Plant sunscreens in the UV-B: Ultraviolet spectroscopy of jet-cooled sinapoyl malate, sinapic acid, and sinapate ester derivatives

Article

pubs.acs.org/JACS

Plant Sunscreens in the UV-B: Ultraviolet Spectroscopy of Jet-Cooled

Sinapoyl Malate, Sinapic Acid, and Sinapate Ester Derivatives

Jacob C. Dean,^†Ryoji Kusaka,^†Patrick S. Walsh,^†Florent Allais,^‡,§,∥and Timothy S. Zwier*^,†

^†Department of Chemistry, Purdue University, West Lafayette, Indiana 47907-2084, United States

^‡AgroParisTech, Chaire Agro-Biotechnologies Industrielles (ABI), 247 rue Paul Vaillant-Couturier, F-51100 Reims, France

^§AgroParisTech, UMR 782 GMPA, Site de Grignon, F-78850 Thiverval-Grignon, France

^∥INRA, UMR 782 GMPA, Site de Grignon, F-78850 Thiverval-Grignon, France

S

* Supporting Information

ABSTRACT: Ultraviolet spectroscopy of sinapoyl malate, an

essential UV-B screening agent in plants, was carried out in the

cold, isolated environment of a supersonic expansion to

explore its intrinsic UV spectral properties in detail. Despite

these conditions, sinapoyl malate displays anomalous spectral

broadening extending well over 1000 cm⁻¹in the UV-B region,

presenting the tantalizing prospect that nature’s selection of

UV-B sunscreen is based in part on the inherent quantum

mechanical features of its excited states. Jet-cooling provides an

ideal setting in which to explore this topic, where

complications from intermolecular interactions are eliminated.

In order to better understand the structural causes of this

behavior, the UV spectroscopy of a series of sinapate esters

was undertaken and compared with ab initio calculations, starting with the simplest sinapate chromophore sinapic acid, and

building up the ester side chain to sinapoyl malate. This “deconstruction” approach provided insight into the active mechanism

intrinsic to sinapoyl malate, which is tentatively attributed to mixing of the bright V (¹ππ*) state with an adiabatically lower ¹nπ*

state which, according to calculations, shows unique charge-transfer characteristics brought on by the electron-rich malate side

chain. All members of the series absorb strongly in the UV-B region, but signiﬁcant diﬀerences emerge in the appearance of the

spectrum among the series, with derivatives most closely associated with sinapoyl malate showing characteristic broadening even

under jet-cooled conditions. The long vibronic progressions, conformational distribution, and large oscillator strength of the V

(ππ*) transition in sinapates makes them ideal candidates for their role as UV-B screening agents in plants.

concentration to prevent penetration of UV-B radiation into

the underlying mesophyll layer where photosynthesis occurs.^1,9

Using genetic mutations in the Arabidopsis thaliana plant, it was

found that the phenylpropanoid pathway could be manipulated

to reduce levels of sinapate esters, resulting in a corresponding

increase in UV-B sensitivity.^1−3,9−12Such experiments provide

unequivocal evidence to implicate sinapate esters (or sinapates)

as the primary class of molecules screening UV-B in

Brassicaceae plants, using such observables as plant growth/

injury,³ﬂuorescence analysis of sinapates and chlorophyll,^10,12

and photosynthetic activity.²Sinapate esters are ester

derivatives of sinapic acid, and sinapoyl malate is the dominant

species concentrated in the leaf epidermis of adult Arabidopsis

plants.^9,10,13The chemical structures of these molecules are

shown in Figure 1a.

1. INTRODUCTION

UV-B radiation comprises the 280−315 nm region of the solar

spectrum, and despite its comparatively small contribution to

the spectrum, its deleterious eﬀects on higher plant life and

function are substantial. Defensive measures to protect

susceptible plant cells/tissues are therefore necessary, given

that plants sustain constant exposure to such radiation. As such,

the impact of UV-B light on plant growth and sustainability has

been the subject of recent research.¹⁻⁵Among the negative

consequences of UV-B exposure is oxidative damage following

free radical formation that leads to the destruction of DNA and

tissues, which ultimately aﬀects the integrity of the plant

structure and inhibits plant growth.^1−3,6Additionally, dis-

ruptions in the eﬃciency of photosynthesis, ﬂowering, and

transpiration have been documented.^7,8As a means of

alleviating these harmful eﬀects, many plants utilize screening

agents, such as phenolics, generated by natural biosynthetic

pathways and ﬁne-tuned to absorb those destructive photons.

These molecules are transported into the upper epidermal layer

of leaves or seedlings, where they accumulate in relatively high

In order for sinapoyl malate/sinapate esters to act as eﬀective

UV-B protectants, they not only must be present in suﬃcient

Received: June 17, 2014

Published: October 8, 2014

14780

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

Figure 1. Chemical structures for the full series of molecules studied including (a) sinapic acid (SA) and sinapoyl malate (SM) and (b) sinapate

derivatives that bridge between these two in chemical complexity, including methyl sinapate (MS), isopropyl sinapate (IS), sinapoyl methyl-lactate

(SML), sinapoyl methyl-butyrate (SMB), and sinapoyl dimethyl malate (SDM). All samples were studied in stereochemically pure (S) form.

abundance and uniformly distributed in the upper epidermis of

plant leaves but must also be able to absorb across much of the

UV-B and with high extinction coeﬃcients.^10,12While the

natural occurrence of biosynthetic pathways that produce these

molecules and their basic UV absorption properties mark them

as candidate UV-B sunscreens, a detailed investigation of the

intrinsic UV absorption properties and excited states of these

compounds has yet to be undertaken. By introducing these

molecules into the cold, isolated environment of a supersonic

expansion, their fundamental absorption properties can be

explored in detail free from solvent eﬀects or other environ-

mental perturbations. By employing this technique, the intrinsic

UV spectroscopy of sinapoyl malate and various smaller

sinapate derivatives are assessed in detail here.

Another class of molecules that is chemically related to

sinapates, and therefore represents a relevant set of analogues,

are the cinnamates. These diﬀer from sinapates only by the lack

of methoxy substituents on the aromatic ring meta to the vinyl

group, and at times a change in substitution of the para hydroxy

group depending on the derivatization. The UV spectroscopy

of cinnamate derivatives and their photophysical characteristics

have been characterized extensively in solution,¹⁴⁻¹⁹and the

derivative octyl methoxycinnamate is a chief ingredient in

synthetic/cosmetic UV-B sunscreens. Excited state decay

1

mechanisms from the excited ππ* state in the cinnamate

class of molecules include internal conversion to the ground

state, trans−cis photoisomerization followed by internal

conversion, and internal conversion to a lower ¹nπ*

state.¹⁴⁻¹⁷Gas phase jet studies of the sunscreen octyl

methoxycinnamate (or 2-ethylhexyl-4-methoxycinnamate,

EHMC) and the simpler derivative methyl-4-methoxycinna-

mate (MMC) were recently undertaken by Buma and co-

workers, and their results revealed picosecond ¹ππ* state

lifetimes (based on resonant two-photon ionization line

widths) yet nanosecond decay proﬁles for the two-photon

ionization yield.²⁰These results were interpreted as signaling

1

fast internal conversion to an adiabatically lower nπ* state

(picoseconds), which then decayed on a nanosecond time scale.

1

Such fast internal conversion to an adiabatically lower nπ*

state was previously characterized in the similar hydroxy-

substituted cinnamates, p-coumaric acid (pCA) and its methyl

ester (4-hydroxycinnamate, MeOpCA), by similar experimental

14781

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

R2PI measurements, two laser pulses were spatially and temporally

overlapped, and the frequency of the excitation photon was scanned

while the ionization photon remained ﬁxed at a frequency lower than

the UV absorption. For 1C-R2PI, UV pulse energies of ∼100 μJ/pulse

were used, while 2C-R2PI was performed with λ_excitation∼ 10 μJ/pulse

and λ_ionization∼ 100 μJ/pulse. Pulse energies were optimized to ensure

that saturation eﬀects were minimized. UV−UV holeburning (UV−

UV HB) spectroscopy was carried out to obtain conformation-speciﬁc

spectra of the conformers present in the expansion. Conformation-

speciﬁc IR spectroscopy was executed by resonant ion-dip infrared

spectroscopy (RIDIRS) on SA, SM, and SDM in the OH stretch, alkyl

CH stretch, and mid-IR regions. Experimental details of the

implementation of these techniques can be found in the Supporting

Information. Dispersed ﬂuorescence (DFL) spectra were recorded for

the electronic origin transitions of MS and SML by collecting the

ﬂuorescence emission, focusing it into the entrance slit (slit width of

50 or 100 μm) of a Horiba Jobin Yvon 750 monochromator, which

disperses it onto an intensiﬁed charge coupled device (iCCD, Andor

iStar).

2.2. Computational Methods. To classify the possible conforma-

tional minima for each molecule, a conformational search was carried

out using the Amber* or molecular mechanics force ﬁeld (MMFF) in

the MacroModel suite of programs.^28,29The ﬁrst 50−100 conforma-

tional minima (or all possible minima if less than 100) were further

optimized by density functional theory (DFT) using the M05-2X

hybrid functional³⁰with the 6-31+G(d) basis set. Ground state

vibrational frequencies and vertical excitation energies were calculated

for all optimized structures using time-dependent DFT (TDDFT) at

the same level of theory. S₁geometries and vibrational frequencies

were also calculated for the assigned conformations. All DFT

calculations were performed in the Gaussian 09 suite.³¹For

comparison with experimental IR spectra, the calculated harmonic

vibrational frequencies were scaled by 0.949, 0.931, and 0.960 for OH

stretch, alkyl CH stretch, and mid-IR vibrations as determined

previously for this level of theory.^27,32The alkyl CH stretch

frequencies were scaled to match the equatorial methoxy CH stretch

transition in guaiacol near 3015 cm⁻¹.

methods, and further conﬁrmed by quantum chemical

calculations.²¹⁻²³Finally, picosecond pump−probe measure-

ments and symmetry-adapted cluster-conﬁguration interaction

(SAC-CI) calculations of MeOpCA performed by Shimada et

al. indicate the possibility of excited state trans−cis isomer-

ization about the propenyl double bond as an additional/

alternative pathway giving rise to the broadening observed in

the UV spectrum of MeOpCA.²⁴

In this work, the UV spectroscopy of a series of sinapate

esters is thoroughly investigated. The nature of the strong ¹ππ*

electronic transition responsible for UV-B screening is analyzed

via the vibrationally resolved UV spectra that are aﬀorded by

cooling the molecules in a supersonic expansion. The series

presented here includes the simplest sinapate chromophore,

sinapic acid (SA), and the biologically relevant sinapoyl malate

(SM) (neutral diacid). In addition, the UV spectroscopy of a

sequence of increasingly complex ester derivatives was carried

out to test the sensitivity of the spectroscopy to various aspects

of the structure of sinapoyl malate. These include methyl

sinapate (MS), isopropyl sinapate (IS), sinapoyl methyl-lactate

(SML), sinapoyl methyl-butyrate (SMB), and the dimethyl

ester of sinapoyl malate (sinapoyl dimethyl malate, SDM)

which are shown in Figure 1b. When possible, single-

conformation UV spectra were attained for those conformers

populated in the expansion, and in all cases, two major

conformers were observed in nearly equal abundance.

Conformation-speciﬁc IR spectroscopy was performed on SA,

SM, and SDM in the OH stretch (3500−3700 cm⁻¹), alkyl CH

stretch (2800−3050 cm⁻¹), and “mid-IR” (1100−1900 cm⁻¹)

regions, to conﬁrm the structures of those conformational

isomers. Of the series of molecules presented, sinapoyl malate

(and its dimethyl derivative) yielded a distinctly broadened

spectrum, suggesting the possibility of excited state processes at

play. Several mechanisms which could give rise to such

anomalous broadening in the gas phase are discussed, leading

to a tentative assignment for the broadening to excited state

mixing of the bright ππ* state with a nearby nπ* state with

substantial charge-transfer character, which is uniquely present

and active in the UV-B region in sinapoyl malate.

3. RESULTS AND ANALYSIS

The lowest singlet excitation calculated for the sinapate

chromophore is a ππ* transition characterized predominantly

as a HOMO → LUMO transition. The energy of this transition

falls near the red edge of the UV-B. An overview of the R2PI

spectra in this region for the seven molecules in this series is

shown in Figure 2. The spectra extend some 1200 cm⁻¹in total,

with the origin transitions spanning a range of 350 cm⁻¹

depending on the side chain substitution. All of the spectra

display considerable Franck−Condon activity, giving rise to

fairly congested UV spectra. As the ester side chain becomes

increasingly complex (going down the series in Figure 2), the

proﬁle of the UV spectrum begins to change with increasing

levels of congestion, accompanied by a shift of the intensity

distribution toward the blue (from the respective origin

transitions). Interestingly, all of the simpler derivatives display

sharp, resolved spectra, whereas sinapoyl malate and its methyl

ester derivative (SM, SDM) both exhibit broad, featureless

spectra that extend well over 1000 cm⁻¹. This observation is of

direct interest to nature’s choice of sinapoyl malate as a UV

sunscreen for plants, and could signal the presence of excited

state dynamics in sinapoyl malate that is diﬀerent from those of

the simpler molecules. The derivatives in the middle of the

ﬁgure (SML and SMB), substituted with separate halves of the

malate side chain, represent an intermediate between the sharp,

well-resolved spectra (sinapic acid, methyl sinapate, and

isopropyl sinapate) and the broad sinapoyl malate spectra. In

these spectra, sharp transitions are clearly observable, but a shift

2. EXPERIMENTAL SECTION

2.1. Experimental Methods. Sinapic acid was purchased from

Sigma-Aldrich, while all other samples were synthesized. Synthesis of

sinapoyl malate (2-O-sinapoyl-^L-malate) was performed from the

procedure reported in Quentin et al. and Allais et al.^25,26Detailed

synthetic procedures for MS, IS, SML, SMB, and SDM are given in the

Supporting Information. All of the chiral samples (SML, SMB, SM,

and SDM) were synthesized in stereochemically pure form with S-

chirality, as is the case for naturally abundant sinapoyl malate. SA,

SMB, SM, and SDM were all introduced into the gas phase by laser

desorption from a thin ﬁlm of the solid on a graphite surface,

necessitated due to extensive decomposition upon heating those

samples. A detailed description of the laser desorption source can be

found elsewhere,²⁷but a brief description is given in the Supporting

Information. MS, IS, and SML were heated in a sample oven behind a

Parker Series 9 general valve, to temperatures of 60, 160, and 190 °C,

respectively, to generate suﬃcient vapor pressure for the experiment.

In this case, neon or helium was used as the buﬀer gas at a backing

pressure of 2−3 bar, instead of argon as used for laser desorption. The

vaporized sample was entrained in the buﬀer gas, forming a supersonic

expansion, which was then skimmed upon entry into the extraction

region of a time-of-ﬂight mass spectrometer to form a molecular beam.

UV excitation spectra were obtained via resonant two-photon

ionization (R2PI) in either a one-color (SA, MS, IS, SM, SDM; 1C-

R2PI) or two-color (SML, SMB, SDM; 2C-R2PI) scheme. For 2C-

14782

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

states. The HOMO and LUMO molecular orbitals (MOs) are

shown in Figure 4a, and they reveal a change in bonding

character across the C₄C₇and C₇C₈bonds, with increased

double bond character in the former at the expense of double-

bond character in the latter. This leads to a quinoidal geometry

change evidenced by the large increase in the C₇C₈bond

length from 1.34 to 1.40 Å and a corresponding decrease in the

C₄C₇bond length, as shown in Figure 4b. An analogous

geometry change was found in the simpler analogue styrene.³⁵

Only three conformational degrees of freedom (Figure 4c)

are needed to describe the conformational preferences of

sinapic acid. The relative orientations of C₇C₈and C₉O

groups have the most impact on the energies of the minima,

with the cis (C₇C₈C₉O ∼ 0°) family falling lower in energy

by approximately 3 kJ/mol than the trans family. Within these

families, syn/anti combinations arise from separate minima

associated with the syringyl OH/C₇C₈orientation. Interest-

ingly, the calculations predict minima for completely planar, C_s

structures, and structures in which the non-H-bonded methoxy

CH₃group is out-of-plane by ∼70° (at a cost of ∼2.3 kJ/mol).

However, there is no experimental evidence for population in

the nonplanar methoxy structures in any of the sinapate esters,

suggesting that collisional cooling to the planar minima is

complete. The lowest energy conformers are therefore the

planar syn/cis (0.28 kJ/mol) and anti/cis (0.00 kJ/mol)

structures diﬀering only by a 180° internal rotation of the

OH group. In what follows, we assign conformers A and B as

syn/cis and anti/cis, respectively. These structures are shown in

Figure 4c.

Figure 2. Summary of the R2PI spectra of the sinapate series. The

spectra are plotted with increasing side chain derivatization from top

to bottom.

in the intensities and the partially resolved nature of the SMB

spectrum indicate properties intermediate to those of sinapoyl

malate. Detailed analyses of these spectra are given in the

following sections.

3.1.2. RIDIR Spectroscopy. To evaluate the structures of

conformers A and B in sinapic acid, single-conformation IR

spectra were recorded in the OH stretch, CH stretch, and mid-

IR regions using RIDIR spectroscopy (Figure 5). Two OH

stretch fundamentals are expected due to the H-bonded phenol

(or syringyl) OH and free carboxylic acid, which are predicted

to be within 10 cm⁻¹of one another. The presence of a single

band in the experimental spectrum (Figure 5, right) suggests

that these frequencies are accidently degenerate.

3.1. Sinapic Acid (SA) Spectroscopy. The simplest of the

sinapate series and the biological precursor to sinapoyl malate is

sinapic acid. We consider in some detail the UV spectroscopy

and conformational assignments of SA as a foundation for the

comparisons with other sinapate esters in what follows.

3.1.1. Conformation-Speciﬁc UV Spectroscopy. The R2PI

spectrum and UV−UV HB spectra for sinapic acid are shown in

Figure 3a. Only two conformers were present in the expansion,

as proven by UV−UV HB, with very similar appearance and

Franck−Condon activity. Their electronic origins (30756 cm⁻¹

for A and 30891 cm⁻¹for B) are separated by 135 cm⁻¹. The

calculated low-energy minima are planar in the ground state,

classifying it in the C_ssymmetry group. As can be seen from the

transitions labeled in the spectra in Figure 3a, essentially all the

Franck−Condon activity can be ascribed to progressions and

combination bands involving four vibrational modes: the in-

plane C₄C₇C₈bending mode, “b”, the C₇C₈C₉(O)

in-plane scissor mode (S) at 193 cm⁻¹, and even overtones and

combinations of the out-of-plane butterﬂy (β) and C₄C₇

torsion/symmetric methoxy torsion (τ). The form of these

modes is shown in Figure 3b. The presence of only even

overtones in the out-of-plane modes proves that the ππ*

The CH stretch spectra are very similar to the syringol

spectrum in this region,³⁶showing three sets of bands

associated with the three OCH₃CH vibrations (CH^eq, CH₂

ax

antisymmetric stretch [AS], and CH₃^axsymmetric stretch [SS]),

which are shifted and split due to Fermi resonances with the

CH bend overtones. Slight diﬀerences are observed in the AS

and SS OCH₃bands between the two conformers. In

conformer A, four bands are clearly observed in the AS

OCH₃region, whereas in B the splittings are less resolved. Of

greatest relevance for the sinapate esters studied here, the

lowest-frequency SS OCH₃bands at ∼2850 cm⁻¹show

diﬀerent intensities for the H-bonded and non-H-bonded

OCH₃SS bands between conformers, with A having a stronger

non-H-bonded band (2845 cm⁻¹) and B a stronger H-bonded

band (2853 cm⁻¹). In what follows, we exploit this diﬀerence in

assessing the presence of syn/anti conformational pairs. The

mid-IR region is quite complex, especially below 1400 cm⁻¹

where CH bend, OH bend, and CO stretch fundamentals

arise. At higher wavenumber, the CO stretch and CC

stretch bands are readily assigned.

excited state is also planar. The close proximity of the β²and

0

τ²overtones with the b¹fundamental suggests the possibility

0

that these levels may be in Fermi resonance with one another.

In addition, calculations predict large changes in frequency and

form of the low-frequency out-of-plane vibrations of sinapic

acid, which could lead to Duschinsky mixing.^33,34The

experimental and calculated vibrational frequencies for con-

formers A and B are given in the Supporting Information

(Table S1).

The long Franck−Condon progression in the C₄C₇C₈

bending mode can be rationalized on the basis of the change in

the electronic conﬁguration between the ground and excited

The calculated IR spectra for the structures assigned to A and

B are shown below the experimental spectra in Figure 5.

Between conformers A and B, the simulated spectra are very

similar, but the overall ﬁt to experiment is relatively good. The

calculation predicts a 6 cm⁻¹splitting between the free

14783

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

Figure 3. (a) R2PI spectrum (black) and UV−UV HB spectra of conformers A (red) and B (blue) for SA. UV−UV HB was taken by ﬁxing the

holeburn laser on the origin transitions of conformers A and B. (b) Franck−Condon active vibrations in sinapic acid: b, C₄C₇C₈bend; S, C₇

C₈C₉(O) scissor; β, butterﬂy; τ, C₄C₇torsion/symmetric methoxy torsion. Refer to Figure 1a for atom numbering.

carboxylic acid OH and syringyl OH fundamentals of the planar

structures, but these bands are not resolved in the spectrum.

The calculated spectra in the 1500−1800 cm⁻¹region match

the experimental spectra well, validating the assignment of the

bands at 1766, 1656, and 1627 cm⁻¹to CO stretch, C

C_vinylstretch, and CC_arstretch fundamentals, respectively.

While only slight diﬀerences are observed between the two

conformers in the IR, further veriﬁcation of the appropriate

syn/anti assignments can be found in the electronic origin

splitting. Vertical excitation energy calculations at the

TDDFT//M05-2X/6-31+G(d) level of theory predict the syn

conformer red of the anti conformer, and when scaled to the

origin transition of conformer A (scale factor = 0.9038), the

calculated splitting is 168 cm⁻¹. As a result, we tentatively

assign conformer A to the syn/cis conformer and B to anti/cis.

The experimental and calculated origin frequencies for

conformers A and B can be found in Table 1. Interconversion

between syn and anti minima of the cis (planar) conformational

family (assigned to conformers A and B) can occur by hindered

rotation about the C₃C₄C₇C₈dihedral. To estimate the barrier

between these two conformers along this coordinate, a relaxed

potential energy scan was carried out and is shown in the

Supporting Information; the barrier height is predicted to be 24

kJ/mol (2006 cm⁻¹), and the second pathway to interconver-

sion by rotation about the C−O(H) bond yields a slightly

higher barrier at 26.8 kJ/mol (2240 cm⁻¹).

3.2. Methyl Sinapate (MS) and Isopropyl Sinapate (IS)

Spectroscopy. 3.2.1. R2PI and UV−UV HB Spectroscopy.

The simplest extension from sinapic acid to its ester derivatives

is to replace the carboxylic acid hydrogen with a methyl group,

forming the methyl ester, methyl sinapate (MS). The R2PI and

UV−UV HB spectra for MS are shown in Figure 6a. This

substitution shifts the electronic origin to the blue by 309 cm⁻¹,

a surprisingly large amount given how remote this substitution

is from the aromatic ring. However, since the electronic

conjugation extends to the end of the molecule, methyl

substitution does play a role in inducing an electronic frequency

shift, and is consistent with the large geometry change in the

vinyl-containing substituent. Three conformers were found in

methyl sinapate, two of which dominate the population, and the

14784

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

Table 1. Assignments, Zero-Point Corrected Energies, and

Experimental and Calculated Origin Frequencies for the

Observed Conformers in the Sinapate Series

a

E

origin^exp

(cm⁻¹

origin^calc

ab

,

conformer assignment

(kJ/mol)

)

(cm⁻¹

)

SA (A)

SA (B)

MS (A)

MS (B)

MS (C)

IS (A)

IS (B)

SML (A)

SML (B)

SMB (A)

SMB (B)

SM

syn/cis

anti/cis

syn/cis

anti/cis

anti/trans

syn/cis

anti/cis

syn/cis

anti/cis

syn/cis

anti/cis

syn/cis

anti/cis

syn/cis

anti/cis

0.28

0

30756

30891

31065

31174

31285

31046

31157

30713

30847

30737

30924

30755

30918

31023

31175

31341

31051

31193

30225

30364

30016

30203

29646

29845

29813

30005

0.3

0

4.26

0.09

0

0.35

0

0.46

0

SM

0.61

0

SDM

0.58

a

b

DFT//M05-2X/6-31+G(d). Scaled by 0.9038.

extending up to seven quanta in conformer A. This band is

assigned to the same bending mode (bⁿ), and the frequency is

0

decreased from sinapic acid due to the heavier methyl group at

the terminus. The splitting between the origins of conformers A

and B is 109 cm⁻¹, an amount similar to that in SA. On the

basis of the assignments for SA, we tentatively assign the red-

shifted A conformer of methyl sinapate to the syn conformer

and the blue-shifted B conformer to its anti counterpart. The

two are calculated to be within 0.30 kJ/mol of one another,

with anti/cis below syn/cis. Interestingly, the syn/cis conformer

displays a longer Franck−Condon progression in the vinyl

bend b, the opposite of that found in sinapic acid.

Figure 4. (a) Molecular orbitals for the S₀→ S₁transition in SA and

(b) optimized bond distances (Å) for S₀and S₁structures calculated at

the TDDFT//M05-2X/6-31+G(d) level of theory. (c) Conforma-

tional degrees of freedom in SA (left) and assigned structures for

conformers A and B in SA.

The other low-frequency bands at 72 (73) and 76 (79) cm⁻¹

in conformer A (B) are assigned to β²and τ²transitions,

0

showing that these conformers still obey C_sselection rules,

proving that both ground and S₁excited states of both

molecules are planar. b¹β²and b¹τ²combination bands

0

third a very minor conformer just barely observed. Once again,

for the two major conformers, long progressions are found with

frequencies of 59 and 57 cm⁻¹for A and B, respectively,

appear just blue of b², and the S¹fundamental emerges at

0 0

+161 cm⁻¹. The electronic origin of conformer C is blue-shifted

from both major conformers, appearing at 31285 cm⁻¹.

Figure 5. RIDIR spectra for conformers A (top, red) and B (bottom, blue) of SA in the OH stretch (right), alkyl CH stretch (middle), and mid-IR

(left) regions with calculated stick spectra below each experimental spectrum, calculated at the DFT//M05-2X/6-31+G(d) level of theory.

14785

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

according to calculation, the isopropyl group asymmetrically

staggers the chromophore plane, placing one methyl group

+80° and the other −22° from the plane. It should be noted

that, among the calculated structures, degenerate pairs were

found for structures that oriented the isopropyl group in the

identical conﬁguration except on opposite sides of the plane,

a

b

a

b

i.e., CH₃(80°)/CH₃(−22°) and CH₃(22°)/CH₃(−80°).

These structures are identical and indistinguishable in the UV

spectrum.

3.2.2. Fluorescence Spectroscopy of Methyl Sinapate. LIF

and DFL spectroscopy was carried out on MS for measurement

of the ﬂuorescence lifetime and a qualitative check on the

ﬂuorescence quantum yield. The ﬂuorescence lifetimes of the

electronic origins of conformers A and B of MS were

approximately 12 and 14 ns, respectively. We surmise that no

fast nonradiative processes are at play in the simple case of MS,

based on its relatively long ﬂuorescence decay and large

ﬂuorescence signal.

The DFL spectra of the origin transitions of conformers A

and B are shown in parts a and b of Figure 7, respectively. As in

Figure 6. R2PI spectrum (black) and UV−UV HB spectra of (a)

conformers A (red), B (blue), and C (green) of MS and (b)

conformers A (red) and B (blue) of IS. Asterisks indicate transitions

used for UV−UV HB.

Calculated vertical excitation energies place the syn/cis

conformer furthest to the red among all calculated conformers,

with the anti/cis predicted at +152 cm⁻¹and the anti/trans at

+318 cm⁻¹. Similar splittings between cis and trans conformers

were observed experimentally for p-coumaric acid and its

methyl ester.^21,22On this basis, conformer C is tentatively

assigned to the anti/trans conformation, calculated to have an

energy of 4.26 kJ/mol above the global minimum. The

experimental and calculated origins can be found in Table 1.

As a next step in building up the full complexity of the

sinapoyl malate side chain, an isopropyl group was substituted

in place of the methyl group in MS, forming isopropyl sinapate.

Notably, the isopropyl group also breaks the planar symmetry,

yet without introducing additional functional groups to the

molecule. The R2PI and UV−UV HB spectra for isopropyl

sinapate (IS) reveal spectra that are resolved but quite

congested, as shown in Figure 6b. As with SA and MS, just

two major conformers were observed, with S₀−S₁origin

transitions split by 111 cm⁻¹, only 2 cm⁻¹larger than in methyl

sinapate; therefore, the two nearly equally abundant conformers

are assigned as the syn/anti pair of the cis CC/CO family,

as in SA and MS. Once again, it is possible to assign the low-

frequency vibronic structure, with excellent agreement between

experiment and calculated frequencies for the excited state

vibrations (Table S1, Supporting Information). In breaking the

mirror symmetry through the chromophore plane with the

isopropyl group, a″ fundamentals are allowed and observed just

above the origin transition, as labeled in Figure 6b. Indeed,

Figure 7. DFL spectra of the S₁origin transitions of conformers A (a)

and B (b) of MS.

the UV excitation spectra, long Franck−Condon progressions

involving the C₄C₇C₈bend are observed, with the pure b⁰

n

progression extending out to ten quanta in conformer A and

potentially six in B. These bands dominate the spectra in Figure

7, hindering detection of the weak β and τ overtone transitions.

The ground state frequency of the bending mode for both

conformers is 59 cm⁻¹, eﬀectively unchanged from the excited

state. Franck−Condon analysis of the b⁰progression in the

n

DFL spectrum of conformers A and B (Figure 7) yielded

displacement parameters of approximately D = 2.0 and 1.3,

respectively.³⁷Transitions to high “n” levels in the ground state

have intensities that likely have unresolved contributions from a

second fundamental, distorting the intensities of these bands.

14786

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

0

The vinyl CH bend fundamental, b_{v.CH 1}, and the CC

0

stretch fundamental, S_{CC 1}, are also prominent in the spectra,

with b⁰_nprogressions built oﬀ of them, as highlighted in Figure

7. The strong Franck−Condon factors for these modes are in

accord with the proposed geometry change associated with the

S₀−S₁electronic transition, favoring those modes that

correspond to changes in the single and double bond character

of the CCCO substituent associated with the ππ*

transition.

3.3. Sinapoyl Methyl-Lactate (SML) and Sinapoyl

Methyl-Butyrate (SMB) Spectroscopy. The studies detailed

in the previous sections on sinapic acid, methyl sinapate, and

isopropyl sinapate have provided a sound basis for extension to

sinapoyl malate-like molecules. As shown in Figure 2, sinapoyl

malate itself possesses a structureless R2PI spectrum that

represents a fascinating, seemingly cataclysmic change in the

excited state spectroscopy and/or dynamics. Therefore, to

search for an onset of such behavior, sinapate ester derivatives

containing only one-half of the dicarboxylate malate side chain

were synthesized and studied. The malate side chain has its two

carboxylic acid groups separated from the ester oxygen by one

and two carbons, respectively. Given that both the dicarboxylic

acid and methyl dicarboxylate SM molecules yielded broadened

UV spectra (Figure 2), we chose to study the methyl

carboxylate derivatives of the separate malate “halves”. In

addition, a methyl group was used at the other end of the side

chain (C₁₁or C₁₂) to better mimic the conformational aspects

of malate by preventing artiﬁcial ﬂexibility of the single side

chain (see Figure 1), forming sinapoyl methyl-lactate (SML)

and sinapoyl methyl-butyrate (SMB), with structures shown in

Figure 1a.

Figure 8. (a) 2C-R2PI spectrum (black) and UV−UV HB spectra for

conformers A (red) and B (blue) of sinapoyl methyl-lactate (SML)

and (b) 2C-R2PI of sinapoyl methyl-butyrate (SMB). Asterisks

indicate UV−UV HB transitions. Bands to the red of the B origin in

part a are due to spectral overlap at the UV holeburn transition.

3.3.1. R2PI and UV−UV HB Spectroscopy. The 2C-R2PI

and UV−UV HB spectra of SML are shown in Figure 8a. The

R2PI spectrum still shows well-resolved vibronic structure, and

an appearance similar to that of its smaller counterparts, but the

spectrum also shows signs of additional congestion/broadening.

UV−UV HB spectroscopy once again identiﬁes the presence of

two conformational isomers. The S₀−S₁origin for conformer A

appears at 30713 cm⁻¹, now shifted back to the red closer to

those of SA and SM. Although the UV−UV HB spectrum of

conformer B clearly contains contributions from A due to

spectral overlap, the large discrepancy in intensities between A

and B transitions and the diagnostic b-progressions allows for

clear identiﬁcation of transitions due to conformer B. The

characteristic splitting between the electronic origins (134

cm⁻¹) and Franck−Condon activity attributed to these

conformers points clearly to their assignment as the syn/cis

(A) and anti/cis (B) conformers, respectively. Computational

predictions also verify this assignment, with the syn/cis

conformer calculated as the global minimum, and its anti

counterpart only +0.35 kJ/mol higher in energy.

The increased side chain ﬂexibility in SML produces

additional low-energy conformational minima associated with

side chain reorientation, predominantly deﬁned by the

C₉OC₁₀C₁₁dihedral, which positions the C₁₁O either cis or

trans to the ester oxygen between C₉and C₁₀. The low energy

structures nominally incorporate a trans C₁₁O conﬁguration,

while analogous structures that incorporate the cis conﬁguration

begin ∼7 kJ/mol higher in energy. Conformers A and B of

SML are tentatively assigned to the global minimum side chain

conformation, with structures shown in detail in the Supporting

Information.

One notable aspect of the UV absorption of SML compared

with the other compounds was its extreme sensitivity to photon

ﬂuence. To properly record an R2PI spectrum without marked

saturation eﬀects, 2C-R2PI was required with excitation laser

powers as low as 4−10 μJ/pulse (Φ ∼ 0.4 mJ/cm²), versus

typical excitation powers of ∼30−100 μJ/pulse for 1C-R2PI

used for the simpler sinapates. The implications of this fact will

be discussed further in the Discussion section.

Fluorescence spectroscopy was also performed on SML, and

it was found that the total ﬂuorescence signal compared to

methyl sinapate was only slightly less, but the lifetime was

noticeably shorter, with an instrument-limited response

providing an upper bound of 8 ns to its S₁lifetime. The DFL

spectra of the conformer A and B origin transitions were

acquired and can be found in the Supporting Information.

While the total ﬂuorescence signal for SML was ∼10% of that

of MS, transitions in the Franck−Condon region of the DFL

spectrum were <1% the size. This suggests that much of the

“missing” emission intensity is shifted well to the red, out of the

Franck−Condon window. Such Stokes-shifted emission is quite

common when coupling of the locally excited ππ* state to a

second excited state (e.g., a charge-transfer state) occurs.³⁸⁻⁴⁰

Dispersed emission was also collected over 50 nm for

conformer B, but no red-shifted emission was detected.

Because detection of broadened, red-shifted emission could

be diﬃcult in DFL spectra, a series of excitation spectra were

recorded with short wavelength cutoﬀ ﬁlters (345, 400, and 425

nm) in front of the PMT detector, to test whether a second

contribution to the absorption proﬁle might be buried

underneath the sharp transitions. A monotonic decrease in

14787

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

ﬂuorescence signal was observed for all transitions, with no

evidence of dual excitation proﬁles found.

The derivative incorporating the longer “half” of the malate

side chain is sinapoyl methyl-butyrate (SMB) shown in Figure

1b. This molecule has one more carbon along its side chain

introducing an additional element of ﬂexibility. The 2C-R2PI

spectrum for SMB taken under the same low-ﬂuence conditions

as SML is shown in Figure 8b. While sharp transitions are

clearly discernible, they are built oﬀ of a large background that

by appearances represents an intermediate case between

sinapoyl methyl-lactate and sinapoyl malate. UV−UV HB

spectra were not obtained for SMB due to low signal levels and

lack of band isolation. If the ﬁrst band is taken as the ﬁrst origin

transition (A), tentatively assigned to a syn/cis conformer, a

progression can be followed for at least two quanta, suggesting

a ν_b′ frequency of 59 cm⁻¹. While this is surprisingly larger in

frequency than in SML, a comparison of the calculated

frequencies between these molecules supports this assignment.

Other bands built oﬀ of this origin are found at +15 and +35

cm⁻¹similar to those low frequency modes in SML. Given the

relative intensities and splittings between the syn and anti

conformers of the molecules presented up to this point,

inspection of the SMB R2PI spectrum leads to a tentative

assignment of the anti/cis origin at 30924 cm⁻¹, 187 cm⁻¹blue

of the A origin. From this origin, there are prominent +37 and

+59 cm⁻¹bands presumed to be analogous features to those

found built oﬀ of conformer A.

Figure 9. R2PI spectra for sinapoyl malate, SM (top), and sinapoyl

dimethyl malate, SDM (bottom).

whether the broadening in the SM/SDM spectra was inherent

to the molecules and not strictly due to a high absorption

coeﬃcient inducing saturation eﬀects. No reproducible

structure was found.

3.4.2. Computational Predictions of Conformational

Families. The sinapoyl malate structure incorporates the full

malate side chain introducing a larger variety of conformational

families available to it. The global minimum syn/cis structure

calculated for sinapoyl malate is shown in Figure 10a. For the

It is possible that other conformers are present within the

spectrum, and calculations predict seven conformers within 5

kJ/mol, with four more structures in this low energy window

than in SML. We must rely on calculations for assignment of

the side chain conformation for the syn/cis and anti/cis

conformers assigned in the spectrum, and those conformational

details are given in the Supporting Information.

3.4. Sinapoyl Malate (SM) and Sinapoyl Dimethyl

Malate (SDM) Spectroscopy. 3.4.1. R2PI Spectroscopy. By

evaluating the vibronic spectroscopy of sinapic acid and its ester

derivatives, and determining the common conformational

distribution among this class of molecules, the spectroscopy

of the biologically relevant sinapoyl malate can be addressed.

The R2PI spectra for SM and its dimethyl carboxylate

derivative SDM are shown in Figure 9. As noted from Figure

2, both of these spectra are broadened, and seemingly

featureless. The UV absorption of these molecules begins at

∼30500 cm⁻¹, further red than any of the ester derivatives

analyzed up to this point. The spectral proﬁle is asymmetric,

with the SM spectrum peaking near 30700 cm⁻¹and extending

past 31200 cm⁻¹with a gradual decay to the blue. SDM was

investigated in order to ensure that the broadening observed in

the UV was not resulting from the acid COOH groups in SM

alone. The R2PI spectrum of SDM is nearly identical to SM,

and it incidentally provided a test case for laser ﬂuence studies

due to a higher signal-to-noise ratio (S/N) for its R2PI

spectrum and therefore a higher sensitivity to spectral features.

As a result, the spectrum for SDM was recorded over a wider

wavelength range (>1200 cm⁻¹), revealing an absorption

extending past 31700 cm⁻¹. While it appears that some

remnant band structure might be present in the sinapoyl malate

spectrum, none were reliably reproducible, and the higher S/N

spectrum of SDM yielded no sign of reproducible structure,

even near the onset of its spectrum where the density of S₁

states is lowest. 2C-R2PI was performed on SDM with the low-

ﬂuence conditions used for SML and SMB, to ﬁrmly determine

Figure 10. Calculated structure for the global minimum conformation

of (a) SM, with top and side views (right), and (b) the lowest energy

structures found for SDM.

lowest-energy syn/cis and anti/cis structures, the syringyl

methoxy groups are planar with the aromatic ring. The longer

alkyl chain (C₁₀C₁₂C₁₃) is extended in a nominally planar

conﬁguration (C₁₀C₁₂C₁₃OH plane), with the ﬁrst carboxylic

acid group (C₁₁OOH) oriented perpendicular to the aromatic

ring. The extended C₁₀C₁₂C₁₃OH part of the side chain is

approximately 20° below the plane of the sinapate itself (side

view of Figure 10a). Interestingly, the bulky size of the malate

group prevents the close approach of either of the carboxylic

14788

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

acid groups to the sinapate moiety, unlike the case of SML and

SMB. For the lowest two structures with this malate

conformation, the CO groups are still nominally pointing

back toward the aromatic ring (Figure 10a), as was the case for

the SML and SMB conformations (Supporting Information).

All of the oxygen-containing groups are well over 3 Å from one

another due to the steric constraints, prohibiting the formation

of any intramolecular H-bonds.

Within 10 kJ/mol of the global minimum, a total of 28

conformational minima were found, diﬀering by their malate

conformation, HO/CC conﬁguration (syn/anti), and degree

of planarity of the non-H-bonded methoxy group on the ring.

Approximately half of those 28 structures incorporated the

planar syringyl ring, with no experimental evidence for the

nonplanar syringyl conformation in any of the sinapate

esters.^41,42No CC/CO trans conformers were found

within this 10 kJ/mol window. The next set of side chain

structures incorporating the all-planar syringyl ring are found at

2.36, 5.82, and 6.11 kJ/mol above the global minimum, and

they diﬀer primarily by their malate dihedral angles summarized

in Table 2. A detailed description of these calculated structures

can be found in the Supporting Information.

syn/anti pair of conformers as were present in the simpler

sinapate esters, here employing the most stable malate

conformation in both, in nearly equal abundance. The

calculations are used to tentatively determine the speciﬁc

malate conformation, shown in Figure 10, relying on the fact

that the employed computational methods have accurately

predicted the same pair of conformations to be present in other

members of the series so far considered, providing evidence of

its predictive capability for SM and SDM.

4. DISCUSSION

As the key UV-B screening agent in many plants, sinapoyl

malate and related sinapate esters are vital to the survival of

plants. Nature’s selection of the sinapate esters for this purpose

is evident in their unique properties as UV-B absorbants. First,

the sinapate esters as a family all possess strong S₀−S₁UV

absorptions that peak in the UV-B region, with oscillator

strengths “f ” for their S₀−S₁transitions approaching 1. Thus,

the per-molecule absorption strength is as eﬃcient as possible,

as one might expect for a UV-B sunscreen. Second, of the

members of the series, sinapoyl malate is itself quite unique in

having an inherently broadened UV absorption that gives near-

complete wavelength coverage even under jet-cooled, isolated-

molecule conditions. The question is why. In what follows, the

characteristics of the sinapate S₀−S₁excitation and potential

mechanisms to account for the broadening in SM will be

explored. Having considered these mechanisms, we will return

to consider the potential implications that this broadening

might have for SM as a sunscreen for plants.

4.1. Nature of the Sinapate Electronic Excitation. The

characteristics of the π−π* transition responsible for the high

sinapate absorbance in the UV-B spectral region were brieﬂy

discussed in section 3.1 for sinapic acid. According to the

TDDFT calculations, the HOMO → LUMO transition is the

prominent contributor to the electronic transition in all of the

sinapates presented in this work, and the MOs shown for SA in

Figure 4a are representative for the excitation in the entire

series. The characteristic geometry change that accompanies

this excitation is reﬂected in the long Franck−Condon

progressions involving the C₄C₇C₈bend vibration.

Table 2. Relevant Dihedrals for Diﬀerent SM Side Chain

Conformations from DFT//M05-2X/6-31+G(d) Structural

Optimizations

E (kJ/mol) C₉OC₁₀C₁₁OC₁₀C₁₁O OC₁₀C₁₂C₁₃C₁₀C₁₂C₁₃O

0

−93.7

−82.2

50.6

95.1

−174

−137

177

65.5

−2.37

25

2.60

5.82

6.11

−161

−177

23.6

−59

145

−118

The broad, featureless UV spectra of the sinapoyl malate

molecules SM and SDM prevent the acquisition of con-

formation-speciﬁc IR data. However, composite IR spectra were

taken by monitoring the depletion of ion signal generated from

UV excitation at the peak of the broad UV absorption (∼30700

cm⁻¹in SM). The IR spectra of SM and SDM in the OH

stretch (SM), alkyl CH stretch, and mid-IR regions, along with

detailed spectral analysis, are given in the Supporting

Information. The OH stretch fundamentals of the CO−OH

groups are not resolved but appear as a single band in the free

COOH OH stretch region, as anticipated on the basis of the

low-energy structures, due to their inability to engage in

hydrogen bonds or interact with the sinapate aromatic ring.

The IR spectra of SM and SDM could not be used to make a

ﬁrm conformational assignment; however, some features of the

spectrum are worth noting. First, the sharp (∼4 cm⁻¹fwhm),

well-resolved IR transitions prove that the probed molecules

are indeed rotationally cold, so that the observed broadening is

not simply a thermal eﬀect. Second, using the IR spectra of

conformers A and B of sinapic acid (Figure 5) as a reference,

particularly in the CH stretch region, the relative intensities of

absorptions speciﬁc to syn and anti rotamers were nearly

equally represented in the SM composite spectrum. Further-

more, RIDIR spectra taken at multiple UV probe wavelengths

(across the SM absorption) were identical, indicating that the

conformers contributing to the spectrum contribute equally

across the UV absorption proﬁle.

p-Coumaric acid (pCA) and methyl 4-hydroxycinnamate

(MeOpCA) are close analogues of the prototypical members of

our sinapate ester series, sinapic acid and methyl sinapate,

lacking only the methoxy groups in the two ortho positions on

the ring. The electronic spectroscopy of pCA and MeOpCA has

been thoroughly investigated both experimentally²¹⁻²⁴and

computationally,⁴³and provides a basis for comparison with the

sinapate esters. In both pCA and MeOpCA, three relevant

singlet states were calculated, corresponding to the HOMO →

LUMO ππ* state discussed here (denoted “V” for “valence”), a

second ππ* of primarily HOMO → LUMO+1 character

(denoted “V′”), and a HOMO−3 n-type orbital → LUMO nπ*

state associated with the nonbonding orbital on the CO

oxygen.^21,22,43These states are diﬀerentiated by their MO

character but also the oscillator strengths for the S₀→ S_n

transition. Calculations for pCA and MeOpCA predicted a

large oscillator strength for transitions from the ground state to

the V state (∼0.65−0.7), and ∼0.02−0.18 for the V′

state.^21,22,43In pCA and MeOpCA, the S₁state was found to

be the V′ state, which possessed only weak Franck−Condon

activity in the b-mode, and dominating Δv = 0 Franck−

Condon factors. The same state ordering was presumed in the

On the basis of this data and our experience with other

members of the series, we surmise that the population in the

expansion consists either primarily or exclusively of the same

14789

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

cinnamates MMC and EHMC, where similar excited state

lifetimes and absorption frequencies were observed.²⁰

In sinapic acid and methyl sinapate, the V/V′ states show

similar characteristics. The Franck−Condon simulations for the

V and V′ excitations are shown for the syn/cis conformer of SA

in the Supporting Information as an example. The simulation

unambiguously veriﬁes the assignment of the S₁state in

sinapate chromophores as the V (ππ*) state, which has a very

large oscillator strength and long FC progressions in the b

mode.

The S₀−S₁absorption of pCA, MeOpCA, MMC, and

EHMC is centered around 33000 cm⁻¹, over 2000 cm⁻¹above

the sinapate absorption. Thus, the two methoxy groups present

in the sinapates shift conjugation out onto the methoxy groups,

adding to the electronic delocalization, and lowering the V state

relative to V′, thereby switching the order of the S₁/S₂states.

TDDFT calculations for pCA and MeOpCA falsely predict the

V state at lower energy than V′; however, this same ordering

was calculated for the sinapate molecules of interest here, and

in this case is in keeping with experiment. The TDDFT//M05-

2X/6-31+G(d) unscaled vertical excitation energies for the ﬁrst

three singlet states are summarized in Table S2 in the

Supporting Information for the assigned conformers in the

sinapate series. The same V (ππ*), V′ (ππ*), and nπ* states are

present in the sinapates, and the calculated V−V′ separation is

approximately equal to that calculated for pCA/MeOpCA with

TDDFT.²¹To illustrate the electronic transitions associated

with each of these three excited states, the calculated MOs

comprising the primary electronic transition for each of these

ground to excited state transitions are shown in Figure 11a, b,

and c for SM.

The bright S₀−V ππ* transition in sinapoyl malate (Figure

11a) promotes an electron from a π HOMO that possesses

C₄C₇C₈C₉O (single−double−single−double) bond

character to a π* LUMO with C₄C₇C₈C₉O bond

character. It should be noted that the electron density in the

LUMO has little contribution on the malate side chain. The

second ππ* state (V′) which was not observed in this

experiment is attributed mostly to the HOMO−1 → LUMO

transition shown in Figure 11b, which yields a smaller

absorption coeﬃcient, with electron density localized on the

aromatic ring in HOMO−1. The primary conﬁguration for the

nπ* state is of interest, as the anticipated localization to the

nominally assigned nonbonding orbital on the C₉O (as in

SA) is lost to some degree in sinapoyl malate, with electron

density extending all the way out into the ﬁrst carboxylic acid of

the malate moiety. This delocalization makes it a peculiar “non-

bonding” MO, diﬀering signiﬁcantly from simpler systems that

do not contain the malate group, such as sinapic acid.

Figure 11. Molecular orbitals for the three relevant singlet state

excitations in sinapoyl malate: (a) V (ππ*), (b) V′ (ππ*), and (c) nπ*.

(d) Electron density diﬀerence maps (EDDMs) of the three singlet

excited states of SM. Blue represents a gain in electron density, while

yellow represents a loss.

esters containing carboxylates/carboxylic acids (SML, SMB,

SM, and SDM) are calculated to be further to the red than the

experimental spectra would suggest they should be. This is

either indicative of a larger geometry change or of deﬁciencies

in the TDDFT calculations in these cases in which the COOX

groups on the side chain are near the chromophore. The scaled

excited state energies and relative oscillator strengths for the

assigned conformers are plotted along with the experimental

R2PI spectra in the Supporting Information (Figure S8),

further illustrating this fact.

As Table S2 (Supporting Information) shows, TDDFT

calculations predict no major changes to the vertical energy

separations between the S₁, S₂, and S₃states along the series of

molecules, with the V state always well below the others.

Furthermore, the direction and magnitude of the shift in the

vertical S₀−S₁transition along the series SA, MS, IS, SML, and

SMB are in qualitative agreement with the experimental

electronic frequency shifts (Table S2, Supporting Information,

c.f. Table 1). The calculations also faithfully reproduce the

experimental splitting between syn and anti OH/CC

orientations, which are not modiﬁed signiﬁcantly along the

series, with syn/cis below anti/cis by about 200 cm⁻¹.

The assignment of the S₀−S₁UV-B excitation to the bright V

(ππ*) state is unequivocal, and validated through experiment

and calculations. The oscillator strength of this transition in

sinapoyl malate is calculated to be ∼0.65, conﬁrming its status

as an eﬃcient absorber of UV-B radiation. Besides this

eﬃciency as an absorber, the absorption strength is spread

out over much of the UV-B region due to strong Franck−

Condon activity involving low-frequency in-plane and high-

frequency bond stretching coordinates associated with the

change in double-bond/single-bond character accompanying

excitation to the V state. It is for this reason that the V state in

pCA has been postulated to mediate trans−cis photoisomeriza-

tion about the C₇−C₈bond.^24,43In addition, the Franck−

Condon active low-frequency out-of-plane butterﬂy, β, and

At the same time, the calculated S₀−S₁vertical splittings

between the syn/cis and anti/cis conformers of the sinapate

14790

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

torsional, τ, modes serve to congest the S₀−S₁absorption

not expected that IVR should play a dominant role in the

excited state of sinapoyl malate near its origin, as it appears to

have minimal eﬀect on the closely related analogue SML, where

sharp structure is readily observed near the electronic origin.

A related possibility is that the Franck−Condon region in

sinapoyl malate begins well above the electronic origin where

the S₁density of states is very high, thereby preventing

resolution of individual transitions. While this idea is in line

with TDDFT calculations which predict a red-shift of larger

magnitude for the onset of the electronic transition in SM, the

electronic origins of the two conformers of SML are clearly

observed, and indicate that the TDDFT calculations over-

estimate the calculated electronic frequency shifts of the malate-

containing derivatives. Furthermore, structural optimization of

the S₁state of SM, at the TDDFT//M05-2X/6-31+G(d) level

of theory, shows no evidence of this larger geometry change

relative to S₀. Therefore, the calculations provide no reason to

suggest a larger change in geometry between S₀and S₁in SM

than in the SML or SMB analogues.

A ﬁnal nondynamical contributor to the broadening could be

the increasing number of populated conformations in sinapoyl

malate. Calculations indeed predict a larger number of low-

energy conformers for SM relative to its simpler derivatives but

not nearly the number necessary to produce a complete loss of

resolved transitions in the spectrum. Furthermore, the

composite IR spectra of SM and SDM suggest nearly equal

contributions from the same two syn/anti−cis conformers

found for the rest of the series. The IR spectra were also

analyzed at various UV excitation wavelengths in an attempt at

shifting the distribution of conformers contributing to the

RIDIR spectrum. However, the IR spectrum remained

unchanged, consistent with the presence of just two conformers

in the jet-cooled spectrum.

through combination bands with intense bⁿtransitions. The

0

activity in these modes is a characteristic feature of the sinapate

chromophore, proven by the extent of their activity in every

molecule studied here. The vibronic congestion brought about

by these active low-frequency modes is manifested in the gas

phase UV spectra, particularly with molecules incorporating a

large side chain, which drops the frequency of the β and τ

fundamentals to as low as 10 cm⁻¹. Coupled to this vibronic

activity is the “doubling” of transitions due to the population of

the syn/anti rotamers in nearly equal abundance. The small

frequency shift of the S₀−S₁absorptions of these conformers

(∼150 cm⁻¹) spreads out the inherent vibronic activity even

further and congests the total excitation spectrum.

4.2. Mechanism for Broadening in Sinapoyl Malate.

The S₀−S₁R2PI spectra of sinapoyl malate and its dimethyl

derivative, SDM, display a broad featureless absorption, which

sets it apart from the simpler derivatives studied in this work.

While this broadening prohibits vibronic analysis, its manifes-

tation points to aspects of the excited states of sinapoyl malate

that are unique (in this series), and may actively play a role in

the selection of sinapoyl malate as a UV-B protectant in plants.

The trends recognized from the UV spectroscopy of the

sinapate ester series studied here, along with the UV and IR

spectroscopy of SM/SDM itself, provide a basis on which to

evaluate the possible mechanisms giving rise to the inherent

broadening in SM.

4.2.1. Vibronic Congestion. One circumstance to be

considered is one already alluded to in evaluating the SA−

SM series: inherent vibronic congestion. One clear trend in the

series presented here is the level of congestion brought on by

low-frequency modes, coupled with the extensive activity of the

b-mode (Figure 3b). The malate side chain presents the largest

reduced mass in the series, resulting in the lowest frequencies

for the set of active vibrations sensitive to the electronic

absorption. The sinapate chromophore itself is inherently

sensitive to these large amplitude vibrations due to the nature

of the S₀−V transition, which extends out to the site of

substitution of the malate side chain (Figure 11a). While a

simple increase in congestion is undoubtedly a contribution to

the broad and extended absorption of SM, it is unlikely that this

alone would lead to a complete washing out of all sharp

structure even at low frequency near the electronic origin,

where the S₁vibrational state density is lowest. For instance,

even in SMB where the spectral broadening ﬁrst becomes

apparent, vibronic analysis can still be accomplished up to the

4.2.2. Excited State Dynamics. The origin of the spectral

broadening in sinapoyl malate, even under jet-cooled

conditions, could be the result of facile, intramolecular decay

pathways. The introduction of the dicarboxylic acid (or

dicarboxylate in SDM), and to some extent the single

carboxylate chain in SMB, instigated the dramatic change

observed in the UV spectrum. With regard to excited state

dynamics, the likely states that would be modulated by this

chemical substitution are the nπ* states that are conjugated to

some degree into the malate side chain, diﬀering from sinapic

acid and the simpler derivatives.

In this scenario, SM and SDM could engage in the same

decay mechanism as the pCA/MeOpCA molecules, namely,

1

b¹band.

direct internal conversion to an adiabatically lower nπ* state

0

(n-nonbonding orbital on C₉O).²⁰⁻²²Vertical excitation

energy calculations of SM (Table S2, Supporting Information)

do not predict a substantial change in the V(¹ππ*)−¹nπ*

vertical splitting compared with the other sinapates, with the

¹nπ* state well above the V′ state in vertical excitation.

However, the adiabatic energy of the ¹nπ* state after geometry

optimization (e.g., involving elongation of the C₉O bond and

other structural distortions) has been shown to drop the energy

of the nπ* state dramatically from the vertical/ground state

geometry, placing it lower than the bright V′ state in pCA and

In addition to vibronic congestion, however, intramolecular

vibrational redistribution (IVR) could also contribute as a

mechanism for vibronic congestion/broadening. While the

number of low-frequency modes (which provide the predom-

inant contribution to the bath for IVR) does increase along the

sinapate ester series, the increase is actually quite modest, with

the isopropyl sinapate/sinapoyl methyl lactate/sinapoyl malate

series possessing seven/eight/nine modes, respectively, with

frequencies below 100 cm⁻¹. Furthermore, the decrease in

frequency of these modes with increasing size of the side chain

is also quite small. Sinapoyl methyl-lactate (SML) provides a

testing ground for the eﬀects of IVR in S₁, since the spectrum is

still resolvable, yet dense. Indeed, some apparent band

broadening is observed beginning about 200−300 cm⁻¹

above the conformer A origin transition; however, this is

attributed more to spectral overlap than from IVR. As such, it is

MeOpCA.^21,22,43Unfortunately, attempts to optimize the nπ*

1

state were unsuccessful for SM at the level of theory presented

here. Nevertheless, this pathway can be evaluated from an

experimental perspective.

If this was a dominant mechanism at play in SM, the

adiabatic energy of the nπ* state would indeed have to be lower

14791

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

than S₁but not well below. The present experiment is sensitive

to the ππ*/nπ* energy diﬀerence through the ionization step

inherent to the R2PI process. According to calculations, the

geometry of the ion produced by the R2PI process is not very

The UV spectra of SML and SMB provide some conﬁrming

evidence for this scenario. First, while the SML UV spectrum

still contains a sharp, resolved vibronic structure like that in the

alkylated sinapate esters, it also displays some characteristics

that hint at the presence of a second state to which the ππ*

state is coupled. First, as Figure 12 demonstrates, the spectrum

1

diﬀerent than the ground state neutral. If the nπ* state is

populated by fast internal conversion, the excess vibrational

energy produced by the process will be carried into the ion. As

a result, ionization will require a photon energy from S₁large

enough to reach into this vertical region of the ion. If the

energy diﬀerence were too large, the R2PI process would be

ineﬃcient. However, all indications are that the R2PI process is

relatively facile throughout the entire series of sinapate esters,

arguing for the state out of which ionization occurs, retaining

most of the electronic excitation energy originally placed in it

by the ﬁrst photon. If coupling to the nπ* state is a dominant

pathway in sinapoyl malate, we estimate that it must occur on a

sub-ps time scale to suﬃciently lifetime-broaden the vibronic

transitions to the extent observed in the spectrum. Calculations

1

of the adiabatic energy of this nπ* state in SM and adiabatic

and vertical ionization energies from that geometry are essential

for evaluating this pathway, oﬀering a challenge to theory.

The geometry change associated with the V (ππ*) state

accessed in the sinapates has been proposed in related

molecules to lead to eﬃcient trans−cis isomerization about

the C₇−C₈double bond.^23,24,43This excited state isomerization

is a pathway by which the molecule can encounter a conical

intersection with the ground state, providing a nonradiative

decay pathway to a vibrationally hot ground state species. While

this pathway is intriguing to consider in explaining SM excited

state dynamics, in the present gas phase experiment, we exclude

it as a possibility, as ionization from the ground state surface

with vibrational energy equivalent to the UV photon energy is

prohibited by Franck−Condon considerations, as just dis-

cussed.

Another possible type of state to which the ππ* state could

be coupled is a charge transfer (CT) state. The presence and

consequences of coupling of a locally excited state (LE) with a

CT state are well-documented in other circumstances.^40,44

When mixing of a locally excited (LE) state occurs with a CT

state, it often leads to complete spectral broadening of the

transition to the LE state due to coupling to the dense set of

vibronic levels in the adjacent, adiabatically lower CT state that

borrows intensity from the bright LE state.⁴⁰The mixing

spreads the ππ* oscillator strength over a manifold of close-

lying CT levels, with a breadth governed by the magnitude of

the average coupling matrix elements between the states.

The intriguing possibility oﬀered by SM is that the ¹nπ* state

itself has some charge transfer character to it. On this basis, we

postulate that the nπ* state and CT state scenarios are not

distinct from one another in SM but that the partial CT

character of the nπ* state contributes to lowering its adiabatic

energy relative to the ππ* state. Thus, the donor−acceptor

groups deﬁning the partial charge-transfer character of the nπ*

state would logically be from the malate side chain to the

sinapate aromatic ring, as illustrated by the nπ* state electron

density diﬀerence map shown in Figure 11d. In the ﬁgure, blue

regions represent gains in electron density while yellow regions

indicate depletions. While the two ππ* transitions involve the

loss of electron density from the ring into an antibonding π*

orbital spanning the length of the sinapate, the nπ* state clearly

shows the transfer of electron density from C₉OO and the ﬁrst

carboxylic acid of the malate group into the antibonding π*

orbital of the sinapate ring.

Figure 12. R2PI spectra for SML at (a) low-ﬂuence excitation

conditions and (b) intermediate ﬂuence conditions (middle) and (c)

R2PI spectrum of SDM at the same intermediate ﬂuence conditions.

of SML showed an unprecedented sensitivity to photon

ﬂuence. While the 2C-R2PI spectrum is clearly unsaturated

and similar to the simpler derivatives (Figure 12a), an increase

in the ﬂuence to that used for acquisition of sharp R2PI spectra

of the alkylated esters SA, MS, and IS yields a signiﬁcant

increase in the intensity of an underlying background (Figure

12b). Plotted against the SDM spectrum shown below, the

proﬁle of the 1C-R2PI of SML bears a striking resemblance to

the sinapoyl malate spectra, perhaps signaling that SML

presents an intermediate coupling case in which the mixed

LE/CT character of the excited state is apparent by the extreme

sensitivity of the spectrum to photon ﬂuence.

A second piece of evidence for the onset of mixing with a

CT-like state is in the emission from SML (DFL spectra can be

found in the Supporting Information). Sharp transitions back

down to Franck−Condon active vibrational levels in the ground

state are observed in the DFL spectra of the origin transitions

in SML, signaling emission from states of predominantly LE

character. However, in comparing the total amount of emission

(in photon counts) in this LE Franck−Condon region with the

“unperturbed” methyl sinapate DFL spectra (Figure 7), a

decrease of at least 2 orders of magnitude in emission is

observed despite the expected diﬀerence of only 1 order of

magnitude based on the total ﬂuorescence signal. One typical

feature of CT state mixing/coupling is that loss of emission in

the LE Franck−Condon region is accompanied by broadened,

red-shifted emission due to the large displacement of the CT

geometry from the ground state.^40,45In order to search for this

component of the emission, the DFL scan was extended out 50

nm to the red of the conformer B origin, yet no discernible

emission was found. We surmise that the shifted emission is

broadened to the point that the photon counts at the iCCD

detector on a per-pixel basis are below the detection limit. It

should be noted that, in our DFL detection scheme, a gated

intensiﬁed CCD with a gate of 10 ns around the laser pulse was

14792

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

used, thereby precluding the possibility of detecting phosphor-

escence, and therefore intersystem crossing mechanisms.

Finally, the SMB R2PI spectrum (Figure 8b) shows an

absorption proﬁle similar to the intermediate-ﬂuence 1C-R2PI

spectrum of SML (Figure 12b), perhaps signaling the initial

onset of enhanced LE/CT mixing. We were unable to record

DFL spectra for SMB, SM, and SDM due to the lack of laser

desorption capabilities in our ﬂuorescence chamber, but direct

detection of broad, red-shifted emission in DFL of SM would

certainly represent an important future step in conﬁrming a

CT-like mechanism. Alternatively, the presence and character

of the second excited state responsible for the broadening

should be apparent in the time evolution of the photoelectron

spectrum.⁴⁶⁻⁴⁸Atomic charges and dipole moments of the

three excited states at the ground state geometry were

calculated by electrostatic potential (ESP) ﬁtting, also revealing

some nπ*/CT character, as summarized in the Supporting

Information.

with ab initio calculations, demonstrate the high degree of

spectral broadening brought on in ππ* transitions by a nearby

charge-transfer state.⁵⁰

As a ﬁrst step toward understanding the spectroscopy and

photophysics of sinapoyl malate in solution, the isolated, gas

phase results presented herein are compared with steady-state

absorption (black) and ﬂuorescence (red) spectra of sinapic

acid and sinapoyl malate in aqueous solution shown in Figure

13. Spectra were acquired with a Cary 100 UV−vis

Finally, while this discussion has focused attention on the

¹nπ* excited state, there is also possibly a role for intersystem

3

crossing to the nπ* state if its energy is also lowered by its

partial CT character. The presence of eﬃcient intersystem

crossing would be detrimental to SM’s role as a UV-B

sunscreen, since triplet states could lead to formation of singlet

oxygen via triplet−triplet annihilation. However, there is no

experimental evidence for formation of a long-lived triplet state

either in ﬂuorescence or in two-color R2PI. The presence of a

long-lived triplet would be detectable in 2C-R2PI as the delay

between the ﬁrst and second photon was increased. Never-

theless, future investigations that search more carefully for the

production of triplet states both in the gas phase and in

solution would be worthwhile.

Figure 13. UV absorption spectra (black) and ﬂuorescence spectra

(red) taken by excitation at the absorption maximum of sinapic acid

(top) and sinapoyl malate (bottom). A solute concentration of ∼8 ×

10⁻⁶M was used to achieve a maximum optical density of ∼0.1 in a 1

cm path length (ε_{324 nm}∼ 12700 cm⁻¹M⁻¹for sinapoyl malate). The

UV-B region spans 280−315 nm.

4.2.3. Condensed Phase Spectroscopy. The anomalous

broadening observed in the UV spectrum of jet-cooled, gas

phase sinapoyl malate provides insight into the decay

mechanism intrinsic to the molecule in isolation. While this

inherent broadening is unusual under such conditions, the

conclusion drawn from this study, namely, that ππ*/nπ*

mixing is likely responsible for the observed broadening, is not

unique to sinapoyl malate, nor to the gas phase. In fact, in the

seminal review by R. M. Hochstrasser more than 40 years ago,

the perturbations imposed on ππ* transitions due to low-lying

nπ* states in the condensed phase are described in some

detail.⁴⁹Here, several examples are given where nπ* states are

in close proximity to ππ* states, such as in nitrogen

heterocycles and ketones, and in some cases the ordering of

the two states can be modulated by the solvent environment. In

cases where an nπ* state falls below the bright ππ* state,

spectral broadening of the resolved vibronic bands occurs, at

times to the point of complete broadening.⁴⁹As discussed for

sinapoyl malate, this is due to vibronic mixing of the ππ* levels

with background nπ* levels. Hochstrasser emphasizes the

importance of these interactions, as band broadening in the

condensed phase is a good indication of the existence of a

lower-energy background state. More speciﬁcally, the diﬀer-

ences in the orbital character of ππ* and nπ* states (or charge-

transfer states) make them primary candidates for mixing, since

the Franck−Condon factors that govern vibronic mixing are

large when the two states are nearby in energy.⁴⁹In the case of

sinapoyl malate, where the nπ* state is itself charge-transfer-like

in character, it is possible that such vibronic interactions are

enhanced. More recent studies on carbonyl-containing

polyenes and carotenoids in the condensed phase, coupled

spectrometer and Fluorolog FL3 ﬂuorescence spectrometer.

Methanol and aqueous solution at neutral pH and pH ∼ 3 (for

increase in [SM]/[SM⁻]; vacuolar pH ∼ 6.1¹²) were used as

solvents, with minimal diﬀerence between the observed spectra.

As anticipated, the well-structured vibronic activity in the gas

phase spectrum of sinapic acid is largely lost in solution. The

shoulder at 288 nm is some 3200 cm⁻¹above the main ππ*

transition peaked at 318 nm, and could be due to the nπ* state.

The main band in sinapoyl malate (323 nm) is peaked just

slightly red of that in sinapic acid, and is somewhat broader but

otherwise similar in appearance. The loss of the shoulder on the

short wavelength edge likely signals a red-shift of the nπ*

transition.

The most striking aspect of the comparison between sinapic

acid and sinapoyl malate is in the emission spectra, where a

substantial Stokes shift of 127 nm is present in sinapoyl malate

(Figure 13, bottom), almost eliminating any spectral overlap

between absorption and emission. The ﬂuorescence bands of

both sinapates are ∼100 nm fwhm, and the sinapoyl malate

band (450 nm) is red-shifted signiﬁcantly from sinapic acid

(∼410 nm). When combined, these results point to the active

involvement of an intramolecular charge-transfer state in

sinapoyl malate in polar solution phase environments, likely

the nπ* state.

Finally, the ﬂuorescence quantum yields for both sinapic acid

and sinapoyl malate in aqueous solution were determined using

9,10-diphenylanthracene as a reference. Both molecules had

ﬂuorescence quantum yields of Φ ∼ 3 × 10⁻³, indicating that

14793

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

nonradiative processes dominate the excited state decay

pathways of both molecules in polar solvents. Future studies

that directly detect the intermediate and ﬁnal states involved in

the decay processes, both in the gas phase and in solution, are

clearly warranted.^50,51Further, while these preliminary

condensed phase results suggest the retention of a dominant

V−CT interaction even in the presence of hydrogen bonding

from polar solvents, jet studies of microsolvated sinapoyl

malate:H₂O clusters would be useful in assessing the eﬀects of

hydrogen bonding from individual water molecules on the

excited states relevant to the nonradiative dynamics. This would

present a test case for comparison with the substituted

cinnamates, where interactions with only a single water

molecule shift the nπ* state out of range of the bright V′

state, at times dramatically aﬀecting the vibronic features and/

or decay pathways.^20,23

Taken as a whole, these initial results in aqueous solution

provide preliminary support for the notion that the excited state

processes ascribed to the isolated sinapoyl malate molecule are

also active in solution. At the same time, their preliminary

nature motivates further photophysical and photochemical

studies that more accurately mimic the in vivo environment of

the plant itself.

taken as a whole suggest a selection of these robust, quantum

mechanical features by nature as a means of biologically

eﬃcient photoprotection. However, further photochemical

studies both in the gas phase and in solution are required to

fully investigate the nature of the excited state dynamics and the

extent to which the inherent intramolecular processes are

modiﬁed in their in vivo environment in plants.

The presence of an eﬃcient mechanism for intramolecular

excited state dynamics leaves open the very important question

of the ﬁnal fate of the energized molecule accessed by

absorption of a UV-B photon in sinapoyl malate. If the

coupling of the ππ* state to a second excited state directs the

electronic energy back to the ground state via internal

conversion, it seems likely that the molecule will collisionally

cool on the ground state potential energy surface without

undergoing signiﬁcant photochemistry. This would complete a

cycle in which absorption of UV-B radiation leads only to heat

dissipated in solution, bringing sinapoyl malate back to its

ground state where it can eﬃciently absorb again. Indeed, this is

what one would anticipate of an eﬃcient UV-B sunscreen, in

which the plant needn’t replenish its supply of the molecule due

to photochemical degradation. However, to the best of our

knowledge, this has not yet been proven experimentally, nor

have the photophysical and photochemical pathways been

explored in any detail from a computational viewpoint, calling

for future eﬀort along both directions.

5. CONCLUSIONS

Even under isolated conditions in the gas phase, the sinapate

esters, and particularly sinapoyl malate, display properties that

justify their role as UV-B screening agents in plants. The

electronic transition involved in providing this protection has

remarkable absorption eﬃciency (large oscillator strength) and

beautifully matched spectral coverage in the UV-B range. The

large absorption coeﬃcient for the transition ensures eﬃcient

absorption of photons incident the leaf epidermis (where SM is

concentrated). Suﬃcient wavelength coverage results from a

ππ* transition involving molecular orbitals which extend over

the full length of the sinapate group, and produce a geometry

change between S₀and S₁states that leads to long Franck−

Condon progressions involving low-frequency vibrational

modes. The result is excitation spectra that are congested and

extend over the entire UV-B region. Indeed, the relatively small

wavelength range studied here is likely to represent only a

fraction of the full absorption proﬁle, with progressions

involving CC and CO stretch modes, as observed in the

DFL spectra of MS (Figure 7).

In increasing the complexity of the ester side chain attached

to the sinapate group, increased spectral complexity was

observed until complete spectral broadening spanning >1000

cm⁻¹was reached in SM and SDM. We have argued that the

broadening in SM/SDM, and the onset in SML and SMB, can

be consistently accounted for by excited state mixing of the V

(ππ*) state and a second excited state, most probably the low-

energy nπ* state, which inherently incorporates signiﬁcant

charge-transfer character. This, coupled with the inherently

congested sinapate UV spectrum, yields complete broadening

in its UV spectrum even under jet-cooled conditions in the gas

phase, a remarkable result. By chemically “deconstructing”

sinapoyl malate, and comparing its UV spectral properties with

the simpler analogues in this series, a clear progression in

behavior is found that illuminates the inherent and unique

excited state properties of sinapoyl malate. A cursory analysis of

the absorption and ﬂuorescence spectra of sinapoyl malate in

aqueous solution demonstrates potential support of this

mechanism even in an aqueous environment. These results

ASSOCIATED CONTENT

* Supporting Information

■

S

The experimental procedures for synthesis of MS, IS, SML,

SMB, and SDM, experimental details of the laser desorption

scheme used in this work, experimental aspects of the double-

resonance spectroscopy employed in this work, vibrational

frequencies for alkylated sinapates, relaxed potential energy

scan about the lowest syn−anti interconversion coordinate in

SA, structural assignments for SML and SMB, analysis of the

small bands to the red of the SML (A) origin, DFL spectra of

SML origin transitions, SM/SDM low-energy structures, RIDIR

spectra of SM and SDM, Franck−Condon simulations of the V

(ππ*) and V′ (ππ*) excitations in SA, and TDDFT scaled

vertical excitations plotted against experimental R2PI spectra.

This material is available free of charge via the Internet at

http://pubs.acs.org.

AUTHOR INFORMATION

Corresponding Author

zwier@purdue.edu

■

Notes

The authors declare no competing ﬁnancial interest.

ACKNOWLEDGMENTS

■

The authors gratefully acknowledge Yohei Matsueda for his

noteworthy contributions to the synthesis of MS, IS, SML,

SMB, and SDM. J.C.D. and T.S.Z. would like to thank Clint

Chapple for discussions about sinapoyl malate and its role in

UV-B protection in plants. The authors also acknowledge

support for this work from the Department of Energy Basic

Energy Sciences, Division of Chemical Sciences (DEFG02-

96ER14656).

REFERENCES

(1) Bieza, K.; Lois, R. Plant Physiol. 2001, 126, 1105−1115.

■

14794

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Journal of the American Chemical Society

Article

(2) Booij-James, I. S.; Dube, S. K.; Jansen, M. A. K.; Edelman, M.;

Mattoo, A. K. Plant Physiol. 2000, 124, 1275−1283.

(3) Landry, L. G.; Chapple, C. C. S.; Last, R. L. Plant Physiol. 1995,

109, 1159−1166.

Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;

Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;

Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09;

Gaussian, Inc.: Wallingford, CT, 2009.

(32) Dean, J. C.; Buchanan, E. G.; James, W. H.; Gutberlet, A.;

Biswas, B.; Ramachandran, P. V.; Zwier, T. S. J. Phys. Chem. A 2011,

115, 8464−8478.

(4) Ren, J.; Dai, W. R.; Xuan, Z. Y.; Yao, Y. N.; Korpelainen, H.; Li,

C. Y. For. Ecol. Manage. 2007, 239, 112−119.

(5) Demkura, P. V.; Ballare, C. L. Mol. Plant 2012, 5, 642−652.

(6) Foyer, C. H.; Lelandais, M.; Kunert, K. J. Physiol. Plant. 1994, 92,

696−717.

(7) Jansen, M. A. K.; Gaba, V.; Greenberg, B. M.; Mattoo, A. K.;

Edelman, M. Plant J. 1996, 9, 693−699.

(8) Jordan, B. R. In Advances in Botanical Research; Callow, J. A., Ed.;

Academic Press Ltd.: London; San Diego, CA, 1996; Vol. 22, pp 97−

162.

(9) Ruegger, M.; Chapple, C. Genetics 2001, 159, 1741−1749.

(10) Chapple, C. C. S.; Vogt, T.; Ellis, B. E.; Somerville, C. R. Plant

Cell 1992, 4, 1413−1424.

(11) Fraser, C. M.; Chapple, C. Arabidopsis Book 2011, 9, e0152.

(12) Sheahan, J. J. Am. J. Bot. 1996, 83, 679−686.

(13) Strack, D.; Pieroth, M.; Scharf, H.; Sharma, V. Planta 1985, 164,

507−511.

(14) Karpkird, T. M.; Wanichweacharungruang, S.; Albinsson, B.

Photochem. Photobiol. Sci. 2009, 8, 1455−1460.

(33) Duschinsky, F. Acta Physicochim. URSS 1937, 7, 551−566.

(34) Muller, C. W.; Newby, J. J.; Liu, C. P.; Rodrigo, C. P.; Zwier, T.

S. Phys. Chem. Chem. Phys. 2010, 12, 2331−2343.

(35) Zilberg, S.; Haas, Y. J. Chem. Phys. 1995, 103, 20−36.

(36) Sibert, E. L.; Tabor, D. P.; Dean, J. C.; Kidwell, N. M.; Zwier, T.

S. Manuscript in preparation.

(37) Henderson, J. R.; Muramoto, M.; Willett, R. A. J. Chem. Phys.

1964, 41, 580.

(38) Cogan, S.; Zilberg, S.; Haas, Y. J. Am. Chem. Soc. 2006, 128,

3335−3345.

(39) Schweke, D.; Baumgarten, H.; Haas, Y.; Rettig, W.; Dick, B. J.

Phys. Chem. A 2005, 109, 576−585.

(40) Vandantzig, N. A.; Shou, H. S.; Alfano, J. C.; Yang, N. C. C.;

Levy, D. H. J. Chem. Phys. 1994, 100, 7068−7078.

(41) Dean, J. C.; Navotnaya, P.; Parobek, A. P.; Clayton, R. M.;

Zwier, T. S. J. Chem. Phys. 2013, 139, 144313.

(42) Rodrigo, C. P.; James, W. H.; Zwier, T. S. J. Am. Chem. Soc.

(15) Lewis, F. D.; Quillen, S. L.; Elbert, J. E.; Schneider, S.;

Geiselhart, P. J. Photochem. Photobiol., A 1989, 47, 173−179.

(16) Pattanaargson, S.; Limphong, P. Int. J. Cosmet. Sci. 2001, 23,

153−160.

2011, 133, 2632−2641.

(43) Gromov, E. V.; Burghardt, I.; Koppel, H.; Cederbaum, L. S. J.

Phys. Chem. A 2005, 109, 4623−4631.

(44) Jiang, S.; Levy, D. H. J. Phys. Chem. A 2002, 106, 8590−8598.

(45) Fleisher, A. J.; Bird, R. G.; Zaleski, D. P.; Pate, B. H.; Pratt, D.

W. J. Phys. Chem. B 2013, 117, 4231−4240.

(17) Promkatkaew, M.; Suramitr, S.; Karpkird, T. M.; Namuangruk,

S.; Ehara, M.; Hannongbua, S. J. Chem. Phys. 2009, 131, 224306.

(18) Singh, T. S.; Mitra, S.; Chandra, A. K.; Tamai, N.; Kar, S. J.

Photochem. Photobiol., A 2008, 197, 295−305.

(46) Schalk, O.; Schuurman, M. S.; Wu, G. R.; Lang, P.; Mucke, M.;

Feifel, R.; Stolow, A. J. Phys. Chem. A 2014, 118, 2279−2287.

(47) Stolow, A. Int. Rev. Phys. Chem. 2003, 22, 377−405.

(48) Wolf, T. J. A.; Schalk, O.; Radloff, R.; Wu, G.; Lang, P.; Stolow,

A.; Unterreiner, A. N. Phys. Chem. Chem. Phys. 2013, 15, 6673−6683.

(49) Hochstrasser, R. M. Acc. Chem. Res. 1968, 1, 266−274.

(50) Enriquez, M. M.; Fuciman, M.; LaFountain, A. M.; Wagner, N.

L.; Birge, R. R.; Frank, H. A. J. Phys. Chem. B 2010, 114, 12416−

12426.

(19) Wang, H. B.; Zhai, B. C.; Tang, W. I.; Yu, J. Y.; Song, Q. H.

Chem. Phys. 2007, 333, 229−235.

(20) Tan, E. M. M.; Hilbers, M.; Buma, W. J. J. Phys. Chem. Lett.

2014, 5, 2464−2468.

(21) de Groot, M.; Gromov, E. V.; Koppel, H.; Buma, W. J. J. Phys.

Chem. B 2008, 112, 4427−4434.

(22) Smolarek, S.; Vdovin, A.; Perrier, D. L.; Smit, J. P.; Drabbels, M.;

Buma, W. J. J. Am. Chem. Soc. 2010, 132, 6315−6317.

(23) Tan, E. M. M.; Amirjalayer, S.; Bakker, B. H.; Buma, W. J.

Faraday Discuss. 2013, 163, 321−340.

(51) Bautista, J. A.; Connors, R. E.; Raju, B. B.; Hiller, R. G.;

Sharples, F. P.; Gosztola, D.; Wasielewski, M. R.; Frank, H. A. J. Phys.

Chem. B 1999, 103, 8751−8758.

(24) Shimada, D.; Kusaka, R.; Inokuchi, Y.; Ehara, M.; Ebata, T. Phys.

Chem. Chem. Phys. 2012, 14, 8999−9005.

(25) Allais, F.; Martinet, S.; Ducrot, P. H. Synthesis 2009, 3571−

3578.

(26) Quentin, M.; Allasia, V.; Pegard, A.; Allais, F.; Ducrot, P. H.;

Favery, B.; Levis, C.; Martinet, S.; Masur, C.; Ponchet, M.; Roby, D.;

Schlaich, N. L.; Jouanin, L.; Keller, H. PLoS Pathog. 2009, 5, 1−9.

(27) Dean, J. C.; Buchanan, E. G.; Zwier, T. S. J. Am. Chem. Soc.

2012, 134, 17186−17201.

(28) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.;

Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J.

Comput. Chem. 1990, 11, 440−467.

(29) Weiner, P. K.; Kollman, P. A. J. Comput. Chem. 1981, 2, 287−

303.

(30) Zhao, Y.; Truhlar, D. G. J. Chem. Theory Comput. 2007, 3, 289−

300.

(31) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;

Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,

B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.

P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;

Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,

T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;

Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,

K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;

Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,

N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;

Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;

Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;

14795

dx.doi.org/10.1021/ja5059026 | J. Am. Chem. Soc. 2014, 136, 14780−14795

Article Doi

Plant sunscreens in the UV-B: Ultraviolet spectroscopy of jet-cooled sinapoyl malate, sinapic acid, and sinapate ester derivatives

DOI: 10.1021/ja5059026

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja5059026

Full text of DOI:10.1021/ja5059026

Products guided by the article

R&D Labs maybe for 1270043-91-8

Relevant to this article

Hot Product