Biochemistry 811
Lecture on pyrimidine metabolism
D. Nelson, last modified Nov. 22, 2002 8AM
Reading: Berg, Tymoczko and Stryer 5th edition, Chapter 25
The pyrimidine bases are cytosine, uracil and thymine. Thymine is 5 methyl uracil.
Pyrimidine metabolism is illustrated on the KEGG pathway maps at Pyrimidine
Metabolism
Under file menu go to page setup and select landscape orientation at 65% to print.
For homo sapiens see Human Pyrimidine Metabolism or use the pull down menu.
The human KEGG map for pyrimidine metabolism has 60 shaded boxes representing 29
Enzymes, compared to 46 in the purine map. Nine of these were seen in the purine map.
20/29 enzymes (69%) cause human diseases.
The KEGG map for pyrimidine metabolism has the de novo pathway for
UMP synthesis at the top left. UMP is the starting point for making the other
pyrimidines. UMP occupies the same central position as IMP in purine metabolism.
The other pyrimidines include the deoxyribose forms for DNA synthesis as well as
the ribose forms for RNA synthesis. The rest of the figure is composed of five
rows. The first row shows the interconversions of the uridine nucleotides on the
left and the breakdown pathway on the right. The second row shows the interconversions
of the cytidine nucleotides and their breakdown pathway. The third row is cytidine
deoxynucleotide interconversions. To make the deoxynucleotides, the enzyme ribonucleotide
reductase is required. This is a very important enzyme since we could not make DNA without
it. The fourth row shows deoxyuridine nucleotides, and the fifth row shows the thymidine
nucleotides. Thymidine is only found in DNA. There is no thymine base in RNA,
so the synthesis of dTTP is a natural target for inhibition of DNA synthesis.
We will begin with the de novo pathway for UMP synthesis (pp. 694-697).
Pyrimidines are smaller than purines, and their biosynthesis is correspondingly easier.
The pyrimidine ring is composed from only two compounds, carbamoyl phosphate and
aspartate. Carbamoyl phosphate is made by the enzymes carbamoyl phosphate
synthetase I and carbamoyl phosphate synthetase II. CPS II is the one we are
interested in. It is in the cytosol, while CPS I is in the mitochondria and is
part of the urea cycle. Since bacteria do not have these compartments, they have only
one CPS enzyme that is used for both amino acid and pyrimidine metabolism.
Therefore it is a branchpoint in a pathway and not the committed step.
The CPS II reaction uses two ATPs (p. 695). The first is used to activate
bicarbonate to carboxyphosphate. In the second step, glutamine donates its amide nitrogen
to displace the phosphate and form carbamate with glutamine instead of ammonia as nitrogen
source (CPSI uses ammonia). In the final step, the second ATP reacts with the carbamate to
form carbamoyl phosphate and ADP. The second enzyme in the de novo pathway is aspartate
transcarbamoylase (ATCase). This step adds aspartate to carbamoyl phosphate. In E. coli,
the ATCase is one of the best studied allosteric enzymes. It is the committed step in bacteria
for pyrimidine biosynthesis, though CPS II is the committed step in eukaryotes.
Dihydroorotase closes the ring. These first three enzymes are all part of a
multifunctional protein called CAD after the first letter of each enzyme name.
The newly formed ring is then oxidized to form orotate. The dihydroorotate dehydrogenase
is one of the many enzymes that pass electrons to the electron transport chain at the level
of ubiquinone. At this point, the orotate base is coupled to PRPP releasing PPi to drive the
reaction forward. The final step is a decarboxylation by orotidine 5' monophosphate
decarboxylase (ODCase) to form UMP. ODCase is the well known yeast selectable
marker gene URA3 that is used on many yeast plasmids. The last two enzymes in the
pathway are part of a bifunctional protein. Defects in either of these last two enzymes cause
oroticaciduria (high levels of orotic acid in the urine, retarded growth and severe anemia).
Treatment with high doses of uridine in the diet corrects the disorder by supplying end products
of the pathway and shutting down the pathway by feedback inhibition. That brings us to the top
center of the KEGG map. Now we need to discuss formation of the di and triphosphate forms.
Uridine monophosphate kinase (UMPK) converts UMP to UDP, the general nucleoside diphosphate
kinase forms UTP. Remember from purine metabolism this enzyme also serves as a tumor suppressor
called NME1. If you look at the KEGG map, there is only one way to get to cytidine nucleotides
from the uridine nucleotides. UTP is converted to CTP by amidation of the 4 position. This is
done by CTP synthetase (p. 697). ATP is required and glutamine donates its amide nitrogen.
Azaserine, a glutamine antagonist, can inhibit this step. The enzyme apyrase converts
the CTP to CDP and CMP by hydrolysis.
Regulation of pyrimidine biosynthesis
As we mentioned earlier, the regulation is different in animals and E. coli
because the committed steps are different. In E. coli, the committed step is
aspartate transcarbamoylase, which is one of the best studied allosteric enzymes.
ATCase is an alpha 6 beta 6 dodecamer. The alpha subunits are catalytic and the
beta subunits are regulatory. ATP and CTP compete for the same site on the
regulatory subunits. ATP is an activator and CTP is an inhibitor. ATP is a purine
and CTP is a pyrimidine, so the ratio of ATP to CTP reflects the balance between
these types of nucleotides. When ATP is higher than CTP there is a need for more
pyrimidine synthesis so ATCase is activated. When CTP is higher, there is
probably not enough energy supply in the cell to be using it for nucleic acid
synthesis so the enzyme is inhibited.
In animals the CPS II enzyme is the committed step since the compartmentation of
CPS I into mitochondria separates the functions of these enzymes.
CPS II is dedicated to pyrimidine synthesis. The enzyme is feedback inhibited by UDP
and UTP, but not CTP. As in the E. coli pathway, ATP is an activator. PRPP
is also an activator in animals.
According to your book (Fig. 25.12, p. 703), the enzyme responsible for conversion of
all four nucleotides to deoxynucleotides (ribonucleotide reductase) only acts on the
diphosphate forms (in most organisms). In the KEGG map this is shown as EC
1.17.4.1. It is on both the purine and pyrimidine maps and acts on all four NDPs.
There is another enzyme in the KEGG maps called EC 1.17.4.2 that acts on all
four of the NTPs. Your book does not mention this enzyme. Looking it up in KEGG
shows it is an anaerobic enzyme in some bacteria. Your book did mention that there
are three different types of ribonucleotide reductases, the EC 1.17.4.2 enzyme is
one of the other types, a class II enzyme with dependence on an adenosylcobalamine
cofactor. The different forms of ribonucleotide reductases seem unrelated
to each other in their sequences. It was thought that they were derived from a
common ribozyme that was replaced by protein enzymes independently after the three
main branches of life (Archaea, Bacteria and Eukarya) separated. However, recent
evidence shows that their 3-dimensional structures are related and this in combination
with their rare free radical mechanism implies they did have a common protein ancestor,
not just a ribozyme that was replaced. In fact the argument is now in favor of an RNA
ribozyme not being able to catyalyze this free radical reaction so that protein had
to evolve before this enzyme and therefore DNA could ever be made.
At least in the E. coli class I enzyme, the conversion of cytidine nucleotides to
deoxycytidine nucleotides must pass through the diphosphate form. However, the
synthesis of cytidine nucleotides only proceeds through the UTP to CTP reaction,
so CDP is not available unless it is made by the hydrolysis reaction of apyrase.
The monophosphate can be formed by a similar step. That completes the formation
of all four ribose nucleotides and their 12 phosphorylated states.
Ribonucleotide reductase
In the last lecture, we did not talk about the formation of deoxypurine
nucleotides. The ribonucleotide reductase acts on all four NDPs, so we will
discuss them all now. The transformation of the ribose ring to deoxyribose at the
2' position is a reduction. An OH is replaced by and H. This requires a source of
electrons. As is often the case, the electrons for biosynthesis come from NADPH.
The electrons move to the ribonucleotide reductase in a short electron transfer
chain composed of thioredoxin and thioredoxin reductase. All three proteins have
paired sulfhydryl groups that undergo reversible oxidation to form a disulfide bond
that can be reduced back to free SH groups.
The E. coli enzyme is shown in Fig. 25.10 and 25.11 p.702. The intact enzyme
is an alpha 2 beta 2 tetramer. Even though the molecule has twofold symmetry, there
appears to be only one active site, made up of the two iron ions and one of the
active site pockets formed by the interface of the alpha and beta subunits. This
is not unheard of. The light reaction center of bacteria has been shown to have a
symmetric array of electron transfering prosthetic groups, but only one side is
actually used.
There is an essential tyrosine in the active site that plays a role in the
mechanism. If phenylalanine is substituted for tyr, the enzyme is dead. It is
thought the OH group on tyr forms a free radical that abstracts a hydrogen from the
3' carbon. (Fig. 25.12, p. 703) then a base donates a proton to the the 2' OH to form a
water that leaves. There is an electrical rearrangment that shifts the free radical to
the 2' carbon. At this point the intermediate is quite oxidized. The paired sulhydryl
groups now become oxidized to form a disulfide and they give up electrons and protons to
the intermediate. The result is the net replacement of the 2' OH group with a hydride
ion to form the deoxynucleotide. At the start of the reaction, the sulfhydryls
were in a reduced state. At the end, they are oxidized. To restore the enzyme to
its initial state, the disulfide must be reduced. This is done by the thioredoxin.
During this interaction, the thioredoxin sulfhydryls become oxidized and the ribonucleotide
reductase becomes reduced. This only moved the problem to another protein. The
thioredoxin must now be reduced, and this is done by thioredoxin reductase. The
thioredoxin reductase contains an FAD that can receive two electrons from NADPH to
reduce the disulfide in it, and this completes the cycle. An alternative to thioredoxin is
glutaredoxin. Glutaredoxin can undergo the same sulfhydryl to disulfide transition as
thioredoxin. It is then reduced by two glutathione molecules. These are tripeptides that
contain cysteine in reduced form (Fig. 24.30, p. 686). They can become
disulfide linked forming oxidized glutathione GSSG, while glutaredoxin becomes reduced again.
The GSSG can be reduced to glutathione by glutathione reductase, which is a flavoprotein that
uses NADPH.
Ribonucleotide reductase is closely regulated so that even amounts of the deoxynucleotides
are produced. As shown in your handout and described on p. 708, there are two
regulatory sites on each alpha subunit. The activity determining site binds either ATP or
dATP. These two molecules compete for binding. If ATP is bound this turns the enzyme on.
ATP is a signal that there is adequate energy available to make DNA. When dATP is bound the
enzyme is shut off. This is a case of feedback inhibition. Remember that dATP accumulates in
adenosine deaminase deficiency and causes SCID by this mechanism. The second site is the
specificity determining site. Four different nucleotides compete for binding at this site,
ATP, dATP, dGTP and dTTP. Each one changes the specificity of the active site.
There is a logic to the selection of the preferred substrates. This logic is
outlined on p. 708. If we assume that the concentration of all dNTPs is zero when we
start, ATP will bind at both the activity site, turning on the enzyme and at the specificity
site. This will set the preference for pyrimidine diphosphates CDP and UDP. As the
concentration of dCTP and dUTP rises, these are the precursors to dTTP so dTTP concentrations
will rise. dTTP will begin to compete with ATP for the specificity site and this
sets the preference for purine nucleotide diphosphates. This causes more dGTP and
dATP to be made. As dGTP concentrations rise the specificity is narrowed to just
ADP and this makes the concentration of dATP rise. Now all five of the nucleotides
have been made and when dATP concentrations are high enough the enzyme is shut
down. Remember that the products of ribonucleotide reductase are dNDPs and they
must be converted to dNTPs by the general nucleoside diphosphate kinase for use in
DNA synthesis.
Quiescent cells do not make ribonucleotide reductase subunits R1 or R2. Growing cells
make both subunits. Cells with DNA damage make an alternative R2 subunit that is induced
by the p53 tumor suppressor protein. This p53R2 subunit is directly involved in the DNA
repair process.
Making thymidine nucleotides
Thymine's only function is to make DNA. It is not used anywhere else in the cell.
If you look at the KEGG map, you will see that there is only one way to make
thymine nucleotides, and that goes through dUMP by the action of thymidylate
synthase. Remember in the de novo pathway for pyrimidines, the first
product is UMP. This cannot be directly converted to any cytidine nucleotide. It
has to be converted in two steps to UTP first, then it can be converted to CTP by
CTP synthase. In humans and aerobic E. coli, the CTP cannot be directly converted
to dCTP. It must first be changed to CDP and then acted on by ribonucleotide
reductase to make dCDP. The process of converting pyrimidine nucleotides seems
to be like a maze with only one door open on each level, and they never seem to be
at the same state of phosphorylation. This means that the phosphorylation state
has to be changed at each level to get to the next level.
From dCDP, the phosphorylation state must again be changed to dCMP before
moving to the next level as dUMP. Finally dUMP can be converted directly to
dTMP by thymidylate synthase. dTMP is then phosphorylated twice to make dTTP.
The shortest path to dTTp is UMP to UDP to dUDP to dUMP to dTMP to dTDP to dTTP.
There is a longer path but this neglects ribonucleotide reductase acting on UDP
(see KEGG map bottom left).
dUTP is a deoxyribose nucleotide, and the DNA polymerase can incorporate it
into DNA, so the concentration of dUTP must be kept fairly low. Two enzymes
act to reduce the concentration of dUTP. One is the general NDP kinase
operating in reverse mode. The other is a special dUTPase that knocks off
pyrophosphate to make dUMP. Why worry about incorporation of dUTP into
DNA? After all, U and T both pair with A, so this is no problem (see pp. 771-772).
That is true until there is a deamination of cytosine to form uridine in DNA. This
happens at a low frequency in the cell. The repair mechanism assumes that a U:G pair
is due to the deamination of C and it restores the C base. If dUTP is incorporated
into DNA, then U:A is accepted as normal and U:G is abnormal as before, except the
repair system cannot decide what should be present in the DNA, should a U:G pair
be replaced by C:G or U:A? That is why dUTP is kept very low in the cell, to
prevent this potential mutation problem.
The synthesis of dTMP from dUMP by thmidylate synthase uses a tetrahydrofolate
intermediate N5N10methyleneTHF. The reaction shown in your handout and in Fig. 25.14, p. 706
results in the oxidation of THF to dihydrofolate. Once the methylene is transfered, the
DHF must be reconverted to THF again by dihydrofolate reductase. (p. 705) The inhibitors
methotrexate, methopterin, trimethoprim and aminopterin that we mentioned in the purine de
novo pathway (p. 707), are also inhibitors of dTMP synthesis because they block
the reformation of N5N10methyleneTHF. In the purine de novo pathway, N10 formyl THF was
used resulting in THF as a product. Dihydrofolate reductase (DHFR) is not required
to reform N10 formyl THF. Why does a DHFR inhibitor block purine biosynthesis?
Because the N10 formyl THF can be converted to N5N10 methyleneTHF to
supply thymidylate synthase. This eventually depletes the N10 formyl THF and
blocks purine biosynthesis. Serine hydroxymethyltransferase (SHMT), a pyridoxal phosphate-
containing enzyme, catalyzes the reversible conversion of serine and tetrahydrofolate to glycine
and 5,10-methylene tetrahydrofolate. It acts just after dihydrofolate reductase. Folate-
dependent one-carbon metabolism is critical for the synthesis of numerous cellular constituents
required for cell growth, and SHMT and DHFR are central to this process.
Resistance often develops to drugs and methotrexate resistance has been studied.
A large number of independent methotrexate-resistant human cell lines all contained
DHFR-bearing double minutes (DMs). DMs are small extrachromosomal elements. By making
35-275 copies of the DHFR gene on these DMs, methotrexate resistance can occur.
See photo p. 693.
The sulfonamide drugs like sulfanilamide are analogs of the para-aminobenzoic acid part of THF.
These drugs block synthesis of folate in bacteria, but have no effect on humans since we do not
make folate. Apparently bacteria cannot acquire folate from their host and so these drugs kill
them.
Fluoro compounds and treatment of cancer
Enzyme mechanisms often require the removal of a proton from a specific site on a
substrate. Fluorine is so electronegative that it makes a very poor leaving group.
It just won't come off as F+. By substituting F for H at a critical site, the
enzyme can become trapped in a ternary complex that won't release the fluorine.
This is illustrated in Fig. 25.15, p. 706. Thymidylate synthase is needed for DNA
synthesis. 5 fluoro uracil gets converted to 5 fluoro-dUMP by the
normal route, but thymidylate synthase cannot extract the F and it becomes
permanently inhibited. This blocks DNA synthesis. Because the inhibition is
dependent on the enzyme mechanism these types of inhibitors are called mechanism
based inhibitors or suicide substrates. 5 fluorouracil is used to treat some
cancers. It should act similarly to methotrexate, except it will be specific for
dTMP synthesis and it won't affect TFH that is needed by other enzymes.
Cancers often develop drug resistance. This is also true for 5 fluorouracil. However,
resistance does not occur at thymidylate synthase. 5 fluorouracil enters the pyrimidine
metabolic map at uracil and has to follow all the way through to dUMP before it reaches
thymidylate synthase. This is a series of nine enzymes. A key enzyme for resistance is CTP
synthetase that converts UTP to CTP. Point mutations at 7 locations in the CTP synthetase gene
have been found that confer drug resistance to 5 fluorouracil and arabinosyl cytosine (ARA C).
These mutants apparently discriminate between the natural substrate and the drugs and
selectively exclude the drugs from the active site.
Salvage pathways for pyrimidines
Free pyrimidine bases can be recovered as nucleotides by a salvage pathway.
Uracil and cytosine can be converted back to nucleosides by addition of ribose.
The nucleosides or deoxynucleosides can be phosphorylated by uridine kinase, to make UMP and CMP
(uridine kinase acts on both uridine and cytidine) deoxycytidine kinase to make dCMP and thymidine
kinase to make dTMP. Note that the free bases are not coupled directly
back to deoxyribose, since there is no equivalent of a deoxyribose PRPP.
Thymidine kinase is an important enzyme because of certain applications
that make use of it (see monoclonal antibodies and knockout mice below).
The drug arabinosyl cytosine (ARA-C) is used in the treatment of acute myeloid leukemia, and the
antileukemic activity of the drug depends on deoxycytidine kinase (a salvage pathway enzyme) to
phosphorylate it. It is eventually converted to the triphosphate form CTP arabinoside which
interferes with DNA synthesis at the DNA polymerase. Drug resistance to cytosine arabinoside is
one of the main problems in the successful treatment of acute myeloid leukemia. Resistance to
the drug has been ascribed to functional deoxycytidine kinase deficiency and to increased
expression of the cytosine deaminase gene which converts it to an inactive uridine form.
Thymidine kinase and monoclonal antibody production
Monoclonal antibodies are made by fusing B cells with myeloma cells to make a
hybridoma cell line that grows better than a B cell can grow and it produces the
monoclonal antibody of the B cell. The difficulty is selecting for the hybridomas,
since the majority of cells after the fusion are myeloma cells that did not fuse.
The solution to this problem was to exploit purine and pyrimidine salvage pathways
in a selection scheme. Kohler and Milstein used myeloma cells that were unable to
use the purine salvage pathway because they were defective in HGPRT
(hypoxanthine guanine phosphoribosyl transferase) When these cells were treated
with amethopterin (a THF analog) they died because purine and dTMP synthesis
was blocked. If they were given hypoxanthine and thymidine they still died
because the HGPRT salvage pathway for hypoxanthine was also blocked.
B cells that had a functioning HGPRT could survive in the HAT medium
(hypoxanthine, amethopterin and thymidine) because they could make dTMP with
thymidine kinase from thymidine, they could make purines by HGPRT salvage and
thus they got around the amethopterin block. However B cells grow poorly in
culture. The hybridomas grow very well, and they will also survive because they
gain the HGPRT enzyme from the B cells. Thus, the hybridoma cells outgrow the
B cells and the unfused myeloma cells die.
Salvage pathways and knockout mice
Making knockout mice requires homologous recombination between a target gene
and a plasmid construction that carries the knockout and some selectable markers.
During the process, there can be some integration of the plasmid at non-
homologous sites (called non-homologous recombination). This is an undesired
outcome and needs to be avoided. Thymidine kinase can be used here to select
against non-homologous recombination. The herpes simplex virus thymidine
kinase HSVtk can phosphorylate a thymidine analog FIAU, but animal cells cannot
do this. The product gets converted into FIAUTP which is toxic. Therefore, any
animal cells that express the HSVtk will be killed by FIAU. In making knockout
mice, the plasmid used for recombination can have an HSVtk added to it. If the
recombination is at the right site and it occurs by homologous recombination the
HSVtk gene is deleted and the cells are not affected by FIAU. If the plasmid
integrates by non-homologous recombination then the HSVtk is present and it will
make the cells susceptible to FIAU. This same trick can be used in studying
development. If a cell specific promoter is placed in front of the HSVtk gene,
then the HSVtk will only be expressed in those cells, causing them to die. This
can be used to study the consequences of deleting certain cell types at early
stages of development. One can also use the diphtheria toxin gene instead of HSVtk
This eliminates the need for FIAU.
Breakdown pathways for pyrimidines
The monophosphate forms of UMP, dUMP, CMP, dCMP and dTMP can be dephosphorylated by
5 prime nucleotidase (3.1.3.5 on your handout). This creates the nucleosides uridine, cytidine,
thymidine, deoxyuridine and deoxycytidine. The nucleosides can be broken into ribose 1
phosphate and free bases by uridine phosphorylase and thymidine phosphorylase. Cytidine and
deoxycytidine are deaminated to uridine and deoxyuridine before they are processed further. The
end products of these reactions are uracil and thymine. These are further broken down to TCA
cycle intermediates acetyl CoA (uracil) and succinyl CoA (thymine). (you are not responsible
for the pyrimidine breakdown pathways)
Modifications of bases in tRNA and the occurrence of modified bases in plants
tRNAs have many modification to the usual nucleotide bases(p. 790). Do not memorize these
structures. They are included to make you aware the tRNAs use a variety of non standard bases,
while DNA does not.
Two favorite beverages around the world are tea and coffee. Another favorite flavoring is
chocolate. These three food and drink items all contain methyl xanthines (shown in your
handout). Chocolate contains all three, however, caffeine concentrations in chocolate
(approximately 22 mg/100 g of milk chocolate compared to 80-100 mg in a cup of coffee) are too
low to have a stimulatory effect on consumers. Theobromine (at 197 mg/ 100 g of milk chocolate)
is the most abundant methylxanthine in chocolate. Theophylline is only present in trace
amounts. It seems unlikely that the concentrations of methylxanthines in chocolate are high
enough to induce significant physiological effects. These compound are purine alkaloids and
they are bitter tasting. The bitter flavor of dark chocolate is due to the higher concentration
of these compounds.
Coffee has caffeine and tea has theophylline. Both of these act as inhibitors of the
phosphodiesterase that breaks down cAMP. This is why they are stimulants. The lethal dose for
caffeine is 150 mg/kg for dogs, cats, and people. (11 grams for a 165 pound person or about 100
cups of coffee).
Theobromine is toxic to dogs. Theobromine when ingested by dogs causes release of epinephrine
(adrenaline) which causes the heart to race and serious cardiac arrhythmias to develop.
The toxic dose in the dog is 100-150 mg/kg The concentration of theobromine varies with the
formulation of the chocolate so: Milk chocolate has 44mg/oz (154mg/100gm): toxic dose for 60 lb
dog - 60 oz of milk chocolate. Semisweet chocolate has 150 mg/oz (528mg/100gm): toxic dose for
60 lb dog - 18 oz of semisweet chocolate Baking chocolate 390mg/oz (1365 mg/100gm): toxic dose
for 60 lb dog - 6 oz of baking chocolate
Theophylline is a bronchodilator drug used to treat people with asthma.
Glutamine analogs as inhibitors
Many of the pathways we have seen involve donation of an amino group from glutamine.
The glutamine analogs azaserine or DON (see handout)
form covalent bonds at the active sites of enzymes that bind glutamine. This has the effect of
blocking the pathways. In purine biosynthesis steps 2 and 5 of the de novo pathway (see handout)
and GMP synthase are affected. In pyrimidine biosynthesis CPSII in the de novo pathway and CPT
synthase in the nucleotide interconversion pathways are blocked. The synthesis of NAD from
nicotinic acid is also a glutamine dependent process.
AZT a nucleotide based inhibitor for AIDS
AZT is a thymidine analog (see handout). It is converted by
thymidine kinase (a salvage pathway enzyme) to the monophosphate form. This is further
phosphorylated to the triphosphate form by thymidylate kinase and NDP kinase. This AZTTP is an
inhibitor of the reverse transcriptase of the AIDS virus. It also gets incorporated into the new
DNA strand and has no 3 prime OH to continue the chain. Drug resistance develops, so modern
treatments use a cocktail of protease inhibitors and DNA synthesis inhibitors to stop virus
replication.