Novel 2,5-hexanedione analogues. Substituent-induced control of the protein cross-linking potential and oxidation susceptibility of the resulting primary amine-derived pyrroles

Article abstract of DOI:10.1021/tx000169q

The neurotoxic γ-diketone, 2,5-hexanedione (2,5-HD), induces neurofilamentous swellings at prenodal sites in proximal axons as a consequence of pyrrolation of lysine ε-amino groups on neurofilament proteins. However, there is disagreement as to whether pyrrole formation and the associated alteration of noncovalent interactions is sufficient to cause neurofilament accumulation, or whether pyrrole autoxidation and subsequent protein - protein cross-linking is an obligatory event. To investigate γ-diketones that might form pyrroles inert to autoxidative-induced cross-linking, we synthesized 1,1,1-trifluoro-2,5-hexanedione, 3-(trifluoromethyl)-2,5-hexanedione (3-TFMHD), and two 3-(dialkylaminocarbonyl)-2,5-diketones and assessed their rates of pyrrole formation with amines, the oxidation susceptibility of the resulting pyrroles, and the protein cross-linking potential in vitro, relative to those of 3-methyl-2,5-hexanedione. 1,1,1-Trifluoro-2,5-hexanedione does not form pyrroles, but the three 2,5-HD analogues with an electron-withdrawing 3-substituent all rapidly formed pyrroles that were inert to autoxidation. Although 3-TMFHD nonetheless still induced cross-linking of ribonuclease A, by a nonoxidative mechanism independent of the pyrrole, the two 3-(dialkylaminocarbonyl)-2,5-diketones did not exhibit any protein cross-linking. As these two γ-diketones possess aqueous-organic partitioning properties similar to those of 2,5-HD, they should serve as useful mechanistic probes in further studies.

Full text of DOI:10.1021/tx000169q

264
Chem. Res. Toxicol. 2001, 14, 264-274
Articles
Novel 2,5-Hexa n ed ion e An a logu es. Su bstitu en t-In d u ced
Con tr ol of th e P r otein Cr oss-Lin k in g P oten tia l a n d
Oxid a tion Su scep tibility of th e Resu ltin g P r im a r y
Am in e-Der ived P yr r oles
Guozhang Xu, Malvinder P. Singh, Damodaragounder Gopal, and
Lawrence M. Sayre*
Department of Chemistry, Case Western Reserve University, Cleveland, Ohio 44106
Received August 7, 2000
The neurotoxic γ-diketone, 2,5-hexanedione (2,5-HD), induces neurofilamentous swellings
at prenodal sites in proximal axons as a consequence of pyrrolation of lysine ꢀ-amino groups
on neurofilament proteins. However, there is disagreement as to whether pyrrole formation
and the associated alteration of noncovalent interactions is sufficient to cause neurofilament
accumulation, or whether pyrrole autoxidation and subsequent protein-protein cross-linking
is an obligatory event. To investigate γ-diketones that might form pyrroles inert to autoxidative-
induced cross-linking, we synthesized 1,1,1-trifluoro-2,5-hexanedione, 3-(trifluoromethyl)-2,5-
hexanedione (3-TFMHD), and two 3-(dialkylaminocarbonyl)-2,5-diketones and assessed their
rates of pyrrole formation with amines, the oxidation susceptibility of the resulting pyrroles,
and the protein cross-linking potential in vitro, relative to those of 3-methyl-2,5-hexanedione.
1,1,1-Trifluoro-2,5-hexanedione does not form pyrroles, but the three 2,5-HD analogues with
an electron-withdrawing 3-substituent all rapidly formed pyrroles that were inert to autoxi-
dation. Although 3-TMFHD nonetheless still induced cross-linking of ribonuclease A, by a
nonoxidative mechanism independent of the pyrrole, the two 3-(dialkylaminocarbonyl)-2,5-
diketones did not exhibit any protein cross-linking. As these two γ-diketones possess aqueous-
organic partitioning properties similar to those of 2,5-HD, they should serve as useful
mechanistic probes in further studies.
to be well accepted that 2,5-HD induces modification of
the primary amine side chains of critical lysine residues
of NF or related cytoskeletal proteins (3) to form pyrroles
(4-11). Evidence that pyrrole formation is associated
with neurotoxicity is that 2,5-HD analogues that form
pyrroles more rapidly are more neurotoxic (5, 10, 12-
16). However, at least as far as the axonal swellings are
observed, there is disagreement as to whether pyrrole
formation and the associated alteration of noncovalent
interactions of NF within the NF-microtubule meshwork
is sufficient to cause neurofilament accumulation (6, 7,
9, 16), or whether autoxidation of the pyrroles to reactive
intermediates and subsequent protein-protein cross-
linking is an obligatory event (4, 17, 18).
With regard to the structure-neurotoxicity relation-
ship for 2,5-HD analogues, it has been observed that 2,5-
HD, 3-methyl-2,5-hexanedione (3-MHD), and 3,4-dimethyl-
2,5-hexanedione (3,4-DMHD) induce a progressively
increased level of neurotoxicity and lead to NF accumula-
tion in progressively more proximal axonal segments
(12-14, 16). The enhanced neurotoxic potency of 3-MHD
and 3,4-DMHD could be accounted for by considering the
faster rate of pyrrole formation alone or, in addition, the
faster oxidative conversion to reactive electrophiles.
Therefore, these methylated γ-diketone derivatives do not
In tr od u ction
Chronic exposure to n-hexane and methyl n-butyl
ketone results in a neurotoxic condition classified as a
central-peripheral distal axonopathy (1). These neuro-
logical deficits are similar to inherited or acquired human
neuropathies. Metabolism studies in vitro and in vivo
have revealed that the neurotoxicity of n-hexane and
methyl n-butyl ketone is associated with their common
metabolite 2,5-hexanedione (2,5-HD)1 (1). Although re-
cent studies have suggested that the axonal atrophy also
observed in 2,5-HD neuropathy may be the most consis-
tent morphological hallmark of this condition (2), most
workers have historically focused on the massive focal
accumulation of neurofilaments (NFs) observed at preno-
dal sites in peripheral axons. In either case, it appears
* To whom correspondence should be addressed: Department of
Chemistry, Case Western Reserve University, Cleveland, OH 44106.
Phone: (216) 368-3704. Fax: (216) 368-3006. E-mail: lms3@po.cwru.edu.
1
Abbreviations: 3-AcHD, 3-acetyl-2,5-hexanedione; CAN, cerium
ammonium nitrate; DMAB, 4-dimethylaminobenzoaldehyde; 3,4-
DMHD, 3,4-dimethyl-2,5-hexanedione; 2,5-HD, 2,5-hexanedione; 3-MHD,
3-methyl-2,5-hexanedione; NF, neurofilament; SDS-PAGE, sodium
dodecyl sulfate-polyacrylamide gel electrophoresis; RNase, ribonu-
clease A; TCA, trichloroacetic acid; TFHD, (1,1,1-trifluoromethyl)-2,5-
hexanedione; 3-TFMHD, 3-(trifluoromethyl)-2,5-hexanedione; TLC,
thin-layer chromatography.
10.1021/tx000169q CCC: $20.00 © 2001 American Chemical Society
Published on Web 02/10/2001

Novel 2,5-HD Analogues and the Derived Pyrroles
Chem. Res. Toxicol., Vol. 14, No. 3, 2001 265
clease A from bovine pancreas (type III-A) was obtained from
Sigma (St. Louis, MO) and used as received.
permit a mechanistic distinction between the two theories
rationalizing NF accumulation. However, the two pos-
sible mechanisms should be distinguishable through an
evaluation of γ-diketones that form pyrroles inert to
autoxidative-induced cross-linking. Graham and co-work-
ers chose the 3-acetyl analogue of 2,5-hexanedione (3-
AcHD), which they showed forms pyrroles on proteins
that are resistant to oxidation (18). They further showed
that 3-AcHD is non-neurotoxic, consistent with a critical
role for protein cross-linking in the initiation of neuro-
toxicity (19). However, despite the ability of the 3-acetyl
analogue to pyrrolate blood proteins (18), the lack of
neurotoxicity of the compound could reflect the inability
of this highly water soluble analogue to partition well
into nerves. High water solubility for 3-AcHD arises
because it is a â-diketone analogue of 2,4-pentanedione,
which is known to exist mainly as its enol tautomer (20)
and which has a relatively low pKa of 9 (21). Thus, we
felt it was important to develop additional γ-diketone
mechanistic probes that retained the aqueous-organic
partitioning properties of the hallmark neurotoxic 2,5-
HD.
5,5,5-Tr iflu or o-3-p en ten -2-on e (1). In a minor modification
of the published method (22), a positive pressure of N2 was used
to sweep the gaseous CF3CHO generated in one assembly
through a reflux condenser packed with dry CaCl2 into the
reaction flask containing a suspension of 1-triphenylphosphor-
anylidene-2-propanone in anhydrous ether. The title product
was purified by fractional distillation. The fraction collected at
93 °C [lit. (22) bp 89-91 °C] was used for the next step without
any attempt to separate the geometric isomers: 1H NMR
(CDCl3) δ 2.37 (s, 3H), 6.59-6.66 (2H, olefin CH); 13C NMR
(CDCl3) δ 28.43, 122.46 (q, 1J C-F ) 270 Hz, CF3), 129.2 (q, 2J C-F
3
) 35 Hz, C4), 134.98 (q, J C-F ) 5.8 Hz, C3), 196.04; 19F NMR
(CDCl3) δ -65.95 (d, J H-F ) 5.3 Hz); HRMS calcd for C5H5F3O
m/z 138.0890, found m/z 139.0325 (M + 1, 15% relative
intensity), 138.0297 (M+, 12% relative intensity), 123.0071 (M
- 15, 100% relative intensity).
5-Nitr o-4-(tr iflu or om eth yl)-2-h exa n on e (2). To a solution
of 1 (14 g, 100 mmol) in 100 mL of CH3CH2NO2 was added 30
mL of aqueous Na2CO3 (1.6 g, 15 mmol) solution, and the
resulting mixture was heated at reflux for 6 h under N2. The
reaction mixture was then cooled and partitioned between water
and CHCl3. The combined organic extract was washed with
water, dried (Na2SO4), and evaporated to afford an oil. NMR
(1H, 13C, and 19F) analysis of this oil showed the presence of 2
in the form of a 1:1 mixture of diastereomers, which was used
for the next step without further purification. An analytical
sample for spectral characterization was obtained by silica gel
flash chromatography (1:1 CH2Cl2/EtOAc, v/v): 1H NMR (dia-
stereomeric mixture in CDCl3) δ 1.64 and 1.67 (2d, 3H), 2.31
(s, 3H), 2.87 and 2.90 (2d, 2H), 3.65-3.90 (1H), 4.90 (m, 1H);
13C NMR (CDCl3) δ 15.57 and 16.27 (C6), 29.86 and 29.89 (C1),
We report here the synthesis of 1,1,1-trifluoro-2,5-
hexanedione (TFHD), 3-(trifluoromethyl)-2,5-hexane-
dione (3-TFMHD), and two 3-carboxamide-substituted
2,5-HD analogues, and reactivity characterization of the
derived pyrroles. It was expected that all the 3-substi-
tuted 2,5-HD analogues would undergo Paal-Knorr
condensation somewhat more rapidly than 2,5-HD itself,
but that the resulting pyrroles would exhibit a minimal
autoxidation tendency due to the electron-withdrawing
3-substituent. We compared the relative propensity of
these pyrroles to undergo in vitro autoxidation and
chemical oxidation by potassium persulfate, as well as
their oxidation potentials by cyclic voltammetry. Modi-
fication of proteins by 3-TFMHD, 3-MHD, and the two
3-carboxamide-substituted 2,5-HD analogues is also de-
scribed, and although unexpected protein cross-linking
is observed for 3-TFMHD, the 3-carboxamide-substituted
2,5-HD analogues are devoid of cross-linking potential.
2
37.45 and 38.1 (C3), 41.50 and 42.25 (2q, J C-F ) 25 Hz, C4),
1
79.75 and 80.12 (C5), 126.01 and 126.41 (2q, J C-F ) 280 Hz,
CF3), 202.74 and 203.09 (C2); 19F NMR (CDCl3) δ -68.91 (d, J H-F
) 8.9 Hz), -69.42 (d, J H-F ) 7.6 Hz); HRMS calcd for C7H11F3-
NO3 (M + H)+ m/z 214.0739, found m/z 214.0684.
3-(Tr iflu or om eth yl)-2,5-h exa n ed ion e (3). Nitroketone 2
(10.7 g, 50 mmol) was dissolved in 350 mL of MeOH, and the
solution was cooled to 5 °C. A freshly prepared solution of KOH
(3.08 g, 55 mmol) in 500 mL of MeOH was then added dropwise
over 2 h with vigorous stirring under N2, while maintaining the
reaction temperature at 5 °C. After the addition was complete,
a solution of KMnO4 (6.32 g, 40 mmol) and MgSO4 (6.98 g, 58
mmol) in 700 mL of water was added, while the reaction
temperature was maintained at 0-2 °C. The brown solution was
then stirred for 3 h at 0-5 °C and then filtered through Celite.
The filtrate was concentrated and extracted with CHCl3. The
organic extract was washed with brine, dried (Na2SO4), and
evaporated to obtain an oily residue which was distilled under
reduced pressure to afford 6.3 g of 3 (isolated yield of 69%): bp
53-55 °C (7 mmHg); 1H NMR (CDCl3) δ 2.20 (s, 3H), 2.42 (s,
3H), 2.76 (dd, J ) 18.4 and 2.8 Hz, 1H), 3.32 (dd, J ) 18.4 and
11.0 Hz, 1H), 3.81 (m, 1H); 13C NMR (CDCl3) δ 29.30, 31.39,
Exp er im en ta l P r oced u r es
1
Gen er a l Meth od s. The H, 13C, and 19F NMR spectra were
recorded on Varian-XL 200 or Varian Gemini 300 multinuclear
instruments. Chemical shifts are reported in parts per million
relative to Me4Si or the solvent peak (for 1H and 13C NMR
signals) or CFCl3 as an internal standard (for 19F signals). In
the 13C NMR line listings, attached proton test (APT) designa-
tions are given as (+) or (-) following the chemical shift. Low-
resolution mass spectra were obtained using an HP 5971A mass
selective detector (EI at 70 eV) in tandem with an HP 5890 gas
chromatograph using an HR-I capillary column (5 mm × 0.53
mm). High-resolution mass spectra were obtained at 20 eV on
a Kratos MS-25A instrument. Electrophoresis studies utilized
an SE 200 series minigel castor (Hoefer Scientific Instruments,
San Francisco, CA) and a Fisher Scientific (Pittsburgh, PA)
Biotech FB 570 power supply. Thin-layer chromatography (TLC)
2
1
39.76, 50.34 (q, J C-F ) 25 Hz, C3), 124.69 (q, J C-F ) 279 Hz,
CF3), 200.69, 204.55; 19F NMR (CDCl3) δ -67.12 (d, J H-F ) 8.6
Hz); HRMS calcd for C7H9F3O2 m/z 182.0555, found m/z
182.0552 (M+, 11% relative intensity), 167.0355 (M - 15, 33%
relative intensity), 147.0550 (M - 35, 100% relative intensity).
4-Oxo-1-p en ta n oic Acid P h en yl Meth yl Ester (4). A
mixture of levulinic acid (116 g, 1 mol), benzyl chloride (126.5
g, 1 mol), and Et3N (201 g, 2 mol) in anhydrous toluene (250
mL) was heated under N2 at reflux for 10 h. The mixture was
then cooled, and the solid (Et3N‚HCl) that appeared was
removed by filtration. The organic filtrate was washed with
water, and was evaporated to afford 152 g (74%) of 4 after
purification by fractional vacuum distillation: bp 90 °C (0.2
mmHg); 1H NMR (CDCl3) δ 2.13 (s, 3H), 2.57 (t, J ) 7.6 Hz,
2H), 2.71 (t, J ) 7.6 Hz, 2H), 5.07 (s, 2H), 7.26-7.34 (m, 5H).
was performed on glass plates precoated with silica gel 60 F254
.
Compounds on the developed plate were visualized by UV light
(λ ) 254 nm) or by placing the plate in a chamber filled with
iodine vapor. 3,4-Dimethyl-2,5-hexanedione (DMHD) was a
diastereomeric mixture from Aldrich. 3-Methyl-2,5-hexanedione
(3-MHD) (16) was prepared by mild hydrolysis of ethyl R,â-
diacetylbutyrate, the latter synthesized from ethyl acetoacetate
and 3-chloro-2-butanone. Neopentylamine was from Karl In-
dustries Inc. (Aurora, OH). All other organic reagents were
obtained from Aldrich Chemical Co. (Milwaukee, WI). Ribonu-
4-Eth ylen ed ioxyp en ta n oic Acid P h en yl Meth yl Ester
(5). A mixture of 4 (69.3 g, 0.34 mol) and ethylene glycol (22.9,

266 Chem. Res. Toxicol., Vol. 14, No. 3, 2001
Xu et al.
0.37 mol) in benzene, containing a catalytic amount of p-
toluenesulfonic acid (300 mg), was heated at reflux, with
azeotropic removal of water using a Dean-Stark trap. After the
calculated amount of water (ca. 6 mL) had been collected, the
reaction mixture was cooled and was washed successively with
10% aqueous NaHCO3 and brine. The organic layer was dried
(Na2SO4) and then evaporated to obtain 5, which was purified
by fractional distillation under reduced pressure (68 g, 80%
isolated yield): bp 132 °C (0.25 mmHg); 1H NMR (CDCl3) δ 1.31
(s, 3H), 2.04 (t, J ) 7.6 Hz, 2H), 2.45 (t, J ) 7.6 Hz, 2H), 3.86-
3.91 (m, 4H), 5.11 (s, 2H), 7.32-7.36 (m, 5H).
47.32 (+), 50.03 (+), 164.57 (+), 202.44 (+); GC/MS m/z 169
(M+), 126 (M - C3H7), 84 (M - C5H11N). Then cerium(IV)
ammonium nitrate (CAN, 5.10 g, 9.3 mmol) was dissolved in
15 mL of a solution of 0.78 g of 3-oxobutanoic piperidide (4.65
mmol) and 0.62 g of isopropenyl acetate (6.2 mmol) in methanol.
After the disappearance of the CAN color (∼20 min at room
temperature), a solution of NaHCO3 (0.75 g) in 10 mL of water
was added, and the mixture was heated at 40 °C for 45 min
with rapid stirring. The mixture was cooled, diluted into ice
water, and extracted with CH2Cl2. The combined CH2Cl2
extracts were evaporated, and the crude residue was purified
by silica gel flash chromatography (EtOAc as the eluent) to
afford 0.88 g of desired product 9 (60%): 1H NMR (CDCl3) δ
1.46-1.66 (6H), 2.11 (s, 3H), 2.16 (s, 3H), 2.95 (dd, J ) 6.61
4-Eth ylen edioxy-2-(tr iflu or oacetyl)pen tan oic Acid P h en -
yl Meth yl Ester (6). A flame-dried three-neck flask equipped
with a reflux condenser, an addition funnel, and a gas inlet was
flame dried under N2. A suspension of oil-free NaH (19.7 g, 0.82
mol) in 200 mL of anhydrous benzene was transferred to the
flask under N2, and the addition funnel was charged with 200
mL of a benzene solution containing 5 (51.35 g, 0.21 mol) and
benzyl trifluoroacetate (50.25 g, 0.25 mol). A catalytic amount
of benzyl alcohol (0.5 mL) was added to the flask, and the
solution in the addition funnel was added dropwise over the
course of 2 h at 10 °C. After the addition was complete, the
reaction mixture was stirred under N2 for 3 h at 80 °C. The
mixture was then cooled, and unreacted NaH was destroyed by
cautiously pouring the reaction solution into ice. The organic
layer was separated, and the aqueous layer was extracted with
ether. The organic extracts were combined, washed with brine,
and evaporated to afford 38 g (crude yield of 52%) of a yellow
oil which was distilled under reduced pressure to afford 6: bp
and 3.01 Hz, 2H), 3.47-3.54 (4H), 4.15 (t, J ) 6.65 Hz, 1H); 13
C
NMR (CDCl3) δ 24.47 (+), 25.59 (+), 26.40 (+), 27.88 (-), 30.00
(-), 41.88 (+), 43.52 (+), 47.46 (+), 51.85 (-), 167.16 (+), 202.90
(+), 206.01 (+); GC/MS m/z 225 (M+); HRMS calcd for C12H19
-
1
NO3 m/z 225.1366, found m/z 225.1362. The H NMR spectrum
of 9 in CD3OD exhibited no evidence of deuterium exchange in
a 4 h period, indicating that the keto form of this compound
does not equilibrate with the enol from in this polar solvent.
N,N-Dim eth yl-2-a cetyl-4-oxop en ta n a m id e (10). A mix-
ture of 0.81 g of 80% aqueous N,N-dimethylacetoacetamide (5
mmol) and 0.55 g of isopropenyl acetate was added together to
15 mL of acetic acid solution containing 2.7 g of Mn(OAc)3‚2H2O
at 70 °C. The reaction was complete in 10 min, and the resulting
solid was filtered off. The brown reaction mixture was concen-
trated and subjected to flash chromatography (6:1 EtOAc/
MeOH, v/v, as the eluent) to afford 360 mg of pure 10 (39%
yield): 1H NMR (CDCl3) δ 2.11 (s, 3H), 2.15 (s, 3H), 2.93 (s,
3H), 2.95 (d, J ) 6.84 Hz, 2H), 3.11 (s, 3H), 4.13 (t, J ) 6.71
Hz, 1H); 13C NMR (CDCl3) δ 27.96 (-), 29.96 (-), 36.03 (-),
37.87 (-), 41.93 (+), 52.02 (-), 169.00 (+), 202.80 (+), 206.10
(+); GC/MS m/z 185 (M+); HRMS calcd for C9H15NO3 m/z
185.1053, found m/z 185.1052. The 1H NMR spectrum of 10 in
CD3OD exhibited no evidence of deuterium exchange in a 4 h
period, indicating that the keto form of this compound does not
equilibrate with the enol from in this polar solvent.
1H NMR Exp er im en ts. All kinetic runs were performed
under N2 in a thermostated probe maintained at 25 ( 0.1 °C.
Standard stock solutions (0.2 M) of 4-(aminomethyl)pyridine,
isobutylamine, or neopentylamine, and of the respective γ-dike-
tone (0.2 M), were prepared in N2-flushed CDCl3 and stored
under N2. The required amounts of amine and γ-diketone
solutions were preequilibriated to 25 °C, mixed, and im-
mediately transferred to a 5 mm NMR tube. Data were
accumulated at regular intervals using the following acquisition
parameters: 64 scans; acquisition time, 1 s/scan; and relaxation
delay, 6 s/scan. For estimating the half-lives, the plots of integral
percentages of N-CH2 signals (amine versus pyrrole) as a
function of time were constructed; the half-life corresponded to
the time when the pyrrole N-CH2 integral equaled half the
original amine N-CH2 integral. For the reactions of isobuty-
lamine and neopentylamine with 9, the N-CH2 signal over-
lapped with the piperidide R-signals, and the reaction course
was instead monitored by observing conversion of the downfield
diketone methyl signal to the downfield pyrrole methyl signal.
No kinetic analyses were performed.
1
140-142 °C (0.2 mmHg); H NMR (CDCl3) δ 1.32 (s, 3H), 2.49
(dd, J ) 14.0 and 4.0 Hz, 1H), 2.67 (dd, J ) 14.0 and 10.0 Hz,
1H), 3.82-3.92 (m, 4H, OCH2CH2O), 4.18 (dd, J ) 10.0 and 4.0
Hz, 1H), 5.15-5.18 (m, 2H, PhCH2), 7.29-7.40 (m, 5H).
4-Oxo-2-(tr iflu or oa cetyl)p en ta n oic Acid P h en yl Meth yl
Ester (7). A solution of 6 (16.75 g, 48 mmol) in 80 mL of an
acetone/water mixture (2:1) containing 100 mg of p-toluene-
sulfonic acid was heated at reflux for 10 h. The mixture was
cooled and evaporated to dryness to afford an oily residue, which
was partitioned between water and CHCl3. The CHCl3 extracts
were dried (Na2SO4) and evaporated to afford 13 g (89%) of the
crude product, which was used for the next step without
purification. An analytical sample for spectral characterization
was obtained by silica gel flash chromatography (1:1 EtOAc/
hexane, v/v, as the eluent): 1H NMR (CDCl3) δ 2.14 (s, 3H),
3.11 (dd, J ) 16 and 4.2 Hz, 1H), 3.35 (dd, J ) 16 and 10.8 Hz,
1H), 4.27 (dd, J ) 10.8 and 4.2 Hz, 1H), 5.10 (s, 2H), 7.30-7.38
(m, 5H).
1,1,1-Tr iflu or o-2,5-h exa n ed ion e (8). To a solution of 7 (13
g, 43 mmol) in 50 mL of EtOAc was added 10% Pd/C (1.8 g),
and H2 was bubbled through the mixture for 8 h at 65 °C. The
progress of the reaction was monitored by TLC (conversion of
Rf from 0.6 to 0.4, CH2Cl2 as the eluent). At the completion of
the reaction, the mixture was filtered through Celite to remove
the catalyst, and the filtrate was concentrated to ∼10 mL by
evaporation at 25 °C. The residue obtained was distilled under
reduced pressure, and the fraction collected at 48-50 °C (5
mmHg) corresponded to the title compound as revealed by 1H
and 19F NMR (4.3 g, 60% isolated yield): 1H NMR (CDCl3) δ
2.22 (s, 3H), 2.89 (t, J ) 7.0 Hz, 2H), 3.02 (t, J ) 7.0 Hz, 2H);
19F NMR (CDCl3) δ -80 (s). Compound 8 has also been prepared
by another method (23).
13C NMR Exp er im en ts. In a typical run, 3 mmol of γ-dike-
tone was dissolved in 0.45 mL of N2-purged CDCl3 and equili-
brated to 22 °C in a 10 mL NMR tube. A CDCl3 solution (0.45
mL) containing the desired amine (3 mmol) was then added,
and the spectra were recorded at regular time intervals. At the
completion of the reactions, indicated by the disappearance of
the signals corresponding to the starting materials, aliquots
were analyzed for pyrrole products by their chromatographic
and 1H NMR characteristics. In the case of 3-TFMHD, ad-
ditional confirmation was afforded by the 19F NMR spectrum.
2-Acetyl-4-oxop en ta n oic P ip er id id e (9). A mixture of 6.5
g of ethyl acetoacetate and 8.5 g of piperidine was refluxed for
48 h, and then distilled under reduced pressure (1 mmHg).
Collection of the bp 119 °C fraction afforded 7 g of 3-oxobutanoic
1
piperidide (83% yield). H NMR (CDCl3) indicated a mixture of
the keto form (86.6%) [δ 1.42-1.61 (m, 6H), 2.19 (s, 3H), 3.27
(t, J ) 5.37 Hz, 2H), 3.48 (s, 2H), 3.49 (t, J ) 5.82 Hz, 2H)] and
the enol form (13.4%) (characteristic singlets at 1.87 and 5.09
ppm in a 3:1 ratio for the methyl and vinyl signal, respectively).
13C NMR (CDCl3) data for the major keto tautomer are as
follows: δ 24.16 (+), 25.30 (+), 26.13 (+), 29.94 (-), 42.68 (+),
Gen er a l P r oced u r e for P r ep a r a tion of P yr r oles fr om
γ-Dik eton es. According to the general procedure (5), a solution
of the appropriate γ-diketone (2 mmol) in 10 mL of EtOH was

Novel 2,5-HD Analogues and the Derived Pyrroles
Chem. Res. Toxicol., Vol. 14, No. 3, 2001 267
mixed with a solution of the required amine (2.1 mmol) in 10
mL of EtOH, and the resulting mixture was stirred under N2
at 50 °C for 10 h. The solvent was then evaporated, and the
residue obtained was partitioned between water and ether. The
ether extracts were washed several times with water, dried (Na2-
SO4), and evaporated to afford the desired pyrrole product,
which was dissolved in 50 mL of EtOAc and passed through a
short column packed with alumina. Evaporation of EtOAc then
provided the pure pyrrole products in the form of colorless oils.
Isolated yield were 80-95%. Pyrrole 14c was prepared in CH3-
CN instead of in EtOH. Storage of these products, even under
N2 and low temperaturs, led to darkening, and thus, the pyrroles
were prepared immediately prior to use for the chemical and
electrochemical oxidation studies.
1-(2-Meth ylp r op yl)-2,5-d im eth ylp yr r ole (11a ) (24): 1H
NMR (CDCl3) δ 0.90 (d, J ) 6.8 Hz, 6H), 1.97 (m, 1H), 2.20 (s,
6H), 3.52 (d, J ) 7.6 Hz, 2H), 5.77 (s, 2H); 13C NMR (CDCl3) δ
12.20, 19.46, 29.32, 50.32, 104.52, 127.01.
1-(4-P yr idin ylm eth yl)-2,5-dim eth ylpyr r ole (11b): 1H NMR
(CDCl3) δ 2.21 (s, 6H), 5.00 (s, 2H), 5.89 (s, 2H), 6.80 (d, J ) 6.2
Hz, 2H), 8.53 (d, J ) 6.2 Hz, 2H); 13C NMR (CDCl3) δ 12.16,
45.60, 105.78, 120.58, 127.57, 147.55, 149.99; HRMS calcd for
C12H14N2 m/z 186.1157, found m/z 186.1158.
1-(2,2-Dim et h ylp r op yl)-2,5-d im et h ylp yr r ole (11c): 1H
NMR (CDCl3) δ 0.97 (s, 9H), 2.33 (s, 6H), 3.75 (s, 2H), 5.80 (s,
2H); HRMS calcd for C11H19N m/z 165.1517, found m/z 165.1513.
1-(2-Meth ylp r op yl)-2,3,5-tr im eth ylp yr r ole (12a ): 1H NMR
(CDCl3) δ 0.90 (d, J ) 6.8 Hz, 6H), 1.98 (m, 1H), 1.99 (s, 3H),
2.10 (s, 3H), 2.17 (s, 3H), 3.49 (d, J ) 7.6 Hz, 2H), 5.80 (s, 1H);
13C NMR (CDCl3) δ 9.98, 11.06, 12.39, 19.95, 29.88, 49.94,
106.91, 112.80, 123.34, 126.08; HRMS calcd for C11H19N m/z
165.1517, found m/z 165.1522.
1-(4-P yr id in ylm eth yl)-2,3,5-tr im eth ylp yr r ole (12b): 1H
NMR (CDCl3) δ 2.01 (s, 3H), 2.02 (s, 3H), 2.09 (s, 3H), 4.96 (s,
2H), 5.78 (s, 1H), 6.80 (d, J ) 6.2 Hz, 2H), 8.52 (d, J ) 6.2 Hz,
2H); 13C NMR (CDCl3) δ 9.53, 11.05, 11.81, 45.71, 107.75, 113.81,
120.63, 123.36, 126.22, 147.89, 149.89; HRMS calcd for C13H16N2
m/z 200.1313, found m/z 200.1315.
1.9 g of neopentylamine (21.6 mmol), and 1.2 g of acetic acid
(21.6 mmol) in 10 mL of CH3CN was stirred under nitrogen.
The mixture slowly turned red over the course of 12 h, and
pyrrole formation was monitored by TLC (1:2 EtOAc/hexane,
v/v, as the eluent). After an additional 12 h, the solvent was
evaporated to yield a brown residue, which was partitioned
between water and ether. The ether extract was washed twice
with water, dried (Na2SO4), and evaporated to give crude pyrrole
14c, which was purified by flash chromatography (1:3 EtOAc/
hexane, v/v, as the eluent) (850 mg yield, 25%). Alternatively,
addition of CsF (1 g) to a solution of 3-TFMHD (110 mg) and
neopentylamine (95 mg) in 10 mL of a 3:1 CH3CN/H2O mixture,
v/v, resulted in the generation of pyrrole 14c in almost quan-
titative yield: 1H NMR (CDCl3) δ 0.98 (s, 9H), 2.21 (s, 3H), 2.30
(s, 3H), 3.64 (s, 2H), 6.05 (s, 1H); 13C NMR (CDCl3) δ 11.96 (-),
13.56 (-), 28.62 (-), 34.81 (+), 54.17 (+), 104.68 (-), 110.23 (+,
2
1
q, J C-F ) 34.5 Hz), 122.93 (+), 125.08 (+, q, J C-F ) 264.8 Hz,
CF3), 128.76 (+); MS m/z 233 (M+), 218 (M - CH3), 214 (M -
F), 176 [M - C(CH3)3]; HRMS calcd for C12H18NF3 m/z 233.1392,
found m/z 233.1395.
1-(2-Meth ylpr opyl)-3-(1-piper idin ocar bon yl)-2,5-dim eth -
1
ylp yr r ole (15a ): H NMR (CDCl3) δ 0.84 (d, J ) 6.7 Hz, 6H),
1.42-1.64 (6H), 1.86-2.01 (m, 1H), 2.11 (s, 3H), 2.21 (s, 3H),
3.45 (d, J ) 7.7 Hz, 2H), 3.51 (t, J ) 5.2 Hz, 4H), 5.77 (s, 1H);
13C NMR (CDCl3) δ 11.53 (-), 12.65 (-), 20.07 (-), 24.91 (+),
26.31 (+), 29.68 (-), 50.93 (+), 105.95 (-), 114.44 (+), 126.90
(+), 129.59 (+), 168.15 (+); HRMS calcd for C16H26N2O m/z
262.2045, found m/z 262.2045.
1-(4-P yr id in ylm et h yl)-3-(1-p ip er id in oca r b on yl)-2,5-d i-
m eth ylp yr r ole (15b): 1H NMR (CDCl3) δ 1.47-1.67 (6H), 2.05
(s, 3H), 2.16 (s, 3H), 3.54-3.59 (4H), 4.96 (s, 2H), 5.93 (s, 1H),
6.79 (d, J ) 5.53 Hz, 2H), 8.49 (d, J ) 5.50 Hz, 2H); 13C NMR
(CDCl3) δ 11.09 (-), 12.10 (-), 24.87 (+), 26.32 (+), 45.88 (+),
106.71 (-), 115.46 (+), 120.78 (-), 127.04 (+), 129.58 (+), 146.90
(+), 150.30 (-), 167.69 (+); HRMS calcd for C18H23N3O m/z
297.1841, found m/z 297.1854.
1-(2,2-Dim eth ylp r op yl)-3-(1-p ip er id in oca r bon yl)-2,5-d i-
1
m eth ylp yr r ole (15c): H NMR (CDCl3) δ 0.98 (s, 9H), 1.52-
1-(2,2-Dim eth ylp r op yl)-2,3,5-tr im eth ylp yr r ole (12c): 1H
NMR (CDCl3) δ 1.01 (s, 9H), 2.04 (s, 3H), 2.17 (s, 3H), 2.25 (s,
3H), 3.60 (s, 2H), 5.79 (s, 1H); 13C NMR (CDCl3) δ 11.47 (-),
13.63 (-), 28.74 (-), 29.02 (-), 34.82 (+), 54.44 (+), 107.75 (-),
113.56 (+), 124.80 (+), 127.55 (+); MS m/z 179 (M+), 164 (M -
CH3), 122 [M - C(CH3)3]; HRMS calcd for C12H21N m/z 179.1675,
found m/z 179.1674; HRMS calcd for C12H20N (M - H)+ m/z
178.1597, found m/z 178.1596.
1-(2-Meth ylpr opyl)-2,3,4,5-tetr a m eth ylpyr r ole (13a ) (24):
1H NMR (CDCl3) δ 0.90 (d, J ) 6.8 Hz, 6H), 1.94 (s, 6H), 2.11
(s, 6H), 1.99 (m, 1H), 3.49 (d, J ) 7.6 Hz, 2H); 13C NMR (CDCl3)
δ 8.84, 9.59, 19.61, 29.64, 50.43, 112.08, 121.85.
1-(4-P yr id in ylm eth yl)-2,3,4,5-tetr a m eth ylp yr r ole (13b):
1H NMR (CDCl3) δ 1.97 (s, 6H), 2.02 (s, 6H), 4.96 (s, 2H), 6.80
(d, J ) 6.2 Hz, 2H), 8.51 (d, J ) 6.2 Hz, 2H); 13C NMR (CDCl3)
δ 9.30, 9.54, 45.71, 113.62, 120.69, 122.26, 148.30, 149.82; HRMS
calcd for C14H18N2 m/z 214.1470, found m/z 214.1480.
1-(2,2-Dim eth ylp r op yl)-2,3,4,5-tetr a m eth ylp yr r ole (13c):
1H NMR (CDCl3) δ 0.99 (s, 9H), 1.99 (s, 6H), 2.17 (s, 6H), 3.6 (s,
2H).
1.67 (m, 6H), 2.20 (s, 3H), 2.28 (s, 3H), 3.57 (br, 4H), 3.59 (s,
2H), 5.87 (s, 1H); 13C NMR (CDCl3) δ 12.70 (-), 13.70 (-), 24.96
(+), 26.36 (+), 28.76 (-), 34.80 (+), 54.23 (+), 106.62 (-), 114.98
(+), 128.12 (+), 130.32 (+), 168.31 (+); GC/MS m/z 276 (M+);
HRMS calcd for C17H28N2O m/z 276.2203, found m/z 276.2203.
1-(2-Meth ylp r op yl)-3-[(N,N′-d im eth yla m in o)ca r bon yl]-
2,5-d im eth ylp yr r ole (16a ): 1H NMR (CDCl3) δ 0.89 (d, J )
6.7 Hz, 6H), 1.99 (m, 1H), 2.16 (s, 3H), 2.28 (s, 3H), 3.04 (s, 6H),
3.51 (d, J ) 7.7 Hz, 2H), 5.87 (s, 1H); 13C NMR (CDCl3) δ 11.67
(-), 12.68 (-), 20.09 (-), 29.70 (-), 50.94 (+), 106.33 (-), 114.27
(+), 126.88 (+), 130.15 (+), 169.42 (+), N-CH3 signal not
observed; HRMS calcd for C13H22N2O m/z 222.1732, found m/z
222.1728.
1-(4-P yr idin ylm eth yl)-3-[(N,N′-dim eth ylam in o)car bon yl]-
2,5-d im eth ylp yr r ole (16b): 1H NMR (CDCl3) δ 2.08 (s, 3H),
2.21 (s, 3H), 3.08 (s, 6H), 4.99 (s, 2H), 6.01 (s, 1H), 6.81 (d, J )
5.50 Hz, 2H), 8.53 (d, J ) 5.94 Hz, 2H); 13C NMR (CDCl3) δ
11.18 (-), 12.09 (-), 45.87 (+), 107.01 (-), 115.23 (+), 120.78
(-), 127.00 (+), 130.05 (+), 146.95 (+), 150.22 (-), 168.99 (+),
N-CH3 signal not observed; HRMS calcd for C15H19N3O m/z
257.1528, found m/z 257.1528.
1-(2,2-Dim eth ylp r op yl)-3-[(N,N′-d im eth yla m in o)ca r bo-
n yl]-2,5-d im eth ylp yr r ole (16c): 1H NMR (CDCl3) δ 0.92 (s,
9H), 2.15 (s, 3H), 2.25 (s, 3H), 3.01 (s, 6H), 3.53 (s, 2H), 5.86 (s,
1H); 13C NMR (CDCl3, -30 °C) δ 12.67, 13.55, 28.48, 34.64,
34.99, 39.30, 53.90, 106.61, 114.21, 127.95, 130.54, 169.22; 13C
NMR (CDCl3, 20 °C) δ 12.60, 13.49, 28.48, 34.62, 39.30, 54.00,
106.76, 114.61, 127.87, 130.56, 169.35, N-CH3 signal not seen
in 13C NMR at room temperature; GC/MS m/z 236 (M+); HRMS
calcd for C14H24N2O m/z 236.1890, found m/z 236.1886.
Kin etic Assessm en ts of th e Oxid a tion of P yr r oles (11c-
16c) by K2S2O8. Kinetic experiments were carried out in 1 cm
quartz cells in the thermostated cell compartment of a Perkin-
Elmer Lambda 3B spectrophotometer. The temperature was
1-(2-Meth ylp r op yl)-2,5-d im eth yl-3-(tr iflu or om eth yl)p yr -
r ole (14a ): 1H NMR (CDCl3) δ 0.89 (d, J ) 6.8 Hz, 6H), 1.95
(m, 1H), 2.16 (s, 3H), 2.25 (s, 3H), 3.52 (d, J ) 7.6 Hz, 2H), 5.98
(s, 1H); 13C NMR (CDCl3) δ 10.38, 12.07, 19.50, 29.37, 50.57,
2
103.62, 105.17 (q, J C-F ) 35.6 Hz), 121.59, 127.42, 129.85 (q,
1J C-F ) 264.8 Hz, CF3); 19F NMR (CDCl3) δ -55.75 (s); HRMS
calcd for C11H16NF3 m/z 219.1235, found m/z 219.1237.
1-(4-P yr id in ylm eth yl)-2,5-d im eth yl-3-(tr iflu or om eth yl)-
p yr r ole (14b): 1H NMR (CDCl3) δ 2.10 (s, 3H), 2.20 (s, 3H),
5.02 (s, 2H), 6.12 (s, 1H), 6.82 (d, J ) 6.2 Hz, 2H), 8.54 (d, J )
6.2 Hz, 2H); 19F NMR (CDCl3) δ -54.6 (s); HRMS calcd for
C13H13N2F3 m/z 254.1031, found m/z 254.1027.
1-(2,2-Dim eth ylpr opyl)-2,5-dim eth yl-3-(tr iflu or om eth yl)-
p yr r ole (14c). A solution of 2.2 g of 3-TFMHD (14.4 mmol),

268 Chem. Res. Toxicol., Vol. 14, No. 3, 2001
Xu et al.
Sch em e 1a
maintained at 30 °C. The reactions were initiated under pseudo-
first-order conditions by addition of freshly prepared MeOH
solutions of pyrrole derivatives (11c-16c) to thermally equili-
brated solutions of a large excess of aqueous K2S2O8 containing
constant buffer concentrations. The final reaction solutions
were: 1:2, v/v, MeOH/H2O (doubly distilled); [pyrrole] ) 0.20
mM; [S2O82-] ) 20 mM; 66.6 mM potassium phosphate buffer
at pH 6.85; 30 °C. All reactions exhibited monotonic spectral
changes with an increase in absorbance near 300 nm. For some
compounds, the increase near 300 nm was immense, and the
rate was more conveniently monitored at an additional λmax seen
in these cases to grow in parallel at a longer wavelength.
Reaction half-lives were determined as the time needed for
reaching half the plateau ∆A at the following λmax values: 476
nm for 11c, 490 nm for 12c, 507 nm for 13c, 291 nm for 14c,
352 nm for 15c, and 306 nm for 16c.
a
(a) Polyphosphoric acid, 180 °C; (b) Ph3PdCHCOCH3, Et2O;
(c) EtNO2, aqueous Na2CO3; (d) aqueous KMnO4, KOH.
H2SO4 and HNO3 for several days. The cell was drained and
repeatedly rinsed with deionized water. Before each electro-
chemical run, the cell was steamed for 1-2 h and then used
directly after cooling. The working electrode was mechanically
polished with R-Al2O3 before every use. Sample solutions
consisted of freshly prepared 2 mM pyrrole (11a -14a ) in CH3-
CN containing 0.1 M tetra-n-butylammonium perchlorate, and
the solutions were purged with Ar for 30 min before each run.
Voltammograms were recorded at a scan rate of 100 mV/s.
Mod ifica tion of RNa se A by γ-Dik eton es. RNase A (final
concentration, 10 mg/mL) was dissolved in 0.1 M sodium
phosphate buffer (pH 7.4) and then incubated at 25 °C in the
dark with 3.75, 7.5, 15, or 30 mM 2,5-HD, 3-MHD, 3-TFMHD
(3), 9, or 10. Detection and quantitation of the resulting pyrrole
adducts (for 2,5-HD, 9, and 10) were performed by the 4-(di-
methylamino)benzaldehyde (DMAB)-based colorimetric assay.
The procedure involved mixing 300 µL of the modified protein
samples from 4 or 24 h incubations with 300 µL of DMAB
reagent (80 mM, prepared freshly by dissolving 1.19 g of DMAB
in 100 mL of 50% aqueous methanol containing 1% concentrated
HCl), incubating the mixture at 40 °C for 10 min, and then
diluting to 1 mL with water. The UV/visible spectrum of each
sample was recorded, and the absorbances at 523 nm were
converted to pyrrole concentrations using an ꢀ of either 301 (for
2,5-HD) or 280 M-1 cm-1 (for 9 and 10) obtained from conducting
the same assay on known concentrations of either 11a or 15c
and 16c (the same value was obtained for the latter two). The
control sample, prepared without γ-diketone, was used as the
blank.
For the analysis of cross-linking, the incubations with 3-MHD
or 3-TFMHD (3) (24 h) or 9 or 10 (48 h) were followed by dialysis
against 500 mL of 0.1 M phosphate buffer (pH 7.4) twice. For
SDS-PAGE analysis, aliquots or solutions of native RNase A
were mixed with an equal volume of standard denaturing buffer
[0.3 M Tris (pH 6.8), 5% SDS, 5% 2-mercaptoethanol, 25%
glycerol, and 0.1% bromophenol], the mixture was heated to 100
°C for 5 min, and aliquots containing 10-13 µg of protein were
loaded onto each lane of an analytical gel (10% acrylamide
separating gel, 4% stacking gel). The gel was run under constant
current (20 mA) at room temperature. Separated proteins were
fixed in 10% TCA and stained with Coomassie Blue G-250.
Deter m in a tion of P a r tition Coefficien ts for 2,5-HD, 9,
a n d 10. A 48.4 mM solution of each diketone in 1-octanol (8
mL) was prepared, and half of each was mixed with an equal
volume of distilled water. After vigorous shaking, the 1-octanol
layer was separated, and the absorbance at 272 nm was
measured in the starting (s) and final (f) 1-octanol layers. The
partition coefficient was calculated as Af/(As - Af).
Au toxid a tion of P yr r oles 11c, 12c, a n d 14c-16c. Each
pyrrole (1 mmol) was dissolved in 10 mL of an acetonitrile/
sodium phosphate buffer solution (0.1 M, pH 7.2) (4:1, v/v). Air
was bubbled through the solutions at 25 °C for up to 168 h in
the dark, and the reaction was monitored by TLC (EtOAc as
the eluent) and GC/MS (described above). For pyrroles 14c-
16c, the solutions remained yellow, and TLC and GC/MS
analysis indicated the complete absence of oxidation during the
168 h incubation period. For pyrrole 11c, some darkening of
the reaction mixture was evident after 48 h, and GC/MS
analysis indicates the production of a low level of dehydrodimer
(2M - 2). 1H NMR analysis of the dimeric material obtained
following chromatographic separation was consistent with an
unsymmetrical dimer, but structural identification was not
pursued. In contrast, for pyrrole 12c, the color of the solution
changed from yellow to red in 20 min, and then to brown. After
16 h, the dark solution contained some insoluble dark residue.
At this time, GC/MS analysis (temperature program, from an
initial temperature of 40 °C, increasing at a rate of 20 °C/min)
indicated the formation of three isomers with a molecular weight
gain of 14 (m/z 193), one compound with a molecular weight
gain of 32 (m/z 211), and several dehydrodimers (m/z 356).
Several products were isolated by preparative TLC (EtOAc as
the eluent), two of which (the m/z 211 compound and one of the
m/z 193 isomers) were obtained in analytically pure form
(below). The dehydrodimer isomers in this case readily under-
went further oxidative decomposition so that no pure samples
could be obtained for structural analysis.
N-(2,2-Dim eth ylp r op yl)-3-a ceta m id o-2-m eth yl-2-bu ten -
a l 18 (E/Z ster eoch em istr y u n k n ow n ): 1H NMR (CDCl3) δ
0.98 (s, 9H), 2.12 (s, 3H), 2.21 (s, 3H), 2.49 (s, 3H), 3.63 (s, 2H),
9.94 (s, 1H); 13C NMR (CDCl3) δ 9.91 (-), 10.77 (-), 12.03 (-),
28.65 (-), 34.83 (+), 116.16 (+), 147.80 (+), 165.50 (+), 185.86
(-); MS m/z 211 (M+), 154 [M - C(CH3)3].
P yr r oleca r b oxa ld eh yd e 17 (on e isom er of u n k n ow n
1
str u ctu r e): H NMR (CDCl3) δ 0.98 (s, 9H), 2.04 (s, 3H), 2.19
(s, 3H), 3.52 (br, 2H), 6.71 (s, 1H), 9.31 (s, 1H); 13C NMR (CDCl3)
δ 11.11 (-), 11.38 (-), 28.08 (-), 34.94 (+), 54.78 (+), 107.2 (-),
118.4 (+), 123.6 (+), 131.2 (+), 177.82 (-); MS m/z 193 (M+),
137, 136 [M - C(CH3)3].
In a separate experiment to study the effect of added thiol
on pyrrole autoxidation, a solution containing 2-mercaptoethanol
(0.7 mL, 10 mmol) and either pyrrole 12c, 14c, 15c, or 16c (2
mmol) in 8 mL of acetonitrile was bubbled with air, and the
reaction progress was monitored by GC/MS.
Electr och em ica l Mea su r em en ts. Cyclic voltammetric mea-
surements were taken with a PAR model 173 potentiostat/
galvanostat and a PAR model 175 programmer attached to a
PAR model EG8G XY recorder. The measurements were made
Resu lts a n d Discu ssion
Syn th esis of F lu or in a ted γ-Dik eton es. 3-TFMHD
(3) was synthesized via the Michael addition strategy
shown in Scheme 1. Thus 1,4-addition of nitroethane
carbanion to 5,5,5-trifluoropent-3-en-2-one (1), prepared
by a modification of a published procedure (22), afforded
a diastereomeric mixture of nitroketone 2, which under
mild “Nef” reaction conditions (25) was oxidized to 3. We
in
a glass cell with three compartments for the working
electrode, a Pt foil counter electrode, and an SCE reference
electrode. The working electrode was a glassy carbon disk (Pine
Instrument Co.) with an area of 0.46 cm2. The glass cell was
initially degreased in 6 M KOH for 1 day followed by thorough
washing with ultrapure water (reverse osmosis and distillation).
The cell was then filled with a 1:1 (v/v) mixture of concentrated

Novel 2,5-HD Analogues and the Derived Pyrroles
Chem. Res. Toxicol., Vol. 14, No. 3, 2001 269
Ta ble 1. Rela tive Ra tes of P yr r ole F or m a tion fr om Con d en sa tion of γ-Dik eton es w ith P r im a r y Am in esa
a
b
Reactions were carried out in CDCl3 at 25 °C with diketone and amine concentrations of 0.1 M. Data for the mixture of dl- and
meso-diastereomers. Thus, the half-life value reflects mainly the rate of pyrrole formation for the more reactive dl-isomer (34). c Integration
in these cases depended on the ratio of i-Pr or tert-butyl CH3 signals (amine vs pyrrole) because the pyrrole N-CH2 signal in the product
overlapped with the piperidide R-signals.
Sch em e 2a
Sch em e 3a
a
(a) Isopropenyl acetate, CAN or Mn(OAc)3/HOAc.
a
(a) PhCH2Br/Et3N, PhCH3, 80 °C; (b) HOCH2CH2OH/p-TsOH,
was achieved through one-pot reaction of the correspond-
ing acetoacetamide and isopropenyl acetate in the pres-
ence of cerium ammonium nitrate (CAN) (29) or manga-
nic acetate (30) (Scheme 3). The selective oxidation of the
enolizable acetoacetamides by Mn(III) or Ce(IV), subse-
quent free radical addition to isopropyl acetate, and facile
elimination of acetyl radical from the acetoxyalkyl adduct
radical are believed to be involved (30) in this simple one-
step synthesis of γ-diketones 9 and 10.
benzene; (c) NaH/CF3COOCH2Ph; (d) acetone, p-TsOH; (e) H2,
catalytic Pd/C.
also attempted to prepare 3 via traditional coupling
methods for the preparation of unsymmetrical γ-dike-
tones, employing 4,4,4-trifluoro-2-butanone prepared elec-
trochemically according to the procedure of Muller (26).
For example, two methods based on enol silyl ethers (27)
were attempted, whereby the enol silyl ether obtained
from 4,4,4-trifluorobutanone by reaction with triethyl-
amine and trimethylsilyl chloride was allowed to react
with either 2-nitropropene or chloroacetone, but these
reactions were unsuccessful.
TFHD (8) was synthesized from commercially available
levulinic acid by the multistep sequence summarized in
Scheme 2. Levulinic acid was converted to its benzyl ester
4, followed by protection of the keto group in the form of
a ketal 5. The sodium hydride-mediated Claisen conden-
sation with benzyl trifluoroacetate (28) followed by
deprotection of the keto group afforded the diketoester 7
containing the trifluoroacetyl group. Removal of the
benzyl group by catalytic hydrogenation occurred with
concomitant decarboxylation, giving the desired product
8, which was purified by fractional distillation in 16.4%
overall yield.
P yr r ole F or m a tion in th e Rea ction of γ-Dik eton es
w ith P r im a r y Am in es. The neurotoxicity of γ-diketones
is known to be associated with modification of critical
lysine residues of NF or related cytoskeletal proteins (3,
9, 11). In view of our goal of selecting γ-diketone
analogues to distinguish the primary chemical events
that are responsible for neurotoxicity, we first monitored
1
pyrrole formation by H and 13C NMR (and by 19F NMR
in the case of 3) for the reactions of the various γ-dike-
tones (2,5-HD, 3-MHD, 3,4-DMHD, 3, 9, and 10) with
three different primary amines. In each case, the dike-
tone and the model amine were mixed together in CDCl3
1
in a constant 1:1 molar ratio, and the H NMR spectra
were recorded at regular time intervals. The spectral
evolution indicated the buildup and decay of intermedi-
ates, consistent with no single step being entirely rate-
limiting in pyrrole formation. We thus used the half-lives
On e-P ot Syn th esis of 3-[(Dia lk yla m in o)ca r bon yl]-
2,5-h exa n ed ion es. Synthesis of γ-diketones 9 and 10

270 Chem. Res. Toxicol., Vol. 14, No. 3, 2001
Xu et al.
Sch em e 4
Sch em e 5
of formation of the final product as a comparative kinetics
index (Table 1). This was most conveniently determined
by monitoring the N-CH2 methylene signal, which
exhibited a chemical shift difference of 1-1.2 ppm
(downfield) upon going from amine to pyrrole, though in
the case of compound 9 reacting with isobutylamine and
neopentylamine, a different signal was utilized (footnote
c in Table 1).
the adducts formed from 8 was not attempted. It is
known that attachment of CF3 directly to carbonyl groups
destabilizes the sp2 hybridization to the point that
molecules such as hexafluoroacetone, CF3CHO, and other
trifluoromethyl ketones exist in solution primarily as sp3
hydrates (36, 37). The initial carbinolamines formed from
attack of the primary amine at the CF3C(dO) group may
resist dehydration and instead cyclize directly to the
corresponding 2,5-dihydroxypyrrolidines or 2-amino-5-
hydroxytetrahydrofurans, which may or may not dehy-
drate at the non-CF3 carbonyl position (Scheme 5).
Ch em ica l a n d Electr och em ica l Oxid a tion of P yr -
r oles. To permit distinction between pyrrole formation
alone and subsequent autoxidation as the crucial factor
governing neurotoxicity of γ-diketones, we evaluated the
relative ease of oxidation of pyrroles 11-16. If pyrrole
oxidation and subsequent cross-linking is the crucial
determinant, then the ease of autoxidation and mecha-
nism of cross-linking would be most physiologically
relevant. In the latter regard, since there is evidence both
for the generation of pyrrole-pyrrole oxidative coupling
(38, 39) and for side chain nucleophile trapping of
electrophilic pyrrole oxidation intermediates (40), we
focused only on the relative ease of oxidation. The nature
of pyrrole autoxidation (41, 42) is a complex, poorly
understood issue, but the rate-limiting step appears to
involve one-electron transfer to molecular oxygen (43).
Furthermore, a direct correlation has been found between
the rate of autoxidative consumption of pyrroles and their
anodic oxidation (43). Increased electron density due to
an increased degree of substitution by alkyl groups is
found to lower the anodic oxidation potentials, whereas
electron-withdrawing substituents were known to lead
to increased anodic oxidation potentials (44).
The detailed mechanism of the Paal-Knorr condensa-
tion still remains controversial (24, 31-33), and the
nature of the rate-limiting step(s) can depend on the
γ-diketone:amine stoichiometry as well as the solvent and
reaction conditions. Although monitoring the reactions
by 13C NMR helped us identify the nature of the inter-
mediates and final pyrrole product, the rate data listed
in Table 1 were derived strictly from integration of the
1H NMR spectra. It should be mentioned that an HPLC-
based method was previously utilized for evaluating the
rates of pyrrole formation under pseudo-first-order condi-
tions with the amine in large excess over the diketone
(34). However, when an excess of amine was utilized
under our conditions, we observed the generation of bis-
1
imines (using TFMHD, by H, 13C, and 19F NMR) and
inhibition of pyrrole formation, suggesting that the bis-
imines proceed to pyrrole more slowly than do the mono-
imines. Thus, we felt that a pseudo-first-order analysis
might not represent useful comparative pyrrole formation
rate data in our case. Final product analysis indicated
that condensation of primary amines with 2,5-HD, 3-MHD,
3,4-DMHD, 9, or 10 in the 1:1 stoichiometry utilized by
us resulted in the quantitative formation of pyrroles (11-
13, 15, and 16) (Scheme 4). However, the yield of pyrrole
1
14 from 3-TFMHD was only 80-90% as indicated by H
NMR analysis of the crude CHCl3 reaction mixture, the
remainder corresponding to side products, characteriza-
tion of which is described elsewhere (35).
As can be seen (Table 1), although the relative rate of
pyrrole formation varies somewhat with the nature of the
primary amine, the general rank order is as follows: 10
g 3,4-DMHD > 3 ≈ 9 > 3-MHD > 2,5-HD. The greater
reactivity of 3,4-DMHD over 2,5-HD is well documented
in previous work by Graham and co-workers (13, 34).
Interestingly, the rate of pyrrole formation was always
greater for the four 3-substituted 2,5-hexanediones than
for 2,5-HD. Our interpretation is that the mechanism
proceeds via an imine f enamine tautomerization that
is partly rate-limiting, and is facilitated by the presence
of a 3-substituent on 2,5-HD.
In our original plan, TFHD (8) was expected to
condense rapidly with amines due to heightened carbonyl
electrophilicity. However, unlike other γ-diketones em-
ployed in the study presented here, the reaction of 8 did
not give a pyrrole product with amines. The final 19F
NMR spectrum indicated a mixture of four F-containing
products. Detailed characterization of the structures of
We chose to evaluate relative pyrrole oxidizability by
measuring the rates of oxidation of the neopentylamine-
derived pyrroles 11c-16c by persulfate, a well-known
electron-transfer oxidant in the neutral-basic pH range
(45, 46). When followed spectrophotometrically in a
MeOH/H2O mixture at pH 6.85 under pseudo-first-order
conditions (excess persulfate), the reactions exhibited
monotonic spectral changes which yielded linear first-
order plots; the calculated half-lives are listed in Table
2.
In addition, cyclic voltammetry was performed on
pyrroles 11a -14a , obtained from isobutylamine and the
respective γ-diketone. These potentials (Table 2) were
measured under anaerobic conditions and reveal the clear
facilitation of oxidation by increasing the number of
methyl substituents, as well as the very large retarding
influence of the strongly electron-withdrawing CF3 sub-
stituent. In each case, the voltammogram was typical of
that expected for an irreversible one-electron oxidation;

Novel 2,5-HD Analogues and the Derived Pyrroles
Chem. Res. Toxicol., Vol. 14, No. 3, 2001 271
Ta ble 2. Su m m a r y of Oxid a tion Da ta on P yr r oles
Sch em e 6
although three pyrrolecarboxyaldehydes (m/z 193) could
be detected by GC/MS, only one isomer could be isolated
to analytical purity, and its spectral data did not allow
establishment of its structure among the three possible
isomers that are shown. In the case of the dehydrodimers
detected by GC/MS analysis of the crude reaction product
mixture (m/z 356), attempted purification of these struc-
tural isomers by flash chromatography over silica gel
resulted in extensive decomposition due to further autoxi-
dation. The generation of 12c autoxidation products was
significantly inhibited by 2-mercaptoethanol, consistent
with what has been observed for autoxidation of 2,5-HD
(39, 40). GC/MS analysis of the reaction mixture with
added 2-mercaptoethanol revealed the occurrence of thiol
dimerization (m/z 154) and generation of at least one
thiol-pyrrole conjugation product (m/z 255), concomitant
with reduced levels of pyrrolecarboxaldehyde 17 (m/z 193)
and only a trace of pyrrole dimerization products (m/z
356).
a
Pyrroles [R ) CH2C(CH3)3] (0.20 mM) were oxidized with 20
mM K2S2O8 in a MeOH/H2O mixture at 30 °C and pH 6.85, and
the rates were measured as described in Experimental Procedures.
From cyclic voltammetry for R ) CH2CH(CH3)2. Data measured
in CH3CN containing 0.1 M tetrabutylammonium perchlorate and
2 mM pyrrole. Glassy carbon disk electrode, scan rate of 100 mV/
s.
b
only a single sweep was performed, and the values for
the peak oxidation potentials were directly read from the
charts. The irreversibility presumably arises because of
the highly reactive pyrrole π-cation radical that is gener-
ated (47). From the data in Table 2, it can be seen that
the half-lives for the persulfate oxidations exhibit the
same rank order as the ease of one-electron oxidation (Ep
values) determined by cyclic voltammetry. Since the
pyrroles arising from 9 and 10 were found to be even
more resistant to persulfate oxidation than 3-TFMHD,
we did not measure the anodic oxidation potentials in
these cases.
Au toxid a tion of P yr r oles. 2,5-HD- and especially
3,4-DMHD-derived pyrroles have been shown to undergo
autoxidation at pH 7.4 (38-40). To determine the autoxi-
dation potential of the pyrroles generated by our new
γ-diketones, we evaluated the autoxidation, under physi-
omimetic conditions, of the neopentylamine-derived 3-sub-
stituted pyrroles (12c and 14c-16c) in comparison to
that of pyrrole (11c) derived from neopentylamine and
2,5-HD. Among all five pyrroles, only the 3-methylpyrrole
12c was found to be readily autoxidized, as evidenced
by the rapid color change. Analysis by TLC, GC/MS, and
NMR revealed that at least three types of products were
formed (Scheme 6): (i) oxygenation of one of the three
methyl groups on the pyrrole ring giving three possible
pyrrolecarboxyaldehydes (17), (ii) a ring-opened N-acetyl-
enaminone 18 (shown as the Z-isomer, though it could
be the E-isomer instead), and (iii) mixtures of pyrrole-
pyrrole dehydrodimers (2M - 2). In the case of 17,
In the case of pyrrole 11c, a low level of dehydrodimer
was revealed by GC/MS and H NMR analysis only after
1
incubation for 48 h. Such dimerization has been previ-
ously investigated (38, 40), and we did not pursue
structural identification here. For pyrroles 14c-16c with
electron-withdrawing substituents, not only did we fail
to observe any signs of autoxidation even after 168 h,
but for reactions run in the presence of 2-mercaptoetha-
nol, no thiol-pyrrole conjugation products could be
detected by GC/MS.
These experiments clearly demonstrate that the elec-
tron-withdrawing trifluoromethyl and N,N′-dialkylamino-
carbonyl groups on the pyrrole ring greatly suppress the
rate of pyrrole autoxidation. On the basis of these
combined results, γ-diketones 3, 9, and 10 should theo-
retically serve as good candidates for distinguishing
whether pyrrole formation alone on the NF protein is
sufficient to induce neurotoxicity, or whether subsequent
autoxidation-induced cross-linking of NF is needed. In
fact, since these γ-diketones would form pyrroles faster
than 2,5-HD, they might be expected to be even more
neurotoxic than 2,5-HD unless pyrrole autoxidation is an
obligatory requirement for development of γ-diketone-
induced neuropathies.
Mod ifica tion of RNa se by 3-MHD a n d 3-TF MHD
(3). Previous studies on γ-diketone-induced neurotoxicity
have resorted to the use of the lysine-rich model protein
ribonuclease A (RNase) in an effort to readily assess
protein cross-linking potential (38, 39), in part because
direct analysis of discrete multimerization of NF is

272 Chem. Res. Toxicol., Vol. 14, No. 3, 2001
Xu et al.
F igu r e 1. 3-TFMHD- and 3-MHD-induced RNase A cross-
linking. Solutions of RNase A (10 mg/mL) containing no γ-dike-
tone (lane 1), 3.75-30 mM 3-MHD (lanes 2-5), or 3.75-30 mM
3-TFMHD (lanes 6-9) were incubated for 24 h in the dark at
25 °C. Protein samples were run on a 10% acrylamide separating
gel in the presence of 2-mercaptoethanol, and the gel was
stained with Coomassie blue.
F igu r e 2. Modification of RNase A by 9 and 10. Solutions of
RNase A (10 mg/mL) containing no γ-diketone (lane 1), 3.75-
30 mM 9 (lanes 2-5), or 3.75-30 mM 10 (lanes 6-9) were
incubated for 48 h in the dark at 25 °C. Protein samples were
run on a 10% acrylamide separating gel in the presence of
2-mercaptoethanol, and the gel was stained with Coomassie
blue.
complicated by the heterogeneity of the molecular masses
of the three NF “triplet” proteins (48, 49). In addition,
as discussed in the introductory section, γ-diketone
neuropathy may involve pyrrolation of lysine groups on
proteins other than NF.
Sch em e 7
On the basis of the knowledge that 2,5-HD and 3,4-
DMHD induce cross-linking of proteins as a consequence
of initial pyrrolation and subsequent pyrrole autoxida-
tion, as well as our finding that the 3-MHD-derived
pyrrole 12c readily undergoes autoxidation, we expected
that 3-MHD would also induce protein cross-linking,
thereby serving as a positive control for investigating
protein cross-linking induced by the other 3-substituted
γ-diketones. We selected 3-TFMHD (3) as our initial
example of a γ-diketone that might form pyrroles on
proteins without subsequent cross-linking. Incubation of
the 13.7 kDa lysine-rich protein, RNase (10 mg/mL), with
3-MHD (3.75-30 mM) for 24 h at 25 °C resulted in a
concentration-dependent appearance of RNase dimer,
trimer, tetramer, and even higher oligomers as assessed
by SDS-polyacrylamide gel electrophoresis (SDS-PAGE)
(Figure 1, lanes 2-5). Surprisingly, incubation of RNase
(10 mg/mL) with 3-TFMHD (3.75-30 mM) under the
same conditions also resulted in RNase cross-linking
(Figure 1, lanes 6-9) that was much more extensive than
in the case of 3-MHD at low concentrations, but the level
of cross-linking increased little with increasing concen-
trations such that the degree of cross-linking by 3-TFM-
HD was lower than that for 3-MHD at higher concen-
trations. The mechanism responsible for the concentration-
independent cross-linking induced by 3-TMFHD (3) is the
subject of an investigation described elsewhere (35), but
it is not an autoxidation-dependent process, and is related
to the finding (vide supra) of non-pyrrole side products
in the reaction of amines with 3. These side products
reflect an E1cb elimination reaction that competes with
pyrrole formation, resulting in the generation of adducts
comprised of one molecule of 3 and two molecules of
amine, thereby explaining the observed protein cross-
linking (35). Needless to say, the observed cross-linking
of RNase by 3 excludes it as a candidate for further
exploration of the mechanism of γ-diketone-induced NF
accumulation.
It is worth pointing out that the “dimer” band gener-
ated by incubation of RNase with 3-MHD or 3-TFMHD
exhibits an apparent molecular mass of 30 kDa rather
than the expected 27.4 kDa, based on the molecular mass
markers that were used. If this were due to the increase
in the overall molecular mass resulting from substantial
pyrrolation, the remaining monomer band would also
shift to a higher molecular mass, which was not apparent.
We conclude that cross-linking of RNase by 3-MHD and
3-TFMHD is accompanied by physicochemical changes
in the protein that confer a higher apparent molecular
mass than is present.
Mod ifica tion of RNa se by γ-Dik eton es 9 a n d 10.
In contrast to what was observed for 3-MHD and
3-TFMHD, incubation of RNase (10 mg/mL) with 3.75-
30 mM 9 or 10 at 25 °C for up to 48 h resulted in no
detectable protein cross-linking, as determined by SDS-
PAGE analysis (Figure 2). However, formation of protein-
based pyrrole was still quite evident as revealed by the
Ehrlich assay, utilizing acidified 4-(dimethylamino)-
benzaldehyde (DMAB). The Ehrlich reagent has been
widely used for detection of protein-bound pyrroles such
as those derived from the γ-keto aldehyde LGE2 (50) and
2,5-HD (5, 13, 18). The basis for such an assay is the
highly conjugated chromophore (λmax ) 526 nm) that
results from acid-catalyzed electrophilic substitution of
the pyrrole nucleus with DMAB (51, 52), the cases of 9
and 10 giving presumably 19 and 20, respectively (Scheme
7).
After incubation for 4 and 24 h, aliquots were assayed
for pyrrole content using Ehrlich’s reagent. Figure 3
shows the relationship between the amount of pyrrole
formed, as a percentage of the number of total RNase

Novel 2,5-HD Analogues and the Derived Pyrroles
Chem. Res. Toxicol., Vol. 14, No. 3, 2001 273
the keto form, with little tendency for enolization, as
evidenced by the lack of any apparent deuterium ex-
change in 4 h. In addition, on the basis of our measure-
ment of the 1-octanol/water partition coefficient (P) of 9
(15.7) and 10 (8.09) relative to that of the hallmark
neurotoxic diketone 2,5-HD (13.3), it can be concluded
that the two new γ-diketones would display neuron-
penetrating ability (one a little better, one a little worse)
comparable to that of 2,5-HD. These results, coupled with
the observations that both 9 and 10 produce pyrroles
faster than 2,5-HD in solution reactions with primary
amines, and that the resulting pyrroles are stable to
autoxidation, suggest that these two γ-diketone ana-
logues should be excellent mechanistic probes for future
studies on the mechanism of γ-diketone-induced neuro-
toxicity.
F igu r e 3. Ehrlich detection of pyrrole adducts on RNase
created by treatment with 2,5-hexanedione (2,5-HD) or γ-dike-
tones 9 and 10 as a function of time and concentration. RNase
A (10 mg/mL, 7.3 mM lysine final concentration) was incubated
with 3.75-30 mM 2,5-HD (b), 9 (]), or 10 (2) for 4 h (dashed
lines) or 24 h (solid lines) at 25 °C, followed by incubation with
DMAB for 10 min at 40 °C. Absorbance was measured at 523
nm; ꢀ ) 301 M-1 cm-1 for 2,5-HD and 280 M-1 cm-1 for 9 and
10.
Ack n ow led gm en t. This work was supported by the
National Institutes of Health (in initial stages by Grant
NS 22688, and in later stages by Grant AG 14249).
Su p p or tin g In for m a tion Ava ila ble: 1H and 13C NMR
spectra of 3, 9, 10, 11b,c, 12c, 13b, 14b,c, 15a -c, 16a -c, 17,
and 18. This material is available free of charge via the Internet
at http://pubs.acs.org.
lysines (10), and the concentration of γ-diketone 9 or 10
in the incubations. At all initial concentrations of 9 or
10 that were employed, more DMAB-detectable pyrrole
is formed with longer incubation times (24 h compared
to 4 h). It appears that when sufficiently high concentra-
tions of 9 or 10 are used, all 10 lysines of RNase can be
modified. Nonetheless, no cross-linking was seen inde-
pendent of the fraction of lysines that were modified.
Refer en ces
(1) Spencer, P. S., Schaumberg, H. H., Sabri, M. I., and Veronesi, B.
(1980) The enlarging view of hexacarbon neurotoxicity. CRC Crit.
Rev. Toxicol. 4, 279-356.
(2) Lehning, E. J ., J ortner, B. S., Fox, J . H., Arezzo, J . C., Kitano,
T.-a., and LoPachin, R. M. (2000) γ-Diketone peripheral neur-
opathy. I. Quantitative morphometric analyses of axonal atrophy
and swelling. Toxicol. Appl. Pharmacol. 165, 127-140.
(3) Lanning, C. L., Wilmarth, K. R., and Abou-Donia, M. B. (1994)
In vitro binding of [14C]2,5-hexanedione to rat neuronal cytosk-
eletal proteins. Neurochem. Res. 19, 1165-1173.
(4) Graham, D. G., Anthony, D. C., Boekelheide, K., Maschmann, N.
A., Richards, R. G., Wolfram, J . W., and Shaw, B. R. (1982)
Studies on the molecular pathogenesis of hexane neuropathy. II.
Evidence that pyrrole derivatization of lysyl residues leads to
protein crosslinking. Toxicol. Appl. Pharmacol. 64, 415-422.
(5) DeCaprio, A. P., Olajos, E. J ., and Weber, P. (1982) Covalent
binding of a neurotoxic n-hexane metabolite: conversion of
primary amines to substituted pyrrole adducts by 2,5-hexanedi-
one. Toxicol. Appl. Pharmacol. 65, 440-450.
(6) DeCaprio, A. P., Strominger, N. L., and Weber, P. (1983) Neuro-
toxicity and protein binding of 2,5-hexanedione in the hen.
Toxicol. Appl. Pharmacol. 68, 297-307.
(7) Sayre, L. M., Autilio-Gambetti, L., and Gambetti, P. (1985)
Pathogenesis of experimental giant neurofilamentous axonopa-
thies. A unified hypothesis based on chemical modification of
neurofilaments. Brain Res. Rev. 10, 69-83.
(8) Sayre, L. M., Shearson, C. M., Wongmongkolrit, T., Medori, R.,
and Gambetti, P. (1986) Structural basis of γ-diketone neurotox-
icity: Non-neurotoxicity of 3,3-dimethyl-2,5-hexanedione, a γ-dike-
tone incapable of pyrrole formation. Toxicol. Appl. Pharmacol. 84,
36-44.
(9) DeCaprio, A. P., and O’Neill, E. A. (1985) Alterations in rat axonal
cytoskeletal proteins induced by in vitro and in vivo 2,5-hexanedi-
one exposure. Toxicol. Appl. Pharmacol. 78, 235-247.
(10) Genter, M. B., Szaka´l-Quin, G., Anderson, C. W., Anthony, D.
C., and Graham, D. G. (1987) Evidence that pyrrole formation is
a pathogenetic step in γ-diketone neuropathy. Toxicol. Appl.
Pharmacol. 87, 351-362.
Qu a lifica tion of γ-Dik eton es 9 a n d 10 a s Releva n t
P r obes for F u tu r e Stu d ies on th e Mech a n ism of
γ-Dik eton e Neu r op a th y. The development here of new
2,5-HD analogues containing electron-withdrawing 3-sub-
stituents was predicated on the basis that such γ-dike-
tones would form pyrroles readily, but that the resulting
pyrroles would be resistant to autoxidation-dependent
cross-linking chemistry. Such diketones could then be
used as probes for the contribution of lysine pyrrolation
in NF and/or other proteins to γ-diketone pathogenesis,
independent of the possible pathogenetic role of NF cross-
linking. In this regard, although the pyrrole formed from
3-TFMHD is indeed resistant to autoxidation, this dike-
tone nonetheless was found to cross-link proteins (by a
nonoxidative mechanism), and is therefore unsuitable for
further study. On the other hand, γ-diketones 9 and 10
rapidly form autoxidation-resistant pyrroles and do not
exhibit any evidence of protein cross-linking.
Another such candidate γ-diketone studied previously,
3-AcHD, was shown to form autoxidation-resistant pyr-
roles on blood-borne proteins (18), but was found to lack
neurotoxicity upon administration to rats (19), suggesting
that cross-linking is required for the manifestation of
γ-diketone neuropathy. However, 3-AcHD is a highly
water soluble â-diketone analogue of 2,4-pentanedione,
which is known to exist mainly as its enol tautomer (20)
and to have a pKa of 9 (21), and therefore would probably
partition poorly into neurons. This complication would
place in doubt a definitive mechanistic conclusion based
on 3-AcHD alone.
(11) DeCaprio, A. P., Kinney, E. A., and Fowke, J . H. (1997) Regiose-
lective binding of 2,5-hexanedione to high-molecular-weight rat
neurofilament proteins in vitro. Toxicol. Appl. Pharmacol. 145,
211-217.
(12) Anthony, D. C., Boekelheide, K., and Graham, D. G. (1983) The
effect of 3,4-dimethyl substitution on the neurotoxicity of 2,5-
hexanedione. I. Accelerated clinical neuropathy is accompanied
by more proximal axonal swellings. Toxicol. Appl. Pharmacol. 71,
362-371.
Unlike 3-AcHD, analogues 9 and 10 are â-ketoamides,
which have substantially reduced CR acidity compared
to â-diketones (21). In fact, whereas 2,4-pentanedione in
methanol (0.1 M) exists 72% in its enol form (20), we
found that 9 and 10 (0.1 M) in CD3OD exist entirely in
(13) Anthony, D. C., Boekelheide, K., Anderson, C. W., and Graham,
D. G. (1983) The effect of 3,4-dimethyl substitution on the

274 Chem. Res. Toxicol., Vol. 14, No. 3, 2001
Xu et al.
neurotoxicity of 2,5-hexanedione. II. Dimethyl substitution ac-
celerates pyrrole formation and protein cross-linking. Toxicol.
Appl. Pharmacol. 71, 372-382.
of pyrroles. 4-Oxoaldehydes. Chem. Res. Toxicol. 8, 234-238.
(33) DeCaprio, A. P., J ackowski, S. J ., and Regan, K. A. (1987)
Mechanism of formation and quantitation of imines, pyrroles, and
stable nonpyrrole adducts in 2,5-hexanedione-treated protein.
Toxicol. Appl. Pharmacol. 32, 542-548.
(34) Szaka´l-Quin, G., Graham, D. G., Millington, D. S., Maltby, D. A.,
and McPhail, A. T. (1986) Stereoisomer effects on the Paal-Knorr
synthesis of pyrroles. J . Org. Chem. 51, 621-624.
(35) Xu, G. (1999) Modification of proteins by neurotoxic and lipoxi-
dation-derived carbonyl compounds. Ph.D. Thesis, Case Western
Reserve University, Cleveland, OH.
(36) Simmons, H. E., and Wiley, D. W. (1960) Fluoroketones. I. J . Am.
Chem. Soc. 82, 2288-2296.
(37) Linderman, R. J ., J amois, E. A., and Tennyson, S. D. (1994)
Synthesis of and analysis of thiol additions to â-alkyl-R,â-
unsaturated trifluoromethyl ketones. J . Org. Chem. 59, 957-962.
(38) Zhu, M., Spink, D. C., Yan, B., Bank, S., and DeCaprio, A. P.
(1994) Formation and structure of cross-linking and monomeric
pyrrole autoxidation products in 2,5-hexanedione-treated amino
acids, peptides, and protein. Chem. Res. Toxicol. 7, 551-558.
(39) Zhu, M., Spink, D. C., Yan, B., Bank, S., and DeCaprio, A. P.
(1995) Inhibition of 2,5-hexanedione-induced protein cross-linking
by biological thiols: chemical mechanisms and toxicological
implications. Chem. Res. Toxicol. 8, 764-771.
(40) Amarnath, V., Valentine, W. M., Amarnath, K., Eng, M. A., and
Graham, D. G. (1994) The mechanism of nucleophilic substitution
of alkylpyrroles in the presence of oxygen. Chem. Res. Toxicol. 7,
56-61.
(41) Smith, E. B., and J ensen, H. B. (1967) Autoxidation of three
1-alkylpyrroles. J . Org. Chem. 32, 3330-3334.
(14) Anthony, D. C., Giangaspero, F., and Graham, D. G. (1983) The
spatio-temporal pattern of the axonopathy associated with the
neurotoxicity of 3,4-dimethyl-2,5-hexanedione. J . Neuropathol.
Exp. Neurol. 42, 548-560.
(15) Rosenberg, C. K., Genter, M. B., Szaka´l-Quin, G., Anthony, D.
C., and Graham, D. G. (1987) dl- versus meso-3,4-Dimethyl-2,5-
hexanedione: a morphometric study of the proximo-distal distri-
bution of axonal swellings in the anterior root of the rat. Toxicol.
Appl. Pharmacol. 87, 363-373.
(16) Monaco, S., Wongmongkolrit, T., Shearson, C. M., Patton, A.,
Schaetzle, B., Autilio-Gambetti, L., Gambetti, P., and Sayre, L.
M. (1990) Giant axonopathy characterized by intermediate loca-
tion of axonal enlargements and acceleration of neurofilament
transport. Brain Res. 519, 73-81.
(17) Graham, D. G., Szaka´l-Quin, G., Priest, J . W., and Anthony, D.
C. (1984) In vitro evidence that covalent crosslinking of neurofila-
ments occurs in γ-diketone neuropathy. Proc. Natl. Acad. Sci.
U.S.A. 81, 4979-4982.
(18) Genter St. Clair, M. B., Amarnath, V., Moody, M. A., Anthony,
D. C., Anderson, C. W., and Graham, D. G. (1988) Pyrrole
oxidation and protein cross-linking as necessary steps in the
development of γ-diketone neuropathy. Chem. Res. Toxicol. 1,
179-185.
(19) Pyle, S. J ., Amarnath, V., Graham, D. G., and Anthony, D. C.
(1992) The role of pyrrole formation in the alteration of neurofila-
ment transport induced during exposure to 2,5-hexanedione. J .
Neuropathol. Exp. Neurol. 51, 451-458.
(20) Moriyasu, M., Kato, A., and Hashimoto, Y. (1986) Kinetic studies
of fast equilibrium by means of high-performance liquid chroma-
tography. Part 11. Keto-enol tautomerism of some â-dicarbonyl
compounds. J . Chem. Soc., Perkin Trans. 2, 515-520.
(21) Pearson, R. G., and Dillon, R. L. (1953) Rates of ionization of
pseudo acids. IV. Relation between rates and equilibria. J . Am.
Chem. Soc. 75, 2439-2443.
(42) Gardini, G. P. (1973) Oxidation of monocyclic pyrroles. Adv.
Heterocycl. Chem. 15, 67-98.
(43) Beaver, B. D., Cooney, J . V., and Watkins, J . M., J r. (1985)
Autoxidation of nitrogen heterocycles. 2. Kinetic measurements
of the autoxidation of 2,5-dimethylpyrrole. Heterocycles 23, 2847-
2851.
(22) Ogoshi, H., Mizushima, H., Toi, H., and Aoyama, Y. (1986) 1,1,1-
Trifluoro-2-penten-4-one as a building block of trifluoromethyl-
substituted compounds. J . Org. Chem. 51, 2366-2368.
(23) Smith, J . O., Mandal, B. K., Filler, R., and Beery, J . W. (1997)
Reaction of ethyl 4,4,4-trifluoroacetoacetate enolate with 3-bromo-
1,1,1-trifluoroacetone: synthesis of 2,4-bis(trifluoromethyl)furan.
J . Fluorine Chem. 81, 123-128.
(24) Katritzky, A. R., Yousaf, T. I., Chen, B. C., and Zeng, G.-Z. (1986)
An H-1, C-13, N-15 NMR study of the Paal-Knorr condensation
of acetonylacetone with primary amines. Tetrahedron 42, 623-
628.
(25) Steliou, K., and Poupart, M.-A. (1985) Synthesis of aldehydes and
ketones from nitro paraffins. J . Org. Chem. 50, 4971-4973.
(26) Muller, N. (1983) Electrochemical synthesis of 4,4,4-trifluoro-2-
butanone. J . Org. Chem. 48, 1370.
(27) Miyashita, M., Yanami, T., and Yoshikoshi, A. (1976) A new 1,4-
diketone synthesis using nitroolefins and trimethylsilyl enol
enters. A convenient regiospecific route to cyclopentenones. J . Am.
Chem. Soc. 98, 4679-4681.
(28) Hagens, A. P., Miller, T. S., Bynum, R. L., and Kapila, V. P. (1982)
Reaction of trifluoroacetic acid with alcohols, phenols, ethers, and
their sulfur analogs. J . Org. Chem. 47, 1345-1347.
(29) Cho, L. Y., and Romero, J . R. (1995) Chemical and electrochemical
oxidative dimerization of carbonyl compounds by cerium(IV) salts.
A comparative study. Tetrahedron Lett. 36, 8757-8760.
(30) Dessau, R. M., and Heiba, E. I. (1974) Oxidation of metal salts.
XII. A novel one-step synthesis of 1,4-diketones. J . Org. Chem.
39, 3457-3549.
(44) Tabba, H. D., and Smith, K. M. (1984) Anodic oxidation potentials
of substituted pyrroles: derivation and analysis of substituent
partial potentials. J . Org. Chem. 49, 1870-1875.
(45) Melik-Ogandzhanyan, L. G., Belleryan, N. M., and Chaltykyan,
O. A. (1974) Decomposition of benzoyl peroxide and potassium
persulfate in the presence of alkylamines and piperidine in a
dioxane-water mixture. Kinet. Katal. 15, 569-572.
(46) Pandav, B. V., Ram, N., and Sidhu, K. S. (1979) Kinetics and
mechanism of oxidation of phenoxazine by persulfate. Indian J .
Chem., Sect. A 17A, 40-44.
(47) Diaz, A. F., Martinez, A., Kanazawa, K. K., and Salmon, M. (1981)
Electrochemistry of some substituted pyrroles. J . Electroanal.
Chem. Interfacial Electrochem. 130, 181-187.
(48) DeCaprio, A. P., and O’Neill, E. A. (1985) Alterations in rat axonal
cytoskeletal proteins induced by in vitro and in vivo 2,5-hexanedi-
one exposure. Toxicol. Appl. Pharmacol. 78, 235-247.
(49) Lapadula, D. M., Irwin, R. D., Suwita, E., and Abou-Donia, M.
B. (1986) Cross-linking of neurofilament proteins of rat spinal
cord in vivo after administration of 2,5-hexanedione. J . Neuro-
chem. 46, 1843-1850.
(50) Iyer, R. S., Kobierski, M. E., and Salomon, R. G. (1994) Generation
of pyrroles in the reaction of levuglandin E2 with proteins. J . Org.
Chem. 59, 6038-6043.
(51) Alexander, R. S., and Butler, A. R. (1976) Electrophilic substitu-
tion in pyrroles. Part 1. Reaction with 4-dimethylaminobenzal-
dehyde (Ehrlich’s reagent) in acid solution. J . Chem. Soc., Perkin
Trans. 2, 696-701.
(31) Amarnath, V., Anthony, D. C., Amarnath, K., Valebtine, W. M.,
Wetterau, L. A., and Graham, D. G. (1991) Intermediates in the
Paal-Knorr synthesis of pyrroles. J . Org. Chem. 56, 6924-6931.
(32) Amarnath, V., Amarnath, K., Valentine, W. M., Eng, M. A., and
Graham, D. G. (1995) Intermediates in the Paal-Knorr synthesis
(52) Mattocks, A. R., and White, I. N. (1970) Estimation of metabolites
of pyrrolizidine alkaloids in animal tissues. Anal. Biochem. 38,
529-535.
TX000169Q