The conversion of a phenol to an aniline occurs in the biochemical formation of the 1-(4-aminophenyl)-1-deoxy- d -ribitol moiety in methanopterin

Article abstract of DOI:10.1021/bi200362w

Recent work has demonstrated that 4-hydroxybenzoic acid is the in vivo precursor to the 1-(4-aminophenyl)-1-deoxy-d-ribitol (APDR) moiety present in the C₁ carrier coenzyme methanopterin present in the methanogenic archaea. For this transformation to occur, the hydroxyl group of the 4-hydroxybenzoic acid must be replaced with an amino group at some point in the biosynthetic pathway. Using stable isotopically labeled precursors and liquid chromatography with electrospray-ionization mass spectroscopy, the first step of this transformation in Methanocaldococcus jannaschii occurs by the reaction of 4-hydroxybenzoic acid with phosphoribosyl pyrophosphate (PRPP) to form 4-(β-d-ribofuranosyl)hydroxybenzene 5′-phosphate (β-RAH-P). The β-RAH-P then condenses with l-aspartate in the presence of ATP to form 4-(β-d-ribofuranosyl)-N-succinylaminobenzene 5′-phosphate (β-RFSA-P). Elimination of fumarate from β-RFSA-P produces 4-(β-d-ribofuranosyl)aminobenzene 5′-phosphate (β-RFA-P), the known precursor to the APDR moiety of methanopterin [White, R. H. (1996) Biochemistry 35, 3447-3456]. This work represents the first biochemical example of the conversion of a phenol to an aniline.

Full text of DOI:10.1021/bi200362w

ARTICLE
pubs.acs.org/biochemistry
The Conversion of a Phenol to an Aniline Occurs in the Biochemical
Formation of the 1-(4-Aminophenyl)-1-deoxy-D-ribitol Moiety
in Methanopterin
Robert H. White*
Department of Biochemistry, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0308, United States
ABSTRACT: Recent work has demonstrated that 4-hydroxybenzoic
acid is the in vivo precursor to the 1-(4-aminophenyl)-1-deoxy-D-ribitol
(APDR) moiety present in the C1 carrier coenzyme methanopterin
present in the methanogenic archaea. For this transformation to occur,
the hydroxyl group of the 4-hydroxybenzoic acid must be replaced with
an amino group at some point in the biosynthetic pathway. Using stable
isotopically labeled precursors and liquid chromatography with electro-
spray-ionization mass spectroscopy, the first step of this transformation
in Methanocaldococcus jannaschii occurs by the reaction of 4-hydroxybenzoic acid with phosphoribosyl pyrophosphate (PRPP) to
form 4-(β-D-ribofuranosyl)hydroxybenzene 50-phosphate (β-RAH-P). The β-RAH-P then condenses with L-aspartate in the
presence of ATP to form 4-(β-D-ribofuranosyl)-N-succinylaminobenzene 50-phosphate (β-RFSA-P). Elimination of fumarate from
β-RFSA-P produces 4-(β-D-ribofuranosyl)aminobenzene 50-phosphate (β-RFA-P), the known precursor to the APDR moiety of
methanopterin [White, R. H. (1996) Biochemistry 35, 3447ꢀ3456]. This work represents the first biochemical example of the
conversion of a phenol to an aniline.
etrahydromethanopterin (H4MPT) and tetrahydrofolate
(H4folate) represent the functional forms of two different
hydroxyphenylacetic acid into tyrosine but not into the APDR
moiety of MPT.14 These observations led to the conclusion that
3-dehydroquinate (DHQ), not chorismate, was the precursor to
AB. A possible alternate pathway for the biosynthesis of AB from
DHQ is shown in Figure 2. Attempts to confirm the operation of
this pathway by incubating cell extracts with L-asparatate semi-
aldehyde and 6-deoxy-5-ketofructose 1-phosphate, DHQ, and
4-amino-5-hydroxy-1,3-cyclohexadiene-1-carboxylic acid and
assaying for an increase in the amount of AB were all proven
negative. One idea not considered in this earlier work was that
chorismate-derived 4-hydroxybenzoic acid (HB) normally
serves as a precursor to APDR and that HB could also be
derived from DHQ. We have recently found that the protein
product of the MJ0807 gene is a chorismate lyase and
produces HB but not AB from chorismate (L. Grochowski
and R. H. White, unpublished results). This finding and the
observed incorporation of the aromatic portion of the HB
molecule into the phenyl portion of the APDR moiety of MPT
indicated that HB is the true in vivo precursor to the APDR
in MPT. Here I show that the first step in this process occurs
by the previously observed condensation of PRPP with HB
to produce 4-(β-D-ribofuranosyl)hydroxybenzene 50-phos-
phate (β-RFH-P)15 catalyzed by the product of the MJ1427
gene. The phenol in the β-RFH-P is then converted to an
aniline to generate β-RFA-P. Here I present evidence of how
this occurs.
T
coenzymes that function as biological C1 carriers (Figure 1).
These coenzymes were both identified and characterized in their
oxidized forms, methanopterin (MPT) and folate, respectively.
MPT was first identified as a coenzyme in the methanogenic
archaea1 and was later identified in bacteria.2 Early experiments
showed that the pterin portion of both of these coenzymes arose
from GTP3ꢀ5 and that the arylamine moieties in both of the
structures could arise from 4-aminobenzoic acid (AB).5,6
These early observations indicated that these coenzymes
could be modified versions of each other. However, despite
their similar structures at the C1 binding sites,7 our current
data on the biosynthesis of these cofactors indicate that they
are not related.8
Despite the fact that the 1-(4-aminophenyl)-1-deoxy-D-ribitol
(APDR) of methanopterin was derived from AB when supplied
to growing cultures of methanogens, no evidence has ever been
obtained that AB is biosynthesized by the methanogens. This is
consistent with the absence in archaeal genomes of the pabA,
pabB, and pabC genes known to be involved in AB biosynthesis
in the bacteria.9ꢀ11 Although the first two steps in the shikimic
acid pathway are modified in the archaeal 6-deoxy-5-ketofructose
1-phosphate (DKFP) pathway,12 the pathway still produces
chorismate, the metabolite used to produce AB using these gene
products. A deletion mutant of AroA0, which encodes the first
step in the DKFP pathway, in Methanococcus maripaludis
was found to require both the aromatic amino acids and
AB for growth,13 yet a 3-dehydroquinate dehydratase (AroD)
deletion mutant did not require AB and incorporated labeled
Received: March 10, 2011
Revised:
June 1, 2011
Published: June 02, 2011
r
2011 American Chemical Society
6041
dx.doi.org/10.1021/bi200362w Biochemistry 2011, 50, 6041–6052
|

Biochemistry
ARTICLE
primers MJ1427-Fwd (50-GGTCATATGATAATTCAAACAC-
CATCG-30) and MJ1427-Rev (50-GCTGGATCCTCACCAA-
ATTTTATG-30). PCR amplification was performed as de-
scribed previously8 using an annealing temperature of 55 °C.
Purified PCR product was digested with NdeI and BamHI
restriction enzymes and ligated into compatible sites in plasmid
pT7-7. The sequence of the resulting plasmid pMJ1427 was
verified by DNA sequencing. pMJ1427 was transformed into
E. coli strain BL21-CodonPlus(DE3)-RiL (Stratagene). The
transformed cells were grown in Luria-Bertani medium (200 mL)
supplemented with 100 μg/mL ampicillin at 37 °C with shaking
until they reached an OD600 of 1.0. Production of the recombi-
nant protein was induced by addition of lactose to a final
concentration of 28 mM.8 After being cultured for an additional
2 h, the cells were harvested by centrifugation (4000g for 5 min)
and frozen at ꢀ20 °C. Induction of the desired protein was
confirmed by sodium dodecyl sulfateꢀpolyacrylamide gel elec-
trophoresis of total cellular proteins. The frozen E. coli cell pellet
(∼0.4 g wet weight from 200 mL of medium) was suspended in
3 mL of extraction buffer and lysed by sonication. The samples
were then centrifuged (16000g for 10 min) to remove cell debris,
and the resulting cell extract was heated for 10 min at 70 °C
followed by centrifugation (16000g for 10 min). This process
allowed for the purification of the protein from the majority of
E. coli proteins, which denature and precipitate under these
conditions. The next step of purification was performed by
anion-exchange chromatography of the 70 °C soluble fractions
on a MonoQ HR column (1 cm ꢁ 8 cm; Amersham Bioscience)
using a linear gradient from 0 to 1 M NaCl in 25 mM TES buffer
(pH 7.5), over 55 min at a flow rate of 1 mL/min and collecting
1 mL fractions. Protein concentrations were determined by
Bradford analysis.18
Enzymatic Synthesis of 4-(β-D-Ribofuranosyl)hydroxy-
benzene 50-Phosphate (β-RAH-P). To 130 μL of extraction
buffer containing 7.8 mM HB and 7.8 mM phosphoribosyl pyro-
phosphate (PRPP) was added 12 μg of recombinant MJ1427
protein, and the mixture was incubated at 70 °C for 0.5 h. After
the sample had cooled to room temperature, 8 μL of 2 M
trichloroacetic acid (TCA) was added and the sample centri-
fuged (16000g for 5 min) to remove the precipitated protein. The
TCA was neutralized with 1 M NaOH, and HPLC analysis of a
20 μL portion of the sample was performed. Chromatographic
separation and analysis were preformed on a Shimadzu HPLC
system equipped with a diode array, a fluorescence detector, and
a C18 reverse phase column (Varian PursuitXRs, 250 mm ꢁ 4.6
mm, 5 μm particle size). The solvent system consisted of 95%
sodium acetate buffer (25 mM, pH 6.0, 0.02% NaN3) and 5%
methanol for 5 min followed by a linear gradient to 20% sodium
acetate buffer and 80% methanol over 40 min at a rate of 0.5 mL/min.
The fluorescence intensity was measured with excitation at
284 nm and emission at 310 nm. The fluorescence data were
quantitated using a known sample of 4-hydroxybenzyl alcohol
separated under the same conditions used for the reference. It
was assumed that the quantum yield of fluorescence for both
compounds is the same.
Figure 1. Chemical structures of the biologically active forms of the
coenzymes tetrahydromethanopterin and tetrahydrofolate.
’ MATERIALS AND METHODS
Chemicals. [2,3,5,6-2H4]-4-Hydroxybenzoic acid (minimum
2
of 98 atom % H), [15N]ammonium chloride (minimum of
98 atom % 15N), [amide-15N]glutamine (minimum of 98
atom % 15N), [15N]-L-aspartic acid (minimum of 98 atom %
15N), 4-aminobenzoic acid, 4-aminophenethyl alcohol, adenylo-
succinic acid, (()-bromosuccinic acid, and phosphoribosyl
pyrophosphate (PRPP) were obtained from Sigma-Aldrich. Water
(minimum of 95 atom % 18O) was obtained from Monsanto
Research Corp. β-RFA-P was prepared enzymatically from the
reaction of AB with PRPP catalyzed by the MJ1427 gene product.
APDR-P was prepared by phosphorylation of APDR as pre-
viously described.16 APDR was either isolated from cell extracts
or prepared synthetically.17
Testing the Chemical Conversion of 4-Hydroxybenzyl
Alcohol to an Aniline. Two separate experiments were con-
ducted to test this idea. The first measured the exchange of the
oxygen(s) of 4-hydroxybenzyl alcohol with 18O-labeled water,
and the second tested for the replacement of the hydroxyl
groups(s) in 4-hydroxybenzyl alcohol with ammonia. In the first
experiment, 4-hydroxybenzyl alcohol (0.5 mg) was heated for
12 h at 100 °C with 30 μL of 95% 18O-enriched water. The water
was evaporated, and the sample reacted with trifluoroacetic
anhydride to form the ditrifluoroacetyl derivative of 4-hydroxy-
benzyl alcohol that was then examined by GCꢀMS to determine
the extent of oxygen exchange. In the second experiment, 15 mg
of 4-hydroxybenzyl alcohol was dissolved in 2 mL of 16 M
ammonia and heated at 100 °C in a sealed bottle for 24 h. After
cooling, the sample was evaporated to dryness with a stream of
nitrogen gas, and the resulting residue was extracted with dilute
HCl. The amino-containing compounds were isolated on a
Dowex 50 50W-X8 Hþ column and analyzed by GCꢀMS as
their ditrifluoroacetyl derivatives.
Cell Extracts of Methanocaldococcus jannaschii. Cell ex-
tracts of M. jannaschii were prepared by sonication under argon
and stored under argon at ꢀ20 °C at a protein concentration of
∼30 mg/mL.12 The buffer used in the extraction consisted of
50 mM TES/Kþ, 10 mM MgCl2, and 20 mM DTT (pH 7.5).
Cloning, Overexpression, and Purification of the M. jan-
naschii M1427 Gene in Escherichia coli. The MJ1427 gene
(Swiss-Prot accession number Q58822) was amplified by polymerase
chain reaction (PCR) from genomic DNA using oligonucleotide
Enzymatic Synthesis of 4-(β-D-Ribofuranosyl)aminobenzene
50-Phosphate (β-RFA-P). To 60 μL of extraction buffer without
DTT containing 0.02 M PRPP and 0.02 M μmol of AB was
added 12 μg of recombinantly produced MJ1427 protein, and the
samples were incubated at 70 °C for 1 h. After the addition of
150 μL of water to the samples, the azo dye derivatives were pre-
pared as previously described.19,20 Because of the large amount of
6042
dx.doi.org/10.1021/bi200362w |Biochemistry 2011, 50, 6041–6052

Biochemistry
ARTICLE
Figure 2. Hypothetical alternate route to AB.
unreacted AB, some of the azo dye derivative of the AB
precipitated and was removed by centrifugation prior to separa-
tion on the BioGel P-4 (200ꢀ400 mesh) column (0.5 cm ꢁ
5.0 cm) as described below. This separation completely sepa-
rated the β-RFA-P azo dye from the AB azo dye and allowed the
measurement of the yield of the reaction and identification of the
product by absorbance spectroscopy. LCꢀESI-MS analysis was
also used to confirm that the desired product was obtained.
Chemoenzymatic Synthesis of 4-(β-D-Ribofuranosyl)-N-
succinylaminobenzene 50-Phosphate (β-RFSA-P). Model re-
actions demonstrated that (()-bromosuccinic acid readily alky-
lated aniline to form N-phenylaspartate when the two com-
pounds were heated together in equal molar amounts at 100 °C
for 20 min. The same product could also be obtained when these
two compounds were heated together at 100 °C at neutral pH in
water. In each case, the product was identified by DI-MS as the
dimethyl trifluoroacetyl derivative (Mþ = m/z 333). Heating
solutions of 4-aminophenylalanine with (()-bromosuccinate
was used to determine the optimal condition for the alkylation.
The progress of the reaction was followed by TLC with UV and
ninhydrin detection of the products. Optimal formation of the
N-succinyl adduct was achieved by heating 25 mM 4-aminopheny-
lalanine and 150 mM (()-bromosuccinate adjusted to pH 7ꢀ8
at 100 °C for 0.5 h. Analysis of the resulting product by TLC
showed one major new UV and ninhydrin positive band. ESI-MS
analysis of the sample showed a trace amount of the starting
4-aminophenylalanine MHþ = m/z 181.3 and MNaþ = m/z
203.2 ions, as well as some of the dialkylated product with MHþ =
m/z 413.1 and MNaþ = m/z 435.1 ions. The major signal was from
the N-monoalkylated product showing the following ions: MHþ =
m/z 297.3, MNaþ = m/z 319.3, and MKþ = m/z 335.1. It is
assumed, on the basis of the pH of the reaction, that only the aniline
nitrogen was alkylated with the bromosuccinate.
A sample of 4-(β-D-ribofuranosyl)hydroxybenzene 50-phos-
phate (β-RFA-P) was prepared as described above but on a
10-fold larger scale. Because the sample contained a large amount
of unreacted AB, this was removed by passing the sample, made
0.1% in formic acid, through a 0.55 cm ꢁ 9.0 cm column con-
taining preparative C18 125A resin (55ꢀ105 μm, Waters)
equilibrated with 0.1% formic acid. The first material to elute
(1ꢀ3 mL) was the β-RFA-P, with the AB being completely
retained on the column. The β-RFA-P fractions were evaporated
to dryness with a stream of nitrogen gas to remove the formic
acid, and the residue was dissolved in 200 μL of water and
adjusted to pH 7ꢀ8 with NaOH. To this sample was added
50 μL of 0.2 M (()-bromosuccinate, and the sample was
heated at pH 7ꢀ8 for 30 min at 100 °C. Analysis of the presence
of an arylamine by the BrattonꢀMarshall assay showed a 95%
reduction in the amount of β-RFA-P. The first incubation
(Table 3) was conducted using a portion of this crude prepara-
tion of β-RFSA-P.
To ensure complete removal of the β-RFA-P from the sample
as well as any fumarate that may have been produced in the
reaction, the sample was purified by FPLC on a MonoQ HR 5/5
column using a 0 to 1 M NH4HCO3 gradient. In this gradient, the
β-RFA-P eluted at 0.5 M NH4HCO3 and the β-RFSA-P eluted at
0.9 M NH4HCO3, approximately the same positions of a known
sample of adenylosuccinate. This is expected because both
compounds contained four negative charges. The production
of the desired compound was identified by its 245 nm absorbance
and DI-ESI-MS in 0.1% HCOOH or 0.1% NH4OAc. DI-ESI-MS
showed in þESI the expected MHþ = m/z 422.0, MNaþ = m/z
444.0 and in ꢀESI the expected (M ꢀ H)ꢀ = m/z 420.1, (M ꢀ
2H þ Na)ꢀ = m/z 442.3. Adenylosuccinate showed in þESI the
expected MHþ = m/z 464.2, MNaþ = 486.1, MKþ = m/z 502.0
and in ꢀESI the expected (M ꢀ H)ꢀ = m/z 462.2, (M ꢀ 2H þ
6043
dx.doi.org/10.1021/bi200362w |Biochemistry 2011, 50, 6041–6052

Biochemistry
ARTICLE
Figure 3. LCꢀESI-MS data and UV analysis of sample 8 in Table 2: (A) UV trace of the separated sample at 545 nm, (B) m/z 425 ion trace for APDR,
and (C) m/z 503 ion trace for R-RFA-P and β-RFA-P.
Na)ꢀ = m/z 444.3, (M ꢀ 2H þ K)ꢀ = m/z 500.0. The total yield
(0.2 μmol) was 5% on the basis of the amount of PRPP used.
Because this product was produced with (()-bromosuccinate, it is
expected to consist of two stereoisomers at the position of the addition
of the succinate. Although the yield was low, sufficient product was
purified to determine if it served as a precursor to β-RFA-P.
Measurement of β-RAH-P Synthesis in Cell Extracts. To
70 μL of cell extract were added 10 μL of 0.1 M PRPP and 10 μL
of 0.1 M HB, and the mixture was incubated at 70 °C for 15 min.
After the sample had cooled to room temperature, 10 μL of 2 M
trichloroacetic acid was added and the sample centrifuged
(16000g for 10 min) to remove the protein. The pellet was
washed by suspension in 44 μL of 0.2 M TCA and centrifuged
(16000g for 10 min), and the combined supernatants were
adjusted to pH 7.0 by the addition of 6 M NaOH prior to HPLC
analysis. A Varian PursuitXRs 250 mm ꢁ 4.6 mm, 5 μm particle
size column was used for the separation using a solvent system
that consisted of 95% sodium acetate buffer [25 mM (pH 6.0)
and 0.02% NaN3] and 5% methanol for 5 min followed by a
linear gradient to 20% sodium acetate buffer and 80% methanol
over 40 min at a rate of 0.5 mL/min.
was added and the insoluble material separated by centrifugation
(16000g for 15 min). The recovered soluble fraction was assayed
directly for β-RAH-P content by HPLC as described above. For
analysis of arylamines, the liquid was separated and diluted with
150 μL of water and the arylamines were converted to their azo
dye derivatives as previously described.19,20 In most cases, the azo
dyes were purified on a small BioGel P-4 (200ꢀ400 mesh)
column (0.5 cm ꢁ 5.0 cm) eluted in 0.1 M HCl and the desired
factions assayed directly by LCꢀESI-MS and/or HPLC. Pro-
ducts were quantified from both the intensities of the specific
MHþ ion from the LCꢀESI-MS data with UV data recorded at
545 nm as well as by HPLC recorded with a diode array detector.
The peak areas of the intensities of the MHþ = m/z 503 ion for
R-RFA-P, the MHþ = m/z 503 ion for β-RFA-P, the MHþ = m/z
505 ion for ADPR-P, and the MHþ = m/z 425 ion for APDR
were used. A typical example of a separation and analytical data
obtained is shown in Figure 3. The calculated values assume that
all these ions are produced from their respective molecules in
proportion to the molar concentration of each compound. This
was confirmed because the measured areas of the 545 nm
absorbance trace corresponded with that obtained from the
m/z 425 and 503 ion traces. The amount of APDR resulting
from oxidative cleavage of MPT in the cell extracts was used as an
internal standard. Measurement of the amount of APDR in the
cell extract was accomplished using the BrattonꢀMarshall assay
of the cell extract with a sample of 4-aminophenethyl alcohol
serving as a standard. This analysis showed that the concentra-
tion of APDR was 3.0 mM in the cell extracts used. This level is
very similar to the level of methanopterin (∼100 mmol/mg of
Incubation of Cell Extracts with Precursors and Quantita-
tion of Products. The procedure consisted of incubating 70 μL
of M. jannaschii cell extract (∼30 mg/mL protein) in a total
volume of 100 μL containing ∼10 mM HB, PRPP, L-aspartate, L-
glutamine, ammonium chloride, and/or the stable isotopically
labeled compounds. When ammonia was tested, the ammonium
concentration was 100 mM. Samples were incubated for 15 min
at 70 °C and cooled to room temperature, and 10 μL of 6 M HCl
6044
dx.doi.org/10.1021/bi200362w |Biochemistry 2011, 50, 6041–6052

Biochemistry
ARTICLE
Table 1. MS/MS Data for the Azo Dye Derivatives r-RFA-P, [15N]-r-RFA-P, β-RFA-P, [15N]-β-RFA-P, and [2H4]-β-RFA-P
observed MS/MS ions originating from MHþ ions
compound
(MHþ ꢀ HPO3)
(MHþ ꢀ H2O)
R-RFA-P, MHþ = m/z 503.3
170.3
170.3
187.2
187.2
333.5
334.5
333.4
337.3
337.3
423.5
424.2
423.3
424.3
427.3
[
15N]-R-RFA-P, MHþ = m/z 504.1
171.2
188.2
486.5
485.3
486.2
489.1
β-RFA-P, MHþ = m/z 503.3
15N]-β-RFA-P, MHþ = m/z 504.1
[2,3,5,6-2H4]-β-RFA-P, MHþ = m/z 507.2
169.2
169.2
169.1
198.2
198.2
198.2
[
protein) known to be present in the methanogens.21,22 Because
this value is close to the measured amount of methanopterin, we
must conclude that most of the methanopterin was cleaved to
APDR during sample preparation. Because this cleavage is known to
occur by air oxidation of 6-substituted tetrahydropterins,23 it is
likely the nitrous acid used in formation of the azo dye is
responsible for the cleavage.20
R-RFA-P and β-RFA-P was prepared as described above, and
each was isolated by preparative HPLC separation using a TFA-
containing solvent system. The compounds were then subjected
to the different acidic conditions used for the production of the
azo dyes, and each was analyzed by HPLC and LCꢀESI-MS to
establish the extent of interconversion and the final position of
equilibrium.
Measurement of Incorporation into Precursors. The meth-
od consisted of forming and purifying the azo dye adducts of the
arylamines present in the cell extracts as described above. The
compounds eluted as a broad purple band that allowed the
separation of the azo dyes from most of the other cellular
constituents. Two additional compounds, 5-aminoimidazole-4-
carboxamide and 5-aminoimidazole-4-carboxamide ribonucleotide,
known to be present in the methanogens,24 were consistently
identified in all samples analyzed that contained cell extracts.
Three different HPLC methods were used to analyze and
quantitate the amounts of these azo dyes. Chromatographic
separation and analysis were performed on a Shimadzu HPLC
system with a diode array and equipped with a C18 reverse phase
column (Varian PursuitXRs 250 mm ꢁ 4.6 mm, 5 μm particle
size, or Agilent Zorbax Eclipse XDB-C-18, 50 mm ꢁ 4.6 mm,
1.8 μm particle size). The elution profile consisted of 5 min at
96.5% buffer A and 3.5% buffer B followed by a linear gradient to
45% buffer B over 50 min at a rate of 1 mL/min.
In solvent system 1, buffer A consisted of 0.1% TFA in water
and buffer B consisted of 0.1% TFA in methanol. In solvent
system 2, buffer A consisted of 0.1% formic acid in water and
buffer B consisted of 1% formic acid in acetonitrile. The
advantage of the TFA solvent is that the azo dyes are completely
protonated and purple with absorbance around 560 nm, whereas
in the formic acid solvent, mixed states of ionization are observed,
resulting in more complex UVꢀvisible spectra. For LCꢀMS, the
formic acid system was used and the samples were separated on a
Zorbax column with mass spectra being recorded using a AB
SCIEX Triple Quad 3200 Qtrap LC/MS/MS instrument. The
following compounds of interest eluted at the indicated times
from the Zorbax column: ZMP, 13.29 min; AIC, 13.66 min;
R-RFA-P, 14.24 min; β-RFA-P, 15.32 min; APDR-P, 17.03 min;
APDR, 17.4 min. A similar elution order was observed in solvent
system 2. Identification of each of these compounds was con-
firmed by its retention time and absorbance spectrum compared
to knowns in each of these solvent systems. Solvent system 3 was
used for the separation of R-RAH and β-RAH, and the elution
consisted of 95% sodium acetate buffer (25 mM, pH 6.0, 0.02%
NaN3) and 5% methanol for 5 min followed by a linear gradient
to 20% sodium acetate buffer and 80% methanol over 40 min at a
rate of 0.5 mL/min.
Analysis and Identification of Known Azo Dyes. In trifluor-
oacetic acid solvent system 1, 4-aminobenzyl-containing com-
pounds such as R-RFA-P, β-RFA-P, and 4-aminobenzyl alcohol
have absorbance maxima between 560 and 561 nm. Reducing the
incipient benzyl alcohol to a methylene as occurs in APDR,
APDR-P, or methaniline, the oxidative cleavage product of
MPT,23 increased the absorbance maximum to 569 nm. Azo
dye derivatives of 5-amino-4-imidazolecarboxamide-1-β-D-ribo-
furanosyl 50-monophosphate (ZMP and AICAR) and 5-amino-
4-imidazolecarboxamide (AIC), 2-aminobenzoic acid, 3-amino-
benzoic acid, AB, and aniline had absorbance maxima at 552, 543,
554, 543, 546, and 554 nm, respectively. Each compound with a
unique λmax can be readily distinguished by its visible color.
These λmax absorbance values are helpful in determining the
chemical nature of the compound being detected. All the azo
dyes that contained polar side chains such as R-RAF-P, β-RAF-P,
APDR, APDR-P, and methaniline were isolated together as a
group after BioGel P-4 chromatographic purification. Eluting
after these compounds were ZMP and AIC followed by the azo
dyes derivatives of AB and aniline.20
Identification of Azo Dye Derivatives of r-RFA-P, β-RFA-P,
APDR-P, and APDR. Each of these compounds was converted to
its azo dye derivative and purified by Bio-Gel P-4 chromatogra-
phy. The resulting samples were used as knowns to establish their
chromatographic retention times, absorbance spectra, and
LCꢀESI-MS and ESI-MS/MS data. Each compound has its
own HPLC retention time. R-RFA-P and β-RFA-P, with the
same MHþ, could be distinguished on the basis of their different
retention times and MS/MS data with the ions at m/z 170 and
187 being present in R-RFA-P and m/z 169 and 198 being
present in β-RFA-P (Table 1). As expected, many of the
observed MS/MS ions are shifted in the labeled compounds.
MS/MS of APDR-P showed the following product ions for MHþ
= m/z 505 at 169, 198, 407, 425 (MHþ ꢀ HPO3), and 487
(MHþ ꢀ H2O) and for MHþ = m/z 506 for the 15N-labeled
material at 146, 169, 198, 408, 426 (MHþ ꢀ HPO3), and 488
(MHþ ꢀ H2O). Even with the observed labeling, it has not yet been
possible to assign structures to any of the observed MS/MS ions.
Measurement of Incorporation of 15N into Soluble Amino
Acids. After incubation of cell extracts with [15N]aspartate, the
proteins were precipitated with TCA and the soluble amino acids
recovered and analyzed by GCꢀMS as their trifluoroacetyl
methyl ester derivatives.25
Chemical Interconversion of Azo Dye Derivatives of
r-RFA-P and β-RFA-P. A mixture of the azo dye derivatives of
6045
dx.doi.org/10.1021/bi200362w |Biochemistry 2011, 50, 6041–6052

Biochemistry
ARTICLE
Figure 4. Pathways to β-RFA-P and Compound 30 from either 4-aminobenzoic acid (AB) or 4-hydroxybenzoic acid (HB).
’ RESULTS
leaving group, can greatly facilitate the reaction.27 To test if an
aniline could be produced from a phenol or an aryl bromide, I
heated both HB and bromobenzoic acid with concentrated
ammonia at 110 °C for 24 h and analyzed the production of
AB by GCꢀMS of their methyl esters. No detectable AB was
produced in either reaction.
Two additional experiments were conducted to establish if it
was chemically possible to convert a phenol to an aniline and to
give mechanistic insight into how this may occur. In these
experiments, the phenol used was 4-hydroxybenzyl alcohol
because it was a closer structural analogue to the phenol in
β-RAH-P than HB. The first experiment addressed the exchange
of the oxygens of 4-hydroxybenzyl alcohol with 18O-labeled
water, and the second tested for the replacement of the hydroxyl
groups with aqueous ammonia. Analysis of the product from
the first experiment by GCꢀMS showed a single peak for the
Defining the Pathway from HB to β-RFA-P. The first step in
this three-step transformation occurs by the reaction of 4-hydro-
xybenzoic acid with PRPP to form β-RAH-P. The β-RAH-P
then condenses in the presence of ATP with L-aspartate to form
β-RFSA-P. Elimination of fumarate from β-RFSA-P produces
β-RFA-P, a known precursor to MPT. The proposed steps in this
reaction are shown in Figure 4. The details of how this was
delineated are presented below.
Testing the Chemical Conversion of 4-Hydroxybenzyl
Alcohol to an Aniline. The chemical conversion of a phenol
to an aniline is not a very facile organic reaction. If ammonia is the
nucleophile, this is an example of aromatic substitution by the
S
RN1 mechanism26 and requires reaction temperatures of >450 °C.
Conversion of the phenol to a phosphate ester, a much better
6046
dx.doi.org/10.1021/bi200362w |Biochemistry 2011, 50, 6041–6052

Biochemistry
ARTICLE
Figure 5. Proposed chemical mechanism for the replacement of the hydroxyl groups of 4-hydroxylbenzyl alcohol with an amine.
ditrifluoroacetyl derivative of 4-hydroxybenzyl alcohol showing a
molecular ion at m/z 316 containing 36% 18O. The base peak in
the spectrum at m/z 203 represented the benzyl cation radical
resulting from the loss of CF3COO• from the molecular ion. This
benzyl ion contained no 18O enrichment, indicating that only the
oxygen of the benzyl alcohol had exchanged with the water.
When the 4-hydroxybenzyl alcohol was heated with 12 M
aqueous ammonia, the generated amine-containing compounds
included 4-hydroxybenzylamine and 4-aminobenzylamine, both
being identified by GCꢀMS as their ditrifluoroacetyl derivatives.
The identification of these compounds confirmed that both the
phenol hydroxyl group and the benzyl alcohol can be chemically
replaced with ammonia to generate an amino group. The
expected reactions leading to these compounds are summarized
in Figure 5. Key to the proposed mechanism of these reactions is
the formation of the methylene quinone by the elimination of
water from 4-hydroxybenzyl alcohol. Addition of ammonia
produces 4-hydroxybenzylamine as the observed major pro-
duct. Conversion of the methylene quinone to a methylene
imine followed by the addition of ammonia would produce
4-aminobenzylamine.
Enzymatic Formation and Characterization of β-RAH-P.
Incubation of HB and PRPP with recombinant MJ1427 protein
(70 °C for 0.5 h) readily produced β-RAH-P. HPLC analysis
showed two fluorescent peaks, a minor peak at 7.5 min for R-
RAH-P and a major peak at 8.60 min for β-RAH-P. The ratio of
the minor peak to the major peak, based on the fluorescence
intensities of the peak areas, was 0.032. Removal of the phosphate
with alkaline phosphatase changed the retention times of the two
peaks to 15.17 and 16.29 min, respectively. The ratio of the minor
to the major peak representing R-RAH and β-RAH, based on the
fluorescence intensities of the peak areas, was 0.031. LCꢀESI-
MS analysis of the R-RFH-P and β-RFH-P peaks showed
(M ꢀ H)ꢀ ions at m/z 305.3, and both of the R-RFH and β-RFH
peaks showed an MHþ ion at m/z227.0 consistent with the proposed
structures. The major peak also showed an (M2 ꢀ H)ꢀ ion at
m/z 611.2 and an (M2 ꢀ H þ Na)ꢀ ion at m/z 633.2. Using
[2,3,5,6-2H4]HB as a substrate increased the observed mass of all
of these products by m/z 4, confirming their origin from HB. The
yield of β-RFH-P in the incubation was 40% based of the amount
of fluorescence observed by HPLC analysis of the sample and the
amount of HB in the incubation. No synthesis of R-RFH-P and
β-RFH-P was observed in the absence of enzyme.
Measurement of β-RAH-P Synthesis in Cell Extracts. Cell
extracts of M. jannaschii incubated with HB and PRPP and
analyzed using HPLC showed R-RFH-P and β-RFH-P with the
same retention times and ratios of R- and β-isomers that were
observed for the sample generated with the recombinant enzyme.
The β-isomer is assumed to be the more stable and major isomer
due to steric hindrance between the C-2 hydroxyl and the phenyl
ring in the R-isomer. LCꢀESI-MS confirmed the identity
of the peaks as described above. The yield of β-RAH-P was
10% based of the amount of fluorescence observed and the
amount of HB used in the incubation. These peaks were
detected in the incubation mixture only after incubation with
these substrates.
Chemical Interconversion of Azo Dye Derivatives of
r-RFA-P and β-RFA-P. The ratio of the R-RFA-P and β-RFA-
P azo dye derivatives prepared from an enzymatically generated
sample of β-RFA-P was 0.11. In the preparation of these azo dye
derivatives, the final acid hydrolysis step was eliminated because
the side chain of the methanopterin did not need to be removed.
This ratio remained stable for samples held at ꢀ20 °C for periods
of up to 4 months, the longest tested. The low ratio indicates that
β-RFA-P is the enzymatic product and that the azo dye derivative
of this compound can equilibrate to the azo dye derivative of
R-RFA-P under the acidic conditions used to prepare the
samples. To test this idea, a chromatographically pure sample
of the azo dye derivative of β-RFA-P was heated with different
concentrations of hydrochloric acid and the ratios of R-RFA-P
and β-RFA-P measured by HPLC. Heating a pure sample of the
β-RFA-P derivative (ratio of R-RFA-P to β-RFA-P of 0.024) in
0.6 M HCl for 5 min at 100 °C produced a sample with a ratio of
R-RFA-P to β-RFA-P of 0.13. Repeating the experiment with
6047
dx.doi.org/10.1021/bi200362w |Biochemistry 2011, 50, 6041–6052

Biochemistry
ARTICLE
Table 2. Incorporation of Labeled Precursors into r-RFA-P, β-RFA-P, and APDR Measured from LCꢀESI-MS Dataa
experimentb
R-RFA-P, MHþ = m/z 503
β-RFA-P, MHþ = m/z 503
APDR-P, MHþ = m/z 505
(1) control
0.0%
0.0%
ndd
∼2 μM
83.2%
∼6 μM
79.6%
(2) [2,3,5,6-2H4]HB and PRPP
ndd
12 μM
39.6%
30 μM
ꢀ
0
(3) [15N]Asp, HB, and PRPP
36.5%
19 μM
33%
60 μM
0.3 μM
29%
(4) [15N]Asp, HB, PRPP, and ATP
(5) [15N]Asp, HB, PRPP, and GTP
(6) [15N]Asp, HB, PRPP, and GTP
(7) H2, [15N]Asp, [2,3,5,6-2H4]HB, PRPP, and ATP
32%
8.7 μM
26%
75 μM
6.4 μM
14%
29%
4.8 μM
28%
18 μM
6.4 μM
10%
31%
4.0 μM
same as β-RFA-P
6.6 μM
91%
19 μM
7 μM
10%
34.4% 2H4, 26.3% 2H4 and 15
N
54 μM
91%
2 μM
∼66%
12 μM
(8) [2,3,5,6-2H4]-R-RFH-P and [2,3,5,6-2H4]-β-RFH-P at
220 μM,c ATP, and L-aspartate
35.1 μM
173 μM
a The percentages are the percentages of the total molecules that were derived from the incubated labeled precursor. The concentrations are the
concentrations of the indicated products measured in the incubation mixtures after incubation. The concentrations were calculated from measured MHþ
ion intensities and confirmed from the UV absorbance data. The MHþ for the APDR was used as the internal standard. b Each component was present at
a concentration of 10 mM in the cell extract in experiments 2 and 3 and 9.2 mM in experiments 4ꢀ7. Incubations were conducted for 15 min at 70 °C.
c The concentration of these compounds in the incubation mixture was based on its fluorescence using phenol as the standard. d Not detected.
6 M HCl increased this ratio to 0.29. Thus, the different values for
the ratios of R-RFA-P to β-RFA-P reported in Table 2, ranging
from 0.12 to 0.32, indicate different extents of chemical conver-
sion of β-RFA-P to R-RFA-P during the preparation of the
samples. Together, these data indicate that β-RFA-P is the
biologically important isomer. Data for both compounds are
listed in Table 2 because they produce two independently
derived sets of incorporation data for each incubation.
[15N]-L-aspartic acid used in the experiment, a reason for this
reduction in the amount of label incorporated was sought. This
reduction could occur either by dilution with unlabeled aspartate
present in the cell extract or through transamination reactions
with R-keto acids and/or amino acids present in the cell extracts.
Extracts of other methanogens generally contain ∼0.2 mM
R-ketoglutarate,28 and M. jannaschii contains ∼0.4 M R-gluta-
mate and ∼0.5 M β-glutamate that serve as osomolites in this as
well as other methanogens.29 Thus, the soluble free amino acids
were recovered from an incubation mixture, and the 15N
abundances in the individual amino acids were measured by
GCꢀMS analysis of their trifluoroacetyl methyl ester derivatives.
The major amino acids found were aspartate, R-glutamate, and
β-glutamate. Aspartate was found to contain 44% 15N and
R-glutamate 38% 15N. β-Glutamate, alanine, valine, proline,
Leu/Ilu, and proline contained no 15N enrichment (<2%).
These results showed that the aspartate nitrogen had under-
gone extensive mixing with the glutamate via transamination
with R-ketoglutarate, thus explaining the observed reduction
of 15N in the aspartate. Thus, the L-aspartate nitrogen was
likely the sole source of the amino group of β-RFA-P.
To test if a nucleoside triphosphate was involved in the
reaction, cell extracts were incubated with [15N]-L-aspartic acid,
PRPP, HB, and either ATP or GTP. In both experiments, ∼30%
15N was incorporated into both R-RFA-P and β-RFA-P (Table 2,
experiments 4 and 5). The amount of both of the isomers,
however, increased in the ATP over the GTP incubation, with the
amount in the GTP incubation being approximately 2ꢀ3 times
that seen in the control. Although the levels of APDR-P remained
approximately the same in both experiments, a 29% incorpora-
tion of 15N into APDR-P was observed in the ATP incubation
and 14% in the GTP incubation. Experiment 6 was a repeat of
experiment 5.
Incubation of M. jannaschii Cell Extracts with Precursors
and LCꢀESI-MS Measurement of Concentrations and Incor-
poration of Isotopes into Products. Incubation of cell extracts
with [2,3,5,6-2H4]HB and PRPP resulted in the incorporation of
the four deuterium atoms of HB into both R-RFA-P and β-RFA-
P to an extent of ∼80% (Table 2, experiment 2). This experiment
also showed an increase in the total amount of R-RFA-P and
β-RFA-P over a control incubated without the addition of
[2,3,5,6-2H4]HB and PRPP (Table 2, experiment 1). This result
demonstrated not only that the cell extracts readily incorporate
HB into β-RFA-P but also that cell extracts were able to catalyze
the replacement of the phenolic hydroxyl group of HB with
an amine. To test for the source of the nitrogen used for the
formation of the amine, incubations were conducted with 10 mM
[15N]-L-aspartic acid, 10 mM [amide-15N]glutamine, and 100 mM
[15N]ammonium chloride each incubated along with HB and
PRPP. 15N was incorporated from only [15N]-L-aspartic acid,
where ∼40% of the R-RFA-P and β-RFA-P molecules were
found to contain 15N (Table 2, experiment 3). Also, a further
increase in the total amount of R-RFA-P and β-RFA-P produced
in the extract was observed. Incubation of extracts with PRPP and
ammonium chloride or glutamine followed by HPLC analysis
using solvent system 1 showed no increase in the amount of
R-RFA-P or β-RFA-P, indicating that these compounds were
not produced from RFH-P using ammonia or glutamine as
the nitrogen source. Because the extent of labeling was much
lower that expected on the basis of the 98% enrichment in the
A cell extract was incubated with [15N]-L-aspartic acid,
[2,3,5,6-2H4]HB, ATP, and PRPP under hydrogen (Table 2,
6048
dx.doi.org/10.1021/bi200362w |Biochemistry 2011, 50, 6041–6052

Biochemistry
ARTICLE
Figure 6. Three possible routes to APDR-P from β-RFA-P (R = H) or Compound 30 (R = pterin). (I) Isomerization to the methylene imine derived
from β-RFA-P or Compound 30 (R = H2pterin), followed by a 1,2 hydride shift and reduction of the resulting ketone. (II) Amadori rearrangement of
methylene imine from β-RFA-P or Compound 30. (III) Direct reduction of methylene imine from β-RFA-P. The keto form is reduced to Compound 40
that could undergo oxidative cleavage to APDR-P. The conversion of APDR-P to methanopterin has not been demonstrated.
experiment 7). Here, 34.4% of the total RFA-P molecules were
derived from the labeled HB, and 26.3% of them had incorpo-
rated a single 15N from the L-aspartate. From the amounts of the
molecules that had increased by m/z 4ꢀ5, the nitrogen incorpo-
rated would have had an 15N enrichment of 43.3%.
of the R-RFA-P or β-RFA-P, indicating that APDR-P was not
derived solely from the β-RFA-P produced enzymatically in the
cell extract. The dilution of label occurs from APDR-P present in
or generated in the extract. The unlabeled material can arise from
oxidative cleavage of the phosphorylated intermediate as shown
in Figure 6.23
To establish that β-RFSA-P was an intermediate in biosynth-
esis, we prepared this molecule, incubated it with a cell extract,
and measured the increase in the amount of R-RFA-P and
β-RFA-P in the cell extract by LCꢀESI-MS of their azo dye
derivatives. The first incubation was conducted with a crude
preparation of β-RFSA-P, and a large increase in the amount of
R-RFA-P and β-RFA-P (Table 3, experiments 2 and 3) was
observed. Incubation in a second experiment with a chromato-
graphically pure sample of β-RFSA-P (82 μM) caused the gene-
ration of ∼45 μM R-RFA-P and β-RFA-P, consistent with half of
the sample being converted into product. This is consistent with
only one of the stereoisomers serving as the precursor. The other
product of the reaction was confirmed to be fumarate by TLC
analysis.
Finally, a mixture of [aromatic-2H4]-R-RFH-P and β-RFH-P
was prepared and incubated along with ATP and aspartate at a
total concentration of 220 μM (Table 2, experiment 8). After
incubation, the recovered azo dyes of R-RFA-P and β-RFA-P
were each found to be labeled to an extent of 91% with four
deuterium atoms. The sum of the R-RFA-P and β-RFA-P
produced was 210 μM, indicating that 95% of the R-RFH-P
and β-RFH-P mixture had been converted into product. Surpris-
ingly, the APDR-P level increased to 12 μM, and it contained
66% 2H4.
APDR-P was detected in many of these experiments on the
basis of the presence of its MHþ = m/z 505 ion at the same
retention time as a known sample of this material. Significant
enrichment of ∼10ꢀ30% 15N was seen in the ADPR-P in
experiments 5ꢀ7 (Table 2). The level of label was less than that
6049
dx.doi.org/10.1021/bi200362w |Biochemistry 2011, 50, 6041–6052

Biochemistry
ARTICLE
Table 3. Conversion of DL-β-RFSA-P to β-RFA-P by Cell
intermediate has delocalized the negative change not on the
oxygen of a carboxyl group but to the C-4 hydroxyl group leaving
from C-1 of the ribose of β-RFH-P. In other words, the opening
of the ribose ring to form the carbocation provides an electron
sink to drive the addition of the aspartate nitrogen to the
phosphorylated phenol. Addition of the C-4 hydroxyl group
back to the ribose C-1 atom with expulsion of phosphate leads to
the formation of β-RFSA-P. The fact that the nitrogen comes
from L-aspartate by this route strongly suggests the stereochem-
istry of the aspartate in the β-RFSA-P is (S), the same as that of
the aspartate in adenylosuccinate.
Enzymes in M. jannaschii annotated to be ATP-grasp enzymes
that presently have no confirmed function include proteins
derived from the MJ0815 and MJ0776 genes. MJ0815 is linked
to many of the genes involved in MPT biosynthesis and likely
encodes the enzyme catalyzing this reaction. We thus propose
that the ATP phosphorylates β-RFH-P and the resulting phos-
phorylated phenol reacts with aspartate to form β-RFSA-P
catalyzed by one of these enzymes. β-RFSA-P, after elimination
of fumarate, then forms β-RFA-P. The enzyme that catalyzes the
loss of fumarate is the adenylosuccinate lyase, the product of
the MJ0929 gene (R. H. White, unpublished observations),
an enzyme that is already known to function not only with
adenylosuccinate but also with 5-aminoimidazole-N-succinylo-
carboxamide ribotide (SAICAR), another intermediate in purine
biosynthesis.35
Extractsa
[R-RFA-P]
(μM)
[β-RFA-P]
(μM)
[APDR-P]
experiment
(1) control
(μM)
2.7
21
6
0.1
0.1
(2) β-RFSA-P incubated with
∼300 μM crude material
(3) 82 μM β-RFSA-P
81
11
42
0.1
a The concentrations reported were calculated from the intensities of the
MHþ = m/z 503 ion for R-RFA-P, the MHþ = m/z 503 ion for β-RFA-P,
the MHþ = m/z 505 ion for APDR-P, and the MHþ = m/z 425 ion for
APDR. The values reported assume that all these ions are produced from
their respective molecules in proportion to the number of moles of each
compound and that the measured concentration of APDR in the cell
extract is 3.0 mM.
’ DISCUSSION
The results presented here are consistent with the pathways
shown in Figure 4 as being involved in the biosynthesis of β-RFA-
P. Earlier in our work on the biosynthesis of methanopterin, we
demonstrated that externally fed AB could serve as a precursor to
the aminophenyl moiety of APDR,6 and from this work, it was
expected that this would also hold true for its in vivo biosynthesis.
This was supported by the observation that cell extracts readily
incorporated AB into β-RFA-P, a precursor of APDR.15,30 The
gene encoding the enzyme catalyzing this incorporation was
identified as MJ1427 in M. jannaschii,31 and the recombinant
enzyme was shown to catalyze the proposed reaction.31 The
mechanism of the enzyme was studied both before30 and after the
gene had been identified.32 One troubling aspect of the involve-
ment of AB in ADPR biosynthesis was that no genes annotated
for the biosynthesis of AB9,10 are present in the genomes of the
methanogens. From detailed biochemical and genetic studies of
the archaeal pathway to aromatic amino acid and AB biosynthesis
in the archaeon M. maripaludis, it was concluded that AB was
derived from an early intermediate in aromatic biosynthesis and
not from chorismate.13 An alternate pathway based on this idea
(Figure 3) was proposed and has now been shown not to be
functional (unpublished data). One idea not considered in this
earlier work was the possibility that chorismate-derived 4-hydro-
xybenzoic acid (HB) normally serves as a precursor to APDR but
that HB could also be derived from DHQ. The conversion of HB
to β-RFA-P resolves this problem because AB is not required to
produce β-RFA-P. AB can, however, be converted into β-RFA-P
by growing methanogens when it is present in the growth
medium. Interestingly, this incorporation occurs with the same
enzyme that incorporates HB and an enzyme that is inhibited by
HB.30
Formation of APDR-P. Early work demonstrated that β-RFA-
P condensed with 6-hydroxymethyl-7,8-dihydropterin pyropho-
sphate to form 7,8-dihydropterin-6-yl-4-(β-D-ribofuranosyl)amino-
benzene 50-phosphate (Compound 30) that was then converted
into 7,8-H2pterin-6-ylmethyl-1-(4-aminophenyl)-1-deoxy-D-ri-
bitol 500-phosphate (Compound 40) in a reaction stimulated by
the addition of FMN and F420.
15 After dephosphorylation, Com-
pound 40 is then converted into methanopterin (Figure 6). The
discovery here that APDR-P was present in the incubation mix-
tures and was labeled by the different precursors was surprising.
This molecule could be produced by the oxidative cleavage of
Compound 40 (Figure 6) and would be expected to be generated
to a small extent in a cell extract. Labeled Compound 40 should
not have been generated in the incubation mixtures because no
6-hydroxymethyl-7,8-dihydropterin pyrophosphate (H2HMP-
PP) (Figure 4) was added to the incubation mixtures. However,
because the extent of labeling was always less than that seen in the
β-RFA-P and the amounts were small, it is possible that the
enzyme(s) responsible for the conversion of Compound 30 to
Compound 40 in Figure 6 could also function to convert β-RFA-
P to APDR-P by one of the pathways shown in Figure 6.
Three possible routes for this conversion can be envisaged.
The first of these routes would be the direct reversible two-
electron reduction of the methylene imine form of β-RFA-P as
shown in reaction III in Figure 6. The formation of this methy-
lene imine would be readily generated by an intramolecular
elimination of an alcohol from the 4-aminobenzyl ether moiety
present within β-RFA-P. Evidence of the formation of such an
intermediate is supported by the facile acid-catalyzed equilibra-
tion of azo dye derivatives of R-RFA-P and β-RFA-P and the
observed exchange of the oxygen of the 4-aminobenzyl alcohol
with H218O. This equilibration explains why we always see the
mixture of R-RFA-P and β-RFA-P in the experiments reported
here because the samples are exposed to acid during the formation
and purification of the azo dyes derivatives of these molecules. As
seen from the data in Table 2, the ratios of the isomers are different
Reaction Mechanism for β-RFSA-P Formation. The avail-
able data indicate the incorporation of the aspartate nitrogen into
the β-RFA-P proceeds as shown in Figure 4. The mechanism of
this reaction is similar to that recently observed by Balskus and
Walsh in the biosynthesis of the mycosporine amino acids that
use a ATP-grasp enzyme33 to couple 4-deoxygadusol with
different amino acids.34 In our scheme, the first step in a series
of reactions is the phosphorylation of the phenol of β-RFH-P by
ATP. Subsequent nucleophilic attack by the aspartate nitrogen
with expulsion of the protonated furan oxygen of the ribose
produces a tetrahedral intermediate similar to that proposed for
ATP-grasp reactions. In our reaction, however, this tetrahedral
6050
dx.doi.org/10.1021/bi200362w |Biochemistry 2011, 50, 6041–6052

Biochemistry
ARTICLE
from sample to sample because not all the samples have come to
equilibrium. This process proceeding through a methylene quinone
(Figure 5) can also explain why a mixture of R-RFH-P and β-RFH-
P is observed. This process is expected to be enhanced by the
addition of electron-donating groups on the aromatic ring as seen in
formation of the carbocation in substituted triphenylmethane
dyes.36,37 Others have prepared molecules either identical or similar
to β-RFA-P, but their anomerization has not been previously
reported.38ꢀ40 This may have been missed because the samples
were likely never subjected to an acid treatment.
In route II, intramolecular loss of water would proceed, which
occurs in the N-(50-phospho-D-ribosylformimino)-5-amino-1-(500-
phosphoribosyl)-4-imidazole carboxamide isomerase (HisA)-cata-
lyzed Amadori rearrangement of the phosphoribosylformimino-5-
aminoimidazole carboxamide ribotide,41 an intermediate in histidine
biosynthesis.42 A very likely candidate for the enzyme to catalyze the
Amadori rearrangement (Figure 6, route II) is the product of the
MJ0703 gene. This protein is homologous to HisA (MJ1532) and is
linked to genes known to be involved in methanopterin biosynth-
esis. The sequence of reactions catalyzed with HisA is the same as
that required for the opening and rearrangement of the ribose ring of
β-RFA-P as shown in Figure 6. Route III would be via a pinacol
rearrangement involving a 1,2 hydride shift. In route II or III, the
resulting keto compound would then be reduced to the alcohol by
an F420-dependent dehydrogenase.15
M. jannaschii has three enzymes homologous to MJ0703
(HisA2). These include HisA (MJ1532), HisF (MJ0411), and
TrpF (MJ0451), each of which conducts the same type of iso-
merase reactions but uses different substrates. The hisA homo-
logous to these enzymes from Mycobacterium tuberculosis and
Streptomyces coelicolor complemented both the hisA and trpF
mutants in E. coli,43 indicating that the HisA protein can catalyze
the same type of isomerase reaction with different substrates. The
wide assortment of reactions catalyzed by these enzymes indicated
that widely different groups can be attached to a ribose 5-phos-
phate moiety and still undergo the same type of reaction reported
here. We propose that one of these isomerases has also evolved
into HisA2 that is involved in the biosynthesis of the APDR.
1-deoxy-D-ribitol; Compound 30, 7,8-dihydropterin-6-methyl-
4-(β-D-ribofuranosyl)aminobenzene 50-phosphate; Compound
40, 7,8-dihydropterin-6-methyl-10-(4-aminophenyl)-1-deoxy-D-ribitol
5-phosphate; ZMP or AICAR, 5-amino-4-imidazolecarboxamide-1-
β-D-ribofuranosyl 50-monophosphate; AIC, 5-amino-4-imidazo-
lecarboxamide; DTT, dithiothreitol; TES, N-tris(hydroxymethyl)-
methyl-2-aminoethanesulfonic acid; HEPES, N-(2-hydroxyethyl)-
piperazine-N-2-ethanesulfonic acid; PPi, pyrophosphate;
LCꢀESI-MS, liquid chromatography with electrospray-ioniza-
tion mass spectroscopy; TFA, trifluoroacetic acid; TFAA, trifluor-
oacetic anhydride; TLC, thin layer chromatography.
’ REFERENCES
(1) van Beelen, P., Stassen, A. P., Bosch, J. W., Vogels, G. D., Guijt,
W., and Haasnoot, C. A. (1984) Elucidation of the structure of
methanopterin, a coenzyme from Methanobacterium thermoautotrophi-
cum, using two-dimensional nuclear-magnetic-resonance techniques.
Eur. J. Biochem. 138, 563–571.
(2) Vorholt, J. A., Chistoserdova, L., Stolyar, S. M., Thauer, R. K., and
Lidstrom, M. E. (1999) Distribution of tetrahydromethanopterin-de-
pendent enzymes in methylotrophic bacteria and phylogeny of methenyl
tetrahydromethanopterin cyclohydrolases. J. Bacteriol. 181, 5750–5757.
(3) Keller, P. J., Floss, H. G., Le Van, Q., Schwarzkopf, B., and
Bacher, A. (1986) Biosynthesis of Methanopterin in Methanobacterium
thermoautotrophicum. J. Am. Chem. Soc. 108, 344–345.
(4) White, R. H. (1986) Biosynthesis of the 7-methylated pterin of
methanopterin. J. Bacteriol. 165, 215–218.
(5) Green, J. M., Nichols, B. P., and Matthews, R. G. (1996) Folate
biosynthesis, reduction and polyglutamylation. In Escherichia coli and
Salmonella Cellular and Molecular Biology (Niedhardt, F. C., Ed.) ASM
Press, Washington, DC.
(6) White, R. H. (1985) Biosynthesis of 5-(p-aminophenyl)-1,2,3,4-
tetrahydroxypentane by methanogenic bacteria. Arch. Microbiol. 143, 1–5.
(7) Maden, B. E. (2000) Tetrahydrofolate and tetrahydrometha-
nopterin compared: Functionally distinct carriers in C1 metabolism.
Biochem. J. 350 (Part 3), 609–629.
(8) Grochowski, L. L., and White, R. H. (2010) Biosynthesis of the
Methanogenic coenzymes. In Comprehensive Natural Products II: Chemistry
and Biology (Begley, T. P., Ed.) pp 711ꢀ748, Elsevier Ltd., New York.
(9) Ye, Q. Z., Liu, J., and Walsh, C. T. (1990) p-Aminobenzoate
synthesis in Escherichia coli: Purification and characterization of PabB as
aminodeoxychorismate synthase and enzyme X as aminodeoxychoris-
mate lyase. Proc. Natl. Acad. Sci. U.S.A. 87, 9391–9395.
(10) Anderson, K. S., Kati, W. M., Ye, Q. Z., Liu, J., Walsh, C. T.,
Benesi, A. J., and Johnson, K. A. (1991) Isolation and structure
elucidation of the 4-amino-4-deoxychorismate intermediate in the
PABA enzymic pathway. J. Am. Chem. Soc. 113, 3198–3200.
(11) Viswanathan, V. K., Green, J. M., and Nichols, B. P. (1995)
Kinetic characterization of 4-amino 4-deoxychorismate synthase from
Escherichia coli. J. Bacteriol. 177, 5918–5923.
’ AUTHOR INFORMATION
Corresponding Author
*Phone: (540) 231-6605. Fax: (540) 231-9070. E-mail: rhwhite@
vt.edu.
Funding Sources
This research was supported by the National Science Foundation
(Grant MCB 0722787 to R.H.W.).
(12) White, R. H., and Xu, H. (2006) Methylglyoxal is an Inter-
mediate in the Biosynthesis of 6-Deoxy-5-ketofructose-1-phosphate: A
Precursor for Aromatic Amino Acid Biosynthesis in Methanocaldococcus
jannaschii. Biochemistry 45, 12366–12379.
(13) Porat, I., Sieprawska-Lupa, M., Teng, Q., Bohanon, F. J., White,
R. H., and Whitman, W. B. (2006) Biochemical and genetic character-
ization of an early step in a novel pathway for the biosynthesis of
aromatic amino acids and p-aminobenzoic acid in the archaeon Metha-
nococcus maripaludis. Mol. Microbiol. 62, 1117–1131.
’ ACKNOWLEDGMENT
I thank Walter Niehaus for assistance with editing the
manuscript and Kim Harich for mass spectral analyses, Laura
Grochowski for assistance in HPLC analyses, and Huimin Xu for
producing the MJ1427 gene product.
(14) Porat, I., Waters, B. W., Teng, Q., and Whitman, W. B. (2004)
Two biosynthetic pathways for aromatic amino acids in the archaeon
Methanococcus maripaludis. J. Bacteriol. 186, 4940–4950.
(15) White, R. H. (1996) Biosynthesis of methanopterin. Biochem-
istry 35, 3447–3456.
’ ABBREVIATIONS
HB, 4-hydroxybenzoic acid; AB, 4-aminobenzoic acid; PRPP,
phosphoribosyl pyrophosphate; β-RFH-P, 4-(β-D-ribofuranosyl)-
hydroxybenzene 50-phosphate; β-RFSA-P, 4-(β-D-ribofuranosyl)
-N-succinylaminobenzene 50-phosphate; β-RFA-P, 4-(β-D-ribo-
furanosyl)aminobenzene 50-phosphate; APDR, 1-(4-aminophenyl)-
(16) White, R. H. (1990) Biosynthesis of methanopterin. Biochem-
istry 29, 5397–5404.
6051
dx.doi.org/10.1021/bi200362w |Biochemistry 2011, 50, 6041–6052

Biochemistry
ARTICLE
(17) White, R. H. (1986) Stereochemistry of the 5-(p-aminophenyl)-1,
2,3,4-tetrahydroxypentane portion of methanopterin. J. Am. Chem. Soc.
108, 6434–6435.
(18) Bradford, M. M. (1976) A rapid and sensitive method for the
quantitation of microgram quantities of protein utilizing the principle of
protein-dye binding. Anal. Biochem. 72, 248–254.
(19) Zhou, D., and White, R. H. (1992) 5-(p-Aminophenyl)-1,2,3,
4-tetrahydroxypentane, a structural component of the modified folate in
Sulfolobus solfataricus. J. Bacteriol. 174, 4576–4582.
(41) Quevillon-Cheruel, S., Leulliot, N., Graille, M., Blondeau, K.,
Janin, J., and van Tilbeurgh, H. (2006) Crystal structure of the yeast His6
enzyme suggests a reaction mechanism. Protein Sci. 15, 1516–1521.
(42) Quevillon-Cheruel, S., Leulliot, N., Graille, M., Blondeau, K.,
Janin, J., and van Tilbeurgh, H. (2006) Crystal structure of the yeast His6
enzyme suggests a reaction mechanism. Protein Sci. 15, 1516–1521.
(43) Barona-Gomez, F., and Hodgson, D. A. (2003) Occurrence of a
putative ancient-like isomerase involved in histidine and tryptophan
biosynthesis. EMBO Rep. 4, 296–300.
(20) White, R. H. (1997) Structural characterization of modified
folates in Archaea. Methods Enzymol. 281, 391–401.
(21) Gorris, L. G., and van der Drift, C. (1994) Cofactor contents of
methanogenic bacteria reviewed. BioFactors 4, 139–145.
(22) Daniels, L. (1993) Biochemistry of methanogenesis. In The
Biochemistry of the Archaea (Archaeabacteria) (Kates, M., Kushner, D. J.,
and Matheson, A. T., Eds.) pp 41ꢀ112, Elsevier, Amsterdam.
(23) White, R. H. (1985) 7-Methylpterin and 7-methyllumizine:
Oxidative degradation products of 7-methyl-substituted pteridines in
methanogenic bacteria. J. Bacteriol. 162, 516–520.
(24) White, R. H. (1997) Occurrence and biosynthesis of 5-aminoi-
midazole-4-carboxamide ribonucleotide and N-(β-D-ribofuranosyl)-
formamide 50-phosphate in Methanobacterium thermoautotrophicum
ΔH. J. Bacteriol. 179, 563–566.
(25) White, R. H. (2010) Identification, source, and metabolism of
N-ethylglutamate in Escherichia coli. J. Bacteriol. 192, 5437–5440.
(26) Bunnett, J. F. (1978) Aromatic Substitution by SRN1 Mechan-
ism. Acc. Chem. Res. 11, 413–420.
(27) Rossi, R. A., and Bunnett, J. F. (1972) General conversion of
phenols to anilines. J. Org. Chem. 37, 3570.
(28) White, R. H. (1989) A novel biosynthesis of medium chain
length R-ketodicarboxylic acids in methanogenic archaebacteria. Arch.
Biochem. Biophys. 270, 691–697.
(29) Robertson, D. E., Roberts, M. F., Belay, N., Stetter, K. O., and
Boone, D. R. (1970) Occurrence of β-glutamate, a novel osmolyte, in
marine methanogenic bacteria. Appl. Environ. Microbiol. 56, 1504–1508.
(30) Rasche, M. E., and White, R. H. (1998) Mechanism for the
enzymatic formation of 4-(β-D-ribofuranosyl)aminobenzene 50-phos-
phate during the biosynthesis of methanopterin. Biochemistry 37,
11343–11351.
(31) Scott, J. W., and Rasche, M. E. (2002) Purification, overpro-
duction, and partial characterization of β-RFAP synthase, a key enzyme
in the methanopterin biosynthesis pathway. J. Bacteriol. 184, 4442–4448.
(32) Dumitru, R. V., and Ragsdale, S. W. (2004) Mechanism of 4-(β-
D-ribofuranosyl)aminobenzene 50-phosphate synthase, a key enzyme in
the methanopterin biosynthetic pathway. J. Biol. Chem. 279, 39389–39395.
(33) Galperin, M. Y., and Koonin, E. V. (1997) A diverse superfamily
of enzymes with ATP-dependent carboxylate-amine/thiol ligase activity.
Protein Sci. 6, 2639–2643.
(34) Balskus, E. P., and Walsh, C. T. (2010) The genetic and
molecular basis for sunscreen biosynthesis in cyanobacteria. Science 329,
1653–1656.
(35) Stone, R. L., Zalkin, H., and Dixon, J. E. (1993) Expression,
purification, and kinetic characterization of recombinant human adeny-
losuccinate lyase. J. Biol. Chem. 268, 19710–19716.
(36) Deno, N. C., and Schriesheim, A. (1955) Carbonium ions. II.
Linear free energy relationships in arylcarbonium ion equilibra. J. Am.
Chem. Soc. 77, 3051–3054.
(37) Goldacre, R. J., and Phillips, R. J. (1949) The ionization of basic
triphenylmethane dyes. J. Chem. Soc. 1949, 1724–1732.
(38) Stefko, M., Pohl, R., Klepetarova, B., and Hocek, M. (2008) A
modular methodology for the synthesis of 4- and 3-substituted benzene
and aniline C-ribonucleosides. Eur. J. Org. Chem. 1689–1704.
(39) Krohn, K., Dorner, H., and Zukowski, M. (2002) Chemical
synthesis of benzamide riboside. Curr. Med. Chem. 9, 727–731.
(40) Matulic-Adamic, J., Beigelman, L., Portmann, S., Egli, M., and
Usman, N. (1996) Synthesis and Structure of 1-Deoxy-1-phenyl-β-
D-ribofuranose and Its Incorporation into Oligonucleotides. J. Org.
Chem. 61, 3909–3911.
6052
dx.doi.org/10.1021/bi200362w |Biochemistry 2011, 50, 6041–6052