Structure-Function Relationships in Serine Hydroxymethyltransferase

A study of enzymes is central to an understanding of biological function. The
binding of substrate(s) to an enzyme and its fit at the active site facilitates a multitude of
chemical reactions. The mechanisms of catalysis include general acid-base, covalent and
metal ion catalysis. The study of structure-function of enzymes has been central to the
elucidation of catalytic mechanism of biochemical reactions. In addition to the factors
mentioned above, coenzymes which are vitamin derivatives have provided several new
insights into biology.
The versatility of vitamin B6 and the wide distribution of pyridoxal 5’-phosphate
(PLP)-dependent enzymes are reflected in the observation that 4% of all catalytic reactions
involve this co-enzyme. The availability of biochemical and structural information on more
than 140 PLP-enzymes gives a good handle to understand the organization of PLP-enzymes
in detail. This information has helped in elucidating the diverse reaction mechanisms of
PLP-enzymes. SHMT, one of the PLP-enzymes, the subject of this study belongs to the
α-family. It links amino acid and nucleotide metabolism. Serine hydroxymethyltransferase
(SHMT) catalyzes THF-dependent hydroxymethyltransfer from L-Ser to tetrahydrofolate
(THF) to yield 5, 10-CH2 THF and Gly. This reaction provides one-carbon fragments for a
wide variety of end products in mammalian systems. SHMT being a part of thymidylate
cycle, suggested that it could be an alternative target for cancer chemotherapy. SHMT, in
addition to L-Ser and Gly inter-conversion, catalyzes THF-independent cleavage of
3-hydroxy amino acids. L-Thr/L-allo Thr cleavage by SHMT results in production of
acetaldehyde which is a major flavoring compound in production of fermented dairy
products. Hence SHMT can be used as a starter culture in the manufacture of dairy
products. SHMT can also be used as a biocatalyst in the synthesis of β-hydroxy-α-amino
acid derivatives.
2
The objectives of the present investigation are: biochemical characterization of
selected residues involved in THF-dependent and -independent reactions; crystallization of
the mutant enzymes with their substrate(s)/inhibitor complexes to understand the role of
these residues; probe the retroaldol and direct displacement mechanisms for the THFdependent
L-Ser cleavage; establish the mechanism of THF-independent cleavage of
3-hydroxy amino acids by mutation of specific amino acid residues; and study the
interaction of chemical inhibitors and compounds from natural sources to understand the
role of SHMT in cancer. With these objectives the present investigation was undertaken and
results and conclusions are presented in the form of thesis entitled “Structure – Function
Relationships in Serine Hydroxymethyltransferase”.
The present study is divided into four chapters:
Chapter 1: The role of lysine226 in the reaction catalyzed by bsSHMT
Lys residue at the active site of PLP-enzymes in addition to anchoring PLP
functions as a proton acceptor or a donor in catalysis. It has been proposed that Lys 229
in eSHMT is crucial for product expulsion, which is a rate determining step of catalysis.
Lys 226 of bsSHMT was mutated to Met and Gln, overexpressed and the mutant enzymes
were purified. The mutant enzymes contained 1 mol of PLP per mol of subunit suggesting
that Schiff’s base formation with Lys was not essential for PLP binding. K226M and
K226Q bsSHMT were inactive for THF-dependent cleavage of L-Ser. However, cleavage
of L-allo Thr and transamination reaction was not abolished completely. K226M bsSHMT
had distinct absorbance maximum at 412 nm and K226Q bsSHMT was similar to that of
bsSHMT. The crystal structure of K226M bsSHMT revealed that PLP was bound at the
active site in an orientation different (16°) from that of the wild-type enzyme. The absence
of reaction intermediate at 388 nm on interaction with methoxyamine (MA) corroborates
the suggestion that the active site of K226M bsSHMT was different from that of bsSHMT.
Both the Lys mutants were capable of forming an external aldimine; this was also supported
by the crystal structure of K226M-Ser/Gly complexes. Spectral studies show the formation
of a small amount of quinonoid intermediate on addition of Gly and THF/FTHF to K226M
bsSHMT. However, stopped-flow studies suggested enhanced quinonoid intermediate
formed on addition of THF was affected drastically. In SHMT, formation of an external
3
aldimine is accompanied by the change in orientation of PLP by 25°. The orientation of
PLP in the external aldimine form (25°) changes to 16° during quinonoid intermediate
formation. The quinonoid intermediate is stabilized by interactions of Lys at the active site.
In the absence of the -NH2 group of Lys, the conversion of the external aldimine to product
quinonoid intermediate, that is, the change in orientation of PLP from 25° to 16°, may not
be possible. This in turn could lead to shifting of equilibrium towards the substrate external
aldimine form (L-Ser form).
These results show that Lys 226 is responsible for flipping of PLP from one
orientation to another, which is accompanied by Cα-Cβ bond cleavage. This flip is important
in the THF-mediated enhanced Cα proton abstraction from Gly in the reverse reaction.
Chapter 2: The involvement of glutamate 53 in binding of L-serine and folate, and
conversion of bsSHMT from ‘open’ to ‘closed’ form
An examination of the crystal structure of bsSHMT binary and ternary complexes
suggested that E53 interacts with L-Ser and FTHF. Glu 53 was mutated to Gln and
structural and biochemical studies were carried out to examine the role of this residue in
catalysis. The mutant enzyme was completely inactive for THF-dependent cleavage of
L-Ser, whereas there was a 1.5-fold increase in the rate of THF-independent reaction with
L-allo Thr. Spectral studies showed that E53Q bsSHMT had absorbance maximum at 425
nm and the addition of L-Ser/Gly resulted in formation of an external aldimine. The crystal
structure of E53Q bsSHMT was similar to that of the wild-type enzyme, except for
significant changes at Q53, Y60 and Y61. E53Q bsSHMT binary complex with L-Ser or
Gly showed that the side chain of L-Ser and carboxyl of Gly were in two conformations in
the respective external aldimine structures. The loss in characteristic decrease in molar
ellipticity on addition of L-Ser and loss of enhanced thermal stability suggested that E53Q
bsSHMT was unable to undergo a conformational change from ‘open’ to ‘closed’ form in
which THF-dependent reaction occurs. Addition of THF/FTHF to E53Q bsSHMT-Gly
complex showed the formation of a quinonoid intermediate. Stopped-flow studies were
performed to obtain rate constants for the formation of quinonoid intermediate with mutant
and wild-type enzymes. However the quinonoid intermediate formed by the mutant enzyme
4
was unstable. Dialysis experiments and dissociation constants for FTHF suggested that, the
affinity for FTHF to mutant binary complex was lower than bsSHMT. This could be due to
loss of interaction of N10 and formyl oxygen of FTHF the enzyme. These results suggested
that Glu plays an important role in folate binding.
The crystal structure of the complex obtained on co-crystallization of E53Q
bsSHMT with Gly and FTHF revealed that it exists in a gem-diamine form with an
orientation of PLP similar to that of wild-type ternary complex. However, electron density
for FTHF was not observed. The formation of gem-diamine in the above conditions was
supported by circular dichroism measurements of E53Q bsSHMT ternary complex in the
visible region. The absence of FTHF in the crystal structure and the formation of quinonoid
intermediate suggest that there was an initial binding of FTHF to the binary complex of
E53Q bsSHMT leading to an alteration in the orientation of PLP. Subsequently, FTHF falls
off from the active site leaving behind the gem-diamine complex.
The differences between the structures of this complex and Gly external aldimine
suggest that the changes induced by initial binding of FTHF are retained, even though
FTHF was absent in the final structure. These observations indicate that mutant enzyme
exhibits a phenomenon known as “enzyme memory”. In this concept, binding of ligands
caused conformational changes in the enzyme and even after removal of ligands such
imprints of ligand binding were retained by the enzyme. Double reciprocal plots of
bsSHMT ternary complex and E53Q bsSHMT-Gly-(FTHF) complex, supports the
suggestion that, mutant enzyme exhibits ‘enzyme memory’. There are not many examples
of structural evidence for ‘enzyme memory’.
The results obtained from these studies suggest that E53 plays an essential role in
THF⁄FTHF binding and in the proper positioning of Cβ of L-Ser for direct attack by N5 of
THF. It does not have an important role in THF-independent reactions.
5
Chapter 3: The role of tyrosine residues in cofactor binding and elucidation of
mechanism for the tetrahydrofolate-independent cleavage of L-allo threonine
Tyr residues play multiple functions in enzyme catalysis. The crystal structure of
bsSHMT showed that hydroxyl group of Y51 interacted with the phosphate group of PLP.
Binary complex of bsSHMT-Ser showed change in conformation in Y61 on binding of
L-Ser. The new conformation Y61 is close to Cβ of the bound ligand, L-Ser. The active site
Tyr’s was mutated to Y51F, Y61F and Y61A bsSHMT to understand their role in bsSHMT
catalysis. Mutation of these residues resulted in a complete loss of THF-dependent and -
independent activities. The PLP content of Y51F and Y61F bsSHMT as isolated was 0.2
mol/mol and 0.6 mol/mol of subunit, respectively, compared to 1 mol/mol of subunit in
bsSHMT. The mutant enzyme could be completely reconstituted with PLP. However, there
was an alteration in the λmax value of the internal aldimine (396 nm) in the case of Y51F
bsSHMT. A decrease in the rate of reduction with NaCNBH3 and a loss of the intermediate
in the interaction with MA suggests that the active site environment is altered in the case of
Y51F bsSHMT due to mutation. The mutation of Y51 to F strongly affects the binding of
PLP, possibly as a consequence of a change in the orientation of the phenyl ring (75°) in
Y51F bsSHMT structure. The results obtained for Y61F were also similar to that of Y51F
bsSHMT.
Y61A bsSHMT, as isolated had 1mol/mol of subunit and a absorption maximum of
425 nm similar to bsSHMT. In addition, Y61A bsSHMT was able to form an external
aldimine; this change was supported by enhanced thermal stability of the mutant enzyme on
addition of L-Ser. However, the formation of the quinonoid intermediate was hindered. An
examination of the active site geometry of bsSHMT and Y51F and Y61A mutants show that
Y51 and Y61 are not suitably placed for the removal of the proton from the hydroxyl group
of L-allo Thr. In bsSHMT, the hydroxyl group of Y51 is at a distance of 3.6 Å and 3.8 Å
from Cα of Gly and Ser, respectively. Of the two residues, Y51 is unlikely to be involved in
proton abstraction from Cα of the bound ligand due to its longer distance and improper
geometry. In contrast, the OH of Y61 is at a distance of 3.3 Å and 3.2 Å from Cα of Gly and
Ser, respectively. In Gly, L-Ser and L-allo Thr complexes of Y51F bsSHMT, the OH of
6
Y61 and Cα of bound ligand are at distances 4.97, 4.39 and 4.39 Å, respectively. Y61
therefore may be involved in Cα proton abstraction in the THF-independent reaction.
This might explain the loss of L-allo Thr cleavage activity of Y51F bsSHMT. It
could therefore, be concluded that Y51 is important for PLP binding and proper positioning
of Y61, while Y61 could be involved in the abstraction of the proton from Cα carbon of
L-allo Thr. Based on these observations, a possible mechanism for SHMT catalyzed
cleavage of L-allo Thr is suggested.
Chapter 4: The interaction of bsSHMT with specific inhibitors from extracts of various
spices
The pivotal role of SHMT in the interconversion of folate coenzymes and its altered
kinetic properties in neoplastic tissues suggested that it could be a potential target for cancer
chemotherapy. This chapter deals with understanding the interaction of aminooxy
compounds with K226M, K226Q, E53Q, Y51F, Y61F and Y61A bsSHMT. Methoxyamine
(MA) and aminooxyacetic acid (AAA) interacted with the mutant proteins in different
manner. K226M, K226Q bsSHMT did not react with MA and E53Q, Y51F, Y61A and
Y61F bsSHMT failed to form 388 nm intermediate on addition of MA. These results
suggest that any change at the active site environment or PLP orientation in bsSHMT could
lead to the loss of intermediate formation on interaction with MA.
Lys mutants were able to interact with AAA, however the interaction rate was
slower compared to bsSHMT suggesting that AAA was a more reactive compared to MA.
The absence of Schiff’s base and a change in the orientation of PLP at the active site of
K226M and K226Q bsSHMTs probably results in poor binding of AAA at the active site.
Y51F, Y61F and Y61A bsSHMT showed a similar interaction pattern as that of bsSHMT.
Only E53Q bsSHMT showed the formation of an intermediate absorbing at 388 nm on
addition of AAA. Stopped-flow studies suggested a drastic decrease in the rate at which
AAA interacts with E53Q bsSHMT, when compared to bsSHMT. It is possible that Glu 53
may be the crucial residue involved in binding of AAA as well as in enhancing the rate of
the reaction.
7
A fruitful approach of identifying compounds which have anti-carcinogenic effect is
to carry out homology modeling and docking of compounds available in several chemical
libraries. Homology modeling and docking of commercially available folate analogues were
performed. Docking of folate analogues resulted in identification of 14 best compounds of
these four water soluble compounds were examined for their ability to inhibit bsSHMT.
Activity studies suggested that folate analogues failed to inhibit bsSHMT. In addition to
synthetic compounds, naturally occurring compounds from spices that can potentially work
as anti-cancer agents were also analyzed for bsSHMT inhibition. The effect of spices such
as garlic, ginger, chilli, turmeric extracts was examined. Among all spices, heat-treated and
lyophilized garlic extract showed 91% inhibition of bsSHMT. D and L-allyl Gly analogues
of substrate L-Ser and S-allyl Cys (probable inhibitor form garlic extract) D and L-allyl
Gly, Alliin did not inhibit bsSHMT activity.
Docking studies with S-allyl Cys, D and L-allyl Gly and alliin suggest that alliin,
L-allyl Gly and S-allyl Cys bind at the active site of bsSHMT with different orientations.
However, alliin and L-allyl Gly bind in similar orientation/position which is different from
that of OADS. Binding of S-allyl Cys was almost similar to that of OADS. These results
suggest that S-allyl Cys could be a potential inhibitor. Chemical synthesis of S-allyl Cys
and further activity measurement has to be carried out to demonstrate the inhibition. These
results emphasize the importance of examining the naturally occurring and synthetic
compounds as possible chemotherapeutic agents.
The results presented in this study bring in the importance of correlating structural
information with catalytic function. The mutation of Lys 226 of bsSHMT demonstrated its
role in facilitating the change in orientation of PLP during catalysis. Glu 53 on interaction
with L-Ser positions the Cα-Cβ bond for attack by N5 of THF, a crucial step in L-Ser
cleavage. The loss of efficient binding of FTHF to E53Q-Gly binary complex up on Glu 53
mutation results in the enzyme exhibiting a phenomenon called ‘enzyme memory’. Tyr 51 in
involved in cofactor binding. Tyr 61 plays an important role in Cα proton abstraction
leading to THF-independent cleavage of 3-hydroxy amino acids. These observations
prompted proposal of a new mechanism for aldol cleavage of 3-hydroxy amino acids. Both
8
commercially available and naturally occurring bioactive compounds from spices that can
serve as anti-cancer agents were analyzed for bsSHMT inhibition. Garlic extract showed
the maximum inhibition suggesting the importance of the naturally occurring compounds as
possible and potential chemotherapeutic agents against cancer.
The summary and conclusions in the thesis briefly highlights the salient features of
the present investigation. The literature sited in the complete text is arranged in alphabetical
order under reference section, which gives all the relevant details including title, journal
year, volume and pagination.