Stereochemistry and mechanism of the conversion of 5-aminolaevulinic acid into porphobilinogen catalysed by porphobilinogen synthase

Article abstract of DOI:10.1039/b302509h

Several (3R)- and (3S)-Deuterated forms of 5-aminolaevulinic acid were synthesized. The resultant (3R)-form shows a significantly larger isotope effect when incubated with porphobilinogen synthase from bovine liver and from Bacillus subtilis. Based on the results of this study and on available crystal structures, a modified mechanism for the enzyme reaction was established.

Full text of DOI:10.1039/b302509h

View Article Online / Journal Homepage / Table of Contents for this issue
Stereochemistry and mechanism of the conversion of
5-aminolaevulinic acid into porphobilinogen catalysed by
porphobilinogen synthase†
Catherine E. Goodwin and Finian J. Leeper*
University Chemical Laboratory, Lensfield Road, Cambridge, UK CB2 1EW.
E-mail: FJL1@cam.ac.uk
Received 5th March 2003, Accepted 26th March 2003
First published as an Advance Article on the web 2nd April 2003
(3R)- and (3S)-Deuteriated forms of 5-aminolaevulinic
acid have been synthesised and the (3R)-form shows a
significantly larger isotope effect when incubated with por-
phobilinogen synthase from bovine liver and from Bacillus
subtilis; based on this and on available crystal structures, a
modified mechanism for the enzymic reaction is proposed.
All the naturally occurring tetrapyrroles (haem, chlorophylls,
vitamin B12, etc.) are biosynthetically derived from the mono-
pyrrole porphobilinogen (PBG, 2).1 PBG is made by the asym-
metric condensation of two molecules of 5-aminolaevulinic
acid (ALA, 1) catalysed by PBG synthase (PBGS, also
known as ALA dehydratase, EC 4.2.1.24).2,3 The overall reac-
tion catalysed by PBGS is shown in Scheme 1. It can be seen
that one ALA molecule provides the acetate side-chain of PBG
whereas the other ALA molecule provides the propionate side-
chain. These ALA molecules must have separate binding sites
in the active site of the enzyme, which are termed the A-site and
Scheme 2 Four possible mechanisms for the C–C bond forming step.
the P-site respectively.
Scheme 1 Reaction catalysed by PBGS.
All PBGSs studied to date form a Schiff’s base (i.e. an imine)
between a conserved lysine residue and the keto-group of an
ALA molecule. Treatment with NaBH4 reduces the imine,
which results in inactivation of the enzyme and allows identifi-
cation of the lysine residue involved as Lys-246 (E. coli number-
ing). There has, however, been some uncertainty as to whether
this Schiff’s base is formed in the A-site (as in Scheme 2a4) or
the P-site (as in Schemes 2b,5 2c6 and 2d2,7,8).
Several crystal structures of PBGSs are now available, includ-
ing ones from yeast,9 E. coli,10 and Pseudomonas aeruginosa.11
These structures show that there are in fact two conserved
lysine residues side-by-side in the active site. Very recently two
crystal structures have been obtained in which both lysine resi-
dues have formed Schiff’s bases.8,12 These are shown in Fig. 1. As
a result of these structures, it has been suggested2,8 that the
normal catalytic mechanism does in fact involve Schiff’s bases
in both the A-site and the P-site and the C–C bond forming step
involves attack of an enamine in the A-site on an imine in the
P-site (Scheme 2d).
Fig. 1 An overlay of the three-dimensional active site structures of
two different PBG synthases, both forming two Schiff’s bases with
substrate analogues. In purple is the Zn2ϩ-dependent E. coli PBGS
complexed with 4,7-dioxosebacic acid (PBD code 1I8J)12 and in green
is the Mg2ϩ-dependent Ps. aeruginosa PBGS complexed with two
molecules of 5-fluorolaevulinic acid and a Naϩ ion (PDB code 1GZG).8
The two lysine residues which form the Schiff’s bases are in front of the
substrate analogue molecules and Phe-203/Phe-214 is behind. The close
alignment of all the active site groups, apart from the differing metal
ions and their ligands, suggests that the Zn2ϩ- and Mg2ϩ-dependent
enzymes both use the same mechanism.
We have previously shown13 that a substantial isotope effect
is observed on the overall rate of reaction when [3,3-2H2]ALA
is used, but no isotope effect is observed with [5,5-2H2]ALA.
Furthermore it was deduced that the isotope effect with
[3,3-2H2]ALA was largely due to the first deprotonation at C-3
rather than the second; this was because if the second de-
protonation was rate-determining, one might expect the first
† Electronic supplementary information (ESI) available: purification
and assay protocols and plots of enzymic activity vs. substrate concen-
tration. See http://www.rsc.org/suppdata/ob/b3/b302509h/
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 4 4 3 – 1 4 4 6
1443

View Article Online
Table 1 Kinetic parameters for the deuteriated and undeuteriated ALA molecules and the derived isotope effects
Isotope effect on
Substrate
Vmax/units mgϪ1
KM/µM
(V/K)/10Ϫ3
Vmax
(V/K)
Bovine liver PBGS
ALA 1
0.683 0.023
0.571 0.014
0.328 0.015
0.337 0.009
118 15
110 11
50 13
5.77 0.63
5.17 0.42
6.63 1.47
4.66 0.51
—
—
[3S-2H]ALA 6S
[3R-2H]ALA 6R
[3,3-2H2]ALA 7
1.20 0.05
2.08 0.12
2.03 0.09
1.12 0.15
0.87 0.21
1.24 0.19
72
9
B. subtilis PBGS
ALA 1
110.9 1.8
56.7 1.0
40.4 0.4
30.5 0.6
273 15
149 10
406 17
381 20
450 17
371 26
—
—
[3S-2H]ALA 6S
[3R-2H]ALA 6R
[3,3-2H2]ALA 7
1.96 0.05
2.75 0.05
3.64 0.09
1.07 0.07
0.90 0.05
1.09 0.09
90
82
4
7
deprotonation to be reversible, resulting in release of ALA
which had lost some of its deuterium, but in fact no such
exchange could be detected. In this paper we report the syn-
thesis of the two chirally monodeuteriated substrates, (3R)-
[3-2H]ALA 6R and (3S)-[3-2H]ALA 6S as a method of probing
the stereochemistry of the deprotonation steps.
thus loss of stereochemical purity. It was important, therefore,
to develop a method to assay the stereochemical integrity of the
deuterium labelling. This was achieved, as shown in Scheme 4,
by initial reduction of ALA with NaBH4 to give, after acidic
work-up, the racemic aminomethyl-lactone hydrochloride 8.
This reduction avoids any further risk of enolisation of the
ketone. The amine was then derivatised by acylation with
(R)-Mosher’s acid to give the diastereoisomers 9 and 10 in a
1 : 1 ratio. It was not necessary to separate these diastereo-
isomers as the protons at C-3 appeared at different chemical
The route used for the synthesis of 6S is shown in Scheme 3.
The synthesis of the unsaturated lactone 4 from -glutamic acid
3 followed literature precedent.14 The double bond of 4 was
hydrogenated using deuterium gas over palladium on carbon.
The 1H NMR spectrum of the product 5 showed that the
deuterium had come entirely from the opposite face to the
bulky CH2OSitBuPh2 group. The deuterium at C-2 was largely
removed by two rounds of deprotonation using LiHMDS
1
shifts in the 500 MHz H NMR spectrum. Thus starting with
undeuteriated ALA 1 four separate signals (all dddd, two of
them overlapping) were seen in the δ 1.80 to δ 2.35 region for the
protons attached to C-3 of 9 and 10. Starting with 6S, however,
two of the four signals (δ 1.85 and δ 2.31) were present and
starting with 6R the other two signals (δ 1.94 and δ 2.29) were
present. In each case a trace of the other isomer (ca. 5%) could
just be detected. This demonstrates that the samples of 6S and
6R were very nearly stereochemically pure and no significant
amount of racemisation had occurred that would interfere with
the enzymic experiments.
1
followed by quenching the enolate with aq. NH4Cl. H NMR
spectroscopy showed 8–14% deuterium remained at C-2 after
this treatment but, as deuterium at C-2 is not expected to affect
the rate of the PBGS-catalysed reaction, no further attempt to
remove the last traces at this site was thought necessary. The
remainder of the synthesis of 6S is essentially as described in
an earlier paper from this laboratory on the synthesis of
2-fluoroALA.15 The enantiomeric 6R was synthesised in the
same way starting from -glutamic acid.
Scheme 4 Reagents: i, NaBH4 then HCl; ii, MTPA, EDCI, HOBt,
Et3N.
The samples of chirally deuteriated ALA 6S and 6R and also
the 3,3-dideuteriated ALA 713 were then tested as substrates for
bovine liver PBGS (purchased from Sigma) and Bacillus subtilis
PBGS (purified from an expressing strain of E. coli). A full
range of substrate concentrations was used, so as to get accur-
ate values for both Vmax and KM. The resulting curves are given
in the supplementary material† and the derived kinetic param-
eters are given in Table 1. It can be seen from the table that the
3R-deuteriated ALA 6R shows a significantly greater isotope
effect on Vmax than 3S-deuteriated ALA 6S. This is true for
PBGS from both sources but the difference is more pronounced
for the bovine liver enzyme, although the magnitude of the
isotope effects is smaller for this enzyme. For neither PBGS
is there any significant isotope effect on V/K for any of the
deuteriated ALA molecules.‡ This suggests that a step with a
higher energy transition state occurs before the first deproton-
ation and so molecules are already committed to proceeding on
to products by this stage.
Scheme 3 Reagents: i, HNO2; ii, BH3ؒSMe2; iii, TBDPSCl, Et3N,
DMAP; iv, LiHMDS, PhSeBr; v, H2O2; vi, D2, Pd/C; vii, LiHMDS
then aq. NH4Cl (twice); viii, TBAF; ix, Tf2O then NaN3; x, KOH;
xi, Me3SiCHN2; xii, PCC; xiii, H2, Pd/C; xiv, HCl.
The final step in the synthesis of 6S and 6R is an acid-
catalysed hydrolysis of the methyl ester (the acid is important as
ALA dimerises readily under alkaline conditions). However
acidic conditions could cause enolisation of the ketone and
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 4 4 3 – 1 4 4 6
1444

View Article Online
Scheme 5 Proposed mechanism for the PBGS-catalysed reaction.
Because the greater isotope effect is observed with 3R-
(N᎐C4–C3–C2) of close to 90Њ; (c) if the oxygen atom of
᎐
deuteriated ALA 6R, we feel it is more likely that removal of
the pro-R hydrogen atom attached to C-3 is the first deproton-
ation step in the enzymic mechanism. Armed with this inform-
ation, one can look again at the crystal structures of this
enzyme to try to determine what base is effecting the deproton-
ation and in which side of the active site.
To try to distinguish between the different mechanisms
shown in Scheme 2, we have inspected the available crystal
structures and have concluded that the mechanism of Scheme
2d is the most probable. This is because (a) there does not
appear to be any base appropriately located to remove the 3R
proton (nor the 3S proton) from an ALA molecule attached to
Lys-246; (b) furthermore, the conformation of the ALA mole-
cule bound to Lys-246 is inappropriate for the formation of a
planar double bond between C-3 and C-4, with a dihedral angle
the enolate in Scheme 2b were coordinated to the Zn2ϩ ion then
the required perpendicular attack from the double bond of the
enolate on the iminium ion with Lys-246 would not be possible;
(d) stereoelectronic reasons also eliminate the mechanism of
Scheme 2c as it is a 5-endo-tet cyclisation which is disfavoured
according to Baldwin’s rules.16 The mechanism in Scheme 2d,
however, gives an enamine attached to Lys-194 which is very
well oriented for attack on the iminium ion between ALA and
Lys-246.
For these reasons we have concentrated on the double-
Schiff’s base mechanism first proposed by Neier.7 We have
based our detailed mechanism on the conformations adopted
by the two 5-fluorolaevulinic acid molecules which form Schiff’s
bases with the two active-site lysine residues in the crystal
structure of the Mg2ϩ-dependent PBGS from Ps. aeruginosa
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 4 4 3 – 1 4 4 6
1445

View Article Online
(see Fig. 1, green).8 However the close similarity between the
active sites of the Zn2ϩ- and Mg2ϩ-dependent PBGSs (shown in
Fig. 1) leaves little doubt that both follow the same mechanism.
In the Ps. aeruginosa structure the 5-fluorolaevulinic acid mole-
cule attached to Lys-205 (equivalent of Lys-194 in E. coli) has
an ideal conformation for formation of the enamine with a
which is lost is the pro-R proton and the amino group of
Lys-246 would be ideally placed to effect this deprotonation.
We feel that the mechanism in Scheme 5 is fully consistent
with the stereoelectronic requirements of the various steps as
well as with the stereochemical and other experimental data
presented here and published previously.
dihedral angle for N᎐C4–C3–C2 of 173Њ. The nearest atom to
᎐
the 3R proton is one of the oxygen atoms of the carboxylate
group of the same ALA molecule, with an oxygen to carbon
distance of 2.77 Å. Again the dihedral angles are close to per-
fect for deprotonation by this carboxylate group—the dihedral
angle for C4–C3–C2–C1 is Ϫ88Њ (Ϫ90Њ would be ideal) and the
dihedral angle for C3–C2–C1–O is Ϫ8Њ (0Њ would be ideal).
Therefore in Scheme 5 we tentatively show this carboxylate
group acting as the base for the first deprotonation. However it
should be noted that intramolecular general base catalysis is
not usually very efficient when it involves as small a ring as
this.17 Intramolecular deprotonation of C-3 of laevulinic acid
derivatives by the carboxylate group has been observed but the
reported effective molarity is only about 0.1 M.18 Another pos-
sibility is that the deprotonation may be directly effected by a
nearby enzymic group, e.g. a hydroxide ion ligated to the zinc
ion in Zn2ϩ-dependent enzymes or the aspartate residue which
replaces the zinc in Mg2ϩ-dependent enzymes.
The full mechanism that we propose for the PBGS-catalysed
reaction is shown in Scheme 5. The zinc ion appears ideally
placed to act as a Lewis acid catalyst in the formation of the
Schiff’s base between the keto group of an ALA molecule in the
A-site and Lys-194 (see A). In contrast there does not appear to
be any acidic group well placed to catalyse attack of Lys-246 on
the keto group of an ALA molecule in the P-site. What we
propose, therefore, is that the first ALA molecule initially forms
a Schiff’s base with Lys-194 in the A-site (C) but then imine
exchange occurs and it is transferred to Lys-246 in the P-site
(E). This would be consistent with earlier pulse-labelling
results.6,19 Following this the second ALA molecule binds in the
A-site (F) and forms the Schiff’s base with Lys-194 (H). De-
protonation of C-3 of the second ALA molecule, possibly by its
own carboxylate group, then gives the enamine (I). Attack of
the enamine on the iminium ion in the P-site then occurs to
form the C–C bond (J). At this stage the second deprotonation
at C-3 might seem a possibility but it is doubtful whether a
planar C3–C4 double bond is attainable due to the position of
the carboxylate binding sites and the attachment to the two
lysine residues. What seems more likely, therefore, is that
another imine exchange reaction occurs first, forming a five-
membered ∆1-pyrroline ring (L). Now the second deproton-
ation at C-3 seems much more feasible and furthermore
Lys-194, which has just been released, would be well placed to
be the base (molecular modelling suggests that the conform-
ational change necessitated by the formation of the five-
membered ring means that the proton is not well placed to be
removed by the ALA carboxylate group, unlike in the first
deprotonation step). The resulting enamine (M) could then
expel the amino group of Lys-246 giving a pyrrolenine (N). The
final step would be loss of a proton from C-5 to generate the
pyrrole ring of PBG (O). It is known20 that the proton at C-5
Notes and references
‡ In our preliminary communication13 we said that there was an isotope
effect on V/K for the B. subtilis enzyme. However, the enzymic assays in
that work did not extend to low enough substrate concentration to get
an accurate measure of the KM value. In this work the assays were
performed with a much greater range of substrate concentrations and
the values for KM, and hence V/K, are therefore much more accurate.
1 A. R. Battersby, Nat. Prod. Rep., 2000, 17, 507.
2 P. M. Shoolingin-Jordan, P. Spencer, M. Sarwar, P. E. Erskine,
K. M. Cheung, J. B. Cooper and E. B. Norton, Biochem. Soc. Trans.,
2002, 30, 584.
3 E. K. Jaffe, Acta Crystallogr., Sect. D: Biol. Crystallogr., 2000, 56,
115; A. R. Battersby and F. J. Leeper, Top. Curr. Chem., 1998, 195,
143.
4 D. L. Nandi and D. Shemin, J. Biol. Chem., 1968, 243, 1236.
5 P. M. Jordan, in Biosynthesis of Tetrapyrroles, ed. P. M. Jordan,
Elsevier, Amsterdam, 1991, p. 1.
6 P. M. Jordan and J. S. Seehra, FEBS Lett., 1980, 114, 283;
P. M. Jordan and J. S. Seehra, J. Chem. Soc., Chem. Commun., 1980,
240.
7 R. Neier, Adv. Nitrogen Heterocycles, 1996, 2, 35.
8 F. Frère, W. D. Schubert, F. Stauffer, N. Frankenberg, R. Neier,
D. Jahn and D. W. Heinz, J. Mol. Biol., 2002, 320, 237.
9 P. T. Erskine, E. M. H. Duke, I. J. Tickle, N. M. Senior, M. J. Warren
and J. B. Cooper, Acta Crystallogr., Sect. D: Biol. Crystallogr., 2000,
56, 421; P. T. Erskine, L. Coates, R. Newbold, A. A. Brindley,
F. Stauffer, S. P. Wood, M. J. Warren, J. B. Cooper, P. M. Shoolingin-
Jordan and R. Neier, FEBS Lett., 2001, 503, 196; P. T. Erskine,
R. Newbold, A. A. Brindley, S. P. Wood, P. M. Shoolingin-Jordan,
M. J. Warren and J. B. Cooper, J. Mol. Biol., 1999, 312, 133;
P. T. Erskine, N. Senior, S. Awan, R. Lambert, G. Lewis, L. J. Tickle,
M. Sarwar, P. Spencer, P. Thomas, M. J. Warren, P. M. Shoolingin-
Jordan, S. P. Wood and J. B. Cooper, Nat. Struct. Biol., 1997, 4,
1025; P. T. Erskine, R. Newbold, J. Roper, A. Coker, M. J. Warren,
P. M. Shoolingin-Jordan, S. P. Wood and J. B. Cooper, Protein Sci.,
1999, 8, 1250.
10 P. T. Erskine, E. Norton, J. B. Cooper, R. Lambert, A. Coker,
G. Lewis, P. Spencer, M. Sarwar, S. P. Wood, M. J. Warren and
P. M. Shoolingin-Jordan, Biochemistry, 1999, 38, 4266.
11 N. Frankenberg, P. T. Erskine, J. B. Cooper, P. M. Shoolingin-
Jordan, D. Jahn and D. W. Heinz, J. Mol. Biol., 1999, 289, 591.
12 E. K. Jaffe, J. Kervinen, J. Martins, F. Stauffer, R. Neier,
A. Wlodawer and A. Zdanov, J. Biol. Chem., 2002, 277, 19792;
J. Kervinen, E. K. Jaffe, F. Stauffer, R. Neier, A. Wlodawer and
A. Zdanov, Biochemistry, 2001, 40, 8227.
13 D. Appleton and F. J. Leeper, Bioorg. Med. Chem. Lett., 1996, 6,
1191.
14 S. Hanessian and P. J. Murray, Tetrahedron, 1987, 43, 5055.
15 F. J. Leeper and M. Rock, J. Fluorine Chem., 1991, 51, 381.
16 J. E. Baldwin, J. Chem. Soc., Chem. Commun., 1976, 734.
17 A. J. Kirby, Adv. Phys. Org. Chem., 1980, 17, 183.
18 R. P. Bell and A. D. Covington, J. Chem. Soc., Perkin Trans. 2, 1975,
1343.
19 P. M. Jordan and P. N. B. Gibbs, Biochem. J., 1985, 227, 1015.
20 M. M. Abboud and M. Akhtar, J. Chem. Soc., Chem. Commun.,
1976, 1007.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 4 4 3 – 1 4 4 6
1446