Docosahexaenoic acid (DHA), a 22:6 ω-3 fatty acid (FA), is abundant in the cell membranes of the human brain, and contributes to the normal development of neural and retinal tissues throughout the human life due to its unique structure and multiple double bonds [1,2]. DHA deficiency in the developing brains of fetuses, newborns, and children is generally linked to neuropathology (e.g., cognitive disorders and anxiety) and disorders related to visual function [3,4]. DHA also plays an important role in maintaining cognitive function and emotional performance during adulthood .
DHA is traditionally obtained by consuming α-linolenic acid (aLNA; 18:3 ω-3)-rich diets and marine foods such as fish and algae. However, the conversion efficiency of α-LNA to DHA in individuals usually cannot meet daily requirements, especially for pregnant women and patients with liver or maple syrup urine diseases [3,6,7]. On the other hand, because the agricultural revolution and food industry have caused a shift in modern diets from marine or α-LNA-rich oils (flaxseed oil, etc.) to ω-6 FA-rich oils (soybean oil, palm olein, and corn oil, etc.) and saturated fats, there is a decreased intake of ω-3 FAs and further decreased concentrations of DHA in human milk [8,9]. Therefore, it has been suggested that preformed DHA from fish oils, algal oils, or high-DHA structured lipids (SLs) be added into foods . Studies have shown that mothers who consumed preformed DHA diets accumulated many times more DHA in their milk, in comparison with the milk of vegans . DHA in vegan milk is primarily synthesized from the ω-3 FAs present in vegetable oils.
In general, DHA is esterified to different positions (sn-1, 2, or 3) in a triacylglycerol (TAG) molecule depending on various food sources. After oral intake, TAGs are hydrolyzed by sn-1,3-specific pancreatic lipase, forming sn-2 monoacylglycerols (MAGs) and free fatty acids (FFAs) . The sn-2 MAGs are then well absorbed through the intestinal mucosa and are preferentially used for the re-synthesis of TAGs or phospholipids (PLs; important components of the brain cell membrane) [13,14]. In contrast, no specific absorption is observed for FFAs hydrolyzed from the sn-1 and sn-3 positions . Therefore, TAGs with DHA located at the sn-2 position are more favorable in terms of absorption and utilization compared with those that have a random DHA distribution . Similarly, sn-2 DHA MAGs showed significantly higher absorption efficiency than other derivatives such as DHA-diacylglycerol (DAG) and DHA-ethyl ester [17,18]. However, most of the current DHA recommendations and supplementations do not pay attention to its positional distribution, and are only focused on the total amount of its daily intake.
Given that the positional arrangement of DHA in TAG and PL structures influences its pharmacological and nutritional benefits for human brain development and maintenance, it is worth providing a background on DHA distribution in common fats and oils, and on the brain benefits provided by high sn-2 DHA lipid diets. The technological procedures of enzymatic syntheses to produce sn-2 DHA-rich SLs and their typical analysis methods are also discussed in this review.
《2. Sn-2 DHA in natural and synthesized lipids》
2. Sn-2 DHA in natural and synthesized lipids
DHA is generally provided by marine fish oils and single-cell oils . There are four main types of DHA lipids from natural sources: sn-2 DHA TAGs, DAGs, MAGs in fish and algal oils, and sn-2 DHA PLs in krill oils and egg yolk (Fig. 1).
Fig. 1. Primary molecular structures of sn-2 DHA lipids. X: ethanolamine, choline, serine, inositol, etc.
The position distribution of DHA on a glycerol skeleton in common fats and oil are summarized in Table 1. Single-cell algal oils (e.g., Schizochytrium sp. oil and Crypthecodinium cohnii oil) contain the highest total DHA levels, ranging from 44.89%–48.20%, followed by various fish oils such as tuna oil, sardine oil, anchovy oil, and salmon oil (9.76%–26.85%). In contrast, the relative percentages of sn-2 DHA were higher in fish oils than in algal oils. Approximately 44.79%–72.99% of the total DHA in fish oil TAGs were esterified at the sn-2 position, while the numerical values were 31.66%–42.09% in algal oil TAGs. This difference might result from the absorption characteristics of sn-2 DHA lipids mentioned above. That is, the DHA synthesized in algal oil is eaten by fish through the food chain; sn-2 DHA MAGs or DAGs are then produced through digestion and absorption, and are further used to resynthesize TAGs, which increases the sn-2 DHA percentages in fish oils to some extent .
Table 1 Position distribution of DHA (%) on a glycerol skeleton in foods and infant formulas.a
a HMF: human milk fat; IFF: infant formula fat; ND: not detectable.
b Relative percentage of DHA at sn-2 position was calculated as [sn-2 DHA percentage / (DHA percentage in TAG × 3)] × 100 , or reported by the literature.
c The data was shown as mol%.
d HMF collected after birth at Days 1–5 was colostrum, at Days 6–15 was transitional, and at more than 15 days was mature.
In particular, the lipids in egg yolk and krill oils are primarily present as PLs (Fig. 1), which are quite different from the lipids in fish and algal oils. Different lipid classes might influence DHA absorption and its concentration in the brain. Diets containing krill oil have been found to increase the DHA levels in rat brain as PLs, and PLs were found to be the major components of both the krill oil and brain cell membranes .
DHA also makes up a small proportion (0.36%–0.70%) of total FAs found in human milk fat (HMF) TAGs, and more than half (52.63%–65.15%) is incorporated at the sn-2 position (Table 1). However, the percentages decreased from colostrum to mature milks (0.56%–0.70%→0.36%–0.44%), while the relative percentages of sn-2 DHA increased from 52.63%–55.71% to 61.39%–65.15%. In addition, DHA levels were found to be progressively lower in nursing mothers who had given birth to twins or had given birth in rapid succession [33,34]. Clinical studies showed that feeding with a-LNA but without DHA over the first 6 months of life cannot sustain normal DHA concentrations in infant brains . The low conversion rates of α-LNA to DHA in newborn and breast-fed infants were also confirmed in this case. It is further concluded from Table 1 that most of the current infant formula fats (IFFs) contain a lower total amount of DHA and sn-2 DHA (the relative percentages were 27.56%–48.17%) in comparison with HMFs. In 11 evaluated IFFs in Spain, only one IFF contained DHA at the sn-2 position . However, 70–80 mg of DHA per day from breast milk is suggested to meet the increasing demand of the rapid growth of a baby’s nervous system . It is therefore suggested that DHA supplementation—especially of sn-2 DHA lipids—in maternal diets may protect infants from deficits in neurodevelopment .
《3. Positive effects of sn-2 DHA on brains 》
3. Positive effects of sn-2 DHA on brains
《3.1. DHA accumulation in brains by utilizing sn-2 DHA lipids》
3.1. DHA accumulation in brains by utilizing sn-2 DHA lipids
Lipids account for approximately 60% of the dry weight of brain tissue . Although DHA is a critical component in maintaining proper brain and nervous functions, its location on a glycerol skeleton exhibits significantly different efficiencies in terms of absorption and utilization. It is much easier for DHA to be absorbed by the intestinal mucosa when it is incorporated at the sn-2 position than when it is randomly distributed at the sn-1,2,3 positions . Further studies have revealed that DHA levels in brain PLs, such as phosphatidylserine and phosphatidylcholine (PC) of newborn rats fed sn-2 DHA diets, were significantly improved compared with those in rats that were fed milk diets (Table 2) . Also, sn-2 lysophosphatidylcholine DHA was preferentially utilized in the rat brains in comparison with unesterified DHA (Table 2) . In addition, large-scale trials have concluded that DHA supplementation through the consumption of large doses of marine oils is safe during pregnancy .
Table 2 Brain benefits of sn-2 DHA lipids
《3.2. DHA supplementation improves brain functions through the gutbrain axis 》
3.2. DHA supplementation improves brain functions through the gutbrain axis
Emotional disorders, which are one of the results of brain function deficits, have been found to be specifically associated with gut microbiota alterations . There has been recent interest in the possible correlation between brain problems (e.g., brain injury, declined cognition, schizophrenia, stroke, anxiety, stress, and depression) and intestinal microflora. The human intestines contain more than 1000 microbiota species with 100 trillion living microorganisms . Bacterial colonization of different species could alter brain functions, and in turn, the central nervous system is speculated to indirectly influence the gut microbial composition. These integrative and bidirectional signaling pathways, which mainly involve the routes of the vagus nerve and spinal pathway, are defined as the gut-brain axis or the brain-gut-microbiota axis (Fig. 2) [41,42].
Fig. 2. The gut-brain axis: Potential multiple bidirectional routes between the brain and the intestinal microflora [41,42]
Previous evidence suggests that gut microbes play an important role in developing therapies for complex brain function disorders. In general, dietary interventions with DHA may have beneficial effects on behavioral and neurophysiological disorders due to alteration of the microbial composition in the intestines [43,44] as seen in Table 3.
Table 3 DHA absorbed through the intestinal mucosa improves brain functions through the gut-brain axis.
EPA: eicosapentaenoic acid; ARA: arachidonic acid.
As shown in Table 3, DHA supplementation for early-life stressed, socially isolated, or aging mice restored and normalized their gut microbiota composition, by increasing the abundance of beneficial species such as Lactobacillus, Bifidobacterium, and Bacteroides, concomitantly decreasing the abundance of Proteobacteria (e.g., Undibacterium) and Cyanobacteria, among others, and subsequently alleviating the mice’s brain-related disorders. In addition, García-Ródenas et al.  has suggested that psychological stress could be reduced by consuming DHA-containing diets through the normalization of gut permeability without the restoration of the intestinal microbiota. This difference indicates that the gutbrain axis includes various bidirectional routes, some of which have not yet been fully elucidated. More studies are required to explain the potential mechanism of the intestinal microbiome on DHA diet-induced effects on the brain. Also, further studies on the impacts of diets with a DHA positional difference (e.g., high sn-2 DHA lipid diets and randomly distributed DHA lipid diets) on the gut-brain axis are necessary.
《4. Enzymatic synthesis of high sn-2 DHA fats and oils 》
4. Enzymatic synthesis of high sn-2 DHA fats and oils
Many infants and pregnant and nursing women consume foods containing only DHA precursors or limited DHA levels . The decreased dietary DHA consumption that results from following a Western diet is responsible for this problem . The production of modified fats and oils with abundant sn-2 DHA using lowpollution and highly efficient techniques such as enzymatic syntheses from saturated fats and the DHA-rich oils listed in Table 1 is encouraged. These processes mainly include the enzymatic reactions of acidolysis, interesterification, ethanolysis, and their combination.
《4.1. Acidolysis reactions》
4.1. Acidolysis reactions
Most of the developed methods to produce high sn-2 DHA SLs focus on the acidolysis of single-cell oils (e.g., DHA single-cell oil from alga Crypthecodinium cohnii (DHASCO)) and FAs (e.g., caprylic acid (C)) in a one-step reaction using sn-1,3 specific lipases or lipases with high activity on DHA.
As shown in the acidolysis reactions in Table 4, optimal reactions are generally carried out with substrate mole ratios of 1:3–1:18 (oils to FFAs) at mild temperatures of 30–55 C with 4%–15% enzymes for dozens of hours [51–54]. The sn-2 DHA levels vary significantly based on the enzyme species . In some cases, the lipases, such as Pseudomonas sp. KWI-56 lipase, showed nonregiospecificity but were active toward DHA and docosapentaenoic acid, and may also cleave the DHA at the sn-2 position, resulting in acyl migration to some extent . This side reaction might easily occur in the presence of caprylic acid and different lipases . It is suggested that possible alternative or better lipases be developed in order to minimize acyl migration. In addition, recovery of the target SLs from these reaction products is usually complicated. Usually, for a small-scale reaction, FFAs are removed by neutralization with alkaline solution, followed by the extraction of TAGs with hexane; the solvent is then further evaporated to obtain the final SLs.
Table 4 Enzymatic syntheses of high sn-2 DHA SLs.
DPA: docosapentaenoic acid; C: caprylic acid.
a Relative percentage of DHA at the sn-2 position was calculated as [sn-2 DHA percentage / (DHA percentage in TAG × 3)] × 100 .
The other typical method to prepare sn-2 DHA SLs is to hydrolyze single-cell oils or marine fish oils to prepare DHA, followed by esterification with TAGs (Table 4). In this context, DHA is first released from the marine oils by saponification using potassium hydroxide and acidification using hydrochloric acid in the presence of antioxidants (e.g., butylated hydroxytoluene). Acidolysis of the prepared DHA and other oils is then conducted at substrate mole ratios of 1:5–1:18 (oils to DHA) and with an enzyme load of 10%, and the reaction is kept at 60–65℃ for around 24 h [25,31,55]. For large-scale and industrial reactions, the extra FFAs are commonly removed through short-path distillation.
《4.2. Interesterification reactions》
4.2. Interesterification reactions
Interesterification between DHA-rich oils/ethyl ester and FA ethyl ester is another method to provide targeted SLs (Table 4). The reactions require strict enzyme selection due to their positional specificities and the steric hindrance of DHA . For example, in a two-step reaction, unspecific DHA-rich oil was first prepared from a nonselective reaction of DHA-ethyl ester and tricapryloylglycerol using Alcaligenes sp. lipase (50℃, 90 h), followed by a sn-1,3 regioselective interesterification of the unspecific DHArich oil and ethyl caprylate using Novozym 435 to produce sn-1,3- dicapryloyl-2-docosahexaenoylglycerol (40℃, 40 h) . Both reactions were carried out in a nitrogen atmosphere to avoid oxidation, and extra esters and tricapryloylglycerol were removed by molecular distillation.
《4.3. From sn-2 DHA MAG to sn-2 DHA lipids》
4.3. From sn-2 DHA MAG to sn-2 DHA lipids
Another typical strategy to obtain sn-2 DHA-rich lipids is to prepare sn-2 DHA MAG from marine oils, followed by the incorporation of needed FAs at the sn-1,3 positions of the MAG (Fig. 3 and Table 4).
To achieve this technical route, preparation of sn-2 DHA MAG from oils is a key step due to the oxidation problems of DHA, acyl migration during enzymatic catalysis, and the cost . Conventional methods were carried out in an ethanol system with enzymes such as Novozym 435, which showed sn-1,3 regiospecificity in the presence of ethanol [59,60]. Recent research has reported a highly efficient approach to produce MAG enriched with x-3 polyunsaturated fatty acids (PUFAs) at the sn-2 position using Candida antarctica lipase A in a more economical way . In similar cases, Candida antarctica lipase A effectively concentrated the sn-2 DHA of anchovy oil from 20.88% in oil to 65.69% at sn-2 MAGs via catalytic reaction at low temperature (35 C) for 12 h; the sn-2 DHA value in microalgae oil was increased from 3.24% to 22.20% in the same way . This research demonstrated that Candida antarctica lipase A exhibits non-regiospecific and non-ω-3 PUFA preference in an ethanol system, and can thus selectively cleave non-target fatty acids and further keep the ω-3 PUFAs such as DHA on the glycerol backbone to form DHA-rich MAGs [21,65,66].
For purification, DHA-containing byproducts such as FFAs and their ethyl esters can be removed by short-path or molecular distillation for further re-utilization . The advantage of this technique is its flexibility in manufacturing different fats and oils such as shortenings, margarines, spreads, IFFs, and bakery and confectionary fats using the sn-2 DHA MAG.
《5. Analytical methods for sn-2 DHA》
5. Analytical methods for sn-2 DHA
Regiospecific analysis of FAs in TAG molecules is generally conducted on a gas chromatograph equipped with a flame ionization detector. In brief, TAGs are first hydrolyzed by sn-1,3-specific lipases to form MAGs, followed by the isolation of sn-2 MAG using thin-layer chromatography and its conversion to fatty acid methyl esters for further analysis . Pancreatic lipase is a widely used lipase, which has been well confirmed through the determination of the sn-2 FA composition of many fats and oils. However, it should be noted that pancreatic lipase exhibits limited ability to hydrolyze all FAs, particularly PUFAs from marine oils . Its ability for selective hydrolysis depends on the FA species and the location of the double bonds . In contrast, Candida antarctica lipase B (Novozym 435 or Lipozyme 435) is suggested to be a better hydrolytic enzyme for this purpose [70,71]. Although Lipozyme 435 is a non-regioselective lipase in many cases, it behaves as sn-1,3-specific in the presence of excess ethanol . Table 5 shows the PUFA compositions of fish oils as detected by the Novozym 435 method and the pancreatic lipase method. Novozym 435 can release PUFAs from fish oils at different rates based on the degree of chain length and unsaturation. For example, eicosapentaenoic acid (EPA) levels detected using the pancreatic lipase method (7.5%–10.8%) were higher than those determined using the Novozym 435 method (6.8%–9.0%), while the contents of DHA exhibited the opposite trends . That is, Novozym 435 shows exclusive selectivity for DHA compared with pancreatic lipase.
Table 5 Sn-2 PUFA compositions of fish oils determined by the pancreatic lipase method and the Novozym 435 method .
In general, the Novozym 435 method needs strict hydrolysis conditions, such as ethanol-to-oil ratio, reaction time, and temperature, to completely release the sn-1,3 FAs from TAGs; otherwise, the hydrolysis reaction might result in lower results compared with C-13 nuclear magnetic resonance (13C NMR) or predicted values. In a cod liver oil test, the result for sn-2 DHA by the Novozym 435 method was 69.4%, which was lower than that measured by 13C NMR (72.5%); however, for analysis of tuna oil, the sn-2 DHA results were similar, at 53.1% for the Novozym 435 method and 52.0% for 13C NMR .
Marine fish and algal oils are typical DHA sources with about half of their FA incorporated at the sn-2 position. Their unique structure makes it easier for DHA to be absorbed by the intestinal mucosa and to be used for the re-synthesis of TAGs or PLs in vivo, in comparison with molecules that have DHA located at the sn-1,3 positions. sn-2 DHA lipids, therefore, play important roles in the development of brain functions and in the mitigation of brain deficits such as anxiety, stress, declined cognition, schizophrenia, and stroke. A focus on the gut-brain axis is the most effective strategy to understand the beneficial effects of DHA supplementation on brain functions. It is suggested that brain problems could be alleviated by restoring and normalizing the gut microbial composition through DHA intervention. However, the multiple bidirectional routes of the gut-brain axis are not yet fully understood or explained. Further research is required on the impacts of dietary sn-2 DHA lipid supplementation on the gut microbiota and brain functions.
DHA accumulates in the human brain at a rapid rate from gestation to age 2. However, although the amount of DHA in HMFs decreases to a low level 15 days after birth, the relative percentages of sn-2 DHA show increased trends, indicating the importance of sn-2 DHA in the brain development of infants and children. Therefore, it is suggested that preformed sn-2 DHA SLs containing sn-2 DHA be included in maternal diets; this could be done by preparing sn-2 DHA MAG from DHA-rich oils, and then incorporating selected FAs at the sn-1,3 positions of the MAG. For further study, it is suggested that novel lipases with high activity at the sn-1,3 positions or with a non-ω-3 PUFA preference be developed, together with mild reaction conditions and purification procedures to make the synthesis techniques and products more efficient and economical.
The work was supported by the Chinese Scholarship Council (201706790068) and the Free Exploration Founded Project of the State Key Laboratory of Food Science and Technology at Jiangnan University (SKLF-ZZA-201705). It was also supported in part by Food Science Research, University of Georgia.
《Compliance with ethics guidelines 》
Compliance with ethics guidelines
Jun Jin, Qingzhe Jin, Xingguo Wang, and Casimir C. Akoh declare that they have no conflict of interest or financial conflicts to disclose.