diets result in greater amounts of methane produced than concentrate diets (Johnson & Johnson 1995). The predominance of pasture in the diet of ruminants in New Zealand presents some specific challenges for mitigation of enteric methane emissions, but also opportunities for some unique solutions. Plant and plant-related attributes that offer mitigation potential would be highly relevant for reducing emissions of methane from forage-dominant diets per kg of dry matter consumed and/or per unit of animal product.

There is preliminary evidence from both laboratory and animal-based studies that enteric methanogenesis may be suppressed by fungi. In a study using an in vitro continuous fermenter, a selected strain of the endophytic fungus (Neotyphodium coenophialum) in tall fescue (Festuca arundinacea) suppressed methanogenesis under a particular set of fermenter conditions (Vibart et al. 2007). Endophytic fungi can have inhibitory effects on other fungi (Christensen 1996) and they have been shown to affect soil biota in a way that slows the decomposition of soil organic matter (Omacini et al. 2004; Antunes et al. 2008; Ryan et al. 2008). The presence of the spoilage fungus Penicillium roquefortii in silage reduced methane production in a dose-dependent manner (Mauruschat 1996). While the effect in that study was probably due to the suppression of fermentation rather than the presence of a specific methane inhibition agent, mouldy silage fed to cattle also reduced methane production to about half of the output normally produced by cattle (Clark & Krause 2007). Methane emissions from sheep grazing pastures composed of kikuyu (Pennisetum clandestinum) were on occasions considerably lower than predicted (Ulyatt et al. 2002a,b). In that study, it was expected that emissions would be higher than from ryegrass, on the basis of the higher fibre content of this subtropical species of grass (with a C4 pathway of photosynthesis, and therefore a greater concentration of cell wall carbohydrates in the dry matter compared with C3 grasses). The fact that they were lower was attributed by Ulyatt et al. (2002a,b) to the possibility of mycotoxins associated with kikuyu, although it was not clear if this was a direct effect on the animal (e.g. kikuyu poisoning, Martinovich et al. 1972) or on rumen microbes and methanogenesis.

Collectively, the evidence above warrants further investigation to establish whether endophytic or saprophytic fungi can reduce rumen methanogenesis by selectively inhibiting methanogenic Archaea, or perhaps indirectly through inhibiting rumen protozoa, fungi or bacteria that also generate hydrogen and contribute to methanogenesis. An advantage of delivery via a fungal endophyte would be the low cost and simplicity associated with the continual dietary supply of a mitigating agent with pasture. An alternative method, particularly if saprophytic fungi were shown to be antimethanogenic, could be extraction of bioactive compounds or synthesis of analogue compounds and delivery by other means, e.g. via drinking water using in-line dispensers, or via intraruminal controlled-release technology. Such methods are already used for delivery of therapeutic compounds to animals.

The objective of this study was to screen commonly occurring endophytic and saprophytic fungi to determine their potential for reducing enteric methanogenesis. Endophytic fungi of the genus Neotyphodium grow in mutualistic association with host grasses. In New Zealand pastures, the two species of fungi of importance are the perennial ryegrass endophyte (N. lolii) and the tall fescue endophyte (N. coenophialum). Within each species of endophyte, there are numerous experimental and commercial strains available. While the prior observation of an antimethanogenic effect (Vibart et al. 2007) related to one strain of tall fescue endophyte, several different strains were included in this study. Perennial ryegrass (Lolium perenne)-endophyte associations were also included because this species of grass is used more commonly than tall fescue in New Zealand (Easton et al. 2001; Easton & Tapper 2005). This enabled a test of whether any antimethanogenic effect is confined to a particular species of endophyte, and to a particular association of grass species and strain of endophyte, or is more general across different genera of grasses and fungi. For each group of fungi, the hypothesis was that the presence of the fungus would reduce methanogenesis compared with its endophyte-free or saprophyte-free control through the action of fungal metabolites.

Materials and methods

This study was conducted at AgResearch Grasslands, Palmerston North during 2009–10, using an in vitro rumen fermentation system.

Selection of candidate fungi

Two groups of fungi, endophytic and saprophytic, were chosen for this project. From these groups, 15 candidate grass-endophyte associations plus two endophyte-free controls and 10 candidate saprophytic fungi were selected for screening.

Endophytic fungi

The 17 candidates comprised 10 perennial ryegrass-endophyte associations, each in the experimental line GA66, and five tall fescue-endophyte associations, each in the cultivar Jesup, plus an endophyte-free control for each species of grass. For each species of grass, the endophyte-free and the wild type (common toxic) endophyte (producing lolitrem B, ergovaline and peramine in ryegrass, and ergovaline, peramine and lolines in tall fescue) provided negative and positive controls. The 10 candidate ryegrass-endophyte associations for screening were chosen according to two criteria: first, to represent the genetic diversity in the available endophyte strains, using a dendrogram based on genetic similarity between strains, determined from molecular marker information (prepared by Dr M Faville, AgResearch); and second, to have a set of candidates that produced the three primary alkaloids—lolitrem B, ergovaline and peramine—in as many combinations as possible. AR50/EL3, EL1, EL2, AR6 and wild type represented five out of seven possible combinations of the presence of a single, a pair and all three alkaloids. In addition, one strain that produces epoxy-janthitrems (AR37) and one strain that does not produce any of the main alkaloids (EL5) were included.

For tall fescue, the three combinations of primary interest, endophyte-free, wild type endophyte, and the commercial strain Max Q® (AR542), were similar to those used in the study reported by Vibart et al. (2007). The experimental strains AR525, AR584 and EL6 (also in cv. Jesup) were included to broaden the range of associations screened. Endophyte strain variation related primarily to the comparison between wild type which produces ergovaline and the selected strains which do not produce ergovaline.

Saprophytic fungi

Many species of saprophytic fungi can colonize and decompose conserved feed. The 10 candidates comprised three strains of each of three species, and a single strain of a fourth species (Table 1). The selection of four species for testing in the in vitro assay was confined to those fungi that have been observed in spoilage of silage (di Menna et al. 1981) and hay, and those commonly seen in association with pasture (M. di Menna, pers. comm.). Although Aspergillus fumigatus was also considered as a possible candidate it was not used because of its potential to cause mycotic infections (aspergillosis) and allergic reactions in humans and other animals. Mortierella wolfii is not known as a human pathogen but can cause mycosis in cattle (Munday et al. 2010). All the fungi tested are known to produce secondary metabolites with a range of biological activities (Storm et al. 2008).

Table 1  The species and strains of saprophytic fungi associated with spoilage of silage and selected for screening, their source and some characteristic production of secondary metabolites that may influence enteric methane production.

Fungus	Strain	Source	Characteristics of fungus
Mortierella wolfii	68/707w2	di Menna^1	Produces a protein nephrotoxin, grows at 37 °C and causes mycosis in cattle (Munday et al. 2010: abortion and pneumonia; encephalitis when injected into brain of mice).
	69/3891	di Menna
	74/2649	di Menna
Penicillium roquefortii	A2	di Menna	Regarded as a complex of three species (Boysen et al. 2000: P. roquefortii, P. paneum and P. carneum—the first two being associated with mouldy silage P. carneum mainly with bread). Produces extensive range of mycotoxins including: • Roquefortine (several) • Mycophenolic acid • Eremofortin C • PR toxin • Penitrem A
	G2	di Menna
	I2	di Menna
Monascus ruber	15220	ICMP^2	A red mould used in fermented food products that produces Citrinin (active against G+ bacteria) and Monacolin K = lovastatin (Wolin & Miller 2006: statins may affect methanogenesis, but also growth of anaerobic fungi).
	503.70	CBS^3
	554.76	CBS
Geotrichum candidum	240.62	CBS	May produce clavine alkaloids (Sallam et al. 1970).
^1 Dr Margaret di Menna, AgResearch, Hamilton.
^2 International Collection of Microorganisms from Plants, Landcare Research, Auckland, New Zealand.
^3 Centraalbureau voor Schimmelcultures, Utrecht, The Netherlands.

Collection of samples and preparation for screening

Endophytic fungi

Established plots of perennial ryegrass at AgResearch Lincoln, and tall fescue at AgResearch Kerikeri, infected with the candidate strains of endophyte were sampled in early autumn 2009. At each site, subsamples of each grass-endophyte association and the endophyte-free controls were taken from each replicate plot by cutting herbage as close to ground level as possible. The replicate subsamples were composited into a single sample of each association providing 200–300 g dry weight for bioassay. Samples were frozen as soon after collection as possible and stored until used for the bioassay or chemical analysis. Samples of the frozen grass-endophyte combinations were freeze-dried and ground (Cyclotec 1093 sample mill, Tecator, Sweden) to pass through a 1 mm sieve.

Saprophytic fungi

The cultures of 10 strains from four candidate species of saprophytic fungi were obtained from culture collections, held in New Zealand (AgResearch, Ruakura; International Collection of Microorganisms from Plants, Landcare Research, Auckland) or in The Netherlands (Centraalbureau voor Schimmelcultures, Utrecht). The cultures received from the culture collections were transferred to potato dextrose agar (PDA) and subcultured on this medium for 3–14 days (reflecting different radial growth rates). Agar plugs taken from the edge of colonies growing on PDA were macerated in sterile water and used to initiate seed cultures grown in 50 ml of Czapek Dox yeast extract liquid medium (CDYE; Sumarah et al. 2005) in 250 ml flasks at 25 °C for 3 days. Macerated fungal mycelium from each seed culture was then used to inoculate two stationary production cultures each of 200 ml of CDYE in a 1 L flask and incubated at 25 °C for 16 days. The resulting mat of fungal mycelium was then separated from the liquid culture medium by straining through nylon cloth (mesh size approximately 218 µm) and the combined fungal material from the two culture flasks macerated in 250 ml sterile water to produce the ‘mycelial mat’ fraction. Remaining small fragments of fungal material were then removed from the liquid medium by centrifugation at 600 g for 15 min. The mycelia mat and supernatant fractions were frozen at -20 °C until required for testing.

Batch-culture fermentation

Batch-culture fermenters that provided controlled conditions to simulate the rumen environment and microbial processes (Cone et al. 1996) were used, with some minor modifications. For the endophytic fungi, a manual batch-culture system was used. Screening of the saprophytic fungi was conducted after the endophytic fungi by which time a newly developed automated system was available for use (Muetzel et al. 2011).

Endophytic fungi

In vitro incubation

Four samples of 600 mg of the freeze-dried substrate were each weighed into 120 ml serum bottles and pre-warmed for 2 h before the incubation started. Artificial saliva (McDougall 1948) was prepared the day before incubation and heated to 39 °C in a water bath. One hour before incubation the buffer was flushed with CO₂ and 20 min before incubation started, 100 mg/L of sodium sulphide (Na₂S) was added to the buffer to completely reduce the solution. Rumen fluid was manually collected at 8:30 am from two rumen-cannulated cows removed from pasture at 8:00 am, and immediately placed in pre-warmed thermos flasks. The rumen fluid samples were then brought to the lab, filtered through one layer of cheese cloth and mixed at a ratio of 1:4 with reduced artificial saliva. The mixture was kept at 39 °C, stirred and continuously flushed with CO₂. For each rumen sample, 60 ml of the mixture was dispensed under a stream of CO₂ into two serum bottles containing the substrate and the bottles were closed with a butyl rubber stopper and placed in an incubator at 39 °C for 24 h. The unit of replication was the rumen fluid from a cow, and the two bottles for each cow were subsamples.

Gas analysis

After 2, 6, 12 and 24 h of incubation, the gas volume was determined by inserting a needle through the rubber stopper and releasing the gas into a calibrated syringe until pressure equilibrated, and a 1 ml aliquot was taken for determination of methane concentration. Two hundred microlitres of the gas sample was injected manually into a gas chromatograph (GC-2010, Shimadzu Corporation, Kyoto, Japan) fitted with a flame ionization detector (FID). Gases were separated in a HP-Plot Molsieve column (length 35 m, ID 0.53 mm, Agilent Technologies Inc, Santa Clara, CA, US) and methane was detected with an FID detector (250 °C). The GC was run under isocratic conditions at 85 °C with N₂ as a carrier gas.

Sampling

At the end of the incubation, a sample was collected from each bottle to determine the total concentrations of short chain fatty acids (SCFA) and the proportions of acetate, propionate and butyrate. An aliquot of 2 ml rumen fluid in duplicate was centrifuged at 21,000 g for 10 min at room temperature. Nine hundred microlitres of the supernatant was transferred into a fresh vial containing 100 µL of internal standard (19.87 mM ethyl-butyric acid, 20% [v/v] ortho-phosphoric acid). The samples were then frozen overnight and centrifuged (21,000 g, 10 min at room temperature) the next day. Eight hundred microlitres of the supernatant was transferred into a crimp cap vial and samples for SCFA were analysed in a gas chromatograph (GC; model HP 6890) equipped with a capillary column (ZB FFAP &HK-G009-22) (Attwood et al. 1998).

Saprophytic fungi

Samples of each fraction of saprophyte, either the macerated mycelia mat of the fungus or the mycelia-free incubation medium the fungus was grown in (supernatant), were added to 600 mg of oven-dried (60 ^oC for 24 h) pasture hay (see below for chemical composition) ground to pass through a 1 mm sieve, as the substrate for fermentation. This allowed independent testing for the effects of the fungus itself and the metabolites of the fungus that were released into the medium. A known inhibitor of methanogenesis, bromoethane sulphonate (BES), was added to a further subsample of the hay substrate at a final concentration of 30 µmoles/L and included in each incubation run as a positive control. The hay with neither BES nor saprophyte sample added served as a negative control.

Incubation, sampling and sample analysis was done as described above for grass-endophyte samples, but an automated analysis system was used to determine gas volume and composition. Each incubation bottle was attached to a pressure sensor (40PC015G1A, Honeywell International Inc, Morristown, New Jersey, US) and gas volume was determined from pressure. Pressure was recorded every minute for each sensor. A calibration curve for every sensor was used to convert pressure into gas volume. When the threshold pressure of 9 kPa (well below the pressure that affects chemical reactions and end-product formation; unpublished data) was reached, a solenoid valve was opened and the gas sample was injected into the GC. Gas analysis was done using the same conditions as described above. The analysis of gas composition was conducted on samples collected after 9 h, reflecting the approximate peak of fibre fermentation activity and at 24 h, the end point of the incubation, reflecting the cumulative production and composition of fermentation products.

For the saprophytic fungi, the negative control was included to assess the effect of the inoculums on overall fermentative capability of the microbial community in each incubation run. The positive control was included to assess the susceptibility of the methanogenic community to an inhibitor.

Other chemical analyses

The concentrations of crude protein (CP), neutral detergent fibre (NDF) and soluble sugars and starch (SSS), and the organic matter digestibility (OMD) of freeze-dried subsamples (dry matter 910±40 g/kg for ryegrass and 910±70 g/kg for tall fescue) of each of the ryegrass- and tall fescue-endophyte associations were determined using near infrared reflectance spectroscopy (feedTECH, AgResearch Grasslands, Palmerston North). A subsample of the hay used as substrate for saprophyte incubations was analysed by standard wet chemistry procedures to determine CP (104 g/kg DM) and NDF (613 g/kg DM). Dry matter of the hay after oven drying was 928 g/kg.

Subsamples of 50 mg of the freeze-dried herbage of each ryegrass- and tall fescue-endophyte association supplied for fermentation bioassay were also analysed for the concentrations of each of the main known alkaloids produced by the endophytic fungi, according to standard methods described for lolitrem B (Spiering et al. 2005), ergovaline and peramine (Spiering et al. 2002), lolines (Kennedy & Bush 1983; Yates et al. 1989) and epoxy-janthitrems (Tapper & Lane 2004). While the presence or absence of lolines and epoxy-janthitrems was not taken into account in the selection of candidates for testing, they were included in the analysis because the epoxy-janthitrems have shown tremorgenic effects similar to lolitrem B in animal studies, and lolines have shown feeding deterrence and toxicity in insect-based studies that is broader than that shown by peramine.

Sample ranking score

Where significant differences in methane production among fungi were detected, to identify candidates of interest for further examination, a ranking score for each candidate was calculated. For each sample, the ranking score was calculated from the sum of the changes in methane as a percentage of total gas production at 9 h and 24 h of fermentation relative to the negative control, plus the change in SCFA concentration relative to the negative control (each of the three components of the ranking score was calculated from the difference between the candidate and the negative control divided by the negative control). The negative control, by definition, has a rank score of zero. Suppression of methanogenesis gives a positive weighting to the score and suppression of SCFA gives a negative weighting; high positive scores indicate promising candidates. The score integrated the effects of a fungus on methane production with the effects on overall fermentation, and so separates out candidates where suppression of methane was related to a specific effect of inhibiting methanogenesis without suppressing fermentation (desirable and ranked high) from those that suppressed fermentation (undesirable and so ranked low).

Statistical analysis

The effects of candidate fungi on total gas, methane gas and SCFA production were compared by analysis of variance (ANOVA), separately for the grass-endophyte associations and the saprophytic fungi. A randomized incomplete block design was used with treatment and incubation run (block) as factors in the model. Each incubation run included a control and a subset of treatments. For grass-endophyte associations, the model included grass species as the main effect and endophyte association as the split-plot effect. For saprophytic fungi, the model included fungal fraction as the main effect and fungal strain as the split-plot effect. In both cases, interactions between the main plot effect and the split-plot effect were identified and ANOVAs were then conducted separately for each grass species and each saprophyte fraction. For the saprophytic fungi, the positive control was excluded from the ANOVAs because of the overriding effect on overall levels of significance. Within each ANOVA, where the treatment effect was significant, each candidate was compared against the negative control based on the least significant difference at P<0.05. For comparing among the ranking scores of the saprophytic fungi, the block effect was excluded from the ANOVA and each species and strain×fraction, including the positive control, was included as a treatment.

Results

Grass chemical composition

The ryegrass-endophyte associations contained 194±13 g/kg DM CP (mean of all associations tested±standard deviation), 105±17 g/kg DM SSS, 492±28 g/kg DM NDF and the tall fescue-endophyte associations contained 229±39 g/kg DM CP, 106±21 g/kg DM SSS and 494±60 g/kg DM NDF. The organic matter digestibility was 773±39 and 745±67 g/kg DM for ryegrass and tall fescue, respectively.

Endophyte alkaloid concentrations

The grass-endophyte associations showed the presence of alkaloids in the expected combinations (Table 2), relating to whether the strain was wild type containing three alkaloids (lolitrem, ergovaline and peramine in ryegrass, and ergovaline, peramine and lolines in tall fescue), a selected strain containing two alkaloids in differing concentrations (ergovaline and peramine in ryegrass, and peramine and lolines in tall fescue), AR37 containing only epoxy-janthitrems, or EL5 containing none of the major classes of alkaloids. The concentration of ergovaline in the tall fescue-wild type endophyte association (0.2 mg/kg DM), and peramine (4.1 mg/kg DM) and lolines (571 mg/kg DM) in all tall fescue associations were generally lower than expected. As expected, the endophyte-free ryegrass contained no alkaloids (not detectable); however, trace concentrations of ergovaline and peramine in the endophyte-free Jesup tall fescue probably indicate a low level of contamination in those plots from tall fescue containing the wild type strain.

Table 2  Concentrations of the major classes of alkaloids lolitrem B, ergovaline (Ev), peramine (Per), lolines and epoxy-janthitrems (Epox) in several strains of Neotyphodium lolii endophyte in perennial ryegrass (GA66) or N. coenophialum endophyte in tall fescue (cv Jesup).

		Alkaloid (mg/kg DM)
Species	Endophyte strain	Lolitrem	Ev	Per	Lolines	Epox
P. ryegrass	Wild type	1.7	0.4	22.7	–^1	–
	EL1	1.3	<0.1	–	–	–
	AR6	–	1.1	14.2	–	–
	EL2	–	1.1	9.1	–	–
	EL3	–	0.6	7.7	–	–
	EL4	–	0.2	9.7	–	–
	AR50	–	<0.1	24.3	–	–
	AR1	–	–	21.3	–	–
	AR37	–	–	–	–	21.7
	EL5	–	–	–	–	–
	Endophyte-free	–	–	–	–	–
Tall fescue	Wild type	–	0.2	3.9	1188	–
	AR525	–	–	5.0	687	–
	AR542	–	–	2.5	217	–
	AR584	–	–	4.0	344	–
	EL6	–	–	5.0	419	–
	Endophyte-free	–	0.1	0.4	–	–
^1 None detected

Fermentation products of grass-endophyte associations

Among the perennial ryegrass-endophyte associations there were no significant differences in total gas production (mean 182 ml/g DM) and the proportion of methane in the total gas production (mean 20.7% of total gas production) (Table 3). For the tall fescue-endophyte associations there were significant effects on total gas production (P=0.002) and, with the exception of AR 525, each association resulted in significantly higher total gas production compared with the endophyte-free tall fescue negative control. However, no tall fescue association reduced the proportion of methane in the total gas production (mean 21%) compared with the endophyte-free negative control (21.5%).

Table 3  Total gas production (GP), methane (CH₄) gas as a proportion of GP, the concentration of total short chain fatty acids (SCFA), and acetate, propionate and butyrate as proportions of SCFA after 24 h in vitro incubation of several strains of Neotyphodium lolii endophyte in perennial ryegrass (GA66) or N. coenophialum endophyte in tall fescue (cv Jesup).

		GP	CH4	SCFA	Acetate	Propionate	Butyrate
Grass	Endophyte strain	ml/g DM	% of GP	mM	%	%	%
P. ryegrass	Endophyte-free	184.1	20.4	41.0	62.5	22.8	10.8
	Wild type	186.2	20.2	41.6	63.2	24.0	9.4
	EL1	182.7	21.6	43.3	62.0	21.4	11.8
	AR6	177.2	20.6	35.9*	58.4	27.6*	11.5
	EL2	183.4	21.6	44.6	61.5	21.4	12.1
	EL3	191.5	20.4	43.0	62.4	24.9*	10.0
	EL4	181.4	20.3	38.4	61.1	24.2	10.8
	AR50	180.8	20.6	42.8	61.5	21.4	12.2
	AR1	175.7	20.6	36.9	63.3	23.5	8.6
	AR37	182.8	20.9	35.6*	63.3	24.3	9.2
	EL5	181.0	20.8	38.3	64.2	24.1	9.0
	Ryegrass mean	182.4	20.7	40.1	62.3	23.6	10.5
	SED	4.60	0.86	2.12	1.55	0.91	1.22
	P value	0.302	0.858	0.013	0.110	<0.001	0.479
Tall fescue	Endophyte-free	141.4	21.5	37.0	63.4	22.6	9.4
	Wild type	164.0*	19.6	48.3*	64.8	22.1	7.8
	AR525	145.8	20.9	37.8	64.3	23.5	7.8
	AR542	172.5*	21.8	46.1*	63.0	23.1	8.4
	AR584	167.9*	20.8	42.5*	62.9	23.2	8.5
	EL6	183.0*	20.4	51.5*	64.0	22.4	8.8
	Tall fescue mean	162.4	21.0	44.2	63.6	22.9	8.5
	SED	6.16	0.99	2.30	1.30	0.60	0.50
	P value	0.002	0.468	0.002	0.700	0.200	0.100
* Indicates grass-endophyte associations that are significantly different from the endophyte-free (negative) control based on Fisher's protected LSD0.05.

For the perennial ryegrass associations there were significant differences among endophyte strains in the total SCFA production (P=0.013) and the proportion of propionate (P<0.001; Table 3). Total SCFA was lower for AR6 (35.9 mM) and AR37 (35.6mM) compared with the negative control (41.0 mM), whereas the proportion of propionate was higher for AR6 (27.6%) and EL3 (24.9%) compared with the negative control (22.8%). The proportions of acetate and butyrate were similar among the strains. In contrast to the ryegrass-endophyte associations, for tall fescue there were no differences among endophyte-associations in either the total concentration of SCFA (mean 44.2 mM), or the proportions of individual SCFA (63.6%, 22.9% and 8.5% acetate, butyrate and propionate, respectively).

Saprophyte fermentation products

In general, the saprophytes had a more pronounced effect on the fermentation than did the grass-endophyte associations. For the fungal mat and supernatant fractions there were significant differences among the fungal species and strains for total gas production and methane as a proportion of gas production at 9 h and 24 h of fermentation (Table 4). For the fungal mat fraction, only P. roquefortii G2 reduced methane as a proportion of gas production compared with the negative control without also reducing total gas production, although for M. ruber 503.7 and 554.76 and G. candidum, which also reduced methane as a proportion of gas production, the suppression of total gas production was apparent only at 9 h of fermentation. For the supernatant fraction, the three strains of M. wolfii, P. roquefortii G2, M. ruber 503.7 and 554.76 and G. candidum each reduced methane as a proportion of gas production at 9 h and 24 h, while increasing or at least not reducing total gas production, compared with the negative control. For these strains the suppression of methane as a proportion of total gas was numerically similar to that by the positive control. Within the mat fraction, some strains decreased total gas production (e.g. M. wolfii 69/3891) or increased the proportion of methane (e.g. P. roquefortii I2 and A2, M. ruber15220) compared with the negative control and both of these effects are undesirable.

Table 4  Total gas production (GP) and methane (CH₄) gas as a proportion of GP after 9 h and 24 h in vitro incubation, and the concentration of total short chain fatty acids (SCFA) and acetate, propionate and butyrate as proportions of total SCFA after 24 h in vitro incubation of negative and positive controls and fungal mat and supernatant fractions of several species and strains of saprophytic fungi.

			Total gas		Methane gas
			ml/g DM		% of GP		SCFA
Fungus			GP	GP	CH4	CH4	Total	Acetate	Propionate	Butyrate
Fraction	Species	Strain	9 h	24 h	9 h	24 h	mM	%	%	%
	Negative control		82.8	153	15.0	17.0	39.7	65.8	22.9	8.3
	Positive control		69.7	133	7.5	12.2	38.8	62.2	25.0	9.9
Fungal mat	Mortierella wolfii	69/3891	75.6*	141.4*	14.6	15.9*	36.4	61.0	27.6*	7.8
	Mortierella wolfii	74/2649	102.7*	180.6*	14.6	16.6	41.6	68.6	20.7*	7.8
	Mortierella wolfii	68/707w2	97.7*	172.5*	14.8	16.9	41.3	68.7	20.6*	7.8
	Penicillium roquefortii	G2	77.0	149.7	13.0*	13.9*	33.7*	58.6	29.4*	8.4
	Penicillium roquefortii	I2	88.7	165.0*	17.3*	19.1*	40.1	66.7	20.6*	9.1
	Penicillium roquefortii	A2	98.5*	174.1*	17.9*	19.2*	41.1	67.3	20.0*	9.0
	Monascus ruber	15220	106.6*	184.0*	17.2*	18.7*	44.3*	66.5	20.4*	9.3
	Monascus ruber	503.7	75.5*	157.9	11.8*	14.0*	34.9*	64.3	23.8	9.0
	Geotrichum candidum	240.62	73.9*	156.9	12.2*	14.3*	33.1*	64.3	23.8	8.8
	Monascus ruber	554.76	72.5*	151.6	11.8*	13.9*	32.4*	63.6	24.5*	9.0
	Fungal mat mean		86.9	163.4	14.5	16.3	38.6	65.2	23.0	8.7
		SED	2.68	5.14	0.55	0.40	1.77	0.93	0.65	0.30
		P value	<0.001	<0.001	<0.001	<0.001	0.032	0.33	0.042	0.153
Supernatant	Mortierella wolfii	69/3891	82.8	155.3	10.0*	11.6*	38.5	61.0	28.5	7.3
	Mortierella wolfii	74/2649	109.8*	185.7*	10.2*	12.2*	46.6*	58.8	30.3	7.9
	Mortierella wolfii	68/707w2	105.3*	185.8*	8.8*	10.6*	44.5*	57.4	30.6	7.7
	Penicillium roquefortii	G2	75.9	142.7	11.1*	13.4*	32.4*	56.4	31.5	8.5
	Penicillium roquefortii	I2	95.2*	172.6	15.6	17.8	43.1*	67.9	20.0	8.6
	Penicillium roquefortii	A2	100.1*	176.0*	15.4	17.9	44.0*	67.5	20.2	8.9
	Monascus ruber	15220	80.6	153.3	14.9	17.5	39.9	68.4	20.1	8.5
	Monascus ruber	503.7	109.7*	197.0*	9.1*	12.4*	45.3*	61.1	26.3	9.9
	Geotrichum candidum	240.62	110.7*	206.0*	8.8*	12.3*	45.1*	61.7	25.9	9.8
	Monascus ruber	554.76	106.7*	194.5*	8.4*	11.9*	45.2*	60.7	26.9	9.8
	Supernatant mean		97.6	176.9	11.2	13.8	42.3	62.7	25.6	8.6
		SED	4.05	9.60	0.86	0.73	1.91	2.23	1.61	0.54
		P value	<0.001	<0.001	<0.001	<0.001	<0.001	0.066	0.062	0.078
* Indicates strains that are significantly different from the negative control based on Fisher's protected LSD0.05.

For the saprophytic fungi, where significant differences were detected in total gas production and methane as a proportion of total gas, the ranking score indicated that the three highest ranked candidates were the supernatant fractions from three strains of M. wolfii (Table 5), although each of the four species were represented among the strains that had ranking scores significantly higher than the negative control.

Table 5  The ranking score of fungal mat and supernatant fractions of several species and strains of saprophyte fungi and the positive control, calculated from the effects on production of methane and short chain fatty acids during in vitro fermentation relative to the negative control (the method of calculation gives the negative control a score of zero).

Fungus	Strain	Fraction	Ranking
Mortierella wolfii	68/707w2	Supernatant	1.08*
Mortierella wolfii	74/2649	Supernatant	0.97*
Mortierella wolfii	69/3891	Supernatant	0.79*
Positive control	–	–	0.77*
Monascus ruber	554.76	Supernatant	0.69*
Geotrichum candidum	240.62	Supernatant	0.63*
Monascus ruber	503.7	Supernatant	0.60*
Penicillium roquefortii	G2	Supernatant	0.46*
Penicillium roquefortii	G2	Fungal mat	0.35*
Mortierella wolfii	69/3891	Fungal mat	0.22*
Penicillium roquefortii	A2	Supernatant	0.20*
Penicillium roquefortii	I2	Supernatant	0.17*
Monascus ruber	15220	Supernatant	0.15
Monascus ruber	15220	Fungal mat	0.07
Negative control	–	–	0.00
Mortierella wolfii	74/2649	Fungal mat	−0.03
Monascus ruber	503.7	Fungal mat	−0.03
Penicillium roquefortii	I2	Fungal mat	−0.06
Mortierella wolfii	68/707w2	Fungal mat	−0.07
Penicillium roquefortii	A2	Fungal mat	−0.08
Monascus ruber	554.76	Fungal mat	−0.10
Geotrichum candidum	240.62	Fungal mat	−0.14
	SED		0.09
	P value		<0.001
* Indicates strains that are significantly different from the negative control based on Fisher's protected LSD0.05.

Discussion

The candidate grass-endophyte associations provided the structured set of presence and absence of several alkaloids required to test whether any possessed antimethanogenic properties. Based on the hypothesis being tested, it was expected that methane as a percentage of total gas production might be suppressed by the presence of various endophyte strains in ryegrass and in tall fescue, relative to the endophyte-free grasses. Furthermore, if suppression was observed, it was expected that comparisons among the five ryegrass associations containing each alkaloid alone (AR1/AR50, EL1 and AR37 for peramine, lolitrem and epoxy-janthitrems, respectively, and EL2 normally containing primarily ergovaline, but some peramine as well) would identify the bioactive alkaloid. There was no suppression observed for any of the endophytes, suggesting that these alkaloids are not antimethanogenic. The samples were collected in March, the time of year when the concentrations of alkaloids would normally be at their seasonal peak (Fletcher 1999; Watson et al. 1999) and were cut as close to the ground as possible so as to include grass tiller bases where the concentrations of lolitrems and ergovaline tends to be highest (Keogh et al. 1996; Lane et al. 1997). The concentrations in ryegrass were within the expected range, whereas for tall fescue the concentrations of the main alkaloids ergovaline and lolines were considered low (unpublished data). This was possibly a consequence of the dry summer conditions at that site prior to sampling. The characteristics of the grass-endophyte associations, and particularly the concentrations of secondary metabolites produced by the fungus, are influenced by both the host grass and the endophyte strain. Using a single line of ryegrass or tall fescue has the advantage of allowing a more systematic evaluation of variation among endophyte strains. However, a potential disadvantage is that this may restrict the extent of variation in the concentrations of alkaloids produced because the expression of strain-specific attributes can be influenced by the grass cultivar in the association. A more rigorous test of the role of alkaloids for suppressing methane production may be to separate grass tillers into leaf lamina, stem and pseudostem, and conduct the assay on the basal herbage containing the higher concentrations of lolitrem B and ergovaline, or on leaves where the concentration of peramine tends to be highest. There may be other explanations for the nil result. For example, there may be other possibly unknown compounds that interfere with an antimethanogenic effect or its detection.

The one previous study that showed suppression of methane production by an endophytic fungus was based on Jesup tall fescue containing the selected strain AR542, but no effect was shown for the same cultivar containing wild type endophyte (Vibart et al. 2007). If the hypothesis that alkaloids are bioactive against methane production is correct, it would be expected that the wild type, containing ergovaline and higher concentrations of lolines than AR542, would show at least the same effect. This also tends to suggest that endophyte alkaloids are not involved. It is hard to explain how a strain selected to eliminate the production of ergovaline would create an effect not seen with the wild type strain that does produce ergovaline, unless there is some other unknown difference between strains. However, further tests using herbage with higher concentrations, or perhaps even extracted alkaloids where a dose-response relationship could be tested, would be required before ruling out any prospects with endophytes.

Compared with the endophytic fungi, the saprophytic fungi showed much greater effects on reducing methane production. Some of the samples reduced methane as effectively as the positive control which was treated with 30 µM of BES. Generally the effects were much more pronounced in the culture supernatant than in the fungal mats, as indicated by the consistently higher ranking score for the supernatant. This suggests that the active component(s) is a soluble fungal product. The ranking procedure indicated that M. wolfii is the most promising species, since the supernatant fraction of each of the three strains of this species were grouped at the top of the ranking list. This appears to be a specific inhibition, in that methane was reduced as a proportion of total gas production, rather than as a consequence of overall suppression of fermentation.

For the saprophytic fungi chosen for this study, there were no chemical analyses undertaken comparable with those conducted on the grass-endophyte associations. The profile of secondary metabolites produced by some of the candidate saprophytic fungi is well known because of their production of mycotoxins or use in production of fermented food or pharmaceuticals (see Table 1). The anecdotal information suggesting strong links between ingestion of mouldy silage and suppression of methane production by ruminants gave no indication of the specific spoilage fungus involved, let alone any specific secondary bioactive metabolite. The suppression of methane production which followed the addition of crude extracts from the mycelium and culture filtrate of two of the four candidate species in the batch test is encouraging. To obtain a positive result in a bioassay of unknown material will always depend on a bioactive constituent being present in the extract at a concentration that will elicit a measurable effect, and without any other constituents of the sample masking the effect. Activity can be greatly affected by fungal strain, growth medium, culture conditions and whether special methods are required to preserve activity of labile components after harvest. Once any degree of activity has been detected, the processes of bioassay-guided fractionation and purification are required to establish the range of bioactive compounds and to establish their chemical identity and essential properties (e.g. specificity, mode of activity). The importance of persevering with bioassay-based fractionation and purification before eliminating potential candidates is highlighted by results with Monascus ruber. This red mould fungus is known to produce monicolin K (lovastatin; Demain & Sanchez 2009) which is reported to suppress growth of methanogenic Archaea but not other rumen bacteria (Wolin & Miller 2006; refer to comment in Table 1), yet extracts of the three strains did not suppress methane production. At this stage it is unclear whether the strains used will produce lovastatin at a detectable concentration, or at all, either under the conditions we used or under one of many other possible combinations of growth medium, incubation conditions or extraction method.

Further investigation is required to show conclusively that fungi can suppress methanogenesis. The range of endophytic and saprophytic fungi used was designed to reflect the diversity between and within species but was extremely narrow (given the constraints on the number that could be screened) and confined to those where there was some prior indication of a possible link with suppression of enteric methane production during the digestion of fresh or ensiled pasture. It is possible there may be species or strains of fungi with antimethanogenic properties, outside of those screened in this study. Furthermore, there is year-to-year and seasonal variation in the concentrations of endophyte alkaloids, so results must be interpreted in the context of the particular concentrations of alkaloids and other secondary metabolites in the year in which these samples were collected. Similarly, as discussed above, for the saprophytic fungi, the culture conditions can influence the fungal growth and the production of secondary metabolites and this could influence the detection of antimethanogenic properties of particular fungi. The possibility of interacting effects between endophytic and saprophytic fungi cannot be ruled out and may have been a factor in the earlier observations by Vibart et al. (2007) and Clark & Krause (2007).

Conclusions

For the species and strains of endophytic fungi tested, none demonstrated potential to significantly reduce methane production in batch-culture assay. On this basis it is unlikely that any suppression would be demonstrated under the more challenging conditions of continuous-culture fermentation and animal-based testing, where even promising short-term suppression may be lost as the rumen microbial population adjusts. However, one species of saprophytic fungus, M. wolfii suppressed methane production without suppressing the beneficial production of short chain fatty acids during fermentation. Simple separation of the mycelia mat from the liquid supernatant and assaying these fractions independently suggests that the bioactive constituent may be a metabolite of the fungus. This should be investigated further in continuous-culture fermentation assays and then in animals to verify that the effect can be sustained, and if so, then undertaking steps to fractionate and identify the bioactive constituent and tests to determine if the effect can be enhanced in a dose-dependent manner by, for example, concentrating the bioactive compound.

Acknowledgements

This project was funded from the Ministry for Primary Industries, Sustainable Land Management and Climate Change fund (C10X0829). David Hume and Marty Faville, AgResearch Grasslands, assisted with the selection of candidate grass-endophyte associations. Samples of ryegrass-endophyte associations were supplied by Phil Rolston, AgResearch Lincoln and samples of tall fescue-endophyte associations were supplied by Bruce Cooper, AgResearch Kerikeri. Lex Foote prepared the grass-endophyte associations for chemical analysis and screening. The advice of Margaret di Menna, AgResearch Ruakura on the selection of saprophytic fungi for screening and supplying the starter cultures of some of the species and strains is gratefully acknowledged.