Trifluoromethyldiazirinylphenyldiazenes: New hemoprotein active-site probes

Article abstract of DOI:10.1021/ja990351h

The photoactivatable trifluoromethyldiazirinylphenyldiazene probes 1a and 2a have been synthesized, and their utility in the mapping of hemoprotein active sites has been validated with myoglobin (Mb). Reaction of the probes with Mb yields Fe-aryl adducts.

Full text of DOI:10.1021/ja990351h

VOLUME 121, NUMBER 20
MAY 26, 1999
© Copyright 1999 by the
American Chemical Society
Trifluoromethyldiazirinylphenyldiazenes: New Hemoprotein
Active-Site Probes
Richard A. Tschirret-Guth, Katalin F. Medzihradszky, and Paul R. Ortiz de Montellano*
Contribution from the Department of Pharmaceutical Chemistry, UniVersity of California,
San Francisco, California 94143-0446
ReceiVed February 4, 1999
Abstract: The photoactivatable trifluoromethyldiazirinylphenyldiazene probes 1a and 2a have been synthesized,
and their utility in the mapping of hemoprotein active sites has been validated with myoglobin (Mb). Reaction
of the probes with Mb yields Fe-aryl adducts. Photolysis of these adducts unmasks a carbene that cross-links
to active-site protein residues. Migration of the aryl group from the iron to a porphyrin nitrogen then attaches
the porphyrin chromophore to the labeled amino acid residue. Tryptic digestion of the labeled proteins followed
by mass spectrometric analysis of the peptides has identified Leu-29, His-64, Ile-107, and Val-68 as active-
site residues. Previous studies with an arylnitrene probe, which appears to only react with nucleophilic groups,
identified His-64 as an active-site residue [Tschirret-Guth, R. A.; Medzihradszky, K. F.; Ortiz de Montellano,
P. R. J. Am. Chem. Soc. 1998, 120, 7404-7410]. These studies have identified all but one of the active-site
amino acids in contact with the probe. The present strategy not only labels active-site amino acids but also,
because the probe is rigidly held in the active site, provides approximate locations for the labeled residues
with respect to the heme iron atom. Validation of the strategy with myoglobin opens the way to use of the
approach with hemoproteins of unknown active-site structure.
Cytochromes P450 are important enzymes involved in the
biosynthetic and metabolic pathways of endo- and xenobiotics.1,2
To date, more than 500 P450 isoforms have been identified,
and many have been characterized with respect to their substrate
specificity or function. Despite this overwhelming number of
proteins, crystal structure coordinates are only available for five
water-soluble bacterial P450 enzymes,3 and information on the
active-site structures of the mammalian membrane-bound iso-
forms has only been obtained by indirect methods such as
homology modeling,4,5 mechanism-based inactivation,6-9 and
photoaffinity labeling.6,10-13 One technique developed by this
(4) Poulos, T. L. Methods Enzymol. 1991, 206, 11-30.
(5) Modi, S.; Paine, M. J.; Sutcliffe, M. J.; Lian, L.-Y.; Primrose, W.
U.; Wolf, C. R.; Roberts, G. C. K. Biochemistry 1996, 35, 4540-4550.
(6) Yun, C.-H.; Hammons, G. J.; Jones, G.; Martin, M. V.; Hopkins, N.
E.; Alworth, W. L.; Guengerich, F. P. Biochemistry 1992, 31, 10556-10563.
(7) Roberts, E. S.; Hopkins, N. E.; Alworth, W. L.; Hollenberg, P. F.
Chem. Res. Toxicol. 1993, 6, 470-479.
(8) Roberts, E. S.; Hopkins, N. E.; Zaluzec, E. J.; Gage, D. A.; Alworth,
W. L.; Hollenberg, P. F. Arch. Biochem. Biophys. 1995, 323, 295-302.
(9) Koenigs, L. L.; Trager, W. F. Biochemistry 1998, 37, 13184-13193.
(10) Obach, R. S.; Spink, D. C.; Chen, N.; Kaminsky, L. S. Arch.
Biochem. Biophys. 1992, 294, 215-222.
(11) Ohnishi, T.; Miura, S.; Ichikawa, Y. Biochim. Biophys. Acta 1993,
1161, 257-264.
(12) Miller, J. P.; White, R. E. Biochemistry 1994, 33, 807-817.
* To whom correspondence should be addressed at the School of
Pharmacy. Phone: (415) 476-2903. Fax: (415) 502-4728. E-mail
ortiz@cgl.ucsf.edu.
(1) Ortiz de Montellano, P. R. In Cytochrome P450: Structure, Mech-
anism and Biochemistry, 2nd ed.; Ortiz de Montellano, P. R., Ed.; Plenum
Press: New York, 1995; pp 473-574.
(2) Guengerich, F. P. Mammalian Cytochromes P450, Vol. I and II; CRC
Press: Boca Raton, FL, 1987.
(3) Poulos, T. L.; Cupp-Vickery, J. R.; Li, H. In Cytochrome P450:
Structure, Mechanism and Biochemistry, 2nd ed.; Ortiz de Montellano, P.
R., Ed.; Plenum Press: New York; pp 125-150.
10.1021/ja990351h CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/06/1999

4732 J. Am. Chem. Soc., Vol. 121, No. 20, 1999
Tschirret-Guth et al.
Scheme 1. Outline of the Strategy for Labeling of
Scheme 2. Conditions: (a) KNO3, Concd H2SO4, 0 °C; (b)
HCO2H/Et3N/(Pd/C), Acetonitrile, Reflux; (c) NaNO2, 6 N
HCl, 0 °C; (d) SnCl2, 6 N HCl, 0 °C; (e) ClCO2Et, Pyridine,
Acetonitrile; (f) NH2NH2, DMSO, 90-100 °C, 30 min; (g)
ClCO2Et, Pyridine, Acetonitrile.
Active-Site Amino Acid Residuesa
Scheme 3. Conditions: (a) NH2OH.HCl, EtOH/Pyridine,
Reflux; (b) Tosyl Chloride, NEt3, Acetone; (c) NH3(liq),
Ether; (d) Pb(OAc)4, CHCl3.
The heme is represented by N-Fe-N, and the side chains of active-
site residues by R1 and R2. X is the photoactivatable group on the probe.
laboratory relies on the reaction of arylhydrazines and aryldia-
zenes with heme centers to give an iron-aryl (Fe-aryl) complex,
followed by oxidation of the Fe-aryl complex to a mixture of
the four possible N-arylprotoporphyrin-IX regioisomers.14,15 The
ratio of the four regioisomers thus obtained can be used to
construct a low-resolution topological model of the active site.
However, this technique does not provide information on the
nature of the amino acids that form the active-site pocket. To
overcome this shortcoming, we have recently introduced the
use of azidoaryldiazenes as photoaffinity probes for the active
sites of heme-containing proteins.16,17 One advantage of this
technique is that initial formation of the iron-aryl complex locks
the probe in a defined geometry and ensures that subsequent
labeling by the photoactivated azide group occurs in a defined
region within the active site of the protein (Scheme 1). Using
sperm whale myoglobin (Mb) as a test protein, both m- and
p-azidophenyldiazene were found to cross-link to His64, which
suggests that these probes are photoactivated to species that
preferentially or exclusively react with nucleophilic active-site
residues. To make this approach suitable for the study of
hemoprotein active sites without strongly nucleophilic residues
and to facilitate the labeling of multiple residues within a single
active site, we have replaced the azido group by a function that
is photoactivated to a more reactive species. The trifluoro-
methyldiazirine group has been chosen because it is relatively
small and displays both remarkable photochemical properties
and chemical stability.18-19 It is activated by irradiation at 350
nm, a wavelength that is removed from both the protein (280
nm) and heme Soret (430-480 nm) absorbance maxima. In
addition, the photogenerated carbene is long-lived and readily
inserts into unactivated C-H bonds. We report here synthesis
of the photoactivatable aryldiazene probes 1a and 2a and
validation of their use to map the active sites of heme containing
proteins by analysis of their reaction with sperm whale myo-
globin.
Results
Synthesis. Since aryldiazenes are only moderately unstable,20
1a and 2a are best generated immediately before use by base
hydrolysis of the corresponding aryldiazenecarboxylate esters
1b and 2b. The strategies used to synthesize 1b and 2b require
that introduction of the acylhydrazine function precede assembly
of the trifluoromethyldiaziridinyl moiety (Schemes 2 and 3).
Probe 1b was prepared from 3-amino-R,R,R-trifluoroacetophe-
none (4), which in turn was obtained by nitration of trifluoro-
acetophenone (3) and reduction of the nitro group with formic
acid/triethylamine/Pd/C.21 Amine 4 was converted to a hydrazine
by successive treatment with sodium nitrite and SnCl2 in
concentrated HCl, and the product was immediately derivatized
to the acylhydrazine 5. Sequential treatment with hydroxylamine,
tosyl chloride, and ammonia gave the diaziridinyl acylhydrazine
9. Standard treatment of oxime 8 with tosyl chloride in pyridine
at reflux for 1-2 h failed to yield the desired product. However,
tosylation in acetone/triethylamine at room temperature for 10
(13) Cvrk, T.; Hodek, P.; Strobel, H. W. Arch. Biochem. Biophys. 1996,
330, 142-152.
(14) Ortiz de Montellano, P. R. Biochimie 1995, 77, 581-593.
(15) Swanson, B. A.; Ortiz de Montellano, P. R. J. Am. Chem. Soc. 1991,
113, 8146-8153.
(16) Tschirret-Guth, R. A.; Medzihradszky, K. F.; Ortiz de Montellano,
P. R. J. Am. Chem. Soc. 1998, 120, 7404-7410.
(17) Tschirret-Guth, R. A.; Ortiz de Montellano, P. R. J. Org. Chem.
1998, 63, 9711-9715.
(18) Brunner, J.; Senn, H.; Richards, F. M. J. Biol. Chem. 1980, 255,
3313-3318.
(19) Frey, A. B.; Kreibich, G.; Wadhera, A.; Clarke, L.; Waxman, D. J.
Biochemistry 1986, 25, 4797-4803.
(20) Huang, P. C.; Kosower, E. M. J. Am Chem. Soc. 1968, 90, 2354-
2362.
(21) Terpko, M. O.; Heck, R. F. J. Org. Chem. 1980, 45, 4992-4993.

Hemoprotein ActiVe Site Probes
J. Am. Chem. Soc., Vol. 121, No. 20, 1999 4733
min gave the tosyloxime 10 in 77% yield. Acylhydrazine 12
was oxidized to the final product 1b using lead tetraacetate in
CHCl3 (or CDCl3 when monitored by NMR). The reaction,
1
which could be conveniently monitored by H NMR (Figure
1), yielded a single product. The reaction is rapid and provides
an excellent alternative to other diaziridine oxidation procedures.
Probe 2b was prepared from 4-fluoro-R,R,R-trifluoroacetophe-
none (6) and hydrazine in dimethyl sulfoxide. The hydrazine
was immediately derivatized with ethylchloroformate to avoid
further intermolecular reaction and hydrazone formation. Acyl
hydrazine 7 was then converted to the final product (2b) as
described above (Scheme 3). However, the yields were some-
what lower than those observed with 1b. The final products 1b
and 2b had UV absorption bands at 350 and 351 nm,
respectively, positions typical for the trifluoromethyldiazyrinyl
group.18
Labeling of Sperm Whale Myoglobin with 1a. Probe 1a
was added to a sample of sperm whale Mb and the course of
the reaction followed by UV spectroscopy (Figure 2A). The
initial Soret maximum of ferric Mb at 408 nm was replaced by
a much less intense band at 430 nm characteristic of histidine-
ligated Fe-aryl complexes.14 Complex formation occurred more
slowly than previously observed with phenyldiazene or m-
azidophenyldiazene, and a large excess of 1a was required to
take the reaction to completion. The reaction of Mb with 2a
was even slower, and only partial formation of the Fe-aryl
complex was achieved (data not shown). To avoid nonspecific
labeling by excess 1a, the protein complex was purified by
passage through a size exclusion PD-10 column before pho-
tolysis. The UV spectrum of the protein after elution from the
column indicated that no loss of the Fe-aryl complex had
occurred (Figure 2B). Irradiation of the complex with an
unfiltered Hg pen lamp for 30 min occasionally caused a red-
shift (with a concomitant narrowing) of the Fe-aryl absorbance
from 430 to 422 nm (data not shown), suggesting that structural
modification of the protein active site was caused by wave-
lengths below 280 nm. However, irradiation at 360 nm for 30
min with a Rayonnet reactor fitted with an additional Pyrex
(<300 nm) filter resulted in only minor loss of the Fe-aryl
absorbance (Figure 2B). After aerobic exposure to acid to
promote iron-to-nitrogen shift of the aryl group, the protein was
subjected to HPLC analysis with the eluent monitored at 214
and 416 nm (Figure 3). Panel A shows the elution profile of
intact Mb. Under HPLC conditions, the heme dissociates from
the protein core and elutes 2 min earlier than the protein. After
complex formation with 1b and oxidative decomposition in the
absence of photolysis, the peak corresponding to free heme
disappeared and was replaced by products that absorb at 416
nm and elute at 19-21 min. These products are attributed to
the free N-trifluoromethyldiazirinylphenylprotoporphyrin IX
isomers (Figure 3B). However, when the sample was photolyzed
prior to denaturation, the protein was covalently attached to a
chromophore that absorbs at 416 nm (panel C). A nonprotein-
bound peak eluting at 16 min is attributed to free N-arylpro-
toporphyrin-IX isomers derived from photolyzed probe that did
not cross-link to Mb. Residual 416 nm-absorbing peaks at ∼20
min suggest that photolysis was not complete.
Figure 1. 1H NMR of the oxidation of 9 to 1b by lead tetraacetate;
(A) no Pb(OAc)4; (B) ∼1 equiv Pb(OAc)4; (C) ∼2 equiv Pb(OAc)4.
Figure 2. (A) (s) native Mb, (- - -) Fe-aryl complex; (B) (- - -) Fe-
aryl complex after PD-10, (- - -), Fe-aryl complex after photolysis.
Figure 3. HPLC profile of (A) Mb, (B) Mb after complex formation
with 1a and oxidative shift, but without photolysis, (C) Mb after
complex formation with 1a, photolysis and oxidative shift.
mass increment (probe + heme) with respect to the mass
predicted for the native tryptic peptides (Table 1).
Mass Spectrometric Analysis of Modified Tryptic Peptides
of Labeled Mb. The labeled Mb was purified by HPLC and
digested with trypsin. The resulting peptides were separated by
HPLC with the detector set to monitor the effluent at 214 and
416 nm (Figure 4). The peaks absorbing at 416 nm were
collected and analyzed by MALDI-TOF and electrospray mass
spectrometry. Labeled species were identified by the 718 Da
The modified peptides eluting at 48-52 min were identified
as derivatives of the tryptic peptide Val17-Arg31. The PSD
spectrum of this peptide, MH+average ) 2312.5 (Figure 5A)
clearly shows the presence of the heme (m/z 563), the heme-
aromatic linker adduct (m/z 720), and an immonium ion at m/z

4734 J. Am. Chem. Soc., Vol. 121, No. 20, 1999
Tschirret-Guth et al.
which the oxidative iron to nitrogen shift did not occur. The
peptide was subjected to ESIMS analysis by nanospray ioniza-
tion performed on a hybrid quadrupole-TOF mass spectrometer.
The 5+ ion of this species (m/z 561.13) was subjected to low-
energy CID analysis (data not shown). Low-energy CID data
revealed that the peptide was His97-Arg118 (Table 1). The
fragment ions y4, y6, y8, y9, y11, and b10 were observed
unmodified and showed that Ile-107 was the site of the
modification. An abundant ion at m/z 244, corresponding to a
modified Ile immonium ion, confirmed this finding. A minor
component was observed with MH+ m/z 3361.84. This mass
corresponds to a modified (heme containing) 97His-118Arg
peptide (MH+ calculated ) 3361.79, -15 ppm). Unfortunately,
it did not produce an abundant enough ion for further (PSD or
CID) analysis by any of the ionization techniques.
MS analysis of the set of three products eluting at 87-90
min gave a protonated molecular mass of 737.3. Since they also
had an absorbance maximum at 416 nm, the products were
identified as the N-arylprotoporphyrin IX isomers 14.
Figure 4. HPLC profile of the tryptic digest of Mb modified with 1a.
805 corresponding to a modified leucine or isoleucine. The
peptide also produced a series of relatively abundant a-, b-, and
(b+H2O)-type internal fragments, all bearing the porphyrin ring
and all containing a Leu and an adjacent Ile residue. The
presence of a modified (C-terminal) y4 and (N-terminal) a13 in
the PSD spectrum indicated that cross-linking occurred either
at Ile 28 or Leu 29. On the basis of other PSD spectra of
porphyrin-cross-linked peptides (see below and ref 16), we
observed that the first (of lowest mass) fragment ion detected
of the N-terminal a-type ion series is formed by a cleavage at
the modified residue. All other a-ions result from cleavage at
positions beyond the modified residue. Thus, since the a12
fragment was absent from the spectrum, we concluded that the
modification had occurred at Leu-29.
Fractions eluting between 59 and 67 min contained the
modified peptides His64-Lys77 and His64-Lys78 or Lys63-Lys77
with masses at 2112.5 and 2239.2, respectively (Table 1). Figure
5B shows the PSD spectrum of the modified peptide His64-
Lys77. The most abundant fragments in the spectrum were the
protonated porphyrin ring at m/z 563 and a modified valine
immonium ion at m/z 791. Modified N-terminal fragment ions
from a5 to b10 and modified internal fragments at m/z 903 (a-
type VL) and m/z 1005 (a-type TVL or VLT) pinpointed Val-
68 as the labeled amino acid. The other peptide is due to a
missed cleavage by trypsin at one terminus of this peptide. PSD
analysis (data not shown) revealed the additional Lys residue
at the C-terminus, and Val-68 as the modified amino acid.
The modified peptides Ala57-Lys77 and Ala57-Lys78, with
MH+ at 2882.62 and 3010.62, respectively, were also identified.
The PSD spectrum of the shortest peptide (Figure 5C) gave an
abundant modified His immonium ion at m/z 828 (unmodified
would be m/z 110) and a series of modified fragment ions with
charge-retention at the N-terminus (a8-a13), the first member of
which was formed by cleavage between the His-64 and Gly-65
residues. In addition, a modified C-terminal fragment ion, y14,
formed via peptide bond cleavage between Lys-63 and His-64,
was formed, indicating that the peptide was modified at His-
64. The longer peptide resulted form a missed cleavage at Lys-
77 due to the poor exoproteolytic properties of trypsin.
Discussion
The novel hemoprotein active-site probes, 3- and 4-(trifluo-
romethyldiazirinyl)phenyl diazenes 1a and 2a, are readily
synthesized from commercially available starting materials
(Schemes 1 and 2), and their acyl precursors (1b and 2b) are
stable for months when kept in the dark at -20 °C. When tested
against sperm whale Mb, we found that the rates of Fe-aryl
formation of 1a or 2a with Mb were slower than the rates for
their azido counterpart. This difference is probably due to the
constricted nature of the Mb active site and the larger size of
the trifluoromethyldiazirine than azido group, although there
may be an electronic contribution to the rate difference due to
the differences in the electron-withdrawing effects of the two
functional groups. It can be expected that the steric contribution
to this difference in rates will be less important with proteins
such as the cytochrome P450 enzymes, which generally have
larger active sites than that of Mb.
Photolysis of the Fe-(3-diazirinylphenyl) complex of sperm
whale Mb followed by oxidative shift of the aryl group to the
porphyrin nitrogens resulted in covalent attachment of the heme
group to the protein via the cross-linking probe (Figure 3). The
levels of heme labeled Mb that are obtained are lower than those
observed with m-azidophenyldiazene.16 Several factors may
contribute to this difference: (a) incomplete photolysis, (b) less
efficient cross-linking by the photoactivated function with
respect to deactivating reactions, (c) insertion into a larger
number of active-site residues, and/or (d) incomplete iron to
nitrogen shift, yielding a protein that is covalently attached to
the probe but not to the heme chromophore. Our data suggest
that all these factors may come into play: (a) peaks between
19 and 21 min in panel C of Figure 3 indicate that not all of
the probe was photolyzed and that a photolysis time longer than
30 min might be advantageous, (b) the presence of an important
The MALDI-TOF spectra of the fraction containing peptide
Ala57-Lys77 also showed the presence of a peptide with MH+
at m/z 2801.63 (data not shown). This mass corresponded to
tryptic peptide His97-Arg118 + 158 Da. The observed mass
increment indicated the presence of the aromatic structure of
the probe without the porphyrin ring, i.e., a labeling reaction in

Hemoprotein ActiVe Site Probes
J. Am. Chem. Soc., Vol. 121, No. 20, 1999 4735
Table 1. Modified Peptides Isolated from Tryptic Digest of Labeled Mb
+
MH+calculated
deviation (ppm)
sequenceb
a
MHobserved
peptide
2310.96
2111.24
2239.32
2882.62
3010.47
2801.31
Val17-Arg31
His64-Lys77
His64-Lys78
Ala57-Lys77
Ala57-Lys78
His97-Arg118
2311.12
2111.13
2239.23
2882.55
3010.64
2801.55
-70
+52
+38
+24
-56
-78
VEADVAGHGQDILIR
HGVTVLTALGAILK
HGVTVLTALGAILKK
ASEDLKKHGVTVLTALGAILK
ASEDLKKHGVTVLTALGAILKK
HKIPIKYLEFISEAIIHVLHSRc
a Masses obtained by MALDI-TOF mass spectrometry, in delayed extraction, reflectron mode, with two point external calibration. b The modified
residue is indicated in bold. c Modified with the probe only.
Figure 6. Side view of the distal side of the active site of the Mb-
phenyl complex. The phenyl group bound to the iron is shown, and
the residues are labeled.
This result illustrates one of the strengths of the present
approach, since the relative position of the labeled residues with
respect to the iron of the heme group can be approximated by
inspection of the geometry of the probe. Examination of the
crystal structure of the Mb iron-phenyl complex revealed that
Phe-43 is the only residue in proximity of the probe for which
a cross-linked peptide was not isolated. This observation is not
surprising, as the phenylalanine side chain is unfavorably
positioned with respect to the probe, but we cannot exclude the
possibility that the phenylalanine was actually labeled but the
iron to nitrogen shift did not occur and the adduct escaped
undetected (Figure 6).
The peptides were obtained as isomeric mixtures. This is not
unexpected given that cross-linking may occur to multiple points
on the same residue (i.e., Leu-29 can be modified at C-â, C-γ,
or C-δ) and that the iron to nitrogen shift can produce up to
Figure 5. Post source decay (PSD) mass spectrum of labeled peptides
four different N-arylporphyrin regioisomers. The retention times
of these closely related peptides are expected to vary only
slightly, so the peptides elute as an isomeric cluster. We have
been unable to detect differences in the mass spectrometric
fragmentation patterns of these isomeric peaks.
(A) Val17-Arg31, (B) His64-Lys77, (C) Ala57-Lys77; the modified frag-
ments are labeled with an asterisk.
nonprotein peak absorbing at 416 nm that elutes between 16
and 17 min (Figure 3, panel C) indicates the formation of free
N-arylprotoporphyrin IX adducts derived from photolyzed
complex that did not cross-link to the protein but underwent
the iron to nitrogen shift, and (c) the isolation of a peptide
modified at Ile-107 that does not bear the heme suggests that
covalent attachment of the probe to the protein may sometimes
hinder the iron to nitrogen shift (see below also).
Tryptic digestion and mass spectrometric analysis of the
HPLC-purified peptides indicated labeling of at least four
different amino acid side chains within the Mb active site (Table
1, Figure 6). This result clearly confirms the superior labeling
properties of the trifluoromethyldiazirine over the azido group
in the context of an active-site mapping strategy. Three out of
the four residues identified by the former probe are nonnucleo-
philic and are located no more than 2-3 Å away from the
estimated position of the photogenerated carbene (Figure 6).
Ile-107 was found covalently bound to the aryl probe but
not to the heme, which suggests that the iron to nitrogen shift
was inhibited or suppressed by covalent attachment of the probe
to this residue. Covalent attachment of the probe to Ile-107 may
interfere with the motion necessary for the shift to occur or may
otherwise interfere with the shift, resulting in decomposition
of the complex rather than the iron to nitrogen shift. The
detection of this peptide was fortuitous because it coeluted with
a heme-bearing peptide. It would not have been detected
otherwise since it lacks the heme chromophore used to monitor
the position of the labeled peptides. This shortcoming can be
circumvented, if desired, by the use of radiolabeled probes that
would render covalent attachment of the heme to the probe
unnecessary.
The presence of the free N-arylprotoporphyrin IX isomers
14, or closely related structures derived from the precursor

4736 J. Am. Chem. Soc., Vol. 121, No. 20, 1999
Tschirret-Guth et al.
mL) and ethyl chloroformate (1.08 mL) were sequentially added. After
10 min, the solvent was removed under vacuum and the residue was
purified by chromatography with hexanes/AcOEt 60:40 as the eluting
solvent. This process yielded 1.73 g (70%) of pale yellow crystals, mp
107-109 °C: 1H NMR δ (CDCl3) 1.283 (t, J ) 7.2 Hz, 3H), 4.209 (q,
J ) 7.2 Hz, 2H), 6.120 (s, 1H), 6.678 (s, 1H), 7.143 (dd, J ) 1.8, 8.1
Hz, 1H), 7.389 (t, J ) 8.1 Hz, 1H), 7.494 (s, 1H), and 7.565 ppm (d,
J ) 8.1 Hz, 1H); 13C NMR δ (CDCl3) 14.448, 62.273, 113.452, 116.620
(q, JC-F ) 1156 Hz), 120.027, 122.651, 129.831, 130.757, 148.883,
157.005, and 180.458 ppm (q, JC-F ) 137 Hz); IR ν (KBr) 3284, 1722,
1696, 1291, 1249, 1204, 1163, 1140 cm-1; HRMS (EI) calcd for
C11H11N2O3F3 (M+) 276.0722, found 276.0716.
Ethyl 2-(4-Trifluoroacetylphenyl)hydrazinecarboxylate (7). p-
Fluorotrifluoroacetophenone (1.02 g, 5.33 mmol) was dissolved in 20
mL of dimethyl sulfoxide. Hydrazine (0.18 mL, 5.73 mmol) was added,
and the mixture was heated to 90 °C for 1 h before the mixture was
cooled, poured into 200 mL of water, and extracted with AcOEt. The
organic layer was washed with brine, dried over anhydrous MgSO4,
and evaporated to dryness. The residue was dissolved in MeCN (20
mL) before pyridine (0.5 mL) and ethylchloroformate (0.55 mL) were
sequentially added. After 10 min, the solvent was removed under
vacuum and the residue purified by chromatography (hexanes/AcOEt,
70:30), yielding a white solid (0.46 g, 31%), mp 138-140 °C: 1H
NMR δ (CDCl3) 1.293 (t, J ) 6.9 Hz, 3H), 4.229 (q, J ) 6.9 Hz, 2H),
6.390 (s, 1H), 6.623 (s, 1H), 6.861 (d, J ) 8.7 Hz, 2H), and 7.971
ppm (d, 8.7 Hz, 2H); 13C NMR δ (acetone-d6) 14.891, 62.216, 112.258,
118.280 (q, JC-F ) 1156 Hz), 121.070, 133.238, 156.666, 157.693,
and 178.475 ppm (q, JC-F ) 133 Hz), IR ν (KBr) 3337, 3298, 1693,
1601 cm-1; HRMS (EI) calcd for C11H11N2O3F3 (M+) 276.0722, found
276.0724.
carbene, in the tryptic digest of modified Mb was unexpected
since the protein had been purified by HPLC prior to tryptic
digestion in order to remove heme products that were not
covalently bound to the protein. It is possible that a fraction of
the probe formed an unstable adduct with Mb that decomposed,
releasing the modified N-arylporphyrins, during the tryptic
digestion at pH 8.3 but not at the highly acidic pH values
employed for the oxidative shift and the protein purification
(both at pH <1). A candidate for such a labile alkylation site is
Thr-67 (Figure 6). It is located close to the probe, and its
hydroxyl group is optimally situated in proximity of the
photogenerated carbene to allow trapping of the carbene by the
alcohol group. It is not clear, however, whether the resulting
trifluoromethyl-substituted ether adduct would be sufficiently
unstable under basic conditions to be cleaved during the tryptic
digestion.
In summary, the use of probes that bind to the heme iron
and are then photoactivated to unmask a carbene functionality
complements and greatly expands the previously reported use
of related probes in which photoactivation unmasked a nitrene
function.16 The nitrene function primarily reacts with nucleo-
philic residues in the hemoprotein active site, whereas the
carbene function successfully inserts into a range of unactivated
amino acids in the same active site. The nitrene probes thus
provide information on reactive amino acids, but the carbene
probes make possible a more complete mapping of the active-
site residues. Furthermore, as the distance and angle of the
photoactivated moiety with respect to the heme iron atom and
the heme plane is fixed, it is possible to assign approximate
positions to the labeled amino acids within the active site. The
topological information provided by this approach is thus much
firmer than that obtainable by normal photoaffinity labeling or
the use of mechanism-based inactivating agents, for which it is
necessary to question whether the labeling even occurs in the
active site. The demonstration that the probes can be used to
map all but one of the accessible active-site residues in sperm
whale Mb confirms the validity of the method and opens the
way to the use of these probes to map unknown hemoprotein
active sites.
Oximes. The trifluoromethyl ketone 5 or 7 (1-2 mmol) was
dissolved in 20 mL of EtOH containing 0.5 mL of pyridine. Hydroxyl-
amine hydrochloride (1.2 eq) was then added, and the mixture was
refluxed for 2 h before the solvent was removed under vacuum and
the residue was purified by chromatography (hexanes/AcOEt 50:50).
The oxime products 8 and 9 are thus obtained as a mixture of isomers
that can be separated by flash chromatography (hexanes/AcOEt 70:
30).
Ethyl 2-(3-Trifluoroacetyloximephenyl)hydrazinecarboxylate,
“Early Fraction” (8a). White solid, mp 139-141 °C: 1H NMR δ
(acetone-δ6) 1.203 (s, 3H), 4.102 (q, J ) 6.9 Hz, 2H), 6.890 (d, J )
7.8 Hz, 1H), 6.949 (d, J ) 7.8 Hz, 1H), 6.961 (s, 1H), 7.194 (s. 1H),
7.309, (t, J ) 7.8 Hz, 1H), 8.266 (s, 1H), and 11.627 ppm (s, 1H); 13
C
NMR δ (CD3OD) 15.047, 62.652, 113.834, 115.134, 120.901, 122.622
(q, JC-F ) 1083 Hz), 129.901, 130.136, 147.487 (q, JC-F ) 129 Hz),
Experimental Procedures
150.672, and 159.926 ppm; IR ν (KBr) 3350, 3309, 1715, 1127 cm-1
;
1
General Procedures. The H and 13C spectra were recorded on a
HRMS (EI) calcd for C11H12N3O3F3 (M+) 291.0831, found 291.0831.
Ethyl 2-(3-Trifluoroacetyloximephenyl)hydrazinecarboxylate, “Late
Fraction” (8b). White solid, mp 140-142 °C: 1H NMR δ (acetone-
d6) 1.205 (s, 3H), 4.102 (q, J ) 6.9 Hz, 2H), 6.898 (d, J ) 7.8 Hz,
1H), 6.944 (d, J ) 7.8 Hz, 1H), 6.957 (s, 1H), 7.175 (s. 1H), 7.260, (t,
J ) 7.8 Hz, 1H), 8.271 (s, 1H), and 11.977 ppm (s, 1H); 13C NMR δ
(CD3OD) 15.0048, 62.652, 113.390, 120.295 (q, JC-F ) 1119 Hz),
120.693, 130.101, 133.177, 147.259 (q, JC-F ) 115 Hz), 150.680, and
159.923 ppm; IR ν (KBr) 3340, 3278, 1698, 1143 cm-1; HRMS (EI)
calcd for C11H12N3O3F3 (M+) 291.0831, found 291.0842.
Ethyl 2-(4-Trifluoroacetyloximephenyl)hydrazinecarboxylate,
“Early Fraction” (9a). White solid, mp 138-139 °C: 1H NMR δ
(acetone-d6) 1.218 (t, J ) 7.2 Hz, 3H), 4.119 (q, J ) 7.2 Hz, 2H),
6.901 (d, J ) 8.7 Hz, 2H), 7.403 (s, 1H), 7.467 (d, 8.7 Hz, 2H), 8.322
(s, 1H), and 11.527 ppm (s, 1H); 13C NMR δ (acetone-d6) 15.004,
61.750, 112.440, 118.036, 122.571 (q, JC-F ) 1081 Hz), 130.845,
146.581 (q, JC-F ) 126 Hz), 152.057, and 157.869 ppm; IR ν (KBr)
3319, 1667, 1173, 1122 cm-1; HRMS (EI) calcd for C11H12N3O3F3 (M+)
291.0831, found 291.0835.
General Electric GE 300 NMR spectrometer, FTIR spectra on an Impact
400 Nicolet infrared spectrophotometer, and UV-vis spectra on an
HP 8452 diode-array spectrophotometer. Mass spectra were obtained
at the University of California, San Francisco Mass Spectrometry
Facility. Reagents were from Aldrich or Fisher and were used without
further purification. The chromatographic purification of synthetic
organic compounds was carried out on Merck 60 (230-600 mesh) silica
gel. m-Nitrotrifluoroacetophenone was prepared according to the method
of Steward and Vander Linden22 and reduced to m-aminotrifluoro-
acetophenone with triethylammonium formate.21 p-Fluorotrifluoro-
acetophenone was prepared as previously reported.23 Melting points
are uncorrected. Sperm whale Mb was purchased from Sigma and
sequencing grade modified trypsin was obtained from Promega.
Ethyl 2-(3-Trifluoroacetylphenyl)hydrazinecarboxylate (5). m-
Aminotrifluoroacetophenone (1.7 g, 9.0 mmol) was dissolved in 30
mL of 6 N HCl and cooled to -5 °C. NaNO2 (0.68 g, 9.8 mmol) in 1
mL of water was added, and the mixture was stirred for 12 min. SnCl2
(3.91 g, 20.6 mmol) in 3 mL of 12 N HCl was added, and stirring was
continued for another 12 min. The mixture was neutralized with Na2-
CO3 and extracted with AcOEt. The organic layer was washed with
brine, dried over anhydrous MgSO4, and evaporated to dryness. The
oily residue was redissolved in 20 mL of MeCN before pyridine (0.92
Ethyl 2-(4-Trifluoroacetyloximephenyl)hydrazinecarboxylate, “Late
Fraction” (9b). White solid mp 156-158 °C: 1H NMR δ (acetone-
d6) 1.214 (t, J ) 7.2 Hz, 3H), 4.109 (q, J ) 7.2 Hz, 2H), 6.861 (d, J
) 8.7 Hz, 2H), 7.345 (d, 8.7 Hz, 2H), 8.310 (s, 1H), and 11.778 ppm
(s, 1H); 13C NMR δ (acetone-d6) 15.010, 61.699, 112.680, 120.183 (q,
JC-F ) 1120 Hz), 122.259, 130.183, 146.763 (q, JC-F ) 106 Hz),
(22) Steward, R.; Vander Linden, R. Can. J. Chem. 1960, 38, 399.
(23) Herkes, F. E.; Burton, D. J. J. Org. Chem. 1967, 32, 1311.

Hemoprotein ActiVe Site Probes
J. Am. Chem. Soc., Vol. 121, No. 20, 1999 4737
152.130, and 157.826 ppm; IR ν (KBr) 3388, 3329, 1715, 1209, 1127
cm-1; HRMS (EI) calcd for C11H12N3O3F3 (M+) 291.0831, found
291.0833.
Tosyl Oxime Preparation. The oximes prepared above (2-4 mmol)
were treated with 4-toluenesulfonyl chloride (1.1 equiv) in acetone (20
mL) and triethylamine (1 mL). After 10 min the solutions were filtered,
evaporated to dryness, and purified by flash chromatography (hexanes/
AcOEt, 80/20). In the case of 10, the product is obtained as a mixture
of isomers that can be separated by chromatography.
Ethyl 2-(3-Trifluoroacetyltosyloximephenyl)hydrazinecarboxylate,
“Early Fraction” (10a). Yellow oil: 1H NMR δ (CDCl3) 1.252 (t, J
) 7.2 Hz, 3H), 2.465 (q, J ) 7.2 Hz, 2H), 4.197 (s, 3H), 6.690 (s,
1H), 6.809 (s, 1H), 6.857, (d, J ) 7.8 Hz, 1H), 6.935 (dd, J ) 1.8, 7.8
Hz, 1H), 7.282 (t, J ) 7.8 Hz), 7.372 (d, J ) 8.1 Hz, 2H), and 7.868
ppm (d, J ) 8.1 Hz, 2H); 13C NMR δ (CDCl3) 14.382, 21.712, 62.156,
112.286, 116.018, 119.500 (q, JC-F ) 1102 Hz), 120.443, 125.269,
129.200, 129.656, 129.830, 131.122, 146.089, 148.419, 154.071 (q,
JC-F ) 134 Hz), and 157.027 ppm; IR ν (KBr) 3401, 1720, 1386, 1196
cm-1; HRMS (EI) calcd for C18H18N3O5SF3 (M+) 445.0919, found
445.0918.
Ethyl 2-(3-Trifluoroacetyltosyloximephenyl)hydrazinecarboxylate,
“Late Fraction” (10b). Yellow oil: 1H NMR δ (CDCl3) 1.269 (t, J )
7.2 Hz, 3H), 2.451 (q, J ) 7.2 Hz, 2H), 4.197 (s, 3H), 6.611 (s, 1H),
6.868 (s, 1H), 6.922, (d, J ) 7.8 Hz, 1H), 6.950 (dd, J ) 1.8, 7.8 Hz,
1H), 7.248 (t, J ) 7.8 Hz), 7.361 (d, J ) 8.1 Hz, 2H), and 7.860 ppm
(d, J ) 8.1 Hz, 2H); 13C NMR δ (CDCl3) 14.433, 21.718, 62.177,
112.923, 116.193, 117.245 (q, JC-F ) 1126 Hz), 121.230, 128.505,
129.074, 129.540, 129.874, 131.412, 145.973, 148.387, 153.484 (q,
JC-F ) 128 Hz), and 156.991 ppm; IR ν (KBr) 3367, 1720, 1386, 1195
cm-1; HRMS (EI) calcd for C18H18N3O5SF3 (M+) 445.0919, found
445.0916.
Diazirinyl Diazene. The diazirinyl hydrazine (12 or 13) was
dissolved in CHCl3 or CDCl3. Lead tetraacetate was added stepwise
and the reaction monitored by TLC (CH2Cl2) or 1H NMR until
completion. The final product was purified by chromatography (CH2-
Cl2) and obtained in ∼90% yield.
Ethyl 2-(3-Trifluoromethyldiazirinylphenyl)diazenecarboxylate
(1). Orange liquid: 1H NMR δ (CDCl3) 1.474 (t, J ) 7.2 Hz, 3H),
4.532 (q, J ) 7.2 Hz, 2H), 7.434 (d, J ) 7.8 Hz, 1H), 7.588 (t, J ) 7.8
Hz, 1H), 7.758 (s, 1H), and 7.965 ppm (d, J ) 7.8 Hz, 1H); 13C NMR
δ (CDCl3) 14.026, 28.204 (q, JC-F ) 156 Hz), 64.642, 121.809, 121.853
(q, JC-F ) 1090 Hz), 124.468, 129.976, 130.724, 131.014, 151.515,
and 161.818 ppm; IR ν (KBr) 1762, 1252, 1234, 1146 cm-1; UV λmax
(MeOH) 350, 282 nm; HRMS (EI) calcd for C8H4N4F3 (M+ - CO2-
CH2CH3) 213.0388, found 213.0389.
Ethyl 2-(3-Trifluoromethyldiazirinylphenyl)diazenecarboxylate
(2). Orange liquid: 1H NMR δ (CDCl3) 1.474 (t, J ) 7.2 Hz, 3H),
4.529 (q, J ) 7.2 Hz, 2H), 7.345 (d, J ) 8.4 Hz, 2H), and 7.948 ppm
(d, J ) 8.4 Hz, 2H); 13C NMR δ (CDCl3) 14.131, 28.453 (q, JC-F
)
164 Hz), 64.714, 121.794, (q, JC-F ) 1096 Hz), 123.856, 127.352,
134.403, 151.665, and 161.860 ppm; IR ν (KBr) 1761, 1246, 1155
cm-1; UV λmax (MeOH) 351, 288 nm; HRMS (EI) calcd for C8H4N4F3
(M+ - CO2CH2CH3) 213.0388, found 213.0398.
Sperm Whale Myoglobin Labeling. The aryl-iron complexes were
formed according to a previously published procedure.16 Prior to
photolysis, the iron-aryl complex was run through a PD-10 (Pharmacia)
column. Irradiation of the complex was performed at 25 °C for 30 min
using a Rayonet photoreactor at 350 nm. The sample was held 2 cm
from the source and an additional Pyrex filter was placed between the
lamp and the samples to avoid protein modification. The iron to nitrogen
shift of the probe was induced by pouring the mixture into acetonitrile/
trifluoroacetic acid (96:4) and keeping the mixture at 4 °C for 16 h.
HPLC analysis and purification of labeled Mb was performed according
to a reported procedure.16 The labeled Mb was lyophilized and
resuspended in 200 µL of 50 mM ammonium carbonate. Sequencing
grade trypsin (Promega) was added at a ratio 2/100 (w/w) and the
digestion carried out at 37 °C for 3 h. HPLC purification of the tryptic
peptides was performed on a Vydac C-18 RP column eluted at a flow
rate of 0.8 mL/min with a gradient of solvent B (acetonitrile/
trifluoroacetic acid 0.085%) into solvent A (water/trifluoroacetic acid
0.1%): 0-1 min 5% B, 1-40 min 5-35% B, 40-80 min 35-45%
B, and 80-90 min 45-80% B. Fractions were collected every minute.
Mass Spectrometric Characterization of the Isolated Tryptic
Fractions. MALDI (matrix-assisted laser desorption ionization) and
MALDI-post source decay (PSD) experiments were carried out as
previously reported.16 Low-energy CID data were obtained on a QSTAR
mass spectrometer (PE Sciex) equipped with an electrospray source.
This instrument afforded sufficient resolution of the fragment ions
allowing for their charge state determination. The peak nomenclature
was taken from the literature.24
Ethyl 2-(4-Trifluoroacetyltosyloximephenyl)hydrazinecarboxylate
(11). Yellow oil: 1H NMR δ (CDCl3) 1.241 (t, J ) 7.2 Hz, 3H), 2.433
(s, 3H), 4.177 (q J ) 7.2 Hz, 2H), 6.471 (s, 1H), 6.782 (d, J ) 8.4 Hz,
2H), 6.947 (s, 1H) 7.346 (d, J ) 8.1 Hz, 2H), 7.360 (d, J ) 8.4 Hz,
2H), and 7.862 ppm (d, J ) 8.1 Hz, 2H); 13C NMR δ (CDCl3) 14.306,
21.557, 62.111, 111.895, 115.333, 119.800 (q, JC-F ) 1104 Hz),
129.719, 130.423, 131.055, 145.997, 150.940, 153.013 (q, JC-F ) 129
Hz), and 157.070 ppm; IR ν (KBr) 3365, 1724, 1608, 1195 cm-1
;
HRMS (EI) calcd for C18H18N3O5SF3 (M+) 445.0919, found 445.0917.
Diaziridinyl Hydrazine. The tosyloxime (10 or 11) was dissolved
in diethyl ether (3-4 mL) and added to 4-5 mL of NH3(l) at -78 °C.
The mixture was stirred at -78 °C for 4 h and brought to room-
temperature overnight. The mixture was then evaporated to dryness
and purified by flash chromatography (hexanes:AcOEt 70:30).
Ethyl 2-(3-Trifluoromethyldiaziridinylphenyl)hydrazinecarboxylate
(12). White solid (60%), mp 112-114 °C: 1H NMR δ (CDCl3) 1.254
(t, J ) 7.2 Hz, 3H), 2.296 (d, J ) 8.4 Hz, 1H), 2.770 (d, J ) 8.4 Hz,
1H), 4.173 (q, J ) 7.2 Hz, 2H), 6.002 (s, 1H), 6.678 (s, 1H), 6.861 (d,
J ) 7.5 Hz, 1H), 7.111 (s, 1H), and 7.253 ppm (d, 7.5 Hz, 1H); 13C
NMR δ (CDCl3) 14.160, 57.971 (q, JC-F ) 141 Hz), 61.841, 112.145,
114.097, 119.926, 123.334 (q, JC-F ) 1105 Hz), 129.329, 132.292,
148.426, and 157.331 ppm; IR ν (KBr) 3278, 3261, 3209, 1728, 1266,
1155 cm-1; HRMS (EI) calcd for C11H13N4O2F3 (M+) 290.0991, found
290.1005.
Acknowledgment. We thank PE Sciex for the opportunity
of obtaining data on a QSTAR mass spectrometer. This research
was supported by National Institutes of Health Grant GM25515.
Mass spectrometric data were obtained at the Mass Spectrometry
Facility (A. Burlingame, director) of the University of California
at San Francisco supported by National Institutes of Health
Grants RR 01614 and Liver Core Center 5 P30 DK26743.
Ethyl 2-(4-Trifluoromethyldiaziridinylphenyl)hydrazinecarboxylate
(13). Yellow oil (7%): 1H NMR δ (CDCl3) 1.254 (t, J ) 7.2 Hz, 3H),
2.198 (d, J ) 9.0 Hz, 1H), 2.746 (d, J ) 9.0 Hz, 1H), 4.190 (q, J )
7.2 Hz, 2H), 5.986 (s, 1H), 6.656 (s, 1H), 6.810 (d, J ) 8.4 Hz, 2H),
and 7.448 ppm (d, J ) 8.4 Hz, 2H): 13C NMR δ (CDCl3) 14.455,
57.654 (q, JC-F ) 139 Hz), 62.128, 112.680, 123.598 (q, JC-F ) 1108
Hz), 123.607, 129.198, 149.492, and 157.046 ppm; IR ν (KBr) 3322,
1719, 1155 cm-1; HRMS (EI) calcd for C11H12N3O3F3 (M+) 291.0831,
found 291.0831.
JA990351H
(24) Biemann, K. Methods Enzymol. 1990, 193, 886-887.