Subscriber access provided by - Access paid by the | UCSB Libraries
A: Spectroscopy, Molecular Structure, and Quantum Chemistry
Conformational Preferences of Isolated Glycylglycine (Gly-Gly) Investigated with IRMPD-VUV Action Spectroscopy and Advanced Computational Approaches Vasyl Yatsyna, Ranim Mallat, Tim Gorn, Michael Schmitt, Raimund Feifel, Anouk M. Rijs, and Vitali Zhaunerchyk J. Phys. Chem. A, Just Accepted Manuscript • DOI: 10.1021/acs.jpca.8b10881 • Publication Date (Web): 04 Jan 2019 Downloaded from http://pubs.acs.org on January 4, 2019
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
Conformational Preferences of Isolated Glycylglycine (Gly-Gly) Investigated with IRMPD-VUV Action Spectroscopy and Advanced Computational Approaches †,‡
Vasyl Yatsyna,
†
Ranim Mallat,
Tim Gorn,
Anouk M. Rijs,
∗,‡
¶
¶
Michael Schmitt,
and Vitali Zhaunerchyk
Raimund Feifel,
†
∗,†
†Department of Physics, University of Gothenburg, 412 96 Gothenburg, Sweden ‡Radboud University, Institute for Molecules and Materials, FELIX Laboratory, Toernoovield 7-c, 6525 ED Nijmegen, The Netherlands
¶Heinrich-Heine-Universität,
Institut für Physikalische Chemie I, D-40225 Düsseldorf, Germany
E-mail:
[email protected];
[email protected] Abstract
1
In this paper we report the results of gas-phase
Gas-phase infrared (IR) spectroscopy of small
IR spectroscopy of neutral glycylglycine (Gly−1 Gly) in the 700 − 1850 cm frequency range.
and medium-sized peptides
A combination of laser desorption, jet-cooling,
level.
and IR multiple-photon dissociation vacuum-
techniques to transfer intact biomolecules into
ultraviolet (IRMPD-VUV) action spectroscopy
the gas-phase, IR spectroscopy became a rou-
is employed, together with extensive quantum-
tine tool for studies of biomolecular systems of
chemical calculations that assist in the analy-
increasing size and complexity.
sis of the experimental data.
Introduction 18
into protein folding forces
allows an insight
912
at the atomic
Thanks to the development of various
13
It has the ad-
As a result, we
vantage of a high sensitivity of molecular vi-
determined that the most favorable conformer
brations to the backbone structure and non-
in the low-temperature environment of the su-
covalent interactions such as hydrogen bonding.
personic jet is the nearly-planar structure with
Since low sample density in the gas phase re-
two C5 hydrogen bonding interactions. Calcu-
stricts direct IR absorption measurements, ac-
lations clearly show that this conformer is fa-
tion spectroscopy techniques are employed in-
vored due to its exibility (considerable entropy
stead.
stabilization), as well as ecient conformer re-
absorption is detected through a change in ion-
laxation processes in the jet.
ization or uorescence yield, electron detach-
To gain more
In action spectroscopy the IR photon
14
understanding into the relative stability of the
ment or fragmentation.
lowest-energy Gly-Gly conformers, the relative
UV spectroscopy is a powerful method that al-
strength of hydrogen bonding and steric inter-
lows recording IR spectra of individual molec-
actions is analyzed using the non-covalent in-
ular conformers.
teractions (NCI) approach.
only be applied to molecules with an aromatic
Double-resonance IR-
This method, however, can
UV chromophore. Among the twenty standard amino acids only three have such a UV chromophore, which signicantly limits the scope
ACS Paragon Plus Environment 1
The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 20
of the peptides which can be studied. For the
lated in our experimental conditions, and com-
spectroscopy of charged species several tech-
pare them to theoretical calculations in order
niques are available, such as tagging (messen-
to identify the most abundant conformers. Al-
ger) technique and IR multiple-photon disso-
though the recent rotational spectroscopy study
ciation (IRMPD), whereas IR spectroscopy of
of Gly-Gly observed three dierent conform-
neutral isolated peptides without an aromatic
ers,
ring remains challenging. One way to circum-
leaves some open questions, for instance, about
vent this problem is to attach an aromatic moi-
the relative stability of the observed folded and
ety, either chemically
planar structures, especially in view of favor-
interactions.
16
15
or through non-covalent
While this approach has found
26
their abundances were not reported. This
able entropic stabilization
28,29
in the latter. In
wide application, the attached aromatic group
this respect, the current work will complement
can modify the intrinsic properties of the pep-
the previous rotational spectroscopy results.
tide by introducing extra non-covalent interac-
Secondly, by studying Gly-Gly we intend to
tions which may also alter its conformational
explore the use of the IRMPD-VUV approach
landscape.
for structural analysis of exible peptides with
Recently, combination
we
have
of
demonstrated
IRMPD
that
spectroscopy
a large degree of conformational heterogene-
the a
ity. Thirdly, we aim to shed some light on the
molecular beam with single-photon vacuum-
strength and importance of non-covalent inter-
ultraviolet (VUV) ionization can be used to
actions that are present in the Gly-Gly con-
record vibrational spectra of cooled neutral
formers. For this purpose, we will employ elec-
molecules of arbitrary structure, without an
tron density topological analysis within atoms
aromatic chromophore in particular.
in molecules (AIM)
17
of
26
In the
current work we apply this approach to IR
3032
and non-covalent inter-
actions (NCI) approaches.
3335
spectroscopy of chromophore-free glycyl-glycine (Gly-Gly).
An
initial
challenge
in
obtain-
2
ing gas-phase vibrational spectra for Gly-Gly
Methods
and other thermolabile molecules is how to
2.1
volatilize an intact molecule. For the studies of charged Gly-Gly species electrospray ionization
The experiments were performed at the free
methods were employed yielding protonated or deprotonated molecular targets.
1822
electron laser FELIX Laboratory at Radboud
In this
University, Nijmegen, The Netherlands, using
case, however, one has to disentangle the ef-
the laser desorption molecular beam set-up.
fects of the proton on the peptide's intrinsic
from Sigma Aldrich (purity of 98%). The sam-
Laser ablation method is commonly
ple was mixed with carbon black and applied
used for rotational spectroscopy of neutral ther-
onto a graphite bar that was placed in front
molabile molecules, though in many cases its
of a pulsed valve nozzle (0.5 mm orice, Jor-
application to peptides and amino-acids is as-
dan TOF Products, Inc) in a source vacuum
sociated with signicant photofragmentation eects.
2325
chamber. The sample molecules were laser des-
Only recently a rotational spec-
orbed from the bar using a Nd:YAG laser (1064
troscopy study of laser-ablated Gly-Gly has been reported.
26
nm, 5 ns, 1−2 mJ/pulse, Polaris II, New Wave
In our work, we employ a
Research) and then seeded into a supersonic-
soft laser desorption process, which imparts minimal
internal
molecules,
14,27
energy
into
the
14
A Gly-Gly sample was obtained commercially
structural properties and its vibrational signatures.
Experiment and data analysis
jet created by a pulsed valve with argon (3
desorbed
bar) as a carrier gas.
followed by collisional cooling
The supersonic-jet ex-
pansion allowed the desorbed molecules to ef-
in an argon supersonic jet.
ciently cool down to the rovibrational ground
The purpose of the current work is three-
state.
fold. Firstly, we aim to obtain the IR signatures
The molecules in the central and cold-
est part of the expansion region were colli-
of the ensemble of Gly-Gly conformers popu-
ACS Paragon Plus Environment 2
Page 3 of 20
mated by a skimmer and delivered to the in-
Ar+
teraction chamber, where the molecular beam
Intensity, arb. units
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
was crossed with a pulsed IR beam from FELIX free electron laser (FEL). The IR frequency of FELIX was scanned in the range of 700-1850 −1 cm , corresponding to IR pulse energies of 30-80 mJ. The exposure time to the IR FEL radiation within the interaction region is di-
IR off IR on
30
30.2
132 132.2 132.4
rectly proportional to the mass of the carrier gas; the heavier the gas the longer the exposure time.
20 30 40 50 60 70 80 90 100 110 120 130 140
mass, amu
With a typical FELIX macropulse
duration of 6-8
µs,
the sample molecules car-
Figure 1:
ried by argon had an estimated residence time of 3
µs
desorbed Gly-Gly molecules under IR on (red
in the interaction region. If a molecu-
curve) and IR o (black curve) conditions.
lar vibrational transition was resonant with the
The inset plots show how the fragment peak
scanned IR light frequency, the molecule dis-
at 30 amu and parent peak at 132 amu change
sociated via InfraRed Multiple Photon Dissociation (IRMPD) process.
VUV (10.5 eV) mass spectra of laser-
their intensities when the IR FELIX wavelength
At the end of the
is resonant with a vibrational transition of Gly-
FELIX macropulse, the IRMPD fragments and
Gly.
parent molecules were ionized by 118 nm VUV laser pulses of ∼2-3 ns duration (hν =10.49 eV, th 3+ 9 harmonic of a Nd :YAG laser, Spectra-
Assuming a constant interaction volume and
The created ions were detected by
a constant time prole of macro-pulses upon
means of a reectron-type time-of-ight mass
scanning the FELIX wavelength, a relative ab-
spectrometer. By measuring the IRMPD frag-
sorption cross section versus IR frequency can
mentation yield versus FELIX wavelength, the
be expressed through the experimentally mea-
vibrational IRMPD-VUV spectrum of Gly-Gly
sured quantities as
Physics).
was obtained. The FELIX laser was operated
σrel
at 5 Hz, while the desorption laser, the pulsed valve and the VUV ionizing laser were operated at 10 Hz.
This enabled measuring refer-
Poff , Pon
(1)
Nph denotes the number of photons in the pulse, P denotes the intensity of the parent
where
ence mass spectra without IR radiation at every
IR
second VUV laser pulse (see Fig. 1).
ion, and the on and o subscripts refer to
The VUV laser light was generated by focus3+ ing 355 nm light (third harmonic of Nd :YAG
the measurements performed with FELIX being on and o, respectively. If all IRMPD frag-
laser) into the center of a gas cell containing a 1:10 mixture of Xe and Ar gases.
1 ln = Nph
ments
The result-
equal
ing 118 nm laser beam was refocused into the
P Fi,on are detected, P to Fi,on + Pon , and
the
Poff
signal is
Eq. 1 can be ex-
pressed as
molecular interaction region through a MgF2 lens. The residual 355 nm beam remained di-
σrel
vergent and was not focused in the interaction
1 = ln Nph
P
Fi,on + Pon . Pon
(2)
region since the focal distance of the MgF2 lens was larger for 355 nm. To further separate the
Eq. 2 is more preferable over Eq. 1 since it only
118 nm and 355 nm beams we used o-axis
involves quantities measured with FELIX be-
alignment through the MgF2 lens.
The esti-
ing on, and hence allows eliminating shot-to-
mated pulse energy of the generated VUV ra-
shot uctuations of the laser desorption source.
≤ 1 µJ, based on the conversion 10−4 in the phase-matched Xe-Ar
To implement Eq. 2 into the current measure-
diation was eciency of mixture.
36,37
ments, it should be corrected for a fraction of fragments produced solely by VUV in the ab-
ACS Paragon Plus Environment 3
The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
H11
sence of the FELIX pulses. The corrected Eq.
Page 4 of 20
H12 N1
2 takes the form
O'
1 2 1 , 1
σrel
1 ln = Nph
P
(1 − α)
Fi,on + Pon Pon
C1
(3)
1
ω1
α
C2
T 2
H''
θTC
N O
N2
C1 where
C
H2
O''
C2
H
denotes the fraction of the parent
O1
molecules that undergo fragmentation when the molecules are not irradiated with FELIX. This
Figure 2:
fraction can be found as
the cis-trans backbone conguration at the N-
P
Fi,off α= P Fi,off + Poff We note that Eq.
Structure of the Gly-Gly molecule in
and C-termini, respectively (ψ1 close to
(4)
φ2
close to
180°).
0°
and
3 neglects the multiple-
photon nature of the IRMPD process.
pulsion in genuine PM6.
Nev-
The resulting struc-
ertheless, it is valid for our measurements as-
tures were sorted according to their relative en-
suming that IRMPD signals are only observed
ergies, and duplicate structures were discarded.
when the IR frequency is resonant with a molec-
The geometries of obtained unique structures
ular vibration and the IR intensity is suciently
(≈80) with relative energies below 1600 kJ/mol
high to dissociate the molecule.
were
optimized
with
higher
accuracy
DFT
methods using the quantum chemistry package
The Gly-Gly raw spectra were obtained by
Gaussian 16.
42
The selection of DFT function-
scanning the FELIX IR laser in the range of −1 700-1850 cm , and measuring the relative in-
als was based on their performance for sim-
tensities of the IRMPD fragments and the par-
ilar dipeptide systems.
ent molecule. The nal spectrum (Fig. 4) was
ods (ω B97X-D/6-311++G(d,p),
obtained by applying Eq. 3 to the data and tak-
31+G(d),
45
43
The selected meth-
44
M06-2X/6-
MP2/aug-cc-pVTZ and B3LYP-
46,47
ing into account the FELIX power prole across
D3BJ/N07D
the measured range.
As there were no varia-
paring theoretical predictions with the Gly-
(see Eq. 4) when scanning FELIX,
Gly experimental rotational constants available
tions in
α
a constant
α
elsewhere
factor was used in the analysis
of each FELIX scan.
≈ 1.6
(see Table S1 of Supporting Infor-
mation (Supp.
The mean width of the
measured IR bands (FWHM of
26
) were further assessed by com-
Info)).
determined that the
% of the
As a result, we have
ω B97X-D
functional with
center frequency) was slightly larger than the
the 6-311++G(d,p) basis set is the most reli-
bandwidth of the FELIX laser (0.5−1 %). This
able for structure calculations of the Gly-Gly
is related to the multiple-photon excitation in-
peptide. The lowest energy Gly-Gly conformers found
volved in the IRMPD process that is known to broaden the vibrational transitions.
38
with our conformational search are shown in Fig. 3.
2.2
Their energies and dihedral angles,
denoted in Fig. 2, are presented in Table 1.
Theoretical treatment
We found that the backbone of most of the conformers deviates from planarity by at least
To explore the conformational space of GlyGly a random search method was applied.
39
10°, with one exception of
A
sc-cis-trans
struc-
large number of structures (≈657000) with ran-
ture that is nearly planar.
dom torsional angles for rotatable bonds was
non-planar conformer exists as a pair of enan-
generated and optimized with a semi-empirical
tiomeric structures with identical energy.
quantum-chemical method PM6-D3H4
40
Moreover, each As
imple-
a result, all structures shown in Fig. 3 have a
program package.
degeneracy factor of 2, except for the planar
PM6-D3H4 corrects for dispersion, hydrogen
sc-cis-trans
conformer which has a degener-
bonding, and the too small hydrogen steric re-
acy factor of 1.
mented in the MOPAC
41
ACS Paragon Plus Environment 4
Page 5 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
C5-trans
C5g-(s)
trans-trans
C5g-(a)
C5g+(s)
C5C7-
C5C7+
sc-cis-trans
g-trans-trans
C5g+(a)
−1 Figure 3: The structures of the lowest energy conformers of Gly-Gly (∆(E +ZP E) below 840 cm , 10 kJ/mol) optimized with the
ω B97X-D/6-311++G(d,p)
method.
Table 1: Zero-point energy corrected electronic energies ∆(E + ZP E), Gibbs free energies ∆G at 400 K, and dihedral angles of the lowest-energy conformers of Gly-Gly shown in Fig. 3, as obtained from the ωB97X-D/6-311++G(d,p) optimized geometries. See Fig. 2 for the description of the dihedral angles. Energies are calculated using the G4MP2 method based on the ωB97X-D/6-311++G(d,p) geometries and harmonic frequencies, and are presented with respect to the C5-trans energies in cm−1 (values in kJ/mol are shown in brackets). ∆(E + ZP E), cm−1 (kJ/mol) 0 (0) 90 (1.1) 179 (2.1) 22 (0.26) 12.5 (0.15) 408 (4.9) 592 (7.1) 617 (7.4) 664 (7.9) 755 (9.0)
C5-trans C5g-(s) C5g+(s) C5C7+ C5C7trans-trans C5g-(a) C5g+(a) g-trans-trans sc-cis-trans
∆G at 400 K,
cm−1 (kJ/mol) 0 (0) 177 (2.1) 199 (2.4) 485 (5.8) 474 (5.7) 378 (4.5) 613 (7.3) 809 (9.7) 876 (10.5) 506 (6.1)
ψ1
ω1
φ2
ψT
θT
φ11
φ21
-13.2 -20.1 -10.5 -10.2 -11.4 -170.1 -18.0 -11.2 -151.5 -0.01
-179.1 170.3 -169.7 175.9 -173.1 -177.5 175.1 -172.6 -176.8 -180.00
-173.5 -91.8 90.3 74.1 -74.6 179.5 -92.6 89.6 -179.1 -179.99
-0.9 -2.0 2.6 123.1 -122.8 -0.1 -176.5 177.6 0.5 0.00
179.7 -179.7 179.7 1.0 -1.2 180.0 178.5 -178.5 -179.8 180.00
150.4 156.0 147.9 142.5 144.0 -52.4 152.6 149.4 -155.1 -64.1
-88.4 -83.4 -91.1 -96.1 -94.7 63.6 -86.9 -89.9 -35.3 64.1
For the assignment of the experimental IR bands,
spectra
of
all
low-energy
of Gly-Gly were calculated.
formers upon laser desorption and subsequent
structures
supersonic-jet expansion is well reected by
For this pur-
a relative Gibbs free energy distribution at
pose, harmonic frequencies and intensities were
300 − 500
calculated with the B3LYP-D3BJ functional
signments based on electronic energies with
(Becke, three-parameter, Lee-Yang-Parr with
zero-point energy (ZPE) corrections should be
D3 version of Grimme's dispersion with Becke-
avoided.
Johnson damping
Gibbs
46
) and the N07D basis set.
47
K, while the conformational as-
2729,5254
free
To calculate highly-accurate
energies
we
used
methods
modication described in Ref. 43:
structures using Vibrational second-order Per-
single-point energy calculations were based on
turbation
molecular geometries and frequency calcula-
48
The
B3LYP-
the
ω B97X-D/6-
harmonic and anharmonic frequency calcula-
311++G(d,p) and B3LYP-D3BJ/N07D meth-
tions as it provides high accuracy within rea-
ods.
sonable computational time.
more reliable for geometry and frequency cal-
These methods were chosen as they are
culations
43
ACS Paragon Plus Environment 5
both
a
tions
It was shown that the population of con-
using
with
high-level
D3BJ/N07D method was chosen both for the
4851
obtained
G4MP2
composite
56
calculated for the most important low-energy (VPT2).
and
the
Anharmonic frequencies and intensities were
Theory
CBS-QB3
55
than the default methods used in
The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
CBS-QB3 and G4MP2. The
lowest-energy
Page 6 of 20
critical points (RCPs) associated with ring-type
vibrational
modes
are
structures correspond to second order saddle points (λ1
approximation,
which can have an adverse
the AIM analysis, useful for the identication
eect on the accuracy of calculated Gibbs free
of strong hydrogen bonds, we have also ap-
energies.
plied the NCI (non-covalent interactions) ap-
57,58
Treatment of some low-energy tor-
33
< 0, λ2
λ3 > 0). 32
known to be poorly described by the harmonic
and
Following
sional vibrations as hindered rotations provides
proach.
a better description.
Therefore, in our analy-
ration of NCI index isosurfaces
sis we applied an automatic procedure available
colormap from sign(λ2 )·ρ data.
in Gaussian
dex corresponds to the reduced density gradient
59
59
to identify hindered rotations,
and corrected the calculated free energy using
(RDG):
This approach is based on the explo-
where
r) =
RDG(
hind hind (∆Evib + ∆Evib ) − T (∆Svib + ∆Svib )
hind ∆Evib
and
hind ∆Svib
where
are the hindered ro-
tor approximation corrections (McClurg
60
lled with a The NCI in-
33
the following relation:
∆Gvib =
33
ρ(r)
|∇ρ(r)| 1 , 2 1/3 2(3π ) ρ(r)4/3
is an electron density map that can
be obtained from DFT calculations.
) to
In our
the vibrational thermal energy and entropy, re-
work we have used the electron density data
spectively.
from the
ω B97X-D/6-311++G(d,p)
method as
this method has been found to provide the
For the calculation of relaxation barriers between the conformers, a transitions state (TS)
best description for Gly-Gly geometries.
search was performed using the QST2 and
is also worth noting that any other DFT or
QST3 methods
as implemented in Gaussian.
wavefunction-based methods with sucient ba-
In the cases where QST2 failed to nd a TS,
sis sets are expected to be equally reliable for
the QST3 method was applied, for which an
the NCI analysis.
61
It
35
initial guess for the TS structure was obtained from the relaxed PES scan connecting two conformers.
3
To obtain the barriers to relaxation,
accurate electronic and Gibbs free energy calcu-
3.1
lations were performed by applying the G4MP2
Results and Discussion IR spectroscopy: experiment,
method to the optimized TS structures and
theory and Gly-Gly conform-
the corresponding conformers.
ers
The geome-
try optimization and hindered rotor frequency analysis were performed with the
The
ω B97X-D/6-
experimental
IRMPD
spectrum
of
jet-
cooled Gly-Gly is presented in Fig. 4 (the black
311++G(d,p) method. The identication of hydrogen bonding and
trace in each plot). The IRMPD-VUV method
other weak non-covalent interactions was per-
is not conformer-selective so all conformers suf-
formed using the electron density
topol-
ciently populated in the molecular beam con-
ogy analysis with the help of the Multiwfn soft-
tributed to the measured spectrum. For com-
ware.
In the rst step, the search of bond
parison, Fig. 4 presents the calculated scaled
critical points (BCPs) within the AIM (atoms
harmonic frequency spectra of several charac-
in molecules)
theory was undertaken. BCPs
teristic lowest-energy Gly-Gly conformers (the
are found as saddle points in the electron den-
62
30
ρ(r)
representing a minimum in the
colored bar spectra). As vibrational transitions −1 at frequencies below 1100 cm have lower in-
bonding direction and maximum in all other di-
tensities, the two dierent frequency ranges are
rections.
Mathematically, this is described by
plotted in panels (a) and (b) of Fig. 4. Com-
λ1 , λ2 , λ3 of the secondorder derivative matrix of ρ(r). At BCPs, λ1 and λ2 are negative while λ3 is positive. Ring
parison with other conformers that are higher
sity map
ρ(r),
32
the sign of eigenvalues
in energy or have similar spectra can be found in Fig. S1 of the Supp.
ACS Paragon Plus Environment 6
Info.
It can be seen
Page 7 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
from Fig. 4 that most of the experimental fea-
spectral features. For example, the calculated
tures are readily reproduced by the calculated Moreover, the mea-
frequency of the NH2 scissoring vibration for trans-trans (1638 cm−1 ) is blue-shifted with −1 respect to the experiment (1620 cm ). More-
sured spectrum shows considerable dierences
over, the CC stretching vibration has a more
with the calculated spectra of all other low-
delocalized character in the
energy conformers (see also Fig. S1 of the Supp.
former than in
Info). This implies that the measured spectrum
quency for
spectrum of
C5-trans
(red bars), which is the
lowest energy conformer.
C5-trans .
is dominated by
Such a conclusion
trans-trans conC5-trans . Its calculated fretrans-trans (843 cm−1 ) does not
agree well with the experimental value of 863 −1 −1 cm , whereas C5-trans (857 cm ) does. The
is elaborated in more detail in what follows. The calculated spectra of all conformers pre-
NH2 wagging (inversion) vibration involves a
dict that the strongest experimental bands at −1 1782 and 1704 cm correspond to the C=O
larger extent of NC and CC stretching in
stretching fundamental transitions. The latter
quency (895 cm
trans-
trans ,
resulting in a higher calculated fre−1 ) that does not match the −1 observed transition at 795 cm . In summary,
band, known as Amide I, originates from amide (peptide link) C=O stretching. Its position −1 at 1704 cm indicates that the the observed
the calculated peak positions and intensities of
spectrum is dominated by conformers with free
ment with the experimental spectrum whereas
amide C=O group, implying no hydrogen bond-
other conformers do not match to various de-
ing at this site. This allows us to conclude that
grees.
the
C5C7-
and
C5C7+
the
conformers, that have
C5-trans
conformer are in favorable agree-
Predominance of the
C5-trans
conformer
a hydrogen bonded amide C=O group, do not
can be further veried by accounting for an-
contribute signicantly to the spectrum.
harmonicity using the VPT2 approach (Fig. 5).
The
question
as
to
what
extent
planar
The VPT2 calculation for the
trans-trans ) and non-planar (C5g-(s) and C5g-(a) ) conformers are popu-
C5-trans
con-
(C5-trans and
former
lated can be assessed using spectral features in
band positions, yielding a mean absolute error −1 (MAE) as low as 6 cm . The calculated over-
accurately
reproduces
experimental
the so-called ngerprint frequency range below −1 1400 cm . For example, the measured strong −1 peak at 1393 cm reects the COH bending
tone and combination bands (Fig. 5) also elim-
vibration of planar Gly-Gly conformers. In the case of non-planar conformers the calculated
For example, a relatively strong band observed −1 at 1224 cm was not predicted by the har-
transition for this vibration is red-shifted and
monic frequency calculations.
signicantly reduced in intensity with respect
a strong band at this frequency corresponding
to the planar counterparts. The normal mode
to the overtone of OH wagging+COH bending
analysis also shows that two modes including
vibration.
both CN and CO stretching vibrations, ob−1 served at 1153 and 1120 cm , are delocalized
of the OH wagging vibration
inate some discrepancies between the scaled harmonic frequency and experimental spectra.
VPT2 predicts
The reported high anharmonicity
50,63
supports this
assignment. The observation of a single conformer,
and are sensitive to the planarity of Gly-Gly.
C5-
The intensities and the frequency spacing be-
trans ,
tween the calculated transitions of these modes
disagrees with theoretical abundances of the
in the non-planar conformers dier from the ex-
lowest-energy conformers (Fig. 6(a)), calculated
periment. The same holds for the CC stretching
from Gibbs free energy data at dierent tem-
vibrations.
peratures.
Based on these considerations, we
C5g-(s)
and
C5g-(a) ,
are not so
500 K,
spectroscopy
study
27
where the calculated population of conformer is less than 40%.
In or-
C5-trans
der to explain this seeming contradiction, con-
can be distinguished by other
formational relaxation processes occurring in
The remaining planar conformers
trans-trans
IRMPD
The temperature of laser-desorbed
C5-trans
important under our experimental conditions. and
our
molecules is expected to be in the range of 350-
can conclude that the non-planar conformers, such as
in
ACS Paragon Plus Environment 7
The Journal of Physical Chemistry NH2 CC wagg. str.
100
CN str. (N term.)
CH2 wagg.
CN,CO str. Am III
400
Experimental C5-trans
0 100
0 400
Experimental C5C7-
NH2 Am I C=O str. (COOH) sciss.
E ( G) [cm-1]
Experimental C5-trans
0 (0)
13 (474) Experimental C5C7-
200
50 0 100
0 400
Experimental C5g-(s)
Experimental C5g-(s)
90 (177)
Experimental trans-trans
408 (378)
Experimental C5g-(a)
592 (613)
200
50 0 100
0 400
Experimental trans-trans
200
50 0 100
0 400
Experimental C5g-(a)
200
50 0 700
0 800
900
1000 1100
1100 1200 1300 1400 1500 1600 1700 1800
wavenumber, cm -1
wavenumber, cm -1
(b)
(a) Figure 4:
Am II
200
50
Relative intensity, arb. un.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 8 of 20
The experimental IRMPD spectra (the black trace in each plot) compared with the
scaled harmonic frequency spectra (B3LYP-D3BJ/N07D, scaling factor 0.976) of several lowest energy conformers of Gly-Gly (C5-trans is shown in red,
trans
in purple, and
C5g-(a)
C5C7- in green, C5g-(s)
in blue,
trans-
in orange, respectively). Note that the y-axis range in panel (a) is
shorter than in panel (b) to facilitate comparison between experimental and calculated vibrational transitions. Abbreviations: Am - Amide, str. - stretching, wagg. - wagging.
52,64
the supersonic jet need to be taken into acIndeed, it is known that if the barri-
have barriers lower than the critical value −1 of 800 cm . This suggests that conformer re-
ers to interconversion are low, ecient confor-
laxation via collisional cooling can occur along
mational relaxations to the lowest energy con-
these pathways in our molecular beam.
formers take place during supersonic-jet expan-
barriers calculated at higher temperatures, e.g.
sion.
Table 2 presents the calculated bar-
at 300 K were found to be much higher than
rier heights for the possible relaxation path-
those calculated at 15 K, with a single exception
ways of Gly-Gly conformers (see also Fig. S4
of
in Supp. Info). The data were obtained using
cold environment of the supersonic jet is hence
count.
52,6468
∆G
∆Eel ,
relaxation.
The
and Gibbs free ener-
favorable for collisional relaxation which would
at 15 K and 300 K. The barriers es-
not take place at higher temperatures. We also
electronic energies gies
sc-cis-trans →C5-trans
The
timated from
∆G
at 15 K are known to ade-
note that the small negative value of
∆G(15 K)
quately describe relaxation for the experimen-
calculated for pathway I is most likely due to a
tal conditions similar to ours,
and therefore
calculation limitation aecting the intrinsic ac-
can be used to assess the likelihood of relax-
curacy of the Gibbs free energy, which becomes
ation. As illustrated in Table 2, pathways I-VI
apparent when structures with very similar ge-
68
ACS Paragon Plus Environment 8
Page 9 of 20
200
C5-trans C5C7+ C5C7C5g-(s) C5g+(s) trans-trans C5g-(a) C5g+(a) sc-cis-trans g-trans-trans
50
Boltzmann population, %
VPT2 fundamental bands VPT2 overtones+combinations VPT2 total spectrum
180 160 140 Relative intensity
120 100
40
30
20
10
80
VPT2 total spectrum Experiment GlyGly
60
0 100
200
300
40
400 500 600 700 Temperature, K
0
800
800
1000
1200 1400 1600 wavenumber, cm−1
1800
80 C5-trans
70
Figure 5:
trans
900 1000
(a)
20
Anharmonic IR spectra for
C5-
Boltzmann populaion, %
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
conformer calculated with VPT2, com-
pared with the experimental IRMPD spectrum of Gly-Gly. The calculated spectra were convoluted with a Gaussian prole with FWHM of 1.5 % of central frequency to match the experimental conditions.
60 50 40 30 20
trans-trans
10
C5C7
0 100
ometries and energies are compared. According
200
300
to the electronic energy calculation, the barrier
400 500 600 700 Temperature, K
for this pathway clearly exists though it is very −1 low (∆Eel = 98 cm ).
800
900 1000
(b)
Our calculations also suggest that the in-
Figure 6: (a) The theoretical Boltzmann popu-
terconversion processes between the lowest en-
lations of the lowest energy conformers of Gly-
ergy conformers
C5C7±
C5-trans , trans-trans
and
Gly, estimated from Gibbs free energies calcu-
are not possible (see pathways VII-IX
lated at dierent temperatures. The Gibbs free
in Table 2 and Fig.
S5 of Supp.
Info).
Such
energy data is taken from G4MP2 calculations
ω B97X-D
processes require several dihedral angles (re-
which employed
laxation coordinates) to be altered and involve −1 high relaxation barriers (>800 cm ), which
quencies.
signicantly reduces the probability of conver-
relaxation processes in the argon jet. The con-
sion.
formers
Assuming
complete
relaxation
geometries and fre-
(b) The calculated Boltzmann pop-
ulations that take into account the conformer
52,64,6971
C5-trans , trans-trans
and
C5C7±
cannot undergo relaxation, but gain population
through pathways I-VI, the populations of the
from other conformers due to relaxation pro-
conformers that do not undergo relaxation, i.e.,
cesses I-VI (Table 2).
C5-trans , trans-trans
and
C5C7±,
temperature are presented in Fig. 6(b). that since
C5C7+
and
C5C7-
versus Note
C5C7 . From this C5-trans structure
as one structure denoted as
are very simi-
gure, it follows that the
lar in structure and energy, they are considered
is expected to be the most abundant under our
ACS Paragon Plus Environment 9
The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 10 of 20
Table 2: The relaxation barriers between the lowest energy conformers of Gly-Gly calculated as the dierences between energies of the corresponding conformers and their transition states, based either on the electronic energy values, ∆Eel (G4MP2 method), or based on Gibbs free energy values at 15 and 300 K, ∆G(15 K) and ∆G(300 K) (G4MP2 method, harmonic frequency and hindered rotation analysis from ωB97XD/6-311++G(d,p)). Note that int1 is the intermediate conformer between the C5C7and C5-trans , and is similar to C5-trans except for the θT dihedral that diers by 180°. Relaxation pathway I. C5g-(s) → C5-trans II. C5g+(s) → C5-trans III. C5g-(a) → C5g-(s) IV. C5g+(a) → C5g+(s) V. sc-cis-trans → C5-trans VI. g-trans-trans → trans-trans VII. C5C7- → int1 VIII. int1 → C5-trans IX. trans-trans → sc-cis-trans
experimental conditions. and
C5C7
Barrier to relaxation, cm−1 ∆G(15 K) ∆G(300 K) 98 -27 410 94 19 327 592 599 980 583 572 825 699 402 374 144 51 228 2023 1857 1999 3763 3466 3605 913 911 1334
∆Eel
φ2 φ2 ψT ψT
NH2 inversion φ1 θT , ψT , φ2 θT , ψT ψ1
3.2
The predicted ratio
between the populations of
trans
Relaxation coordinate
C5-trans , trans-
Energy dierence ∆(Eel +ZPE) 90 179 502 438 755 256 -1481 1494 -347
Hydrogen bonding in Gly-Gly conformers
conformers is 78:10:12 at 400
K. This result is in a good agreement with our
There are some debates in the literature about
previous conclusion, drawn from the measured
the presence and importance of hydrogen bond-
IRMPD spectrum, that
C5-trans
ing in glycine and alanine structures.
is the most
31,7274
The
functional groups of Gly-Gly structures are sim-
abundant. The three conformers that do not relax in
ilar to glycine and alanine and can show similar
the supersonic jet (Fig. 6(b)) were observed
weak non-covalent interactions. Therefore, we
in the recent rotational spectroscopy study of
employed AIM (atoms in molecules) and NCI
Gly-Gly,
(non-covalent interactions) approaches to iden-
26
though relative abundances of these
conformers were not reported.
tify weak interactions in the Gly-Gly conform-
The spectro-
ers.
scopic observation of presumably low abun-
C5C7-
conformers (see
Two types of relatively strong hydrogen bond-
Fig. 6(b)) is most probably related to the
ing interactions were identied using the AIM
high resolution and sensitivity of the microwave
approach (see Fig. S7 of Supp.
dant
and
trans-trans
(MW) spectroscopy technique.
Info).
The
rst type corresponds to H2 N· · · HN hydrogen
Although it
is not possible to directly compare conformer
bonding (r (N· · · H) = 2.2 Å) that was found in
abundances between the MW experiment and
most of the conformers with cis conguration at
ours, one can expect the two methods to yield
the N-terminus (see e.g. structures
dierent populations of the conformers for two
C5g-(s)
reasons.
action leads to the formation of a ring involv-
pulses,
The laser ablation method (ps laser
λ = 355
and
C5C7+
C5-trans ,
in Fig. 3).
Such inter-
ing ve atoms, a so-called C5 interaction.
nm) employed in the MW
If
study is known to result in evaporated molec-
such a hydrogen bond is absent, the cis cong-
ular ensemble with higher temperatures than
uration is less stable than trans. For example,
the one achieved with laser desorption method
the
employed by us (ns pulses,
λ = 1064
nm).
sc-cis-trans trans-trans (see
27
Table 1). Another strong in-
teraction identied by the AIM approach cor-
Moreover, the MW study used neon (15 bar) as
responds to OH· · · O=C hydrogen bonding at
a carrier gas, which is known to result in less ef-
69,71
structure is less stable than
compared
the C-terminus (r (H· · · O) = 1.7 Å) that forms
to argon that is used in the current measure-
a seven-membered ring which was denoted with
ments.
C7. This interaction stabilizes the
cient conformational relaxation
ACS Paragon Plus Environment 10
C5C7+
and
Page 11 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
C5C7-
conformers. The relative distances be-
character
(e.g.
sc-cis-trans and transC7 character (C5C7+ ). The
tween the BCPs and RCPs identied by the
trans ),
AIM analysis indicates that the C7 interaction
NCI analysis showed that the formation of the
is stronger than C5.
C5 ring is associated with considerable steric
or strong
Weaker hydrogen bonding interactions that
repulsion eects which weaken the structure
could not be identied with the AIM method
stabilization by the hydrogen bond. It was also
were
found that the C5 interactions at the N- and
found
with
the
help
of
the
NCI
ap-
proach introduced by Jonson and co-workers.
33
C-termini show some competition due to shar-
Fig. 7(a,c) illustrates the identied hydrogen
ing the same donor (amide NH group).
bonding interactions (blue and green colored is-
example, the C5 interaction at the N-terminus
lands) for the
C5C7+
and
trans-trans
struc-
of the
C5-trans
conformer (ρ
tures. Each interaction corresponds to a min-
slightly weaker than in the
imum in the plots of RDG versus sign(λ2 )·ρ
tures (ρ
shown in Fig. 7(b,d). In such plots, the strength
ing.
of non-covalent interactions is related to the
terminus of
sharpness of the minima and the values of
ρ
transC5C7+.
clearly shows that hydrogen bonds in are much weaker than those in
= 0.020)
C5g±(s)
is
struc-
that do not have donor shar-
Similarly, the hydrogen bond at the C-
than those (ρ
at the minima. For example, the NCI analysis
trans
= 0.022)
For
C5-trans (ρ = 0.015) is weaker in trans-trans and sc-cis-trans
= 0.017).
Finally, as can be seen from Table 3, the
cis-trans
sc-
conformer is associated with a weak
The results of the NCI analysis applied to dif-
interaction at the N-terminus, H2 N· · · HN. It
ferent Gly-Gly conformers are summarized in
is almost twice weaker than the corresponding
Table 3 (see also Fig. S8 of Supp.
C5 interaction in the structurally similar
Info).
For
C5-
each interaction found at the N- and C-termini,
trans
Table 3 lists the RDG values in dierent minima
spectroscopy. If we neglect the weak H2 N· · · HN
and the corresponding sign(λ2 )·ρ values. These
interaction in
two numbers provide information on the type
lization due to the C5 interaction at the N-
and strength of the non-covalent interaction.
terminus of
35
Based on the data from Table 3, several conclu-
conformer, observed by IRMPD-VUV
sc-cis-trans ,
C5-trans
the energy stabi-
can be estimated from
the the energy dierence between the two con−1 formers that is equal to 750 cm (9 kJ/mol).
sions can be drawn. Firstly, the planar conformers ordered by increasing strength of their N-terminus hydro-
4
transtrans < sc-cis-trans < g-trans-trans < C5trans (Table 3). Interestingly, the interaction of relatively stable trans-trans conformer gen bonding attractive interaction are:
is the weakest (see also Fig. 7(c,d)).
IRMPD-VUV spectroscopy of a cold molecular beam of laser-desorbed Gly-Gly revealed that the most dominant conformer of this dipep-
It is
tide is
even weaker than CH· · · O=C interactions in the
C5g±(s)
structure.
the contribution of the
On
trans-trans
trum.
with reduced entropy contributions
drance contributes to the relatively high stabilhydrogen
bonding
trans .
interac-
com-
C5-
It is worth noting that Gibbs free en-
ergy calculations (Fig. 6(a)) clearly capture the
tions at the C-terminus can have either weak CH· · · O=C character (C5g±(s) ),
28,29
pared to more exible structures such as
trans-trans . the
The explanation can be found in the
rigidity of this structure which is associated
Fig. 7(a,c)). Therefore, the reduced steric hin-
Secondly,
conformer, sta-
found to be negligible in the measured spec-
are less pro-
nounced than in the case of cis structures (see
ity of
C5C7
Interestingly,
bilized by a strong C7 hydrogen bond, was
the other hand, the steric repulsion eects at the N-terminus of
with the nearly planar cis-
hydrogen bonding interactions.
does not contribute signicantly to the sta-
trans-trans
C5-trans ,
trans backbone structure stabilized by two C5
structures and most probably it
bilization of the
Conclusions
stabilization eect due to entropy, predicting
weak C5
ACS Paragon Plus Environment 11
The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 12 of 20
Table 3: The minimum values of RDG and the corresponding sign(λ2 )·ρ values associated with dierent attractive interactions at N- and C-termini of Gly-Gly structures. The values denoted with ∗ correspond to the weak CH· · · O=C interaction. Note that the interactions at the N- and C-termini of the g-trans-trans conformer have nearly identical strength, which hindered nding the individual RDG and sign(λ2 )·ρ values. Conformer
C5-trans C5C7+ C5g-(s) C5g+(s) trans-trans g-trans-trans sc-cis-trans
sign(λ2 )·ρ
RDG(min) N-terminus
C-terminus
N-terminus
C-terminus
0
0.145
-0.0202
-0.0151
0
0
-0.0219
-0.038
0
0.137∗
-0.022
-0.0148∗
0
0.153∗
-0.0216
-0.0147∗
0.312
0.115
-0.012
-0.0165
0.081
0.081
-0.0165
-0.0165
0.08
0.089
-0.0139
-0.0169
C5-trans structure is more favorable than C5C7 at non-zero temperatures. that
molecule.
14
The current study shows that the
structure of the most stable conformers of pep-
Conformational relaxation in the supersonic
tides without aromatic moiety can successfully
jet is another eect that was found to play
be studied with IRMDP-VUV spectroscopy in
an important role in the observation of the
combination with the jet-cooling method.
C5-trans
we observed, jet-cooling is particularly ecient
conformer as the most dominant.
As
The accurate determination of relaxation barri-
for conformational relaxation.
ers between dierent conformers indicated that
ables reducing the number of conformers pop-
C5-trans
The latter en-
can gain population from a number
ulated, signicantly simplifying the measured
of low-energy conformers, making it the most
spectra. This is highly advantageous for spec-
abundant in a wide range of temperatures. It is
troscopy of exible peptides with a large degree
worth noting that owing to extremely low bar−1 riers to interconversion (