¹H-NMR Profiling as a Tool for Assessing Frass Hygienization and Storage Stability

1. Introduction

The insect-based protein sector has expanded in recent years as part of the transition towards more resilient, resource-efficient, and circular food and feed systems. This deve-lopment is particularly relevant in Europe, where approximately 66% of high-protein feed is imported, including most of the soybean meal used in the European livestock sector [1,2]. Insect farming may contribute to greater protein autonomy by converting selected agri-food by-products into valuable biomass while reducing the loss of nutrients from the food chain. Its environmental performance, however, depends on factors such as the feed substrate, energy requirements, rearing conditions, and processing methods [3,4].

The growth of insect farming also increases the need to valorize its residual streams. One of the main by-products is frass, a heterogeneous material comprising insect excreta, feeding substrate, parts of farmed insects, and dead eggs. Depending on the insect species, diet, rearing conditions, and subsequent processing, frass may contain organic matter, mineral nutrients, chitinous residues, soluble carbohydrates, nitrogen-containing compounds, and other metabolites with potential relevance for soil fertility and plant growth [2,5]. Its use as a fertilizer or soil amendment could therefore provide an additional valorization route for insect-production residues while contributing to nutrient recycling in agricultural systems [6].

However, frass is not a chemically uniform material. Its composition and agronomic properties vary according to the insect species and the substrate used for rearing [1,6]. For example, Castillo et al. reported differences among frass from four insect species in nutrient composition, dissolved organic carbon, organic-matter characteristics, ecotoxicity, and phytotoxicity. Their results also showed that Tenebrio molitor frass had a high dissolved organic carbon content but produced phytotoxic effects in germination assays, which were reduced by dilution [5]. This variability highlights the need for reliable analytical methods to characterize frass composition and evaluate its stability before agricultural application [7].

The commercialization of insect-derived materials is subject to European regulatory requirements. Regulation (European Union—EU) No 142/2011 [8], as amended by Regulation (EU) 2021/1925 [9], establishes the regulatory definition of frass and the conditions applicable to certain insect-derived products. Regulation (EU) 2019/1009 establishes the framework for EU fertilising products and includes requirements concerning their composition and safety [10]. For frass intended for use in fertilising products, the applicable provisions include thermal treatment at 70 °C for at least 1 h and compliance with defined microbiological criteria [2,9]. These requirements are intended to reduce microbiological hazards, but the effect of the treatment on the subsequent metabolic stability of frass remains insufficiently characterized [2,11].

Analytical approaches that can monitor several metabolites simultaneously may therefore complement conventional microbiological testing [2,12,13]. ¹H-NMR spectroscopy provides rapid and reproducible metabolic profiling with limited sample preparation and has been applied to the characterization of complex biological and organic matrices [13,14,15]. Although NMR does not directly quantify viable microorganisms or identify microbial taxa, treatment- and storage-dependent changes in metabolite profiles may provide complementary information on the biochemical evolution and stability of frass [2,13].

In this study, ¹H- NMR profiling was applied to aqueous extracts of Tenebrio molitor frass subjected to three thermal conditions: no heat treatment, 70 °C for 1 h in accordance with the applicable European requirements, and an intensified treatment at 70 °C for 18 h. Samples were analyzed at different storage times to assess the evolution of their metabolite profiles during storage. The aim was to determine whether ¹H NMR could distinguish treatment-dependent differences in frass stability and provide a complementary approach for evaluating thermal hygienization and storage-related metabolic changes in frass intended for use as a fertilizer or soil amendment.

2. Materials and Methods

Tenebrio molitor larvae were obtained from a commercial supplier and maintained in separate containers under dark conditions at room temperature (22–24 °C). Two dietary models were used: a control diet based on oat flakes and an experimental diet consisting of by-products from the stone-fruit processing industry. The latter comprised milled peach stone fractions, including seed and shell material. Carrots were supplied to all groups as a moisture source. The rearing period was set to 71 days to monitor survival and the complete larva-to-adult developmental cycle. During this time, the larvae were fed weekly as dietary substrates were transformed. At the end of the rearing period, all living and dead organisms were removed from the containers. The remaining material was separated using sieves with mesh sizes of 3.15 mm, 1 mm, and 0.5 mm. Frass was operationally defined as the fraction passing through the 0.5 mm sieve. This fraction was collected and stored until analysis.

A total of 41 aqueous extracts were analyzed. The extracts were prepared from frass samples previously subjected to three treatments: no thermal treatment, heating at 70 °C for 1h, and heating at 70 °C for 18 h. The 70 °C for 1 h treatment was selected to reproduce the temperature–time condition specified in the applicable European regulatory framework for insect frass [10]. The 70 °C for 18 h treatment was included as an intensified experimental condition for comparison. Samples were analyzed at different storage times (1, 7 and 14 days) to monitor the evolution of their metabolite profiles during storage.

For the aqueous extractions, 25 mg of dried frass was suspended in 2 mL of deutered water (D₂O > 99.9 atom% D, CAS no. 7789-20-0, CortecNet) in capped glass vials. The suspensions were agitated for 30 min at room temperature, after which 700 μL of each extract was transferred to a conventional 5 mm NMR tube (Wilmad, CortecNet). The ¹H NMR spectra were acquired using a Bruker AVIII 500 MHz spectrometer equipped with a conventional liquid-state probe. Measurements were performed at room temperature using the Bruker one-dimensional water-suppressed NOESYPR1D pulse sequence. A total of 128 scans were accumulated for each spectrum, corresponding to an acquisition time of approximately 10 min per sample.

Additionally, five non-polar extractions were performed as a control for the lipid fraction, following the same procedure as for the aqueous fraction but using deuterated chloroform (CDCl₃, > 99.8 atom% D, CAS no. 865-49-6, Scharlau) as the solvent. For these extracts, a standard single-pulse one-dimensional ¹H NMR experiment (Bruker zg) with 128 scans under identical instrumental conditions.

For selected aqueous extracts, ¹H–¹H COSY, ¹H–¹³C HSQC, ¹H–¹³C HMBC experiments were acquired using standard Bruker pulse sequences to support metabolite assignment. The resulting ¹H–¹H and ¹H–¹³C correlations were evaluated together with chemical-shift values, resonance multiplicities, reference spectra, and previously established protocols for the analysis of polar fractions from complex organic matrices [16]. The Biological Magnetic Resonance Data Bank (BMRB) and Human Metabolome Database (HMDB) databases were employed as reference sources for metabolite confirmation [17,18]. One- and two-dimensional spectra were acquired and processed using TopSpin 3.7 (Bruker BioSpin, Rheinstetten, Germany) and MestReNova 11.0.4 (Mestrelab Research, Santiago de Compostela, Spain). Processing comprised Fourier transformation, phase correction, baseline correction, and chemical-shift referencing.

During the preparation of this manuscript, generative artificial intelligence was used to assist with English-language editing and manuscript organization. The authors reviewed and edited the generated text and take full responsibility for the accuracy, originality, and integrity of the final manuscript. No artificial intelligence tool was used to generate experimental data or replace data analysis and scientific interpretation.

3. Results

Representative ¹H NMR profiles of aqueous extracts from untreated Tenebrio molitor frass obtained after 1 day of storage from larvae fed either an oat-based diet or peach stone-derived material containing seed and shell fractions are shown in Figure 1. The two dietary models displayed distinct spectral profiles over the 0.6–4.5 ppm region, reflecting differences in the composition of the corresponding aqueous extracts. The most noticeable variation was observed in the carbohydrate-associated region (3.0–4.5 ppm), where the oat-based frass extract showed a greater relative contribution of resonances attributed to soluble carbohydrate-related compounds than the extract obtained from the peach-derived diet (Figure 1, dashed box).

In both spectra, the most intense signal was a singlet marked with an asterisk and assigned to glycine betaine. Its detection in the aqueous extracts is consistent with the high water solubility and zwitterionic character of glycine betaine. This assignment agrees with the ¹H-HRMAS NMR study by Gil-Gonzalo et al., in which betaine-related resonances were detected directly in intact frass samples, indicating that these metabolites can be observed both in the whole semi-solid material and in its water-soluble fraction [13].

In the spectrum of Figure 1 (polar fraction), broad signals are also observed in the aliphatic region (approximately 0.8–2.3 ppm), compatible with fatty acid residues. These signals are attributed to shorter-chain or partially oxidized fatty acid derivatives that are sufficiently polar to co-extract into the aqueous phase together with the water-soluble metabolites. Complementarily, non-polar extractions of the frass samples were performed using CDCl₃ to evaluate the status of the fatty acids after digestion by Tenebrio molitor. In all samples, the glycerol signals of triglycerides (sn-1,3 and sn-2) were totally or almost totally absent, whereas the fatty acid profile was preserved (Figure S1, Supplementary Information). Moreover, the triplet corresponding to the α-CH₂ group adjacent to the carboxylic function was shifted downfield relative to its typical position in triglycerides, in agreement with the expected chemical shift change upon conversion of esterified fatty acids to free fatty acids [19]. This behavior is consistent with extensive hydrolysis of triglycerides during metabolization by T. molitor.

As described in Materials and Methods, frass samples were stored for 1 and 7 days and either analyzed directly or after a thermal treatment at 70 °C for 1 h, in accordance with Regulation (EU) 2019/1009. Figure 2 presents an expanded view (4.3–5.5 ppm) of the

¹H-NMR spectra of aqueous Tenebrio molitor frass extracts, comparing untreated samples with those subjected to this thermal treatment at each storage time. This spectral region contains the characteristic doublet resonances assigned to the anomeric hydrogens of α-glucose (α-Glc) and β-glucose (β-Glc), allowing changes in the soluble sugar profile to be monitored over time.

Clear differences were observed between the two processing conditions. In untreated samples, the glucose-related doublets clearly detected after 1 day of storage were no longer visible after 7 days. By contrast, samples treated at 70 °C for 1 h retained the characteristic doublets associated with both glucose anomers after 7 days, although with lower intensities than those observed at day 1.

Figure 3 presents the temporal evolution of glycerol and selected glycerol-associated metabolites in aqueous extracts of Tenebrio molitor frass. The expanded ¹H NMR region between 3.44 and 3.97 ppm was examined for samples subjected to three processing conditions: no thermal treatment, thermal treatment at 70 °C for 1 h in accordance with Regulation (EU) 2019/1009, and extended thermal treatment at 70 °C for 18 h. For each condition, spectra were recorded at 1, 7, and 14 days of storage.

At day 1, the selected spectral region was dominated by the glycerol resonances, particularly the doublet of doublets corresponding to the terminal CH₂–OH groups, together with the singlet assigned to glycine betaine at 3.92 ppm. The singlet at 3.567 ppm, assigned to the CH₂ group of glycine, was also present throughout the storage period but became more clearly resolved at later time points as the overlapping glycerol contribution decreased. Between days 1 and 7, the glycerol resonance remained essentially unchanged under all processing conditions, and no marked appearance of additional resonances was observed within the monitored spectral region.

By day 14, clear time- and treatment-dependent differences were observed. In untreated samples, the glycerol resonance was no longer detectable and was accompanied by the marked appearance of signals assigned to glycerol transformation products. These included the triplet at 3.699 ppm assigned to the terminal hydroxymetylene groups (CH₂–OH) of 1,3-propanediol, and the triplet at 3.805 ppm assigned to the CH₂–OH group of 3-hydroxypropionate (3-HP).

Samples treated at 70 °C for 1 h retained a detectable glycerol resonance at day 14, although the same product-associated signals observed in untreated samples were also present. In contrast, samples subjected to the extended treatment at 70 °C for 18 h showed a substantially more stable spectral profile throughout storage, with only minor differences between days 1, 7, and 14. Across all processing conditions and storage times, the glycine betaine resonance at 3.92 ppm remained essentially unchanged in both chemical shift and relative intensity, whereas glycerol and its associated transformation products showed clear treatment- and time-dependent changes.

The corresponding triplet at 2.447 ppm, tentatively assigned to the methylene group adjacent to the carboxylate moiety (CH₂–COO⁻) of 3-hydroxypropionate, and the quintet at 1.809 ppm assigned to the central methylene group of 1,3-propanediol (–CH₂–), were also observed outside the spectral range displayed in Figure 3. These resonances, together with other metabolites identified during the time-dependent evolution of the aqueous frass extracts, are listed in Table S1 in the Supplementary Information.

4. Discussion

The present results show that ¹H NMR profiling can resolve diet-, treatment-, and storage-dependent changes in the aqueous metabolite profile of Tenebrio molitor frass. The distinct spectra obtained from oat-fed larvae and from larvae fed peach seed and stone residues indicate that the rearing substrate strongly influences the water-soluble metabolome, particularly the carbohydrate-associated region. This variability likely reflects both the carry-over of diet-derived compounds and metabolic contributions from the insect and its associated microbiota. With minimal sample preparation, ¹H NMR therefore provides a rapid and non-destructive method for monitoring abundant low-molecular-weight metabolites and their stability in complex biological matrices, as previously demonstrated for food and related systems [15].

The broad aliphatic resonances observed in the aqueous extracts suggest the co-extraction of relatively polar lipid-derived compounds. Complementary analysis of the non-polar fraction showed near-complete loss of triglyceride glycerol resonances, while the fatty-acid profile was retained and the α-methylene signal shifted downfield. These observations are consistent with extensive triglyceride hydrolysis and the accumulation of fatty acids predominantly in their free form in T. molitor frass.

The evolution of the carbohydrate-associated region during storage provides an initial indicator of postharvest changes in the aqueous frass fraction. In untreated samples, the α-Glc and β-Glc anomeric resonances visible at day 1 were no longer detectable after 7 days. In contrast, both signals were still present after the 70 °C/1 h treatment, but with lower intensity. This indicates that the thermal treatment slowed the loss of soluble glucose-related compounds during storage. This effect may reflect reduced microbial activity after heating.

The glycerol region provided a second, longer-term marker of the changes occurring during frass storage. In untreated samples, the glycerol resonance disappeared by day 14, whereas the singlets assigned to glycine betaine and glycine remained visible. The disappearance of glycerol in untreated frass at day 14 coincided with the appearance of resonances assigned to 1,3-propanediol and 3-hydroxypropionate. Their coordinated appearance as glycerol disappeared supports the conversion of glycerol, or of glycerol intermediates, during frass storage. In microbial systems, glycerol can enter reductive and oxidative pathways that yield low-molecular-weight C₃ metabolites, including 1,3-propanediol, 3-hydroxypropionaldehyde, and 3-hydroxypropionate, together with other fermentation end products, such as acetate, lactate, and ethanol, all of which were detected in the spectra (Table S1, Supplementary Information). The formation of 1,3-propanediol from glycerol is well established in anaerobic and facultatively anaerobic bacteria, including members of the genera Clostridium, Klebsiella, Citrobacter, and Enterobacter [20].

The response to the 70 °C/1 h treatment was intermediate between that of untreated frass and the extended 70 °C/18 h condition. Glycerol remained detectable at day 14 after the 1 h treatment, but resonances assigned to 1,3-propanediol and 3-hydroxypropionate were also present. This suggests that the 70 °C/1 h treatment delayed, but did not fully prevent, the subsequent chemical or biological evolution of the aqueous fraction. This interpretation is compatible with the intended purpose of the thermal treatment: hygienisation rather than sterilisation. Previous studies on insect frass support this distinction, showing that heat treatments sufficient to meet microbiological safety criteria (e.g., reduction of Enterobacteriaceae and absence of Salmonella) can still leave surviving microorganisms, spores, enzymes, or non-biological reaction pathways that contribute to continued chemical evolution during storage [2,11].

The extended treatment at 70 °C for 18 h produced the most stable ¹H NMR profiles, with only minor spectral differences between days 1, 7, and 14. Under the experimental conditions examined, prolonged heating was therefore associated with the greatest preservation of the initial glycerol-containing spectral profile and the lowest apparent formation of glycerol-associated transformation products. As the 18 h treatment was included as an intensified experimental condition rather than a regulatory requirement, this result should not be interpreted as evidence that prolonged heating is necessary for regulatory compliance. Instead, it highlights the versatility of ¹H NMR profiling for monitoring metabolic changes in frass extracts and for evaluating and optimizing heat-treatment protocols.

Taken together, Figure 1, Figure 2 and Figure 3 show that ¹H NMR can detect both initial diet-related variability in T. molitor frass and the subsequent changes caused by thermal processing and storage. The data identify soluble carbohydrates, glycerol, 1,3-propanediol, and 3-hydroxypropionate as informative spectral markers of frass evolution, while glycine betaine appears comparatively stable under the conditions evaluated. These results support the use of NMR as a complementary screening tool for assessing the stabilization of frass-based fertilising materials. However, NMR monitoring cannot replace microbiological testing required for regulatory compliance, because the observed metabolite profile does not directly establish pathogen absence, viable microbial counts, or the taxonomic identity of the surviving microbial community.

The presence of glycine betaine, together with its apparent stability and its reported agronomic potential as an osmoprotectant and biostimulant [21], highlights the possible value of aqueous frass extracts for agricultural applications. Future work should combine targeted NMR quantification with microbiological analyses and dose–response plant bioassays against appropriate controls.

5. Conclusions

¹H NMR profiling is a rapid tool to assess frass hygienisation and storage stability. The observed metabolite changes provide indirect evidence of residual microbial activity and treatment efficacy, while also capturing the biotransformation state of the aqueous fraction and diet-dependent effects. Overall, this approach offers a practical framework for evaluating the microbiological and chemical stability of frass-based fertilizers.