On the identity of the radiation-induced stable alanine radicalw

Ewald Pauwels,*aHendrik De Cooman,

aMichel Waroquier,

aEli O. Hole

band

Einar Sagstuenb

Received 17th March 2010, Accepted 27th May 2010

DOI: 10.1039/c004380j

Using periodic DFT calculations, it is concluded that the stable

radiation-induced alanine radical most probably is the result of

reductive deamination and protonation of the detached amino

group, yielding an NH4+ ammonium ion and a negatively

charged radical.

Alanine dosimetry is an established technique for measuring

absorbed doses of ionizing radiation.1 When crystals or

powders of the amino acid L-a-alanine are exposed to ionizing

radiation, free radicals are formed and stably trapped in the

matrix. The number of stable radicals is proportional to the

absorbed dose, radiation dose read-out being enabled by

Electron Paramagnetic Resonance (EPR) spectroscopy.

Several radicals have been found to contribute to the overall

EPR spectrum of alanine, although a well-defined resonance

pattern due to one of the radiation-induced radical species,

often called the ‘Stable Alanine Radical’ (SAR), dominates the

spectrum.2 The structure of this radical is commonly assumed

to be that of B (Scheme 1), the stable end product of several

radiation-induced reductive processes including deamination

and protonation.3 This structure was proposed on the basis of

careful analysis of the g- and hyperfine couplings in several

EPR and ENDOR studies. In Table 1, the results of ref. 2 are

reproduced. The a- and b-proton hyperfine interactions clearly

indicate a p-type carbon-centred radical, with one hydrogen

atom and one methyl group directly bound to the radical

centre. The protonation of the carboxyl group in structure B

was first (indirectly) suggested by Kuroda and Miyagawa4

who found an additional weak hyperfine coupling (indicated

by the shorthand H(N) in Table 1) and attributed this to a

proton located close to one of the oxygens of the –COO�

group of the radical. This is consistent with chemical intuition,

as the proton compensates for the net negative charge induced

by the radiation-induced reduction event.

Since the advent of Density Functional Theory (DFT),

several theoretical studies have been performed for the

purpose of reproducing the EPR properties of alanine

radicals5,6 and the SAR in particular.7–9 However, only the

isotropic parts of the hyperfine couplings were calculated and

the radicals were considered being in the gas phase or in a very

rudimentary solution model. In the present work, a complete

theoretical determination of the g- and hyperfine coupling

tensors of the SAR using periodic DFT calculations is

reported. The application of periodic boundary conditions to

simulate the solid-state environment of a radical has proven

particularly successful in the reproduction of these EPR

properties.10,11 In addition, periodic calculations allow an

evaluation of the stability of a radical structure within its

solid-state environment and can also give insight into its

formation mechanism.12,13 The calculations strongly suggest

that B is not the correct structure of the SAR. Instead,

convincing evidence is presented in support of structure C, a

radical structure in which the abstracted amino group is

protonated, instead of the carboxyl group of the radical.

All DFT calculations were performed with the CP2K

software14 and the BLYP functional.15,16 The basic unit for

the periodic calculations was chosen to be the crystallographic

unit cell of alanine,17 duplicated along all directions [2a2b2c].

In this way, the interaction of a radical with its periodic images

is prevented. The Gaussian and plane waves (GPW) dual basis

set method18 was used in all geometry optimizations, employing

a TZVP triple-z Gaussian basis set19 and plane waves (300 Ry

density cut-off) with GTH pseudopotentials.20,21 For the

subsequent g- and hyperfine coupling tensor calculations, we

relied on recent implementations10,11 in the CP2K code,

employing the all-electron Gaussian and augmented plane

wave (GAPW) method.22 The density cut-off for the auxiliary

plane wave basis set was 200 Ry and the all-electron TZVP

basis23 was used. This methodology has proven successful in

other studies.24

Since the SAR presumably is an end product of reductive

radiation damage to the alanine crystal, we started by

simulating a periodic box with a net negative charge. One of

the 16 alanine molecules in this box was altered in accordance

with structure A (Scheme 1) by elongating the C–N bond.

Subsequent optimization proved this structure to be stable,

with an absolute energy of �1010.16267 a.u. for the entire

periodic box. Only deamination took place during the

optimization of this structure, yielding a separate NH3

molecule. Intriguingly, the carboxyl group of this radical

Scheme 1 The chemical structure of L-a-alanine and several suggested

structures for the Stable Alanine Radical.

a Center for Molecular Modeling, Ghent University, Technologiepark903, B-9052 Zwijnaarde, Belgium, QCMM-alliance Ghent-BrusselsBelgium. E-mail: ewald.pauwels@UGent.be

bDepartment of Physics, University of Oslo, P.O. Box 1048 Blindern,N-0316 Oslo, Norway

w Electronic supplementary information (ESI) available: CalculatedEPR properties for B-like structures. See DOI: 10.1039/c004380j

This journal is �c the Owner Societies 2010 Phys. Chem. Chem. Phys., 2010, 12, 8733–8736 | 8733

COMMUNICATION www.rsc.org/pccp | Physical Chemistry Chemical Physics

Publ

ishe

d on

18

June

201

0. D

ownl

oade

d by

Bos

ton

Col

lege

on

05/0

9/20

13 1

5:03

:01.

View Article Online / Journal Homepage / Table of Contents for this issue

remained not protonated, thus maintaining a net negative

charge on the radical. Yet, Kuroda and Miyagawa4 argued

in their experimental paper that the carboxyl group is proto-

nated, and that this COOH proton is transferred from the

NH3 group of a nearby, hydrogen-bound molecule. In the

alanine crystal, the carboxyl group of each molecule is

involved in three hydrogen bonds with the amino groups of

three separate neighbouring molecules. In structure A (Fig. 1),

the radical participates in the same three hydrogen bonds.

Proton transfer along each of these three routes was explored.

Starting from structure A, the O–H distance between the

carboxyl oxygen of the radical and the amino proton of a

nearby molecule was constrained to a value of 0.9 A and the

geometry was optimized. This resulted in three structures,

similar to structure B, in which the negative charge of the

radical was effectively transferred to the neighbouring

molecule. However, these constrained species proved to be

severely less stable than structure A (with energy differences of

about 600 kJ mol�1), and they all reverted back to structure A

in a second, unconstrained optimization.

The instability of these structures suggests that structure B

is not valid. However, comparison of the calculated EPR

properties for structure A with the experimental values gives

no convincing evidence that A is the true structure of the SAR

either. As is evident from Table 1, two major hyperfine

interactions are present in this structure. Their anisotropic

couplings are characteristic for the typical patterns of a- andb-type hyperfine interactions. The corresponding isotropic

couplings, on the other hand, are largely underestimated with

respect to the experimental data. Even though this particular

EPR property depends significantly on the chosen density

functional or basis set,25 a difference of 30 MHz altogether

for the Hb interaction is unrealistically large. In addition, the

eigenvectors of the anisotropic hyperfine interactions do not

agree well with their experimental counterparts: the angles

between corresponding eigenvector directions are at least 251.

The calculated g-tensor is somewhat better: the g-values are

quite comparable and the eigenvector corresponding to the

maximum g-value deviates only 131 from the corresponding

experimental direction. Nevertheless, previous studies with a

similar methodology13,24,26,27 have clearly shown that valid

radical structures give rise to calculated EPR properties that

are in much better accordance with experimental data. Here,

the overall agreement of all EPR parameters is disappointing

and strongly suggests that structure A is not the appropriate

structure of the SAR. Structure B is also rejected, from

stability considerations.

In the previous simulations, the net negative charge of the

periodic cell (induced by the reductive radiation event) was not

counteracted. Proton transfers were considered, but only

between the radical and one of its surrounding molecules. In

that situation, the negative charge is still localized in the direct

vicinity of the radical and, in addition, the hydrogen-bonding

network of the radical is modified. Proton transfer between the

radical and an alanine molecule further away in the crystal

would leave this network intact and would also restore the

local charge neutrality. Such long-range proton transfers do

Table 1 Isotropic and anisotropic hyperfine couplings (in MHz) andg-tensor values of the stable alanine radical (experiment) and twomodel structures. The last column indicates the angle (in degrees)between corresponding experimental and calculated eigenvectordirections.

Aiso/giso Aaniso/ganiso Angle

Experiment Hb 69.9 �2.6 Ref. 2�2.34.8

Ha �56.1 �31.83.927.9

g 2.0033 2.00242.00342.0041

H(N) 0.2 �4.0 Ref. 4�1.75.7

A Hb 37.4 �2.8 47�2.2 475.0 25

Ha �36.1 �26.2 81�2.4 9028.6 33

g 2.0036 2.0023 262.0039 232.0047 13

C Hb 71.5 �2.7 3�2.0 34.7 2

Ha �42.6 �30.1 1�2.6 432.6 4

g 2.0041 2.0023 32.0045 42.0056 5

H(N) 0.0 �4.2 3�1.0 35.2 6

Fig. 1 Three-dimensional view of the periodic cell of alanine with the

optimized structures A and C. The radical centres are represented by

balls and sticks, and are additionally expanded at the bottom of the

Figure. Green lines indicate hydrogen bonds between the radical

carboxyl group and amino groups of neighbouring alanine molecules.

8734 | Phys. Chem. Chem. Phys., 2010, 12, 8733–8736 This journal is �c the Owner Societies 2010

Publ

ishe

d on

18

June

201

0. D

ownl

oade

d by

Bos

ton

Col

lege

on

05/0

9/20

13 1

5:03

:01.

View Article Online

occur in irradiated biomolecular crystals. Not only are protons

formed as side products of oxidative radiation damage,

‘hopping’ of protons along hydrogen-bond chains has been

observed in both experimental28–30 and theoretical studies.31

The effects of such a process can most easily be simulated by

just adding one proton to the periodic cell, neutralizing the net

negative charge. Reconsidering structure A, there are two

likely sites for this proton: the –COO� group and the detached

NH3 group. Addition to either oxygens of the radical carboxyl

group resulted in stable structures similar to B (with the

difference that the COOH proton was now simply added

and not transferred from a neighbouring molecule). Proton

addition on the detached amino group resulted in the

formation of an NH4+ ammonium ion next to the negatively

charged radical (structure C in Scheme 1). This species proved

significantly more stable than the other two, with an absolute

energy of �1010.8671 a.u. at least 100 kJ mol�1 lower than the

others.

The calculated EPR properties for the B-like structures

(presented in the ESI as B-1 and B-2w) are in poor corres-

pondence with the experimental values for the SAR. Those of

structure C, on the other hand, are in near perfect agreement

with experiment (Table 1). The isotropic Hb coupling differs

by less than 2 MHz, whereas the difference between the

calculated and experimental Ha isotropic coupling is reduced

to 13 MHz. Most strikingly, however, is that all experimental

eigenvector directions are reproduced to within 61. In previous

works, these directional parameters proved to be particularly

useful to gauge the accuracy of a proposed radical model.32

The g-tensor is also nicely reproduced. Even though the actual

g-values differ somewhat more from their experimental

counterparts than for structure A, the eigenvector directions

are almost perfectly aligned with experiment. Comparing the

radical structures A and C in the lower part of Fig. 1, it is clear

that the additional amino proton has induced a reorientation

of the Ca–Ha bond in structure C. This feature is clearly

responsible for the improved agreement of calculated EPR

properties for structure C.

In view of its stability and the overall exceptional agreement

between calculated and experimental EPR properties, there is

little doubt that structure C to date represents the most

probable candidate for the structure of the stable alanine

radical. Since this structure has both a positively and

negatively charged part (illustrated in Scheme 1), it closely

resembles the zwitterionic form of undamaged alanine

molecules, which might explain the observed stability of this

radical. Finally, the additional H(N) dipolar hyperfine inter-

action that was observed by Kuroda et al.4 is clearly not due to

a proton attached to the carboxyl group. Rather, the proton is

bound to the amino group but still is in close proximity to the

radical (Fig. 1). The calculated hyperfine tensor for this

additional amino proton in structure C (also shown in

Table 1) is in perfect agreement with the experimental H(N)

tensor.

It is quite likely that the additional proton in structure C

actually originates from the other half of the radiation-

induced ionization process. A reduction site is generated for

every oxidation site. In alanine, oxidative radical formation

invariably involves deprotonation, giving rise to excess

protons.3 It is conceivable that these protons eventually

migrate towards the reduction sites, where they protonate

the detached NH3 group, effectively restoring charge

neutrality in both the oxidation and reduction site.

In conclusion, we have convincingly determined the true

structure of the stable alanine radical using periodic DFT

calculations. In structure C, the amino group of alanine

is detached and is further protonated, yielding an NH4+

ammonium ion trapped next to the negatively charged

radical.

Acknowledgements

This work is supported by the Fund for Scientific Research—

Flanders (FWO), the Research Board of the Ghent University

and BELSPO in the frame of IAP 6/27. The authors E. P. and

H. D. C. acknowledge a Postdoctoral Fellowship with the

FWO. Part of the computational resources and services used

in this work were provided by Ghent University.

References

1 D. F. Regulla and U. Deffner, Int. J. Appl. Radiat. Isot., 1982, 33,1101.

2 E. Sagstuen, E. O. Hole, S. R. Haugedal and W. H. Nelson,J. Phys. Chem. A, 1997, 101, 9763–9772.

3 E. Sagstuen, A. Sanderud and E. O. Hole, Radiat. Res., 2004, 162,112–119.

4 S. Kuroda and I. Miyagawa, J. Chem. Phys., 1982, 76, 3933–3944.5 E. Pauwels, V. Van Speybroeck, P. Lahorte and M. Waroquier,J. Phys. Chem. A, 2001, 105, 8794–8804.

6 T. L. Petrenko, J. Phys. Chem. A, 2002, 106, 149–156.7 C. Adamo, V. Barone and A. Fortunelli, J. Chem. Phys., 1995, 102,384–393.

8 F. Q. Ban, S. D. Wetmore and R. J. Boyd, J. Phys. Chem. A, 1999,103, 4303–4308.

9 P. Lahorte, F. De Proft, G. Vanhaelewyn, B. Masschaele,P. Cauwels, F. Callens, P. Geerlings and W. Mondelaers,J. Phys. Chem. A, 1999, 103, 6650–6657.

10 R. Declerck, E. Pauwels, V. Van Speybroeck and M. Waroquier,Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 74, 245103.

11 V. Weber, M. Iannuzzi, S. Giani, J. Hutter, R. Declerck andM. Waroquier, J. Chem. Phys., 2009, 131, 014106.

12 H. De Cooman, G. Vanhaelewyn, E. Pauwels, E. Sagstuen,M. Waroquier and F. Callens, J. Phys. Chem. B, 2008, 112,15045–15053.

13 E. Pauwels, H. De Cooman, G. Vanhaelewyn, E. Sagstuen,F. Callens and M. Waroquier, J. Phys. Chem. B, 2008, 112,15054–15063.

14 http://cp2k.berlios.de.15 A. D. Becke, Phys. Rev. A: At., Mol., Opt. Phys., 1988, 38,

3098–3100.16 C. T. Lee, W. T. Yang and R. G. Parr, Phys. Rev. B: Condens.

Matter, 1988, 37, 785–789.17 M. S. Lehmann, T. F. Koetzle and W. C. Hamilton, J. Am. Chem.

Soc., 1972, 94, 2657.18 G. Lippert, J. Hutter and M. Parrinello, Mol. Phys., 1997, 92,

477–487.19 J. VandeVondele and J. Hutter, J. Chem. Phys., 2007, 127,

114105.20 S. Goedecker, M. Teter and J. Hutter, Phys. Rev. B: Condens.

Matter, 1996, 54, 1703–1710.21 C. Hartwigsen, S. Goedecker and J. Hutter, Phys. Rev. B: Condens.

Matter Mater. Phys., 1998, 58, 3641–3662.22 M. Krack and M. Parrinello, Phys. Chem. Chem. Phys., 2000, 2,

2105–2112.23 N. Godbout, D. R. Salahub, J. Andzelm and E. Wimmer,

Can. J. Chem., 1992, 70, 560–571.24 M. Tarpan, E. Sagstuen, E. Pauwels, H. Vrielinck, M. Waroquier

and F. Callens, J. Phys. Chem. A, 2008, 112, 3898–3905.

This journal is �c the Owner Societies 2010 Phys. Chem. Chem. Phys., 2010, 12, 8733–8736 | 8735

Publ

ishe

d on

18

June

201

0. D

ownl

oade

d by

Bos

ton

Col

lege

on

05/0

9/20

13 1

5:03

:01.

View Article Online

25 R. Improta and V. Barone, Chem. Rev., 2004, 104, 1231–1253.26 H. De Cooman, E. Pauwels, H. Vrielinck, E. Sagstuen,

M. Waroquier and F. Callens, J. Phys. Chem. B, 2010, 114,666–674.

27 R. Declerck, E. Pauwels, V. Van Speybroeck and M. Waroquier,J. Phys. Chem. B, 2008, 112, 1508–1514.

28 W. A. Bernhard, J. Barnes, K. R. Mercer and N. Mroczka, Radiat.Res., 1994, 140, 199–214.

29 D. M. Close, L. A. Eriksson, E. O. Hole, E. Sagstuen andW. H. Nelson, J. Phys. Chem. B, 2000, 104, 9343–9350.

30 W. H. Nelson, E. Sagstuen, E. O. Hole and D. M. Close, Radiat.Res., 1998, 149, 75–86.

31 E. Pauwels, R. Declerck, V. Van Speybroeck and M. Waroquier,Radiat. Res., 2008, 169, 8–18.

32 E. Pauwels, V. Van Speybroeck andM.Waroquier, J. Phys. Chem. A,2004, 108, 11321–11332.

8736 | Phys. Chem. Chem. Phys., 2010, 12, 8733–8736 This journal is �c the Owner Societies 2010

Publ

ishe

d on

18

June

201

0. D

ownl

oade

d by

Bos

ton

Col

lege

on

05/0

9/20

13 1

5:03

:01.

View Article Online