Advancing d-amino acid-containing peptide discovery in the metazoan

Abstract

The discovery of enzyme-derived d-amino acid-containing peptides (DAACPs) that have physiological importance in the metazoan challenges previous assumptions about the homochirality of animal proteins while simultaneously revealing new analytical challenges in the structural and functional characterization of peptides. Most known DAACPs have been identified though laborious activity-guided purification studies or by homology to previously identified DAACPs. Peptide characterization experiments are increasingly dominated by high throughput mass spectrometry-based peptidomics, with stereochemistry rarely considered due to the technical challenges of identifying l/d isomerization. This review discusses the prevalence of enzyme-derived DAACPs among animals and the physiological consequences of peptide isomerization. Also highlighted are the analytical methods that have been applied for structural characterization/discovery of DAACPs, including results of several recent studies using non-targeted discovery methods for revealing novel DAACPs, strongly suggesting that more DAACPs remain to be uncovered.

1. Introduction

Prior to the 1980s, it was well-accepted that free d-amino acids played important roles in animal biology, but the presence of d-amino acid residues in animal-derived peptides/proteins had not been reported [1]. Aside from age-related, spontaneous isomerization of amino acid residues in peptides and proteins, a long-standing scientific consensus was that animal proteins/peptides were comprised of l-amino acids. However, in 1981 a d-Ala residue was detected in dermorphin, a gene-derived opioid-like hepta-peptide isolated from the skin secretions of two frog species, Phyllomedusa rhodei [2] and Phyllomedusa sauvagei [3]. In the years following the discovery of dermorphin, physiologically active d-amino acid-containing peptides (DAACPs) were uncovered in a variety of species, including spiders, crustaceans, mollusks, and even in the platypus (Figure 1). Unlike previously identified DAACPs in animals, these DAACPs were found to be the product of an enzyme-catalyzed post-translational modification (PTM), and not a result of spontaneous, age-related isomerization.

Figure 1.

Phylogeny of organisms in which enzyme-derived DAACPs have been discovered, highlighting kingdom, phylum, class, and order. The roles of the DAACPs are presented beneath the order.

Despite the increasing prevalence of animal DAACPs, modern peptide characterization workflows seldom measure stereochemistry. Even in cases when experimental results suggest isomerization, these are sometimes mis-interpreted as proline cis-trans isomerization [4,5] or disulfide bond pairing isomers [6,7]. Furthermore, it is a difficult analytical task to distinguish peptide isomers that result from spontaneous isomerization/epimerization of Asp and Ser residues from enzyme-derived DAACPs, which have a different set of biological consequences than age-related DAACPs (for relevant publications on spontaneous peptide isomerization see [8–11]). This review focuses on presumed enzyme-derived DAACPs in the metazoan (although the specific isomerases are often yet to be described). Because greater numbers of enzyme-derived DAACPs with biological functions are being discovered, it becomes increasingly important for stereochemistry to be considered when identifying new peptides and proteins. For this reason, the routine identification and discovery of novel enzyme-derived DAACPs in animals requires understanding the biosynthesis, and the biological consequences, of isomerization in functional peptides.

2. DAACPs throughout the metazoan

DAACPs have been discovered in animals across three phyla: chordates, arthropods, and mollusks. Figure 1 depicts the phylogeny—including the kingdom, phylum, class, and order—of the organisms known to have DAACPs. The DAACPs discovered across phyla differ substantially in peptide sequence, length, position and identity of the d-residue, and their physiological properties/functions. Most of these DAACPs were found to be secretory peptides, encoded by protein precursors targeted to the secretory pathway.

The first indication that DAACPs are the product of an enzyme-catalyzed PTM came from the discovery that the standard codon for l-Ala was found to be present in the cDNA sequence of dermorphin [12,13], suggesting that the d-residue was not incorporated by the ribosome. Subsequently, the discoveries of peptide isomerase activity in the extracts of spider venom [14–16], frog skin secretions [17–19], and platypus venom [7,20–22] provided more evidence that DAACPs in animals are produced by post-translational isomerization, with the enzymes yet to be characterized. Isomerase activity has also been detected in extracts of male echidna venom [23] and even mouse heart [24], providing significant evidence for the presence of l/d isomerases and DAACPs in several species.

Initial studies on the biosynthesis of DAACPs in frogs suggested that isomerization occurs at the early stages of post-translational processing, before the peptides are cleaved from the precursor [25]. However, subsequent studies on cellular and subcellular localization of DAACPs in frogs [26], crustaceans [27–29], and mollusks [30,31] have shaped the current understanding of isomerization as a late stage, enzymatic PTM that occurs in secretory granules, after the peptide has been proteolytically cleaved from its precursor by prohormone convertases [26,29,32]. Importantly, it has been suggested that l/d isomerization of cone snail toxin peptides precedes proteolytic cleavage of the precursor [33]. Exactly where l/d isomerization occurs in the secretory pathway has not been the subject of rigorous study, and it could vary depending on the species and enzymes involved. Whether isomerization occurs in the secretory pathway before or after cleavage by prohormone convertases may indeed vary in different species.

To date, two protein sequences for l/d isomerases have been reported: one l/d isomerase from the venom of the funnel web spider [34] and a second from frog skin secretions [17]. Although these two proteins have little sequence similarity and have different substrate specificities [15,16,18], both appear to catalyze the l/d isomerization of amino acid residues in peptides by similar mechanisms [16,19]. As a final note, the reported isomerase sequence for the funnel web spider has been questioned and may be a potential misidentification [35]. Similarly, while a modest amount of enzymatic activity from the frog isomerase could be recapitulated in a Xenopus oocyte system [17],to date, no l/d isomerase has been successfully purified from a recombinant system.

DAACPs can generally be classified into two categories: DAACP toxins and cell-cell signaling DAACPs. These categories are defined by the organism/tissue the peptide was isolated from, and the function/physiological activity of the peptide. The biosynthetic mechanisms of cell-cell signaling peptides and peptide toxins are similar and it is assumed that DAACPS under both categories are products of enzymatic post-translational isomerization.

2.1. DAACP toxins

For the purposes of this review, a DAACP toxin can be generally defined as a gene-derived DAACP produced by an organism that has poisonous effects on another organism. DAACPs under this category have been isolated from mammalian platypus venom, frog skin secretions, spider venom, and marine cone snail venom, and comprise a significant proportion of discovered DAACPs. Table 1 lists the DAACPs classified as toxins that have been confirmed experimentally.

Table 1.

DAACP toxins discovered in the metazoan. References correspond to the publication in which the DAACP was first confirmed experimentally. d-amino acid residues are indicated in bold and underlined (X).

Peptide	Sequence	Species	Ref
Leu-Contryphan-P	---GCVLLPWC	Conus purpurascens	[41]
Leu-contryphan-Tx	----CVLYPWC	Conus textile	[42]
Contryphan-R	---GCOWEPWC-NH2	Conus radiatus	[43]
Bromocontryphan	----COWEPwC-NH2	Conus radiatus	[33]
Contryphan-Sm	---GCOWQPWC-NH2	Conus stercusmuscarum	[44]
Contryphan-P	---GCOWDPWC-NH2	Conus purpurascens	[44]
Contryphan P/Am	---GCOWDPWC-NH2	Conus amadis	[45]
Contryphan-Lo	---GCPWDPWC-NH2	Conus loroisii	[45]
Contryphan-Vn	--GDCPWKPWC-NH2	Conus ventricosus	[46]
Contryphan-R/Tx	---GCOWEPWC-NH2	Conus textile	[42]
Contryphan-Tx	----COWQPYC-NH2	Conus textile	[42]
Glacontryphan‐M	NγSγCPWHPWC-NH2	Conus marmoreus	[47]
gld-v*	-------AOANSvWS	Conus gladiator	[48]
gld-v	-------AOANSVWS	Conus gladiator
mus-V*	-------SOANSvWS	Conus gladiator	[48]
mus-V	-------SOANSVWS	Conus gladiator
Conomarphin	DWEYHAHPKONSFWT	Conus marmoreus	[49]
R11a	----…GCSTSSFFKI	Conus radiatus	[50]
R11b	----…GCSTSSFFRI	Conus radiatus	[51]
R11c	----…GCSTNVFLT	Conus radiatus	[51]
Conomap vt	AFVKGSAQRVAHGY-NH2	Conus vitulinus	[52]
Dermorphin related	YIFHLMD	Pachymedusa dacnicolor	[53]
Dermorphin	YAFGYPS-NH2	Phyllomedusa rhodei	[2]
[Hyp6]-dermorphin	YAFGYOS-NH2	Phyllomedusa rhodei	[2]
Dermorphin	YAFGYPS-NH2	Phyllomedusa sauvagei	[3],[13]
Met-Deltorphin	YMFHLMD-NH2	Phyllomedusa sauvagei	[36,54]
Dermorphin-like1	YAFWYPN	Phyllomedusa bicolor	[12]
Dermorphin-like2	YAFGYPK	Phyllomedusa bicolor	[12]
pentapeptide	YAFWN	Phyllomedusa bicolor	[12]
Deltorphin1	YAFDVVG-NH2	Phyllomedusa bicolor	[37],[12]
Deltorphin2	YAFEVVG-NH2	Phyllomedusa bicolor	[37],[12]
Leu-deltorphin	YLFADVASTIGDFFHSI-NH2	Phyllomedusa burmeisteri	[55]
Bombinin H3	IIGPVLGMVGSALGGLLKKI-NH2	Bombina variegate	[56]
Bombinin H4	LIGPVLGLVGSALGGLLKKI-NH2	Bombina variegate	[56]
Bombinin H5	IIGPVLGLVGSALGGLLKKI-NH2	Bombina variegate	[56]
GH-2	ILGPVLDLVGRALRGLLKKI-NH2	Bombina orientalis	[57]
Phenylseptin	FFFDTLKNLAGKVIGALT-NH2	Hypsiboas punctatus	[58]
ovCNPb	LLHDHP…	Ornithorhynchus anatinus	[5]
DLP	IMFEMQ…	Ornithorhynchus anatinus	[7]
ω-agatoxin	…LIMEGLSFA	Agelenopsis aperta	[40]

w = 6-bromotyrptophan; v = gamma d-hydroxy valine; O = hydroxyproline; γ = γ‐carboxylated glutamate; I = d-allo-isoleucine. Note: this list is curated to include only DAACPs that have been confirmed experimentally. Some DAACPs may have been excluded unintentionally. More details on these DAACPs are reported in Tables S2–S4.

Toxin DAACPs can cause potent interspecies biological effects. For example, the DAACPs isolated from frog skin secretions, dermorphin and deltorphin, showed significantly greater affinity and selectivity for mammalian δ-opioid and μ-opioid receptors than their all-l counterparts [36–39]. The DAACP isolated from the funnel web spider, ω-agatoxin, is able to block Purkinje neuron P-type calcium ion channels and has modestly higher potency than the all-l-form (4- to 90-fold, depending on the assay) [14,40].

While the d-residue in many toxins appears to be required for bioactivity of the toxin, this is not always the case. The bombinin H family of DAACPs from frog skin secretions have potent antimicrobial and hemolytic activity [56,71,72], but the biological effects of the all-l and d-amino acid-containing bombinin H family of peptides do not differ significantly. Many toxins remain functionally and structurally uncharacterized, and many DAACPs have yet to be identified in venoms/toxins produced by several species. As researchers continue to characterize the active components of venoms, it will be exciting to investigate if DAACPs toxins are found to be present in other species.

2.2. Cell-cell signaling DAACPs

We define cell-cell signaling DAACPs as peptides that have been either shown or are expected to be involved in physiological processes within the organism that produces them. Cell-cell signaling DAACPs have been identified in arthropods and mollusks, but to our knowledge, have not been identified in chordates. A list of cell-cell signaling DAACPs that have been confirmed in mollusks and arthropods are reported in Table 2. For DAACPs involved in cell-cell signaling that have been functionally characterized, the d-residue is often critical for the peptides’ biological activity. For example, DAACPs found in gastropod mollusks, including achatin and achatin-related peptides (GFAD, GFFD, and GYFD; d-amino acid residues indicated in bold and underlined) [59,66,67,73,74], fulicin (FNEFV-NH₂) [60,75], and fulyal (YAEFL-NH₂)[61], show potent activity from physiological measurements of neurons, muscle tissue or behavior, whereas their all-l-residue counterparts show little or no activity in the same experiments. Similarly, the d-residue in the cardioactive DAACP Ocp-1 (GFGD) isolated from octopus was determined to be critical for bioactivity. However, the opposite pattern was observed with another DAACP isolated from octopus, Octp-4 (GSWD), wherein bioactivity was present only for the all-l form of the peptide [62].

Table 2.

Cell-cell signaling DAACPs discovered in the metazoan. References correspond to the publication in which the DAACP was first confirmed experimentally in the reported species. d-amino acid residues indicated in bold and underlined (X).

Name	Sequence	Species	Ref
Achatin-1	GFAD	Achatina fulica	[59]
Fulicin	FNEFV-NH2	Achatina fulica	[60]
Fulyal	YAEFL-NH2	Achatina fulica	[61]
Ocp-1	GFGD	Octopus minor	[62]
Ocp-4	GSWD	Octopus minor	[62]
FMRF-related decapeptide	ALAGDHFFRF-NH2	Mytilus edulis	[63]
NdWFa	NWF-NH2	Aplysia kurodai	[64]
NdWFa	NWF-NH2	Aplysia californica	[30,65]
ApALNP-1	GFFD	Aplysia californica	[66]
ApALNP-2	GYFD	Aplysia californica	[67]
ApALNP-3	SYADSKDEESNAALSDFA	Aplysia californica	[67]
[d-Phe2]ATRP	GFRLNSASRVAHGY-NH2	Aplysia californica	[68]
[d-Phe2]Plrn1	MFYTKGSDSDYPRI-NH2	Aplysia californica	[69]
[d-Phe2]Plrn2	SFYTTGNGNHYPRI-NH2	Aplysia californica	[69]
[d-aIle2]Plrn3	GIFTQSAYGSYPRV-NH2	Aplysia californica	[69]
[d-Leu2]Ip1	YLDHLGSSLV	Aplysia californica	[69]
[d-Leu2]Ip2	YLDGIASSLI	Aplysia californica	[69]
[d-Met2]FMRGFa	FMRGF-NH2	Aplysia californica	[69]
[d-Tyr3]Plrn2	SFYTTGNGNHYPRI-NH2	Aplysia californica	[69]
[d-Phe3]Plrn3	GIFTQSAYGSYPRV-NH2	Aplysia californica	[69]
[d-Phe3]CHH	qVFDQACK…	Homarus americanus	[70]
[d-Trp4]VIH	ASAWFTND…	Homarus americanus	[29]

q = pyro glutamic acid; CHH = crustacean hyperglycemic hormone, VIH = vitellogenesis inhibiting hormone; ApALNP = Aplysia achatin-like neuropeptide. Note: this list is curated by the authors to include only DAACPs that have been confirmed experimentally. Some DAACPs may have been excluded unintentionally. More details on these DAACPs are reported in Tables S2 and S4.

To our knowledge, receptors have been identified for only two families of endogenous cell-cell signaling DAACPs: the achatin family (achatin-1, ApALNP-1, ApALNP-2 and homologues) [76,77] and Aplysia allatotropin-related polypeptides (ATRP) [68]. In the case of achatin family peptides, the DAACP form is a potent agonist of the receptor, while the all-l-residue analogues are devoid of biological activity [76,77]. These findings demonstrate that the d-residue is critical for receptor activation in the achatin family, which is consistent with previous results from in vivo and ex vivo assays showing that the d-residue is required for biological activity [59,66,67,73]. These functional studies suggest that in many cases the post-translational incorporation of a d-residue is a key step in generating the biologically active form of the peptide used for cellular communication.

In contrast to the results described above, there are also several cases of DAACPs that display similar biological activity to their all-l-residue counterparts. For example, both forms of Aplysia ATRP are potent agonists of the identified receptor, and the all-l-residue form is somewhat more potent at activating this receptor than the DAACP form (Figure 2A) [68]. Additionally, both the all-l and d-amino acid-containing forms of ATRP showed similar effects in neuronal electrophysiology experiments [68], and similar potency in myoactivity assays of both diastereomers of the bivalve mollusk FMRFa-related decapeptide were also observed [63]. As a final example, the DAACP and all-l-residue forms of the crustacean hyperglycemic hormone (CHH) demonstrated similar effects on hyperglycemia (albeit with different kinetics) [70] and on ecdysteroid synthesis [78].

Figure 2.

Biological consequences of l- to d-residue isomerization among Aplysia neuropeptides. (A,B) Representative dose-response curves showing the ability of peptides to activate either the (A) achatin receptor or (B) ATRP receptor, when transiently transfected in CHO-K1 cells. Points represent mean ± standard deviation for duplicate wells in an experiment. (C) Time-course of peptide degradation in Aplysia hemolymph plasma, as assessed by LC-multiple reaction monitoring. Points represent mean ± standard deviation for two biological replicates, and curves show fit of the data to an exponential decay model. Panel (A) adapted from J.W. Checco, G. Zhang, W.D. Yuan, K. Yu, S.Y. Yin, R.H. Roberts-Galbraith, P.M. Yau, E.V. Romanova, J. Jing, J.V. Sweedler, Molecular and physiological characterization of a receptor for d-amino acid-containing neuropeptides, ACS Chem. Biol., 13 (2018) 1343–1352. Copyright 2018, American Chemical Society [76]. Panels (B) and (C) originally published in the Journal of Biological Chemistry, and adapted from J.W. Checco, G. Zhang, W.D. Yuan, Z.W. Le, J. Jing, J.V. Sweedler, Aplysia allatotropin-related peptide and its newly identified d-amino acid-containing epimer both activate a receptor and a neuronal target, J. Biol. Chem., 293 (2018) 16862–16873, © the Authors [68].

Although physiological measurements of biological activity using in vivo or ex vivo approaches are powerful tools for elucidating peptide functions, they are somewhat limited in their ability to determine the full consequences of l/d isomerization. Physiological experiments output a response that is significantly downstream of the initial signaling event and dependent on many factors. In addition, the choice of bioactivity experiment for many newly discovered peptides is often somewhat arbitrary and may not reflect a primary biological function in the animal. In contrast, studying a peptide’s direct activation of cognate receptors can provide important information regarding their signaling capabilities that decouples receptor potency from other factors.

2.3. Alternative roles for l/d isomerization

Although the several DAACP toxins and cell-cell signaling DAACPS mentioned above do not differ substantially in bioactivity from their all-l counterparts, the presence of a d-residue can have other biological consequences. The presence of d-residue(s) within a sequence can slow native and exogenous protease action relative to an all-l-residue peptide, presumably because the substrate-altered backbone does not allow it to be positioned correctly into the protease active site [79,80]. For these reasons, DAACPs may have greater stability against proteolytic degradation and an increased biological lifetime. For some DAACP toxins that have been identified, studies have shown a higher protease stability for the DAACP relative to the all-l-residue peptide in the venom environment [14,17,37], suggesting the d-residue provides stability while the toxin is stored and waiting to be released. For cell-cell signaling DAACPs, increased resistance to protease degradation may allow these peptides to engage in hormone-like signaling, activating receptors at distances far from the point of release. Consistent with this hypothesis, in experiments measuring the degradation of naturally occurring Aplysia DAACPs in hemolymph plasma, the DAACP forms were found to be >70-fold more stable to degradation than their all-l-residue analogues (Figure 2C) [68]. These results are consistent with the fact that hemolymph is rich in aminopeptidases [81] and for some aminopeptidases, activity is hindered by a single d-residue close to the N-terminus of a peptide sequence [82,83]. The increased stability of DAACPs may indicate another possible role for peptide l/d isomerization: to modify peptide stability and thus, duration of action. It will be important for future studies to carefully examine the effects of l/d isomerization on the lifetime of peptides in their natural context (e.g., upon release at the synapse for neuropeptides, or after exposure to prey/predators for toxins) to gain full insight into the consequences on stability.

2.4. Structural diversity of DAACPs

All known gene-derived animal DAACPs discovered to date are epimers of their all-l-residue isomerase substrate peptide (Table 1, Table 2). In other words, l/d isomerases in animals appear to invert the stereochemistry of only one amino acid residue in the peptide, usually at a defined position in the backbone. This important characteristic makes animal DAACPs largely distinct from those found in bacteria, which often have more than one d-residue at multiple positions, and are diastereomers of the all-l residue peptide [84]. Very recently, DAACPs were identified from the whole bodies of the annelid Whitmania pigra that contained two d-residues [85]. However, it is unclear if these DAACPs were gene-derived peptides and produced via the biosynthetic machinery of this animal, or through other mechanisms (e.g., spontaneous isomerization or bacterial derived). Interestingly, the isomerized residue in animal DAACPs is oftentimes located near the amino or carboxy termini of the peptides, albeit with several exceptions (Table 1, Table 2). Another interesting observation is that for most DAACPs across the metazoan, the isomerized residue is often an amino acid with aromatic or hydrophobic side chains. To date, peptides containing the hydrophobic residues d-Ala, d-Leu, d-aIle, d-Val, d-Met, and aromatic residues, d-Phe, d-Tyr, d-Trp, have been isolated from animals (Table 1, Table 2). Additionally, peptides containing d-Asn and d-ser residues have also been discovered (Table 1, Table 2). Studies reporting the substrate specificity of l/d isomerase enzymes indicate the activity of the isomerases towards a specific substrate is influenced by the composition of the isomerized residue, but it is also influenced by the substrate sequence [16,18,22]. To our knowledge, enzyme-derived DAACPs containing d-Thr, d-Arg, d-Lys, d-Glu, d-Asp, d-Cys, d-His, or d-Gln, d-Pro residues have not been isolated from animals, highlighting an important gap in the current knowledge regarding the structural diversity of DAACPs in the metazoan. Studies on peptide isomerase activities are extremely limited, and likely do not represent a complete picture of the isomerase activity and specificity in the metazoan.

Importantly, because peptide stereochemistry is not routinely evaluated in modern peptide characterization workflows, the structural diversity of DAACPs may be much greater. Although the functions of many DAACPs have not been fully explored, the structural features common to known DAACPs (e.g., position and composition of the d-residue(s)) can aid in designing approaches to facilitate future discovery efforts for novel DAACPs by targeting unique structural features common to many substrates of l/d isomerization.

3. History of DAACP discovery and the challenges moving forward

Many of the first animal DAACPs were discovered between the 1980s and early 2000s, before the widespread usage/availability of modern mass spectrometric technologies. At that time, peptide sequencing required large quantities of peptide to be purified, sometimes from hundreds of animals. Oftentimes, the bioactivity of the native peptide of interest was known before its structure, and this activity guided the purification [3,33,36,37,43,55,59,60,62,63]. Following purification of a native peptide, the amino acid composition and sequence were determined by techniques such as acid hydrolysis/amino acid analysis or Edman degradation. Finally, the structure of the native peptide was confirmed by comparing the bioactivity, chromatographic retention, and other properties of the purified native peptide to a synthetic peptide standard. Following this workflow, many of the first animal DAACPs were discovered after the biological activities [40,59,60,62], chromatographic [5,7,40,43,49,50,52,59,60,63,64] or spectroscopic properties [5,7,48,59] of the purified native peptide(s) were not reproduced by the synthetic all-l peptide standards [5,7,40,43,48–50,52,59,60,62–64]. The presence of a d-residue was only revealed upon further investigation into the native peptide structure using various chiral amino acid analysis techniques [86–89].

Figure 3 is a schematic representation of the peptide characterization workflow that led to the discovery of many DAACPs, a process that has continued to evolve since the discovery of dermorphin. With the increasing capabilities and availability of mass spectrometry (MS) instrumentation, large-scale peptide purifications and chemical amino acid sequencing techniques have generally been replaced by high throughput liquid chromatography (LC)-tandem MS (MS/MS) sequencing. LC-MS/MS data can be searched against genomic- and transcriptomic-derived protein databases, enabling accurate identification of hundreds to thousands of sequences in a complex mixture in a single experiment. However, because MS cannot easily differentiate peptide stereoisomers, stereochemistry is not routinely evaluated in modern peptidomics studies. As a result, DAACPs may be overlooked simply because the practice of confirming native peptide structures by comparing their properties to synthetic standards has become increasingly limited. Consequently, the abundance of DAACPs among animal peptides may be severely underestimated.

Figure 3.

Peptide characterization workflow historically used that resulted in the discovery of DAACPs in animals. (A) Large-scale peptide extraction and (B) purification and amino acid sequencing was performed to determine amino acid sequence and composition without consideration of stereochemistry. (C) Once the sequence was known, physical/chemical properties or biological activities of the native peptide and synthetic peptide sequences (all-l-residue) were compared. (D) If the properties of the native peptide were not reproduced by the synthetic peptide, further structural characterization of the native peptide was performed using chiral amino acid analysis techniques or other structural characterization methods, and the native peptide was discovered to contain a d-residue.

It is important to note that the major challenge facing modern DAACP discovery is not the characterization of stereochemistry. The real challenge is predicting which few peptides, among the thousands that are detected in LC-MS/MS experiments, likely contain a d-residue. As soon as a DAACP is suspected, identifying its stereochemistry is straightforward and can be accomplished with chiral amino acid analysis techniques, or by comparing physical and chemical properties to synthetic peptide standards. Various modern methods used to characterize the structure of DAACPs, including amino acid analysis, LC-MS/MS, ion mobility (IM)-MS, MS/MS, capillary electrophoresis, etc., have been discussed in recent reviews [90–92], and a comprehensive list of methods applied to characterize native DAACPs can be found in Table S1. Despite the number and quality of analytical tools now available for differentiating peptide stereoisomers, they are being underutilized by modern peptidomics workflows for the identification of native peptide stereoisomers.

Although native peptide diastereomers do not differ by mass, they can be differentiated based on other physical or chemical properties such as gas phase structure, enzymatic stability, electrophoretic mobility, chromatographic retention, and even MS/MS fragmentation. Putative DAACPs may also be identified based on sequence similarities to known DAACPs. Modernization of DAACP discovery requires analytical methods capable of predicting which peptides in a complex mixture are DAACPs before the biological activity is evaluated. As discussed further in Section 4.3., Modern DAACP discovery, the goal is to develop standardized workflows to identify novel DAACPs.

4. Methods for identifying DAACPs

4.1. Predicting DAACPs based on homology and peptide sequence

In some cases, peptides that undergo post-translational l/d isomerization can be predicted based on the l/d isomerase specificity (if known) [18], the sequences of known DAACPs from the organism, or by sequence similarity to known DAACPs in other organisms. For example, if a DAACP is discovered in one organism, one might predict that similar peptide sequences from the same organism, or in a related species, may also exist as DAACPs. This homology-guided discovery can help narrow the list of potential DAACPs that will be subsequently characterized.

Until recently, most modern DAACP discovery efforts relied heavily on homology to find DAACPs in different organisms. For example, a d-residue has been detected at position 2 from the N-terminus in dermorphin, deltorphin, and related peptides from a variety of frog species [2,12,13,25,53,55,93,94]. Because this peptide family has been confirmed to undergo post-translational l/d isomerization in so many instances across species, it is largely assumed that a d-residue is present in all of the dermorphin/deltorphin-related peptides in frog skin secretions [53,93]. Many cone snail DAACPs in the contryphan conotoxin family were also discovered based on homology [33,41,42,44–47]. In crustaceans, the DAACP forms of CHH and vitellogenesis inhibiting hormone (VIH) were originally discovered in the American lobster, Homarus americanus [29,70], but the d-residue has been shown to be conserved in VIH and CHH peptides from a variety of other lobster and crayfish species (Table S2) [27,78,92,95–99]. The achatin-like peptides from A. californica, GFFD and GYFD, were both initially targeted and discovered because these peptides had similar sequences to the known DAACP, GFAD from A. fulica, and were encoded by homologous prohormones [66,67]. Similarly, the A. californica DAACP, allatotropin-related polypeptide ([d-Phe²]ATRP), was targeted because of its sequence similarity to the cone snail DAACP toxin, conomap-Vt [68].

Once a DAACP has been discovered in one organism, there is a high likelihood that other peptides in the organism (or related species) also have DAACPs. For instance, Mignogna et al. [56] tested the bombinin H peptides in Bombina variegata for d-residues simply because “peptides from the skin of frogs belonging to the sub-family Phyllomedusinae contain a d-amino acid in position 2,” despite having virtually no sequence similarity with dermorphin-related peptides. If one peptide encoded by a prohormone is identified as a DAACP, it is likely (but not guaranteed) that similar peptides encoded by the same prohormone also exist with a d-amino acid-containing form. The finding that multiple peptides encoded by a common precursor undergo l/d isomerization was first observed with peptides from the A. fulica and P. sauvagei, and later observed in other organisms [25,61,67,69]. Recently, three peptide sequences encoded by the A. californica pleurin precursor, three peptides from the ApALNP precursor, and two peptides from the Aplysia isomerized peptides precursor (AIPP) were discovered to be DAACPs [66,67,69]. However, just because multiple similar peptides are encoded on the same precursor does not guarantee every peptide will have a d-amino acid-containing form [61,69]. The latter observation may be due to the l/d isomerase substrate specificity, or native substrate concentration [69].

Importantly, prohormone homology and sequence similarity are not always predictors of DAACPs between organisms. For example, although fulyal peptide (YAEFL-NH₂) was found to be a DAACP in A. fulica, the identical sequence in A. californica was only detected in the all-l-residue form [66]. As another example, the sequence of the DAACP from platypus venom, ovCNPb, closely resembles the sequence for the mammalian C-type natriuretic peptide (CNP) [4,5]. Although the possibility that mammalian CNPs may exist in multiple stereoisomers has not been evaluated, it should not be assumed that they have a d-residue simply because of this homology. The conservation of d-residues in homologous peptides across species is likely dependent on the conservation of the l/d isomerase enzyme.

Although DAACPs are common among certain peptide families across species, homology-guided discovery alone significantly limits the ability to discover novel DAACPs. To uncover novel and more structurally diverse DAACPs in the metazoan, peptide characterization efforts would benefit from developing methods to swiftly characterize native peptide stereochemistry instead of relying on previous findings. In the future, cellular localization of peptides may be an important factor for determining whether a peptide may exist as a DAACP. As more DAACPs are discovered, and single cell characterization methods advance, it may be possible to predict novel DAACPs by targeting peptides that colocalize with known DAACPs, or with the l/d isomerase (if known) in single cells.

4.2. Analytical methods for identifying DAACPs

DAACPs oftentimes coexist in biological extracts with the all-l-residue form of the peptide [5,30,56,59,62,69,70,100]. Therefore, substrates of l/d isomerization may be identified without the need for synthetic standards by attentively searching for evidence of multiple diastereomers of peptide sequences using separations techniques or other specialized approaches. Native peptide diastereomers are often observed in chromatographic separations as multiple peaks of the same sequence eluting at different retention times, a key signature indicating that a peptide could have a d-amino acid-containing form [7,69,92,101]. The DAACPs from platypus and lobsters/crayfish were initially discovered because identical peptide sequences were found to elute at more than one retention time [4,5,29,70].

Six peptide sequences detected with split peak elution profiles by LC-MS/MS were recently discovered to have d-amino acid-containing forms (Figure 4) [69]. Modern LC-MS/MS peptide characterization studies could implement routine identification of DAACP candidates by analyzing chromatograms for peptides that elute at multiple retention times, helping identify peptides to target for downstream structural characterization. Some peptidomics studies have already begun to report peptides with split eluting peaks, even when no follow-up chiral analysis was performed [102]. In cases where standard conditions fail to separate DAACPs (i.e., C18 stationary phase, water/acetonitrile mobile phase), a separation of peptide stereoisomers can be accomplished by modifying gradient conditions, solvent system, stationary phase, or chromatographic technique [60,68,69,103]. Future DAACP discovery workflows may include LC-MS/MS analysis using several separation conditions to identify peptide stereoisomers that coelute under one chromatographic condition, but separate under different conditions. However, DAACPs could easily be overlooked if their discovery relies solely on chromatographic separation approaches. Co-elution of native peptide stereoisomers may be common. Therefore, alternative methods that can be easily worked into LC-MS based workflows must also be considered to assist in DAACP identification.

Figure 4.

Identification of peptide stereoisomers in A. californica central nervous system extracts by LC-MS/MS based on multiple peak elution. Each plot shows LC-MS/MS extracted ion chromatograms corresponding to native peptide stereoisomers of (A) Plrn1, (B) Plrn2, (C) Plrn3 (D) Ip1, (E) Ip2, (F) FMRGF-NH₂ in a peptide extract of cerebral ganglia. Within each plot, each peak is a peptide stereoisomer, and the peaks that were identified as DAACPs are labeled with the d-amino acid residue that was detected in the peptides. Peaks labeled all-l correspond to the all-l-residue form of the peptide. The full sequences for the peptides can be found in Table 2. The peak labeled with an “X” in (E) was not determined to be a stereoisomer of Ip2. Adapted from D.H. Mast, J.W. Checco, J.V. Sweedler, Differential post-translational amino acid isomerization found among neuropeptides in Aplysia californica, ACS Chem. Biol., 15 (2020) 272–281. Copyright 2020, American Chemical Society [69].

IM-MS can achieve rapid gas phase separation of peptide stereoisomers [58,90,104–111] and is a potentially valuable approach for identifying native peptide stereoisomers. IM-MS methods have been widely applied for the structural characterization and separation and of known naturally occurring DAACPs and other peptide stereoisomers [58,104–109,111]. Recently, a workflow utilizing IM-MS in conjunction with MS/MS was developed to perform site-specific localization of d-amino acid residues in a known DAACP [110]. The latter was a milestone for DAACP discovery techniques not only because native peptide diastereomers could be differentiated by IM-MS, but more importantly, the position of the d-residue in the peptide could be localized by comparing the mobilities of product ions between the two diastereomers. However, IM-MS methods have seldom been used in the direct discovery of native DAACPs. Recently, trapped ion mobility spectrometry (TIMS) combined with LC-MS/MS was used to confirm and differentiate native stereoisomers of Plrn2 ([d-Phe²]Plrn2 and [d-Tyr³]Plrn2) in different regions of the A. californica central nervous system and validate stereochemical assignments [69]. At present, IM-MS is underutilized in native DAACP discovery workflows, but exciting new discoveries will certainly be made as IM-MS technology improves, resolving power increases, and the approach becomes more common. Future DAACP discovery efforts may integrate IM-MS into LC-MS/MS-based peptidomics workflows to perform gas phase separations of native peptide stereoisomers that may coelute by LC.

MS/MS can be a powerful tool for differentiating peptide stereoisomers by comparing the relative abundances of the fragment ions intensities produced by this measurement approach [112–115]. DAACPs can be identified, confirmed or quantified by comparing the fragment ion spectra of native peptide stereoisomers to each other, or comparing the fragment ion ratios of native peptides and synthetic standards [31,48,57,66,69]. The low detection limits of MS/MS-based methods are a benefit for confirming or identifying possible DAACPs. However, a drawback is that structural confirmation requires a synthetic standard for comparison, which is not ideal for non-targeted discovery efforts when hundreds of potential DAACPs are possible. MS/MS has been extremely useful in recent years for structural characterization of peptide isomers resulting from spontaneous/age-related peptide isomerization/epimerization. Looking ahead, peptide characterization protocols may increase confidence in peptide stereoisomer assignments using LC-MS/MS by comparing the product ion ratios of peptide sequences eluting at multiple retention times.

Another successful approach for identifying candidate DAACPs is by analysis of the native peptide’s enzymatic stability. In combination with chromatographic separations and MS detection, enzymatic digestions can simultaneously aid in the identification of peptide stereoisomers and in localization of the position of the d-residue within a peptide sequence. The presence of d-amino acid residues in peptides can make them poorer substrates for proteolytic enzymes [116]. Therefore, DAACPs can be differentiated from their all-l counterparts by the rate of proteolysis, or by the degradation products of proteolytic enzymes [3–5,49,50,52,70,78].

The exopeptidase, aminopeptidase M (APM), has been widely used for identifying native DAACPs. APM sequentially cleaves l-amino acid residues from the N-termini of peptides, but peptides with a d-residue present near the N-terminus tend to resist degradation by APM [82,83]. Because APM has a well-defined stereoselectivity and substrate cleavage specificity, it has historically been used to help confirm the position of the d-residue in position 2 in naturally occurring DAACPs [3,5,56,65,67–69,85]. This was done simply by testing the resistance of the native peptide to APM digestion compared to the synthetic all-l peptide and synthetic DAACP standards. The DAACP is expected to be resistant to APM, while the all-l peptide is expected to degrade. While APM is commonly used to identify native DAACPs, future discovery efforts could benefit from using other proteolytic enzymes with different substrate specificities to aid in identifying native DAACPs.

4.3. Modern DAACP discovery methods

Modern DAACP discovery approaches seek to identify potential DAACPs in animals using the methods discussed above. Workflows have only just begun to be used to evaluate the chromatographic/mass spectrometric properties and enzymatic stability of peptides in biological extracts to identify possible DAACP candidates for follow up analysis (Figure 5). These approaches take into consideration peptide homology, enzyme stability, and MS properties to look for possible DAACPs in peptide extracts.

Figure 5.

Proposed general workflow for modern DAACP discovery. (A) Peptide extraction followed by (B) LC-MS/MS analysis and sequencing by database search to identify native peptides. (C) Identification of putative isomeric peptides detected in LC-MS/MS analysis by analyzing ion mobility migration, MS/MS, protease stability, chromatographic retention, or homology. (D) Finally, after identifying peptide stereoisomers, the DAACPs are confirmed by comparing the properties of native peptides to using synthetic standards, or by chiral AA analysis.

Recently, an LC-MS/MS based, non-targeted workflow termed “the DAACP discovery funnel” was developed that used APM to identify potential peptides with a d-residue in position 2 based on stability against APM digestion [67]. Peptides that resist degradation by APM—and thus may have a d-residue near the N-terminus—are purified and the stereochemistry evaluated using chiral amino acid analysis techniques. If a d-residue is detected in the peptide, the native peptide’s stereochemistry is confirmed by comparing the chromatographic retention of the native peptide with synthetic standards. More recently, the discovery funnel was improved upon by adapting the APM screening step to large scale LC-MS/MS peptidomics [69]. Using this optimized discovery funnel workflow, the resistance to APM of hundreds of peptides in peptide extracts was evaluated using LC-MS/MS and comparing multiple chromatographic and mass spectrometric properties of peptides before and after APM digestion. This led to the discovery of eight novel DAACPs, including three from previously uncharacterized prohormones in the A. californica central nervous system [69]. DAACP candidates were assigned if they met a strict set of criteria, e.g., peptide sequences needed to be resistant to APM degradation and be observed at multiple retention times before APM treatment. Because the position of the d-residue could be localized to the N-terminus based on the stereospecificity of APM, the purification of the peptides and the chiral amino acid analysis step could be skipped and the predicted stereochemistry of the native peptides confirmed by comparing their chromatographic retention, mobilities (via TIMS), and MS/MS fragmentation patterns to synthetic peptide standards [69]. This method also showed that by considering multiple properties (i.e., mass spectrometric and chromatographic properties, and enzyme stability) it is possible to rapidly discover structurally diverse DAACPs in a peptidomics experiment.

5. Conclusions

Among the enzyme-derived DAACPs that have been identified, l/d isomerization has been shown to play a functional role in toxins, hormones, and neuropeptides. However, despite this functional importance, stereochemistry is rarely considered in modern peptide characterization experiments due to the technical challenges of differentiating l- and d-residues using LC-MS-based methods. Because of these challenges, animal DAACPs may be missed and thus, more diverse than has been currently reported. Importantly, aside from the DAACP toxins identified in the platypus, DAACPs have not been identified in mammals. The current prevalence of DAACPs in the metazoan and the observation of l/d isomerase activity in mouse heart homogenates [24], suggests mammals, including humans, may have functional DAACPs that have not yet been discovered. As more non-targeted methods are developed to detect DAACPs, it will be exciting to see if this modification is found in mammalian cell-cell signaling peptides. Characterization of the l/d isomerases responsible for producing native DAACPs will provide important insight into their biosynthesis, and the conservation of DAACPs throughout the metazoan could help identify new DAACPs. Many of the genomes and transcriptomes of animals known to have DAACPs are incomplete, and discovery of the isomerase(s) will benefit from more comprehensive genetic sequencing and proteomic analysis. Further, while nearly all known DAACPs in the metazoan have been encoded by proteins in the secretory pathway, the enzymatic incorporation of d-residues into non-secretory, gene-derived proteins/peptides has not been studied. It is interesting to speculate that proteins larger than those identified thus far might be isomerized, although this would be extremely challenging to detect. Modern peptide characterization studies could benefit from integrating new approaches to evaluate stereochemistry, including methods outlined in this review. Future developments in non-targeted DAACP identification will ultimately help reveal the true prevalence and biological consequences of l/d isomerization among animals, and may even uncover new structurally diverse DAACPs in the metazoan.