Aracyl triflates as derivatizing agents for biological betaines

Article abstract of DOI:10.1071/CH03133

Several trifluoromethanesulfonates of general structure ArCOCH ₂OTf were prepared and their suitability for use as derivatizing agents for the HPLC analysis of betaines was assessed. Four of them (Ar=2′-naphthyl, 2′-fluorenyl, 6′-methoxynaphtha

Full text of DOI:10.1071/CH03133

Full Paper
CSIRO PUBLISHING
www.publish.csiro.au/journals/ajc
Aust. J. Chem. 2004, 57, 467–472
Aracyl Triflates as Derivatizing Agents for Biological Betaines
D. Alan R. Happer,A,C Colin M. Hayman,A,B Malina K. Storer,B and Michael LeverA,B
A Department of Chemistry, University of Canterbury, Christchurch, New Zealand.
B Biochemistry Unit, Canterbury Health Laboratories, Christchurch, New Zealand.
C Author to whom correspondence should be addressed (e-mail: a.happer@chem.canterbury.ac.nz).
Several trifluoromethanesulfonates of general structureArCOCH2OTf were prepared and their suitability for use as
derivatizing agents for the HPLC analysis of betaines was assessed. Four of them (Ar = 2ꢀ-naphthyl, 2ꢀ-fluorenyl,
6ꢀ-methoxynaphthacyl, and 2ꢀ-phenanthrenacyl) showed promise for use with UV and/or fluorescence detectors,
with the last potentially the most promising.
Manuscript received: 15 May 2003.
Final version: 20 October 2003.
Introduction
while fairly high (ε ≈ 14000), come at a relatively short
wavelength (λ 260 nm). There is an obvious case to be
made for seeking an alternative to the 4ꢀ-bromophenacyl
chromophore that offers superior optical characteristics
(absorption at longer wavelengths and/or a higher extinc-
tion coefficient), or, even better, is a good fluorophore, as
fluorescence detection is normally more sensitive.
The major limiting factor for any potential aracyl
chromophore/fluorophore is that it needs to be capable of
surviving the conditions required to prepare and to use the
reagent. Such a constraint particularly affects fluorophores,
as some of the most useful and widely used are chemi-
cally quite reactive. There are several synthetic routes to
α-ketotriflates such as 4ꢀ-bromophenacyl triflate available.
The one that appears the simplest and most logical, the
reaction of an aracyl halide with silver triflate, fails. Less
convenient, but effective, alternatives that have been, or could
be, used include the acylation of α-hydroxy ketones by triflic
anhydride,[11,12] theoxidationofsilylenolethersusinghyper-
valentiodinecompounds,[13] orthetreatmentofdiazoketones
with triflic acid.[12] From the point of view of simplicity and
the availability of suitable precursors the last is the method
of choice.
The diazoketone intermediate is conveniently prepared by
one of two routes (see Scheme 1). Method A is best suited
to making small amounts of material, as larger quantities
require the preparation and use of diazomethane on a large
scale, which, in view of the latter’s toxicity, carcinogenic-
ity, and instability, constitutes a potential hazard. The second
route (method B) makes use of the Regitz diazo transfer
reaction[14] and uses the methyl ketone rather than the acid
chloride.The ketonic methyl is subjected to a crossed Claisen
condensation with methyl formate to give the sodium salt
of the β-keto aldehyde, which in the next step is reacted
with p-toluenesulfonyl azide to give the diazoketone. While
Many betaines play important roles in physiology and bio-
chemistry. For example, carnitine and its acyl esters are key
intermediates in acyl group transport across the mitocho−n-
drial membrane,[1–3] and glycine betaine (Me3N+CH2CO2 )
is involved in cell volume regulation and in the metabolism of
methylgroups.[4–6] Thereisademandforanaccurateandreli-
able analytical method for compounds of this type. There are
several methods that have been used to determine carboxylic
acids in biological systems. Mainly they are based on treat-
ing the carboxyl function with a reagent bearing a suitable
chromophore/fluorophore, followed by high-performance
liquid chromatography (HPLC) analysis of the product using
UV or fluorescence detection. However, most of these
reagents fail to react with the less reactive carboxylate group
of a betaine.An ideal derivatizing reagent for the latter would
be one that reacted rapidly and quantitatively with betaines,
and possessed a strong UV absorbance or fluorescence at a
wavelength distinct from that of common HPLC buffers and
mobile phases. It would also be an advantage if it could be
synthesized in a small number of steps from commercially
available starting materials.While a lack of reactivity towards
other species likely to be present would also be an advan-
tage, it is not essential, provided the products formed did not
interfere with the HPLC analysis.
Of the common reagents employed to analyze carboxylic
acids in biological systems, the one that comes closest to ful-
filling these requirements is 2-(4ꢀ-bromophenyl)-2-oxoethyl
trifluoromethanesulfonate (4ꢀ-bromophenacyl triflate), and
this has been used by several groups.[7–10] Unfortunately, this
compound, while stable, easily prepared in two steps from
commercially available 4-bromobenzoyl chloride, and capa-
ble of reacting rapidly and quantitatively with betaines to give
their 4ꢀ-bromophenacyl esters, does not fluoresce, and the
absorptionmaximaofitsbetainederivativesintheUVregion,
© CSIRO 2004
10.1071/CH03133
0004-9425/04/050467

468
D. A. R. Happer et al.
this method is not itself entirely hazard-free (it involves the
preparation and use of p-toluenesulfonyl azide), it is the
one of choice when larger amounts of material are involved.
It has the added advantage that p-toluenesulfonyl azide,
unlike diazomethane, can be stored indefinitely in a freezer
until required.
available. Good fluorophores normally also absorb strongly
in the UV region, so our selections were oriented towards
these. The most successful turned out to be ones based on
fused carbocyclic ring systems.
Results and Discussion
Conversion of diazoketones to α-ketotriflates can be
achieved by treatment with triflic acid in liquid sulfur
dioxide.[12] Although this method gave acceptable results,
dichloromethane at low temperatures proved more conve-
nient and gave equal or superior yields. In our search for
suitable chromatophore/fluorophores, we limited ourselves
mostly to those that already contained either an acetyl or
a carboxyl functional group and which were commercially
The following triflates were selected for screening as poten-
tial candidates for derivatizing agents: 4ꢀ-methoxyphenacyl
triflate, 2ꢀ-fluorenacyl triflate, 3ꢀ,4ꢀ,5ꢀ-trimethoxyphenacyl
triflate, 2ꢀ-naphthacyl triflate, 9ꢀ-xanthenacyl triflate, 6ꢀ-
methoxynaphthacyl triflate, 4ꢀ-phenoxyphenacyl triflate,
and 2ꢀ-phenanthrenacyl triflate. Of the eight, 3ꢀ,4ꢀ,5ꢀ-
trimethoxyphenacyl triflate and 4ꢀ-phenoxyphenacyl tri-
flate decomposed during attempts to purify them, and
4ꢀ-methoxyphenacyl triflate, while obtained pure and suc-
cessfully derivatized representative betaines, had too short a
shelf life to justify detailed investigation for possible labora-
tory use. This left five reagents, namely the 9ꢀ-xanthenacyl,
2ꢀ-fluorenacyl, 2ꢀ-naphthacyl, 6ꢀ-methoxynaphthacyl, and
2ꢀ-phenanthenacyl triflates, for further investigation. A cur-
sory check revealed that 9ꢀ-xanthenacyl triflate did not absorb
strongly in the UV region, and fluoresced only weakly.
In view of this, and the observation that the yield in
the derivatization step was poor, this compound was not
investigated further. This left four potential reagents for
detailed study. All were used to derivatize a representa-
tive betaine, namely glycine betaine. Derivatization involved
mixing the betaine and the triflate together in aceto-
nitrile containing a small amount of magnesium hydrox-
ide. The spectroscopic properties of the derivatives obtained
were determined, and the relevant data are summarized in
Table 1. Also included for comparison purposes are the
Method A
Method B
Ar
Ar
Ar
Cl
Ar
CH3
O
O
O
O
NaOMe
HCO2Me
CH2N2
CHN2
TsN3
Ar
H
O
ONa
CF3SO3H
CH2OTf
Scheme 1.
Table 1. Extinction coefficients and fluorescence properties of aracyl triflates
Solvents: propan-2-ol/water (75%), acetonitrile/water (90%)
4ꢀ-Bromophenacyl
2ꢀ-Naphthacyl
2ꢀ-Fluorenacyl
6ꢀ-Methoxynaphthacyl 2ꢀ-Phenanthrenacyl
PriOH/ MeCN/
PriOH/ MeCN/ PriOH/ MeCN/ PriOH/
H2O H2O H2O H2O H2O
MeCN/ PriOH/
H2O H2O
MeCN/
H2O
H2O
H2O
Aracyl triflate
λmax [nm]
εmax
260
14000
257
15900
249
35400
252
418
100
249
39600
314
17500
312
25900
241
26900
247
431
1400
243
266
267
58500
265
418
733
30500 49000
λEx [nm]
231B
306B
260
409
550
266
437
1666
λEm [nm]
Relative
fluorescenceA
350B
282B
350B
516B
C
–
Glycine betaine derivative
λ [nm]
260
38
260
21
249
100
252
431
100
249
52
250
413
1.5
314
53
335
396
12.5
314
31
327
379
<0.2
241
60
245
451
212.5
241
27
260
431
12.5
266
97
267
463
178.5
266
60
260
433
82
Peak area (UV)D
λEx [nm]
λEm [nm]
Peak area
(fluorescence)D
A Arbitrary units relating to peak height on fluorescence spectrophotometer (for 2ꢀ-naphthacyl triflate this value is 100 in 75%
propan-2-ol).
B These values may be due to an impurity (see text).
C None detected.
D Area of derivative peak following HPLC separation relative to 2ꢀ-naphthacyl triflate derivative (area 100) in 75% propan-2-ol.

Aracyl Triflates as Derivatizing Agents for Biological Betaines
469
corresponding properties of 4ꢀ-bromophenacyl triflate and
its glycine betaine derivative.
of the fluorescence of the aromatic system and decreases
the separation of the excitation and emission wavelengths.
2ꢀ-Naphthacyl, 6ꢀ-methoxynapthacyl, and2ꢀ-phenanthrenacyl
triflates had excitation and emission wavelengths consider-
ably separated which is desirable for a derivatizing reagent
to be used with fluorescence detection. As the bandwidth
of fluorescent detectors in most HPLC instruments is
approximately 10–15 nm, well separated wavelengths lead to
decreased background interference due to scattered light, and
hence to an improved signal-to-noise ratio. However, in the
case of the 2ꢀ-fluorenacyl triflate, the excitation and emission
wavelengths are close enough to cause problems.
The fluorescence properties of the glycine betaine deriva-
tives were similar to those of the derivatizing reagent used,
and significant fluorescence was observed when the exci-
tation and emission wavelengths of the reagents were used
for their detection. However, in solution, differences in the
solvation of the derivatives and the reagents can affect the
positions of the fluorescence maxima, so the wavelengths
used for detection were optimized. It was found that, in gen-
eral, the excitation maxima of the glycine betaine derivatives
lay close to those of the reagents, while those for emission
shifted to a longer wavelength. The wavelengths of the 2ꢀ-
fluorenacyl triflate and its betaine derivative were not related
and the strength of the fluorescence of the former was much
stronger than that of the latter. This discrepancy suggests that
the reagent may contain a strongly fluorescent impurity. If
this impurity did not derivatize glycine betaine, the wave-
lengths recorded would be the true maxima of the fluorenacyl
derivative.
For UV detection, the ideal derivatizing reagent would
have a maximum absorption wavelength of greater than
260 nm and an extinction coefficient at least as great as 4ꢀ-
bromophenacyl triflate. Of the reagents tested, 2ꢀ-naphthacyl,
2ꢀ-phenanthrenacyl, and 6ꢀ-methoxynapthacyl had absorp-
tion maxima at wavelengths close to or shorter than
4ꢀ-bromophenacyl triflate, but the extinction coefficients
were larger. The glycine betaine derivatives of these reagents
had greater UV absorbance than the bromophenacyl deriva-
tive, so although the wavelengths used for detection did
not avoid the interference from the HPLC buffer and sol-
vents, the signal-to-noise ratios were improved. The other
reagent, 2ꢀ-fluorenacyl triflate, had an extinction coeffi-
cient that was only 1.2 times that of 4ꢀ-bromophenacyl
triflate;however, theabsorptionmaximum(314 nm)waslong
enough to be distinct from interference from common HPLC
buffers and solvents. The potentially most useful reagent
is 2ꢀ-phenanthreneacyl triflate, as this had a maximum
absorbance at 266 nm and an extinction coefficient more than
three times that of 4ꢀ-bromophenacyl triflate.
All four of the tested reagents successfully derivatized
glycine betaine under the reaction conditions used for deriva-
tization with 4ꢀ-bromophenacyl triflate.The resulting glycine
betaine derivatives had virtually identical absorbance max-
ima to those of the parent reagents. Extinction coefficients
of these derivatives were not determined, but the measured
peak areas, detected by UV, are related to these. The deriva-
tives of 4ꢀ-fluorenacyl triflate absorbed only slightly less
than derivatives of 4ꢀ-bromophenacyl triflate, and less than
those of 2ꢀ-naphthacyl triflate. However, the former would
be more useful in practice because of the longer wavelength
of their maximum. Detection at 314 nm is less subject to
interference than detection at 260 nm, the position of the
maximum for 4ꢀ-bromophenacyl triflate, and this results in
a superior signal-to-noise ratio. As would be expected, the
peak areas of the glycine betaine derivatives of 2ꢀ-naphthacyl,
6ꢀ-methoxynapthacyl, and 2ꢀ-phenanthrenacyl triflates were
all greater than that for the 4ꢀ-bromophenacyl derivative.
6ꢀ-Methoxynaphthacyl triflate would be less useful for UV
detection than the other reagents, as detection at 241 nm can
be subject to a considerable interference from buffer compo-
nents, and this affects the signal-to-noise ratio. The deriva-
tives of 2ꢀ-phenanthrenacyl triflate have a strong absorption
maximum at 266 nm. This is very useful for UV detection, as
the background at this wavelength is low.
6ꢀ-Methoxynaphthacyl triflate and 2ꢀ-phenanthrenacyl tri-
flate had greater fluorescence than 2ꢀ-naphthacyl triflate, and
this was also true of their betaine derivatives. However, the
extent of the increased fluorescence was dependent on the
solvent used. This means that choice of solvent becomes
potentially an important consideration when designing an
HPLC system. It was found that for all the reagents less
fluorescence was observed in acetonitrile, to the extent that
it was negligible for 2ꢀ-naphthacyl triflate in that solvent.
A considerable decrease was also noted for its derivatives.
As the HPLC system is based on an acetonitrile mobile
phase, this made 2ꢀ-naphthacyl triflate unsuitable as a reagent
for fluorescence-based detection.[15,16] The reagents were
more strongly fluorescent in propan-2-ol; similar behaviour
was noted for the derivatives. Using a propan-2-ol mobile
phase, the derivative of 2ꢀ-phenanthrenacyl triflate was twice
as fluorescent as the same derivative analyzed with an
acetonitrile one. Changing the solvent affected the fluores-
cence of the derivatives to varied extents. In acetonitrile,
the 2ꢀ-phenanthrenacyl derivative was almost 50 times more
fluorescent than the derivatives of the other reagents, but in
propan-2-ol the increase is only twofold. The emission max-
ima of the reagents and the derivatives changed with solvent.
Tamaki has shown that the effect of protic solvents on the
fluorescence spectra of 2ꢀ-acylanthracenes arises from a
combination of the non-specific dipolar and the specific
hydrogen-bond interactions.[17,18] The decrease in dielec-
tric constant from acetonitrile (38.8) to propan-2-ol (18.3)
The derivatization conditions used were ones optimized
for 2ꢀ-naphthacyl triflate and are not necessarily optimal
for the other reagents. This may explain the less than
expected sensitivity increase for the 2ꢀ-phenanthrenacyl and
6ꢀ-methoxynaphthacyl compounds. Larger peak areas could
be expected if the derivatization conditions used were opti-
mized for each reagent.
The sensitivity of the assay could be increased consid-
erably if fluorescence detection was possible. All of the
reagents in Table 1 had some fluorescence. However, the car-
bonyl of the COCH2OTf group both diminishes the intensity

470
D. A. R. Happer et al.
decreases the Stokes shift and results in a longer emis-
sion maximum. Hydrogen-bond formation in propan-2-ol
enhances this shift, but interaction is less in acetonitrile. The
absorbance of the derivatives was only changed to a small
extent by the solvent.
ether, and allowed to air-dry. The yield (about 22 g) was approximately
quantitative, and the product could be used without further purification.
If not used immediately, it could be stored until needed.
The sodium salt of the β-keto aldehyde (22 g, 100 mmol) was sus-
pended in ethanol (250 mL) and cooled to 0–5◦C in an ice bath. To this
was added dropwise, with stirring, p-toluenesulfonyl azide[21] (20 g,
100 mmol). (The rate of addition was not critical, as heat did not appear
to be evolved.) The ice bath was removed, and stirring continued at
room temperature for 4 h. During this time the suspension turned yel-
low. The ethanol was then evaporated at 30◦C under reduced pressure,
approximately 250 mL of ether and 200 mL of 10% sodium hydroxide
was added to the residue, and the contents of the flask were shaken until
any solid present had dissolved in the ether layer. This layer was then
separated, washed with water, and dried. Evaporation of the ether gave
the diazo ketone as a yellow solid. This material was sufficiently pure
for conversion into the triflate.The yield was approximately 18 g (92%).
Diazoketonespreparedbythisroutewere4ꢀ-phenoxydiazoacetophe-
none (54%), 2ꢀ-diazoacetylfluorene (50%), 6ꢀ-methoxy-2ꢀ-diazoacetyl-
naphthalene (93%), and 2ꢀ-diazoacetylphenanthrene (85%).
In the case of the last two diazo ketones, the method was modified
slightly because of the low solubility of both the methyl ketone starting
material and the products in the solvents used. For the 6ꢀ-methoxy-
2ꢀ-diazoacetylnaphthalene compound, it proved necessary to use about
three times as much ether in the Claisen step, and in the extraction
of the diazo compound dichloromethane was used rather than ether. For
the 2ꢀ-phenanthrenyl compound, the starting material and products were
even less soluble. The condensation step was allowed to continue for 2
days, and the diazo transfer step for 3 days. The diazo ketone product
was then isolated by multiple extractions of the reaction mixture with
dichloromethane.
Conclusions
All four of the potential reagents that were synthesized
and investigated in detail represented an improvement on
the widely used 4ꢀ-bromophenacyl triflate.[7–10,19] All could
be readily prepared from commercially available start-
ing materials. Each had a chromophore/fluorophore with
a stronger extinction coefficient and fluorescence than 4ꢀ-
bromophenacyl, and all derivatized betaines rapidly and
quantitatively. Of the four, 2ꢀ-phenanthrenacyl triflate shows
the most promise as a derivatizing reagent for both UV
and fluorescence detection in biological samples. The other
reagents studied are also useful, as having a range of reagents
to choose from increases the range of buffers and solvents
that can be utilized. A range of fluorescent reagents is also
advantageous for capillary electrophoresis coupled with flu-
orescent detection, as the reagent can be chosen that best suits
the fixed extinction and emission wavelengths available.
Experimental
The carboxylic acid chlorides required for method A were either com-
mercially available or were prepared from the corresponding acids by
treating them with either thionyl chloride or oxalyl chloride. The methyl
ketones required for method B were all commercially available. 1H and
13C NMR spectra were obtained on a 300 MHzVarian Unity or 500 MHz
Varian Inova spectrometer.
Conversion of the Diazoketones into α-Ketotriflates
The diazo ketones prepared by methodA or B were converted into α-keto
triflates by treatment with triflic acid, either by the method of Vedejs[12]
or by treatment with triflic acid in dichloromethane. The main differ-
ence between the two methods is the different solvent used. The first
uses liquid SO2 at −78◦C and the second, dichloromethane. The results
were comparable, but the second method was more convenient.[22] The
procedure for the CH2Cl2 method is given below.
Dichloromethane (30 mL) was placed in a 100-mL conical flask.
The flask and its contents were cooled to about −20◦C by means of
a dry ice/ethylene glycol bath, and triflic acid (4.5 mL, 50 mmol) was
added slowly to the stirred solution. A solution of the diazo ketone (5 g,
25 mmol) dissolved in approximately 20 mL of dichloromethane was
then added dropwise over 10–15 min to the well stirred solution. The
solution initially turned orange and then darkened. After the addition
was complete, stirring under nitrogen was continued for about 1 h at
−20◦C. The mixture was then allowed to warm to room temperature
and stand overnight.
Excess triflic acid was neutralized by the cautious addition of aque-
ous sodium bicarbonate (10 mL) to the stirred solution. The mixture
was then poured into a 250-mL separating funnel, washed several
times with aqueous sodium bicarbonate, and then with water. The dark
dichloromethane layer was separated, dried with anhydrous sodium sul-
fate, and the dichloromethane evaporated to give the triflate as a dark,
but normally crystalline, product. At this point the yield was usually
quite high, and normally a single crystallization gave a product suf-
ficiently pure for subsequent use. Purer products could be obtained by
furtherrecrystallizationofthismaterial, butthisoftenledtoconsiderable
losses.
Synthesis of the Diazo Ketones
The two methods used for the preparation of the diazo ketones are
illustrated below for the case of 2ꢀ-diazoacetylnaphthalene.
Method A[20]
A solution of 2-naphthoyl chloride (11 g, 60 mmol) in anhydrous ether
(20 mL) was added dropwise over 20 min to a stirred solution of approx-
imately 0.1 moles of diazomethane in ether (300 mL) cooled to 0◦C in
an ice bath. Stirring was continued for 3 h, and the mixture was then
allowed to stand unstoppered in a fume hood overnight to allow evapora-
tion of excess diazomethane. Evaporation of the solution gave a product
that was sometimes dark. This could be removed by dissolving it in
warm benzene and filtering it through a short (15 cm) silica column.
The major impurities were small amounts of methyl 2-naphthoate and 2-
chloroacetylnaphthalene. Crystallization of the material from petroleum
ether containing a small amount of dichloromethane gave a sufficiently
pure product for the next step. The yield of diazo ketone was usually
10–11 g (80–90%).
Other diazo ketones prepared by this route were 4ꢀ-methoxydiazo-
acetophenone (86%), 3ꢀ,4ꢀ,5ꢀ-trimethoxydiazoacetophenone (85%),
9ꢀ-diazoacetylxanthene (93%), and 2ꢀ-diazoacetylfluorene (81%).
Method B[14]
Details for individual compounds are given below. Unless otherwise
stated the conversion was carried out in dichloromethane.
A well stirred slurry of sodium methoxide (6 g, 110 mmol) in anhydrous
ether (80 mL) was cooled in an ice bath and to this was added, drop-
wise over approximately 15 min, a solution of 2-acetonaphthone (17 g,
100 mmol) and methyl formate (6.6 g, 110 mmol) in sufficient ether to
dissolve the former (approximately 50 mL). Stirring was continued for
approximately 15 h at room temperature. During this period, the suspen-
sion of the sodium methoxide was replaced by one of the sodium salt of
the β-keto aldehyde.The precipitate was filtered, washed with anhydrous
2ꢀ-Naphthacyl Triflate
Crystallization of the crude product from benzene/petroleum ether gave
colourless leaflets, mp 109–110◦C, in approximately 30% yield (Found:
C 49.0, H 3.1. C13H9F3O4S requires C 49.1, H 2.9%). δH (300 MHz,
CDCl3 and Me4Si) 8.35 (1 H, s, ArH), 7.88–7.97 (4 H, m, ArH), 7.56–
7.68 (2 H, m, ArH), 5.77 (2 H, s, CH2).

Aracyl Triflates as Derivatizing Agents for Biological Betaines
471
2ꢀ-Fluorenacyl Triflate
0.2
0.1
0.0
Crystallization from dichloromethane/pentane gave light purple flakes
(30%), mp 140–142◦C (Found: C 53.8, H 3.2. C16H11F3O4S requires
C 53.9, H 3.1%). δH (300 MHz, CDCl3 and Me4Si) 8.08 (1 H, s, ArH),
7.86–7.91 (3 H, m, ArH), 7.59–7.63 (1 H, m, ArH), 7.42–7.46 (2 H, m,
ArH), 5.71 (2 H, s, CH2), 3.99 (2 H, s, H29ꢀ).
2
1
3
5
6ꢀ-Methoxy-2ꢀ-naphthacyl Triflate
4
This product was extremely difficult to free from coloured impurities.
Crystallization from benzene/petroleum ether gave white flakes with
a greenish tinge, mp 132–134◦C, in approximately 30% yield. More
strongly coloured fractions obtained (ranging up to dark green) appeared
substantially pure by 1H NMR spectroscopy (Found: C 55.7, H 2.9.
C15H13F3O6S requires C 55.4, H 3.0%). δH (300, MHz, CDCl3 and
Me4Si) 8.28 (1 H, s, ArH), 7.91 (1 H, d, J 8.8, ArH), 7.81–7.86 (2 H, m,
ArH), 7.24 (1 H, d, J 8.8, ArH), 7.17 (1 H, s, ArH), 5.76 (2 H, s, CH2),
3.96 (3 H, s, OCH3).
10
20
30
min
Fig. 1. HPLC trace of standards of natural betaines derivatized
with 2ꢀ-phenanthrenacyl triflate. [Peak 1, acetylcarnitine (Me3N+
CH2CH(OAc)CH2CO−); peak 2, arsenobetaine (Me3As+CH2CO−2 );
peak 3, glycine b2etaine; peak 4, dimethylsulfoniopropionate
(Me2S+CH2CH2CO2−); peak 5, carnitine (Me3N+CH2CH(OH)
CH2CO−2 ).]
2ꢀ-Phenanthrenacyl Triflate
Because of the low solubility of the diazo ketone in dichloromethane,
the diazo ketone was added as a slurry. This appeared to have no
effect on the yield. The crude product was substantially pure by NMR
spectroscopy, and formed in close to quantitative yield. Recrystalliza-
tion from dichloromethane gave light brown crystals, mp 120–122◦C
(Found: C 55.7, H 2.9. C17H11F3O6S requires C 55.4, H 3.0%). δH
(300 MHz, CDCl3 and Me4Si) 8.79 (1 H, d, J 8.8, ArH), 8.71 (1 H, d, J
7.3, ArH), 8.39 (1 H, s, ArH), 8.12 (1 H, d, J 8.8, ArH), 7.94 (1 H, d, J
6.8, ArH), 7.86 (1 H, d, J 8.8, ArH), 7.81 (1 H, d, J 8.8, ArH), 7.71–7.74
(2 H, m, ArH), 5.82 (2 H, s, CH2).
5 min at 11600g. Samples were removed immediately from centrifuge
and 200 µL of supernatant was transferred to an HPLC vial, and the sam-
ples were diluted with acetonitrile (200 µL). Vials were capped tightly
to avoid evaporation of sample and 10 µL of sample was injected onto
the column.
HPLC was performed on a Shidmadzu 10 AVP. The maximum
ultraviolet absorption of derivatives was determined using data gath-
ered during analysis of standards. Fluorescence data were obtained
by collecting fractions containing glycine betaine derivatives after
HPLC.
HPLC analyses used as the mobile phase either 75% propan-2-ol/
water with 1.25 mmol L−1 succinic acid, 0.625 mmol L−1 triethylamine
eluting buffer, or 90% acetonitrile/water with 7 mmol L−1 glycolate
eluting buffer. Columns were either a Brownlee 5-µm silica 100 by
4.6 mm cartridge or a Phenosphere (SCX 80A 5 µM, 250 by 4.6 mm)
with a guard cartridge of the same packing. Columns were maintained
at 40◦C during analyses.
9ꢀ-Xanthenacyl Triflate
The crude product was obtained in approximately 90% yield using
Vedejs’ method.[12] Recrystallization from pentane/dichloromethane
gave white fibrous crystals, mp 98–100◦C, in low yield (Found: C 51.8,
H 3.0. C16H11F3O4S requires C 51.6, H 3.0%). δH (300 MHz, CDCl3
in Me4Si) 7.34–7.41 (2 H, m, ArH), 7.10–7.26 (6 H, m, ArH), 5.15
(1 H, s, H9ꢀ), 4.83 (2 H, s, CH2).
4ꢀ-Methoxyphenacyl Triflate
The crude product, also obtained using Vedejs’ method,[12] crystallized
from dichloromethane/pentane as fine white needles (49%), mp 73–
74◦C (Found: C 40.5, H, 3.2. C10H9F3O5S requires C 40.3, H 3.0%).
δH (300 MHz, CDCl3) 7.84 (2H, d, J 9.0,ArH), 6.96 (2H, d, J 9.0,ArH),
5.58 (2 H, s, OCH2), 3.87 (3 H, s, OCH3). This compound decomposed
over a few weeks.
A chromatogram obtained from an analysis using a Phenosphere
column and 90% acetonitrile/glycolate eluent of the product obtained
following derivatization of a standard mixture of natural betaines by
2ꢀ-phenanthrenacyltriflateisgiveninFig. 1. (Thetwounidentifiedpeaks
were also present in a blank.)
Acknowledgments
4ꢀ-Phenoxyphenacyl Triflate
The authors thank Chris McEntyre for expert technical
assistance. Financial assistance from the Foundation of
Research Science and Technology, and the Canterbury Med-
ical Research Foundation is gratefully acknowledged.
All attempts to purify the crude product led to its rapid decomposition.
Spectrophotometric Measurements
Ultraviolet adsorption data were recorded on a Philips PU 8730 UV-
visible spectrophotometer. Fluorescence spectra were recorded on a
Varian Cary Eclipse fluorescence spectrophotometer. A quartz (10-mm
square) sample cell was used and the slit width for both excitation and
emission was 5 nm. Standard solutions (10 µM) were prepared in 75%
propan-2-ol, 25% water, or 90% acetonitrile, 10% water.
For the formation of all derivatives, an aqueous standard containing
glycine betaine was used.The final concentration of sample injected into
the HPLC port after derivatization was 7 µM. The procedure used was
essentially the same as that used previously by us for the derivatization
of betaines using p-bromophenacyl triflate.
References
[1] N. Siliprandi, L. Sartorelli, M. Ciman, F. di Lisa, Clin. Chim.
Acta 1989, 183, 3. doi:10.1016/0009-8981(89)90267-2
[2] R. Haeckel, E. Kaiser, M. Oellerich, N. Siliprandi, J. Clin.
Chem. Clin. Biochem. 1990, 28, 291.
[3] J. H. Mar, J. B. McMillin, Adv. Lipobiol. 1997, 2, 195.
[4] D. M. J. Dickson, G. O. Kirst, New Phytol. 1987, 106, 645.
[5] A. Garcia-Perez, M. B. Burg, Physiol. Rev. 1991, 71, 1081.
[6] S. T. Chambers, M. Lever, Nephron 1996, 74, 1.
[7] (a) S. T. Ingalls, P. E. Minkler, C. L. Hoppel,
J. E. Nordlander, J. Chromatogr. 1984, 299, 365. doi:10.1016/
S0021-9673(01)97852-5
To 50 µL of standard in acetonitrile (250 µL) in an eppendorf tube
was added derivatizing reagent (50 µL, 10 mM) in acetonitrile and
Mg(OH)2 (5 µL, 0.1 g L−1 suspension of MgO in water). This mixture
was vortexed briefly and shaken for 5 min, before being centrifuged for

472
D. A. R. Happer et al.
(b) P. E. Minkler, S. T. Ingalls, L. S. Kormos, D. E. Weir,
C. L. Hoppel, J. Chromatogr., Biomed. Appl. 1984, 336, 271.
doi:10.1016/S0378-4347(00)85150-6
(c) P. E. Minkler, S. T. Ingalls, C. L. Hoppel, J. Chro-
matogr., Biomed. Appl. 1987, 420, 385. doi:10.1016/0378-
4347(87)80194-9
[14] M. Regitz, F. Menz, Chem. Ber. 1968, 101, 2622.
[15] D. O. McGregor, W. J. Dellow, M. Lever, P. M. George,
R. A. Robson, S. T. Chambers, Kidney Int. 2001, 59, 2267.
doi:10.1046/J.1523-1755.2001.0590062267.X
[16] W. J. Dellow, M. Lever, S. T. Chambers, H. Lunt, R. A. Robson,
Diabetes Res. Clin. Pr. 1999, 43, 91. doi:10.1016/S0168-8227
(98)00115-6
[17] T. Tamaki, Bull. Chem. Soc. Jpn 1982, 55, 1761.
[18] T. Tamaki, Bull. Chem. Soc. Jpn 1980, 53, 557.
[19] P. E. Minkler, E. P. Brass, W. R. Hiatt, S. T. Ingalls, C. L. Hoppel,
Anal. Biochem. 1995, 231, 315. doi:10.1006/ABIO.1995.0057
[20] J. L. E. Erickson, J. M. Dechary, M. R. Kesling, J. Am. Chem.
Soc. 1951, 73, 5301.
[21] T. J. Curphey, Org. Prep. Proced. Int. 1981, 13, 112.
[22] We are grateful to Dr J. C. Morris for suggesting this
modification.
(d) P. E. Minkler, S. T. Ingalls, C. L. Hoppel, Anal. Biochem.
1990, 185, 29.
[8] M. Lever, L. Bason, C. Leaver, C. M. Hayman, S. T. Chambers,
Anal. Biochem. 1992, 205, 14.
[9] M. H. Mar, T. W. Ridky, S. C. Garner, S. H. Zeisel, J. Nutr.
Biochem. 1995, 6, 392. doi:10.1016/0955-2863(95)80008-Z
[10] L. Vernez, W. Thormann, S. Krahenbull, J. Chromatogr. A 2000,
895, 309. doi:10.1016/S0021-9673(00)00703-2
[11] X. Creary, A. J. Rollin, J. Org. Chem. 1977, 42, 4226.
[12] E. Vedejs, D. A. Engler, M. J. Mullins, J. Org. Chem. 1977,
42, 3109.
[13] R. M. Moriarty, W. R. Epa, R. Penmasta, A. K. Awasthi, Tetrahe-
dron Lett. 1989, 30, 667. doi:10.1016/S0040-4039(01)80277-X