Glycine homopeptides: The effect of the chain length on the crystal structure and solid state reactivity

Article abstract of DOI:10.1039/c4ce00630e

A series of linear oligoglycines has been studied using X-ray diffraction, thermal analysis complemented with simultaneous FTIR spectroscopy, gas chromatography-mass spectrometry analysis, and ¹H/¹³C NMR. The new and previously reported data are rationalized to reveal the effect of the chain size on the crystal structure, molecular conformation and thermal stability of the oligopeptides in the solid state, as well as thermally induced transformations (pyrolysis) in the solid state and the gaseous phase. Tetraglycine (4G) and pentaglycine (5G) form triclinic crystals, P1, where they adopt a zwitterionic form and fully extended chain conformation. The antiparallel β sheet arrangement is similar to that in β-triglycine (β3G) and α-diglycine (α2G) but differs from the Form I of polyglycine (polyG). The stability sequence α2G < β3G < 4G < 5G has been attributed to the stabilizing environment in the solid phase. For all studied glycines, the thermal degradation is a kinetic process complicated by multiple parallel pathways and sensitivity to the experimental conditions. These pathways include defragmentation reactions (mainly deamination and decarboxylation) and condensation reactions (formation of new peptide bonds), the latter yielding cyclo-diglycine (c2G, or 2,5-diketopiperazine or DKP) as a major product. Further pyrolysis results in polymerization (solid samples) or further fragmentation (gaseous phase). The overall cyclization reaction α2G(solid) = c2G(solid) + H₂O(gas) is endothermic with ΔH = 63.7(4) kJ mol^-1 at 492 K. The standard enthalpy of the reaction α2G(solid) = c2G(solid) + H₂O(liquid) was estimated to be ΔH²⁹⁸ ≈ 18 kJ mol^-1 meaning the reaction is thermodynamically unfavorable under ambient conditions. Finally, the cyclization reaction mechanisms are discussed for the whole series studied. This journal is the Partner Organisations 2014.

Full text of DOI:10.1039/c4ce00630e

CrystEngComm
View Article Online
View Journal
PAPER
Glycine homopeptides: the effect of the chain
length on the crystal structure and solid state
reactivity†
Cite this: DOI: 10.1039/c4ce00630e
*
Aaron J. Smith, Farukh I. Ali and Dmitriy V. Soldatov
A series of linear oligoglycines has been studied using X-ray diffraction, thermal analysis complemented
with simultaneous FTIR spectroscopy, gas chromatography – mass spectrometry analysis, and 1H/13C NMR.
The new and previously reported data are rationalized to reveal the effect of the chain size on the crystal
structure, molecular conformation and thermal stability of the oligopeptides in the solid state, as well as
thermally induced transformations (pyrolysis) in the solid state and the gaseous phase. Tetraglycine (4G)
and pentaglycine (5G) form triclinic crystals, P¯1, where they adopt a zwitterionic form and fully extended
chain conformation. The antiparallel β sheet arrangement is similar to that in β-triglycine (β3G) and
α-diglycine (α2G) but differs from the Form I of polyglycine (polyG). The stability sequence α2G < β3G < 4G < 5G
has been attributed to the stabilizing environment in the solid phase. For all studied glycines, the thermal
degradation is a kinetic process complicated by multiple parallel pathways and sensitivity to the experimental
conditions. These pathways include defragmentation reactions (mainly deamination and decarboxylation)
and condensation reactions (formation of new peptide bonds), the latter yielding cyclo-diglycine (c2G, or
2,5-diketopiperazine or DKP) as a major product. Further pyrolysis results in polymerization (solid samples)
or further fragmentation (gaseous phase). The overall cyclization reaction α2G(solid) = c2G(solid) +
H2O(gas) is endothermic with ΔH = 63.7(4) kJ mol−1 at 492 K. The standard enthalpy of the reaction
α2G(solid) = c2G(solid) + H2O(liquid) was estimated to be ΔH298 ≈ 18 kJ mol−1 meaning the reaction is
thermodynamically unfavorable under ambient conditions. Finally, the cyclization reaction mechanisms
are discussed for the whole series studied.
Received 27th March 2014,
Accepted 3rd May 2014
DOI: 10.1039/c4ce00630e
www.rsc.org/crystengcomm
molecules in a desired mutual arrangement within the
solid state is a challenging task preventing wide spread adap-
Introduction
A new found interest in short peptides has stemmed from
the idea of utilizing these unique molecules in the design of
template matrices for solid state reactions. Solid state organic
synthesis (also known as “solvent-free synthesis”) has emerged
as a new domain of synthetic organic chemistry1 and to date,
almost all main types of organic reactions have been conducted
in the solid state.2,3 The main advantages of such synthetic
methods – simplicity and economy (solvent-free, often yield a
single product without a need for separation), high (usually
quantitative) yields, as well as specificity of the product
(including stereospecificity) – resonate well with the princi-
ples of Green Chemistry.4 However the placement of reactant
tation. One approach to resolve this issue is based on the
co-crystallization of the reactant with suitable template mole-
cules as demonstrated previously in some oligo- and polymer-
ization reactions.5,6 Peptides may be used to build these key
template matrices as they readily form inclusion compounds
and co-crystals with various organics.7 Peptides are advanta-
geous over other molecules due to the high degree of predict-
ability of their crystal structure motifs and potential for
modification simply by changing the side groups on the peptide
backbone.8 Short peptides are also non-toxic and ecofriendly
and therefore the solids they form would be a suitable media
to host “green” synthetic reactions.
The utilization of the peptide matrix in solid state syn-
thesis requires a comprehensive knowledge of the molecular
versus crystal structure relationship among peptides of differ-
ent length and type, as well as the pathways of their solid
state transformations induced by external stimuli, and the
stability limits of the overall solids. These key solid state
characteristics have not been sufficiently and systematically
studied in their mutual relation when it comes to peptides.
Department of Chemistry, University of Guelph, 50 Stone Rd. E., Guelph, ON,
N1G 2W1, Canada. E-mail: dsoldato@uoguelph.ca; Fax: +1 519 766 1499;
Tel: +1 519 824 4120, ext 53548
† Electronic supplementary information (ESI) available: Crystal structure data
in CIF format for 4G and 5G. Representative FTIR spectra of volatile products
released during TGA experiments of γ1G, α2G, β3G, 4G, 5G and 2G(p-h). CCDC
993137 and 993138. For ESI and crystallographic data in CIF or other
electronic format see DOI: 10.1039/c4ce00630e
This journal is © The Royal Society of Chemistry 2014
CrystEngComm

View Article Online
Paper
CrystEngComm
“Pre-heated” 2G, abbreviated here as “2G(p-h)”, was
prepared by heating α2G (270 mg, 2.0 mmol) isothermally at
250 °C in a ventilation oven for 30 min. The mass loss of
18.3% was significantly higher than 13.6% calculated for the
escape of water associated with the condensation reaction.
The recovered sample was black in color due to a minor
black fraction insoluble in water and DMSO. The soluble frac-
tion revealed the presence of c2G (cyclo-diglycine) as the
main product (>60% of the whole sample) as attested by
1H NMR (the singlet of two NH protons at 8.04 ppm and dou-
blet of four CH2 protons at 3.72 ppm, J = 3.2 Hz) and
13C NMR (a signal from two CO at 166.95 ppm and two
CH2 at 44.77 ppm). The presence of c2G as a dominant prod-
uct was also confirmed by PXRD analysis of the recovered
sample as its powder pattern corresponded to the crystal struc-
ture of c2G reported previously.34 Finally, a crystal was selected
from the sample (thin needle, 0.2 mm) and a single crystal
XRD test (unit cell only) also confirmed c2G.34 Another prod-
uct of the soluble fraction (~25%) was not identified.35
Heating 2G for 30 min at 280 °C resulted in similar results
with a greater amount of insoluble fraction, while heating at
200 °C showed no reaction (PXRD pattern of α2G30); therefore
only the product prepared at 250 °C was used in further
experiments.
Fig.
1 Linear glycines in zwitterionic form: general formula and
torsion angles describing molecular conformation.
For example, most studies on the solid state reactivity are limited
to specific, pharmaceutically relevant peptides and proteins
and pay minimal attention to their crystal structure.9–11
This work focuses on a series of linear glycine homopeptides
(Fig. 1), their crystal structure, molecular conformation and solid
state stability, as well as their thermally induced transforma-
tions in the solid versus gaseous phase. Since glycine is the
simplest amino acid, oligoglycines are frequently used as models
in experimental and theoretical studies on amide linkages
and conformational preferences in peptides and proteins12–23
as well as the degradation pathways of these important bio-
molecules.24–27 Most of these studies refer to the molecules
in vacuum or solution. Experimental solid state data are
available only either for the smallest members of the series
(1G, 2G and 3G) or the polypeptide (polyG). The medium-size
oligomers are scarcely studied due to their low solubility and
other limitations.28 This study rationalizes, based on both
newly obtained and previously reported data, several trends
that occur in the series. In particular, we show how the chain
length of the peptide affects the resulting crystal structure, as
well as stability and thermally induced chemical transforma-
tions. Further, we compare the properties and behavior of
the peptide molecules in the solid phase with the same mole-
cules under other conditions.
1H NMR and 13C NMR spectra were collected on a Bruker
Avance 400 MHz spectrometer. The samples (~2 mg) were
dissolved in DMSO-d6 (1 mL; D 99.9%, Cambridge Isotope
Laboratories, Inc). The data were analyzed using the TopSpin
3.1 software.
X-Ray diffraction (XRD)
All experiments were conducted at room temperature on a
SuperNova Agilent single crystal diffractometer equipped
with a microfocus CuKα (λ = 1.54184 Å) radiation source
and Atlas CCD detector. The data were processed using
CrysAlisPro software.36
Experimental section
General
All glycine oligomers were obtained from Chem-Impex
(assays >99%) and used without further purification. The
polymorphic forms of all previously studied glycines were
determined using PXRD: γ1G,29 α2G30 and β3G.31 These poly-
morphs are the most stable forms of the compounds at room
temperature and atmospheric pressure.
Single crystals of 4G and 5G were obtained by crystalliza-
tion from water.32 The room temperature solubility of glycine
oligomers in water decreases dramatically with the number
of residues: 201.10, 72.35, 5.25 and 1.50 mg ml−1 for 2G, 3G,
4G and 5G, respectively.33 Due to this fact large volumes
(>15 mL) of saturated solutions of 4G and 5G were prepared
at ~90 °C and then were allowed to simultaneously evaporate
and cool down to room temperature during two or three
days, until <5% of the original volume remained. The crys-
tals grew on the walls of the vial as extremely thin sheets
aggregated into clusters. Samples suitable for XRD analysis
were obtained after several (4G) or numerous (5G) crystalliza-
tion attempts.
Powder X-ray diffraction (PXRD) experiments were performed
with powder microsamples enclosed in 0.3 mm capillary tubes
(Hampton Research) which were mounted on a goniometer
head. Four frames were collected from different angular posi-
tions of the goniometer using φ-scan to cover 2θ range of
5–60°. Points of the powder pattern were generated from the
original images with the step of 0.02° in 2θ, with each of the
point representing readings from 200 to 2000 pixels.
Single crystals of 4G and 5G were mounted on the
tip of a glass fiber. Diffraction intensity data were collected
using ω-scan to the maximum 2θ angle of 150.6° and 133.2°
(resolution of 0.80 and 0.84 Å) for 4G and 5G, respectively,
with the redundancy factor ~4 and completeness >99%. The
unit cell parameters were refined using the entire data sets.
The structures were solved (direct methods) and refined
(full-matrix least-squares on F2) using SIR-92 (ref. 37) and
SHELXL-97.38 Non-hydrogen atoms were refined anisotropi-
cally, while hydrogen atoms were introduced at calculated
positions and refined isotropically as riding, with Uiso values
CrystEngComm
This journal is © The Royal Society of Chemistry 2014

View Article Online
CrystEngComm
Paper
set to 1.2Ueq of the carrier atom. Three hydrogen atoms were
located near each terminal nitrogen atom in a tetrahedral
geometry implying the zwitterionic form in all cases. This
form was also evident from the system of hydrogen bonds
present. The distances and angles in the NH3 groups then
were fixed but the groups were allowed to rotate around the
N–C bond. Detailed analysis of the structures and prepara-
tion of graphics were carried out using WinGX39 and Olex.40
Crystal data and experimental parameters are summarized in
Table 1. For further details see ESI.†
spectrometer through a stainless steel transfer line (2 mm
internal diameter, coated with fused silica) heated to 220 °C
and a TGA-FTIR accessory with 23 mL gas cell and 100 mm
beam path length heated to 250 °C. The FTIR spectra were
continuously collected in parallel with a TGA experiment at
8 cm−1 resolution, 8 scans per spectrum, and temporal resolu-
tion of 5.43 s. The FTIR experiments were controlled and ana-
lyzed using OMNIC Specta software.
Gas chromatography with mass spectrometry (GC-MS)
The samples were analyzed on an HP-5890 Series II gas chro-
matograph coupled to an HP-5971A mass selective detector.
Hot aqueous solutions of the studied peptides were prepared
immediately prior to the injection, with the concentration of
2.5 mg mL−1 for 5G and 5 mg mL−1 for all other compounds.
About 8 μL sample was injected in a split mode when injector
was set to 300 °C (Restek 2 mm Gooseneck Splitless liner). A
DB-5 ms column from Agilent was used (ID 0.25 mm, length
30 m, film depth 1 μm). The oven was programmed at 250 °C
isothermally for 15.2 min. Purge valve was opened at 0.5 min
at an 8 mL min−1 flow rate for 15 minutes, where helium
(Grade 5, Linde) was used as a carrier gas. Solvent delay was
set to 2.5 min. The detector, fitted with electron impact ioni-
zation source, was used at 300 °C. Total ion chromatograms
were acquired scanning for 30–400 a.m.u. Compounds were
identified and characterized by comparison of their retention
times and mass spectra with mass spectral database (G1033A
NIST/EPA/NIH-Library) associated with the instrument software.
Thermal analysis (DSC, TGA) and infra-red spectroscopy (FTIR)
The differential scanning calorimetry (DSC) and thermo-
gravimetric analysis (TGA) measurements were conducted
on a DSC Q-2000 and TGA Q-5000 IR analyzers, respectively
(TA Instruments). The experiments were controlled using
Q Advantage software and analyzed using Universal Analysis
software associated with the instruments.
The DSC measurements were conducted with linear heating
(5° min−1) in the 30–500 °C range under the flow of nitrogen
(50 mL min−1). The samples of α2G and 2G(p-h) of various size
(1.5 to 10.5 mg) were non-hermetically closed into Tzero alu-
minum pans. For the enthalpy measurements, the instrument
was calibrated with indium before each experiment and two
independent determinations were conducted, with 2.65 and
10.45 mg samples of α2G.
The TGA measurements were conducted with ~10 mg sam-
ples placed in 100 μL platinum pans and heated linearly
from 30 to 800 °C in the flow of nitrogen. Two sets of the
heating rate – N2 flow rate were used: 5° min−1 – 25 mL min−1
and 20° min−1 – 100 mL min−1. The second set was used when
the TGA measurements were accompanied by the FTIR analy-
sis of the released volatiles. For this purpose, the TGA furnace
outlet was connected to Thermo Scientific Nicolet iS10 FTIR
Results and discussion
Crystal structures of 4G and 5G
¯
Both 4G and 5G are triclinic, space group P1. The centro-
symmetry is possible due to the lack of a chiral center in gly-
cine. In both structures, there are two crystallographically
independent molecules (labeled A and B), both in a zwitter-
ionic form and almost ideal fully extended chain conforma-
tion (Fig. 2 and 3; Table 2). The molecules assembly into a flat
antiparallel β sheet banded arrangement, with N–H⋯OC
hydrogen bonds between amide groups of the adjacent mole-
cules (Fig. 2 and 3) as well as weak Cα–H⋯OC hydrogen
bonds. The β sheet bands are linked into infinite layers
through charge-assisted +N–H⋯O− hydrogen bonds on the
edges of the bands, between the terminal amine and carboxyl
groups. Also, the same type of hydrogen bonds cross-links the
layers into a 3D framework. A more detailed comparison of
the two crystal structures and the structures of other glycine
homopeptides is provided in the following sections.
Table 1 Summary of crystallographic data
Compound
4G
5G
Empirical formula
M
C8H14N4O5
246.2
C10H17N5O6
303.3
Crystal system
Space group
a (Å)
b (Å)
c (Å)
Triclinic
P1
Triclinic
P1
¯
¯
4.8042(2)
14.8701(5)
15.6243(5)
67.284(3)
85.298(3)
88.190(3)
1026.14(7)
4
4.7995(3)
14.9108(15)
17.6905(19)
90.203(8)
94.866(7)
91.976(7)
1260.7(2)
4
α (°)
β (°)
γ (°)
V (Å3)
Z
Dcalc (g cm−3
T(K)
)
1.594
295
1.598
295
Crystal color and habit
Crystal size (mm)
Reflections collected
Unique reflections (Rint
Refined parameters
White sheet
0.4 × 0.2 × 0.03
19 837
4181 (0.030)
309
0.039/0.110
−0.23/+0.26
White sheet
0.3 × 0.2 × 0.01
17 005
4425 (0.063)
381
0.086/0.235
−0.31/+0.66
Molecular conformations
The conformational behavior of glycine oligomers is of spe-
cial interest. As the only unsubstituted amino acid, glycine
is an important contributor to the conformational flexibility
of peptide fragments in proteins. Collagen and silk fibroin
)
R1/wR2 (data with I > 2σI)
Residual min/max (e Å−3
)
This journal is © The Royal Society of Chemistry 2014
CrystEngComm

View Article Online
Paper
CrystEngComm
Staphylococcus aureus' cell walls are responsible for the walls
high stability, as well as surface protein attachment.43–45 In
the intact bacterium cells, the pentaglycine bridges predomi-
nantly adopt a compact, coil conformation46 which easily
rearranges into the extended conformation in response to a
temperature change leading to swelling of the cell walls.47
There is a slow exchange between the two conformations at
0 °C with the compact to extended form ratio of approxi-
mately 9 : 1.48
There are a number of computational studies in the litera-
ture comparing different conformations of oligoglycines and
their analogs. For an isolated oligopeptide, some studies sup-
port the extended form as the most stable,18,41,49,50 while
others suggest a different conformation type.13,51 Computa-
tions in solution environment do not support the extended
conformation as energetically favorable.13,52 Finally, experi-
mental studies suggest the dominance of helical20,23,53 or
compact cyclic21 forms in aqueous solutions.
In this context, it is very surprising that the 4G and 5G
peptides in the crystal structures reported in this work adopt
almost ideal fully extended conformation (Table 2). This is
also true for the β polymorph of 3G.31 There is a reasonably
good agreement of the observed conformations with those
calculated by Ivanova et al.41 In contrast, the proposed con-
formation of polyG (form I) corresponds to a classical anti-
parallel β sheet (Table 2).42 (Another polymorph of polyG,
form II, contains a helical conformation.54) Moreover, the
low molecular weight fraction of polyG may have a unique
structure with the molecular conformation drastically
Fig. 2 Antiparallel β sheet arrangement of molecules in the crystal of
4G: ORTEP drawing and labeling scheme. The molecules in the
sequence BBAA… form
a flat band where the next neighbor is
antiparallel and only the fourth neighbor is related by a translation.
are among many examples of highly flexible glycine-rich
polymers occurring naturally. Conformationally flexible
pentaglycine bridges cross-linking peptidoglycan layers in
Fig. 3 Antiparallel β sheet arrangement of molecules in the crystal of 5G: ORTEP drawing and labeling scheme. The molecules in the sequence
BBAA… form a flat band where the next neighbor is antiparallel and only the fourth neighbor is related by a translation.
CrystEngComm
This journal is © The Royal Society of Chemistry 2014

View Article Online
CrystEngComm
Table 2 Torsion angles (°) in some glycine homopeptides with a β sheet structurea b
Paper
,
φ12
ψ12
φ23
ψ23
φ34
ψ34
φ45
ψ45
Ref.
α2G
2G
−152.4
(145.3)
−149.9
−161.8
(148.5)
−149.3
−159.2
(169.2)
−158.9
+148.9
−149.9
+149.9
−155.1
(156.8)
+177.6
−165.9
(177.7)
+178.1
−165.2
(170.4)
−163.9
−178.7
+146.5
−146.5
Crystal30
Single molecule, calculatedc 41
Crystal31
β3G, mol. A
β3G, mol. B
3G
4G, mol. A
4G, mol. B
4G
5G, mol. A
5G, mol. B
I-polyG, chain 1
I-polyG, chain 2
−171.5
+174.5
(172.8)
−171.2
+172.0
(175.3)
+171.4
+172.2
+172.8
+172.8
(173.1)
+173.7
−175.6
(178.9)
−175.0
−172.6
Crystal31
Single molecule, calculatedc 41
Crystal, this work
−177.4
+172.8
(170.6)
+173.0
+174.6
−172.3
−174.6
(176.1)
−176.0
−175.1
Crystal, this work
Single molecule, calculatedc 41
Crystal, this work
+177.3
+172.8
+174.2
−173.7
Crystal, this work
Crystal structure model 42
Crystal structure model 42
a
φij and ψij are the torsion angles (as defined in Fig. 1) between i and j residues. For example, for the molecule B in 4G (Fig. 2) φ12 is
b
N1B–C1B–C2B–N2B and ψ12 is C2B–N2B–C3B–C4B. Standard (φ, ψ) angle sets: (180,180) fully extended chain; (−142, 145) antiparallel-chain
pleated β sheet; (−119, 113) parallel-chain pleated β sheet; (−47, −57) right-handed α helix; (−49, −26) right-handed 310 helix. Simultaneous
change of both signs produces a mirrored conformation. c Only absolute values reported.41
different from both fully extended and antiparallel β types.55
These facts may suggest that although the same structural
organization is observed in the β3G – 4G – 5G series, these
molecules are not long enough to observe the formation of a
structure similar to that of polyG. A possible explanation for
the difference between oligoglycines studied in this work and
polyG is a high concentration of charge-bearing amine and
carboxyl terminal groups in the former. These groups set
additional requirements on the crystal structure due to the
necessity of optimal charge distribution within the crystal as
well as the formation of numerous energetically favorable
charge-assisted hydrogen bonds.
not resemble any of the classical protein structure types. α2G
is the most stable polymorph at 105 °C58 and, most likely, at
ambient conditions. It has a monoclinic structure with one
molecule in the asymmetric unit.30 This structure is related
to the family of antiparallel β sheet structures reported and
discussed in this work (Table 3). 3G was isolated in two poly-
morphic forms59 but no report of α3G has appeared since
1931 and the structure remains unknown. β3G, which is
apparently the stable form of 3G, crystallizes in a triclinic
structure with two independent molecules (A and B, see Table 2).
The structure is an antiparallel β sheet type. 4G and 5G crys-
tallized and studied in this work always grew as a single poly-
morph (as attested by PXRD tests). Higher glycine oligomers,
to the best of our knowledge, have not been characterized for
their crystal structure. PolyG is known to exist as two poly-
morphs (forms I and II) of which form II is trigonal, P31, with
the molecules having a helical conformation.54 Form I is actu-
ally two polymorphs, both of the β sheet type.55 The relative
stability of the forms I and II is not known and would be diffi-
cult to determine experimentally due to the polymeric nature
of the compound.
Solid state structure trends in glycine homopeptides
Polymorphism. 1G forms three polymorphs29 of which
γ1G is the most thermodynamically stable at ambient
conditions.56 2G forms three polymorphs,57 of which α2G
seems to be the most stable. No report of γ2G has appeared
since 1931 (ref. 57) and the full structure remains unknown.
The crystal structure of β2G58 has H-bonded dimers and does
Table 3 Comparison of the β-sheet-like structures of solid oligoglycines and polyG
α2G30
β3G31
4Ga
5Ga
PolyGb
Max van der Waals size of the molecule in
the crystal (Å)
9.9
13.5 (A); 13.6 (B)
17.2 (A and B) 20.8 (A);
20.7 (B)
3.52 per monomer
Repeat distance in the chain–axis direction (Å) 9.42
24.07 (two molecules) 15.62 (=c)
17.69 (=c)
AA′B′B…
[3 1 0]
7.04 (=b; two monomers)
AA′…
[1 0 0] (along a)
4.77 (=a/2)
Order of molecules in the β sheet bandc
Direction of the β sheet band in the crystal
AA′…
AA′B′B…
AA′B′B…
[3 1 0]
¯
¯
[0 1 0] (along b) [0 1 3]
Average distance between adjacent molecules 4.78
5.11
5.09
5.09
in the β sheet band (Å)
Crystal planes defining the orientation of
the β sheet layers
¯
¯ ¯
¯ ¯
(1 3 2)
(2 0 2)
(2 3 1)
(1 3 0)
(0 0 1)
Distance between adjacent β sheet layers (Å)
d2 0 ¯2 = 3.21d
1.516
d2 ¯3 ¯1 = 3.29e
1.572
d1 3 0 = 3.31e
1.594
d1 ¯3 ¯2 = 3.32e d0 0 1 = 3.38e
1.598 1.67
Calculated density f (g cm−3
)
a
c
b
Data of this work. Form I; monoclinic cell (P21/c) with a = 9.54, b = 7.044, c = 3.67 Å and β = 113°; antiparallel rippled β sheet model used.42
A and B are two crystallographically independent molecules, while A and A′ are enantiomorths (related by an inversion center) and therefore
d
e
antiparallel. One of two most intense peaks on the simulated powder diffractogram. The most intense peak on the simulated powder
diffractogram. f Room temperature for all the structures.
This journal is © The Royal Society of Chemistry 2014
CrystEngComm

View Article Online
Paper
CrystEngComm
Trends in the β sheet structures. The comparison of
structure).54 The peak is absent for 3G*HCl61 or c6G*0.5H2O
(cyclo-hexaglycine hemihydrate)62 as none of these structures
have β sheet layers present. The powder patterns of antiparallel
β sheet polyalanines all show the strongest peak correspond-
ing to the interlayer distance of 5.1 Å, a substantial increase
due to the methyl group of alanine.63 One should note there
are some variations within the narrow 3.2–3.4 Å range. As a
consequence of not fully extended conformation in polyG,
adjacent molecules in the β sheet band come slightly closer
(4.77 Å) causing some increase in the interlayer distance (3.38 Å).
As expected, the density of ideal structure grows with the
number of residues, as the peptide covalent links effectively
replace van der Waals contacts in the crystal.
structures of the antiparallel β sheet type is presented in
Table 3. α2G shows some resemblance to the higher oligomers
and may be regarded as a simplified prototype in the series.
β3G, 4G and 5G have a remarkably similar crystal structure
which may extend to the higher oligomers not studied yet. In
all the three structures, there are two molecules (A and B)
arranged in an antiparallel β sheet band as illustrated in
Fig. 2 and 3. The band is very flat due to the fully extended
molecular conformation of both A and B molecules and the
H-bonds connecting neighboring molecules. The sequence of
molecules in the band is AA′B′B…, where the two adjacent
molecules of the same kind are related by an inversion center
(e.g. A and A′) while only the fourth neighbor is related to the
initial by a translation. PolyG, form I, may exist in at least two
variations42,55 of which the most stable (for high molecular
weight) has two enantiomorphic chains as H-bonded neigh-
bors in the β sheet, resulting in the AA′… sequence.60
Table 3 reveals crystal packing similarities throughout the
whole series, in spite of differences in the unit cell symmetry
and dimensions. The maximum van der Waals size of the
molecule changes by an increment of 3.5 to 3.7 Å per residue,
as expected for an extended conformation. The low value of
3.5 Å for polyG is consistent with the chain conformation not
being fully extended. The repeat distance in the chain–axis
direction grows stepwise, with smaller steps from even to
odd number of residues (2.62 Å from 2G to 3G; 2.07 Å from
4G to 5G) and greater otherwise (3.58 Å from 3G to 4G). The
average distance between H-bonded molecules in the β sheet
band falls into the narrow range of 4.8–5.1 Å defined by inter-
molecular H-bonding.
Stability trends in solid glycine homopeptides
Thermal stability of glycine oligomers under nitrogen atmo-
sphere may be quantified using the onset temperatures of
the initial mass loss on TGA curves (Fig. 4 and Table 4). As
shown in the next section, the process associated with the
initial mass loss is irreversible and therefore the onset tem-
peratures are related to the kinetic stability of the com-
pounds. Due to the kinetic nature of the process, the onset
temperature depends on the heating rate; therefore only rela-
tive stability is discussed here.
One can clearly see that the stability limit increases with
the chain length in the series of β sheet structures as α2G <
β3G < 4G < 5G. The main factor seems to be the stabiliza-
tion environment in the solid phase. The peptide molecule
must acquire a certain degree of freedom in order for any
decomposition reaction to take place. Each additional residue
adds two extra H-bonds as well as additional van der Waals
contacts restricting the molecular motion. Another possible
factor is the stability of the molecule itself. Although according
to calculations the energy of the peptide bond does not depend
significantly on the number of residues for a neutral oligoglycine
in extended conformation,51 in the studied solids the peptides
exist in zwitterionic form, with two charges located on the
opposite ends of the molecule. It is known that the zwitter-
ionic form of amino acids and peptides is less stable for an
A very characteristic feature throughout the whole series
is the distance between β sheet layers of 3.2–3.4 Å. It is
defined by the interlayer van der Waals contacts (cf. C⋯C
van der Waals contact of 3.4 Å). The distance accounts for
the strongest (or second strongest) peak on the simulated pow-
der diffractograms of the peptides which should be a signature
in all β sheet glycines. For example, it appears as a strong peak
at 3.40 Å42 or 3.45 Å54 on the powder diffractogram of polyG,
form I, while there is no such a peak for polyG, form II (helical
Fig. 4 TGA thermograms of studied glycines (stable polymorphs). The onset temperatures are listed in Table 4. Five bars on the right show
calculated mass losses for the release of one mole of water for each of the five compounds. Purge: N2; heating rate: 5° min−1
.
CrystEngComm
This journal is © The Royal Society of Chemistry 2014

View Article Online
CrystEngComm
Paper
Table 4 Onset temperatures (°C) of the initial mass loss of studied
glycines subjected to TGA (average of three determinations)
The pyrolysis of the studied compounds (Fig. 5) shows
similarity in the sequence the volatile products are released
although the temperatures of the major steps vary. At the
decomposition onset, two types of processes occur. The first
type is defragmentation reactions, specifically deamination
and decarboxylation (i.e. loss of NH3 and CO2) and possibly
dehydrazination (the IR spectrum of hydrazine H2N–NH2 is
almost identical to that of NH3). These reactions should be
followed by polymerization of remained less volatile frag-
ments. The second type is the condensation reaction with the
release of H2O and formation of new peptide bonds. The
intermolecular condensation can lead to higher linear peptides
while intramolecular condensation to cyclic oligopeptides of
which the major product is the cyclo-diglycine, c2G, also
known as 2,5-diketopiperazine (DKP).
At the temperatures above 350 °C the major volatiles
released are NH3, CO2, HNCO and CO while H2O is not sig-
nificant. The formation of NH3 and CO2 is still due to the
defragmentation of linear peptides, through the stripping the
amino and carboxyl terminals. Some of these terminals were
present in the original molecules and survived the initial con-
densation step, while others could form in the reaction of the
original molecules with c2G (e.g. 1G + c2G = 3G26,75). The
condensation reaction is not noticeable here, probably because
the concentration of pairs of amino and carboxyl terminals in
close proximity to each other drops dramatically after the ini-
tial step. The release of HNCO and CO results from thermal
decomposition of c2G (Fig. 6).27,76,77
γ1G
α2G
β3G
4G
5G
Heating rate 5° min−1
221(2)a
216(1) 231(1) 250(1) 289(1)
226(1) 237(2) 262(2) 292(2)
Heating rate 20° min−1 244(2)a
a
The value is derived from the tangent at the maximum rate mass
loss but the gradual onset may indicate a delay due to an induction
period (see text).
isolated molecule, but becomes stabilized in solution or the
crystalline state. For example, from both experimental64,65
and theoretical66–68 studies glycine is not a zwitterion in the
gaseous phase but is zwitterionic in any of its three crystalline
forms.29 Better separation of charges in the fully extended
conformation of the studied oligoglycines may reduce the
charge density and additionally stabilize longer-chain zwitter-
ions. Finally, as shown in the next section, the principal
decomposition pathways differ for the glycines of different
chain length. In particular, the cyclization reaction for 1G
cannot occur in a single step that may explain why the ther-
mogram of γ1G shows a more gradual onset, with the decom-
position process delayed due to an induction period.
Pyrolysis of glycine homopeptides in the solid state
In order to examine the pathways of thermal decomposition
under nitrogen (pyrolysis), the TGA experiments with simulta-
neous FTIR analysis of released volatiles were performed. The
Gram–Schmidt curves showing total IR absorbance vs. tem-
perature are shown in Fig. 5. Representative spectra are given
in the ESI.† In all cases, the decomposition proceeds in sev-
eral overlapped steps, with the same set of species produced:
H2O, NH3, CO2, HNCO, CO and HCN. The formation of these
volatiles has been observed and explained for 1G and 2G.27,69–74
Samples recovered at 600 or 800 °C were black glassy residues
which is consistent with the formation of partially decom-
posed polymeric matter.
Finally, at the highest temperatures some amounts of
HCN are observed. This is also a product of c2G decomposi-
tion (Fig. 6). Although c2G produced at lower temperatures is
too volatile to remain in the TGA crucible, some amounts of
c2G can form through depolymerization reaction of higher
oligopeptides24–27,77–80 (see the last section). The formation
of HCN over HNCO is caused by the higher temperature, as
demonstrated in the literature.77
The differences in the decomposition temperature profiles
along the series (Fig. 4 and 5) can be explained assuming the
formation of c2G is the main process at the very beginning
which in turn could accelerate other decomposition routes.
The formation of c2G from 2G is the easiest as it requires the
formation of only one peptide bond and the only other prod-
uct is the H2O molecule. In contrast, two peptide bonds
should form in order to yield c2G from 1G, and so the onset
is the slowest in the series. For higher oligomers the onset
temperature gradually increases with the chain length. This
increase is due to the dependancy of the depolymerization
mechanism on the mobility of the molecules in the solid.
In order to confirm these conclusions, a sample of α2G
was pre-heated in a ventilation oven at 250 °C and the prod-
uct referred here as 2G(p-h) was characterized and further
studied. The NMR and XRD analyses confirmed c2G as the
main component of 2G(p-h) but the observed mass loss
implied other reactions with a release of volatiles as well (see
Experimental section for details). A sample of 2G(p-h) was
subjected to a TGA-FTIR experiment under similar conditions
Fig. 5 Gram–Schmidt total IR absorbance vs. temperature for gaseous
species evolved in TGA experiments (purge: N2, 100 mL min−1; heating
rate: 20° min−1).
This journal is © The Royal Society of Chemistry 2014
CrystEngComm

View Article Online
Paper
CrystEngComm
Fig. 6 Thermally induced decomposition pathways of c2G. The (a) and (b) defragmentation routes of β-lactam are the low- and high-temperature
routes, respectively.77 Molecular masses are shown in square brackets.
(Fig. 5). It is clear that although some condensation still
occurs for 2G(p-h), it is unlikely to be the cyclization of 2G
but rather depolymerization of longer chains, as the first
peak is positioned similar to that of 5G.
4-imidazolidinone (3-, 4- and 5-membered rings respectively).
The formation of these cyclic molecules and their decom-
position produces smaller species: HNCO, H2CNH, H2CCO,
H2CCH2, CO, H2 and HCN. Some of these species appear on
the FTIR spectra of volatiles produced in TGA of solid gly-
cines (Fig. 5): HNCO and CO on moderate heating and HCN
at high temperatures, while other species either do not form
from the solid samples or quickly polymerize.
Pyrolysis in the gaseous phase and comparison
In order to compare the thermally induced transformations
in the solid and gaseous phases, the same set of glycines was
subjected to GC-MS analysis. At the injector site, a very small
amount (~8 μL) of diluted peptide is instantly heated to
300 °C allowing a reasonable assumption that all transforma-
tions induced at this temperature occur in the gaseous phase.
The gas chromatograms (Fig. 7) reveal only one peak.
Although the intensity of the peak varies, the retention time
and mass spectrum are practically identical in all cases. The
mass spectra unequivocally indicate the formation of c2G
from all the studied glycine homopeptides at 300 °C in the
gaseous phase. The absence of other peaks on the chromato-
grams is because other products that can form in the injector
reaction space are either too light (e.g. H2O) or too heavy
(e.g. higher oligomers), so that their retention times either
too short or very long. In other words, c2G is the only product
in the middle molecular mass range formed from the gly-
cines gasified at 300 °C. One could see that the intensity of
the c2G peak is maximal for 2G(p-h), somewhat lower for 2G,
3G and 4G, relatively low for 5G and very low for 1G. This
observation is consistent with the easiness of c2G formation
from different molecules in the series. In particular, the 2G(p-h)
sample already contains c2G, while the formation of c2G from
1G is difficult being more than a one-step reaction.
In general, one can see both similarities and differences
in the pyrolysis of the glycine homopeptides in the solid state
and gaseous phase. In both cases a number of parallel reac-
tions occur, with TGA-FTIR and GS-MS being complementary
in detecting these reactions and species produced. A similar
set of final volatile products forms from the solid and gas-
eous glycines and c2G forms on early stages in both cases.
At the same time, there are evident differences that must
be mentioned here. First, the pyrolysis of solid glycines pro-
duces polymeric materials due to high concentration of reac-
tive species formed. In the gaseous phase, fragmentation is
more likely. Consequently, a number of reactive species seen
in the GC-MS experiments do not show in TGA-FTIR but poly-
meric residues are recovered instead. Second, the same reac-
tion (such as the formation of c2G) in the solid state could
be easier conducted and scaled. In the gaseous phase, only
small amounts may be reacted and very big amounts of solvent
and energy are required. These considerations bring forth the
idea of using the solid state pyrolysis as a possible synthetic
method for c2G and other DKPs as discussed in the next section.
Formation of the cyclic dipeptide and cyclization mechanisms
Fig. 6 shows the major decomposition pathways of the
c2G molecule leading to species observed in the mass spec-
tra. The three major products are α-lactam, β-lactam and
The formation of cyclo-diglycine (c2G, or DKP) in the course
of thermal treatment of the studied glycines is of special
interest as it may suggest a general synthetic procedure for
CrystEngComm
This journal is © The Royal Society of Chemistry 2014

View Article Online
CrystEngComm
Paper
cyclic dipeptides (DKPs). These molecules attracted attention
due to a wide spectrum of their biological activity and the
use in medicinal chemistry.81–84 On the other hand, the food
industries interest in DKPs is due to their contribution to
flavors, such as the bitter taste of roasted coffee,85 meats,86
beer87 and chocolate.88
The formation of c2G, the simplest DKP, was observed in
all thermal experiments of this work. However the reaction is
almost always accompanied by various decomposition pro-
cesses including the degradation of c2G itself. Therefore a
better understanding of the reaction mechanism and thermo-
dynamics are essential in the development of future solid
state synthetic methods of DKPs.
Thermodynamics. The enthalpy of the cyclization reaction (1)
was measured in DSC experiments:
α2Gsolid = c2Gsolid + H2Ogas
(1)
ΔH = 482(3)J g−1 = 63.7(4) kJ mol−1 (at 492 K)
The DSC thermograms of α2G and 2G(p-h) are compared
in Fig. 8. The only difference between the intact and pre-
heated samples is the presence of an endotherm with the
onset temperature of 219 °C due to the cyclization reaction (1).
The complex endotherm at 280–330 °C, almost identical for
both samples, corresponds to the sublimation of the cyclic
form with evident partial degradation. The absence of any
other effects on the α2G thermogram below 280 °C was sur-
prising indicating the cyclization reaction was the only pro-
cess in this range.89 This observation made it possible to
accurately measure the enthalpy of the reaction (1). The non-
hermetic seal of the DSC capsules allowed for vaporization of
water released during the reaction but prevented early subli-
mation by maintaining sufficient vapor pressure of c2G over
the sample. Therefore the enthalpy of 63.7(4) kJ mol−1 can be
assigned to the pure reaction (1) at 219 °C. In order to esti-
mate the standard enthalpy of the reaction, the vaporization
enthalpy of water at 220 °C of 33.468 kJ mol−1 (ref. 90) should
be subtracted as well as the energy to heat water to 219 °C,
14.7 kJ mol−1 (taking heat capacity of water 4.21 J g−1 K−1
=
75.9 J mol−1 K−1).90 The contribution due to heating the solids
may be assumed to be −2.5 kJ mol−1 (the heat capacities
reported for 2G and c2G at 298 K are 149 (ref. 91) and
136 J mol−1 K−1,92 respectively). With this assumption, the
standard enthalpy of the reaction (2) is
α2Gsolid = c2Gsolid + H2Oliquid
(2)
ΔH298 ≈ 18 kJ mol−1
As one can see, the cyclization reaction is thermodynami-
cally unfavorable under ambient conditions (the entropy term
is expected to be small) but becomes possible at elevated
temperatures due to growing entropy term.
Fig. 7 Gas chromatograms of 1G–5G (a–e) and 2G (p–h) (f) and mass
spectra corresponding to the major peak at ~4.3 min. The weight
amounts of injected compounds are the same except for 5G which is
half the amount (see Experimental section).
This journal is © The Royal Society of Chemistry 2014
CrystEngComm

View Article Online
Paper
CrystEngComm
terminal amino group on the carbonyl to close the ring.108
Finally other, less common mechanisms were proposed.109
One reason why the tetrahedral intermediate mechanism
(Fig. 9) is believed to take place in this study is because it
agrees with all experimental observations and explains the
similarity of the decomposition process for the whole series.
Another reason is that the hydroxyl ion is not a good leaving
group making the SN1 mechanism very unlikely.
As described in the crystal structure part, the peptide
solids of this work have their molecules in the zwitterionic
form and fully extended conformation packed in a β sheet
arrangement. For intramolecular aminolysis, the peptide
molecule should be in its canonical rather than zwitterionic
form. Further complications include restricted motion of
the molecule and the necessity to break a number of inter-
molecular hydrogen bonds as the reaction proceeds in the
solid state. These considerations imply a multi-step process
much more complicated than would appear from Fig. 9. The
thermal energy appears to be essential in generating a neutral
peptide molecule, breaking hydrogen bonds and disrupting
the crystal packing in order for the molecule to fold and close
the cycle. Although more detailed studies are needed to eluci-
date the mechanism further, it is relevant to mention here a
remarkable flexibility observed previously in crystalline pep-
tides.110–114 Significant local distortions of the crystal struc-
ture and possibly concerted mechanisms may be expected.
The estimated activation energies for cyclization of aspartame
and aspartyl-phenylalanine of 268(8) and 242(8) kJ mol−1
respectively115 may illustrate the kinetic requirements for the
reactions to occur in the solid state, the values exceeding the
cyclization enthalphy determined in this work by about one
order of magnitude.
Fig. 8 DSC thermograms of 2G(p-h) and α2G (sample mass: 1.47 and
2.65 mg, respectively).
Kinetics. Aside from standard synthetic methods,93 the
formation of DKPs was observed in the mass spectrometry
experiments on dipeptides,94–96 pyrolyzed oligopeptides76 and
thermally degraded proteins,97 upon sublimation of dipeptides,98,99
from thermally100,101 or microwave102–104 treated solutions and
in few cases in the solid state.105 In all cases the process
involves cyclization through intramolecular aminolysis that
can proceed in different pathways. The type of the leaving
group, amino acid residue, pH and temperature are the factors
that greatly affect the mechanism and rate of the reaction in
solution.100 The most common mechanism observed for short
peptides is a nucleophilic addition of the terminal amino
group to the carbonyl moiety of the next residue in the peptide
sequence. It outcomes in a tetrahedral intermediate followed
by proton transfer from the amine nitrogen onto carbonyl
oxygen. The resultant intermediate decomposes after the proton
transfer to the leaving group.26,75,106,107
This mechanism is illustrated in Fig. 9 for the oligoglycines
of this work. The tetrahedral intermediate is a cyclic zwitterion
which splits into c2G and HR, where the leaving group HR
is H2O for 2G or a shortened glycine chain for higher
oligoglycines. The former case can be referred to as condensa-
tion, while the latter as depolymerization. Aside from the reac-
tions in solution, the tetrahedral intermediate mechanism
was proposed for mass spectrometry experiments,95 pyrolysis
of a melt,97 or solid state pyrolysis of short peptides.78,79
In contrast, the thermally-induced intramolecular cyclization
of solid aspartame (Asp–Phe–OCH3) follows the SN1 mecha-
Conclusion
This work reveals the effect of the chain size on the properties
of solid peptides made of glycine residues, including their
crystal structure, conformation, spatial arrangement and non-
valent interactions of the peptide molecules in the crystal, as
well as thermal stability and chemical changes in the solid
state induced by heating.
The antiparallel β sheet structure with the molecules in
zwitterionic form and fully extended conformation is the most
stable crystalline form for lower glycine oligomers although
up to three polymorphs may be isolated. For polyglycine the
crystal structure and the molecular conformation differ, but
one of the polymorphs (form I), likely the most stable, is still
a β sheet structure. In spite of dissimilar unit cell parameters,
the crystal packing motifs are very similar throughout the series.
The thermal stability of crystalline oligoglycines increases
with the chain length due to the stabilization environ-
ment of the molecules in the crystal and other factors. The
decomposition onset is subtle and involves a number of possi-
ble pathways which depend, to a lesser extent, on the peptide
chain length but, to a much greater extent, on the experimen-
tal conditions. The two types of reactions at the onset are
−
nism. The first step involves the release of OCH3 leaving
group to form acyl cation, followed by the attack of the
Fig. 9 Thermally induced intramolecular cyclization of oligoglycines
through a tetrahedral intermediate.
CrystEngComm
This journal is © The Royal Society of Chemistry 2014

View Article Online
CrystEngComm
Paper
defragmentation and condensation. These two reactions take
place both in the solid state and the gas phase but polymeri-
zation of fragments dominates on a later stage for solid state
samples only.
12 R. A. Scott and H. A. Scheraga, J. Chem. Phys., 1966, 45,
2091–2101.
13 C. Alemán and J. J. Perez, J. Comput.-Aided Mol. Des., 1993,
7, 241–250.
A major condensation product in thermal treatment of all
oligoglycines is cyclo-diglycine (DKP). This formation occurs
as intramolecular cyclization from diglycine and a depolymer-
ization reaction from higher oligomers. From diglycine, the
reaction proceeds quickly and DKP may be obtained as a
pure product but the reaction is too sensitive to experimental
conditions and difficult to scale. The solid state synthesis of
DKP could be a synthetically important reaction if conducted
without side products. More studies are needed in order to
understand the effect of various factors on the cyclization
reaction and to prevent undesired degradation processes which
occur in parallel. In particular, the mechanism of the cycliza-
tion in the solid state appears to be complex and needs to be
understood. The outcome of such studies could be the devel-
opment of a general synthetic procedure for various DKPs.
Finally, similar studies on other peptide series would be
useful to verify how the trends observed in our study apply to
solid peptides of various compositions.
14 T. V. Chalikian, A. P. Sarvazyan, T. Funck and K. J. Breslauer,
Biopolymers, 1994, 34, 541–553.
15 C. Alemán, R. Roca, F. J. Luque and M. Orozco, Proteins,
1997, 28, 83–93.
16 R. Improta, V. Barone, K. N. Kudin and G. E. Scuseria,
J. Chem. Phys., 2001, 114, 2541–2549.
17 G. Nandini and D. N. Sathyanarayana, J. Phys. Chem. A,
2003, 107, 11391–11400.
18 A. Chatterjee, L. Zhao, L. Zhang, D. Pradhan, X. Zhou and
K. T. Leung, J. Chem. Phys., 2008, 129, 1–6.
19 R. Wu and T. B. McMahon, J. Phys. Chem. B, 2009, 113,
8767–8775.
20 S. Bykov and S. Asher, J. Phys. Chem. B, 2010, 114, 6636–6641.
21 S. A. Allison, H. Pei, U. Twahir, H. Wu and H. Cottet, J. Sep.
Sci., 2010, 33, 2430–2438.
22 A. M. Ferrari, B. Civalleri and R. Dovesi, J. Comput. Chem.,
2010, 31, 1777–1784.
23 S. V. Bykov and S. A. Asher, J. Phys. Chem. Lett., 2010, 1,
269–271.
24 Y.-C. Oh, C.-K. Shu and C.-T. Ho, J. Agric. Food Chem., 1991,
39, 1553–1554.
Acknowledgements
We thank the Canada Foundation for Innovation (CFI), Ontario
Ministry of Research and Innovation (MRI) and Department
of Chemistry at the University of Guelph for funding the X-ray
diffraction facility, and The National Science and Engineering
Research Council Canada for the thermoanalytical instrumen-
tation. DVS is grateful to Prof. W. Gabryelski (U. of Guelph)
for a useful discussion. The assistance of M. Ignacio in some
of the PXRD tests is appreciated.
25 V. A. Basiuk and J. Douda, J. Anal. Appl. Pyrolysis, 2000, 55,
235–246.
26 J. Bujdák and M. Rode, J. Pept. Sci., 2004, 10, 731–737.
27 J. Li, Z. Wang, X. Yang, L. Hu, Y. Liu and C. Wang, J. Anal.
Appl. Pyrolysis, 2007, 80, 247–253.
28 The oligomers dissolve only in highly polar solvents, easily
become ionic in acidic or basic media and gradually
degrade with time.
29 E. V. Boldyreva, T. N. Drebushchak and E. S. Shutova,
Z. Kristallogr., 2003, 218, 366–376.
30 S. A. Moggach, D. R. Allan, S. Parsons and L. Sawyer, Acta
Crystallogr., Sect. B: Struct. Sci., 2006, 62, 310–320.
31 T. Srikrishnan, N. Winiewicz and R. Parthasarathy, Int. J.
Pept. Protein Res., 1982, 19, 103–113.
32 The solubility in other solvents, such as methanol, ethanol
or their mixtures with water, was too low.
33 J. Lu, X.-J. Wang, X. Yang and C.-B. Ching, J. Chem. Eng.
Data, 2006, 51, 1593–1596.
Notes and references
1 G. M. J. Schmidt, Pure Appl. Chem., 1971, 27, 647–678.
2 K. Tanaka and F. Toda, Chem. Rev., 2000, 100, 1025–1074.
3 G. Kaupp, Top. Curr. Chem., 2005, 254, 95–183.
4 P. T. Anastas and J. C. Warner, Green chemistry: theory and
practice, Oxford University Press, New York, 1998.
5 L. R. MacGillivray, G. S. Papaefstathiou, T. Friščić, D. B. Varshney
and T. D. Hamilton, Top. Curr. Chem., 2004, 248, 201–221.
6 J. W. Lauher, F. W. Fowler and N. S. Goroff, Acc. Chem. Res.,
2008, 41, 1215–1229.
34 R. Degeilh and R. E. Marsh, Acta Crystallogr., 1959, 12,
1007–1014.
7 D. V. Soldatov, in Nanoporous Materials, ed. A. Sayari
and M. Jaroniec, World Scientific, New Jersey, 2008,
pp. 213–224.
35 1H NMR: triplet at 1.19 ppm, singlet at 2.01 ppm and
quartet at 4.04 ppm (approx. intensity ratio 2 : 1.5 : 1).
PXRD: 2θ = 20.46, 21.28 and 25.85°.
8 C. H. Görbitz, Acta Crystallogr., Sect. B: Struct. Sci., 2010,
66, 84–93.
9 M. C. Lai and E. M. Topp, J. Pharm. Sci., 1999, 88, 489–499.
10 S. Jacob, A. A. Shirwaikar, K. K. Srinivasan, J. Alex, S. L. Prabu,
R. Mahalaxmi and R. Kumar, Indian J. Pharm. Sci., 2006, 68,
154–163.
36 Xcalibur CCD System, CrysAlisPro Software, Version
1.171.35.8, Agilent Technologies, Yarnton, Oxfordshire,
UK, 2011.
37 A. Altomare, G. Cascarano, C. Giacovazzo and A. Gualardi,
J. Appl. Crystallogr., 1993, 26, 343–350.
38 G. M. Sheldrick, SHELXL-97, Program for refinement of
crystal structures, University of Göttingen, Germany, 1997.
39 L. J. Farrugia, J. Appl. Crystallogr., 1999, 32, 837–838.
11 J. Patel, R. Kothari, R. Tunga, N. M. Ritter and B. S. Tunga,
BioProcess Int., 2011, 9(1), 20–31.
This journal is © The Royal Society of Chemistry 2014
CrystEngComm

View Article Online
Paper
CrystEngComm
40 O. V. Dolomanov, L. J. Bourhis, R. J. Gildea, J. A. K. Howard
and H. Puschmann, J. Appl. Crystallogr., 2009, 42, 339–341.
41 B. B. Ivanova, T. Kolev and S. Y. Zareva, Biopolymers, 2006,
82, 587–596.
42 B. Lotz, J. Mol. Biol., 1974, 87, 169–180.
43 O. Schneewind, A. Fowler and K. F. Faull, Science, 1995,
268, 103–106.
69 P. G. Simmonds, E. E. Medley, M. A. Ratcliff and
G. P. Shulman, Anal. Chem., 1972, 44, 2060–2066.
70 K. J. Voorhees, F. Basile, M. B. Beverly, C. Abbas-Hawks,
A. Hendricker, R. B. Cody and T. L. Hadfield, J. Anal. Appl.
Pyrolysis, 1997, 40–41, 111–134.
71 N. Gallois, J. Templier and S. Derenne, J. Anal. Appl.
Pyrolysis, 2007, 80, 216–230.
44 X. Zhou and L. Cegelski, Biochemistry, 2012, 51, 8143–8153.
45 A. P. A. Hendrickx, J. M. Budzik, S.-Y. Oh and O. Schneewind,
Nat. Rev. Microbiol., 2011, 9, 166–176.
46 A. Lapidot and C. S. Irving, Proc. Natl. Acad. Sci. U. S. A.,
1977, 74, 1988–1992.
72 V. Y. Yablokov, I. L. Smel'tsova, I. A. Zelyaev and
S. V. Mitrofanova, Russ. J. Gen. Chem., 2009, 79, 1704–1706.
73 S.-S. Choi and J.-E. Ko, J. Anal. Appl. Pyrolysis, 2010, 89, 74–86.
74 S. C. Moldoveanu, Pyrolysis of Organic Molecules with
Applications to Health and Environmental Issues, in
Techniques and Instrumentation in Analytical Chemistry,
Elsevier, Amsterdam, 2010, ch. 18, vol. 28.
47 A. Lapidot and C. S. Irving, Biochemistry, 1979, 9,
1788–1796.
48 G. Tong, Y. Pan, H. Dong, R. Pryor, G. E. Wilson and
J. Schaefer, Biochemistry, 1997, 36, 9859–9866.
49 F. A. Momany, R. F. McGuire, J. F. Yan and H. A. Scheraga,
J. Phys. Chem., 1971, 75, 2286–2297.
75 N. Nagayama, O. Takaoka, K. Inomata and Y. Yamagata,
Origins Life Evol. Biosphere, 1990, 20, 249–257.
76 G. Chiavari and G. C. Galletti, J. Anal. Appl. Pyrolysis, 1992,
24, 123–137.
50 L. L. Shipman and R. E. Christoffersen, Proc. Natl. Acad.
Sci. U. S. A., 1972, 69, 3301–3304.
77 K.-M. Hansson, J. Samuelsson, L.-E. Åmand and C. Tullin,
Fuel, 2003, 82, 2163–2172.
51 P. Chaudhuri and S. Canuto, J. Mol. Struct.: THEOCHEM,
2008, 849, 25–32.
52 R. Importa, V. Barone, K. N. Kudin and G. E. Scuseria,
J. Chem. Phys., 2001, 114, 2541–2549.
78 K. J. Voorhees, W. Zhang, A. D. Hendricker and
B. Murugaverl, J. Anal. Appl. Pyrolysis, 1994, 30, 1–16.
79 K.-M. Hansson, L.-E. Åmand, A. Habermann and F. Winter,
Fuel, 2003, 82, 653–660.
53 S. Ohnishi, H. Kamikubo, M. Onitsuka, M. Kataoka and
D. Shortle, J. Am. Chem. Soc., 2006, 128, 16338–16344.
54 F. H. C. Crick and A. Rich, Nature, 1955, 176, 780–781.
55 A. V. Kajava, Acta Crystallogr., Sect. D: Biol. Crystallogr.,
1999, 55, 436–442.
80 K.-M. Hansson, J. Samuelsson, C. Tullin and L.-E. Åmand,
Combust. Flame, 2004, 137, 265–277.
81 C. Prasad, Peptides, 1995, 16, 151–164.
82 M. B. Martins and I. Carvalho, Tetrahedron, 2007, 63,
9923–9932.
56 E. V. Boldyreva, V. A. Drebushchak, T. N. Drebushchak,
I. E. Paukov, Y. A. Kovalevskaya and E. S. Shutova, J. Therm.
Anal. Calorim., 2003, 73, 409–418.
83 A. D. Borthwick, Chem. Rev., 2012, 112, 3641–3716.
84 Y. Wang, P. Wang, H. Ma and W. Zhu, Expert Opin. Ther.
Pat., 2013, 23, 1415–1433.
57 J. D. Bernal, Z. Kristallogr., Kristallgeom., Kristallphys.,
Kristallchem., 1931, 78, 363–369.
85 M. Ginz and U. H. Engelhardt, J. Agric. Food Chem., 2000,
48, 3528–3532.
58 E. W. Hughes and W. J. Moore, J. Am. Chem. Soc., 1949, 71,
2618–2623.
59 F. V. Lenel, Z. Kristallogr., Kristallgeom., Kristallphys.,
Kristallchem., 1931, 81, 224–229.
60 F. Colonna-Cesari, S. Premilat and B. Lotz, J. Mol. Biol.,
1974, 87, 181–191.
86 M. Z. Chen, M. L. Dewis, K. Kraut, D. Merritt, L. Reiber,
L. Trinnaman and N. C. Da Costa, J. Food Sci., 2009, 74,
C100–C105.
87 M. Gautschi, J. P. Schmid, T. L. Peppard, T. P. Ryan,
R. M. Tuorto and X. Yang, J. Agric. Food Chem., 1997, 45,
3183–3189.
61 V. Lalitha and E. Subramanian, Cryst. Struct. Commun.,
1982, 11, 561–564.
88 T. Stark and T. Hofmann, J. Agric. Food Chem., 2005, 53,
7222–7231.
62 I. L. Karle and J. Karle, Acta Crystallogr., 1963, 16, 969–975.
63 T. Asakura, M. Okonogi, K. Horiguchi, A. Aoki, H. Saitô,
D. P. Knight and M. P. Williamson, Angew. Chem., Int. Ed.,
2012, 51, 1212–1215.
64 R. D. Suenram and F. J. Lovas, J. Mol. Spectrosc., 1980, 72,
372–382.
65 M. J. Locke and R. T. McIver Jr., J. Am. Chem. Soc., 1983,
105, 4226–4232.
66 Y. Ding and K. Krogh-Jespersen, Chem. Phys. Lett., 1992,
199, 261–266.
89 As shown in this study, a number of reactions may be
induced upon thermal treatment of oligoglycines. In most
experiments cyclization is accompanied by side degradation
reactions which are difficult to prevent. The observation of
cyclization as a single process in DSC experiments may be
attributed to the small size and purity of the samples, as
well as the dynamic nature of this technique.
90 CRC Handbook of Chemistry and Physics, ed. D. R. Lide,
CRC Press, Boca Raton, 88th edn, 2007–2008, p. 6–4.
91 V. G. Badelin, O. V. Kulikov, V. S. Vatagin, E. Udzig,
A. Zielenkiewicz, W. Zielenkiewicz and G. A. Krestov,
Thermochim. Acta, 1990, 169, 81–93.
67 D. Yu, D. A. Armstrong and A. Rauk, Can. J. Chem., 1992,
70, 1762–1772.
68 J. H. Jensen and M. S. Gordon, J. Am. Chem. Soc., 1995, 117,
8159–8170.
92 G. Barone and R. Puliti, J. Therm. Anal. Calorim., 1999, 57,
119–132.
CrystEngComm
This journal is © The Royal Society of Chemistry 2014

View Article Online
CrystEngComm
Paper
93 E. Giralt, R. Eritja, J. Josa, C. Kuklinski and E. Pedroso,
Synthesis, 1985, 181–184.
105 S.-Y. Lin and S.-L. Wang, Adv. Drug Delivery Rev., 2012, 64,
461–478.
94 H. J. Svec and G. A. Junk, J. Am. Chem. Soc., 1964, 86,
2278–2282.
95 I. Lüderwald, M. Przybylski, H. Ringsdorf and D. Silberhorn,
Z. Naturforsch., B: Anorg. Chem., Org. Chem., 1978, 33, 805–808.
96 A. S. Noguerola, B. Murugaverl and K. J. Voorhees, J. Am.
Soc. Mass Spectrom., 1992, 3, 750–756.
106 N. F. Sepetov, M. A. Krymsky, M. V. Ovchinnikov,
Z. D. Bespalova, O. L. Isakova, M. Soucek and M. Lebl,
Pept. Res., 1991, 4, 308–313.
107 P. Xia, C. Wang and C. Qi, Chin. J. Chem., 2013, 31, 813–818.
108 Y.-D. Cheng and S.-Y. Lin, J. Agric. Food Chem., 2000, 48,
631–635.
97 M. Langhammer, I. Lüderwald and A. Simons, Fresenius' Z.
Anal. Chem., 1986, 324, 5–8.
109 J. Templier, N. Gallois and S. Derenne, J. Anal. Appl. Pyrolysis,
2013, 104, 684–694.
98 D. Gross and G. Grodsky, J. Am. Chem. Soc., 1955, 77, 1678–1680.
99 V. A. Basiuk and T. Y. Gromovoy, Collect. Czech. Chem.
Commun., 1994, 59, 461–466.
100 P. M. Fisher, J. Pept. Sci., 2003, 9, 9–35.
101 P. Gomez, N. Vale and R. Moreira, Molecules, 2007, 12,
2484–2506.
110 C. H. Görbitz, Acta Crystallogr., Sect. B: Struct. Sci., 2002,
58, 849–854.
111 M. Akazome, A. Hirabayashi, K. Takaoka, S. Nomura and
K. Ogura, Tetrahedron, 2005, 61, 1107–1113.
112 D. V. Soldatov, I. L. Moudrakovski, E. V. Grachev and
J. A. Ripmeester, J. Am. Chem. Soc., 2006, 128, 6737–6744.
113 T. J. Burchell, D. V. Soldatov, G. D. Enright and J. A. Ripmeester,
CrystEngComm, 2007, 9, 922–929.
102 L. Pérez-Picaso, J. Escalante, H. F. Olivo and M. Y. Rios,
Molecules, 2009, 14, 2836–2849.
103 M. Jainta, M. Nieger and S. Bräse, Eur. J. Org. Chem., 2008,
5418–5424.
104 Nonappa, K. Ahonen, M. Lahtinen and E. Kolehmainen,
Green Chem., 2011, 13, 1203–1209.
114 R. Anedda, D. V. Soldatov, I. L. Moudrakovski, M. Casu and
J. A. Ripmeester, Chem. Mater., 2008, 20, 2908–2920.
115 S. S. Leung and D. J. W. Grant, J. Pharm. Sci., 1997,
86, 64–71.
This journal is © The Royal Society of Chemistry 2014
CrystEngComm