Tracking lipid synthesis using ²H₂O and ²H-NMR spectroscopy in black soldier fly (Hermetia illucens) larvae fed with macroalgae Free

Breeding insects is a sustainable approach to obtain high-quality protein and lipids for food and/or feed (Ameixa et al., 2020; van Huis et al., 2013). From about 2000 edible insect species (https://www.wur.nl/en/research-results/chair-groups/plant-sciences/laboratory-of-entomology/edible-insects/worldwide-species-list.htm), the black soldier fly (Hermetia illucens) stands out because of its ability to convert low-value feedstuffs, such as co-products from several industries, into high-value insect protein that can be included in fish feed formulations (Basto et al., 2020; Belghit et al., 2019; Guerreiro et al., 2020; Nogales–Mérida et al., 2019; Roques et al., 2020).

Black soldier flies accumulate energy reserves as triacylglycerols (TAGs) in the fat body during the larval stage (Pimentel et al., 2017), providing sustenance for subsequent adult life and reproduction (Liu et al., 2017), resulting in a high fat content (∼20–50% dry mass), with saturated fatty acids (SFAs) comprising between 60% and 80% of the fatty acid (FA) profile (Caligiani et al., 2018; Rodrigues et al., 2022b; Surendra et al., 2020). However, marine aquaculture animals require diets enriched in polyunsaturated fatty acids (PUFAs), especially long chain omega-3 (n-3), which are typically absent in black soldier flies. The usual approach is defattening insect meals, which can then be incorporated into aquafeeds in considerable amounts if additional n-3 PUFAs are supplied from fish oil (Belghit et al., 2018; Belghit et al., 2019), and more recently algal oil (Zatti et al., 2023). Another option, as the FA profile of the growing larvae is closely tied to their diet, is to manipulate it by using different rearing substrates (Ameixa et al., 2023; Rodrigues et al., 2022b), including marine-based n-3 enriched substrates, such as fish offal (St-Hilaire et al., 2007), expired aquafeed (Rodrigues et al., 2022a), macroalgae (Belghit et al., 2018; Liland et al., 2017) and microalgae (El-Dakar et al., 2020; Erbland et al., 2020). Nevertheless, larval growth rate and final mass are often impaired and the enrichment in n-3 FA is deficient (Ewald et al., 2020).

The FA profile is also a consequence of lipid metabolism, as FAs may be modified or synthesized through de novo lipogenesis (DNL). Lipid metabolism can be tracked using stable isotopes, such as deuterium (²H) and carbon-13 (¹³C) (Belew and Jones, 2022; Visser et al., 2017). ²H is usually supplied in deuterated water (²H₂O), which is injected intraperitoneally (Duarte et al., 2014; Murphy, 2006; Viegas et al., 2017) and/or supplied in the drinking/feeding water, or added to the growth medium, as in the case of aquatic animals such as fish or mosquitoes (Tose et al., 2021; Viegas et al., 2011). Labelled water hydrogen atoms in the animal's body water will then label a variety of biosynthetic products via multiple reactions (Jones et al., 2001). In the lipids, ²H is incorporated by fatty acid synthase (FAS) and ELOVL enzymes from labelled acyl-CoA, NADPH and the body water itself (Fig. 1). ²H enrichment in the lipids can be evaluated by mass spectrometry (MS) (Hoc et al., 2020; Visser et al., 2017) or nuclear magnetic resonance (NMR) methods. ¹H/²H NMR spectroscopy enables the characterization of the major groups of FAs and reveals lipid fluxes, including de novo lipogenesis, elongation, desaturation and glycerol backbone synthesis, by measuring the resonance of specific functional groups within the triacylglycerol (TAG) molecules in the lipid extract (Fig. 1). This ²H-NMR approach has been successfully employed in vertebrates, including fish (Viegas et al., 2016), birds (Viegas et al., 2017), mice (Duarte et al., 2014) and humans (Belew and Jones, 2022). MS relies on the mass difference of specific lipids (M+1, M+2, etc.) caused by isotope enrichment. The most common approach involves analysing FA methyl esters via gaseous chromatography (GC-MS) after lipid transesterification (Hoc et al., 2020), while liquid chromatography methods enable the analysis of whole lipid molecular species. Information on positional enrichment may be obtained from the fragmentation patterns, although derivatization procedures and mathematical modelling of the fragment ions may be required (Castro-Perez et al., 2011; Deja et al., 2020). Regardless of the system, DNL is inferred from the enrichment of the acyl chain, not the methyl terminal, as in NMR (Fig. 1). This allows a distinction between DNL and elongation of dietary FAs. Furthermore, NMR enables the application of two stable isotope tracers simultaneously (Silva et al., 2019), which would be challenging to confidently assign by MS (Deja et al., 2020), and does not require sample derivatization. To the best of our knowledge, ²H₂O tracer studies in insects thus far have employed MS-based approaches (Hoc et al., 2020; Tose et al., 2021; Visser et al., 2017).

This study presents a ¹H/²H-NMR-based methodology for measuring lipid synthesis in black soldier fly larvae using ²H₂O as a stable isotopic tracer, building upon previously established NMR lipidomics. Specifically, we demonstrate the delivery of ²H from diluted ²H₂O present in the feeding substrate, first through the successful ²H-labelling of the body water and subsequently the ²H-incorporation into fatty acids esterified into TAG. The feasibility of this approach was validated by exploring lipid synthesis in black soldier fly larvae fed with a macroalgae-based diet.

The trials were carried out at the ENTOLAB rearing facilities located at ECOMARE – Laboratory for Innovation and Sustainability of Marine Biological Resources of the University of Aveiro, Portugal.

In this facility, a black soldier fly, Hermetia illucens (Linnaeus 1758), colony is continuously maintained under controlled conditions (photoperiod 16 h:8 h light:dark, relative humidity 40±5%, temperature 28±3°C) (and larvae are fed with a control diet consisting of chick feed (Rações Valouro SA, Ramalhal, Portugal) and tap water (1:1 v/v). When the larvae reach the prepupae instar they are transferred into net cages where adult emergence occurs. After emergence, males and females are allowed to mate, and after a few days, females lay eggs in clutches, which are recovered every 2 days. After hatching, neonate larvae used in the feeding trials are fed with control diet for 5 days (Bosch et al., 2020).

For the body water enrichment over time, 10 day old larvae, were collected and placed in an enclosed plastic flask containing control diet prepared with a 10% solution of deuterated water (deuterium oxide, ²H₂O, 99.8%, CortecNet, Les Ulis, France). Three samples, consisting of two random larvae each, were collected over a 24 h period at several time intervals (20, 40, 60 and 90 min, 2, 4, 6, 9, 12 and 24 h). Larvae were cleaned using tap water, dried in paper towels and stored at −80°C.

Macroalgae Gracilaria vermiculophylla (Ohmi) (Gracilariales, Rhodophyta) fresh biomass was collected in the Aveiro coastal lagoon (Aveiro, Portugal, 40°38′07.2″N 8°39′41.8″W). All harvested macroalgae were washed with tap water to eliminate impurities and epibionts, and stored at −20°C until further use. Experimental diets consisted of partial replacement of chick feed with G. vermiculophylla biomass as follows: 0% (control), 25% and 50% (of wet mass).

For the feeding trial, 5 day old larvae were counted into groups of 500 individuals each, weighed and placed in labelled plastic containers (500 ml). Three replicates were used for each treatment, and the feeding ratio considered for the trial was 15 mg larva⁻¹ day⁻¹. After a 10 day period, a sample of 10 individuals was randomly selected from each container and the remaining larvae were placed in cages for mating purposes. The 10-larvae groups were placed in closed plastic cups containing the respective diets prepared with 10% ²H₂O. After 24 h, the larvae were collected, cleaned and stored at −20°C. From each 10-larvae group, two were used for body water extraction and three were subjected to lipid extraction, as detailed in the next section.

Solvents were of HPLC grade and purchased from Sigma-Aldrich (St Louis, MO, USA). Body water was recovered by placing a microcentrifuge vial with the insect samples in a heat block (80°C, 90 min; TD 150 P1, FALC Instruments, Treviglio, Italy; ref. 621.0422.01) and transferring the water that condensed on the lid to another vial. Condensation was further promoted by rubbing ice on the lids. Three body water samples per dietary treatment (2 larvae per sample) were analysed in duplicate. ²H enrichment in the body water was determined as described by Jones et al. (2001). Larvae lipid extraction was performed following the MTBE protocol (Matyash et al., 2008). Samples (n=9 per diet,), consisting of single larvae (∼150 mg), were stripped from their skin by squeezing, and the lipids were extracted using methyl tert-butyl ether, methanol and water (MTBE:CH₃OH:H₂O, 10:3:2.5, 24 ml g⁻¹).

NMR spectra were obtained at 25°C with a Bruker Avance III HD system with an UltraShield Plus magnet (11.7T, ¹H operating frequency 500 MHz) equipped with a 5 mm ²H-selective probe with ¹⁹F lock and ¹H-decoupling coil. Body water ²H enrichment was determined as weight per cent from 10 μl aliquots of insect BW by ²H NMR as described in Jones et al. (2001), where water content was assumed to be 99%. TAG samples were reconstituted in chloroform containing a pyrazine standard and hexafluorobenzene (for field frequency lock) as previously described (Viegas et al., 2016). Briefly, ¹H NMR spectra were acquired with a 90 deg pulse, 3 s of acquisition time and 8 s of delay for 16 scans. ²H NMR spectra were acquired with a 90 deg pulse, 0.67 s of acquisition time and 8 s of delay, with the number of scans ranging from 1500 to 2500, corresponding to approximately 5 h of collection time. TAG ²H enrichments were quantified from the ¹H and ²H NMR spectra by measuring the ¹H and ²H intensities of selected signals relative to the ¹H and ²H intensities of the pyrazine standard (representative spectra presented in Fig. S1; detail of ²H spectrum in Fig. 1 inset) according to Duarte et al. (2014). The ²H enrichment in the FA terminal methyl site (0.9 ppm, Fig. 1 and Fig. S1 signal a) for TAG-bound FA derives from de novo lipogenesis. The ²H enrichment in the sn-1,3 glycerol site (4.15 and 4.30 ppm; Fig. S1 signal l) derives from cycled TAG-bound glycerol. FA desaturation, leading to the formation of monounsaturated FAs (MUFAs), is quantified from ²H enrichment of the allylic hydrogens (1.9 ppm, Fig. 1 and Fig. S1 signal f), i.e. the hydrogen atoms bonded to the carbons adjacent to the ones forming the double bond. Excess TAG positional ²H enrichments were calculated after systematic subtraction of the values with 0.015% taken as the mean background ²H enrichment. Fractional synthesis rates (FSRs; in % day⁻¹) were estimated by dividing these positional TAG enrichments by that of body water. The spectra were processed by applying exponential multiplication to the free-induction decay (¹H: 0.1 Hz; ²H: 1.0 Hz) and analysed using the curve-fitting routine supplied with ACD Labs 1D NMR processor software 2.4 (Advanced Chemistry Development, Inc.).

Data are presented as means±s.d. Data from lipid FSRs between diets were tested for normality and homoscedasticity using the Shapiro–Wilk and Bartlett's tests, respectively. Differences were compared using the non-parametric Kruskal–Wallis test and Dunn's test, corrected for multiple comparisons. Differences were considered statistically significant at P<0.05. Analyses and figure generation were performed in GraphPad Prism 8 (GraphPad Software, Inc.).

The delivery of the stable isotope tracer, i.e. deuterated water, via the feeding media was adapted to the biology of this animal model. Injection, routinely done with vertebrates (Belew and Jones, 2022; Murphy, 2006; Viegas et al., 2022), was not an option owing to the limited size of the individuals. However, because the larvae have a high surface area-to-volume ratio, ²H assimilation in the body water occurs rapidly.

As shown in Fig. 2, ²H enrichment in the body water was detectable from 20 min of exposure onwards, rapidly increasing in the first hours and then slowing down, gradually reaching a plateau (e.g. ∼1% at 2 h, ∼2% at 4 h, ∼3% at 12 h). Variance in the enrichment values can mostly be attributed to individual variation in larval activity and feeding rate, e.g. during moulting. A logarithmic curve was fitted to all data points, as shown in Fig. 2, as follows: y=1.360×ln(t+1)+0.015 (R²=0.938), where t is time, in hours, and the interception was constrained at ²H natural abundance, assumed to be the 0.015% at t=0. According to this equation, ²H enrichment stabilizes at ∼5% after 38 h, whereas our experimental data show that these values were achieved at 24 h (Fig. 2 and next section).

A theoretical increase of up to 6% was estimated only after 80 h (3.3 days), which is not viable. The exposure to the enriched media started after 15 days of larval growth, roughly corresponding to the beginning of prepupae stage, when the larvae cease feeding, thus halting the ingestion of ²H₂O. Therefore, if higher ²H labelling is required, the tracer could be delivered earlier in larval development or in slightly higher concentration.

The standard control diet used in black soldier fly experiments is dry chick feed, to which water is added at 50–70% total mass (Bosch et al., 2020), so it is only logical but also practical to provide ²H₂O there. This can be replicated easily for similar substrates, and potentially for studying other invertebrate species. A limitation of this method is when testing biowaste substrates naturally high in water (e.g. vegetables), where adding more water would render the substrate inappropriate for the larvae. Additionally, if the larval feeding rate is impacted (e.g. by a low palatability substrate), ²H labelling may also slow down.

Incorporation of ²H at significantly higher concentrations (>20%) is not recommended as toxic and even lethal effects may be observed in the animals, besides making the experiment more expensive. This is related to the distinct solvent properties of ²H₂O compared with H₂O and isotopic effects, that render deuterated molecules less reactive (Kushner et al., 1999). In insects, high ²H concentrations slow down metabolism, decreasing growth rate and final larval mass (Hoc et al., 2020; White et al., 1992). In fact, when black soldier flies were grown with >99% deuterated water in the feeding media, various FAs were not completely metabolized, effectively skewing the FA profile and lowering the total amount of newly synthesized FAs in the larvae, although survival was not affected (Hoc et al., 2020). A lower ²H dosage, as proposed here, avoids these issues. A final body water ²H enrichment around 5% was actually the target in this study, as previous work from our group achieved similar enrichment values: 5% in fish, with ²H provided in the tank water (Viegas et al., 2011), ∼8% in birds (Viegas et al., 2017) and 3–5% in humans (Belew and Jones, 2022; Jones et al., 2001), with ²H provided by injection (99%) and drinking water (5% enriched). The ²H enrichment of lipids would theoretically be observed after 24 h by allowing isotopic equilibrium between feed and body water ²H content (Dufner et al., 2005), and subsequently adequate labelling of lipids, a prerequisite for delivering an accurate and representative quantification of de novo lipogenesis.

This experiment confirmed the assimilation of ²H in black soldier fly tissues. The larvae evaluated in the feeding trial for lipid synthesis had a body mass ²H enrichment of 4.95±1.78%, which is consistent with the body mass ²H enrichment curve for 24 h.

FA FSR measured from the terminal CH₃ (Fig. 1; signal a) averaged 1.82±1.75% day⁻¹. ²H enrichment on the glycerol backbone (Fig. S1; signal l) was mostly below detection levels, with a broad signal envelope only visible in three ²H spectra. Likewise, ²H enrichment in the MUFA allylic hydrogens (Fig. 1; signal f) was only barely detectable in the same three spectra (signal-to-noise ratio below 1). This means that the formation (by synthesis or turnover) of glycerol was negligible for the duration of the experiment. Moreover, the contribution of desaturation to the FA profile was insignificant. Thus, during the exposure time, black soldier fly larvae synthesized essentially SFAs and used pre-existing (unlabelled) glycerol molecules, which reinforces the capability of this NMR approach to further characterize lipid synthesis patterns. Black soldier fly larvae essentially synthesize SFAs, in agreement with other studies (Hoc et al., 2020). This was expected, as the FA profile is typically dominated by lauric acid, the 12 carbon-long saturated FA (Rodrigues et al., 2022b) that is virtually absent from the diet.

According to the Kruskal–Wallis test, DNL measured through ²H enrichment was significantly different between dietary treatments (P=0.017). The larvae reared on the control diet presented the highest DNL values (Fig. 3), which were statistically different from those of the 50% diet group (P=0.014), while the 25% diet group had intermediate values, with no significant differences from the other groups (control P=0.722, 50% diet P=0.289). Hence, lipid synthesis declined with increasing G. vermiculophylla content in the diet. Macroalgae (seaweeds) are not ideal feed ingredients for many animals because of the elevated water, salt and complex carbohydrates they contain (Liland et al., 2017), and incorporating algae into the diet of black soldier flies impacts larval development. For example, Ascophyllum nodosum (Phaeophyceae) incorporation caused a reduction in larvae size and mass at incorporation levels lower than 50%, while higher incorporation increased mortality (Liland et al., 2017). Similar impacts on larval growth, survival and final biomass were observed with microalgae (Erbland et al., 2020). Lower lipid content was also observed in these studies, together with a reduction in growth (larval mass), and possible root causes include various physicochemical substrate traits, including dietary imbalances, particle size and the occurrence of phenolic compounds and transition metals (Biancarosa et al., 2018; Liland et al., 2017). The nutritional value of G. vermiculophylla from the Portuguese coast was recently published (Afonso et al., 2021). Like other red seaweeds (Rhodophyta), the protein content is relatively high (17% dry mass basis). However, both carbohydrates and proteins have low digestibility, and some amino acids, such as methionine and lysin, may become limiting (Afonso et al., 2021). Therefore, it is probable that black soldier flies supplied with macroalgae will gain mass slower than those on control diets because of the low nutrient availability, particularly lipids and highly digestible carbohydrates, leading to reduced lipid synthesis rates. Moreover, the proportionally higher n-3 FA content of black soldier flies fed with macroalgae or other marine-sourced substrates (Liland et al., 2017; Rodrigues et al., 2022a) could partially result from a reduction in SFA synthesis. As de novo FAs are saturated, with limited nutritional value, lower lipid synthesis rates may actually be preferable for insect meal applications, especially for aquafeeds. The reduced ability of black soldier flies to adapt to certain substrates is a major reason for diversifying the insect species that are used in insect meals (Duarte et al., 2021; Parry et al., 2021).

There are significant gaps in our understanding of lipid metabolism in black soldier flies, particularly when exposed to various rearing substrates. Most studies underline the close relationship between the FA profile of the larvae and their diet, but the significance of shifting lipid synthesis patterns may have been overlooked (Hoc et al., 2020; Rodrigues et al., 2022b).

Our proposed methodology, combining innovative diluted ²H₂O delivery through the feeding media with well-established analysis of ²H enrichment using ¹H/²H NMR spectrometry, was proven adequate for research into the lipid metabolism of black soldier fly larvae. The possibility of analysing both body water and lipids using a single tool, without requiring any derivatization, sets it apart from MS studies. Furthermore, in NMR, ¹H and ²H nuclei resonate at distinct frequencies, enabling the independent quantification of ²H enrichment with greater precision compared with MS. Depending on the specific needs of future investigations, the administration of ²H₂O should be adjusted. When using natural substrates, the water content must be carefully monitored to allow a constant ²H enrichment between diets. If higher enrichment is required, e.g. to accurately quantify FA desaturation, ²H may be provided earlier in larval development.

We thank Catarina Castro for her technical assistance with insect care. NMR data was collected at the UC-NMR facility which is supported in part by FEDER - European Regional Development Fund through the COMPETE Programme (Operational Programme for Competitiveness) and by National Funds through FCT - Fundação para a Ciência e a Tecnologia (Portuguese Foundation for Science and Technology) through grants RECI/QEQ-QFI/0168/2012, CENTRO-07-CT62-FEDER-002012, and also through support to Rede Nacional de Ressonância Magnética Nuclear (RNRMN) and to Coimbra Chemistry Centre through grant UID/QUI/00313/2019.