X-Ray Absorption Near-Edge Structure
of Amino Acids and Peptides
Jan Boese1 , Angelika Osanna1 , Chris Jacobsen1 , Janos Kirz1 , Ed Tall2 ,
Xiaodong Zhang3
1
2
3
Physics Department, SUNY at Stony Brook, Stony Brook, NY 11794-3800, USA
Department of Physiology and Biophysics, SUNY at Stony Brook, Stony Brook,
NY 11794-8661, USA
Department of Molecular and Cellular Biology, Harvard University, Cambridge,
MA 02138, USA
Abstract. Carbon edge x-ray absorption near edge structure (XANES)
of amino acids and peptides has been investigated for the first time. Because of their heat sensitivity, the samples were analyzed in solid form.
In order to obtain quantitative measurements using transmission spectroscopy, the illuminated spot of the sample has to be of uniform thickness. Therefore a special sample preparation technique was used, and the
spectra were taken using the small spot size of the Stony Brook scanning transmission x-ray microscope. XANES spectra of six amino acid
monomers, four dipeptides and one tripeptide were compared. The results show that peptide spectra can be explained as a sum of the spectra
of the monomers they consist of, indicating that the peptide bond has
only a weak effect on the spectra. If this is true in general, XANES spectra of proteins could be predicted from the spectra of the constituent
amino acids. This could be used in protein specific imaging using x-ray
microscopes.
1
Introduction
A very interesting feature of x-ray microscopes is the possibility to combine imaging with x-ray absorption near-edge structure (XANES) analysis. Comparison
of near-edge absorption resonances with reference spectra makes it possible to
map the concentrations of different double and triple bonds at the high spatial
resolution of an x-ray microscope. This approach has been used to map phases
in polymer blends [1] and to quantify protein and DNA concentrations in sperm
[2]. We want to investigate the applicability of this method to protein specific
imaging.
All proteins found in living systems consist of only 20 different amino acids,
linked together via the peptide bond. It would be interesting to know if a protein
spectrum can be predicted from spectra of the amino acids. We wanted to find
out if an amino acid dimer spectrum is just the sum of the spectra of the two
monomers it consists of. If the effect of the peptide bond on a XANES spectrum
is only weak, a protein spectrum will be just a weighted sum of the spectra
III - 90
J. Boese et al.
of the constituent amino acids. Preliminary work has been done by X. Zhang
and S. Williams [3]. They showed that the above is true for the nitrogen-edge
XANES spectrum of the dipeptide consisting of arginine and aspartic acid. In
this work, carbon-edge XANES spectra of amino acids were taken for the first
time. Spectra of six amino acid monomers, four dipeptides and one tripeptide
were measured and compared.
Similar investigations have been done on nitrogen edge XANES spectroscopy
of DNA [4]. The result was that a DNA spectrum is equivalent to a weighted
sum of polynucleotide spectra, indicating that the differences in base stacking
interactions produces only negligible spectral effects.
2
Experimental Methods
Most of the XANES spectra of molecules have been measured using gas phase
samples. However, evaporation of amino acids and peptides is difficult because
of their heat sensitivity. Therefore they have been analyzed in solid form.
Absorption spectroscopy of solid samples involves a significant difficulty: A
basic requirement is that the thickness of the sample is uniform across the illuminated area.
2.1
Effect of Sample Inhomogeneity
A transmission x-ray absorption spectrum is measured by detecting the transmitted x-rays while varying the photon energy E of the incident beam. The linear
absorption coefficient µ(E) can be determined by taking a spectrum I0 (E) without sample in addition to a spectrum I(E) of the sample of thickness t. The
transmitted intensity is then given by
ZZ
1
e−t(x,y)µ(E) dx dy
I(E) = I0 (E)
A
A is the illuminated area, and x, y the position on the sample. Only if t is
constant across A, the absorption can be determined as
I
1
(1)
µ(E) = − ln .
t
I0
This effect is present in other types of transmission spectroscopy as well. It has
been analyzed for electron energy loss spectroscopy (EELS) by R.Ho et al. [5].
They found that non-uniform specimen thickness has significant effects on the
measurements of elemental concentrations with EELS. The errors in estimates
of concentrations may be as high as 30%. They suggested that the use of a
very small probe, namely the focussed spot of a scanning transmission electron
microscope, can be used to overcome the problem. We take the corresponding
approach here by using a scanning transmission x-ray microscope (STXM) to
acquire XANES spectra. This has the advantage that by taking high resolution
micrographs of the sample, a uniform spot suitable for spectroscopy can be
located. For spectroscopy, the illuminated spot size on the sample was set to
about 10 µm by defocusing to minimize radiation damage to the sample.
X-Ray Absorption Near-Edge Structure
2.2
III - 91
Sample Preparation
The samples were all obtained commercially (from Sigma Chemical Co., St.
Louis, MO) in the form of crystalline powder. To obtain a thin film, the powder
has to be dissolved and then dried on a sample window. There are two problems.
First, some amino acids cannot be dissolved in water. Second, as described above,
it is important for the sample films to be of uniform thickness across the area of
the probe (10µm). A difficulty arises from the fact that the sample film tends to
crystallize during drying. The crystals formed are on the order of microns in size,
creating very uneven samples. To approach this problem, pure trifluoroacetic acid
(TFA) was used as solvent. Because of its very high volatility, the solution dries
quickly enough without crystallization in most cases. TFA is commonly used as
protein solvent. It is a highly corrosive substance and therefore has to be handled
with special care.
2.3
Data Acquisition and Analysis
The spectra were obtained using the Stony Brook scanning transmission x-ray
microscope, which is operated at the National Synchrotron Light Source at
Brookhaven National Laboratory. For each spectrum of a sample, a normalization spectrum of a clean silicon nitride window was taken.
To be able to compare spectra from different samples, it is necessary to obtain an absolute measure of the mass absorption coefficient µm (E) from the
transmission signal. In principle this is possible with Eq. 1 using the signal from
the clean window as I0 and the signal from the sample window as I. However,
the thickness of the samples is not known. It can be calculated using the fact
that away from the edge, molecular bonds do not affect the spectrum. Therefore the absorption coefficient of a molecule is just the sum of the absorption
coefficients of the atoms it consists of. Atomic cross sections have been tabulated by B.L. Henke et al.[6]. These tables were used to obtain mass absorption
coefficients µm (E1 ) and µm (E2 ) at E1 = 279 eV and E2 = 316 eV. The mass
absorption coefficient at energies in between can then be calculated as
I(E1 )
I(E)
µm (E2 ) − µm (E1 )
· − ln
+ µm (E1 ) (2)
+ ln
µm (E) = I0 (E)
I0 (E1 )
ln I(E1 ) − ln I(E2 )
I0 (E1 )
I0 (E2 )
The spectra were calibrated using a spectrum of carbon dioxide gas that was
taken at the beginning of each run. Carbon dioxide has sharp π ∗ resonances
of accurately known energy. This calibration method has been developed by
A.P. Smith et al. [7].
3
Results
Carbon edge XANES spectra of six amino acid monomers, four dipeptides, and
one tripeptide were taken. Spectra of glycine, tyrosine, histidine and phenylalanine are shown in Fig. 1. All spectra show characteristic π ∗ resonances at about
III - 92
J. Boese et al.
20
Mass Absorption Coefficient / µm−1(g/cm3)−1
Mass Absorption Coefficient / µm−1(g/cm3)−1
289 eV. These are a result of the C=O double bond of the carboxyl group. Tyrosine, histidine and phenylalanine also show strong resonances at 285 eV from
the C=C double bonds of their aromatic rings. Above the ionization edge, only
rather broad σ ∗ resonances are expected. However, most of the spectra show
three small but narrow peaks at about 293 eV, 295 eV and 297 eV. These are
probably due to a small amount of TFA that did not evaporate.
Also shown in Fig. 1 are spectra of the glycine-tyrosine and histidine-phenylalanine dipeptides. They are compared to a sum of the two monomer spectra,
added together corresponding to their molecular weight. A small effect of the
peptide bond can be seen: The π ∗ resonance resulting from the C=O bonds near
289 eV is shifted about 0.3 eV towards lower energies. This can be explained
because a neighboring atom to the carbon atoms gets changed by the peptide
bond: one oxygen atom is replaced by a nitrogen atom, which causes a shift
toward lower energies. The other peak positions of the π ∗ resonances between
285 eV and 289 eV match very well. Peak heights of some spectra are not in
perfect agreement. This is probably a result of the various effects that introduce
distortions in XANES spectra of solid samples [3].
(a)
15
Gly
10
Tyr
5
0
Dipeptide
Gly+Tyr
270
280
290
300
Energy/eV
310
12
(b)
10
His
8
6
Phe
4
2 Dipeptide
His+Phe
0
270
280
290
300
Energy/eV
310
Fig. 1. (a) Spectra of glycine (Gly) and tyrosine (Tyr) and a weighted sum (Gly+Tyr)
compared to a spectrum of the Gly-Tyr dipeptide. (b) Spectra of histidine (His) and
phenylalanine (Phe) and a weighted sum (His+Phe) compared to a spectrum of the
His-Phe dipeptide. The monomer spectra are vertically offset for clarity.
X-Ray Absorption Near-Edge Structure
4
III - 93
Conclusions
The peptide bond has only a very small effect on the π ∗ resonances in the
spectra analyzed. This suggests that protein spectra can indeed be predicted
from the amino acid spectra, which could be useful for XANES using x-ray
microscopes. Being able to predict a protein spectrum from a database of all 20
amino acids would eliminate the need to measure reference spectra of purified
protein samples, at least for preliminary experiments. Furthermore, it would
allow one to estimate whether it is possible to distinguish two different proteins
by their XANES structure. Most of the information of a protein lies in the order
of the different amino acids. XANES spectroscopy is not sensitive to this order,
nor to the folding of the protein chain. However, this method will be useful to
distinguish structure building proteins that differ in their overall amino acid
content.
Acknowledgements
We would like to thank Sue Wirick for technical assistance with the STXM
and Prof. Nicole Sampson for helpful comments. We are also grateful to Shawn
Williams, who had the original idea for this project. This work was supported by
the Office of Health and Environmental Research, U.S. Department of Energy
under contract FG02-89ER60858, by the National Science Foundation under
grants BIR-9316594 and BIR-9112062, and by the Presidential Faculty Fellow
Award RCD-9253618 (CJ). The measurements were performed at the National
Synchrotron Light Source, which is supported by the Department of Energy.
References
1. H. Ade, A. P. Smith, S. Cameron, R. Cieslinski, G. Mitchell, B. Hsiao, and
E. Rightor: Polymer 36(9), 1843–1848 (1995).
2. X. Zhang, R. Balhorn, J. Mazrimas, and J. Kirz: Journal of Structural Biology 116,
335–344 (1996).
3. X. Zhang: PhD thesis, Department of Physics, State University of New York at
Stony Brook (1995).
4. S. M. Kirtley, O. C. Mullins, J. Chen, J. van Elp, S. J. George, C. T. Chen,
T. O’Halloran, and S. P. Cramer: Biochimica et Biophysica Acta 1132, 249–254
(1992).
5. R. Ho, Z. Shao, and A. P. Somlyo: Journal of the Microscopy Society of America
2(2), 87–97 (1996).
6. B. L. Henke, E. M. Gullikson, and J. C. Davis: Atomic Data and Nuclear Data
Tables 54, 181–342 (1993).
7. A. P. Smith, T. Coffey, and H. Ade: In J. Thieme, G. Schmahl, E. Umbach, and
D. Rudolph, editors, X-ray Microscopy and Spectromicroscopy, Berlin 1996, Springer
Verlag.
This article was processed using the LATEX macro package with LLNCS style