Role of Hb in lipid oxidation development during ensilaging
The role of Hb in PV and TBARS development during ensilaging of herring filleting co-products is shown in Fig. 1. Significantly (p < 0.05) increased PV´s were noticed in Hb-fortified samples after 6 h of ensilaging, compared to the controls (Fig. 1A,B). The PV´s of both Hb-fortified samples then remained around 500–1000 µmol/kg higher than the control samples throughout the whole studied period, despite the fact that the Hb levels were very different. At the end of the ensilaging, the relative differences in PV´s between two- and threefold Hb-fortified samples and their controls were 15.93% and 29.84%, respectively; however, a higher absolute level of PV was noticed in the twofold Hb-fortified samples, compared to the threefold fortified one, i.e. 7826.58 vs. 5887.89 µmole peroxide/kg, respectively.
Effect of Hb fortification on PV (A-B) and TBARS (C-D) development during ensilaging of herring filleting co-products at 22 °C. Controls refer to samples without Hb fortification, and time point zero (0) refers to the sample before ensilaging. Filleting co-products from batch-2 was used in this experiment, and their endogenous Hb-level was 68.68 µmole Hb/kg. Control-1 and 2 contained 57.23 and 49.05 µmole Hb/kg silage, respectively; and, level-1 and 2 contained 114.46 and 147.17 µmole Hb/kg silage, respectively. Star (*) sign represents significant (p < 0.05) difference between the control and Hb-fortified samples at specified time points; and, different lower-case letters along the same line denote significance (p < 0.05) difference. Results are expressed as mean ± SEM (n = 2).
In case of TBARS, significantly (p < 0.05) higher levels were noticed after 2 days and 1 day of ensilaging in 2- and threefold Hb-fortified samples, respectively, compared to the controls, and the difference between fortified and control samples gradually increased throughout the studied period (Fig. 1C,D). Contrary to PV-data, a higher level of TBARS was noticed in the sample with threefold Hb-level, compared to the one with twofold Hb-level, i.e. 1954.96 vs. 1777.02 µmole TBARS/kg, respectively, after 7 days of ensilaging. The relative difference to the control at the end of the ensilaging was also larger for the former sample, i.e. 137.60% vs. 66.51% for the sample with threefold vs. twofold Hb-level, respectively.
The fact that both PV and TBARS values increased significantly (p < 0.05) over time in all samples was in line with our earlier findings11, confirming that herring co-product ensilaging is a system with high sensitivity to oxidation, despite the gradual formation of short-chain, potentially antioxidative, peptides3. Several earlier studies have shown that fish protein-derived peptides are efficient chelators for low molecular weight (LMW) iron27; however, their inability to prevent oxidation in herring silage points at heme-bound iron or other compounds as being of higher relevance as pro-oxidants. The importance of Hb/heme was evident in this study as an elevated level of Hb resulted in an increased level of TBARS. The effect on PV was much smaller, illustrating the ability of Hb and heme to react with lipid hydroperoxides generating ferryl heme protein radicals, lipid radicals and hydroxyl radicals28, altogether preventing a significant buildup of peroxides. Earlier studies in washed cod mince models (pH 6.5–6.8) have shown a strikingly constant ratio between maximum TBARS values reached and the Hb levels added, i.e. 13.6 ± 4.8 µmole TBARS/µmole Hb tetramer (range 6.4–24, n = 15)6,15. In this present study, where Hb was both endogenous and added, the ratio was however surprisingly similar, i.e. 16.05 ± 2.25 µmole TBARS/µmole Hb tetramer (range 13.28–18.64, n = 8), even though this system was acidified and contained hydrolyzed proteins. This implies that Hb/heme strongly controls oxidation development both when present in situ and when added, and its quantification could provide a basis for predicting the subsequent degree of oxidation in both fish muscle and silage derived thereof.
Effect of ensilaging on Hb change—the trout hemolysate model system
To further understand the role of Hb in lipid oxidation during ensilaging, we used a simple trout hemolysate model system and simulated ensilaging conditions by adjusting the pH to 3.50, and then followed the shift in Hb spectra (Fig. 2A), heme group release from Hb (Fig. 2B), and Hb precipitation (Fig. 2C) over time. As can be seen in Fig. 2A, the trout hemolysate had a bright red color before ensilaging, reflecting the oxygenated state of Hb (i.e. the oxyHb form), which is also visible from its distinguishable peaks around 415, 540, and 580 nm. However, the bright red color turned into brown immediately after adjusting the pH to 3.50, illustrating the change from oxyHb to metHb, which was also evident by the disappearance of peaks around 540 and 580 nm, a shift of the peak around 415 nm to around 405 nm, as well as development of a peak around 630 nm. A peak around 375 nm also appeared in the spectra, which was probably due to release of the heme group from Hb. The decreasing intensity of this peak over time most likely reflects gradual heme group degradation with subsequent release of free iron. The significant (p < 0.05) release of the heme group from Hb immediately after adjusting the pH to 3.50 and its gradual degradation over time was also supported by the analysis of heme groups present in the hemolysate-buffer model system (Fig. 2B). Hargrove, et al.29 reported that the release of heme from Hb is around 60 times faster from metHb, compared to oxyHb and deoxyHb, which thus supports our observed increase in heme group release immediately after the acid-induced shift in Hb spectra from oxyHb to metHb. The exposure of the heme group to the surrounding environment stimulates autoxidation and subsequent heme-loss30. Most likely, these changes explain the rapid increase in PV, followed by an increased TBARS value, which took place during ensilaging of herring filleting co-products (Fig. 1). The protein content in the hemolysate-buffer model system remained constant throughout the studied time period (Fig. 2C), suggesting that Hb did not precipitate during ensilaging.
Effect of adjusting a solution of trout Hb (68.68 µmole/kg) in 0.1 M Tris buffer from pH 8.0 to pH 3.50 on shift in Hb absorption spectra (A), heme group release (B), and soluble protein content (C) over time at 22 °C. The soluble protein content was determined in the supernatant after centrifugation of the sample at 16,000 × g for 5 min (4 °C). The insets in (A,B) show the changes in Hb color from red to dark brown and heme group release from the Hb, respectively, upon adjusting pH from 8.0 to 3.50. Time point zero (0) refers to a sample before adjusting pH to 3.50 (i.e. pH 8.0). Different lower-case letters denote significance (p < 0.05) difference. Results are expressed as mean ± SEM (n = 2).
Effect of pre-incubating the co-products in water or salt solutions with and without antioxidants on lipid oxidation
To minimize Hb-mediated lipid oxidation during ensilaging, we investigated the possibility of pre-incubating the herring co-products in different treatment solutions. The hypothesis was that removing Hb from the outer layer of the co-products, with or without covering the surface by antioxidants, could possibly minimize lipid oxidation. In our first trial, we incubated the co-products in physiological salt solution (i.e. 0.9% NaCl) to prevent the lysis of erythrocytes31. In addition, we also investigated the possibility of using 3.0% NaCl to simulate seawater, which is traditionally used during pre-storage of herring prior to processing. Controls were incubation in tap water (i.e. 0% NaCl) and non-treated. Incubating herring co-products in 0.9% NaCl for 20 min has previously been shown to remove 6.6 – 18.0% Hb; the exact amount varied with the specific co-product part i.e. 6.6, 10.3, 17.9, and 18.0% from fin, head, backbone, and residuals, respectively5. However, in this study, no significant (p > 0.05) differences in TBARS were noticed after incubating the co-products for 30 s or 2 h in water or salt solutions, except for the 2 h incubation in 3.0% NaCl which resulted in a significant (p < 0.05) increase in TBARS, compared to the non-treated control (Fig. 3A). This was possibly due to the hypertonic nature of 3.0% NaCl, causing lysis of erythrocytes32, as well as the known pro-oxidative ability of NaCl in muscle tissue33. Figure 3B illustrates the actual change in ionic strength as a result of the incubations, and it can be seen that co-products incubated in 3.0% NaCl for 30 s or 20 min obtained an ionic strength of around 1.11 and 1.85% NaCl-equivalents, respectively. Analysis of the endogenous antioxidant α-tocopherol in the co-products revealed that incubating in 0.9% and 3.0% NaCl resulted in a significantly (p < 0.05) lower level of α-tocopherol content, compared to the non-treated control (Fig. 3C). Thus, the slight removal of pro-oxidative Hb was counteracted by a simultaneous loss of endogenous α-tocopherol during the incubations, which otherwise could have provided antioxidative effect by donating a hydrogen atom to free radicals such as L● or LOO● to disrupt the propagation process and ultimately reduce the formation of hydroperoxides34.
Effect of pre-incubating the co-products for 30 s or 20 min in TBARS (A), ionic strength (B), and α-tocopherol (C) immediately after incubation and after 7 days of ensilaging. Control refers to sample without any treatment. Filleting co-products from batch-1 was used in this trial. Star (*) sign represents significant (p < 0.05) difference between incubated and ensilaged samples subjected to the same treatment; and, different lower-case letters among incubation treatments or silages denote significance (p < 0.05) difference between treatments. Results are expressed as mean ± SEM (n = 3).
Increased TBARS values were noticed after 7 days of ensilaging in all treatments, and there were no significant (p > 0.05) effects from different treatments (Fig. 3A). Further, a significant (p < 0.05) increase in the ionic strength was noticed in silages (Fig. 3B), compared to the incubated co-products, which was possibly due to the addition of acid. The latter was confirmed by an increase in the ionic strength immediately after adding acid to the minced co-products, which later remained relatively constant throughout the studied ensilaging period (see supporting information; Fig. 1). Thus, in accordance with Fig. 1, these data confirmed that the ensilaging process itself, partly due to the acidic conditions, promotes lipid oxidation, which is also apparent from a significant (p < 0.05) consumption of α-tocopherol after 7 days of ensilaging (Fig. 3C). The negative correlation between TBARS and α-tocopherol level over time was in agreement with our previous study11.
Based on the outcomes of our first trial, antioxidants were added to the treatment solutions to compensate for the loss of endogenous α-tocopherol and to add extra protection. Incubating the co-products in solutions made from 0.2% isoascorbic acid with 0.044% EDTA in tap water or in 0.9% NaCl for 20 min significantly (p < 0.05) lowered the TBARS values both immediately after the incubation and after 7 days of ensilaging, compared to the non-treated control, incubated only in tap water or 0.9% NaCl (Fig. 4A). Isoascorbic acid provides antioxidative effect by scavenging free radicals and reducing hypervalent iron35, while EDTA works by chelating metal ions like ferrous and ferric iron36. The commercial antioxidant mixture Duralox MANC-213 at 5%, gave the same low TBARS values immediately after incubation, but provided a significantly (p < 0.05) stronger inhibitory effect after 7 days of ensilaging (Fig. 4A). Among the treatments used, Duralox MANC-213 provided the highest TBARS inhibition, compared to the non-treated control, both directly after incubation and after 7 days of ensilaging, i.e., 79.50 and 70.95%, respectively (Table 2). This was probably because of its multiple ingredients with rosemary extract being the major one, further to tocopherols, ascorbic acid and citric acid, providing synergistic effects37. The rosemary extract of Duralox MANC-213 contains many phenolic compounds including rosmarinic acid, carnosic acid, and carnosol; the latter two which have been reported as the most active compounds of rosemary extract38. The same authors also reported that both carnosic acid and carnosol provided better inhibitory effect at pH 4.0 than at pH 7.0, in a corn oil-in-water emulsion system oxidized at 60 °C for 4 days, which is probably due to their stability, better reducing capacity, and partitioning either in the oil phase or in the oil–water interface at lower pH values38. Our own analyses revealed a carnosic acid level of 1.55 ± 0.06 mg/g Duralox-MANC-213 and a TPC of 69.63 ± 2.29 mg gallic acid eq/g; the latter indicating there were a lot of phenolic compounds beyond the carnosic acid. Further, Duralox MANC-213 contains both hydrophilic (e.g. ascorbic acid, citric acid and aqueous rosemary-derived compounds) and lipophilic compounds (e.g. tocopherol and rosemary-derived lipophilic compounds), which possibly aided its partitioning both into the oil phase and oil–water interface under acidic ensilaging conditions, supporting its strong antioxidative protection during ensilaging.
Effect of pre-incubating the co-products in water, 0.9% NaCl or antioxidant solutions for 20 min (A,B), and reusing the solution for 4 or 10 treatments, lowering the solution to co-products ratio from 5:1 to 3:1 or 2:1, shortening the treatment time from 20 min to 30 s, and, using a rosemary extract alone (C) on TBARS development immediately after incubation and after 7 days of ensilaging. Control refers to sample without any treatment. Filleting co-products from batch 2, 3, and 4 were used for trials 2, 3, and 4 as shown in panel (A–C), respectively. Star (*) sign represents significant (p < 0.05) difference between incubation and silage samples within the same treatment; and, different lower-case letters among incubation treatments or silages denote significance (p < 0.05) difference between treatments. Results are expressed as mean ± SEM (n = 3). MANC: Duralox MANC-213; incub. incubation.
In the next trial (i.e. trial 3) the amount of Duralox MANC-213 in the incubation solution was lowered from 5 to 2% and it was also compared with 2% isoascorbic acid. Further, the effect on TBARS from pre-storing the co-products at 4 °C for 24 h after incubation was investigated to simulate a case where co-products would need to be stored/transported to another plant before being ensilaged. Results show that incubating the co-products in 2% Duralox MANC-213, with or without pre-storage at 4 °C for 24 h, was sufficient to significantly (p < 0.05) inhibit TBARS development compared to the tap water-treated control and the 2% isoascorbic acid-treated sample, both immediately after the incubation and after 7 days of ensilaging (Fig. 4B). It also gave better TBARS inhibition compared to the incubation in 5% Duralox MANC-213 (Table 2). This could however partly be due to high-quality starting raw material with less pre-formed oxidation products used in this trial. Even though there was a very slight, but still significant (p < 0.05), increase in TBARS in the pre-stored co-products after 7 days of ensilaging, our data suggest that pre-incubated co-products can be stored for some time prior to ensilaging without compromising the oxidative stability of the co-products or the silage. However, microbial stability during this pre-storage time without acidification should be considered and requires further investigation. Trial 3 also revealed that 2% isoascorbic acid alone, in relative terms, was as efficient as 0.2% isoascorbic acid with 0.044% EDTA in preventing oxidation during the incubation and subsequent ensilaging (Fig. 4A vs B, and Table 2). Isoascorbic acid alone provided 70.76% TBARS inhibition after 7 days ensilaging, while the combination with EDTA provided 34.21% TBARS inhibition (Table 2). This illustrates that LMW-iron is of minor importance as a pro-oxidant in herring silage, rather heme-bound iron is the dominant pro-oxidant.
In the last trial, (i.e. trial 4), we investigated the possibilities of reusing the antioxidant solution for up to 10 incubation treatments, lowering the solution to co-products ratio from 5:1 to 3:1 or 2:1, shortening the treatment time from 20 min to 30 s, and using a rosemary extract alone in TBARS inhibition (Fig. 4C). There were no significant (p > 0.05) differences in the immediate TBARS-values after reusing the solution 4 and 10 times, compared to using it in one incubation treatment. Similarly, there were no significant (p > 0.05) differences in TBARS developments after 7 days of ensilaging, except that a slight but significant (p < 0.05) increase in TBARS was noticed after reusing the solution for 10 times. The use of a 3:1 solution to co-products ratio resulted in a significantly (p < 0.05) lower TBARS value immediately after incubation, compared to the 2:1 ratio, however, there was no significant (p > 0.05) difference after 7 days of ensilaging. Further, the use of 0.5% rosemary extract and 30 s incubation time were very effective and significantly (p < 0.05) inhibited TBARS development in silage, compared to the non-treated control and to 2% isoascorbic acid. It was also evident that among all the treatments from trial 1–4, both Duralox MANC-213 and rosemary extract, in relative terms, provided the highest TBARS inhibitory effects after 7 days of ensilaging, i.e., inhibitions were 70.95–98.53% and 88.26–93.07%, respectively (Table 2). This confirms that rosemary extract remains stable and active even under the acidic ensilaging conditions, and it can constitute a cheaper option than Duralox MANC-213, also requiring less labelling. The carnosic acid level of the rosemary extract was much higher than in Duralox MANC-213 (57.84 ± 1.54 vs. 1.55 ± 0.06 mg/g), explaining its high activity even in the absence of other antioxidants such as those present in the Duralox MANC-213 mixture (i.e. ascorbic acid, citric acid and tocopherol). Also, its TPC was higher, but to a much smaller extent (102.53 ± 0.04 vs 69.63 ± 2.29 mg gallic acid eq/g), revealing that other phenolics than carnosic acid played a dominant role in Duralox MANC-213.
The last trial also yielded several pieces of information which can make the incubation technology more scalable, provided that the used antioxidant is effective in inhibiting TBARS during ensilaging, e.g. (i) the incubation time could be shortened, (ii) the same solution could be reused for several incubation treatments, and (iii) the ratio between solution and co-products could be lowered. However, it is very important to also consider that just as we can remove a portion of Hb during the incubations, we are also losing some proteolytic enzymes by such treatments, which slightly reduced the protein hydrolysis rate during ensilaging (see supporting information; Fig. 2). We believe though that the advantage of using incubation treatments to minimize lipid oxidation during ensilaging outweighs the slightly reduced rate of protein hydrolysis as the final silage will have a much higher quality. Further, the proteins lost into the incubation solution can be recovered by using techniques such as flocculation combined with flotation or ultrafiltration, as reported elsewhere39,40,41, ensuring as complete use as possible of the herring raw material.
