Opportunities for Probing the Structure and Mechanism of Porphobilinogen Synthase by Raman Spectroscopy

Full text of DOI:10.1021/ja971357e

11556
J. Am. Chem. Soc. 1997, 119, 11556-11557
Opportunities for Probing the Structure and
Mechanism of Porphobilinogen Synthase by Raman
Spectroscopy
†
†
John Clarkson, Eileen K. Jaffe, Robert M. Petrovich,
Jian Dong, and Paul R. Carey*
Department of Biochemistry
Case Western ReserVe UniVersity
Figure 1. The PBGS-catalyzed asymmetric condensation of two
molecules of ALA to form porphobilinogen. C labeled atoms are
marked 2.
1
3
1
0900 Euclid AVenue, CleVeland, Ohio 44106-4935
Institute for Cancer Research, Fox Chase Cancer Center
7
701 Burholme AVenue, Philadelphia, PennsylVania 19111
ReceiVed April 29, 1997
Although Raman spectroscopy is capable of providing
1
molecular detail on protein ligand contacts and on enzyme-
substrate intermediates,2 its utility has been severely limited
by problems associated with low sensitivity and spectral
interference from luminescent chromophores. The latter may
be intrinsic to the system under study or present as trace
,3
4
impurities. Due to technical innovations these problems have
been largely removed. We show here the first Raman difference
spectroscopic data for product bound to the enzyme porpho-
bilinogen synthase (PBGS); it is possible to detect changes in
the chemistry of the product molecule upon binding to the active
site and, in addition, to observe many Raman features which
occur due to amino acid side chains being perturbed by product
binding. The present data demonstrate the potential for obtain-
ing Raman spectra for even recalcitrant systems. PBGS is a
large enzyme, an octamer of total molecular mass ∼300 kDa
involved in the tetrapyrrole biosynthetic pathway.5 It has an
intense fluorescence emission, due to impurities which may be
oligopyrroles or their fragments bound to the enzyme in minute
amounts, when excited by laser wavelengths in the visible
spectrum. However, here we show that high-quality Raman
data can be acquired for PBGS complexes using deep red
excitation at 752 nm.
Porphobilinogen synthase is a metalloenzyme that catalyzes
the asymmetric condensation of two molecules of 5-amino-
levulinate (ALA) to form porphobilinogen as illustrated in
6
Figure 1. This reaction is common to all tetrapyrrole biosyn-
Figure 2. The Raman difference spectra of: A, product labeled with
C at the 3 and 5 positions bound to PBGS; B, bound, unlabeled
product complex; and C, unlabeled free product.
13
theses, e.g. for porphyrin, chlorophyll, and vitamin B12, and is
essential for cellular life. The PBGS octamer contains four
active sites, each of which binds two molecules of ALA that
have different chemical fates. Although many details of the
reaction mechanism are not well established, it is known that
there is a Schiff base formed between a universally conserved
lysine and one of the two ALA molecules at the active site. In
this study, we have reacted ALA with PBGS from Bradyrhizo-
bium japonicum in stoichiometric amounts and examined the
resultant product complex by Raman difference spectroscopy.
By running the reaction with smaller amounts of enzyme
followed by ultrafiltration to remove the protein, we were able
to generate the Raman spectra of free, unbound product for
comparison. Spectral assignments for the product were sup-
1
3
ported by also using [4- C]ALA as substrate that gave rise to
1
3
C at the 3 and 5 positions of the product (Figure 1).
The Raman difference spectra of porphobilinogen, labeled
5
and unlabeled, bound to PBGS are compared in Figure 2 to the
spectrum of the free unlabeled product. The difference spectra
were obtained by undertaking a computer subtraction of the
spectra: (enzyme-bound product in buffer) - (enzyme in buffer
at the same concentration). For the difference spectrum of
“free”, a spectrum of buffer was subtracted from that of the
product in buffer. Each set of spectral data was acquired in
approximately 10 min using 1 W of 752 nm Kr+ excitation.
Under these conditions the features seen in the difference spectra
7
†
Fox Chase Cancer Center.
To whom correspondence should be addressed: telephone, 216-368-
are highly reproducible. In the spectrum of the free product
*
-1
(Figure 2, bottom), the intense features at 1524 and 1416 cm
0
031; fax, 216-368-4544; e-mail, carey@biochemistry.cwru.edu.
1) Callender, R.; Deng, H. Annu. ReV. Biophys. Biomol. Struct. 1994,
are assigned to pyrrole ring stretching modes on the basis of
quantum mechanical calculations (J.D. and P.R.C., unpublished
work). For the bound product, the equivalent features can be
(
2
3, 215-245.
(
(
(
2) Tonge, P. J.; Carey, P. R. AdV. Spectrosc. 1993, 20, 129-161.
3) Carey, P. R.; Tonge, P. J. Acc. Chem. Res. 1995, 28, 8-13.
4) Kim, M.; Owen, H.; Carey, P. R. Appl. Spectrosc. 1993, 47, 1780-
-
1
seen at 1516 and 1412 cm , unlabeled, and 1497 and 1403
-
1
13
1
783.
cm for enzyme-bound [3,5- C]porphobilinogen with the
(
(
(
5) Jaffe, E. K. J. Bioenerg. Biomembr. 1995, 27, 169-179.
-1
reduction in wavenumbers of 19 and 9 cm supporting the
6) Shemin, D.; Russell, C. S. J. Am. Chem. Soc. 1953, 75, 4873-4874.
assignments to pyrrole ring modes. The downshifts in the ring
mode frequencies upon binding are evidence for a significant
7) Petrovich, R. M.; Litwin, S.; Jaffe, E. K. J. Biol. Chem. 1996, 271,
8
692-8699.
S0002-7863(97)01357-7 CCC: $14.00 © 1997 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 119, No. 47, 1997 11557
change of electron distribution in the pyrrole ring upon active
site binding. Similar shifts are seen in modes for rings which
are part of highly polarized π-electron systems such as thien-
ylacryloyl acyl groups bound in the active sites of cysteine
proteases8 and, in the present case, could be caused by
propinquity to charged groups or dipoles. The third major
feature for the free product occurs at 942 cm-1 and may involve
set and the 752 nm excited Raman spectrum of PBGS in hand
(data not shown), we tentatively assign the features near 1620,
-1
1275, 1213, and 840 cm to the phenol ring of tyrosine, which
indicates that one or more tyrosines are perturbed by product
-1
binding. Similarly, inflections in the baseline near 1005 cm
imply phenylalanine side chain(s) perturbation. PBGS contains
seven phyllogenetically conserved tyrosines and six such
-
5
-1
a contribution from the C-C stretch from the C-COO groups
phenylalanines. The negative peak near 1441 cm is intrigu-
ing. In some inorganic compounds, the symmetric stretch from
(
J,D, and P.R.C. unpublished work). The feature appears to
-
1
-
shift to near 955 cm in the spectra of the bound products,
without undergoing a significant isotope shift with 3,5- C
substitution. The role of the -COO groups in the 942 cm
mode can be put on a firmer footing by using isotopic sub-
stitutions in one or more of the atoms involved.
a -COO group bound to a cation has been assigned to this
region. Thus, the negative feature in Figure 2 may mean that
a -COO side chain on the protein is being converted to
-COOH when the product binds since it implies that a putative
-COO symmetric stretch (from an amino acid side chain)
disappears in the complex. This may be a result of the catalytic
removal of one of the four protons which are lost from substrate
and formally produce water. Interpretation of this kind is still
in its infancy, and we expect it to become better defined as the
number of experiments on families of related proteins increases.
However, it is apparent that a wealth of molecular detail will
soon become available from systems that have until now been
1
3
9
-
-1
-
-
9
B. japonicum PBGS uses a catalytic magnesium ion in
contrast to other PBGSs which use zinc. 13C NMR studies show
high similarity between product bound to different zinc utilizing
5
PBGSs but very different chemical shifts for PBG bound to B.
japonicum PBGS (R.M.P and E.K.J., unpublished results).
Comparative Raman studies promise to assist in the interpreta-
tion of these chemical shift differences.
1
0
inaccessible.
It is clear that the difference spectra for the enzyme-product
complex contain a plethora of features which do not originate
from the product. Instead, they are due to protein modes which
emanate from parts of the protein which are perturbed by product
binding. Since the side chains of aromatic amino acids are
strong Raman scatterers, features from these are expected if the
environments of these chains change. With the convention we
are using to undertake Raman difference spectroscopy, an
increase in intensity of a Raman band in the product complex
will result in a positive feature in the difference trace, while a
decease will result in a negative (below the baseline) peak. A
small shift in frequency gives rise to a derivative-type peak.
We have recorded the Raman spectra of the aromatic amino
acids using 752 nm excitation, since we are unaware of spectra
in the literature recorded at this wavelength. With this standard
Acknowledgment. This research was supported by NIH Grants
ES03654 (to E.K.J.), CA09035 (to Fox Chase Cancer Center) and
CA06927 (to Institute for Cancer Research).
JA971357E
(
9) Nakamoto, K. Infrared and Raman Spectra of Inorganic and
Coordination Compounds; Wiley: New York, 1986; pp 233.
(10) Experimental conditions: 4 Raman spectra were recorded with
+
equipment described by Kim et al., using 1 W, 752 nm Kr excitation and
collecting each data set for 10 min. Concentrations are (A) 1.7 mM, (B)
1
.7 mM, and (C) 21 mM. B. japonicum PBGS was purified according to
7 13
the method of Petrovich et al. [3,5- C]PBG was prepared in a 0.5 mL
13
buffer using 10 µmol of [4- C]ALA-HCl and 1.5 mg of E. coli PBGS in
2
2
0 mM Kpi (pH 7.0), 10 µM ZnCl2,, 1 mM MgCl2, and 10 mM
-mercaptoethanol. KOH (10 µmol) served to neutralize the ALA-HCl. The
reaction was monitored by periodic removal of aliquots and workup with
13
Ehrlich’s reagent. After 3 h the [3,5- C]porphobilinogen was centrifuged
through a 10 kDa ultrafiltration membrane to remove the protein.
(8) Doran, J. D.; Carey, P. R. Biochemistry 1996, 35, 12495-12502.