General lack of structural characterization of chemically synthesized long peptides

Abstract

Many peptide chemistry scientists have been reporting extremely interesting work on the basis of chemical peptides for which the only characterization was their purity, mass, and biological activity. It seems slightly overenthusiastic, as many of these structures should be thoroughly characterized first to demonstrate the uniqueness of the structure, as opposed to the uniqueness of the sequence. Among the peptides of identical sequences in the final chemical preparation, what amount of well‐folded peptide supports the measured activity? The activity of a peptide preparation cannot prove the purity of the desired peptide. Therefore, greater care should be taken in characterizing peptides, particularly those coming from chemical synthesis. At a time when the pharmaceutical industry is changing its paradigm by moving substantially from small molecules to biologics to better serve patients' needs, it is important to understand the limitations of the descriptions of these products and to start to apply the same “good laboratory practices” to our peptide research. Here, we attempt to delineate how synthetic peptides are described and characterized and what will be needed to describe them in regards to how they are well‐folded and homogeneous in their tertiary structure. Older studies were done when the tools were not yet discovered, but more recent publications are still lacking proper descriptions of these peptides. Modern tools of analysis are capable of segregating folded and unfolded peptides, even if the preparation is biologically active.

Introduction: Peptides? Which Peptides?

In the newest modern approach to pharmacology and therapeutics, a growing space is reserved for biologics. Sharfstein1 stated that one can separate biologics into several categories: proteins, cells, peptides, nucleic acids, carbohydrates, and viruses. This is a change of paradigm compared to small molecules, which used to form ~95% of the pharmacopeia for the last few decades. Without stating whether this trend is temporary or progress, these new molecular entities have changed the landscape of some areas of the pharmacology world, including structural biology, industrial chemistry, governmental agency recommendations, and molecular research. Peptides also form a family of tools used to better understand physiopathological processes, even before they are turned into potential drugs. Such examples are numerous and frequent in the literature. Peptides have also been seen as a family of chemicals that can fill the gap between small molecules and antibodies, with special mention of their theoretical capacities to drug the undruggable regions, such as the protein–protein interfaces, a surface that is often too large for small molecules, with scarce hot spots, too distant from one another to be covered by a single chemical molecule. Biochemists also put forward a very simple idea: proteins are made of bricks (amino acids) more or less modified in living organisms (by post‐translational modifications). Most biological research, including therapeutic research, attempts to understand how these proteins function. For such a task, we use compounds that fall into two categories: peptides and nonpeptides. The former are formed of the same bricks as proteins, the latter are not. Which category is best suited to complement the architecture of proteins—the ones made of the same bricks or those made mostly of flat aromatic moieties? Our main concern, here, is proteins and peptides, particularly those obtained chemically. Not only can they be made of the 20 natural amino acids, but they can also benefit from the integration of several hundreds of exotic (i.e., nonproteinogenic) amino acids with the availability of protected amino acids of increasing purity and increasing diversity (see examples of such diversity in Xiao and Schultz2 or in Gates et al.3). The current status of peptides as drugs in therapeutics has been covered by interesting reviews, such as those from Henninot et al.,4 Erak et al.,5 and Kaspar and Reichert.6 Peptides containing ~50 amino acid residues form an important family of biologics, with 70 peptide therapeutics on the market in 2016. Over the past few years, they have been accepted by the Food and Drug Administration (FDA) with a frequency of 2–3 per year. The number of new preclinical studies on peptides has grown to more than 500. Interestingly, the FDA made a cutoff of 100 amino acids for peptide/protein.

As the technologies have progressed over the years, the lengths of the peptides entering clinical development have grown, though moderately for those beyond 50 amino acids (Fig. 1)7 up to a point where it is probable that in the next decade or so, chemically obtained proteins will be generated at both the research and therapeutic levels.

Figure 1.

Length of peptides entering clinical development. Adapted from Lau and Dunn.7

In the present review, we chose to survey the nature of the characterization of the synthetic peptides as opposed to recombinant ones. We rather limited ourselves to synthetic peptides, essentially because (i) most of the peptides in the Pharmacopeia are from chemical origins; (ii) research peptides are often isolated from living sources (insect skin, various poisons, organ homogenates), purified, sequenced, and then synthesized; and (iii) their analogs are mostly obtained through chemical synthesis. Nevertheless, the existence of many problems linked to the nature and characterization of peptides used in research, whether from chemical or recombinant origins, must be addressed because the often limited quality of these peptides contributes to the limited reproducibility of biological results.

Bad Reputation

Less than 10 years after the revolution of solid‐phase synthesis by Merrifield,8 it was often thought that protein chemistry was in its decline.9 Peptides have a bad reputation of being expensive to produce, not following Lipinski rules, and not easily crossing membranes, and that even if they do, they are fairly unstable in the human body. This reputation has blocked progress, including in research areas in which the peptide chemistry has often been considered the Cinderella of medicinal chemistry, if not in academia, then at least in industry. However, 64 therapeutic peptides are available in the pharmacopeia in the USA and/or Europe and more than 150 are in active development, even when excluding insulin from this catalog. They have generated more than 10 billion dollars in market value. By being “easier” to produce, and with longer structures becoming available to research laboratories, we were interested in outlining some of the limitations associated with the descriptions of such “long” peptides. In contrast to common beliefs, the in‐depth characterization of compounds remains a high standard for good science and research. Peptides that are synthesized, purified by reversed phase high pressure liquid chromatography (RP‐HPLC), and characterized by mass spectrometry (MS) are seen frequently. On this basis, the peptide can be determined to be 95% pure. Furthermore, a simple biological test in which the peptide expectedly exhibits a given activity leads the reader to assume that once purified, peptides presenting the right mass in a mass spectrum are pure and homogeneous, despite indirect evidence that some S‐S bridges are not formed spontaneously and/or wrongly assembled.

Thus, peptides have a poor reputation among most medicinal chemists. Among the liabilities attributed to this modality, most can be addressed by the skills of peptide chemists, as demonstrated by the steady flow of approved peptide therapeutics.7 Here are some counterexamples:

1. “Peptides will ultimately be replaced by small molecules discovered by high throughput screening”: Following the discovery of the first nonpeptide antagonist of cholecystokinin receptors derived from a naturally occurring benzodiazepine in 1985,10 random screening of small molecule libraries has been predicted to provide direct access to both antagonists and agonists for any peptide receptor. Although it has been shown to be true mainly for antagonists of Class A receptors with great success, such as angiotensin 2 receptor antagonists, peptides still dominate the field of gonadotropin‐releasing hormone (GnRH) modulators. In addition, peptide analogs dominate the Class B family of G protein‐coupled receptor therapeutic agonists. Notoriously, the incretin family with GLP‐1 and GLP‐2 agonists has still not seen a small molecule competitor on the market. Furthermore, several attempts to find small molecule MCHR1 antagonists have also failed because most of these small molecules that were extremely powerful at the receptor were hErg‐positive, and this association characteristic was very difficult to read (see discussion in Johansson and Löfberg11).

2. “Peptides are not orally available”: desmopressin, a synthetic analog of the naturally occurring vasopressin, has been developed as an orally disintegrating tablet, although oral availability is very low. Since its initial approval in Finland in 2005, desmopressin has been approved in more than 80 countries. More recently, Novo‐Nordisk has reported several positive clinical trials of its oral Semaglutide®, a once‐daily lipopeptide GLP‐1 analog.12, 13

3. “Peptides are not metabolically stable and are too short lived”: Semaglutide® is currently approved by both the EMA and FDA as a once‐weekly subcutaneous injection for the treatment of patients with type 2 diabetes.14 This astonishing half‐life extension is attributable to protection against proteolysis and binding of its lipid moiety to serum albumin. Another example is Degarelix®, a selective GnRH receptor antagonist. After subcutaneous injection, it forms a gel from which the drug is released over a period of 1 month, achieving testosterone levels corresponding to medical castration in 97–98% of patients.15 Another example is Lanreotide, a somatostatin analog initially formulated as a PLGA nanoparticle formulation (Somatuline LA) that provided 10 days of coverage. The observation that at high concentration of peptide became an amyloid semisolid led to the approval of Somatuline autogel, which provides 42–56 days of coverage following subcutaneous injection.16

4. “Peptides require multistep synthesis and cannot be produced cost‐effectively on a large scale.” This led to the assumption that only highly active peptides, active at microgram doses, can be cost‐effective. For example, teriparatide, a 34‐amino‐acid peptide approved in 2002 for the treatment of osteoporosis, was produced by recombinant technology. Dosed at 20 μg once a day, it required only 7.3 mg per patient per year. The next year, the 34‐amino‐acid HIV fusion inhibitor enfuvurtide was approved. It requires 90‐mg injections twice a day, corresponding to 66 g of peptide per patient per year. Major improvements in both solid‐phase synthesis and RP‐HPLC purification allow large batches to be made and make the active pharmaceutical ingredient (API) available at an affordable cost.17

Thus, more and more studies are being published to describe longer and longer peptides. Because FDA approvals go together with up‐to‐date technical characterizations of a drug candidate, particularly biologics, more and more techniques have been developed to characterize these chemical entities. More and more evidence strongly suggests that Malcolm's statement9 was far from what really happened over the last few years and will happen in the coming decades.

Thoughts of the Agencies

Interestingly, a chemically synthesized polypeptide is not defined as a “biological product” and will be regulated as a drug, regardless of its length. Due to continuous improvements in synthetic methods and analytical tools, chemical synthesis appears to be a viable approach for manufacturing generic peptide drugs. This led the FDA to clarify its definition of biological products in the Biologics Price Competition and Innovation Act of 2009 (https://www.fda.gov/downloads/drugs/guidances/ucm444661.pdf). In particular, the definition of a “biological product” was amended to include “a protein (except any chemically synthesized polypeptide).” In the absence of scientific consensus, the FDA decided to base the statuary distinction on size only. According to these regulatory definitions, a “protein” means any alpha amino acid polymer with a specifically defined sequence >40 amino acids in size (the total number of amino acids is not limited to the number of amino acids in a contiguous sequence); thus, peptides <40 amino acids in size are excluded independent of their mode of production (synthetic or recombinant). This definition excludes insulin but includes peptides such as glucagon, liraglutide, nesiritide, teriparatide, and teduglutide (of rDNA origin). Current peptide synthesis technologies are a viable alternative to producing generic copies of these drugs that, due to the availability of orthogonal techniques, can be characterized extensively (https://www.fda.gov/Drugs/ScienceResearch/ucm578111.htm). The FDA now considers it possible to demonstrate that the active ingredient in a proposed generic synthetic peptide is the same as the reference active ingredient of rDNA origin, allowing the submission of an abbreviated new drug application. A “chemically synthesized polypeptide,” according to the FDA definition, is not a “biological product” and will be regulated as a drug under the Federal Food, Drug, and Cosmetic (FD&C) Act. Although several definitions exist, the FDA interprets the statutory exclusion for “chemically synthesized polypeptide” to mean any molecule that is made entirely by chemical synthesis and composed of up to 99 amino acids. Such molecules will be regulated as drugs under the FD&C Act. A “chemically synthesized polypeptide” composed of more than 99 amino acids, according to these definitions, will be considered a biological product. Thus, the FDA is prepared to examine polypeptides produced by chemical synthesis that contain between 41 and 99 amino acids as drugs and not biologics. This acknowledges the current and future achievements of polypeptide chemical synthesis and characterization.

Peptides? Which Structure, Which Refolding?

Roughly half a decade ago, we decided to enter a series of programs aimed at synthesizing a series of proteins of growing length from ubiquitin18 and calstabin19 to quinone reductase (which contains 226 amino acids and with which we failed). The difficulties encountered in developing a universal methodology to obtain such proteins, most of which were enzymes, led us to limit ourselves to ~200 amino acids. The most advanced program concerned calstabin, a 120 amino‐acid proline isomerase. We successfully obtained it by solid‐phase synthesis after struggling with multiple steps of native chemical ligation solutions. In brief, this native chemical ligation technique20 involves the chemoselective conjugation between a pair of unprotected peptide fragments, one functionalized as a C‐terminal thioester and the other with an N‐terminal cysteine (Cys) residue leading to a conjugate dipeptide in a single step after several steps, including a rearrangement, leading to the recovery of the initial cysteine side chain. Although the original description was limited to cystein‐containing peptides, it permitted the development of a series of techniques that are less limited to this point of view (see Conibear et al. for a complete review21). The process that we used for these chemical approaches was the standard process used for shorter peptides over the last decade.22 We treated the large, unfolded peptide like a regular peptide: precipitation, lyophilization, purification by RP‐HPLC, and MS analysis. We obtained a unique liquid chromatography (LC) peak, strictly symmetrical, strongly suggesting that the peptide was pure and ready to enter the refolding process. We realized that nothing can be farther from the notion of purity than such a “simple” observation of the RP‐HPLC profile and/or mass spectrogram. The refolding process, used on the basis of several years of refolding recombinant proteins expressed in insoluble fractions of bacteria, led to a mere 50% of the total protein being correctly refolded. Separation onto a gel filtration column led to two pools: one active and the other not.19 These two fractions could not be distinguished from one another using these analytical approaches. A closer analysis with circular dichroism (CD) spectrometry showed a slight difference between both, whereas MS, as expected, gave a single mass corresponding to the desired product. Furthermore, modification of the CD spectrum of the mixture due to the presence of the unfolded portion was not significantly altered compared to the pure, recombinant sample.19

After many years in the peptide area, these results were unexpected, as most reports indicate minimal analytical characterization of the samples. For long peptides, generally >40 amino acids, the main characterization steps performed these last few years are MS and CD. Table 1 provides some of the reports on “long” peptides randomly picked over the last three decades as examples of the way such synthetic peptides were characterized over the years. Furthermore, in no way this table could be considered as exhaustive. From the research standpoint, though, the situation is less constrained than from the therapeutic one. A close examination of Table 1 shows that peptides followed the historical changes in technical skill, with more and more characterizations in recently published papers. Nevertheless, due to the lack of obligations and rules, it is not rare to find papers describing peptides with minimal characterization of the peptide (obtained chemically) when it is RP‐HPLC/MS pure and the expected biological activity is at the rendezvous because the peptide is pure and forms a homogeneous entity. For example, the folding of the synthetic Anaphylatoxin C3a was demonstrated because the peptide has the same activities than the commercial (recombinant) one on two biological tests.52 None of those methods deliver any information on the peptide conformation (3D structure). However, in contrast to recombinant expression, in peptide synthesis, no chaperone is present to assist in the refolding of the nascent peptide chain. Thus, refolding is either spontaneous (e.g., ubiquitin18) or a long and partial process in which the end product must be thoroughly characterized. As a particularly interesting example, the recent paper by Kuroha et al.69 on a new lasso peptide essentially aimed at describing the structure of the material. Such papers are not easy to publish, essentially because they do not “tell a story,” but they are important to notice, as they pave the road to a standard procedure leading to accurately described pure peptides. Of course, other complete characterizations of synthesized peptides have been reported (e.g., Wei et al.44 for the HNP1 peptide analogs or Henriques et al.43 for LEAP2). In these two instances, the peptides were even crystallized.

Table 1.

Examples of chemically obtained peptides and their characterization(s)

Year	Peptide/protein	Reference	Number of amino acids	Biophysical characterizationa	Biological assay/activity
1983	hPTH	23	84	LC	Functional binding
1988	TGF alpha	24	~15 fragments	LC/tlc	Cell growth
1992	Elafin	25	57	LC	Enzyme inhibition
1992	NPY	26	36	LC	Binding
1992	Calciseptine	27	60	LC	Channel blocking
1994	IL8	20	72	LC/MS	NRb
1994	Ubiquitin	28	76	Crystallization	NR
1994	Ubiquitin	29	76	MS	NR
1996	Midkine	30	121	LC	Neurite extension
1997	Secretoneurin	31	33	NMR, CD	NR
1998	TSR and EGF1 modules	32	29 and 41	LC/MS, CD	NR
1998	GFP precursor	33	238	Fluorescencec
1999	Peptide E	34	25	CD	Anesthesia
2001	Octadecaneuropeptide	35	18	LC/MS, NMR	Calcium
2002	Transthyretin	36	127	LC/MS, NMR	NR
2005	26Rfa	37	26	CD, NMR	NR
2006	Seleno‐glutaredoxin 3	38	82	LC	Activation of reduction rate
2008	CGRP	39	37	LC/MS	Binding
2009	Kisspeptin‐52	40	52	LC/MS, NMR	Calcium
2010	Polytheonamide B	41	48	LC, NMR	Cellular toxicity
2010	Insulin	42	51	LC/MS, NMR	Binding
2010	LEAP‐2	43	40	NMR	Antimicrobial among others
2010	HNP1	44	30	LC/MS, NMR, crystallization	Antimicrobial
2011	Tetraubiquitin	45	304	LC/MS	Enzymatic degradation
2012	EPO	46	166	LC/MS, CD	Colony formation
2012	D‐VEGF	47	101	LC/MS, Crystallization	Binding
2012	Cyclotide 2v	48	42	LC/MS	NR
2013	EPO	49	166	CD	Colony formation
2013	α‐Scorpio toxin OD1	50	65	Crystallization	Na+ Current
2013	PYP	51	134	LC/MS, absorbance	Fluorescence
2013	Anaphylatoxin C3a	52	77	LC/MS	Binding
2013	HGF	53	127	LC/MS, CD	Binding
2013	Dengue capsid Prot C	54	80	LC/MS, CD	Dimerization
2013	INSL3	55	46	LC/MS, CD	Binding
2014	Hepcidin	56	25	NMR (Supp. Inf.)	Ferroportin degradation
2014	GM‐CSF	57	127	LC/MS, CD	Cell proliferation
2014	Lucifensin	58	40	LC/MS, CD	Antimicrobial
2014	Influenza virus M2d	59	97	CD	Single channel current
2014	SUMO	60	91	LC/MS	Used as substrate of SUMO E1
2014	SUMO	61	96	LC/MSe	NR
2015	G‐CSF	62	174	SDS‐PAGE, MS	Cell proliferation
2015	Caenopore‐5	63	82	LC/MS, 1D 1H NMR, CD	Permeabilization
2015	Histone 2B	64	125	LC/MS, CD	NR
2016	HGF	65	85	LC/MS	NR
2016	d‐ASFV pol Xf	66	174	LC/MS, CD	d‐Polymerase activity
2016	NK1	65	180	LC/MS	NR
2016	AS48	67	70	LC/MS, CD	Antimicrobial
2016	K27 Ubiquitin	68	151	LC/MS, CD, crystallography	Biochemistry
2017	Subterisin	69	16	NMR, MS/MS	NR
2018	CIGB‐330	70	25	LC/MS	Cell proliferation
2018	GsMTx4	71	34	LC/MS, CD, crystallography	NR

^aIt should be reminded that LC and LC/MS are not structural techniques. The information issued from such experiments is only used to purify the material. NMR, CD, and crystallization could be considered as structural indicating techniques.

^bNR means not reported in this particular reference.

^cIn this context, fluorescence is both a biophysical characterization and a biological activity (fluorescent protein).

^dThis is an integral membrane protein.

^eOnly the chemical conjugate between this protein and a peptide (RASIKAEGR) was analyzed for its structure.

^f d‐ASFV pol X stands for all D‐aminoacid African swine fever virus polymerase X.

Peptide Characterizations: What Is at Our Disposal?

The main point of peptide synthesis, especially since the use of robots became common, as well as the use of capping at every cycle of the step‐by‐step synthesis of such compounds, is that the sequence is fine, corresponding to a list of iterative steps entered into the robot by the operator and easily independently checked. In other words, the sequence entered should correspond to the desired sequence, with a minimal risk of mistake. Interestingly, the solid‐phase synthesis of peptides has covered a number of coupling sequences between almost any natural amino acids, which is still rare to find in short sequences (e.g., ~20 amino acid length), with massive failure in the succession of these coupling sequences, though we showed that the rates of coupling may be extremely different from one amino acid to another.72 Nevertheless, these observations concerned few, if any, exotic amino acids. We observed that coupling rates between natural and exotic or between exotic amino acids could be extremely slow. The possible incorporation of nonencoded amino acids into a pseudo‐peptide sequence remains one of the vastest possibilities of peptides, with no limit in terms of the nature of exotic amino acids that can be included. In contrast, using recombinant techniques, the incorporation of amino acids of various structures is possible, but limited to two or three different amino acids per sequence due to the limitations of genetic code manipulations.73, 74 Thus, for the analysis of the newly formed peptides, the routine initial step is purification via chromatography, up to a point in which the main peak in RP‐HPLC is purified and usually corresponds to the desired product. MS added the qualitative notion that the overall mass of the purified product corresponds to the theoretical mass, as calculated by the succession of amino acids in the sequence. The next two steps are verification of the peptide sequence and assessing its conformational homogeneity. The landscape for full amino acid sequence peptides dramatically evolved during the last decade with two well‐established techniques, both based on MS: fully automated amino acid sequence verification with LC–MS/MS of digest developed in the context of proteomic analyses, and conformational/folding information by ion mobility MS (IM‐MS) developed in the field of native spectrometry.75

The input of proteomic analysis methods

The development of proteomic analysis provided a very powerful tool that allows the rapid and nonambiguous full sequence verification of a given peptide or protein. The LC–MS/MS analysis performed on a series of digestion products using different enzymes allows determination of not only the amino acid sequence, but also the presence of exotic amino acids. The power of this method is well illustrated by its routine use in sequence verification of monoclonal antibodies containing ~1500 amino acids and establishing the heterogeneity of N‐glycosylations for the evaluation of the consistency of batch‐to‐batch production. A good example is the quantitative mass spectrometry multi‐attribute method (MAM)76 widely used by many manufacturers for the full characterization of monoclonal antibodies. A single amino acid substitution can be detected.77 Thus, the tools assessed for the recombinant approach, particularly medicinal antibodies, can be applied to the much simpler situation of synthetic peptides, even if their sizes are in the 100 amino acid range. Amino acid sequence determination based on proteomic analysis78 can easily be adapted to synthetic peptides <100 amino acids and provides in depth characterization when used in conjunction with already well‐established techniques, such as chromatography with absorbance or mass spectroscopy detection. In addition, they yield information on both the purity of the desired peptide and the possible impurities present in the preparation. This is well illustrated in the case of identifying and quantifying hundreds of contaminating host cell proteins present at ppm in recombinant proteins.79 However, the most basic tool is gel filtration because it can be used in a preparative mode. Because of the potential unfolded nature of some of the molecules underneath the purified product, the behavior of these species will be different in a simple gel filtration analysis, as we demonstrated previously with our synthetic version of the protein calstabine.19

After sequence verification, detection, quantification, and identification of possible impurities, one cannot claim that the peptide is active in whatever biological assay, because one does not know unequivocally the quantity of peptide that is folded appropriately. Several conformations are possibly formed for a given sequence, together with the completely, nonfolded linear peptide, and the differences between these cannot be appreciated by either RP‐HPLC or basic MS. Various tools exist for such structural characterization. The most common ones involve NMR, with the major limitation that these experiments require a large amount of material. Nevertheless, observation of the shapes of the peaks in the spectrum of a given peptide will tell if the compound is unique, regardless of whether it is well‐folded, or if the preparation is composed of several entities more or less folded. Despite providing a unique method for assessing the identity of a peptide, it is questionable whether it will be able to detect small amounts of micro‐heterogenic impurities.

Another possibility is the use of CD. Again, the main obstacle for the general use of this technique will be the need of a control peptide, such as a known, well‐folded peptide of the same sequence. Studying the CD spectra will reveal the various forms the peptide has adopted (e.g., barrels, coils, etc.). This technique also requires quite a large amount of product. In addition to these well‐established methods, IM‐MS can provide crucial information.

Conformation information by IM‐MS

This method provides a measurement of the collision cross section (CCS) of a molecule in the gas phase. This CCS is obviously linked to the size and shape of the molecular ions in the gas phase. If experimental parameters are selected carefully, the gas phase CCS can provide reliable information on the conformation in the liquid phase. For example, IM‐MS allowed the characterization of peptide−synthetic polymer conjugates,80 the study of conformer preferences for hydrophobic antimicrobial peptides,81 the identification of lasso peptide topologies,82, 83 and the separation of d‐amino‐acid‐containing peptides.84 IM‐MS can even be coupled with 1D or 2D LC as shown for the characterization of antibody–drug conjugates,85 in which initial separation by size is achieved by hydrophobic interaction chromatography.

Crystallization

There is a common belief that peptides are not easy to crystalize, though crystallization is the ultimate way to acquire information on their 3D structures, as opposed to the more cumbersome ways of gaining information by NMR. We provide a series of examples of such reports mostly extracted from the Protein Data Base (www.rcsb.org) in Table 2, ranging in length from 20 to 106 amino acids. This approach does not solve our problem (segregation between unfolded and folded peptides) because only the final product (in the crystal) will be present, by definition, and will show one or the other of possible forms that “accepted” to crystalize. Furthermore, crystallization is itself a method of purification.

Table 2.

Examples of crystallographic data on peptides (extracted from the Protein Data Bank)

PDB ID	Name of the peptide (if any)	Date	MW	Residue count	Primary citation author	Reference
1COI	COIL‐VALD	1996	3302.8	31	Ogihara et al.	86
1ALG	P11	1997	2442.84	24	Nordhoff et al.	87
1ZDC	Stable miniprotein A domain, Z34C	1997	4188.71	35	Starovasnik et al.	88
2A3D	Protein (de novo 3‐helix bundle)	1999	8120.21	73	Walsh et al.	89
1E0M	WW‐prototype	2000	4358.77	37	Macias et al.	90
2JWU	Gb88	2007	6457.42	56	He et al.	91
2KIR	Designer toxin	2009	3890.8	34	Takacs et al.	92
2FD7	[V15]crambin	2009	4707.0	46	Bang et al.	93
3OVJ	KLVFFA hexapeptide segment from amyloid beta	2010	3716.48	24	Landau et al.	94
2L96	LAK160‐P7	2011	2666.54	24	Vermeer et al.	95
4H8M	CC‐Hex‐H24‐A5/7C	2012	6648.03	64	Zaccai et al.	96
2LR2	Immunoglobulin G‐binding protein A	2012	9875.64	88	Barb et al.	97
2MT8	MTAbl13 of grafted MCoTI‐II	2014	4148.76	39	Huang et al.	98
2MSQ	Conotoxin cBru9a	2014	2809.19	27	Akcan et al.	99
4P6K	Computationally designed transporter of Zn(II) and proton	2014	3016.33	26	Joh et al.	100
4D5M	Triptorelin	2014	5534.81	44	Valéry et al.	101
4OWI	p53LZ2	2014	7861.36	66	Lee et al.	102
5ET3	Fullerene organizing protein (C60Sol‐COP‐3)	2015	7097.77	60	Kim et al.	103
–	[V15 A]crambin	2015	4708.5	46	Tang et al.	104
5KWX	Designed peptide NC_EEH_D1	2016	58989.76	25	Bhardwaj et al.	105
2NAU	Entity	2016	3283.89	28	Datta et al.	106

Conclusions

A trend has emerged in government agencies recommending that the regulation of market‐oriented peptides (and proteins, including antibodies) be reinforced, with more validation of sequences and minute details of structures. Although these recommendations are peripheral to research, one must realize that many peptides synthesized/discovered these last few years have been poorly described from the point of view of their structure.

We stress that, at the research level, the characterization of long peptides should be enforced imperatively in order not to try/test preparations in which only an unknown portion of the peptide is properly folded and thus active. It is clear for us that the introduction of proteomic methodologies and IM‐MS offers a rather easy way to produce batches of peptides with much better characterization than in the last decade.