Amino acid nitrogen and carbon isotope data: Potential and implications for ecological studies

Abstract Explaining food web dynamics, stability, and functioning depend substantially on understanding of feeding relations within a community. Bulk stable isotope ratios (SIRs) in natural abundance are well‐established tools to express direct and indirect feeding relations as continuous variables across time and space. Along with bulk SIRs, the SIRs of individual amino acids (AAs) are now emerging as a promising and complementary method to characterize the flow and transformation of resources across a diversity of organisms, from microbial domains to macroscopic consumers. This significant AA‐SIR capacity is based on empirical evidence that a consumer's SIR, specific to an individual AA, reflects its diet SIR coupled with a certain degree of isotopic differences between the consumer and its diet. However, many empirical ecologists are still unfamiliar with the scope of applicability and the interpretative power of AA‐SIR. To fill these knowledge gaps, we here describe a comprehensive approach to both carbon and nitrogen AA‐SIR assessment focusing on two key topics: pattern in AA‐isotope composition across spatial and temporal scales, and a certain variability of AA‐specific isotope differences between the diet and the consumer. On this basis we review the versatile applicability of AA‐SIR to improve our understanding of physiological processes as well as food web functioning, allowing us to reconstruct dominant basal dietary sources and trace their trophic transfers at the specimen and community levels. Given the insightful and opportunities of AA‐SIR, we suggest future applications for the dual use of carbon and nitrogen AA‐SIR to study more realistic food web structures and robust consumer niches, which are often very difficult to explain in nature.


| INTRODUC TI ON
Ecosystems and life on Earth are subject to global environmental stressors resulting from human activities, including overfishing, invasive species, species loss, and habitat degradation. These natural and anthropogenic environmental stressors affect the number and portions of basal components, such as living autotrophs (to support "green" food webs) and decomposed organic matter (OM) (to support "brown" food webs or "microbial loops"; Quévreux et al., 2021;Wolkovich et al., 2014;Zou et al., 2016). These ongoing global environmental changes have also led to ecosystem-wide impacts, including trophic connections, numbers of trophic levels (Lynam et al., 2017;Maureaud et al., 2017), the consumer population abundance, and omnivory levels (Gibert, 2019). However, it is not simple to say how a dynamic climate change is related to ecosystem-wide impacts in the face of complex temporal and environmental variations, inherently combined with biogeochemical and ecological processing McMeans et al., 2015). A more elaborate methodology could enable us to assess environmental risks and vulnerabilities (e.g., natural and/or anthropogenic impacts) and to develop conservation management plans.
Naturally occurring stable isotope variations, particularly those of carbon (C) and nitrogen (N), are powerful tools for describing trophic interactions, food web structures, and biogeochemical processes. This is due to (1) the fixed isotope signals in the sources, which are isotopically distinct, and (2) the predictable isotope fractionation factor at the organism and environment levels. The isotope fractionation factor is the factor by which the abundance ratio of two isotopes will change during a physiochemical process. The isotope fractionation factors occur due to reaction rate differences between lighter (e.g., 12 C and 14 N) and heavier isotopes (e.g., 13 C and 15 N) during physicochemical and biogeochemical process, which leaves the consumer with a higher ratio of heavier than light isotopes in their diet (or source element) (i.e., Fry, 2006;Fry & Carter, 2019;Martínez del Rio et al., 2009). It has long been established that consumers become more enriched with heavy isotopes than with light isotopes relative to their dietary resources; for C, the isotope difference is small at 0-1‰, and for N it is significantly higher at 2-4‰ (DeNiro & Epstein, 1978, 1981Fry, 2006;Minagawa & Wada, 1984). Thus, stable isotope ratios (SIR) can provide the evidence related to the trophic connections as well as trophic hierarchy between consumer and potential (or actual) dietary resource. The applications of SIR are extended to estimate nutritional or basal resources (also referred to as food web end-members or baselines) as well as the varying reliance on those different resources quantitatively for consumers which support local food webs (Middelburg, 2014;Newsome et al., 2007).
Many studies assert that consumers' isotope values (i.e., the whole-tissue parts) are confounded by variable temporal and spatial traits (reviewed in Shipley & Matich, 2020), inducing isotope variations at the base of local food webs. For instance, stable C isotope ratios ( 13 C: 12 C, represented by δ 13 C) are useful to separate photosynthetic pathways (C 3 , C 4 , and CAM plants), inorganic sources (e.g., CO 2 and bicarbonate), but also can be affected by environmental change  as well as regional variances (e.g., marine vs. terrestrial, benthic vs. pelagic algae, nearshore vs. offshore productivity, and high vs. low latitude; Glibert et al., 2019, references therein). Similarly, stable N isotope ratios ( 15 N: 14 N, represented by δ 15 N) can differ among sources such as N 2 , NH 4 + , NO 3 − , and amino acids (AAs) as well as other environmental characteristics of locations or sites (e.g., precipitation, altitude, temperature, and salinity; Hyodo, 2015;Kelly, 2000). It indicates that C and N bulk isotope ratios would be variable in consumers and producers, as do baseline isotope ratios vary over time and regions (Drobnitch et al., 2018;Zhang et al., 2021). Thus, it is notably important for assessing correct trophic interactions in local food webs to distinguish effect of dietary sources and/or environmental differences on consumer isotope ratios (Blattmann & Ishikawa, 2020;Middelburg, 2014).
Ecologists have considered amino acid (AA)-SIR as a complementary and straightforward approach of bulk SIR analysis (Table 1). AAs, the building blocks of protein as essential macronutrients, are the persistent and abundant organic compounds (Ruess & Müller-Navarra, 2019), and are found in diverse sample types from biological samples as well as environmental samples. Thanks to advances in continuousflow isotope analyses of single compounds in the late 1990s, applications of AA δ 15 N and AA δ 13 C have become more widespread and common over the last decades ( Figure 1). AA-SIR analysis has been extensively applied to elucidate trophic connections from field as well as lab-simulated experiments, designed to examine simplified food chains (e.g., consumer-direct diet interactions) and multi-trophic level (e.g., producer-herbivore-carnivore). There have been recent reviews of AA-SIR applications in ecological and geochemical studies  (Close, 2019;Ohkouchi et al., 2017). Table 2 presents AA lists grouped by patterns of AA-SIR variances, as follows: 1. The degree of AA ΔN is generally high or changeless during trophic transfer. For instance, AAs such as glutamic acid/glutamine (Glx) and alanine (Ala), proline (Pro), isoleucine (Ile), leucine (Leu), valine (Val), and aspartic acid/asparagine (Asx) usually become 15 N enriched during trophic transfer (Table 2).
2. The degree of AA ΔC is highly variable or changeless. The variable pattern from (+) to (−) ΔC is commonly detected in Ala, Gly, Ser, and Glx, belonging to non-essential AAs (NEAAs) that consumers can biosynthesize de novo from diverse dietary non-protein biomolecules (carbohydrates and fats). However, other AA δ 13 C in consumers generally resembles that of their food source with relatively modest alterations during trophic transfer (i.e., ΔC ≈ 0-1‰). The less fractionating patterning is observed mostly in essential AAs (EAAs) encompassing Phe, Lys, methionine (Met), Leu, Ile, threonine (Thr), and Val ( Table 2) that consumers cannot biosynthesize de novo and should obtain from dietary resources. Figure 2, the overall magnitude of ΔN and ΔC variances in AAs is caused by net results of complex physiological responses. The specified metabolic processes to control N and C variances in AAs are different. For instance, the 15 N isotope enrichment in consumer AA relative to dietary AA would be connected to active biochemical reactions such as deamination or transamination during N metabolism (Chikaraishi et al., 2007(Chikaraishi et al., , 2009McMahon & McCarthy, 2016;O'Connell, 2017;Ohkouchi et al., 2015Ohkouchi et al., , 2017, while 13 C enrichment in consumer could be associated with decarboxylation during C metabolism (Fry & Carter, 2019). Deamination and transamination are N-involved reactions that convert an amine group (NH 2 ) into ammonia (NH 3 + ) and transfer amine group to keto acids for synthesizing other AAs, respectively. On the other hand, decarboxylation is C-involved reaction that removes a carboxyl group from a molecule (e.g., AA) and releases carbon dioxide (CO 2 ). The N and C involved reactions on AAs, TA B L E 1 Comparing strategies of traditional bulk and amino acids isotope approaches for diet tracing and food web reconstruction

Major themes
Bulk δ 13 C and δ 15 N analysis Amino acid δ 15 N analysis Amino acid δ 13 C analysis

Diet resolution
If used to two isotopes, max.

items
Assumed limited number (e.g., terrestrial plant or aquatic algae)

3-6 phylogenetically separated groups in autotrophs
like nearly all physiological reactions, depend selectively on the lighter stable isotope via kinetic fractionation. That is, as the lighter 14 N or 12 C-AAs are preferentially used (or removed) during the metabolic processing, the heavier 15 N or 13 C-AAs would remain more abundantly in consumer body. This is particularly understandable for explaining (highly fractionating) trophic AA δ 15 N patterns listed in  However, the heavier isotope enrichment in several AAs (grouped into trophic AAs and NEAAs) is not always detected, and rather the depletion is frequently observed via showing (-) responses (e.g., 38 m , 39 n , 40 n , 41 n , 51 l , and 52 l in Figure 3b). Generally, we assume that (-) Δ responses might relate to biosynthetic process of AAs, as the lighter 14 N or 12 C substrates are selectively used to make products (e.g., AAs) than the heavier isotope 15 N or 13 C. If AA biosynthesis occurs frequently and is stimulated than anabolic response, AA-SIR becomes lower than its substrate. In terms of C substrates, intermediates of glycolysis and citric acid cycle, routed from carbohydrates, lipids, and proteins, can be used to synthesize NEAA carbon skeletons. Accordingly, the variability in NEAA ΔC is likely associated with the complex interplay between the relative proportions of dietary macronutrients and the metabolic (catabolic and anabolic) demands within consumer body.
The C isotopic composition of the many AAs in consumers generally resembles that of their food source ( Figure 3b, Table   S1), with relatively modest alterations during trophic transfer (i.e., ΔC ≈ 0-1‰). The less fractionating patterning is observed mostly in essential AAs (EAAs) ( Table 1). This is commonly found in F I G U R E 1 Trend of numbers of trophic ecology-based publications that have employed compound-specific stable isotope analysis till 2020, which include all peer-reviewed publications in the Web of Science database published before this date. We conducted this review by first searching for amino acid isotope and narrowed the search terms to food web, resource, diet, and consumer. We then removed all publications that were listed twice, which resulted in a total of 267 references Aspartic acid/Asparagine a Tyrosine and glycine are conditionally essential AA (Reeds, 2000). Threonine offset was not close to 0 (e.g., McMahon et al., 2010).

TA B L E 2
Classification of common amino acids (AA) according to their isotope offset patterns between diet and consumers during trophic transfer, which can be triggered by AA nutritional essentiality and metabolism in consumers. Classifications based on data from zooplankton to invertebrates, fish, and mammal diverse consumers under field sampling and diet-controlled experiments, such as with sea slugs-sponges (Takizawa et al., 2020), lady bug beetles-aphids (Whiteman et al., 2019), fish-seaweeds (Wang et al., 2019), and penguin-herring (McMahon, Polito, et al., 2015).
Since EAA δ 13 C values remain largely invariable during trophic transfer, a food chain encompassing the first consumer (herbivore) to apex predator would have similar EAA δ 13 C values. For instance, in a food chain experiment with the green alga Chlorella-copepod Calanus (first consumer) and the anchovy Engraulis (second consumer), EAA δ 13 C values are unchanged in the herbivorous copepod and alga, the carnivorous anchovy (Liu et al., 2018). Similarly, Phe ΔN of source AAs which belongs to EAA group is changeless (e.g., <1.0‰ in Figure 3) during trophic transfer, although it is still unclear why other source AAs (e.g., Gly and Ser belonging to NEAA). Overall, the fractionation magnitude of the EAAs and the source AAs is not significantly affected by the nutritional content

| Is AA-SIR variable across spatiotemporal scales?
SIR in compound level is more tightly associated with the exact biomolecules (e.g., AAs), whereas bulk SIR of whole tissue or total protein depends on inter-molecular isotope differences and the mo- Nonetheless, organisms that share similar biosynthetic/metabolic pathways tend to exhibit analogous patterns of isotope fractionation (e.g., Hayes, 2001;Larsen et al., 2009;Scott et al., 2006).
It means that overall patterns of AA-SIR variability are assumed to be highly consistent, although actual AA-SIR values themselves are significantly different among space and time ( Figure 4). The robust AA-SIR patterns have been presented by placing the isotope variables around the centerline of the average of multiple AAs.
The multiple AAs involved are generally the group of essential AAs such as Thr, Ile, Leu, Phe, Val, and Lys. In a case of Larsen et al.
Moreover, the mean-centering procedure there showed no clear difference between the cultured microalgal groups (consisting of blue-green algae and diatoms) and the field-collected microalgal complex (Figure 4c), although their actual AA δ 13 C variables showed dissimilar patterns (> 3‰ difference, on average). In this F I G U R E 2 A simplified overview of AA isotope variation in consumer relative to diets. The range of ΔX variations, from (+) to (−), is related to the nutritional essentiality as well as the net balance between losses and biosynthesis via diverse physiochemical processes. (De-) Transamination and decarboxylation are one of the most common processes related to protein metabolism, which increase δ 15 N and δ 13 C variables of AA, respectively. Biosynthesis process of non-essential AA originated from non-protein macronutrients is known to decrease isotope value in a view of substrate-product relation (Fry & Carter, 2019;McMahon & McCarthy, 2016) regard, AA-SIR patterns detected in the lab-cultured and greenhouse samples can be referred to interpret basal sources in local food webs in nature.
Similar to the AA δ 13 C variability, variances of AA δ 15 N in microalgae are also observed in controlled culture conditions, i.e., growth stage, light regime, herbivory pressure, and N-and P-limited within a range similar to that of field-collected macroalgae as well as cyanobacteria and cultured microalgae, i.e., 3.4 ± 0.9‰ (Chikaraishi et al., 2009. It implies that AA-SIR approaches can help us define an isotopic reference of trophic bases and validate basal resources over times and regions in natural environments.

F I G U R E 3
Mean isotope offset variability of AA-N (A) and AA-C (B) between diverse consumers and diet obtained from controlled feeding studies. Numbers in X-axis mean consumer-diet combination pairing (summarized in Table S1). Bolitic font means pairing of consumer and nutritionally poor diet. Diets are grouped to animal-based protein (marked as closed symbol) vs. producer (including plant, microalgae, fungi, and bacteria)-based protein (   respectively, and the producer-specific isotope offset of trophic and source AAs from the isotope baseline information as fixed factors, and is followed by:

| Estimating consumer TP
The δ 15 N trophic AA and δ 15 N source AA represent the nitrogen isotope ratio of trophic AA and source AA analyzed in the consumer; the TDF trophic AA and TDF source AA represent the TDF of each AA; α represents the trophic level of the isotope baseline (if a plant is selected, α = 1; if a primary consumer is involved, α = 2); and the β value indicates the δ 15 N difference between the trophic AA and the source AA in the trophic base. Generally, Glx (the canonical trophic AA) and Phe (the canonical source AA) are included in the TP AA estimate.
Commonly, a difference between TDF Glx and TDF Phe of 7.6‰ is not variable and generally applied as a constant (Bradley et al., 2015;Chikaraishi et al., 2009;McMahon & McCarthy, 2016;. The TDF was found consistently in recent meta-analyses of wild-caught marine consumers (Blanke et al., 2017;, i.e., 6.6 ± 1.7‰, and consumers under controlled feeding ex- Functioning of normalization to average of multiple essential AAs in autotroph Thalassiosira weissflogii under diverse culture treatments from Larsen et al. (2015): (a) Actual data and (b) after isotope value normalization of each AA to the average of five essential AAs, and (c) comparing the variation in essential AA δ 13 C between mean single algal complex and algal complex collected in field (data adopted from Larsen et al., 2013) value based on Glx-Phe is also known as fixed factor (Chikaraishi et al., 2009;Ramirez et al., 2021) Determining the TP AA of consumer is frequently generated by Glx and Phe pair, trophic AA and source AA, respectively, and recent studies examined Ala or Pro relative to Phe, and the mean of multiple trophic AA δ 15 N variables relative to the mean of multiple source AA δ 15 N variables. It is very important to obtain proper TDF and β parameters, fitted to AA pair, from lab-/field-collected or literature data. Such diverse trials produce more realistic TP estimates matched with feeding habits of the consumer widely known (Kjeldgaard et al., 2021;Ledesma et al., 2020;Matthews et al., 2020;Ohkouchi et al., 2017;Ramirez et al., 2021;Troina et al., 2021).
For example, in point of δ 15 N Glx , protists (heterotrophic microconsumers) show quite similar to phytoplankton (microalga Dunaliella) (Decima & Landry, 2020;Decima et al., 2017;Gutierrez-Rodriguez et al., 2014). This is an unexpected point from their study design that the TP Glx-Phe of the protists can be like its diet source. In contrast to δ 15 N Glx , TP Ala-Phe in the protists is in the middle between phytoplankton (microalga Dunaliella) and metazooplankton (copepods and krill), which is observed consistently in the lab (Decima et al., 2017) and the field (Decima & Landry, 2020), which is close to the realistic trophic relation. It implies that Ala δ 15 N works out effectively in reconstructing the protist involved food webs. Furthermore, Bode et al. (2021) found that the cascading effects from the microplanktonic trophic step enrichment are transferred to a higher trophic level (predator fish), and it makes TP Ala-Phe 0.5 to 0.6 unit higher than TP Glx-Phe even for micronecktonic fish. Although the physiological background of the invariable Glx and the variable Ala as trophic AAs in such micro-sized consumers is still unknown, the comparison of TP Ala-Phe to the original TP Glx-Phe estimate could reveal the role of microzooplankton protists in marine food web functioning, which is still poorly understood.
F I G U R E 5 (a) A scheme for describing trophic position using AA-N variables and (b) example of AA-C isotope patterns, focused on essential AAs, from cultured species of marine microalgae, terrestrial plants, and heterotrophic bacteria in different nutrient sources, as data were compiled from Larsen et al. (2013). Mean AA isotope values with standard deviation across multiple experimental species are presented. These unique grouping patterns can be applicable as end-member fingerprints in stable isotope mixing models. Red, blue, and green points within a big triangle indicate consumers which feed mostly on bacteria, microalgae, and terrestrial plants as diets, respectively, while grey point in the middle of triangle is expected to feed mixing of the three sources. The dashed lines help get approximate diet proportion composed of different basal members derived from each organic C resources 3.2 | Identifying multiple producers using AA isotope fingerprints AA-SIR has been applied for identifying trophic bases (nutritional origins) in local habitats (Table 1). This is due to several empirical evidences that naturally occurring AA δ 13 C variables are unique to distinguish phylogenetically different producers (Larsen et al., 2009(Larsen et al., , 2013Scott et al., 2006) (Figure 5a), which is known as "stable isotope fingerprints" (Larsen et al., 2009(Larsen et al., , 2013. The stable isotope fingerprints based on AAs are related to phylogenetically distinctive characters of producers, such as how to fix inorganic C substrate (e.g., HCO 3 − and CO 2 ) and how to synthesize C skeletons of AAs (Brett et al., 2017;Macko et al., 1987;Scott et al., 2006) in diverse routes (e.g., glycolysis, the citric acid cycle, and the pentose phosphate pathway). Currently, 13 C EAA patterning can separate basal C sources from terrestrial plants, fungi, and heterotrophic bacteria Pollierer et al., 2019;Scott et al., 2006), between seagrass and bacteria (Larsen et al., 2013), and between subtidal kelp (brown macroalga Laminaria) and phytoplankton/ephemeral green macroalga (Ulva) (Smith et al., 2018). The EAA fingerprint pattern among producer groups can be conceptualized and visualized using dimensionality reduction techniques such as principal component analysis (PCA) and linear discriminant analysis (LDA).
Please note that the unique EAA isotope fingerprints can be utilized to estimate their relative contribution in an environment.
For instance, if a food wed is composed of three producers (e.g., bacteria, microalgae, and terrestrial plants, which EAA isotope fingerprints are distinct), represented by endpoints of a big triangle, the relative contribution is assessed by stable isotope mixing model. The unique isotope fingerprints in trophic bases are passed to consumers, as actual variables of multiple EAA δ 13 C from nutritional sources are transferred on to consumers. In this regard, the pattern of primary producers in trophic base (TP = 1) is expected to be kept over higher TPs (TP > 1), if there is no extra input of dominant producers. Moreover, the stable isotope fingerprints can also be available with mean-centered EAA δ 13 C compositions (i.e., difference of AA δ 13 C to average of multiple EAA δ 13 C) to reduce uncertainty related to temporal and regional effect than actual isotope variable (Whiteman et al., 2019). In these regards, the EAA isotope pattern has been a crucial tool to diagnose major basal sources in local food webs (Larsen et al., 2009(Larsen et al., , 2013(Larsen et al., , 2020Liu et al., 2018).

| Identifying nutritional sources from non-AA components
NEAA δ 13 C variability can reflect physiological responses of consumer to changes in diets or nutrient content in diet source. This is because (1)  plants (−12‰), which is not directly associated with Ala δ 13 C in diets (Jim et al., 2006;McMahon et al., 2010;Newsome et al., 2011Newsome et al., , 2014O'Brien et al., 2002). Even at high protein levels, controlled feeding experiments with varying diet proportions of isotopically distinct protein/carbohydrates (Jim et al., 2006;Newsome et al., 2011) and protein/lipids (Newsome et al., 2014) showed that the NEAA δ 13 C variability is closely associated with non-protein dietary sources.
Similarly, feeding experiments with rats fed diets with varying pro- true for animals that often undergo seasonal or spatial changes in macronutrient availability (Magozzi et al., 2021), and also for animals that consume lipid-rich prey, which contributes ~25-50% of the C source of marine animals (Newsome et al., 2014), or carbohydraterich fruit, which contributes ~50-90% of the C source for migratory birds (Gomez et al., 2018).

| INTERPRE TATI ON OF A A-S IR FOR DEFINING TROPHI C TR AN S FER B E T WEEN G REEN AND B ROWN FOOD WEBS
There is considerable interest to reveal ecological functions of green-based source (mainly composed of non-detrital OM by photoautotrophs) and brown-based source (mostly detrital OM) for consumer community. However, current reviews on AA-SIR applicability scope focus more to describe the trophic ecology for small

| Defining trophic interactions of consumers with green-as well as brown-based resource
Since consumer AA δ 15 N is tightly associated with diet (or basal sources) AA δ 15 N (Topic 1 in Table 3), the consumer SIR can reveal the change in diet (or resource) shifts (Figure 5a). This application is particularly useful when the isotope variables between the basal sources are significantly distinguishable (Phillips et al., 2014). Several studies tested whether AA δ 15 N variables themselves (Ishikawa, 2018;Naito et al., 2016)  Although dead OM (i.e., detritus combined with living bacteria and/or fungi) is more heterogeneous and dynamic than living OM, AA 15 N analysis is useful to determine the trophic links of "brown" detritus-based food webs (Topic 2 in Table 3). regarded as the brown diets, respectively, were offered to herbivores and carnivores for 1 month. In perspective of TP Glx-Phe estimate, the brown (decomposed) foods were about half a TP up compared to fresh diets, which is observed consistently for both herbivores and carnivores; this suggests that the characteristics of detritus are subsequently transferred to increase the TP of detritus-feeding consumers (i.e., nutritional absorption from the detritus into consumer).
Thus, TP AA can provide a perspective for reconstructing the complex food web structure derived from the mixture of brown and green sources for macrosized consumers in natural environments.
AA δ 15 N approach is powerful to detect the transformation of detrital OMs dominated by microbes (bacteria) of marine environments in context of biogeochemical processing (Topic 3 in Table 3). Colored arrows indicate different trophic transfer from sources A, B, and C, respectively, from trophic base members. Straight lines represent trophic relations between consumer (TP belongs to α) and source (TP becomes α-1), whereas curved lines indicate trophic links with multi-hierarchy sources (TP could be either α-1 or even smaller) TA B L E 3 Use of AA-C and N isotope application and its relevant research topics at individual, species, community, and ecosystem level  difference between the two OM fractions, especially in summer. In a microscale, ultra-filtered particulate OM (a large particle fraction) in ocean water column contains slightly low AA δ 15 N relative to dissolved OM fractions (Yamaguchi & McCarthy, 2018).
Not only AA-SIR value itself but also a proxy (the ∑V index, called total "heterotrophic AA resynthesis") can express the magnitude of microbial (bacterial) activity, which describes the δ 15 N variability of multiple trophic AAs from the average of δ 15 N values of multiple trophic AAs (e.g., Leu, Val, Ile, Ala, Pro, and Glx), as originally proposed by McCarthy et al. (2007). The defined ranges of ∑V values in phytoplankton and metazoans are 0-1 and 1-2, respectively, whereas ∑V values >2 indicate detrital OM (brown resource) with an increasing proportion of degrading algae and zooplankton (Arthur et al., 2014;Hannides et al., 2020;McCarthy et al., 2007). Notably, ∑V values are greatly increased in smaller dissolved OM compared to particulate OM (Yamaguchi & McCarthy, 2018). As food web base components, detrital (non-living) OM and its small molecular size can be linked to its microbial reactivity and degradation rate in middledepth water; these might be more active and fresher near the ocean surface (Doherty et al., 2021;Stücheli et al., 2021).
With developing analytical method (i.e., coupling of polar and nonpolar column in gas chromatography system), inclusion of the δ 15 N of multi-source AAs, Phe, and Met could be an option in the isotope mixing model . This approach can be effective because (1) inclusion of multiple isotope variables can handle with distinguishing more diverse sources (e.g., Choi et al., 2017;Ishikawa et al., 2014) and (2) the source AAs provide a more direct proxy indicating the δ 15 N base relative to the local habitats (e.g., offshore vs. nearshore and/or benthic vs. pelagic) rather than the consumer-specific metabolic effect on isotope variability (Quillfeldt & Masello, 2020).

| Quantifying the relative importance of green vs. brown sources in consumer diets
As actual variables of multiple EAA δ 13 C from nutritional sources are transferred on to consumers (Figure 6b), EAA isotope fingerprints are also passed on to consumers. In this aspect, EAA isotope pattern in consumer can determine major basal sources via conducting stable isotope mixing models (Larsen et al., 2009(Larsen et al., , 2013(Larsen et al., , 2020Liu et al., 2018). Thus, the EAA fingerprint approach is regarded as a reliable tool to diagnose the basal EAA sources and its relative proportion in food webs, particularly even when direct sampling of trophic base components is difficult or unknown.
EAA isotope fingerprints of higher TP consumers can elucidate the basal C origins in green sources (autotroph such as microalgae, cyanobacteria, and plants-relevant OM) vs. brown sources (nonautotroph such as bacteria-and fungi-originated OM) for sessile (Ferrier-Pagès et al., 2021;Wall et al., 2021) and mobile macrobenthos (Smith et al., 2018, and fish as tertiary (or upper) TP consumers in aquatic ecosystems (Larsen et al., 2020;Thorp & Bowes, 2017). Recently, the applicability of EAA δ 13 C fingerprint has been extended to distinguish two brown groups, heterotrophic bacteria and fungi, which both are well-known potential decomposers in underground soil and sediment. Since EAA δ 13 C fingerprints in heterotrophic bacteria are notably different from those in fungi (Larsen et al., 2013;Scott et al., 2006), they can determine major C sources for detritivores such as earthworms and microarthropods (Larsen  (Bradley et al., 2015;Chikaraishi et al., 2009Chikaraishi et al., , 2014Steffan et al., 2017). It can further vary in the specific consumers in aquatic ecosystems (Figure 3a).
The trophic AAs Δ 15 N values can be affected by diet quality (e.g., protein insufficient foods, marked as 2 b , 4 d , and 18 k in Favella fed with alga Heterocapsa (Decima et al., 2017) showed small variances in Glx (<1‰) but large variability in Ala of trophic AAs (> 8‰). In extreme cases of prolonged lack of food (for months), free-ranging mammals (elephant seals) showed higher AA δ 15 N values compared to successful hunting periods, when their metabolic response was to transform AAs into energy sources during the fasting period (Lübcker et al., 2020).
Although relatively high variance of Glx and/or Ala Δ 15 N is commonly reported in aquatic consumers, this is not usually the case in terrestrial ecosystems (Figure 3a). For instance, Steffan  (Figure 3a). The depletion of AA 15 N (particularly, Thr as one of the source AAs in Table 2) is also reported in omnivore rat which feeds on high protein diets compared to one with low protein diets (Fuller & Petzke, 2017), but more studies are needed to understand how Thr δ 15 N variations are affected by diet quality (Whiteman et al., 2021).
Such unexpected fractionation patterning in terrestrial consumers (detritivores and omnivores) is not fully understood, and can suggest a mismatch of nutrient content to nutritional demand from a consumer (e.g., Pollierer et al., 2019).
However, the range of the δ 15 N offset in plants that have a true vascular system, root, and leaf is highly dependent on their morphology (reviewed in Ramirez et al., 2021).

| Beyond dietary EAA δ 13 C estimates: Effects of the microbiome
There are exceptional cases wherein the EAAs Δ 13 C between consumers and their diets are substantial (Figure 3b), ranging from 2.5‰ to >7.5‰ (Table S1). Such phenomena are reported even in designed feeding studies with diverse consumers such as rats , enchytraeid worm , and coral (Ferrier-Pagès et al., 2021). The mismatch of δ 15 C between consumer EAAs and dietary EAAs is usually interpreted as the nutritional contribution of de novo synthesized microbes (i.e., by the gut microbes) (Arthur et al., 2014;Larsen et al., 2016;Newsome et al., 2011Newsome et al., , 2014Newsome et al., , 2020. This is usually a feature in consumers adapted to live off nutritionally imbalanced diets, such as detritus, wood, and other low protein diets with low digestibility (Arthur et al., 2014;Ayayee et al., 2016;Larsen et al., 2016;Newsome et al., 2011;del Rio et al., 2009). Larsen et al. (2016) also reported by using EAA δ 13 C fingerprints that enchytraeids (oligochaetes) in Arctic peatlands derived 80% of their EAAs gut symbiotic bacteria EAAs. 16S rRNA sequencing of enchytraeids from a subsequent feeding trial found that EAA microbiome contribution to host and the taxonomic composition of the gut microbiome were significantly associated with the diets' nutritional contents, such as the proportions of starches relative to fibers. A similar finding was found for mice living off protein-deficient diets with low digestibility . By comparing Δ 13 C values and the relative abundance of specific gut microbial taxa, Newsome et al.
(2020) concluded that gut microbiome (e.g., Firmicutes) made a greater contribution of Val (~60%) to host than Phe (<6%) when the mice were fed protein-deficient and hardly digestible diets. These findings emphasize that for some animals, the microbiome can play an important role in supplementing macronutrients that are otherwise insufficient in the diet.

| COLL ABOR ATI ON OF N AND C-S IR IN A A s : FUTURE S TUDY
There are increasing trends of dual analysis of N and C isotope variables in AAs, but generally their main roles in ecological contexts are highly differentiated. For instance, AA-N assessment as a reliable tool has commonly elucidated trophic pathways in detail, while AA-C assessment improves our understanding about multiple basal C resources in trophic energy flow as well as tracking biosynthetic C transfer (Table 1). We assume that change in the AA-C isotope composition is inherently different from that of the AA-N variable.
However, the potential of collaborating AA-N and C-SIR applicability to characterize population niche and community structure is still less concerned, and thus we suggest some possible applications in the following section.

| Suggestion 1. Proxies for assessing the ecological niche of a focal species
A δ 13 Cδ 15 N bi-plot, relying on bulk analysis, is powerful to describe food web structure and population niche in a local habitat, but there are difficulties in comparing different populations (or communities) spatially and temporally. This is because bulk isotope ratios in the basal resources often show spatial and temporal variations, directly reflected to those of consumer (Shipley & Matich, 2020). Furthermore, information of basal resources (primary producers) and their sampling in local food webs are not always reliable. As explained in Section 2.2, AA-SIR variables are relatively more robust to temporal and spatial variations in local habitats than bulk δ 13 Cδ 15 N variable, informing basal N or C sources (belonging to trophic bases) by source AA and essential AA, respectively. Accordingly, we believe that AA δ 15 N and AA δ 13 C compositions in the bi-plot shown in Figure 6a might help comparing temporally and spatially different populations (or communities) as well as informing trophic links more effectively.
For instance, basal sources (multi-producers and OM) can be well separated in the lowest part and consumers are in relatively higher positions within the AA-SIR plot. As the direction changes, the size of the organism is larger and AA isotope composition is generally higher, so this plot helps visualize trophic structure and trophic dependence in complex environments in high resolution. In fact, wherein the X-axis was δ 13 C of Gly and the Y-axis was the relative trophic position (RTP = δ 15 N Glxδ 15 N Phe of consumers). The two isotopic indices help characterize the trophic distances among the consumer species, and ultimately disentangle species-specific isotopic niches. In this way, community-wide isotope parameters such as food (trophic) chain length, trophic base diversity, and ecological niche space can also be replaced by AA-SIR, which has been originally approached by bulk isotopic values (e.g., Layman et al., 2007). A large isotopic niche length or width (area) of a consumer might imply a wider range of diet choice. When the isotope niches between consumers are not distinct in local habitats, we might interpret it as overlap of consumer niches.

| Suggestion 2. Estimating complex diet composition of omnivores
To uncover the real diet compositions of consumers in local habitats, it is first necessary to disentangle heterotrophic diets and multiple basal sources. The best known δ 13 C fingerprinting approach, discussed here, involves characterizing the source mixtures composed of primary producers as well as bacteria to infer the trophic pathways of the basal C resources for general consumers ( Figure 6b). However, δ 13 C fingerprinting works poorly to distinguish diet contributions in mixtures of multi-trophic diet sources, since the C isotope values of EAAs in a plant, a planteating prey animal, and a consumer (e.g., an omnivore and its diet items) cannot be separated. Instead, AA δ 15 N measurement is a possible solution that accounts for multi-trophic sources and consumers. Combining AA δ 15 N and δ 13 C compositions in AAs can be an option in a stable isotope mixing model (Figure 6b). The traditional approach of mixing bulk C and N isotopes allows the ecologist to calculate the proportional contributions of two or four food resources in a consumer diet (Phillips & Koch, 2002). Instead, the new model can be fitted with AA δ 13 C and δ 15 N variables, accommodating more diverse food resources by including the numbers of more useful isotope variables. Consequently, estimating the proportional contributions in multiple diets of heterotrophic-level prey animals has important implications for interpreting the ecological functions of consumers that are surrounded by complex diet sources in nature.

| CON CLUS ION
In conclusion, the traditional SIR approach can be a valuable tool for source determination, but still struggles with environmental variation, limited numbers of sources determination, and ambiguous trophic transfers. Consequently, bulk SIR is still difficult to address trophic interactions, food web dynamics, and ecological processes objectively and logically in local habitats. Future work should utilize compound-specific SIR approach to describe ecological and biogeochemical processes, particularly if sampling of trophic base is too difficult or unknown to get enough sample size. To extend AA-SIR applicability in the future, analytic technique should be improved to produce more reliable data of certain essential AAs (histidine, tryptophan, and methionine), which are substantially underestimated in ecological/geochemical context. Interpretation of the AA-SIR pattern with underlying knowledge of AA metabolism allows us to properly place consumers among the diet candidates, and accordingly bring to bear ecologically meaningful ecological proxies such as TP estimates, trophic chain lengths, and dietary source portions, and ultimately viewing food web structures and ecosystem functions more clearly. of Korea (project number R2022073).

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data in this article were obtained from publically available pub-

S U PP O RTI N G I N FO R M ATI O N
Additional supporting information may be found in the online version of the article at the publisher's website.