Synthesis and physicochemical characterization of (6: S)-5-formiminotetrahydrofolate; A reference standard for metabolomics

Article abstract of DOI:10.1039/c8ob00245b

The one-carbon carrier of the formate oxidation level derived from the interaction of tetrahydrofolate and formiminoglutamate, which has been tentatively identified as 5-formiminoltetrahydrofolate, has been prepared by chemical synthesis. Treatment of a solution of (6S)-tetrahydrofolate in aqueous base with excess ethyl formimidate in the presence of anti-oxidant under anaerobic conditions afforded a gummy solid which, based on mass spectral analysis, conformed to a monoformimino derivative of tetrahydrofolate. Further physicochemical characterization by validated methods strongly suggested that the product of chemical synthesis was identical to the enzymatically produced material and that it was, in fact, (6S)-5-formiminotetrahydrofolate. Conditions and handling methods toward maintaining the integrity of this highly sensitive compound were identified and are described, as is analytical methodology, useful for research studies using it.

Full text of DOI:10.1039/c8ob00245b

Organic &
Biomolecular Chemistry
View Article Online
View Journal
PAPER
Synthesis and physicochemical characterization of
(6S)-5-formiminotetrahydrofolate; a reference
standard for metabolomics†
Cite this: DOI: 10.1039/c8ob00245b
A. H. Lewin, * P. Silinski,
H. H. Seltzman
D. Zhong,
A. Gilbert,
S. W. Mascarella
and
The one-carbon carrier of the formate oxidation level derived from the interaction of tetrahydrofolate and
formiminoglutamate, which has been tentatively identified as 5-formiminoltetrahydrofolate, has been pre-
pared by chemical synthesis. Treatment of a solution of (6S)-tetrahydrofolate in aqueous base with excess
ethyl formimidate in the presence of anti-oxidant under anaerobic conditions afforded a gummy solid
which, based on mass spectral analysis, conformed to a monoformimino derivative of tetrahydrofolate.
Further physicochemical characterization by validated methods strongly suggested that the product of
chemical synthesis was identical to the enzymatically produced material and that it was, in fact, (6S)-5-
formiminotetrahydrofolate. Conditions and handling methods toward maintaining the integrity of this
highly sensitive compound were identified and are described, as is analytical methodology, useful for
research studies using it.
Received 29th January 2018,
Accepted 17th July 2018
DOI: 10.1039/c8ob00245b
rsc.li/obc
pared product could be (6S)-5-formiminotetrahydrofolate or
(6R)-10-formiminotetrahydrofolate.
Introduction
Formiminotetrahydrofolate, formiminoglycine and formimino-
glutamic acid, are among the few natural compounds contain-
ing the formimino functionality; all three metabolites are said
to be short-lived.1 This one-carbon carrier of carbon at the
formate oxidation level is derived from the interaction of tetra-
hydrofolate and formiminoglutamate;2 it is converted to 5,10-
methenyltetrahydrofolate, which is involved in the purine and
methionine cycles,3 by formiminotransferase cyclodeaminase.2
As such, it is an important tool in pharmacological and meta-
bolomics studies. Along with the one-carbon carriers 5,10-
methenyl- and 10-formyltetrahydrofolate, formiminotetrahy-
drofolate has been identified as being unstable in solution
over the physiological pH range.4 On the other hand, enzymati-
cally prepared formiminotetrahydrofolate has been reported to
be relatively stable when stored in the cold.2 Using the Cahn–
Ingold rules to define the stereochemistry at C-6, one-carbon
carriers derived from (6S)-tetrahydrofolate that are substituted
at the 5-position retain the (6S)-designation (e.g., (6S)-5-formyl-
tetrahydrofolate) but substitution at the 10-position alters the
substituent hierarchy leading to (6R)-designation (e.g., (6R)-
5,10-methenyltetrahydrofolate). Hence the enzymatically pre-
Results and discussion
Synthesis
The literature report wherein a formyl group was aminated to
the corresponding formimino group by the action of N-(chloro-
carbonyl)isocyanate in the presence of antimony pentachlor-
ide5 was attractive as the most unambiguous and straight-
forward way to approach the chemical synthesis of (6S)-5-form-
iminotetrahydrofolate (2). Starting from commercially available
(6S)-5-formyltetrahydrofolate (1) the product 2 would be iso-
lated as a hexachloroantimonate (Scheme 1).
However, considering the difficulties associated with the
preparation and handling of N-(chlorocarbonyl)isocyanate,
this approach was not pursued. Instead, the preparation of 2
by treatment of tetrahydrofolate (3) with ethyl formimidate1,6
was explored. Based on the regeneration of the 1,4-bisform-
imino grouping in acanthoine1 by treatment of the 1,4-diamine
obtained by hydrolysis of acanthoine with boiling 10% sodium
hydroxide with ethyl formimidate (ESI Scheme 1†), it was
anticipated that the reaction of a large excess of ethyl form-
imidate with tetrahydrofolate (3) could produce 5-formimino-
tetrahydrofolate (2), 10-formiminoltetrahydrofolate (4), or
5,10-bisformiminotetrahydrofolate (5), but not the cyclic
RTI International, 3040 E. Cornwallis Road, Research Triangle Park, NC 27709, USA.
E-mail: ahl.emeritus@rti.org
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c8ob00245b
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
first 30 minutes of the reaction, and then to decrease, with
concomitant appearance of a product with m/z 456 and two
products with m/z 474. This suggested that 5-formiminotetra-
hydrofolate (RS-2) was formed, hydrolyzed to 5-formyltetrahy-
drofolate (RS-1) (m/z [M + H]+ calcd for C20H23N7O7: 474.14)
under the basic conditions, and also cyclized to 5,10-methenyl-
tetrahydrofolate (RS-7) (m/z [M]+ calcd for C20H22N7O6: 456.16).
The latter is known to rapidly react with base7,8 to form 10-for-
mylterahydrofolate (RS-8) (m/z [M + H]+ calcd for C20H23N7O7:
474.14), which would convert to its thermodynamically favored
isomer, 5-formyltetrahydrofolate (RS-1) as shown in Scheme 3.
HPLC and UPLC/MS analysis (Fig. 1) of the reaction
mixture confirmed the presence of 5,10-methenyltetrahydrofo-
late (7), as well as of 5- and 10-formyltetrahydrofolate (1 and 8,
respectively), consistent with this explanation (Scheme 3).
Isolation of a small amount of the product with m/z [M + H]+
473 by preparative HPLC gave a gummy solid that, based on
HPLC analysis, appeared to be stable for at least a week, when
kept refrigerated in neutral buffer.
To isolate the product obtained by chemical synthesis on a
preparative scale, the conditions utilized in the enzymatic syn-
thesis were optimized. Thus, to mimic the literature con-
ditions,2 the reaction mixture obtained by treatment of (6RS)-
tetrahydrofolate (RS-3) with excess ethyl fomimidate in
aqueous base containing BME for 30 minutes (maximum
product) was diluted with 0.2 M potassium phosphate buffer
(pH 6.8), applied directly to a Dowex column, and eluted with
a gradient of 0.3 to 1 M acetic acid at 2 °C. Although the col-
lected fractions were highly enriched in the desired product
assumed to be RS-2, these conditions were deemed inappropri-
ate for preparative scale. A modified protocol in which the
reaction mixture was rapidly diluted with cold (0 °C) water
and, working rapidly at 2 °C, directly chromatographed on a
solid phase extraction cartridge, eluting with 0.1 M acetic acid,
afforded pure RS-2. These conditions were used to prepare the
desired diastereoisomer, (6S)-5-formiminotetrahydrofolate (2).
Because isolation of 2 had given a gummy solid, the pure
product was formulated in 0.1 M acetic acid solution. Aliquots
Scheme 1 Unambiguous chemical synthetic route to (6S)-5-formimi-
notetrahydrofolate (2).
material (6) considered as a possible structure of the enzymatic
product (Scheme 2).2
Based on cost considerations probe experiments were
carried out using (6RS)-tetrahydrofolate (RS-3), as a model for
the more expensive (6S)-tetrahydrofolate (3). Treatment of a
solution of RS-3 in aqueous sodium hydroxide with ethyl for-
mimidate, under strictly anaerobic conditions and in the pres-
ence of the oxidation inhibitor β-mercaptoethanol (BME), gave
a product with mass consistent with a formimino derivative of
tetrahydrofolate (m/z [M + H]+ calcd for C20H24N8O6: 473.1892;
found: 473.1926). Monitoring the reaction progress by HPLC
showed the concentration of this product to increase for the
Scheme 2 Putative reaction products of treatment of tetrahydrofolate
(3) with excess ethyl formimidate.
Scheme 3 Degradation of 5-formimnotetrahydrofolate (2).
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
Physicochemical characterization
Purified (6S)-5-formiminotetrahydrofolate (2) that had been
stored ampouled in 0.1 M acetic acid at −80 °C for two weeks
was characterized using a combination of UPLC/UV and
LC-MS (Fig. 2). As part of the characterization, a gradient
RP-UPLC/UV method for determination of related organic
impurities was developed and qualified for use in determining
the chromatographic purity of 2. The UPLC/UV impurity
method was found to be acceptable for analysis of 2 in the
presence of related organic impurities 6. The UPLC analysis
(Fig. 2) revealed that 2 eluted as a single peak with a chromato-
graphic purity of 94.2% (280 nm). The peak corresponding to 2
was well-resolved from (6R)-5,10-methenyltetrahydrofolate (7),
which was present at 1.8%. The presence of a UV-transparent
dimer related to BME was also confirmed as a late-eluting
peak in the chromatogram (m/z [M + NH4]+ and [M + Na]+
calcd for C4H10O2S2:172 and 177; found: 172, 177).
The high-resolution ESI+ mass spectrum (Fig. 2, inset) of 2
exhibited the (M + H)+ ion (m/z 473.1926) as the base peak,
which is within 7.2 ppm of the theoretical value for
C20H24N8O6, obviating the possibility of the product being a
bis-formimino adduct, but leaving open the question of
whether the chemically synthesized product is one of the
expected monosubstituted products, 5-formiminotetrahydro-
folate (2), or 10-formiminoltetrahydrofolate, or the cyclic
material 6 considered as a possibility by Tabor.2
Fig. 1 LC-MS analysis of (6S)-5-formiminotetrahydrofolate. (A) UV trace
at 280 nm. (B) Extracted ion chromatograms are shown for the expected
ions (m/z) for 5-formiminotetrahydrofoate (2) (black, 473), tetrahydro-
folate (1) (red, 446), 5-formyl/10-formyltetrahydrofoate (3/5) (blue, 474)
and 5,10-methenyltetrahydrofolate (4) (green, 456).
(10 mL) of the solution of 2 were dispensed into vials (20 mL),
and argon gas was bubbled into the solution before the vials
were crimp-sealed and stored at −80 °C.
Quantitation
To accurately determine the concentration of 2 we followed the
literature approach2 wherein 2 was quantitated by converting it
to 7. To determine the optimal conditions for the conversion
of 2 to 7 the reaction progress at room temperature was moni-
tored by HPLC using UV/vis absorbance at 368 nm (λmax of 7)
and 280 nm (λmax of 2). It was found that – after addition of a
drop of hydrochloric acid to a solution of 2 in 0.1 M acetic acid
– the amount of 2 (280 nm) decreased over a 3-hour period,
with a concomitant increase in the amount of 7 (368 nm) (ESI
Fig. 1†). Separately, a Beer’s Law9,10 dose–response curve was
prepared (ESI Fig. 2†) using commercially procured (6RS)-5,10-
methenyltetrahydrofolate (RS-7) chloride. The ampouled solu-
tion of 2 in 0.1 M acetic acid was quantitated by conversion to
7 (under the optimized conditions), and determining the con-
centration of 7 in the resulting solution by HPLC from the
Beer’s Law dose–response curve for RS-7. The converted
sample was found to contain 0.000145 mmol mL−1 of 7,
Fig. 2 Analysis of (6S)-5-formiminotetrahydrofolate (2) by UPLC/UV
corresponding to 0.069 mg mL−1 of 2 in the ampouled sample. and LC-MS.
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
The UV/vis spectrum (ESI Fig. 3,† Panel A) corresponding to However, it could be argued that this comparison may be
2 exhibited two maxima at 222 nm and 286 nm, consistent invalid since, due to the extreme insolubility of 2 in water, the
with the reported UV for the enzymatically produced product.2
1H NMR spectrum was recorded in deuterated dimethyl-
The 1HNMR spectrum of 2 was consistent but not confirma- sulfoxide, whereas the 1H NMR spectra of 1 and 8, which as
tory of the structure (data not shown) and the 13C NMR spec- the calcium salt and the potassium salt, respectively, are in-
trum could not be obtained due to the limited solubility of 2.
soluble in dimethyl sulfoxide, were recorded (on the same
instrument) in deuterated water. In an attempt to address this
concern, the 1H NMR spectrum of a saturated solution of 1
was recorded in deuterated dimethylsulfoxide. The noisy spec-
trum (ESI Fig. 7†) showed the aromatic resonances to occur at
6.60 and 7.59 ppm; addition of 5 drops of deuterated water sig-
nificantly improved the appearance of the spectrum (ESI
Fig. 8†) but did not change the chemical shifts of the aromatic
protons, which were 6.60 and 7.62 ppm, in good agreement
with the chemical shifts observed by us for 1 in deuterated
water, 6.73 and 7.66 ppm. The noisy spectrum of 5-formyltetra-
hydrofolate (1) in dimethyl sulfoxide (ESI Fig. 7†) also showed
multiple downfield resonances that disappeared upon
addition of 5 drops of water (ESI Fig. 8†) consistent with the
observation of exchangeable protons in dimethyl sulfoxide but
not in water and suggesting that, similarly, the downfield reso-
nances observed in the spectrum of 2 are associated with
exchangeable protons. In summary, we therefore conclude that
the chemically prepared product is, indeed, 5-formiminotetra-
hydrofolate 2.
But are the chemically and enzymatically prepared form-
iminotetrahydrofolates one and the same? The fact that the
product obtained by formimidation of tetrahydrofolate, and
the enzymatically produced formiminotetrahydrofolate
possess very similar properties suggests that they are the same.
In particular, the UV spectrum (λmax 286.3 nm) for the chemi-
cally synthesized compound matches the reported value,2 as
does the overall behavior of the two compounds. Our study,
therefore, confirms the structural assignment of the enzymatic
product as (6S)-5-formiminotetrahydrofolate, as proposed by
Tabor.2
Structural characterization
Of the four possible products of the treatment of (6S)-tetra-
hydrofolate (3) with excess ethyl formimidate in the presence
of base, the bis-adduct can be eliminated based on the mole-
cular formula determined mass spectrometrically. To un-
ambiguously determine the site of attachment of the form-
imino group NMR correlation studies (COSY, ASQC, HMBC)
were carried out; unfortunately, these studies, performed at
room temperature to minimize sample degradation, proved
1
inconclusive. In fact, the HNMR of the synthetic product was
difficult to interpret, possibly due to conformational and con-
figurational heterogeneity. If the chemically produced product
were 5-substituted, two rotamers may be detected by NMR: one
rotamer with the formimino nitrogen anti to the isocytosine
ring and syn to the 9-methylene (A, R = NH, Fig. 3), and a
second rotamer with the formimino nitrogen syn to the iso-
cytosine ring and anti to the 9-methylene (B, R = NH); these
rotamers may, or may not, be present in equal concentrations.
Likewise, if 10-formiminotetrahydrofolate (4) were the chemi-
cally produced product it could exist as a mixture of the rota-
mers C (R = NH) and D (R = NH) (Fig. 3) where the imino nitro-
gen would be syn to the 9-methylene and anti to the phenyl
ring in C and vice versa for D. This situation, resulting from
restricted rotation around the carbon–nitrogen bond in
amides, has been described for the analogous 5- and 10-for-
myltetrahydrofolates (1 and 8, resp)11–13 and can be observed
in the 1H NMR spectrum of 5-formyltetrahydrofolate (1) shown
in ESI Fig. 8.† At the same time, comparison of the chemical
shifts of the aromatic protons of the synthetic product
(at 6.6 and 7.6 ppm, ESI Fig. 4†) to the shifts reported11–13 and
observed (6.4–6.8 and 7.6–7.9 ppm) for 5-formyltetrahydro-
folate (1, ESI Fig. 5†) and not those reported8,9 and observed
(7.5 and 7.9 ppm) for 10-formyltetrahydrofolate (8, ESI Fig. 6†),
strongly supports attachment of the formimino group at N-5
and not at N-10 in the chemically synthesized product.
Stability
Overall, chemically prepared (6S)-5-formiminotetrahydrofolate
(2) has been found to be reasonably stable. Lyophilization of
high purity chromatographic fractions affords a gummy solid
that is relatively stable in the cold (−20 to −80 °C), in agree-
ment with the literature.2 However, handling of this material
is extremely difficult, particularly for quantitation and physico-
chemical characterization, with anaerobic conditions being
essential to maintain the integrity of this material. At room
temperature solutions of 2 were found to degrade at 15% per
day, even under anaerobic conditions but brief (2-hour)
exposure to room temperature conditions led to minimal
(ca. 2%) degradation. Solutions in dilute acetic acid afford
good compound stability at −80 °C, but not at −20 °C.
The observed decomposition of 2 in base is worthy of some
discussion. It has been shown that treatment of formimidated
amines with strong base while heating, removes the formimino
functionality, generating the corresponding amine.1 For 2, this
would regenerate the starting (6S)-tetrahydrofolate (3). While
Fig. 3 Conformational heterogeneity of (6S)-5-formyl-/(6S)-5-formimino-
tetrahydrofolate (1/2, resp) and (6R)-10-formyl-/(6R)-10-formiminotetra-
hydrofolate (8/4, resp).
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
this study did not explore this reaction, the fact that the con-
(6S)-5-Formiminotetrahydrofolate (2). To (6S)-tetrahydrofo-
version of 3 to 2 (which is carried out in basic medium) never late (3, Scheme 2) (140 mg, 0.31 mmol) and ethyl formimidate
went to completion despite the use of a large excess of ethyl hydrochloride (679 mg, 6.2 mmol) in a centrifuge tube (50 mL)
formimidate, is consistent with this possibility. On the other was added BME (1.6 mL), followed by aqueous NaOH (2 N,
hand, since the conversion of 3 to 2 was carried out at low 3.4 mL, 6.8 mmol). Argon gas was bubbled into the mixture
temperature, deformylation may not take place although for 1 min. All the solids dissolved. After stirring at RT under
ammonia may be stripped from the formimino functionality, argon for 30 min the yellow solution was mixed with cold
leaving behind a formyl group, thereby accounting for the (0 °C) H2O (400 mL). The resulting slightly yellow solution was
formation of 5-formyltetrahydrofolate (1) even before 5,10- passed through a Phenomenex Strata-XL-A (strong anion) solid
methenyltetrahydrofolate (7) is detected. The observed phase extraction (SPE) tube (10 g per 60 mL), activated with
decomposition products of 2 in dilute acid are consistent with CH3OH (60 mL) and equilibrated with H2O (60 mL, 0 °C) before
slow cyclization and hydrolysis to 7 and 1, respectively.
use, under N2 pressure. The SPE tube was washed with H2O
(50 mL, 0 °C), CH3OH (30 mL) then with 0.1 M HOAc (4 ×
100 mL, 0 °C); the eluent was collected into Erlenmeyer flasks
that were cooled in an ice bath during and after elution. The
combined eluent, containing 95% pure 2, (HPLC: Phenomenex
Kinetex EVO C18, 150 × 4.6 mm, 2.6 μm: 25 mM Na2HPO4/H2O,
0.8 mL min−1, λ 280 nm, λ 298 nm) was divided into two por-
tions that were lyophilized overnight to give pale brown solids
(16 mg and 15 mg respectively). Repeat of the preparation on
the 140 mg scale of 3 gave consistent results but attempted
scale-up to 280 mg of 3 led to a significantly lower yield of 2.
Purification. Batches of 2 with different purities, obtained
from several preparations and stored at −80 °C, were allowed
to warm to RT, and then combined in 0.1% NH4OH (392 mL,
ice-cold). HPLC analysis (vide infra) showed the resulting solu-
tion to contain 90% pure 2. The solution was passed through a
Phenomenex Strata-XL-A (polymer strong anion) solid
exchange Giga tube (10 g per 60 mL), activated with CH3OH
(50 mL) and equilibrated with H2O (50 mL) before use drop-
wise (under nitrogen pressure). The SPE tube was washed with
H2O (50 mL, ice-cold), and then eluted with 0.1 M HOAc–H2O
(ice-cold) to give several fractions containing 2 with different
purities. Highest purity fractions were combined to give 2 with
96.3% purity (λ 280 nm) in 0.1 M HOAc solution. MS m/z calcd
for C20H24N8O8: 472; found: 473 (M + H)+; 471 (M − H)−.
Aliquots (10 mL) of the solution of 2 were dispensed into vials
(20 mL); argon gas was bubbled into the solution before the
vials were crimp-sealed and stored at −80 °C.
Determination of the solution concentration of (6S)-5-formi-
minotetrahydrofolate (2). A solution of 2 in 0.1 M HOAc con-
taining 1 M BME was de-aerated with argon gas and then
treated with HCl. Aliquots were analyzed by HPLC
(Phenomenex Kinetex EVO C18 150 × 4.6 mm, 2.6 μm, A:
[25 mM Na2HPO4/H2O], B: CH3CN, 2% B, 0.7 mL min−1, λ
368 nm, λ 280 nm) (ESI Fig. 1†). Separately, authentic (6RS)-
5,10-methenyltetrahydrofolate (RS-7) chloride (2.494 mg,
5.07 × 10−3 mmol) was weighed in an aluminum weighing
boat that was then placed into a volumetric flask (25 mL).
Methanol (3 mL) and 0.1 M HOAc were added to volume and
the sample was sonicated to achieve solution (2.028 × 10−4 M).
A Beer’s Law dose–response curve (ESI Fig. 2†) was generated
by HPLC (vide infra) of aliquots (0.5, 1, 2, 3, 4, and 5 µL) of
the solution. Plotting the injected amount (µmol, Y)
Experimental
Instrumentation
1HNMR spectra were recorded on a Bruker Avance 300 MHz
spectrometer or a 500 MHz NMR spectrometer. Low-resolution
LC-MS data were obtained using a PerkinElmer API 150 EX
mass spectrometer outfitted with an ESI (turbospray) source,
coupled to a PerkinElmer 200 Series liquid chromatography
system. High resolution LC/MS analysis was performed on an
Agilent system consisting of a 1290 Infinity UPLC coupled to a
6230 accurate-mass time-of-flight mass spectrometer with a
Dual AJS ESI source. Mass spectral data were acquired in posi-
tive-ion mode over the range of 100–1700 m/z using a gas
temperature of 350 °C, a nozzle potential of 1000 V, and a
capillary potential of 3500 V. HPLC analyses were performed
either on a dual pump system consisting of two HPLC pumps
(Varian Prostar 210 solvent system delivery system),
a
Rheodyne injector and a Varian ProStar 335 DAD UV detector
(or a Varian ProStar 320 UV detector) controlled by Varian Star
Workstation software or on Agilent HPLC 1100 system (two
HPLC pumps, an autosampler, and a diode-array detector)
controlled by ChemStation HPLC software.
Materials
The chemicals, reagents and solvents used in the synthesis
were inspected and released for use based on visual inspection
and conformance of the individual lot with the manufacturer’s
Certificate of Analysis. (6S)-Tetrahydrofolate and 5,10-methenyl-
tetrahydrofolate chloride were purchased from Schircks
Laboratories; (6RS)-tetrahydrofolate and ethyl formimidate
were from Sigma; solvents were from Sigma Aldrich. UPLC
analyses were performed using HPLC-grade acetonitrile
(Fisher), HPLC-grade ammonium acetate (Fisher), and in-
house de-ionized water. For high-resolution LC-MS, water and
acetonitrile were obtained from Fisher (Optima LC/MS grade),
and ammonium acetate was obtained from Fluka (LC-MS Ultra
grade). All other buffers and reagents were reagent grade or
better.
Synthesis
All solvents and solutions were purged with argon immediately against the UV response (integrated area, X) gave a straight
before use and were kept ice-cold.
line (Y = 107X + 0.0394, R2 = 0.9995).
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
The concentration of the purified sample of (6S)-5-form- of detecting related organic impurities at ≥0.2% of the target
iminotetrahydrofolate (2) in 0.1 M HOAc solution was determined assay concentration level for 2.
by adding BME (70 µL) to an aliquot (0.93 mL), bubbling
Specificity. Specific confirmation of detection and separation
argon gas into the solution and adding concentrated of potential impurities was performed for 5,10-methenyltetra-
HCl (20 µL). Bubbling with argon gas and stirring at RT was hydrofolate (7). Authentic (6RS)-5,10-methenyltetrahydrofolate
continued for 4 hours. To determine the concentration of 7 (RS-7) chloride, as a solid, was used as a reasonable surrogate
generated from 2 an aliquot (5 µL) of the reaction mixture was for (6R)-5,10-methenyltetrahydrofolate (7), which is the
quantitated by HPLC under the conditions utilized to generate expected degradation product of (6S)-5-formiminotetrahydro-
the Beer’s Law dose–response curve and the peak area folate (2). A control solution at ∼0.1 mg mL−1 in CH3OH was
was used to calculate the molar amount of the injected prepared and analyzed (single injection) to determine the
material. After correcting for the initial amount of 7 in 2 the retention time of 7 relative to 5-formiminotetrahydrofolate (2).
concentration of 2 was calculated to be 0.069 mg mL−1 UPLC/UV results (Fig. 1) showed that 7 eluted after 2 (RRT
(0.145 mM).
1.16) and was baseline-resolved from 2 by visual inspection.
High-resolution LC-MS (Fig. 1 inset) confirmed the identity of
both peaks. These results demonstrated the ability to discrimi-
nate between these two components.
Analysis
UPLC/UV analysis was performed using a Waters Acquity UPLC
system with a photodiode array detector, Waters Acquity HSS
T3, 2.1 × 100 mm, 1.8 μ, A: 10 mM NH4OAc pH 6.8, B: CH3OH,
0.2 mL min−1, with a linear gradient of 0–20% B from
0–4 min, 20–90% 4–5 min, hold at 90% B 5–6 min, and re-
equilibration to 0% B 6.1–10 min. The column temperature
was 25 °C, and the auto-sampler was chilled to 5 °C to mini-
mize sample degradation during analysis. The nominal injec-
tion parameters were 2 μL of a 0.1 mg mL−1 solution. UV detec-
tion was at 280 nm, and the UV spectrum (190–400 nm) was
included to support peak identification.
Solution stability
A vial of 2 in 0.1 M HOAc was removed from storage at −20 °C,
placed in an auto-sampler at 5 °C, and analyzed by LC-MS to
monitor purity over a 72-h period. Based on peak area at
280 nm, the purity of 2 was found to decrease by ca. 15% per
day, with the major degradation product being 7, accompanied
by 1. Analysis of an ampouled solution of 2 in 0.1 M HOAc by
LC-MS after 11-month storage at −80 °C showed sample purity
of >94%. At 5 °C the sample purity decreased to 91.7% after
2 hours; the largest impurity was 5,10-methenyltetrahydrofo-
late (7) (∼5% total area). The IDs of the main component (2)
and of the known impurity 7 were confirmed by mass
spectrometry.
UPLC method qualification
Linearity. To verify that analyzed sample concentrations were
well within the linear range of the detector response for the
analyte, method linearity was determined by single analysis of
three test concentrations of 5-formiminotetrahydrofolate (2)
over the nominal concentration range of 0.01 mg mL−1 to
0.15 mg mL−1 (i.e., 10% to 150% of the target analyte concen- Conclusions
tration of 0.10 mg mL−1). To simulate sample concentrations
in the range of 0.05–0.15 mg mL−1 (50%, 100% and 150% of
target concentration) three separate direct injections of the
diluted stock solution at 1, 2 and 3 μL were made. An
additional solution was prepared by 10-fold dilution of the
stock sample solution (nominally at 0.1 mg mL−1 in 0.1 M
HOAc) into 0.1 M HOAc to produce a test solution at 0.01 mg
mL−1 (10% of the target concentration). This solution was also
analyzed for linearity using a 2 µL injection. For this qualifica-
tion, the main peak corresponding to 2 was integrated for each
linearity injection, and analyte response (peak area) was
plotted versus concentration. The data were fit using linear
regression techniques and the results showed a best-fit line
with r2 = 0.9999, and % error within 5.0% at all 4 concen-
trations, establishing a linear range (peak area) for 5-formimino-
The one-carbon carrier (6S)-5-formiminotetrahydrofoalte (2)
has been prepared by chemical synthesis. Based on its pro-
perties, it appears to be the same as the enzymatic product
derived from the interaction of tetrahydrofolate and form-
iminoglutamate, thus confirming the structural assignment of
the enzymatic product as (6S)-5-formiminotetrahydrofolate, as
proposed by Tabor.2 Conditions for handling this sensitive
compound have been determined, facilitating its use as a refer-
ence standard as well as in biochemical studies.
Conflicts of interest
There are no conflicts to declare.
tetrahydrofolate from 0.01–0.15 mg mL−1
.
Limit of detection (LOD). The limit of detection for the impur-
ity method was visually confirmed at a level ≥0.2% of the
target assay concentration (0.1 mg mL−1), corresponding to
Acknowledgements
≥0.2 μg mL−1. The analyte peak height at 0.2% of the target Research
was
supported
through
Contract
No.
assay concentration was measured and the signal-to-noise HHSN268201300021C to Health and Human Services (HHS)
level was determined to be 22 : 1. Thus, the method is capable from the National Heart, Lung, and Blood Institute (NHLBI).
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
7 J. C. Rabinowitz, in Method Enzymol, ed. N. O. K. S. P.
Colowick, Academic Press, San Diego, 1963, vol. 6, pp.
814–815.
Notes and references
1 B. Frydman and V. Deulofeu, Tetrahedron, 1962, 18, 1063–
1072.
8 A. H. Lewin, P. Silinski, J. Hayes, A. Gilbert,
S. W. Mascarella and H. H. Seltzman, Metabolomics, 2017,
13, 1–8.
2 H. Tabor and L. Wyngarden, J. Biol. Chem., 1959, 234,
1830–1846.
3 P. V. Danenberg, B. Gustavsson, P. Johnston, P. Lindberg,
9 A. Beer, Ann. Phys., 1852, 86, 78–88.
R. Moser, E. Odin, G. J. Peters and N. Petrelli, Crit. Rev. 10 J. D. J. Ingle and S. R. Crouch, Spectrochemical Analysis,
Oncol. Hematol., 2016, 106, 118–131.
Prentice Hall., New Jersey, 1988.
4 P. Stover and V. Schirch, Anal. Biochem., 1992, 202, 82–88.
11 M. Poe and S. J. Benkovic, Biochemistry, 1980, 19, 4576–4582.
5 E. Muller, O. Orama, G. Huttner and J. C. Jochims, 12 J. Feeney, J. P. Albrand, C. A. Boicelli, P. A. Charlton and
Tetrahedron, 1985, 41, 5901–5912. D. W. Young, J. Chem. Soc., Perkin Trans. 2, 1980, 176–180.
6 M. Alpegiani, A. Bedeschi, E. Perrone, F. Zarini and 13 L. Li, T. Cai, Z. Wang, Z. Zhou, Y. Geng and T. Sun,
G. Franceschi, Heterocycles, 1988, 27, 1329–1340.
Spectrochim. Acta, Part A, 2014, 120, 106–118.
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.