MEDI: Macronutrient Extraction and Determination from invertebrates, a rapid, cheap and streamlined protocol

Abstract Macronutrients, comprising carbohydrates, proteins and lipids, underpin many ecological processes, but their quantification in ecological studies is often inaccurate and laborious, requiring large investments of time and bulk samples, which make individual‐level studies impossible. This study presents Macronutrient Extraction and Determination from Invertebrates (MEDI), a protocol for the direct, rapid and relatively low‐cost determination of macronutrient content from single small macroinvertebrates. Macronutrients were extracted by a sequential process of soaking in 1:12 chloroform:methanol solution to remove lipid and then solubilising tissue in 0.1 M NaOH. Proteins, carbohydrates and lipids were determined by colorimetric assays from the same individual specimens. The limits of detection of MEDI with the equipment and conditions used were 0.067, 0.065 and 0.006 mg/ml for proteins, carbohydrates and lipids respectively. Adjusting the volume of reagents used for extraction and determination can broaden the range of concentrations that can be detected. MEDI successfully identified taxonomic differences in macronutrient content between five insect species. Macronutrient Extraction and Determination from Invertebrates can directly and rapidly determine macronutrient content in tiny (dry mass ~3 mg) and much larger individual invertebrates. Using MEDI, the total macronutrient content of over 50 macroinvertebrates can be determined within around 3 days of collection at a cost of ~$1.35 per sample.


| INTRODUC TI ON
Despite the relevance of macronutrients, comprising proteins, carbohydrates and lipids, to a broad range of applications, few ecological studies quantify them. Many studies concerned with the macronutrient content of invertebrates use analogues, such as nitrogen as a surrogate for protein (e.g. crude protein = N × 6.25; Jones, 1931) or in lieu of protein (Bryer et al., 2015;Finke, 2005;Pekár & Mayntz, 2014). This allows broad-scale studies of ecological stoichiometry in trophic networks, focusing on the ratios of analogous elements such as carbon, nitrogen and phosphorous (Anderson & Hessen, 2005;Frost et al., 2005;Raubenheimer et al., 2009). While broadly useful, these analogues can produce inaccurate results since, for example, nitrogen is present in many non-protein constituents of invertebrates, including exoskeleton (Janssen et al., 2017;Jones, 1931;Raubenheimer et al., 2009). Correction factors may circumvent these issues, but one correction factor is unlikely to work on all species given the diversity of invertebrates (Janssen et al., 2017). Additionally, some analyses of macronutrient content use gravimetric methods (e.g. Pekár & Mayntz, 2014), which require either bulk samples (~1 kg insect material for Finke, 2013) or very fine, often expensive, scales for the determination of macronutrient mass, long waiting times, and often still rely on analogues. Bulk samples are laborious to collect and process, impeding multi-taxon or individual-level analyses (Bryer et al., 2015).
Methods have previously been developed for determining the macronutrient content of small single macroinvertebrate samples (e.g. Lu et al., 2008), but these are standalone protocols each tailored to only one macronutrient, tripling the collection effort necessary to determine the content of each macronutrient from a population and making individual-level studies impractical. By implementing a uniform extraction method and streamlining a protocol to determine all three macronutrient contents from a single specimen, information output would increase while reducing sampling effort. No protocol has yet been published which uses direct measures of all three macronutrients taken in parallel from single small invertebrate specimens.
Standardised adoption of such a protocol would also ultimately benefit future meta-analyses. For individual-level determination of macronutrient content, or studies involving particularly small or scarce invertebrates, there is a need for a standardised approach to directly determine macronutrient content in parallel from single macroinvertebrate specimens to better understand ecological nutrient dynamics.
Our protocol determines the content of all three macronutrients from the same individual specimen. Presented herein is Macronutrient Extraction and Determination from Invertebrates (MEDI), a streamlined, rapid, cheap and simple protocol for the extraction and determination of macronutrient content that can be applied at the scale of individual invertebrates (≥~3 mg dry mass).
Using MEDI, the total carbohydrate, lipid and protein content of over 50 macroinvertebrates can be determined within around 3 days of collection at a cost of ~$1.35 per sample using standard laboratory equipment. This protocol will enhance the study of macronutrient content in invertebrates and other small samples in contexts including trophic interactions, parasitology and development.

Pierce BCA Protein Assay reagents and Pierce Modified Lowry Protein
Assay reagents were obtained from Thermo Fisher Scientific. Ribbed, skirted 1.5 ml screwcap microtubes and caps were obtained from STARLAB. Sulphuric acid (95%) and phosphoric acid (85%) were obtained from Fisher Scientific.

| Macronutrient extraction
Macronutrient extraction is a two-step process that first involves extracting lipid and then solubilising the remaining tissue for carbohydrate and protein analysis (Figures 1 and 2). Details of the methods will vary depending on the size of arthropod used. There are many important considerations when analysing the macronutrient content of arthropods ( Samples were first weighed and lipids were extracted by soaking whole arthropods in 1 ml of 1:12 chloroform:methanol for 24 hr (smaller specimens such as those <0.5 mg dry mass could be soaked in 0.5 ml for increased detectability, and larger specimens in larger volumes ~5× their body volume to ensure full submersion and to prevent saturation of the solvent). Half of the added volume of supernatant was then pipetted into a fresh tube for later lipid determination, the rest of the supernatant discarded, and any residue allowed to evaporate. This procedure for soaking arthropods was repeated for another 24 hr, but discarding all supernatant, to ensure any residual lipids were removed from the sample prior to protein and carbohydrate extraction. The change in dry mass of a sample before and after soaking in the solvent can also be used as an estimate of the lipid content of samples where practicable (i.e. gravimetric assay).
Following the lipid assay, the soft tissue of samples was digested to facilitate quantification of protein and carbohydrates. This procedure only measures the macronutrient content of the soft tissue of arthropods and not any protein that may be bound in the chitinous matrix of the exoskeleton during sclerotisation. Whole arthropods from 1 to 10 mg lean mass (i.e. mass after lipid extraction) were weighed, added to a microcentrifuge tube along with a stainless-steel bead (~3-7 mm diameter) and lysed at room temperature using a TissueLyser II (Qiagen) for 8 min at 30 Hz in 2-min increments. Larger samples were ground (e.g. bead beating or mortar and pestle) and an approximately 5 mg subsample was weighed into a clean tube. To each tube was added 1 ml of 0.1 M NaOH (or 0.5 ml for smaller specimens, e.g. <1 mg). Tubes were placed in a thermo-shaker at 80°C and 250 RPM for 30 min, then removed and left at room temperature overnight (~16 hr). Samples were centrifuged for 10 min at 13,000 rpm and 600 µl of supernatant pipetted into a separate tube for protein and carbohydrate determination. Supernatant was diluted prior to assaying such that the concentration of lean tissue (approximately 25%-75% protein for arthropods) was approximately 1-2 mg/ml to allow protein values to fall within the range of the protein assay kit (most commercial protein assay kits can measure 0.025-2 mg/ml protein). Dilution of supernatant or change in volume of NaOH used, along with the mass of sample used, must be accounted for in subsequent calculations of protein content.

| Exoskeletal mass determination
The exoskeleton content of samples can also be measured, which may be of interest in morphological studies or those concerned with the nutritional quality of arthropods for consumers ( Figure 2). A separate sample was used for this measurement in this study since lysis of tissues was carried out during macronutrient extraction to facilitate rapid dissolution of all soft tissues. Preliminary work suggests that exoskeleton measurements of lysed tissue result in lower values than measurements on intact arthropod bodies (S.M. Wilder, unpubl.). To maintain intact exoskeletons, the exoskeletal measurements instead included a second round of NaOH treatment and longer heated incubations.

F I G U R E 2
Protocol for the extraction of macronutrients and measurement of exoskeletal mass from invertebrate bodies. Only the first two rows are carried out for macronutrient extraction. For exoskeletal measurement, separate samples were used in this study. It is advised to lyse the specimen for protein extraction because this allows efficient extraction of protein in one wash of NaOH. But it is not advised to lyse the specimen for exoskeletal measurement because preliminary work suggests that exoskeleton measurements on lysed samples are significantly lower than measurements on intact samples (S.M. Wilder, unpubl.); it is also advised to heat the specimen for a longer period (i.e. 2 hr) and to repeat the central steps (addition of NaOH, heating, incubation and discarding of supernatant) to ensure removal of all soft tissues. Figure created using Biorender.com Exoskeletal measurement could theoretically be carried out on the same specimens used for macronutrient determination, but appropriate care must be taken to ensure that the soft tissue is appropriately dissolved; separate specimens should thus be used where possible.
First, lipid should be completely extracted from the sample as described above. Then, the exoskeleton of the sample should be lightly cracked and 0.1 M NaOH (a volume approximately 5-10 times that of the sample) should be added to a vial with the sample. Samples should be heated for 2 hr at 80°C and then allowed to soak overnight after which the NaOH should be removed and discarded. Centrifugation may help move the exoskeleton to the bottom of the vial. An additional volume of NaOH is added to the tubes and allowed to soak for 24 hr at room temperature, after which the NaOH can again be removed and discarded. Similar volumes of water should then be added to samples and removed twice to rinse any remaining NaOH from samples.
Exoskeleton content is then the mass of sample remaining in the vial. Assumes that all nitrogen in a sample is in the form of protein with 16% nitrogen The estimated protein content of a sample will vary depending on the method used and each has biases. Ideally, analysis of hydrolysed amino acids could be used to determine which assay is most appropriate for a group of organisms. Alternatively, users can measure samples using multiple assays and take the average of those estimates

| Macronutrient determination
Colorimetric assays were selected for the determination of macronutrients, given their ease-of-use and capacity for high-throughput assaying of samples in 96-well plates (Cheng et al., 2011;Rodrı et al., 2008). All absorbance measurements were obtained from a Tecan Infinity M200 Pro plate reader ( for total lipid content, gravimetric methods are the most appropriate option, but difficult for small invertebrates without specialised scales. Samples for lipid analysis comprised the initial supernatant taken after chloroform/methanol extraction.
Given the range of available protein assays, each with different benefits, the same samples from the five species analysed were put through two different protein-based colorimetric assays: bicinchoninic acid (BCA) and Lowry assays (Figure 4; Supporting Information 2). These assays followed the manufacturer protocols for BCA and Lowry assays. Samples for protein analysis comprised the supernatant taken after NaOH extraction.
For carbohydrate determination, the anthrone method, originally proposed by Dreywood (1946), was adapted ( For each assay, the absorbance measurement of the blank standard was subtracted from all other absorbance measurements and a standard calibration curve prepared by plotting the blank-corrected measurement of the standards against their known concentrations.
The regression equation of the standard curve was used to determine macronutrient concentration in each sample in mg/ml, which was then used to calculate the total macronutrient concentration in the sample based on the sample weight used for analysis and any dilution that was applied to the sample.

| Limits of MEDI
Limit of blank (

| Statistical analysis
All statistical analyses were performed using R version 3.5.3 (R Core Team, 2020). To compare the macronutrient content of the species analysed, multivariate linear models (MLMs) were fitted using the 'manylm' function of the mvabund package (Wang et al., 2012).

| Calculation of methodological boundaries
The gravimetric lipid mass and exoskeletal mass were determined for the three focal species for which body mass could be accurately measured (Table 4).

Macronutrient Extraction and Determination from Invertebrates
(MEDI) successfully measured macronutrient content directly and rapidly from the same macroinvertebrate, even as small as a single aphid or collembolan, or as large as a tenebrionid larva or German cockroach. Aphid macronutrient content exceeds the LoDs except for carbohydrate, confirming a sensitivity broadly appropriate for small arthropods and other samples. The relatively low concentration of lipid and carbohydrate estimated in many invertebrate bodies may result in difficulties quantifying at least carbohydrates in such invertebrates (Bryer et al., 2015;Finke, 2005), but the extraction procedure could overcome this by using smaller solvent volumes (e.g. 0.5 ml) to increase the solution concentration, leaving enough material to complete all three assays, or altering the plate incubation times, reagent con- TA B L E 2 Limit of blank (LoB), limit of detection (LoD) and smallest detectable difference (SDD) calculated from repeat methanol blanks, and single aphid macronutrient content (mean ± SD)  (Vashist & Dixit, 2011) and indeed most assays. Such inhibitors could be eliminated by introducing a purification step such as trichloroacetic acid protein precipitation, but this is unlikely to be necessary in most cases. Rather than highlighting an optimal assay, this emphasises the importance of selecting assays and standards to best match the context of the work being carried out. The detection of different amino acids by each assay, their consequently differential relevance to protein standards and their variable performance in the presence of inhibitory compounds thus warrants a case-by-case consideration of the optimal assay to use, or the averaging of values from a range of assays or standards. relationships can be determined from many individuals, but the accuracy, particularly for smaller invertebrates, could be poor.
Such pooling, if maintained for assay preparation, could ensure sufficient concentrations to overcome the limits of detection for smaller invertebrates.
Micronutrients were not considered in this protocol, despite their biological importance (Jing et al., 2014), as they do not comprise a single detectable or quantifiable group. Without considering a specific micronutrient, or a subset of them, their quantification can be laborious and, given the expectedly minute content of micronutrients in each invertebrate, detection, much less quantification, of micronutrients may be unfeasible for all but the largest macroinvertebrates without specialised equipment.

| SUMMARY
Macronutrient Extraction and Determination from Invertebrates accurately detects macronutrients for a broad range of potential experimental applications involving invertebrates and other tissues, improving upon existing protocols for macronutrient determination. The protocol is relatively cheap, fast and simple and could present a uniform standard to be used across ecological studies.

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.