X-ray structure of a putative reaction intermediate of 5-aminolaevulinic acid dehydratase.

Abstract

The X-ray structure of yeast 5-aminolaevulinic acid dehydratase, in which the catalytic site of the enzyme is complexed with a putative cyclic intermediate composed of both substrate moieties, has been solved at 0.16 nm (1.6 A) resolution. The cyclic intermediate is bound covalently to Lys(263) with the amino group of the aminomethyl side chain ligated to the active-site zinc ion in a position normally occupied by a catalytic hydroxide ion. The cyclic intermediate is catalytically competent, as shown by its turnover in the presence of added substrate to form porphobilinogen. The findings, combined with those of previous studies, are consistent with a catalytic mechanism in which the C-C bond linking both substrates in the intermediate is formed before the C-N bond.

Full text

Page 1

Biochem. J. (2003) 373, 733–738 (Printed in Great Britain)
733
X-ray structure of a putative reaction intermediate of 5-aminolaevulinic
acid dehydratase
Peter T. ERSKINE*, Leighton COATES*, Danica BUTLER*, James H. YOUELL*, Amanda A. BRINDLEY†, Steve P. WOOD*,
Martin J. WARREN†, Peter M. SHOOLINGIN-JORDAN* and Jonathan B. COOPER*1
*Division of Biochemistry and Molecular Biology, School of Biological Sciences, University of Southampton, Bassett Crescent East, Southampton, SO16 7PX, U.K., and
†School of Biological Sciences, Queen Mary, Mile End Road, London, E1 4NS, U.K.
The X-ray structure of yeast 5-aminolaevulinic acid dehydratase,
in which the catalytic site of the enzyme is complexed with a
putative cyclic intermediate composed of both substrate moieties,
has been solved at 0.16 nm (1.6 Å) resolution. The cyclic inter-
mediate is bound covalently to Lys263with the amino group of
the aminomethyl side chain ligated to the active-site zinc ion in
a position normally occupied by a catalytic hydroxide ion. The
cyclic intermediate is catalytically competent, as shown by its
turnover in the presence of added substrate to form porphobi-
linogen. The findings, combined with those of previous studies,
are consistent with a catalytic mechanism in which the C–C bond
linking both substrates in the intermediate is formed before the
C–N bond.
Key words: 5-aminolaevulinic acid dehydratase, catalytic mech-
anism, reaction intermediate, trapped intermediate, X-ray
structure.
INTRODUCTION
Theenzyme5-aminolaevulinicaciddehydratase[ALAD;porpho-
bilinogen (PBG) synthase, E.C. 4.2.1.24] catalyses an early step
in the biosynthesis of tetrapyrroles involving the condensation of
two 5-aminolaevulinic acid (ALA) molecules to form the pyrrole
PBG (see Scheme 1) [1–5]. Subsequent enzymes in the pathway
cyclizefourofthesePBGmoleculestomakethecoretetrapyrrole
framework uroporphyrinogen III, which undergoes various modi-
fications to form a variety of essential metallo-prosthetic groups,
including haem, chlorophyll and the cobalamins. ALADs share a
high degree of sequence identity, contain about 350 amino acids
per subunit and are usually octameric. In humans, hereditary defi-
ciencies in ALAD give rise to the rare disease Doss porphyria [6],
and the exquisite sensitivity of the human enzyme to inhibition
by lead ions is a major factor in acute lead poisoning [7,8].
The structures of ALADs from several species have been deter-
mined, and several inhibitor complexes have been studied [9–12].
In these structures the enzyme is a homo-octamer with each
subunit adopting a (β/α)8 or TIM (triosephosphate isomerase)
barrel fold with an N-terminal arm which, in yeast ALAD, is 39
residues in length (Figure 1). Within the octamer, subunits form
dimers in which each subunit has its N-terminal arm wrapped
around the barrel of the other monomer. The active site of each
subunit is located in a pronounced cavity formed by loops at the
C-terminalendsoftheβ-strandsformingtheTIMbarrel.Alleight
active sites are oriented towards the outer surface of the octamer
and are independent.
Atthebaseofeachactivesitearetwolysineresidues(Lys210and
Lys263in yeast ALAD), one of which (Lys263) is known to form
a Schiff-base link to the first molecule of substrate that binds
to the active site [13,14]. The substrate bound to Lys263is in-
corporated into the propionic acid side, or P-side, of the product
PBG whereas the second substrate molecule forms the acetic
acid side, or A-side, of PBG (see Scheme 1). Thus catalysis pro-
ceeds by formation of a Schiff-base link at the P-site between the
Abbreviations used: ALA, 5-aminolaevulinic acid; ALAD, 5-aminolaevulinic acid dehydratase; PBG, porphobilinogen; TLS, translation-libration-screw.
1To whom correspondence should be addressed (e-mail J.B.Cooper@soton.ac.uk).
4-oxo group of substrate ALA and an invariant lysine residue,
equivalent to Lys263in yeast ALAD. In many ALADs, an active-
sitezincionhasbeenimplicatedinthecatalyticmechanismandis
thought to act on substrate bound to the A-site. The zinc-binding
site is formed by the following residues in yeast ALAD: Cys133,
Cys135andCys143.Incontrast,theplantandsomebacterialALADs
require magnesium and lack the triple-cysteine sequence motif,
whichisreplacedbytwoaspartateresiduesandanalanineresidue
[15]. A more detailed characterization of the metal requirement
of ALADs is given in [5].
Previously, the structures ALADs in which the P-site is occu-
pied by a number of bound inhibitors have been analysed, and
the structure of substrate ALA bound in this site has also been
determined[16].AllligandsformaSchiffbasewithLys263andare
held with their carboxy groups making hydrogen bonds with the
conserved residues Ser290and Tyr329. Hydrophobic interactions
are made with the side chains of residues Tyr216, Phe219, Tyr287
and Val289, which effectively form four sides of this substrate-
binding subsite. In yeast ALAD, ALA appears to bind in two
conformations, but in both of these conformers the C-5 amino
group interacts with the side chains of Asp131and Ser179. This
implicates these two residues as having potential catalytic roles
in the reaction mechanism. The A-site is a solvent-filled cavity
bounded on one side by the P-site ALA molecule and on the other
by two arginine residues (Arg220and Arg232) which are strongly
conserved. The base of the A-site is formed by the zinc ion held
by three cysteine residues. There is a solvent molecule bound to
the zinc ion which is within H-bonding distance of the amino
group of P-side ALA. This water molecule may be a zinc-bound
hydroxide which abstracts a proton from C-3 of A-side ALA as
a prerequisite for formation of the C–C bond between the two
substrates.
Hitherto, mechanistic proposals for ALAD have been do-
minated by the experimental finding that the enzyme only forms
a Schiff base with one of the two identical substrate mole-
cules. This was based on labelling studies which showed only
c ?2003 Biochemical Society

Page 2

734
P. T. Erskine and others
Scheme 1 Reaction catalysed by ALAD
Figure1
siteLys263alongwiththeboundintermediateinball-and-stickrepresentation
TertiarystructureofamonomerofyeastALADshowingtheactive-
The zinc ion is also shown (the larger, separate, black sphere). Residues 213–233 over the
active site are omitted in order to show the position of the putative intermediate more clearly.
the P-side substrate becoming covalently trapped upon reduction
with NaBH4 [13,14]. Recently, the structures of the yeast and
Escherichia coli enzymes complexed with the irreversible inhibi-
tor 4,7-dioxosebacic acid have been solved [17,18]. Intriguingly
this inhibitor forms two Schiff bases at the active site, involving
Lys210as well as Lys263. One half of the molecule resides in the P-
site and the other half fills the putative A-site with its carboxy
group interacting with invariant arginine residues Arg220and
Arg232The structure of Pseudomonas aeruginosa ALAD com-
plexed with 5-fluorolaevulinic acid [19] shows that both A- and
P-sites are occupied by an inhibitor molecule covalently bound
through a Schiff base with Lys210and Lys263respectively. The
structural evidence that both invariant lysine residues may form
Schiff bases with substrate at the active site is suggestive of a
double-Schiff-base mechanism such as that proposed by Neier
[20].
Previously,wefoundthatALAwascovalentlyboundasaSchiff
base to Lys263at the P-site of yeast ALAD, despite the fact that
ALA had not been added to the enzyme [16]. It has also been
observed that human ALAD, highly purified from erythrocytes,
possesses a ligand resembling PBG at the active site, despite
the fact that no ligands had been externally added [12]. In the
present study we have found that crystallizing the purified yeast
dehydratase in the presence of added 5-aminolaevulinic acid
results in the trapping of what appears to be a similar cyclic inter-
mediatecomplex,arisingfromthereactionoftwosubstratemole-
culesattheactivesite.Theintermediateiscatalyticallycompetent
and can turn over to form the product, PBG. The complex, re-
finedat0.16 nm(1.6 Å)resolution,showsacovalentlinkbetween
Lys263andtheintermediate,adirectlinkwiththezincionandpro-
vides important stereochemical information giving an unpreced-
ented view of active-site interactions. A catalytic mechanism
takingintoaccountthestructureoftheenzyme–intermediatecom-
plex is proposed.
EXPERIMENTAL
Recombinant yeast ALAD was purified as described previously
[9]. Crystals of yeast ALAD with the bound intermediate were
obtained by co-crystallizing the enzyme in the presence of
substrate (ALA). The hanging-drop vapour-diffusion method was
used with ALA present at a concentration of 0.2 mM and enzyme
present at a concentration of 2 mg/ml. Apart from the presence of
ALA, the crystallization conditions [2–10% (w/v) poly(ethylene
glycol) of molecular mass 6000 Da/200 mM Tris/HCl, pH range
7.0 – 8.0] are otherwise identical with those used to crystallize
the native enzyme [21]. Crystals of the complex, which ap-
peared within 3–4 weeks, belong to space group I422 with
unit cell parameters of a=b=10.32 nm (103.2 Å), c=16.7 nm
(167.7 Å). The crystals were flash-cooled in liquid ethane and
X-ray data were collected at the European Synchrotron Radiation
Facility(ESRF;Grenoble,France)usingtheID-29beamlinewith
the crystal maintained at a temperature of 100 K using an Oxford
Cryosystems (Long Hanborough, Oxford, U.K.) cooler. The data
were processed using MOSFLM [22] and the CCP4 suite [23].
The structure was refined using SHELX [24] and the program
RESTRAIN [25] was used for refinement of translation-libration-
screw (TLS) tensors. Graphical rebuilding was done using
TURBO-FRODO (Bio-Graphics, Marseille, France) running on
Silicon-Graphics(SGI)computers.Theco-ordinatesandstructure
factors have been deposited in the Protein Data Bank with
accession codes 1ohl and r1ohlsf respectively.
RESULTS AND DISCUSSION
Interpretation of the electron-density difference map
The initial difference map for yeast ALAD co-crystallized with
ALA showed strong electron density for a five-membered ring
c ?2003 Biochemical Society

Page 3

Putative intermediate of 5-aminolaevulinic acid dehydratase
735
Figure 2Stereo view of the electron-density map of the putative intermediate (labelled *) covalently bound to Lys263at 0.16 nm (1.6 A ˚) resolution
The map was weighted as described by Read [32], and is contoured at 1.0 root mean square. The P-site is on the right-hand side and the A-site is formed by residues shown on the left-hand side
including the zinc ion. The other invariant lysine (Lys210) is in the foreground just to the left of Lys263, but it has been omitted for clarity. The side chains of the two lysine residues lie approximately
parallel with each other.
intermediate resembling PBG bound to Lys263in the active site.
Theligandwasbuiltintothedensityaccordinglyandthestructure
of the complex refined for several cycles after each round of
manual rebuilding. The refined electron-density maps at 0.16 nm
(1.6 Å) resolution established that the carbon atoms of the five-
memberedringtowhichtheaceticacidandpropionicacidgroups
are attached were trigonal. The final electron-density map is
shown in Figure 2. Refinement of the occupancy of the ligand
yielded a value of 81% confirming that the putative intermediate
hasbeentrappedinthemajorityoftheenzymemoleculesforming
the crystal. The atoms of the bound intermediate have a mean iso-
tropic B-factor (Biso) of 0.4 nm2(40.0 Å2), which is comparable
with the B-factors of the entire complex [mean Biso=0.383 nm2
(38.3 Å2)]. This presumably reflects on the buried state of the
ligand,whichiscoveredbytheenzyme’sactive-siteflap(residues
215–235).
The final structure, which has been refined using data between
4.5 nm (45 Å) and 0.16 nm (1.6 Å), has an R-factor of 18.8%
and an R-free of 24.1% (see Table 1). Refinement of the TLS
tensors for the TIM barrel and arm domains gave a significant
improvementinR-freeofover7%.Thestructurehas90.8%ofits
residues within the ‘most favoured’ regions of the Ramachandran
plotbythePROCHECKcriteria[26]and8.6%ofresidueswithin
theso-called‘additionalallowed’boundary.Noresiduesareinthe
disallowed regions. The temperature factors are reasonable, ex-
cept for a few residues in the active-site flap (most notably 228–
230), where the electron density remains poor despite extensive
effortstorebuildtheseresiduesthroughouttherefinementprocess.
Table 1
intermediate after TLS refinement
X-ray statistics for the complex of yeast ALAD with the putative
Abbreviation: RMSD, root-mean-square deviation.
ParameterValue
Statistics for entire dataset
Resolution range [nm (A˚)]
Rmerge(%)
Completeness (%)
Multiplicity
Mean I/σ(I)
%I >3σ(I)
Statistics for outer shell
Resolution range [nm (A˚)]
Rmerge(%)
Completeness (%)
Multiplicity
Mean I/σ(I)
Refinement
R-factor (%)
R-free (%)
Number of reflections
RMSD bond lengths [nm (A˚)]
RMSD 1–3 distances [nm (A˚)]
RMSD bumps [nm (A˚)]
RMSD chiral tetrahedra [nm (A˚)]
RMSD main-chain planes [nm (A˚)]
RMSD side-chain planes [nm (A˚)]
4.47–0.16 (44.7–1.6)
5.5
96.2
7.2
21.7
77.5
0.17–0.16 (1.7–1.6)
38.8
96.2
2.6
2.5
18.8
24.1
57 545
0.0005 (0.005)
0.0013 (0.013)
0.0015 (0.015)
0.0009 (0.009)
0.0009 (0.009)
0.0007 (0.007)
c ?2003 Biochemical Society

Page 4

736
P. T. Erskine and others
Nonetheless the electron density for the active-site loop in this
complex is substantially better than that of the native enzyme.
Interactions made by the putative intermediate
The active site of ALAD is located in a large cavity at the C-
terminal end of the eight-stranded all-parallel β-barrel (Figure 1).
It involves two invariant lysine residues (Lys263and Lys210in
yeast ALAD) as well as a zinc ion which is co-ordinated by
three cysteine residues (Cys133, Cys135and Cys143) in the yeast
enzyme.
The trapped intermediate is held by a covalent bond to the
invariant lysine residue in the P-site, namely Lys263(Figure 2).
The P-side of the intermediate is surrounded by the hydrophobic
side chains of Tyr216, Phe219, Tyr287and Val289. The P-side carboxy
group forms hydrogen bonds with the side chains of Ser290and
Tyr329and the P-side amino group is close to the side chains of
Ser179and Asp131. The side chains of Phe89and Tyr207are also
involved in forming this binding site. All of the above residues
areinvariantorstronglyconserved,implyinganimportantrolefor
theminsubstratebindingtoALADsfromallspecies.Theputative
intermediate defined in this analysis makes essentially the same
interactions on its P-side as a number of inhibitors which occupy
the P-site only [10,11,16,27]. The analysis confirmed that the
intermediate is not bound covalently to the other active-site lysine
residue(Lys210),eventhoughthishasbeenshowntomakeaSchiff
base in complexes with inhibitors that occupy the A-site [17–19].
In previous studies of inhibitor and substrate binding at the
P-site alone [10,11,16,27], the location of the A-site of ALAD
was inferred from the presence of a solvent-filled cavity adjacent
to the P-site ligand. This cavity is formed by the zinc ion and a
number of invariant residues, including Arg220, Arg232and Gln236.
These predictions are largely confirmed by the currently deter-
mined structure in which it was found that the arginine side
chains make electrostatic interactions with the A-side carboxy
group(Figure2).Thiswasalsofoundintheboundstructureofthe
inhibitor4,7-dioxosebacicacid,whichoccupiestheA-andP-sites
[17–18]. However, in the structure of the putative intermediate
there is an additional interaction between the A-side carboxyl
group and the side-chain amino group of Lys210– one of the two
invariant catalytic lysine residues. It was speculated that the zinc
ion plays an important mechanistic role by interacting with the
A-side ALA via a solvent molecule which is bound datively to
the metal ion. The requirement for a base in the mechanism
to abstract hydrogen atoms from the C-3 of A-side ALA suggests
that the zinc-bound solvent molecule is a hydroxide ion. In the
presentlydeterminedcomplex,theputativehydroxideionappears
to be substituted by the A-side amino group of the trapped inter-
mediate, which binds datively to the metal ion.
Implications for the mechanism
Various mechanisms for ALAD based on the X-ray structures of
bound ligands have been proposed (e.g. [10]). These relied on the
experimental finding that only one of the substrate moieties (P-
side ALA) forms a Schiff base with one of the active-site lysines
(Lys263in yeast ALAD) [13,14]. However, the recent findings
that A-side ALA may also bind by forming a Schiff base with
Lys210[17–19] suggest that proposals for the mechanism need to
be reconsidered accordingly.
Another major issue of discussion with regard to the ALAD
mechanism has been the order in which the bonds linking the two
ALA molecules are formed, i.e. whether C–C bond formation
occurs before or after C–N bond formation. The structure of the
putative intermediate reported here indicates that A-side ALA
must be in contact with the putative zinc-bound hydroxide, a
strongly basic group which could act on the C-3 carbon of the A-
side substrate. The formation of a Schiff base between this ALA
moiety and Lys210would further help to labilize the C-3 hydrogen
atomsofA-sideALAandassistdeprotonationofthiscarbonatom,
yielding a bound enamine (see Scheme 2). This could lead to
nucleophilic attack of the enamine at the A-site on the C-4 of
P-side ALA, thus forming the C–C bond linking the substrates.
The inter-substrate C–N bond then forms followed by the second
deprotonation of the C-3 of A-side ALA, thereby yielding
the intermediate found in the crystal structure (labelled ‘*’ in
Scheme 2). Although it is possible that the C–N bond could
form first, thus yielding a Schiff base linking both substrates,
this would imply that the Schiff base between A-side ALA and
Lys210observed in recent studies [17–19] has no role in catalysis
otherthanperhapstoanchortheA-sidesubstratepriortocatalysis.
If instead it is assumed that the Schiff base linking A-side ALA
with Lys210has a catalytic function (rather than just a passive
binding role), then it is likely that inter-substrate C–C bond for-
mation would occur before the C–N bond formation, as shown in
Scheme 2.
In the catalytic reaction, the putative zinc-bound hydroxide is
likely to deprotonate the C-3 of A-side ALA [shown as base B(1)
in Scheme 2]. The fact that this hydroxide is apparently displaced
by the C-5 amino group of A-side ALA, which binds to the zinc
instead,maycontributetothestabilityofthetrappedintermediate.
Previous NMR studies using isotopically labelled PBG have
suggested that the amino group of the reaction product binds
to the active-site zinc ion [28]. Two invariant residues, Asp131and
Ser179, which are likely to catalyse various proton transfers in the
reaction, are shown in Figure 2 below the putative intermediate.
Two potential catalytic roles for these residues [shown as bases
B(2) and B(3) in Scheme 2] could include deprotonation of the
P-side C-5 amino group and proton transfer to the amino group
of Lys210. The fact that the side chains of Lys210and Lys263are
adjacent would allow one to transfer a proton to the other during
the reaction. For example, the base [labelled B(4) in Scheme 2]
which protonates Lys263to facilitate breakdown of the putative
intermediate could be Lys210.
The apparent stability of the postulated intermediate, under
conditions where no particular attempts have been made to trap
it, begs the question of how it is stabilized in the crystal structure.
Although it is possible that crystal packing contributes in some
way to its stabilization, we believe that the intermediate may
actually represent a bona fide step in the catalytic cycle that
is awaiting the completion of the final stages of the catalytic
mechanism (last three stages in Scheme 2). This premise is sup-
ported by preliminary observations that enzyme, allowed to turn
over in the presence of [4-14C]ALA, results in
enzyme in which two substrate-equivalents are bound per subunit
[29]. Addition of unlabelled ALA to the labelled enzyme results
in the release of [14C]PBG and generates unlabelled enzyme,
presumably with the same (although unlabelled) intermediate at
the active site. This is further supported by the observation from
MS (N. Mills-Davies, D. Butler and P. M. Shoolingin-Jordan,
unpublishedwork)thatthehumanenzymeformsastablecomplex
insolutionwithamoleculeofmassclosetothatofenzyme-bound
PBG. In all cases, the enzyme intermediate complex is Ehrlich’s-
reagent-negative[30],indicatingthattheboundintermediatedoes
nothavethesamechemicalpropertiesasPBG.TheX-raycrystal-
lography and biochemical studies are thus mutually rein-
forcing.
A further interesting observation arising from the X-ray ana-
lysis of the enzyme–ligand complex is that the high resolution of
14C-labelled
c ?2003 Biochemical Society

Page 5

Putative intermediate of 5-aminolaevulinic acid dehydratase
737
Scheme 2 Proposed mechanism for ALAD
It is assumed that the Schiff base made by Lys210with A-side ALA facilitates proton abstraction at its C-3 position. This leads to nucleophilic attack on P-side ALA forming the first C–C bond linking
the substrates. The inter-substrate C–N bond then forms, yielding the putative intermediate found in the crystal structure (labelled *). Base (1)B is likely to be the putative zinc-bound hydroxide.
the structure allows the stereochemistry of the only chiral centre
in the intermediate, the 3-position which is attached to Lys263,
to be assigned as S. Following the cleavage of the bond between
Lys263,thesamelysineresidueisintheoptimalpositiontoremove
stereospecifically the proton from the 2-position (that becomes
the free a-position) [31] and to complete the enzymic synthesis.
The precise mechanism by which added substrate facilitates these
last stages that culminate in the transformation of the bound
intermediate into PBG is currently under detailed investigation.
We gratefully acknowledge the ESRF for access to synchrotron beam time and associated
travelfunds.WethankProfessorMuhummadAkhtar,FRS,andDrM.Montgomery(bothof
the University of Southampton) for many useful discussions. We thank the Biotechnology
and Biological Sciences Research Council (BBSRC) and the Engineering and Physical
Sciences Research Council (EPSRC) for financial support.
REFERENCES
1 Jordan, P. M. (1991) Biosynthesis of tetrapyrroles. New Compr. Biochem. 19, 1–65
c ?2003 Biochemical Society