Biochemical Journal

Document Sample
Biochemical Journal Powered By Docstoc
Biochem. J. (2009) 417, 757–764 (Printed in Great Britain)   doi:10.1042/BJ20081247                                                                                           757

Disruption of the crossover helix impairs dihydrofolate reductase activity
in the bifunctional enzyme TS–DHFR from Cryptosporidium hominis
Melissa A. VARGO*, W. Edward MARTUCCI*† and Karen S. ANDERSON*1
*Department of Pharmacology, Yale University School of Medicine, 333 Cedar Street, New Haven, CT 06520, U.S.A., and †Department of Molecular Biophysics and Biochemistry,
Yale University School of Medicine, 333 Cedar Street, New Haven, CT 06520, U.S.A.

                                                                                                                                                                                      Biochemical Journal
In contrast with most species, including humans, which have                               of the helix were mutated to alanine. These mutant enzymes were
monofunctional forms of the folate biosynthetic enzymes TS                                studied using a rapid transient kinetic approach. The mutations
(thymidylate synthase) and DHFR (dihydrofolate reductase),                                caused a dramatic decrease in the DHFR activity. The DHFR
several pathogenic protozoal parasites, including Cryptospor-                             catalytic activity of the alanine-face mutant enzyme was 30 s−1 , the
idium hominis, contain a bifunctional form of the enzymes on                              glycine-face mutant enzyme was 17 s−1 , and the all-alanine helix
a single polypeptide chain having both catalytic activities. The                          enzyme was 16 s−1 , all substantially impaired from the wild-type
crystal structure of the bifunctional enzyme TS–DHFR C. hominis                           DHFR activity of 152 s−1 . It is clear that loss of helix interactions
reveals a dimer with a ‘crossover helix’, a swap domain between                           results in a marked decrease in DHFR activity, supporting a role
DHFR domains, unique in that this helical region from one                                 for this swap domain in DHFR catalysis. The crossover helix
monomer makes extensive contacts with the DHFR active site                                provides a unique structural feature of C. hominis bifunctional
of the other monomer. In the present study, we used site-directed                         TS–DHFR that could be exploited as a target for species-specific
mutagenesis to probe the role of this crossover helix in DHFR                             non-active site inhibitors.
catalysis. Mutations were made to the crossover helix: an ‘alanine-
face’ enzyme in which the residues on the face of the helix close
to the DHFR active site of the other subunit were mutated to                              Key words: bifunctional, Cryptosporidium hominis, dihydrofolate
alanine, a ‘glycine-face’ enzyme in which the same residues were                          reductase, protozoal parasite, thymidylate synthase, transient
mutated to glycine, and an ‘all-alanine’ helix in which all residues                      kinetics.

INTRODUCTION                                                                              apicomplexan TS–DHFR enzymes, there is no N-terminal tail
                                                                                          in C. hominis and T. gondii and only a five-amino-acid tail in
Protozoal parasites, such as Cryptosporidium hominis, Leish-                              P. falciparum, and the linker region between the TS and DHFR
mania major, Toxoplasma gondii and Plasmodium falciparum,                                 domains is long (89 residues in P. falciparum, 72 residues in
are unusual in that the TS (thymidylate synthase) and DHFR                                T. gondii and 58 residues in C. hominis). This linker region begins
(dihydrofolate reductase) enzymes exist on a single polypeptide                           in the DHFR domain of one monomer (monomer A), crosses to
chain to form the bifunctional enzyme TS–DHFR (TS–DHFR                                    the other monomer (monomer B), forms the crossover helix that
is a functional designation as dihydrofolate is produced at TS                            makes extensive contacts with the opposite DHFR domain, then
and used at DHFR; the bifunctional enzyme is also referred to as                          crosses back to monomer A to form the TS domain (Figure 1).
DHFR–TS because the DHFR domain is at the N-terminal portion                                 Apart from structural differences, these enzymes also display
of the bifunctional enzyme) [1]. These are essential enzymes and                          unique kinetic behaviours in terms of how the DHFR catalytic
have been established as drug targets. TS catalyses the conversion                        activity may be modulated. Moreover, each protozoal species
of dUMP (2 -deoxyuridine monophosphate) and CH2 H4 folate                                 exhibits distinct modes of modulation. The catalytic activity of
(methylene tetrahydrofolate) into dTMP (2 -deoxythymidine                                 DHFR from L. major and P. falciparum is enhanced upon TS
monophosphate) and H2 folate (dihydrofolate). DHFR then                                   ligand binding, whereas C. hominis DHFR activity is unaffected
catalyses the reduction of H2 folate by NADPH to form H4 folate                           by the presence of TS ligands at the TS active site (Table 1) [3–
(tetrahydrofolate), which is used for one-carbon transfer reactions                       5]. Despite sharing a linker and crossover helix, P. falciparum
in many biochemical processes.                                                            and C. hominis clearly differ in terms of DHFR kinetics. A
   After the determination of the crystal structure of C. hominis                         closer look at the P. falciparum structure shows that, although
TS–DHFR, it was suggested that there are two families of                                  the enzyme does form a crossover helix in the same general
bifunctional TS–DHFR: a short-linker family with an N-terminal                            orientation as C. hominis, it does not contact the DHFR active
tail, as in the kinetoplastids, which includes L. major and the                           site of the other monomer; however, the crossover helix in
trypanosomes; and a long-linker family which contains a donated                           C. hominis DHFR makes extensive contacts with the catalytically
or crossover helix, as in the apicomplexan family, containing                             important helix B of the DHFR active site. This unique structural
C. hominis, P. falciparum and T. gondii [2]. The short-linker                             characteristic led us to hypothesize that although there is no
family has a linker length of two residues (L. major) and an                              domain–domain modulation of catalytic activity between the TS
N-terminal tail of 22 residues, which stretches from the DHFR                             and DHFR domains of the same subunit, the crossover helix
domain and wraps around the TS domain. However, in the                                    swap domain may be responsible for modulating catalysis for

   Abbreviations used: CH2 H4 folate, methylene tetrahydrofolate; DHFR, dihydrofolate reductase; dTMP, 2 -deoxythymidine monophosphate; dUMP,
2 -deoxyuridine monophosphate; FRET, fluorescence resonance energy transfer; H2 folate, dihydrofolate; H4 folate, tetrahydrofolate; TS, thymidylate
synthase; TS–DHFR, bifunctional form of TS and DHFR on a single polypeptide chain having both catalytic activities.
     To whom correspondence should be addressed (email

                                                                                                                       c The Authors Journal compilation c 2009 Biochemical Society
758                M. A. Vargo, W. E. Martucci and K. S. Anderson

Table 1     Structural and kinetic comparison of TS–DHFR from L. major [5], P. falciparum [4] and C. hominis [3]

                      N-terminal tail?             Linker region                                          DHFR activity in the absence          DHFR activity in the presence          DHFR rate
Species               (amino-acid length)          (amino acid length)          Crossover helix?          of TS ligands (s−1 )                  of TS ligands (s−1 )                   enhancement

L. major              Yes (22)                     Short (2)                    No                         14                                   120                                    ∼ 10-fold
P. falciparum         Yes (5)                      Long (89)                    Yes                        61                                   135                                     ∼ 2-fold
C. hominis            No                           Long (58)                    Yes                       152                                   152                                    None

Figure 1        Cryptosporidium hominis TS–DHFR structure (PDB ID: 1QZF)
(A) Dimer structure of TS–DHFR. The TS and DHFR domains are labelled. Crossover helix and helix B are also labelled in the DHFR domains. The DHFR ligands, NADP+ and H2 folate are shown
in stick formation. (B) Close-up of the crossover helix region. Residues on the crossover helix (light grey) are displayed as well as residues on the active-site helix (dark grey). (C) Space-filling
representation highlighting the close interactions of the crossover helix (light grey) and helix B (dark grey) residues. DHFR active-site ligands are shown in stick formation.

the C. hominis DHFR. The residues of this ‘crossover helix’ were                                     NADPH was determined by using a molar absorption coefficient
mutated in order to determine whether these structural differences                                   (ε) of 6220 M−1 · cm−1 at 340 nm. [3 H]-labelled H2 folate and
might account for some of the mechanistic differences between                                        CH2 H4 folate were synthesized as previously described using
enzymes from different species.                                                                      tritiated folic acid as the starting material [11,12]. [3 ,5 ,7,9-3 H]-
   Cryptosporidiosis, caused by C. hominis infection, is one of the                                  folic acid was purchased from Moravek Biochemicals.
major causative agents of the diarrhoeal diseases in AIDS patients
[6–8]. There have been several outbreaks of C. hominis infections
                                                                                                     Plasmids and site-directed mutagenesis
from contaminated water supplies in the past few years that have
sickened thousands including an episode in a New York water                                          Full-length C. hominis TS–DHFR was encoded in the pTrc99A-
amusement park [9,10]. There is currently no effective treatment                                     rHCp (the ‘genotype 1’ TS–DHFR gene derived from a human
for this disease, thus there is an urgent need for new drugs. Further                                parasite clone), kindly provided by Dr Richard G. Nelson
understanding of the mechanistic and structural characteristics of                                   (formerly of the Division of Infectious Diseases, University
the enzyme may reveal key features of catalytic function that could                                  of California, San Francisco, CA, U.S.A.) and Dr Amy C.
be exploited in the design of potential species-specific inhibitors.                                  Anderson (Department of Pharmaceutical Sciences, University
                                                                                                     of Conneticut, Storrs, CT, U.S.A.). Site-directed mutagenesis
                                                                                                     was performed using the Stratagene QuikChange® kit. The
MATERIALS AND METHODS                                                                                mutations for the alanine-face mutant (K194A, D198A, L202A,
                                                                                                     D205A, I206A and R210A) were all introduced using a single
Chemicals and reagents                                                                               oligonucleotide that encoded the changes. The same residues
All buffers and reagents were of the highest purity. NADPH                                           were mutated to glycine in the glycine-face mutant. In order
and dUMP were purchased from Sigma. The concentration of                                             to form the all-alanine helix mutation, a second round of

c The Authors Journal compilation c 2009 Biochemical Society
                                                                                    Role of crossover helix in C. hominis TS–DHFR                     759

PCR was used to introduce the remaining alanine mutations              HPLC analysis
at positions 195, 196, 197, 199, 200, 201, 203, 204, 207
                                                                       The tritiated products from the rapid chemical quench experi-
and 208. DNA sequencing confirmed the presence of all of
                                                                       ments were analysed using reversed-phase HPLC connected to
the mutations. CD spectra of wild-type and all three mutant
                                                                       a radioactivity flow detector as described previously [17]. The
enzymes were globally the same suggesting overall folding of
                                                                       isocratic separation was performed using a BDS-Hypersil C18
the proteins is maintained; however, because the expected change
                                                                       reversed-phase column (250 mm × 6.4 mm) (Alltech) with a flow
in the percentage of helical content for the mutant enzymes
                                                                       rate of 1 ml/min using 10 % methanol in 180 mM triethylammon-
were all within error of wild-type, we could not determine
                                                                       ium bicarbonate (pH 8.0). The elution times for the products were
by CD whether the crossover helix is maintained as a helix.
                                                                       as follows: 9 min for H4 folate, 17 min for H2 folate and 20 min for
The poly-alanine sequence has a high propensity to form α-
                                                                       CH2 H4 folate.
helix [13], and therefore the alanine-face and all-alanine mutant
enzymes will presumably retain the α-helical structure. In addi-
tion, based on nnPREDICT, secondary structure prediction soft-         Stopped-flow fluorescence experiments
ware, the alanine sequences are predicted to be helical; however,      Stopped-flow experiments were performed using a Kintek SF-
this software would predict the glycine-face enzyme would not          2001 apparatus (Kintek Instruments). To determine the rate for the
maintain the helical structure [14]. Therefore, the glycine-face       DHFR reaction, coenzyme FRET (fluorescence resonance energy
mutant was made to determine whether simply the presence of            transfer) was measured at an excitation of 290 nm with an output
a helix, and not specific interactions, is necessary for maximal        filter at 450 nm. The signal measured at 450 nm would decrease as
catalytic activity.                                                    bound NADPH involved in the FRET is converted into NADP+
                                                                       and released from the enzyme. In single turnover experiments,
                                                                       enzyme (50 μM) was incubated with 500 μM NADPH and then
Protein expression and purification                                     mixed with 10 μM H2 folate. The data was collected over a
                                                                       given time interval using software provided by Kintek. For burst
The proteins were overexpressed in BL21 Escherichia coli and           experiments, 7.5 μM enzyme was incubated with 50 μM H2 folate
purified using previously described methods [5,15]. The protein         and then mixed with 500 μM NADPH. In order to determine
was further purified using a PD-10 column from Amersham Bio-            whether ligands bound at the TS site would result in an activated
sciences to remove residual H2 folate. The concentration of            DHFR rate, as is seen with other TS–DHFR enzymes from other
purified C. hominis TS–DHFR was determined spectrophotomet-             species, DHFR burst experiments were also examined in the pres-
rically using a molar absorption coefficient of 80 722 M−1 · cm−1 .     ence of TS ligands. For these experiments, enzyme was incubated
The DHFR activity was determined by following the decrease             in the presence of 100 μM FdUMP (5-fluoro-dUMP; a TS
in absorbance at 340 nm ( ε = −12.8 mM−1 · cm−1 ) which                inhibitor), 200 μM CH2 H4 folate and 1 mM NADPH, and the
corresponds to the conversion of NADPH and H2 folate into              reaction was initiated by rapidly mixing with 200 μM H2 folate.
NADP+ and H4 folate. The TS activity was determined by                 The data was collected over a given time range. Runs (4–7) were
following the increase in absorbance at 340 nm ( ε = 6.4               collected and averaged. The data were fitted to either a single
mM−1 · cm−1 ) as the substrates CH2 H4 folate and dUMP are             exponential or burst equation to obtain the rate constants.
converted into H2 folate and dTMP [16]. Mutant enzymes were
purified in a similar manner to wild-type enzymes.
                                                                       Expression and steady-state activity of helix mutants
Rapid chemical quench experiments
                                                                       All of the helix mutants were purified using the previously
Rapid chemical quench experiments were performed using                 published protocols [5,15] and all enzymes were ∼ 95 % pure as
a Kintek RFQ-3 rapid chemical quench apparatus (Kintek                 measured by SDS/PAGE gel electrophoresis. The DHFR steady-
Instruments). Single turnover experiments were initiated by            state rates were 2.7 + 0.1 s−1 , 1.7 + 0.1 s−1 , 1.9 + 0.4 s−1 and
                                                                                             −                −              −
mixing 15 μl of enzyme solution [enzyme + 2 × buffer (1 mM             1.1 + 0.4 s−1 , for wild-type, alanine-face, glycine-face and all-
EDTA, 50 mM MgCl2 and 50 mM Tris/HCl, pH 7.8)] with 15 μl                  −
                                                                       alanine mutant enzymes respectively. The TS steady-state rates
of the tritiated substrate (approx. 20 000 d.p.m.). The DHFR single    were 3.5 + 0.4 s−1 , 3.0 + 0.2 s−1 , 1.6 + 0.4 s−1 and 1.7 + 0.1 s−1
                                                                                 −              −               −                 −
turnover reaction was monitored by the addition of [3 H]H2 folate      for wild-type, alanine-face, glycine-face and all-alanine mutant
(10 μM) to enzyme (50 μM) and NADPH (500 μM); all concen-              enzymes respectively. All rates were determined using a spectro-
trations shown in the text are after mixing. The TS–DHFR               photometric assay.
reaction was monitored by the addition of [3 H]CH2 H4 folate
(10 μM) to enzyme (50 μM), dUMP (500 μM) and NADPH
(500 μM). The reactions were terminated by quenching with              Single turnover of the DHFR reaction
67 μl of 0.78 M KOH to give a final concentration of 0.54 M             Single turnover experiments were conducted to directly assess
KOH [5]. The quenching solution also contained 10 % sodium             the effects of the mutations on the rate of catalysis at the
ascorbate and 200 mM 2-mercaptoethanol to prevent degradation          DHFR site. Experiments were performed using stopped-flow fluo-
of the products. To confirm complete quenching of the enzymatic         rescence (Figure 2). Bifunctional TS–DHFR from wild-type or
reactions, controls in which substrate was added to a premixed         mutant enzymes was pre-incubated with saturating amounts of
solution of enzyme and quench were performed for each                  NADPH and then mixed with limiting amounts of H2 folate. The
experiment, showing stability of the CH2 H4 folate. In addition, a     alanine-face mutant catalytic rate was 30 + 1 s−1 , the glycine-
control in which enzyme was allowed to react with substrates for       face catalytic rate was 17 + 1 s−1 and the all-alanine catalytic rate
1 min was performed to show complete conversion into products          was 16 + 1 s−1 , compared with a catalytic rate of 152 + 7 s−1 for
                                                                               −                                                −
and ensure the stability of the formed H4 folate. The rate constants   the wild-type enzyme. Similar experiments were performed using
were determined by fitting the data to either a single or double        rapid chemical quench. Bifunctional TS–DHFR was pre-incub-
exponential equation using Kaleidagraph.                               ated with a saturating amount of unlabelled NADPH and then

                                                                                               c The Authors Journal compilation c 2009 Biochemical Society
760               M. A. Vargo, W. E. Martucci and K. S. Anderson

Figure 2     DHFR catalysis as measured by coenzyme FRET
Fluorescence excitation was at 290 nm and emission was at 450 nm. In all panels (A–D), 50 μM enzyme was incubated with 500 μM NADPH and then rapidly mixed with 10 μM H2 folate. The
traces show the change in fluorescence signal after mixing as a function of time. (A) A representative stopped-flow trace of wild-type TS–DHFR. Data were fitted to a single-exponential with a rate of
152 + 7 s−1 . (B) A representative stopped-flow trace for the alanine-face TS–DHFR mutant. Data were fitted to a single exponential with a rate of 30 + 1 s−1 . (C) A representative stopped-flow trace
     −                                                                                                                                              −
of the all-alanine TS–DHFR mutant. Data were fitted to a single exponential with a rate of 16 + 1 s−1 . (D) A representative stopped-flow trace for the glycine-face TS–DHFR mutant. Data were fitted
to a single exponential with a rate of 17 + 1 s−1 .

rapidly mixed with a limiting amount of radiolabelled H2 folate.                                     subsequent conversion of the H2 folate into H4 folate can be
Similar rates were obtained as compared with the stopped-flow                                         measured at the DHFR site. The bifunctional TS–DHFR was
results. Both methods confirm a reduced catalytic rate for all                                        pre-incubated with saturating amounts of unlabelled dUMP and
mutant enzymes. Doubling the enzyme concentration did not                                            NADPH, and then rapidly mixed with a limiting amount of
change the catalytic rate, demonstrating that binding was not rate-                                  radiolabelled CH2 H4 folate. The time course for the disappearance
limiting in these assays.                                                                            of CH2 H4 folate, the appearance and disappearance of H2 folate,
                                                                                                     and the formation of H4 folate is shown in Figure 4. The build-up of
                                                                                                     H2 folate in all enzymes is consistent with previously determined
Pre-steady-state burst experiments of the DHFR reaction                                              characteristics of C. hominis TS–DHFR. The accumulation of
The DHFR reaction was also studied under pre-steady-state burst                                      H2 folate was seen to a greater extent in the all-alanine helix
conditions (the substrate in slight excess over enzyme), using                                       than in wild-type or mutant-face enzymes (6.2 μM compared
stopped-flow fluorescence. Bifunctional TS–DHFR was incub-                                             with 4.1 μM for wild-type enzyme) (Table 2). The H2 folate also
ated with H2 folate and then mixed with saturating NADPH. A                                          persisted for longer times in all of the mutant enzymes than in
burst was observed for all enzymes (Figure 3), suggesting that                                       the wild-type enzyme (Table 2). Interestingly, the H2 folate build-
chemistry is not the rate-limiting step, but rather a later step,                                    up lasted for much longer times (∼ 0.7 s) in the glycine-face
perhaps product release, limits the catalytic cycle. DHFR burst                                      enzyme than either of the alanine mutants. The end result, the
reactions were also performed in the presence of TS ligands in                                       rate of H4 folate formation, was decreased in all of the mutants:
order to determine whether the DHFR rate would be activated                                          wild-type, 10.1 + 0.9 s−1 ; alanine-face mutant, 6.7 + 0.8 s−1 ; all-
                                                                                                                       −                                   −
in the presence of TS ligands. The DHFR rate did not change in                                       alanine mutant, 4.0 + 0.4 s−1 ; and glycine-face mutant, 1.0 +
                                                                                                                             −                                         −
the presence of ligands for any of the mutant enzymes, indicating                                    0.3 s .
that there is no enhancement in the DHFR catalytic rate even if                                         The rates of catalysis for the TS reaction were derived from the
the TS active site has bound ligands.                                                                full TS–DHFR reaction time courses. Rates were determined from
                                                                                                     CH2 H4 folate consumption. The rates for the alanine-face and the
                                                                                                     all-alanine helix mutant enzymes were not significantly different
Single turnover of the TS–DHFR reaction                                                              from that of wild-type (all within an error of 15 s−1 ). However, in
The full bifunctional TS–DHFR reaction was examined using                                            the case of the glycine-face enzyme, the rate for the disappearance
rapid chemical quench. In this experiment, the formation of                                          of CH2 H4 folate was much slower than the other enzymes, with a
H2 folate from CH2 H4 folate can be measured at the TS site and                                      rate of 2.7 + 0.5 s−1 .

c The Authors Journal compilation c 2009 Biochemical Society
                                                                                                                      Role of crossover helix in C. hominis TS–DHFR                          761

Figure 3     C. hominis DHFR stopped-flow pre-steady-state burst experiments
Fluorescence excitation was at 290 nm and emission was at 450 nm. In all panels (A–D), 7.5 μM enzyme was incubated with 50 μM H2 folate and then rapidly mixed with 500 μM NADPH. The
change in fluorescence was measured over a given time range. (A) A representative stopped-flow trace for wild-type TS–DHFR. Data were fitted to the burst equation with the fast rate of 60 + 2 s−1
and the slower rate consistent with the steady-state rate. (B) A representative stopped-flow trace for the alanine-face TS–DHFR mutant. Data were fitted to the burst equation with a faster rate of
22 + 1 s−1 and a slower phase similar to the steady-state rate. (C) A representative stopped-flow trace for the all-alanine TS–DHFR mutant. Data were fitted to a burst equation with a faster rate
of 9.5 + 0.1 s−1 . (D) A representative stopped-flow trace for the glycine-face TS–DHFR mutant. Data were fitted to the burst equation with a faster rate of 8.9 + 0.3 s−1 .
       −                                                                                                                                                       −

Table 2     Data from the TS–DHFR single turnover reactions                                         gesting that the interactions between the crossover helix and the
                                                                                                    residues of the opposite DHFR domain are important for maximal
                                                   Persistence of H2 folate                         DHFR catalytic activity. The crossover helix packs against helix
                       Maximum accumulation        peak (amount of H2 folate   TS–DHFR              B in the opposite DHFR domain. Residues on helix B opposite
TS–DFHR form           of H2 folate (μM)           at 0.2 s in μM)             reaction (s−1 )
                                                                                                    from the crossover helix form a portion of the DHFR active
                                                                                                    site and contain many highly conserved residues including Phe36 ,
Wild-type              4.1                         1.5                         10.1 + 0.9
Alanine-face mutant    5.1                         2.6                          6.7 + 0.8
                                                                                                    which is universally conserved and a key residue in the catalytic
                                                                                    −               mechanism of DHFR, specifically the conformational change and
Glycine-face mutant    3.8                         3.6                              + 0.3
                                                                                1.0 −
All-alanine mutant     6.2                         3.9                          4.0 + 0.4
                                                                                                    hydride transfer [18,19]. Helix B and the crossover helix form at
                                                                                                    least seven tight interactions, some of which are as follows: Asp198
                                                                                                    makes a salt bridge to Asn42 , Leu202 is involved in hydrophobic
                                                                                                    interactions with Ile39 , Asp205 makes a salt bridge with Lys38 , and
                                                                                                    Ile206 makes hydrophobic interactions with the aliphatic portion
                                                                                                    of the Lys34 side chain and also the side chain of Phe35 . Therefore
                                                                                                    mutating residues on the crossover helix may cause slight shifts
Based on currently known structures and sequence alignments,                                        in the helix of the active site. As shown in E. coli, the DHFR
the crossover helix appears to be present only in TS–DHFR                                           domain proceeds through a series of conformational changes
of the apicomplexan family, including C. hominis, P. falciparum                                     along the reaction pathway, some of which are distal to the
and T. gondii [2]. The linker and crossover helix in C. hominis                                     active site, and it is therefore possible that small perturbations
brings the two DHFR domains in much closer proximity than                                           in these conformational changes, caused by mutations altering
bifunctional TS–DHFR enzymes with very short linker regions,                                        helix B-crossover helix packing, could affect catalysis [18,20].
such as L. major. Based on the structural differences between                                       The additional loss in activity for the all-alanine and glycine-
the families, we tested the role of the crossover helix using a                                     face mutant enzymes can be explained by the loss of additional
site-directed mutagenesis approach. Our kinetic characterization                                    interactions. Although they do not involve helix B, several of these
of C. hominis helix mutations has provided new insight into the                                     interactions are between the two DHFR domains. Residues on an
role of the crossover helix.                                                                        orthogonal face of the crossover helix interact with a β-sheet of
   The alanine-face, glycine-face and all-alanine helix enzymes                                     the DHFR domain. There are also interactions between the cross-
have lower DHFR catalytic rates than wild-type enzyme, sug-                                         over helix and residues located in the flexible tethers of the

                                                                                                                                     c The Authors Journal compilation c 2009 Biochemical Society
762               M. A. Vargo, W. E. Martucci and K. S. Anderson

Figure 4     C. hominis TS–DHFR single enzyme turnover reaction time course
Full TS–DHFR reaction: , CH2 H4 folate; , H2 folate; and , H4 folate. In all panels (A–D), 50 μM enzyme is incubated with 500 μM dUMP and 500 μM NADPH, followed by rapid mixing
with 10 μM H2 folate. (A) Wild-type enzyme: H4 folate production is 10.1 + 0.9 s−1 . (B) Alanine-face mutant enzyme: H4 folate production is 6.7 + 0.8 s−1 . (C) All-alanine mutant enzyme: H4 folate
                                                                          −                                                                      −
production is 4.0 + 0.4 s−1 . (D) Glycine-face mutant enzyme: H4 folate production is 1.0 + 0.3 s−1 (note the change in the time-scale for D).
                  −                                                                       −

crossover domain. It is likely that these interactions are necessary                                 region. Based on the mutant enzymes made in the present
for a maximal DHFR catalytic activity, possibly by positioning                                       study, it appears that the specific interactions of the crossover
the crossover helix in an optimal orientation. The returning tether                                  helix are necessary for a fully active DHFR domain, whereas
makes several interactions and hydrogen bonds with the TS                                            simply the presence of a stable helix is important for full TS
domain. The loss of these interactions could explain the reduction                                   activity.
in TS steady-state activity for the all-alanine mutant enzyme.                                          Interestingly, L. major, which has a very short linker (∼ two
Cumulatively, these lost interactions result in an additional 2-fold                                 amino acids, and therefore no crossover helix), has a very low
loss in activity compared with the alanine-face mutant enzyme.                                       DHFR activity of 14 s−1 . However, when ligands are bound at the
In fact, initial experiments mutating residues Leu203 and Phe207 on                                  TS site, the DHFR activity is enhanced approx. 10-fold to a rate
the orthogonal face of the crossover helix cause a loss of activ-                                    of 120 s−1 . This enhanced rate is comparable with the activity of
ity to 30 s−1 . The argument could be made that a stretch of alanine                                 C. hominis DHFR. Interestingly, the C. hominis all-alanine mutant
mutations introduced anywhere in the protein could cause this                                        enzyme has an activity (16 s−1 ) equivalent to that of L. major in
reduction in rate; however, when the analogous mutations are                                         the unliganded, unenhanced state (14 s−1 ). The bifunctional TS–
made in the P. falciparum enzyme, there is no reduction in activity                                  DHFR enzyme from P. falciparum is an interesting combination
[4]. The crossover helix appears to be necessary to retain the                                       of C. hominis and L. major both structurally and catalytically.
productive conformation of the active-site helix and to allow for                                    Structurally, P. falciparum has a long linker containing a crossover
proper co-ordinated movement, and thus maximal activity. In the                                      helix between the TS and DHFR domains (similar to C. hominis),
alanine-face and all-alanine mutant enzymes, the crossover helix                                     but also has an N-terminal tail similar to L. major. Unlike
would presumably still be present; however, in the case of the                                       C. hominis, P. falciparum DHFR activity increases 2-fold when
glycine-face mutant, we would predict that the crossover helix is                                    TS ligands are bound, to reach an enhanced activity of 130 s−1 ,
no longer maintained as a helix. The glycine-face enzyme results                                     similar to the inherent rate of C. hominis DHFR. P. falciparum
in a similar DHFR rate to the all-alanine mutant enzyme, but,                                        does have a crossover helix; however, upon mutation of the helix
surprisingly, substantially alters the TS rate. Because the linker,                                  face residues to alanine, there is no reduction in DHFR activity
upon returning to its own domain, makes many contacts with                                           in contrast with that observed for C. hominis [4], as expected
the TS domain, this entire region could be disrupted by the                                          since the helix does not contact the DHFR active site, but rather
lack of a structurally stable helix. Although we do not observe                                      has electrostatic residues which make contacts with several lysine
that ligands binding to TS enhance DHFR activity, there may                                          residues scattered throughout the DHFR domain. It appears that
be a reciprocal modulation of TS activity by DHFR mediated                                           the crossover helix plays a different role in P. falciparum than in C.
through proper positioning of the crossover helix and linker                                         hominis, offering further evidence that these bifunctional enzymes

c The Authors Journal compilation c 2009 Biochemical Society
                                                                                                             Role of crossover helix in C. hominis TS–DHFR                          763

have developed different modes of modulating or enhancing                                  REFERENCES
activity. TS–DHFR from T. gondii also shows an enhanced
                                                                                            1 Yuvaniyama, J., Chitnumsub, P., Kamchonwongpaisan, S., Vanichtanankul, J.,
DHFR activity whether or not ligands are bound at TS, similar
                                                                                              Sirawaraporn, W., Taylor, P., Walkinshaw, M. D. and Yuthavong, Y. (2003) Insights
to C. hominis TS–DHFR. This parasite is also a member of                                      into antifolate resistance from malarial DHFR–TS structures. Nat. Struct. Biol. 10,
the apicomplexan family; and, therefore has a long linker and                                 357–365
crossover helices, based on secondary structure prediction.                                 2 O’Neil, R. H., Lilien, R. H., Donald, B. R., Stroud, R. M. and Anderson, A. C. (2003)
   Based on the information in the present study, we would predict                            Phylogenic classification of protozoa based on the structure of the linker domain in the
that the crossover helix in T. gondii is much like C. hominis, and                            bifunctional enzyme, dihydrofolate reductase–thymidylate synthase. J. Biol. Chem. 278,
makes extensive contacts with the DHFR active site of the opposite                            52980–52987
monomer. Crystallography and mutational studies are currently                               3 Atreya, C. E. and Anderson, K. S. (2004) Kinetic characterization of bifunctional
underway to validate this hypothesis.                                                         thymidylate synthase-dihydrofolate reductase (TS–DHFR) from Cryptosporidium
   It appears that for all species studied, there is a maximal                                hominis : a paradigm shift for TS activity and channeling behavior. J. Biol. Chem. 279,
DHFR catalytic rate. An optimized DHFR catalytic rate may be                                  18314–18322
essential in order to help the parasites compete for host folates.                          4 Dasgupta, T. and Anderson, K. S. (2008) Probing the role of parasite-specific, distant
It is important to note that folate-dependent enzymes are present                             structural regions on communication and catalysis in the bifunctional thymidylate
                                                                                              synthase–dihydrofolate reductase from Plasmodium falciparum . Biochemistry 47,
in much greater concentrations than folates in the cell (making
the concentration of free folates in the cell negligible) and                               5 Liang, P. H. and Anderson, K. S. (1998) Substrate channeling and domain–domain
therefore these single-enzyme turnover rates are also probably                                interactions in bifunctional thymidylate synthase–dihydrofolate reductase. Biochemistry
relevant when studying these enzymes [21,22]. Rather than TS                                  37, 12195–12205
to DHFR modulation of catalytic activity, as seen in L. major or                            6 Current, W., Reese, N., Ernst, J., Bailey, W., Heyman, M. and Weinstein, W. (1983)
P. falciparum, our results suggest that an optimal rate is affected                           Human Cryptosporidiosis in immunocompetent and immunodeficient persons. Studies
by a swap domain between subunits. In addition to influencing                                  of an outbreak and experimental transmission. N. Eng. J. Med. 308, 1252–1257
the DHFR activity of the monomer to which it crosses, another                               7 Guerrant, R. (1997) Cryptosporidiosis : an emerging, highly infectious threat.
potential role for the crossover helix is to provide a structural motif                       Emerg. Infect. Dis. 3, 51–57
which is used for interdomain communication, first proposed by                               8 Tzipori, S. (1998) Cryptosporidiosis : laboratory investigations and chemotherapy.
O’Neil et al. [2]. The crossover helix is connected to its own                                Adv. Parasitol. 40, 187–221
domain through two linkers. The subdomains of the DHFR active                               9 Gupta, M. and Haas, C. (2004) The Milwaukee Cryptosporidium outbreak: assessment of
site (including helix B) move during the catalytic cycle, resulting                           incubation time and daily attack rate. J. Water Health 2, 59–69
                                                                                           10 O’Connor, A. (2005) Tainted water at a state park claims victims in 20 counties.
in as great as a 1 Å (1 Å = 0.1 nm) shift in position at the active
                                                                                              New York Times, 20 August, 2005
site [23]. Because helix B and the crossover helix are so closely
                                                                                           11 Blakley, R. L. (1960) Crystalline dihydropteroylglutamic acid. Nature 188,
associated, movements of helix B during the catalytic cycle                                   231–232
may be transferred to the crossover helix and this ‘information’                           12 Johnson, E. F., Hinz, W., Atreya, C. E., Maley, F. and Anderson, K. S. (2002) Mechanistic
can then be relayed to its own active site through the flexible                                characterization of Toxoplasma gondii thymidylate synthase (TS–DHFR)-dihydrofolate
linker. The information may be used to activate or co-ordinate                                reductase. Evidence for a TS intermediate and TS half-sites reactivity. J. Biol. Chem. 277,
activity with the other DHFR active site. This would be a unique                              43126–43136
type of domain–domain communication among the characterized                                13 Dwyer, D. S. (2001) Electronic properties of the amino acid side chains contribute
bifunctional TS–DHFR enzymes.                                                                 to the structural preferences in protein folding. J. Biomol. Struct. Dyn. 18,
   The mutational analysis and mechanistic information we have                                881–892
gained in the present study opens an exciting new avenue of                                14 Kneller, D. G., Cohen, F. E. and Langridge, R. (1990) Improvements in protein
investigation that may be exploited for inhibitor design. Ligands                             secondary structure prediction by an enhanced neural network. J. Mol. Biol. 214,
targeted to bind near or adjacent to the crossover helix may also                             171–182
interfere with domain–domain communication and thereby inhibit                             15 Grumont, R., Sirawaraporn, W. and Santi, D. V. (1988) Heterologous expression of the
                                                                                              bifunctional thymidylate synthase-dihydrofolate reductase from Leishmania major .
enzyme activity. A unique pocket is formed by the crossover
                                                                                              Biochemistry 27, 3776–3784
helix and the tethers that can be targeted for non-active site inhib-
                                                                                           16 Schlichting, I., Yang, X.-J., Miles, E. W., Kim, A. Y. and Anderson, K. S. (1994) Structural
itors using a molecular docking strategy coupled with virtual                                 and kinetic analysis of a channel-impaired mutant of tryptophan synthase. J. Biol. Chem.
library screening [24]. A non-active site TS–DHFR inhibitor                                   269, 26591–26593
would be advantageous because this site would be specific for the                           17 Atreya, C. E., Johnson, E. F., Williamson, J., Chang, S.-Y., Liang, P.-H. and Anderson,
bifunctional enzyme and therefore not inhibit the monofunctional                              K. S. (2002) Probing electrostatic channeling in protozoal bifunctional thymidylate
(host) enzymes. The successful use of non-active site inhibitors                              synthase-dihydrofolate reductase using site-directed mutagenesis. J. Biol. Chem. 278,
as a component of combination therapy is well-established in                                  28901–28911
the treatment of HIV infections [25]. A potential non-active site                          18 Agarwal, P. K., Billeter, S. R., Rajagopalan, P. T., Benkovic, S. J. and Hammes-Schiffer, S.
inhibitor of C. hominis could work in two ways: (i) by interfering                            (2002) Network of coupled promoting motions in enzyme catalysis. Proc. Natl.
directly with the crossover helix and helix B or (ii) by binding to                           Acad. Sci. U.S.A. 99, 2794–2799
a site adjacent to the crossover helix. These non-active site                              19 Chen, J. T., Taira, K., Tu, C. P. and Benkovic, S. J. (1987) Probing the functional role of
inhibitors could be used in conjunction with potential C. hominis                             phenylalanine-31 of Escherichia coli dihydrofolate reductase by site-directed
DHFR active-site inhibitors as combination therapy [26,27],                                   mutagenesis. Biochemistry 26, 4093–4100
                                                                                           20 Watney, J., Agarwal, P. and Hammes-Schiffer, S. (2003) Effect of mutation on
theoretically decreasing the onset of resistance mutations. De-
                                                                                              enzyme motion in dihydrofolate reductase. J. Am. Chem. Soc. 125, 3745–3750
velopment of effective species-specific inhibitors for C. hominis
                                                                                           21 Schirch, V. and Strong, W. B. (1989) Interaction of folylpolyglutamates with enzymes in
is essential as there are currently no effective treatments for                               one-carbon metabolism. Arch. Biochem. Biophys. 269, 371–380
cryptosporidiosis.                                                                         22 Strong, W. B., Tendler, S. J., Seither, R. L., Goldman, I. D. and Schirch, V. (1990)
                                                                                              Purification and properties of serine hydroxymethyltransferase and C1-tetrahydrofolate
                                                                                              synthase from L1210 cells. J. Biol. Chem. 265, 12149–12150
FUNDING                                                                                    23 Sawaya, M. R. and Kraut, J. (1997) Loop and subdomain movements in the mechanism of
This work was supported by the National Institutes of Health [grant numbers AI 44630 (to      Escherichia coli dihydrofolate reductase: crystallographic evidence. Biochemistry 36,
K. S. A.), 5T32-AI07404 (to W. E. M.)].                                                       586–603

                                                                                                                            c The Authors Journal compilation c 2009 Biochemical Society
764               M. A. Vargo, W. E. Martucci and K. S. Anderson

24 Atreya, C. E., Johnson, E. F., Irwin, J. J., Dow, A., Massimine, K. M., Coppens, I.,       26 Pelphrey, P. M., Popov, V. M., Joska, T. M., Beierlein, J. M., Bolstad, E. S., Fillingham,
   Stempliuk, V., Beverley, S., Joiner, K. A., Shoichet, B. K. and Anderson, K. S. (2002) A      Y. A., Wright, D. L. and Anderson, A. C. (2007) Highly efficient ligands for dihydrofolate
   molecular docking strategy identifies Eosin B as a non-active site inhibitor of protozoal      reductase from Cryptosporidium hominis and Toxoplasma gondii inspired by structural
   bifunctional thymidylate synthase-dihydrofolate reductase. J. Biol. Chem. 278,                analysis. J. Med. Chem. 50, 940–950
   14092–14100                                                                                27 Popov, V. M., Chan, D. C., Fillingham, Y. A., Atom Yee, W., Wright, D. L. and Anderson,
25 Basavapathruni, A. and Anderson, K. S. (2006) Developing novel non-nucleoside HIV-1           A. C. (2006) Analysis of complexes of inhibitors with Cryptosporidium hominis DHFR
   reverse transcriptase inhibitors: beyond the butterfly. Curr. Pharm. Des. 12,                  leads to a new trimethoprim derivative. Bioorg. Med. Chem. Lett. 16,
   1857–1865                                                                                     4366–4370

Received 17 June 2008/3 October 2008; accepted 13 October 2008
Published as BJ Immediate Publication 13 October 2008, doi:10.1042/BJ20081247

c The Authors Journal compilation c 2009 Biochemical Society

Shared By: