A Simplified Alkane Analysis Method for the Determination of Herbage Intake by Dairy Cows

Abstract

An accurate dry matter intake (DMI) measurement method is of great importance for determining the performance of dairy cows. While n-alkane technique is commonly used to estimate the DMI of ruminants, the protocols used to measure alkanes are complex, time consuming and low throughput. Herein we present a simplified and miniaturised protocol that combines a smaller sample matrix (60 mg of faeces or herbage), a shorter (3 h) saponification step and a direct purification of the crude extract by filtration through silica powder or mixing with silica powder prior to GC-FID analysis. In addition to a much higher throughput compared with the existing methods, by skipping the commercial SPE columns and reducing the organic solvent volume, our miniaturised protocol should significantly reduce the sample processing cost and is thus suitable for alkane determination for large-scale experiments.

1. Introduction

Measuring the dry matter intake (DMI) of grazing animals is a challenging task, but one that is crucial for evaluating the feed conversion efficiency of individual animals. Although various methods have been reported for estimating the DMI of different ruminants [1], the n-alkane technique is the most accurate for individual animals [2,3]. Due to the widespread presence of epicuticular wax on the adaxial surface of grass species and the quasi-indigestibility of long-chain alkanes present in the wax, the possibility of estimating herbage intake in ruminants using dosed C32 and natural C33 n-alkanes was demonstrated over three decades ago [4].

Although n-alkanes can be easily analysed by gas chromatography (GC), sample preparation prior to GC analysis is a lengthy and costly process [5]. The earliest protocol involved petroleum extraction (in a Soxhlet), followed by over-night saponification, and then sample clean-up with silica columns [4]. To reduce the workload involved in sample preparation, various modifications of the initial protocols were proposed, such as direct saponification of samples and shortened (from 16 to 3–4.5 h) incubation time [6,7], direct purification of the crude extract by silica columns (without prior solvent evaporation) [8] and reducing the sample processing scale from 1–2 g to 0.1–0.2 g [9]. While direct saponification has been widely adopted since, other simplification approaches have not found widespread use. Until now, there has been no consensus on a simple method for the alkane analysis of feed and faecal samples.

Numerous sample preparation protocols have been reported in the literature in the past two decades [10,11,12,13,14,15,16]; nearly all of them were derived from Mayes et al. [4] and Vulich et al. [8]. The common steps of these methods include: (1) mixing 0.5 (faeces) or 1.5 g (herbage) sample and C34 (internal standard, IS) with 7–14 mL 1 M ethanolic potassium hydroxide (KOH) reagent; (2) 3–16 h saponification at 90 °C; (3) extraction of released wax components twice by hexane (using up to a total 28 mL hexane per sample); (4) evaporation of the solvent and purification of n-alkanes by silica solid-phase extraction (SPE) columns before GC analysis. The entire procedure, referred to as the reference method in this report, is less complex compared to the initial protocol [4], but is still labour intensive, low throughput and requires a large amount of hazardous solvent (hexane).

The purpose of this study was to scale down the alkane analysis procedure and simplify each step of the protocol to increase the overall sample processing throughput. In addition, the feasibility of using a bulk silica powder to replace the commercial SPE columns for n-alkane purification was explored with the aim to reduce the sample processing cost.

2. Materials and Methods

2.1. Herbage and Faecal Samples

Herbage samples (mainly perennial ryegrass) and faecal samples of dairy cows following ingestion of C32 dosed herbage were collected from the research farm of the Department of Jobs, Precincts and Regions, Ellinbank Centre in Victoria, Australia. The experiment was conducted in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. Herbage samples were freeze-dried for a minimum of 72 h and faecal samples were oven-dried for 72 h at 40 °C. All samples were ground to a fine powder (<1 mm) before analysis.

2.2. Chemicals

Standards of n-alkanes (a mix of 40 alkanes from C8 to C40) and C34 (IS), ethanol and potassium hydroxide were purchased from Sigma-Aldrich. Chromatographic silica media (40–63 µm) was purchased from Grace, n-hexane from Ajax Finechem, and SPE columns (Strata SI-1, 55 µm, 500 mg/3 mL tubes and 1 g/6 mL tubes) from Phenomenex.

2.3. Sample Preparation for N-Alkane Analysis

2.3.1. Reference Method

Faeces (0.5 g) or herbage (1.5 g) samples were saponified in thick-walled screw-topped Pyrex tubes (25 mm × 150 mm) after adding 0.25 mg C34 (IS), and 7 mL (for faeces) or 14 mL (for herbage) of 1 M ethanolic KOH for 3 h at 90° C (in a water bath). Alkanes liberated were extracted twice with 7 mL or 14 mL of hexane and the combined organic phase (14 mL or 28 mL) was evaporated to dryness under a stream of N₂. The crude extract was then redissolved in 3.5 mL hexane and applied onto a pre-conditioned silica SPE column (1 g/6 mL). The n-alkanes were then eluted with 8 mL hexane, and the eluent was evaporated to dryness again and reconstituted in 0.3 mL hexane before GC-flame ionization detector (GC-FID) analysis in splitting mode.

2.3.2. Miniaturised Method

Our miniaturised n-alkane analysis protocol was adapted from that reported by Dove and Mayes (2006) [9], in which the saponification of samples was performed with 100 mg faeces and 200 mg plant material in 1.5 mL and 2 mL reagent (1 M ethanolic KOH), respectively, for 16 h at 90 °C. Our preliminary tests found 60 mg of herbage and faeces (in 2 mL reagent) to be optimal for both alkane yield and signal-to-noise ratio in n-alkane quantification by GC-mass spectrometry (MS) and GC-FID. In order to determine the optimal saponification time under this scale, faeces (60 mg) and herbage (60 mg) samples were incubated in thick-walled screw-topped 5-mL Wheaton vials after adding 10 µg C34 (IS) and 2 mL 1 M ethanolic KOH for 1, 2, 3 and 16 h at 90 °C (heating block). Alkanes liberated were extracted twice with 2 mL hexane and the combined organic phase (4 mL) was evaporated to dryness under a stream of N₂. The crude extract was then redissolved in 0.4 mL hexane and applied onto a pre-conditioned silica SPE column (500 mg/3 mL). The alkanes were eluted with 2 mL hexane, and the eluent was evaporated to dryness again and reconstituted in 0.9 mL hexane before GC-FID analysis in non-splitting mode.

2.3.3. Purification of Crude Extract by Silica Powder

Two alternative protocols for n-alkane purification were tested for the miniaturised method, both using chromatographic grade silica powder to replace the commercial silica SPE columns. Protocol 1: the crude extract (4 mL) was filtered directly through a 5 mL disposable syringe tube (TERUMO Corporation) pre-filled with 7 mm-thick silica powder (around 450 mg); the filtrate (about 3.5 mL) was dried and redissolved in 0.7 mL hexane before GC analysis in splitless mode. Protocol 2: the crude extract (4 mL) was thoroughly mixed (by vortex for 2 min) with 1 g silica powder in a 15 mL centrifuge tube; after centrifugation (5000× g for 10 min), the supernatant (about 2 mL) was collected, evaporated to dryness, and redissolved in 0.4 mL hexane before GC analysis in splitless mode.

2.4. GC-FID Analysis

The n-alkanes of different carbon chain length were separated on a capillary HP-1MS column (30 m, 0.25 mm ID and 0.5 µm film, Agilent Technologies) with a constant flow of 2.4 mL/min helium as the carrier gas and the following temperature program: initial temperature of 170 °C and held for 4 min, increased by 30 °C/min to 215 °C and held for 1 min, further increased by 6 °C/min to 325 °C and held for 11 min. The injection inlet and the detector temperature were 280 °C and 300 °C, respectively. The air, H₂ and makeup gas (N₂) flow rates were 400, 30 and 25 mL/min, respectively. The injection volume was 1 µL.

2.5. Statistical Analysis

All experiments were performed with 2–4 technical replicates. The results were subjected to ANOVA or Student’s t-test (Excel, Microsoft 365) depending on the number of treatments. All experiments were repeated at least twice to confirm the findings and only one set of results are presented.

3. Results and Discussion

3.1. Saponification Time Optimization

Using the lower sample loading (30 mg per mL for both faeces and herbage), we found that a lengthy over-night (or 16 h) saponification was not necessary, because no significant difference was observed in the yield of any of the n-alkanes (from C27 to C33) between 16 h and 1–3 h for both faeces and plant samples (Figure 1A,B). As the 3 h incubation gave a slightly higher overall yield of the major n-alkanes and the least variation between technical replicates, a standard 3 h saponification regime with a fixed sample to reagent ratio of 30 mg per mL was adopted for our miniaturised n-alkane analysis protocol.

3.2. Reproducibility of the Miniaturised Method

The reproducibility of our miniaturised method was assessed by comparing the results of two independent operators (named staff A and staff B) following the same protocol using the same faeces and herbage samples. For both operators the error bars (SD) of four technical replicates were rather small for all the major n-alkanes surveyed and for both sample types. The relative standard deviation (RSD) is generally used to evaluate the reproducibility of a method; RSD < 5% is considered to be satisfactory and RSD < 3% to be excellent, for biological samples. In this study, the average RSD was 2.7% and 2.4% for faeces and 1.5% and 1.9% for herbage (Figure 2A,B). In addition, there was no significant difference between the mean values (across all the n-alkane species and for both sample types) obtained by the two operators.

While the measurement precision or reproducibility is the major concern for any miniaturised method due to a small amount of sample used, these results prove that the current miniaturised method generates reproducible results and thus is suitable for n-alkane quantification in well ground and homogenised faeces and herbage samples.

3.3. Comparison between the Miniaturised Protocol and the Reference Method

Having demonstrated the precision of our miniaturised method, we have further assessed its measurement accuracy by comparing the n-alkane content results with those generated by the well-established and widely used reference method. The reference method for sample saponification is of a much greater scale and the sample/reagent ratio is around 71 mg and 107 mg per mL for faeces and herbage, respectively.

Both methods show a satisfactory reproducibility. Overall, the miniaturised method and the reference method produced very similar results with both the herbage and the faecal sample (Figure 3A,B); a slight difference observed in some cases was not statistically significant. This indicates that the miniaturised method is reliable for alkane analysis. The significantly reduced sample mass when using the ministurised protocol led to a substantial reduction in solvent consumption (a total of 10 mL vs. 32–46 mL hexane per sample for the reference method), a significant saving in SPE column cost (about 35%), and most importantly, a 3–4-fold reduction in the sample processing time (i.e., a 3–4-fold increase in sample throughput).

3.4. Alternative Methods for Sample Clean-Up

For n-alkane analysis by GC-FID, purification of the crude extract through a SPE column is an indispensable step, because plant epicuticular wax contains a range of interfering compounds that are also released upon saponification. Unfortunately, sample clean-up using SPE columns is a labour-intensive and costly step, which cannot be circumvented by using GC-MS, as no specific ions can be found for the targeted n-alkane species (results not shown). Although the SPE eluting process can be shortened by applying a positive or negative pressure, the remaining steps that include evaporation of the solvent, reconstitution of the sample and conditioning of the SPE column (prior to sample loading) are still time consuming. To simplify this bottleneck step, two alternative sample purification methods were tested in this work, both involving the use of silica powder to replace the commercial SPE columns.

Figure 4 shows that direct filtration of the entire crude extract (i.e., without evaporation of the solvent and reconstitution) through a layer of silica powder in a syringe tube produced similar results in terms of n-alkane content for both faeces and herbage samples, as compared to the standard SPE column clean-up. Moreover, comparable results were obtained when the crude extract was purified by directly mixing with silica powder in a centrifugation tube (followed by centrifugation) (Figure 4). For all the n-alkanes surveyed in both sample types, no significant difference was observed across the three purification protocols.

It is worth mentioning that in both cases, the purified eluent recorded a substantial reduction in volume (from 4 mL to around 3.5 mL for the filtration method and from 4 mL to around 2 mL for the direct mixing method) due to adsorption and gelatinization, but this did not appear to cause any differential loss of targeted n-alkanes as compared to the IS, nor any distortion of the results. In addition, both methods generated reproducible results, as judged by the low RSD across all the alkanes measurements (averaging 0.8 and 2.5% for faecal and 1.1 and 2.0% for herbage samples). Using these two alternative methods for sample clean-up is expected to reduce the cost related to SPE columns substantially, and also increase significantly the sample processing throughput.

The GC-FID profiles of long-chain alkanes of the faeces and herbage samples before and after purification by the three methods are shown in Figure 5. While numerous interfering peaks which were not fully separated from the targeted alkane peaks were observed especially in the crude extract of the faecal sample, purification by filtration through silica powder or by directly mixing with silica powder generated a much cleaner alkane profiles for both sample types, very similar to those achieved by the commercial SPE column (Figure 5A,B).

Clearly, the filtration method requires an extra 5 mL syringe tube, but uses a smaller amount of silica powder (<500 mg). By comparison, the procedure of directly mixing the crude extract with silica powder is very simple to perform, but employing a sufficient amount of silica (no less than 1 g) and thorough mixing by vortex (for 2 min) is essential to achieve a satisfactory outcome.

When these alternative sample cleaning methods are combined with the miniaturized saponification and liquid-liquid extraction steps, we estimate the sample processing throughput can be further increased by 25%.

4. Conclusions

We have optimised a miniaturised alkane measurement method for herbage and faeces samples with the aim of determining the DMI of individual dairy cows. Compared to the current n-alkane analysis protocols, our method features a reduced sample to reagent ratio (30 mg per mL), a shorter saponification step (3 h) and two simpler sample purification procedures based on silica powder. With reduced consumable and labour cost, and increased sample throughput, our method is suitable for the determination of n-alkane based DMI of dairy cows in large-scale experiments.