ChapterPDF Available

Chapter 11. Determination of the Identity, Content and Purity of Therapeutic Peptides by NMR Spectroscopy


Abstract and Figures

This chapter describes the application of NMR spectroscopy to the determination of the identity, content, structure and purity of peptides. Both 1D and 2D NMR examples are presented for both ¹H and other nuclei. NMR spectroscopy is a nondestructive technique requiring limited sample pretreatment and development time. The limited sample pretreatment reduces the experimental time and variations due to handling, thereby increasing robustness. NMR spectroscopy can distinguish between very closely related peptides. It is especially suited for the determination of unrelated impurities such as process-related impurities and extractables/leachables. NMR spectroscopy is very sensitive to higher-order structure. Although NMR equipment in itself is expensive, the actual cost of an NMR spectrum is low owing to automation and limited consumables.
Content may be subject to copyright.
Determination of the Identity,
Content and Purity of
Therapeutic Peptides by
NMR Spectroscopy
Aspen Oss BV, Kloosterstraat 6, P.O. Box 98, 5340 AB Oss, The Netherlands;
Swedish Medical Products Agency, P.O. Box 26, 751 03 Uppsala, Sweden,
11.1 Introduction
NMR spectroscopy has been applied to peptides for a very long time and
some rather old and perhaps slightly outdated, but still very useful, books
on the topic are available.
NMR spectroscopy, together with mass spec-
trometry, is the most powerful technique for structure elucidation. In con-
trast to mass spectrometry, NMR spectroscopy is a nondestructive technique
that also yields information on the three-dimensional structure of peptides.
In recent decades, multidimensional, inverse pulsed field gradient stimu-
lated heteronuclear NMR was developed for proteins to increase both sen-
sitivity and resolution.
These techniques are equally applicable to peptides.
In this chapter, we limit ourselves to analytical and quality control (QC)
applications of 1D and 2D
H and
C NMR spectroscopy to small (o100
amino acids) therapeutic peptides. Reviews on NMR spectroscopy using less
Drug Discovery Series No. 72
Peptide Therapeutics: Strategy and Tactics for Chemistry, Manufacturing and Controls
Edited by Ved Srivastava
rThe Royal Society of Chemistry 2019
Published by the Royal Society of Chemistry,
abundant nuclei such as
have appeared elsewhere, in addition
to reviews on the 3D structure determination of proteins.
Articles on the
P NMR spectroscopy of phosphorylated peptides
F NMR spec-
troscopy of fluorinated peptides and proteins have also been published.
F NMR spectroscopy will also sensitively detect trifluoroacetate.
11.2 Peptide Identity
Currently, a number of monographs for synthetic peptides included in the
European Pharmacopoeia (Ph. Eur.) require amino acid analysis (AAA) ac-
cording to Moore and Stein
as an identity test. Recently, NMR spectroscopy
was introduced as an alternative method in the Ph. Eur.
The United States
Pharmacopeia (USP) contains a method for oxytocin identification by NMR
These NMR methods are meant as an alternative to AAA
as an identity test for peptides. Briefly, AAA entails the acid hydrolysis of a
peptide to yield the related amino acids followed by chromatographic sep-
aration, ninhydrin derivatization and detection.
The relative amount
(ratio) and (absolute) content of amino acids in a peptide can be determined
to be used both as an identity test and to quantify peptides. Some drawbacks
of this method are the degradation of amino acids by the harsh hydrolysis
conditions (typically overnight at 100 1C) and the laborious and time-
consuming sample preparation.
On the other hand, NMR analysis in solution is widely recognized,
accepted and established as a powerful tool for the structural elucidation
and characterization of peptides and proteins.
1,2,14 1
H and
C NMR spectra,
for instance, readily distinguish all 20 naturally occurring amino acids and
also unnatural amino acids present in synthetic peptides.
H NMR spectrum of a peptide depends primarily on the relative
amounts and identities of the amino acids present in the sequence, i.e. the
same information as obtained from an AAA.
H and
C NMR spectra readily distinguish all 20 naturally occurring
amino acids and unnatural amino acids present in synthetic peptides and
the random coil chemical proton chemical shifts in water have been
Quantification of the proton signals of peptides is relatively
straightforward. The intensity of the proton signal is directly proportional
to the relative number of protons at a particular position in a peptide.
a result, the proton NMR spectrum of a peptide depends primarily on the
relative amounts and identities of the amino acids present in the sequence.
This information concerning the identity of a peptide is the same as that
obtained from an AAA. Therefore, from a scientific point of view, NMR
spectroscopy is a suitable alternative to AAA.
H NMR spectroscopy at 400 MHz was recently introduced as an identity
test for eight peptides that are described in Ph. Eur. Monographs (Table 11.1).
The identification consists of a straightforward and simple comparison with
the spectrum of a (well-characterized) peptide standard and no calculations
are performed. Even peptides that differ only very slightly can be readily
382 Chapter 11
distinguished. Moreover, since no hydrolysis is performed prior to analysis, a
distinction can be made between Glu and Gln, Asp and Asn, something that
AAA is not able to do. Also, amino acids prone to hydrolysis and/or oxidation,
such as Trp and Cys in AAA, can be detected without any problems by NMR
spectroscopy. Table 11.2 summarizes the benefits and drawbacks of NMR
spectroscopy and AAA.
If we only consider peptides containing the 20 naturally occurring L-amino
acids, there is a staggering number (20
) of potential decapeptides, rising to
potential decapeptides if D-amino acids are included. Obviously, a strongly
discriminating technique is therefore required to distinguish between these
for identification. One-dimensional
H NMR spectroscopy is extremely sensi-
tive to composition sequence and is at least in principle capable of
Table 11.2 Comparison between NMR spectroscopy and AAA. Reproduced from
ref. 11 with permission from rPharmeuropa Scientific Notes, 2008-1.
Method Advantages Drawbacks
AAA Widely available; classical, accepted
technique; also applicable for
larger peptides/proteins
Destruction of Asn and Gln; Cys,
Met and Trp; hydrolysis conditions
differ from one peptide to another;
time-consuming and laborious
sample preparation
NMR General, robust technique; no
hydrolysis; data acquisition fast and
interpretation straightforward;
allows identification/quantitation
of proton-bearing organic counter
Less widely available
NMR analyses can, however, be readily outsourced to laboratories working under GMP for
prices in the order of hundreds of euros, i.e. far less than the price for outsourcing of a classical
Table 11.1 Composition of the peptides from Figure 11.1. Reproduced from ref. 13
with permission from rPharmeuropa Scientific Notes, 2008-1.
Peptide No. of AA
residues Sequence
Oxytocin 9 H–Cys-Tyr-Ile-Gln-Asn-Cys-Pro-Leu-Gly–NH
Desmopressin acetate 9 Mpa–Tyr-Phe-Gln-Asn-Cys-Pro-D-Arg-Gly–NH
Gonadorelin acetate 10 Glp–His-Trp-Ser-Tyr-Gly-Leu-Arg-Pro-Gly–NH
Gonadorelin diacetate 10 Glp–His-Trp-Ser-Tyr-Gly-Leu-Arg-Pro-Gly–NH
Buserelin acetate 10 Glp–His-Trp-Ser-Tyr-D-Ser(
Goserelin 10 Glp–His-Trp-Ser-Tyr-D-Ser(
Protirelin 3 Glp–His-Pro–NH
Tetracosactide 24 H–Ser-Tyr-Ser-Met-Glu-His-Phe-Arg-Trp-Gly-Lys-
Identity, Content and Purity of Therapeutic Peptides by NMR 383
discriminating between different sequences with the same amino acid com-
position whereas, in contrast, conventional AAA by its very nature is not.
Figure 11.1 demonstrates the strength of the
H NMR method by showing
the spectra of eight different peptides all recorded in the same way at
400 MHz, a moderate field strength widely available within the pharma-
ceutical industry. All peptides can be readily distinguished, even in the case
of goserelin and buserelin, two peptides showing only very small differences
in their sequence (Table 11.1). The method is robust and reproducible and
brought great simplification to the final release testing of peptides.
Na or
Cl NMR spectroscopy, counter ions such as tri-
fluoroacetate (TFA), sodium and chloride may be identified and quantified.
This is illustrated for TFA in Figure 11.3.
F NMR spectroscopy may also be
employed for the quantification of TFA.
A further example of the selectivity of NMR is presented in Figure 11.4, in
which a simple comparison of
H NMR spectra allows the differentiation of
the peptide goserelin acetate from its serine diastereoisomer named goser-
elin related compound A that differs from the former by the epimerization
of a single serine a-carbon (marked with an arrow in Figure 11.4).
Note the
signal dispersion and the differences in the spectra of the two compounds
(major differences are in the boxed areas). This is not feasible by con-
ventional AAA because no distinction is made between L- and D-amino acids
by the liquid chromatographic (LC) method.
AD/Lmodification of a single amino acid can result in large chemical shift
differences. For the pharmaceutical nonapeptide buserelin, distinct differ-
ences were observed in the
C-HSQC spectrum when L-His in buserelin
was changed to D-His. Spectra were acquired using B0.5 mg of material that
was dissolved in 0.2 mL of D
O–acetic acid-d
(8 : 2) mixture in a 3 mm NMR
tube. For the aromatic amino acids, large proton and carbon shift differ-
ences of about 0.2 ppm were observed for the histidine residue (Figure 11.5).
Small but distinct differences were observed for some of the tryptophan
and tyrosine signals. For the a-protons, shift differences of 0.00–0.07 ppm
were observed, the largest for the His and Tyr residues (not shown).
The assignment of
C chemical shifts of the buserelin D-His
stereoisomer was performed by a combination of standard NMR experiments:
H, two-dimensional
H-total correlation spectroscopy
(TOCSY) and rotating frame nuclear Overhauser spectroscopy (ROESY) and
C-heteronuclear single quantum coherence (HSQC)
and heteronuclear multiple bond correlation (HMBC). The amino acid com-
position was determined by evaluation of a-proton cross-peaks in the TOCSY
spectrum. For each amino acid, magnetization is transferred from the a-proton
to other protons in the amino acid side chain. Two different spin-lock delays of
30 and 60 ms were used. At the shorter delay, TOCSY cross-peaks were detected
mainly for the b-proton(s), while additional gand dcross-peaks were observed
at the longer delay. The sequence of amino acids was determined from inter-
residue carbonyl cross-peaks in the HMBC spectrum. Figure 11.6 clearly shows
that sequence information was detected between all residues except Arg-Pro.
384 Chapter 11
Figure 11.1 Aliphatic region of the 400 MHz
H NMR spectra of the eight peptides
listed in Table 11.1. The spectra were recorded in D
O in 200 mM
phosphate buffer at pH 5. Also the two salt forms of gonadorelin
(acetate and diacetate), using the areas of the acetate counter ion
(Table 11.2) based on the intensity of the acetate signal. This is illus-
trated in Figure 11.2. More generally, counter ion identification and
stoichiometry can be readily performed when nonexchangeable proton-
bearing counter ions are involved.
Reproduced from ref. 11 with permission from rPharmeuropa
Scientific Notes, 2008-1.
Identity, Content and Purity of Therapeutic Peptides by NMR 385
Fortunately, sequence information could be obtained from the ROESY
spectrum that showed distinct inter-residue through-space correlation peaks
between the Arg a-proton and the Pro b-protons (Figure 11.5). A mixing time of
200 ms was used.
Complete NMR chemical shift assignments are given in Table 11.3. All
carbons and nonexchangeable protons were detected and interpreted.
The systematic approach to determining the
H spin systems for every
amino acid, the C–H correlations at each position and inter-residual sequence
determination is a powerful way to determine the primary structure of a pep-
tide and also for modifications – and this without the need for reference data!
It was shown in 1988 that human, porcine and bovine insulins (51 amino
acids, differing in 1–4 amino acids; Table 11.4) can be readily distinguished
both by one-dimensional
H NMR (Figure 11.7) and by homonuclear
Figure 11.2
H NMR spectra of gonadorelin acetate (top) and gonadorelin diacetate
(bottom). The intensity of the acetate peak (singlet at about 1.9 ppm)
allows distinction between the two peptides.
Reproduced from ref. 11 with permission from rPharmeuropa Scien-
tific Notes, 2008-1.
Figure 11.3
F NMR spectrum of residual TFA in an aqueous solution of oxytocin.
A singlet peak at 76.5 ppm is observed. The NMR sample contained
6 mg of oxytocin material and 1.5 mg of TFA, corresponding to
0.25 mg g
. The spectrum was acquired on a 600 MHz instrument
equipped with a cryoprobe, the total experiment time was 4 min and
the signal-to-noise ratio was 200 : 1. Reasonably, 10-fold lower amounts
of TFA can be detected under these experimental conditions.
386 Chapter 11
Figure 11.4
H NMR (600 MHz, 95% H
O) spectrum of goserelin acetate and of a serine diastereoisomer (goserelin related compound A).
Regions with clear spectral differences are indicated by the boxes.
Identity, Content and Purity of Therapeutic Peptides by NMR 387
Figure 11.5 Overlay of the aromatic region of
C-HSQC spectra for D-His-buserelin (blue) and buserelin (red). The chemical structure
of buserelin is shown. Large
H and
C chemical shift differences were observed for the D- and L-isomers of the His residue.
Smaller differences were observed for some of the signals from the other aromatic amino acid residues.
388 Chapter 11
Figure 11.6 Informative two-dimensional NMR spectra for D-His-buserelin. Expansions from (a) the
C-HSQC (alpha region) and the
HMBC (carbonyl region) and (b) the alpha to aliphatic region of the
H-ROESY spectrum. Sequence information was
obtained from cross-peaks between the
C backbone carbonyl peak and the a-proton of the next residue. The ROESY
spectrum showed through-space correlation between Arg and Pro, as indicated by ellipses in (b).
Identity, Content and Purity of Therapeutic Peptides by NMR 389
Table 11.3 Interpretation of NMR data for D-His-buserelin in D
O–acetic acid-d
(4 : 1) at 25 1C.
Residue Shift, da
bgdOther CO H,H-TOCSY
C,H-HMBC from CO to
pyro-Glu-1 d
/ppm 3.90 2.09, 1.42 1.99 b2.09/1.42;
/ppm 56.4 24.9 28.9 174.3;
CO: 4.34/(3.90)/2.09/1.42
CO sc: 3.90/2.09/1.99/(1.42)
D-His-2 d
/ppm 4.34 2.72, 2.59 Ar: 8.08, 6.70 b2.72/2.59
/ppm 52.0 25.9 Ar: 133.1, 116.7,
127.8 170.9 CO: (4.43)/4.34/4.26/2.72/
Ar: 8.08/2.72/2.59; 8.08/
Trp-3 d
/ppm 4.43 2.97, 2.82 Ar: 7.28, 7.14, 6.90,
6.86, 6.80
/ppm 54.2 26.9 Ar: 117.9, 111.5,
121.6, 124.0,
119.0, 136.0,
126.7, 108.5
173.0 CO: 4.43/4.11/2.97/2.82
Ar: 7.28; 7.14; 7.28/6.86;
Ser-4 d
/ppm 4.11 3.45, 3.41 b3.45/3.41 CO: 4.15/4.11/3.45/3.41
/ppm 55.1 60.9 170.4
Tyr-5 d
/ppm 4.15 2.66, 2.63 Ar: 6.78, 6.51 b2.66/2.63 CO: 4.15/4.02/2.66/2.63
Ar: 2.66/2.63; 6.78/6.51;
/ppm 55.4 36.0 Ar: 130.2, 115.2,
154.5, 127.2 172.3
390 Chapter 11
/ppm 4.02 3.20, 2.92 tert-Butyl: 0.79 b3.20/2.92 CO: 4.06/4.02/3.20/2.92
tert-Butyl: 3.20/(2.92)/0.79
/ppm 53.5 60.7 tert-Butyl: 26.1, 74.4 171.2
Leu-7 d
/ppm 4.06 1.32, 1.23 0.59; 0.53 b1.32/1.23;
g0.59/0.53 CO: 4.26/4.06/(1.32)/(1.23)
/ppm 51.9 39.6 22.1; 20.3 173.6
Arg-8 d
/ppm 4.26 1.51, 1.39 1.32 2.84 b1.51/1.39; g
1.32; d2.84
CO: (4.43)/4.34/4.26/2.72/
Ar: 2.84
/ppm 50.8 27.5 39.6 40.4 Ar: 156.3 170.9
Pro-9 d
/ppm 3.95 1.87, 1.54 1.69 3.40,
b1.87/1.54; g
1.69; d3.40/
CO: (3.95)/2.85/1.87/1.54
/ppm 60.6 29.1 24.3 47.6 173.2
/ppm 2.85; 0.76
/ppm 34.2; 13.3
Atom position.
Correlations from Hathat reports on amino acid identity.
Inter-residual ROE from Ha.
Inter-residual HMBC in bold and weak interactions in parentheses.
C-terminal modification.
Identity, Content and Purity of Therapeutic Peptides by NMR 391
two-dimensional NMR at a relatively moderate field strength of 360 MHz.
The sequence features for a selection of insulins are summarized in Table 11.4.
As indicated in Table 11.4, there are differencesbetween human, porcine, and
bovine insulins. Glargine and lispro are human insulin variants, with identical
Ala and Thr distribution. Despite this, the small structural differences between
various insulin types are reflected in the Ala and Thr H
region of the
correlation spectroscopy (COSY) spectra (Figure 11.8).
The insulin materials that were used in Figure 11.7 were prepared in
O–acetic acid-d
solvent mixture. Similar spectra were also obtained when
was used, as shown in Figure 11.8. The pattern of the
H and 2D
H correlation spectroscopy (COSY) spectra of human,
porcine, bovine, glargine and lispro insulin clearly show differences.
Human, glargine, and lispro insulins have very similar peptide sequences,
differing only at few positions: A21 (Asn vs. Gln) for human vs. glargine, and
B28-B29 (Pro-Lys vs. Lys-Pro) for human vs. lispro. Glargine insulin also has
two additional Arg residues at positions B31 and B32. Despite the structural
similarity, the COSY spectra of these materials show distinct differences for
reliable identification of the respective insulin type.
In conclusion, distinction between various insulin types is possible in
short time from standard NMR spectra on the material. Comparison to ref-
erence spectra is convenient if standardized sample conditions are employed.
Human, porcine and bovine insulins can be distinguished by comparison of
proton and/or COSY spectra. Notably, the expansions from the proton spectra
of human and glargine insulins are very similar, whereas the patterns in the
COSY spectra are distinctly distinguishable. The differences between proton
as well as COSY spectra of human and lispro insulins are small but fair for
both solvent mixtures.
11.2.1 Protamine Sulfate
Protamine sulfate from chum salmon is composed of four major peptides,
A, B, C and D (Table 11.5), each containing 30–32 residues of the amino
acids arginine, serine, proline, isoleucine, alanine, glycine and valine.
NMR spectroscopy can be used to determine the amino acid composition,
to obtain sequence information and for quantification of the total amount.
Furthermore, there are multiple possible sources of protamine sulfate and
the origin of the material can be determined by NMR spectroscopy.
Table 11.4 Position and presence of threonine and alanine residues in human,
porcine, bovine, glargine, and lispro insulin A and B chains.
Insulin species A-8 A-10 B-14 B-27 B-30
Human Thr Ile Ala Thr Thr
Porcine Thr Ile Ala Thr Ala
Bovine Ala Val Ala Thr Ala
Glargine Thr Ile Ala Thr Thr
Lispro Thr Ile Ala Thr Thr
392 Chapter 11
Figure 11.7 High-field region of the 360 MHz 1D
H NMR spectra of (a) human, (b) porcine and (c) bovine insulin. The species can easily
be distinguished and identified from the spectra.
Reproduced from ref. 22 with permission from John Wiley and Sons, Copyright r1988 Royal Pharmaceutical Society of Great
Identity, Content and Purity of Therapeutic Peptides by NMR 393
Figure 11.8 Methyl region of proton and COSY spectra for five insulins. The upper row shows spectra in D
O–acetic acid-d
at pH 3 and
37 1C, for (a) human, (b) porcine, (c) bovine, (d) glargine, and (e) lispro insulin. The lower row shows the corresponding
spectra (f)–(j) in D
2 : 1 mixture at pH 2.5–3 and 25 1C. Samples were prepared using 2–4 mg insulin
dissolved in 0.7 ml solvent mixture, and spectra were acquired on a 600 MHz instrument. The chemical shift was set to
0.00 ppm for internal TSP-d
394 Chapter 11
A simple test for the determination of the source of the protamine sulfate
material was proposed by Gucinski et al.
An assay was developed based on
the ratios of the alanine, glycine and arginine NMR peaks (relative to the
arginine C
proton signal). The test also confirms that no aromatic amino
acids are present, but does not report on the exact molar or weight ratio
between peptides A–D. Selected peaks in the proton NMR spectrum were
integrated (Figure 11.9) and the experimental integral ratios calculated. The
results are reported in Table 11.6, together with the proposed limits.
The determined ratios were well within the limits and showed that the
source of the material is protamine sulfate from salmon.
As can be seen in Table 11.6, the experimental ratio for arginine C
was lower than the expected ratio of 50%. When the minor arginine C
peak at 4.59 ppm (see below) was included in the Arg C
H area, an experi-
mental ratio of 50.0% was obtained (3.3 þ46.7%).
The amino acid composition and sequence information have been studied
by various kinds of homo- and heteronuclear two-dimensional NMR experi-
ments. Two protamine sulfate samples were prepared in D
respectively, at a concentration of 2.5 mg mL
. Spectra were acquired at 25 1C
on a 600 MHz NMR spectrometer equipped with a 5 mm H(F)CN cryogenically
chilled probe. The amino acids were confirmed by interpretation of side-chain
correlation peaks in the two-dimensional
C-HMBC spectra using the D
O sample. Sequence information was
obtained by interpretation of inter-residue correlations in the
C-HMBC spectra using the H
O sample (allowing
detection of exchangeable amide protons). A typical fingerprint of chum sal-
mon protamine sulfate is shown in the HSQC spectrum in Figure 11.10.
The assigned
H and
C chemical shifts were found to correlate well with
reference data for the respective amino acids in various proteins.
interpreted NMR data were in accord with the amino acid types and their
relative occurrence in the peptide sequences for chum salmon protamine
sulfate (Table 11.5): (i) major NMR peaks were from Arg, (ii) larger peaks
were observed for Ser, Pro and Gly and (iii) smaller peaks were observed for
the less frequent amino acids Val, Ile and Ala (see Figure 11.9, HSQC).
No additional signals from other amino acids were observed. Two or more
different sets of proton/carbon resonances (spin systems) were observed for
Arg, Pro, Ser and Val, reflecting the fact that these amino acids occur at
several and structurally different positions in peptides A–D. On the other
Table 11.5 Single-letter peptide sequences
from chum
salmon protamine sulfate.
Peptide Sequence
P, proline; R, arginine; S, serine; I, isoleucine; A, alanine; G, glycine;
V, valine.
Identity, Content and Purity of Therapeutic Peptides by NMR 395
Figure 11.9 Proton NMR spectrum of protamine sulfate from salmon in D
O solution. The source of the material was confirmed by
calculation of selected ratios using peak areas and the number of protons in the respective peak.
396 Chapter 11
hand, only one set of resonances was observed for Ile, Ala and Gly (except for
the additional carbonyl signal for Gly).
Sequence information was obtained from correlations between adjacent
amino acid residues: (a) nuclear Overhauser effect (NOE)/rotating frame
Overhauser effect (ROE) correlations from the a-proton (C–H) of one amino
acid to an amide proton (N–H) of the next residue and (b) HMBC correlations
from a carbonyl (C
O) of one amino acid to the a-proton (C–H) of the next
residue. This is illustrated in Figure 11.11.
The observed peptide backbone correlations were -Pro-Arg-, -Arg-Ser-,
-Ser-Arg-, -Val-Arg-, -Val-Ser-, -Pro-Val-, -Arg-Val- (from NOE/ROE) and
-Gly-Arg-, -Arg-Arg-, -Val-Ser-, -Gly-Gly-, -Ser-Ser- (from HMBC). The NOE/ROE
correlations for protamine sulfate are illustrated in the ROESY spectrum
shown in Figure 11.12. The arrows indicate inter-residue correlations that
were used for confirmation of the peptide sequence.
Correlations between different amino acids are illustrated in Figure 11.13,
in agreement with the structures of peptides A–D. No conflicting correlations
were observed in the NMR spectra. Note that the HMBC spectrum also allows
the detection of correlations between amino acids of the same type, e.g.
Arg-Arg. Such correlations are present in the HMBC spectrum but are not
illustrated in Figure 11.13.
11.3 Peptide Content by qNMR Spectroscopy
In addition to identity determination, quantitative NMR (qNMR) spec-
is an excellent technique to determine peptide content.
For the
content determination of, e.g., reference standards, qNMR spectroscopy is a
well-established method described in both the Ph. Eur. and the USP.
tide content can be evaluated relative to an unrelated but well-characterized
standard. Related impurities may be distinguished, in contrast to, e.g., Kjel-
dahl, titration and amino acid analysis; however, for a reference standard,
material is usually selected that has undergone an additional purification.
Moreover, most methods applied for content declaration are nonspecific: im-
purities and degradants are not distinguished. Proton-bearing counter ions can
be quantified simultaneously. The variation is usually o1%. Limited method
development is required for qNMR for this generic method, making it very
suitable for peptides applied in personalized peptide-based treatment vaccines.
As an illustrative example, Figure 11.14 shows the determination of
gonadorelin acetate using a maleate standard. Five independent weighings
of both the peptide and standard were performed.
Table 11.6 NMR test for salmon source assay.
Amino acids No. of H ratio Area ratio/% Experimental
ratio/% Proposed
Ala C
H/Arg C
H 3 : 2 3.0 2.0 1.4–3.4
Gly C
H/Arg C
H 2 : 2 8.8 8.8 6.4–12.4
Arg C
H/Arg C
H n.a.
46.7 46.7 45–55
Not applicable.
Identity, Content and Purity of Therapeutic Peptides by NMR 397
Figure 11.10 Expansion of the two-dimensional
C-HSQC spectrum of protamine sulfate in D
O solution. The spectrum shows the
backbone a-H,C peaks for all amino acids (arginine, serine, proline, isoleucine, alanine, glycine and valine) and also some Ser,
Pro, Arg side-chain peaks. The corresponding proton spectrum is shown above the HSQC spectrum. The peaks are annotated
with the respective amino acid and side-chain positions 1 (a), 2 (b), 3 (g), 4 (d), ... Note that a certain amino acid type may
display multiple peaks, i.e. different chemical shifts, as a consequence of different positions in the peptide chains.
398 Chapter 11
Figure 11.11 NMR inter-residue correlations from NOE, ROE and HMBC experi-
ments provide sequence information.
Figure 11.12 Part of the 600 MHz ROESY spectrum for protamine sulfate in H
solution at 25 1C. Inter-residue correlations between the alpha and
amide hydrogens are indicated by arrows. These cross-peaks provided
valuable sequence information.
Identity, Content and Purity of Therapeutic Peptides by NMR 399
Figure 11.13 Observed inter-residue correlations in NOESY and ROESY spectra (red ellipses) and HMBC spectra (blue ellipses) for
protamine sulfate peptides A–D. No other, unexpected, correlations were observed in the NMR spectra.
400 Chapter 11
Figure 11.14 Part of the
H NMR (500 MHz, 95% D
O) spectrum of gonadorelin acetate and maleate displaying the signals and their
integrals used in content determination.
Identity, Content and Purity of Therapeutic Peptides by NMR 401
Using the equation
where P(sample) ¼searched purity (wt%), P(std) ¼purity of standard (wt%),
MW ¼molecular weight (g mol
), nH¼number of equivalent protons in the
NMR peak (1, 2, 3, ...), m¼weight of material (mg) and A¼area of NMR
peak (arbitrary units), a content of 89.6% (m/m) with an SD of 0.3% could be
calculated. The total experimental time was several hours.
This method can also be applied to assess the concentration of solutions
as shown in the example in Figure 11.15. Lanreotide was studied in solutions
with direct addition of an exact volume of an internal standard solution of
3-sulfolene. The content was determined as 1.02 mg per vial with an RSD
of 0.13% (n¼3). The labeled content was 1.016 mg.
The method is also applicable to larger peptides such as insulin, but some
challenges may be encountered. Insulin is a polypeptide with a well-defined
structure composed of two peptide chains named A and B. NMR sample
preparation is problematic, since insulin is prone to aggregation, which
results in poor NMR spectra with very broad signals. Several useful
approaches have been reported, e.g. using mixtures of water and an organic
solvent such as acetic acid
or acetonitrile
at pH 3. Solutions of
Figure 11.15 NMR spectrum used to determine the concentration of lanreotide
(structure shown at top right) vials applying 3-sulfolene as an internal
standard. The Tyr and Val signals are suitable for integration.
402 Chapter 11
monomeric insulin show well-resolved NMR spectra that permit quantifi-
cation. The a-protons of the two His signals of the B chain are adequately
resolved at a chemical shift of around 8 ppm. A problem is that the
exchangeable NH protons are also situated in this spectral region, resulting
in spectral overlap. Protons situated on nitrogen and oxygen are normally
exchanged by deuterium in a perdeuterated solvent such as D
O. For smaller
molecules, the H-to-D exchange process is usually immediate or very fast.
Unfortunately, this is not the case for insulin, particularly at lower pH.
This is shown in Figure 11.16, which illustrates that some of the amide ex-
changeable protons were not completely exchanged after 7 days.
The H–D exchange rate is much faster at neutral pH and it may be a good
idea to keep the solution at a higher pH for some time before pH adjustment
to a low pH.
When adequate H–D exchange was obtained, quantification was easily
performed using 3-sulfolene as an internal standard, as shown in Figure 11.17.
In principle, either of the two His signals could be integrated, but for this
sample His-B10 at 8.9 ppm was chosen owing to the smaller linewidth.
Residual ethanol could be determined simultaneously.
C NMR spectroscopy can also be applied to measure proline peptide
bond cistrans isomers, which is very difficult to do otherwise because of the
rapid isomerization.
Figure 11.16 H–D exchange for porcine insulin in D
CN (2 : 1) at pH 3, 25 1C
and 600 MHz. The protons signals of the B-chain His-10 (8.87 ppm)
and His-5 (8.72 ppm) are suitable for quantification. Very slow H–D
exchange was observed for the amide signals at 8.85 and 8.65 ppm
in spectra acquired immediately after sample preparation (a) and after
2 days (b), 3 days (c), 7 days (d) and 43 days (e).
Identity, Content and Purity of Therapeutic Peptides by NMR 403
Figure 11.17 Quantification of porcine insulin by
H NMR spectroscopy in D
(2 : 1). The purity was calculated as 91 wt%,
using a well-defined amount of 3-sulfolene as the internal standard. The sample contained about 2.5 wt% of residual ethanol.
404 Chapter 11
The major benefit of qNMR spectroscopy lies in its robustness and its
long-term precision,
and also its selectivity and general applicability.
Quantification of peptide mixtures such as protamine sulfate can be
performed by qNMR spectroscopy. The material consists of four slightly
different polypeptides A–D, with an undefined molar ratio between them. As a
consequence for the calulations, the molecular weight may vary between
4064 and 4319 gmol
, with an average of 4217 g mol
. The number of
amino acid residues varies between materials for Arg (21–22 residues) and Ala
(0–1 residues), while all four different peptides contain two Gly residues each.
This also affects the number of protons in the respective NMR peak to be used
for integration. The measurement error due to the unknown distribution of
peptides A–D is only 4toþ2% when the Gly signal is used, but slightly larger
when other amino acid peaks are used. In practice, integration of different
peaks in the proton spectrum resulted in an underestimated result when the
Ala peak was used, compared with when Arg or Gly peaks were used
(Table 11.7). The assay is reported as wt% of protamine for three independent
determinations, using 3-sulfolene as the internal standard. The content of
sulfate and water has to be added to the reported values.
The low value obtained when the Ala peak was used is probably a con-
sequence of less peptides A and/or D in the protamine material, as indicated
from analyses by high-performance liquid chromatography-mass spec-
trometry (HPLC-MS) and HPLC-UV detection. The values for Arg and Gly
were similar and reasonably accurate when compared with the results
obtained with LC-UV analysis.
Quantification may also be performed by 2D HSQC qNMR. In general,
2D spectra provide much better resolution than 1D proton spectra. A problem
with 2D experiments is that they are not truly quantitative, since various
compounds show differences in relaxation properties and coupling constants.
The method proposed by Hu et al.
compensates for the differences. The true
peak volume of each 2D peak can be obtained by extrapolation from three
HSQC experiments with incremented repetition times. Absolute concen-
trations for each compound can then easily be calculated by the use of an
internal standard. Poly(ethylene glycol) (PEG-4000) was selected as the internal
standard in order to allow a reasonably short relaxation delay between
repetitions (4 s) and thereby provide reasonably short measurement times
(4.5 h per experiment). The protamine sulfate concentration was 5–6 mg mL
Practically, three different HSQC experiments were acquired on each sample:
and HSQC
, where 1, 2 and 3 define the number of repetitions
Table 11.7 qNMR spectroscopy of protamine sulfate.
Alanine Arginine Glycine
Molecular weight/g mol
4217 4217 4217
NMR peak/ppm 1.40 3.20 3.98
Assay, mean/% 58.2 70.9 68.2
RSD/% 2.4 3.1 3.2
Identity, Content and Purity of Therapeutic Peptides by NMR 405
of the HSQC pulse sequence within one cycle. This procedure results in three
2D spectra with decreasing peak amplitudes (Figure 11.18).
For each compound, the respective peak amplitudes 1, 2 and 3 were
determined. HSQC
displayed weak peaks and the protamine content was
calculated using the peak volume and amplitude of HSQC
and the peak
amplitude extrapolated from all three experiments according to Hu et al.
The resulting protamine content as calculated from the glycine signal and
the arginine signal is given in Table 11.8 and compared with the respective
1D qNMR results. Similar results were obtained for 2D and 1D qNMR spectra
and also irrespective of the NMR signal that was used (Gly versus Arg).
2D qNMR spectroscopy may provide an alternative to chemometric
methods for calculation of content when the poor resolution in 1D NMR
spectra becomes an issue. The fact that a universal internal standard can be
employed means that the methodology can be applied in principle to any
material. Recently a survey of peptide quantification methods including
NMR and comparison of their reproducibility using oxytocin was pre-
It concluded that since qNMR has the advantages of simpler
Figure 11.18 HSQC cross–peaks for two different amino acids of protamine sulfate
and the internal standard PEG-4000. The peaks are used for calcu-
lation of protamine content. Shown are HSQC
(blue), HSQC
shifted upwards) and HSQC
(green, shifted upwards). The peak
amplitudes clearly decrease when the HSQC pulse sequence is re-
peated in HSQC
(two times) and HSQC
(three times). The use of 2D
spectroscopy considerably increases the distinction between peaks
compared with the proton spectrum.
Table 11.8 Protamine content determined by 2D qNMR spectroscopy
and comparison with 1D qNMR spectroscopy. Three repli-
cates were analyzed.
Gly Arg Gly Arg
Content/% 69.3 67.7 69.2 71.9
RSD/% 4.6 2.5 1.3 0.9
406 Chapter 11
operation and shorter analytical time, it deserves further exploration as a
primary method for peptide assay and that NMR would need to be intro-
duced as more readily in pharmaceutical quality control (QC) laboratories.
11.4 Peptide Purity
NMR spectroscopy is not the method of choice for routine impurity analysis of
peptides because of the large number of similar, related impurities at similar,
low levels potentially present. However, NMR spectroscopy is a powerful, uni-
versal detection method to identify unrelated and maybe unforeseen impur-
ities such as process-related impurities, leachables and extractables.
impurities may not be easily detected in peptides by HPLC analyses because
they lack UV absorbance and/or retention in the chromatographic methods
used. Small protonated impurities such as coupling agents and scavengers
used in deprotection can be sensitively detected because they tend to have
many equivalent protons resulting in sharp signals. Moreover, NMR spec-
troscopy yields a mole/mole response resulting in a high mass/mass response
of these small molecules relative to the larger peptide molecules. Moreover,
process-related impurities tend to be significantly smaller than peptides
and will therefore have longer relaxation times and faster diffusion, which
can be exploited for their detection.
These qualities can be employed
for their detection and characterization. Figure 11.19 shows the
spectrum of a monoclonal antibody obtained with and without employing a
Figure 11.19 Sensitive detection of process-related small-molecule impurities in a
solution of a monoclonal antibody (mAb) employing a CPMG sequence
to suppress the signals of high molecular weight resonances.
Reproduced from ref. 35 with permission from John Wiley and Sons,
Copyright r2012 American Institute of Chemical Engineers (AIChE).
Identity, Content and Purity of Therapeutic Peptides by NMR 407
Carr–Purcell–Meiboom–Gill sequence (CPMG) to suppress the signals of high
molecular weight resonances. In this way, 1 ppm (mass/volume) levels of
propylene glycol could be detected.
In contrast, pulsed field gradient experiments can be used to suppress the
signals of low molecular weight, rapidly diffusing molecules as illustrated in
Figure 11.20.
Another example in which a gradient stimulated echo pulse sequence
effectively suppresses low molecular weight compounds is illustrated in
Figure 11.21. The somatotropin spectrum is highlighted, while the strong
signals from the excipient, mannitol, and signals from other low molecular
weight components are suppressed.
The dipeptide N(2)-Ala-Gln is used for glutamine supplementation. A USP
monograph is available,
with limits for impurities cyclo-(Ala-Gln), alanine,
glutamine, Ala-Ala-Gln and Ala-Glu.
The proton spectrum of an Ala-Gln material that has been kept at room
temperature for some years showed a number of small peaks from unknown
compounds (Figure 11.22). These impurities could be specified or unspeci-
fied related substances, degradation products, process related compounds
or contaminants.
Figure 11.20 Suppression of the NMR signals of low molecular weight, rapidly
diffusing molecules by the application of pulsed field gradients. In
spectrum B, no gradient was applied and signals of low molecular
weight impurities are visible at B8.3 and 1.1 ppm. The application of
two opposite pulsed field gradients separated by a 50 ms delay sup-
presses the signals of rapidly diffusing, low molecular weight species.
Reproduced from ref. 38 with permission from American Chemical
Society, Copyright 1999.
408 Chapter 11
The methyl region around 1.5 ppm is suitable for a detailed impurity
evaluation, since methyl peaks generally are intense and not too split
(J-coupled). Low-power
C decoupling can generally be applied to simplify
spectra by eliminating
C satellites, as illustrated in Figure 11.23 for alanine
methyl signals.
Comparison with reference spectra of selected impurities showed that the
largest impurity doublet peak at 1.48 ppm could be either alanine or cyclo-
(Ala-Glu). The methyl signal of alanine residues is J-coupled to the respective
a-hydrogen and a
H-COSY spectrum provides valuable information
about the chemical environment of the particular alanine residue. This is
illustrated in the expansion of the COSY spectrum of Ala-Gln shown in
Figure 11.24. COSY is a fairly sensitive experiment and the second dimen-
sion permits the detection of impurity peaks that are hidden under larger
Figure 11.21 Proton NMR spectra of an aqueous solution of somatotropin medicinal
product. The ordinary water-suppression proton spectrum is shown in
(a) (red), showing mainly signals from mannitol at 3.5–4 ppm and several
sharp peaks from other low molecular weight compounds. The gradient-
edited spectrum is shown in (b) (blue). Notably, the somatotropin signals
are strongly enhanced in the lower spectrum, whereas the peaks of
mannitol and other compounds are greatly or completely suppressed.
Figure 11.22 Proton spectrum at 600 MHz of Ala-Gln in D
Oat251C. In addition
to the
C satellites on both sides of each main peak, a number of
impurity peaks can also be observed.
Identity, Content and Purity of Therapeutic Peptides by NMR 409
peaks in the one-dimensional proton experiment. For Ala-Gln, the most
abundant impurity showed a cross-peak at 1.48/3.78 ppm, in complete ac-
cord with reference NMR shifts for alanine. Smaller impurity cross-peaks
were observed at 1.43/4.37 and 1.36/4.15 ppm. Further evaluation of these
impurity peaks is possible by recording selective one-dimensional NOESY
or ROESY experiments, which may provide sequence information about
through-space H–H contacts from Ala to neighboring amino acids.
The cross-peak at 4.02/1.18 ppm is due to small amounts of 2-propanol.
Quantification of residual solvents is easily performed by comparison of the
integral of the solvent peak with that of the
C satellite peak of the main
compound. The natural abundance of
C is 1.1%, which means that each
C satellite signal corresponds to 0.55% of the main peak of the compound.
This results in the following equation for the quantification of residual
solvent (or any specified impurity) in the compound:
0:0055 nHðcompoundÞþAðsolventÞMWðsolventÞ
Figure 11.23 The upper spectrum shows the methyl region of Ala-Gln in D
O. In the
lower spectrum, the
C satellites have been quenched by low-power
C decoupling during the acquisition. This clearly simplifies the
evaluation with respect to the five different impurity peaks.
410 Chapter 11
where Ais the peak integral area, MW is the molecular weight (g mol
) and
nH is the number of hydrogens in the peak. For levels lower thanB100 ppm
the solvent term in the denominator can be omitted. For the present case,
the areas of 2-propanol and Ala-Gln were 13.5 and 100, respectively, which
resulted in a 2-propanol level of B100 ppm (Figure 11.25).
NMR spectroscopy can also be used to elucidate the structure of isolated
peptide impurities in conjunction with MS. Furthermore, peptide modifi-
cations difficult to elucidate by MS (e.g. diastereomeric impurities) may be
elucidated by NMR spectroscopy. In stability studies on syringes of a
Figure 11.24 The alanine methyl region from a 600 MHz COSY spectrum of Ala-Gln
in D
O solution at 25 1C. Four different cross-peaks related to impur-
ities are observed (see text), as indicated by asterisks.
Identity, Content and Purity of Therapeutic Peptides by NMR 411
Figure 11.25 Residual 2-propanol at 1.18 ppm in Ala-Gln. The solvent level can be determined from the integral ratio between the solvent
peak (six hydrogens) and the
C satellite peak of the compound (three hydrogens).
412 Chapter 11
pyridylalanine-containing peptide, an HPLC peak of a degradant impurity was
detected that increased with time. MS analysis of this peak show that it had
the same mass as the peptide itself. Based on this information, all diaster-
eomeric impurities of the peptide were synthesized, but none of them
coeluted with the degradant impurity. NMR analysis of a larger amount of
the isolated impurity immediately revealed, from the chemical shift of the
aromatic pyridyl signals, that the pyridyl nitrogen had reacted with residual
acrylic acid monomers from the glue in the syringe to form an N-pyridyl de-
rivative by a Michael condensation. Under the conditions of the mass spec-
trometer, this impurity reverted to the original peptide by a b-elimination.
This clearly demonstrates the strength of the nondestructive nature of
NMR spectroscopy. NMR spectroscopy is therefore an excellent technique for
root-cause analysis and troubleshooting rather than routine analysis.
11.5 Higher-order Structure
Most small linear hydrophilic peptides fail to adopt a defined three-
dimensional structure in water.
(Deuterated) trifluoroethanol,
protective solvents
and micelles
have been widely used to induce/
stabilize higher-order structures of peptides that otherwise would exist as a
random coil. Basically, four indicators of the presence of secondary and
tertiary structure in peptides can be detected by NMR spectroscopy:
1. Non-random coil chemical shifts (chemical shift dispersion):
chemical shifts are extremely sensitive to peptide conformation due to
anisotropy. An example is given in Figure 11.26.
Figure 11.26 Plot of the differences between the observed a-proton chemical shifts
and the corresponding random coil values versus the amino acid
sequence of glutaredoxin 3.
Reproduced from ref. 48 with permission from American Society
for Biochemistry and Molecular Biology, Copyright r1996 by The
American Society for Biochemistry and Molecular Biology, Inc.
Identity, Content and Purity of Therapeutic Peptides by NMR 413
2. Non-average three-bond Jcouplings, (chemical shift dispersion):
exclusively averaged Jcouplings are expected to occur in random coil
peptides. Non-random coil Jcouplings indicate secondary structure.
3. (2D) NOE/ROE contacts: these arise if protons are oB5 Å apart. If these
protons are more than one residue apart in the peptide sequence, this
suggests the presence of secondary and tertiary structure elements.
4. (Amide) NH and OH protons will readily exchange with deuterium in a
deuterated solvent containing exchangeable protons such as D
Residual signals due to unexchanged protons can be due to NH and/or
OH protons protected from exchange because they are involved in strong
hydrogen bonds or protected from exchange because they are inaccess-
ible to exchange since they are buried in the interior of the structure.
The determination of the three-dimensional structure of peptides and
proteins by heteronuclear, multidimensional NMR spectroscopy has reached
maturity and has been the subject of several books
and reviews
and is
outside the scope of this chapter. The reader is referred to the literature cited.
Aggregation of peptides has been evaluated by concentration dependence,
relaxation times and decreased diffusion rates using pulsed field
The 600 MHz
D spectrum of a 17 mg ml
solution of oxytocin shown
in Figure 11.27 displayed small, additional signals at ca. 7 ppm. These sig-
nals increased with concentration, which rules out proline cis/trans isomers
and/or impurities.
The 2D
H DOSY shown in Figure 11.28 obtained on the sample showed
that the additional signals corresponded to species with a longer diffusion
time. Also different NOE buildup curves were observed for the assumed
aggregate (Figure 11.29).
The concentration dependence, the increased diffusion time and NOE
buildups of the additional signals are all consistent with the presence of
aggregated oxytocin material in this samples.
11.6 Conclusion
This chapter has described the application of NMR spectroscopy as a com-
prehensive technique to determine the identity, content, purity and struc-
ture of synthetic therapeutic peptides, including a number of practical
examples. However, NMR spectroscopy still suffers from a number of
popular prejudices with ‘‘traditional’’ analytical chemists, regretfully limit-
ing its use. NMR spectroscopy is thought to be expensive, insensitive, slow,
difficult, specialist and not quantitative. However, the following points
should be taken into account:
Thanks to highly developed automation, 24/7 equipment operation.
Electronics can be upgraded.
414 Chapter 11
Figure 11.27 (a) The 600 MHz
H 1D spectrum of a 17 mg ml
solution of oxytocin. Additional, minor signals were noted at B7 ppm. In
(b), overlay spectra acquired at different concentrations show that the small signals at 7.03–7.10 ppm clearly increase with
increasing sample concentration, while other minor peaks remain unaffected. The Tyr main peak at 7.22 ppm was used for
Identity, Content and Purity of Therapeutic Peptides by NMR 415
Figure 11.28 2D
H DOSY spectrum obtained on the sample from Figure 11.27.
416 Chapter 11
Figure 11.29 NOE build-up curves for oxytocin in D
O at 600 MHz and 25 1C: Tyr-Ar (B7 ppm) to Tyr-CH
(B3 ppm). Peak heights for
monomer (orange) and aggregate (blue) cross-peaks at mixing times of 100, 150, 300, 500 ms. (a) shows absolute intensities
and (b) shows relative intensities that have been normalized against the auto-peak. Note that the normalized NOE factors are
negative by definition; close to zero for oxytocin monomer but distinctly negative for the aggregate.
Identity, Content and Purity of Therapeutic Peptides by NMR 417
Magnet lasts for decades.
No or limited sample pretreatment.
Analysis of intact samples without extensive method development.
In-depth knowledge not required for application of NMR spectroscopy.
Accuracy and precision as good as in chromatographic techniques.
Long-term robustness excellent.
No problems due to variations introduced by the use of different
No response factors required.
There are many examples of NMR detection of trace molecules where
NMR spectroscopy not only confers sensitivity but also selectivity.
NMR spectra can routinely be obtained on milligrams of sample but
also, with some more effort, on sample quantities several orders of
magnitude lower.
Many 1D and 2D NMR techniques can yield spectra within minutes due
to technological developments (e.g. cryoprobes).
Hence NMR spectroscopy is an excellent and general tool for the de-
termination of the identity, content and purity of therapeutic peptides. It is
expected that in the future other, more advanced NMR methods will be
implemented (e.g., multidimensional and/or heteronuclear) enabling further
application of the use of NMR in the analysis of therapeutic peptides and
their formulations.
1. K. Wu
¨thrich, NMR in Biological Research: Peptides and Proteins, North-
Holland, Amsterdam, 1st edn, 1976.
2. R. A. Dwek, Nuclear Magnetic Resonance in Biochemistry, ed. W. Harrington
and A. R. Peacock, Clarendon Press, Oxford, 1st edn, 1973, pp. 44–54.
3. J. Cavanagh, W. J. Fairbrother, A. G. Palmer III, M. Rance and N. J.
Skelton, Protein NMR Spectroscopy: Principles and Practice, Academic
Press, New York, 2nd edn, 2010.
4. J. G. Beck, A. O. Frank and H. Kessler, NMR of Biomolecules: Towards
Mechanistic Systems Biology, ed. I. Bertini, K. S. McGreevy and G. Parigi,
Wiley and Blackwell, Chichester, 1st edn, 2012, pp. 328–344.
5. B. Valentine, A. Steinschneider, D. Dhawan, M. I. Burgar, T. S. T. Amour
and D. Fiat, Int. J. Pept. Protein Res., 1985, 25, 56.
6. R. Hoffmann, I. Reichert, W. O. Wachs, M. Zeppezauer and
H. R. Kalbitzer, Int. J. Pept. Protein Res., 1994, 44, 193.
7. J. T. Gerig, Prog. Nucl. Magn. Reson. Spectrosc., 1994, 26, 293.
418 Chapter 11
8. P. Mykhailiuk, S. Afonin, G. V. Palamarchuk, O. V. Shishkin, A. S. Ulrich
and I. V. Komarov, Angew. Chem., Int. Ed. Engl., 2008, 47, 5765–5767.
9. S. Afonin, R. W. Glaser, M. Berditchevskaia, P. Wadhwani, K.-H. Gu
U. Mo
¨llmann, A. Perner and A. S. Ulrich, ChemBioChem, 2003, 4, 1151–
10. D. H. Spackman, W. H. Stein and S. Moore, Anal. Chem., 1958, 30, 1190.
11. European Pharmacopoeia 9.8 General Chapter 2.2.64 Peptide Identifi-
cation by Nuclear Magnetic Resonance Spectroscopy.
12. United States Pharmacopeia 41 Oxytocin and Goserelin Acetate
13. E. Kellenbach, et al.,NMR Spectroscopy in Pharmaceutical Analysis,ed.
U. Holzgrabe, I. Wawer and B. Diehl, Elsevier, Amsterdam, 1st edn, 2008,
ch. 5, pp. 429–436; E. Kellenbach, et al.,Pharmeuropa, 2008, 2008, 1–8.
14. K. Wu
¨thrich, NMR of Proteins and Nucleic Acids, John Wiley and Sons,
New York, 1st edn, 1986.
15. A. Bundi and K. Wu
¨thrich, Biopolymers, 1979, 18, 285.
16. D. S. Wishart and B. D. Sykes, Methods Enzymol., 1994, 239, 363.
17. G. Maniara, K. Rajamoorthi, S. Rajan and G. W. Stockon, Anal. Chem.,
1998, 70, 4921.
18. F. Malz and H. Jancke, J. Pharm. Biomed. Anal., 2005, 38, 813.
19. Y. B. Monakhova and B. W. K. Diehl, J. Pharm. Biomed. Anal., 2018,
154, 332.
20. A. Preiß, J. Kruppa, J. Buschmann and C. Mu
¨gge, J. Pharm. Biomed. Anal.,
1998, 16, 1381.
21. D. I. Freedberg, Y. Aubin and C. Jones, BioPharm Int, 2010, 6(Supplement
August), 28.
22. C. W. Funke, J.-R. Mellema, P. Salemink and G. N. Wagenaars, J. Pharm.
Pharmacol., 1988, 40, 78.
23. J. A. Hoffmann, R. E. Chance and M. G. Johnson, Protein Expression
Purif., 1990, 1, 127.
24. A. C. Gucinski, M. T. Boyne II and D. A. Keire, Anal. Bioanal. Chem., 2015,
407, 749.
25. U. Holzgrabe, R. Deubner, C. Schollmayer and B. Waibel, J. Pharm.
Biomed. Anal., 2005, 38, 806; M. Malet-Martino and U. Holzgrabe,
J. Pharm. Biomed. Anal., 2011, 55,1.
26. C. K. Larive, D. Jayawickrama and L. Orfi, Appl. Spectrosc., 1997,
51, 1531.
27. European Pharmacopoeia 9.8 General Chapter 2.2.33. Nuclear magnetic
resonance spectrometry.
28. United States Pharmacopoeia 41 General Chapter: o7614Nuclear
Magnetic Resonance Spectroscopy.
29. D. J. Craik and K. A. Higgins, Pept. Res., 1991, 4, 177.
30. A. D. Kline and R. M. Justice Jr., Biochemistry, 1990, 29, 2906.
31. C. Grathwohl and K. Wu
¨thrich, Biopolymers, 1976, 15, 2025.
32. W. de Wildt, H. Kooijman, C. Funke, B. U
¨n, A. Leika, M. Lunenburg,
F. Kaspersen and E. Kellenbach, Molecules, 2017, 22, 1362.
Identity, Content and Purity of Therapeutic Peptides by NMR 419
33. K. Hu, W. M. Westler and J. L. Markley, J. Am. Chem. Soc., 2011, 133,
34. C. Li, et al.,J. Pharm. Biomed. Anal., 2019, 166, 105.
35. K. Skidmore, D. Hewitt and Y.-H. Kao, Biotechnol. Prog., 2012, 28, 1526.
36. N. Magarian, K. Lee, K. Nagpal, K. Skidmore and E. Mahajan, Biotechnol.
Prog., 2016, 32, 718.
37. Y.-H. Kao, J. Bender, A. Hagewiesche, P. Wong, Y. Huang and
M. Vanderlaan, PDA J. Pharm. Sci. Technol., 2001, 55, 268.
38. E. Kellenbach, M. Burgering and F. Kaspersen, Org. Process Res. Dev.,
1999, 3, 141.
39. United States Pharmacopoeia 41. L-Alanyl-L-glutamine monograph.
40. E. Kellenbach and P. Dani, NMR in Pharmaceutical Sciences,ed.
J. R. Everett, R. K. Harris, J. C. Lindon and I. D. Wilson, John Wiley and
Sons, Chichester, 2015, ch. 31, pp. 441–451.
41. M. Buck, Q. Rev. Biophys., 1998, 31, 297.
42. A. Jasanoff and A. R. Fersht, Biochemistry, 1994, 33, 2129.
43. A. Motta, D. Picone, T. Tancredi and P. A. Temussi, Tetrahedron, 1988,
44, 975.
44. R. Spadaccini, O. Crescenzi, D. Picone, T. Tancredi and P. A. Temussi,
J. Pept. Sci., 1999, 5, 306.
45. D. Picone, A. D’Ursi, A. Motta, T. Tancredi and P. A. Temussi, Eur. J.
Biochem., 1990, 192, 433.
46. D. A. Kallick, M. R. Tessmer, C. R. Watts and C. Y. Li, J. Magn. Reson., Ser.
B, 1995, 109, 60–65.
47. M. Lindberg, J. Jarvet, U
¨. Langel and A. Gra
¨slund, Biochemistry, 2001,
40, 3141.
48. F. Åslund, K. Nordstrand, K. D. Berndt, M. Nikkola, T. Bergman,
H. Ponstingl, H. Jo
¨rnvall, G. Otting and A. Holmgren, J. Biol. Chem.,
1996, 271, 6736.
49. D. Wishart, Curr. Pharm. Biotechnol., 2005, 6, 105.
50. S. L. Mansfield, D. A. Jayawickrama, J. S. Timmons and C. K. Larive,
Biochim. Biophys. Acta, Protein Struct. Mol. Enzymol., 1998, 1382, 257.
51. S. L. Mansfield, A. J. Gotch, G. S. Harms, C. K. Johnson and C. K. Larive,
J. Phys. Chem. B, 1999, 103, 2262.
52. M. Iqbal and P. Balaram, Biopolymers, 1982, 21, 1427.
53. A. Phillips, Genotoxic Impurities: Strategies for Identification and Control,
ed. A. Teasdale, John Wiley and Sons, Hoboken, New Jersey, 1st edn,
2010, ch. 13, 351–384.
420 Chapter 11
ResearchGate has not been able to resolve any citations for this publication.
  • B Valentine
  • A Steinschneider
  • D Dhawan
  • M I Burgar
  • T S T Amour
  • D Fiat
B. Valentine, A. Steinschneider, D. Dhawan, M. I. Burgar, T. S. T. Amour and D. Fiat, Int. J. Pept. Protein Res., 1985, 25, 56.
  • R Hoffmann
  • I Reichert
  • W O Wachs
  • M Zeppezauer
  • H R Kalbitzer
R. Hoffmann, I. Reichert, W. O. Wachs, M. Zeppezauer and H. R. Kalbitzer, Int. J. Pept. Protein Res., 1994, 44, 193.
  • J T Gerig
J. T. Gerig, Prog. Nucl. Magn. Reson. Spectrosc., 1994, 26, 293.
  • P Mykhailiuk
  • S Afonin
  • G V Palamarchuk
  • O V Shishkin
  • A S Ulrich
  • I V Komarov
P. Mykhailiuk, S. Afonin, G. V. Palamarchuk, O. V. Shishkin, A. S. Ulrich and I. V. Komarov, Angew. Chem., Int. Ed. Engl., 2008, 47, 5765-5767.
  • S Afonin
  • R W Glaser
  • M Berditchevskaia
  • P Wadhwani
  • K.-H Gührs
  • U Möllmann
  • A Perner
  • A S Ulrich
S. Afonin, R. W. Glaser, M. Berditchevskaia, P. Wadhwani, K.-H. Gührs, U. Möllmann, A. Perner and A. S. Ulrich, ChemBioChem, 2003, 4, 1151-1163.
  • E Kellenbach
E. Kellenbach, et al., NMR Spectroscopy in Pharmaceutical Analysis, ed. U. Holzgrabe, I. Wawer and B. Diehl, Elsevier, Amsterdam, 1st edn, 2008, ch. 5, pp. 429-436;
  • E Kellenbach
E. Kellenbach, et al., Pharmeuropa, 2008, 2008, 1-8.
  • A Bundi
  • K Wüthrich
A. Bundi and K. Wüthrich, Biopolymers, 1979, 18, 285.
  • D S Wishart
  • B D Sykes
D. S. Wishart and B. D. Sykes, Methods Enzymol., 1994, 239, 363.