Synthesis and characterization of 3',4'-anhydroadenosylcobalamin: A coenzyme B12 analogue with unusual properties

Article abstract of DOI:10.1021/ja0013780

The question of how coenzyme B₁₂-Dependent enzymes facilitate the cleavage of the Co-C bond of the cofactor is of interest. We have synthesized an analogue of 5'-Deoxyadenosylcobalamin (AdoCbl¹) designed to stabilize the 5'-Deoxyadenosyl radical (5'-Deoxyadenosine-5'-Yl) that is produced upon homolysis of the Co-C bond. By replacement of the upper axial ligand of AdoCbl by a 3',4'-Anhydro-5'-Deoxyadenosyl moiety, the radical formed on the nucleoside analogue is stabilized by allylic delocalization. The compound, 5'-Deoxy- 3',4'-Anhydroadenosylcobalamin (3',4'-AnAdoCbl) was synthesized by chemical and enzymatic methods. The final step was coupling of cob(I)alamin and 3',4'-AnhydroATP catalyzed by CobA, an ATP: corrinoid adenosyltransferase. 3',4'-AnAdoCbl displays interesting properties. The compound has not been purified to homogeneity due to its thermal and oxygen sensitivity. It was characterized by UV-Vis spectroscopy, ESI-MS, and NMR spectroscopy. The bond dissociation energy of the Co-C bond of the analogue was measured by radical trapping techniques. A significantly weaker bond (24 ± 2 kcal/mol) as compared to AdoCbl (30 kcal/mol) was observed, as was homolytic cleavage at ambient temperature. Photolysis experiments conducted under anaereobic conditions reveal no formation of cob(II)alamin, whereas the compound breaks down rapidly under aerobic conditions as measured by cob(III)alamin formation. We postulate that the weak Co-C bond is cleaved reversibly by photolysis, where recombination of the allylic radical and cob(II)alamin occurs efficiently in the absence of a radical scavanger. Activation of the coenzyme B₁₂-Dependent enzymes dell dehydrase and ethanolamine ammonia-Lyase was observed with the cofactor analogue. The measured activity was low and no formation of cob(II)alamin could be detected in the steady-State of the reaction for either enzyme. Comparative interactions of AdoCbl and 3',4'-An AdoCbl with dell dehydrase and ethanolamine ammonia-Lyase suggest that cleavage of the Co-C bond is facilitated by enzyme-coenzyme binding contacts that are remote from the Co-C bond.

Full text of DOI:10.1021/ja0013780

VOLUME 122, NUMBER 37
SEPTEMBER 20, 2000
©
Copyright 2000 by the
American Chemical Society
Synthesis and Characterization of 3′,4′-Anhydroadenosylcobalamin:
A Coenzyme B Analogue with Unusual Properties
1
2
Olafur Th. Magnusson and Perry A. Frey*
Contribution from the Department of Biochemistry, UniVersity of Wisconsin-Madison,
1
710 UniVersity AVenue, Madison, Wisconsin, 53705
ReceiVed April 19, 2000
Abstract: The question of how coenzyme B12-dependent enzymes facilitate the cleavage of the Co-C bond
1
of the cofactor is of interest. We have synthesized an analogue of 5′-deoxyadenosylcobalamin (AdoCbl )
designed to stabilize the 5′-deoxyadenosyl radical (5′-deoxyadenosine-5′-yl) that is produced upon homolysis
of the Co-C bond. By replacement of the upper axial ligand of AdoCbl by a 3′,4′-anhydro-5′-deoxyadenosyl
moiety, the radical formed on the nucleoside analogue is stabilized by allylic delocalization. The compound,
5
′-deoxy- 3′,4′-anhydroadenosylcobalamin (3′,4′-anAdoCbl) was synthesized by chemical and enzymatic
methods. The final step was coupling of cob(I)alamin and 3′,4′-anhydroATP catalyzed by CobA, an ATP:
corrinoid adenosyltransferase. 3′,4′-anAdoCbl displays interesting properties. The compound has not been
purified to homogeneity due to its thermal and oxygen sensitivity. It was characterized by UV-vis spectroscopy,
ESI-MS, and NMR spectroscopy. The bond dissociation energy of the Co-C bond of the analogue was measured
by radical trapping techniques. A significantly weaker bond (24 ( 2 kcal/mol) as compared to AdoCbl (30
kcal/mol) was observed, as was homolytic cleavage at ambient temperature. Photolysis experiments conducted
under anaereobic conditions reveal no formation of cob(II)alamin, whereas the compound breaks down rapidly
under aerobic conditions as measured by cob(III)alamin formation. We postulate that the weak Co-C bond is
cleaved reversibly by photolysis, where recombination of the allylic radical and cob(II)alamin occurs efficiently
in the absence of a radical scavanger. Activation of the coenzyme B12-dependent enzymes diol dehydrase and
ethanolamine ammonia-lyase was observed with the cofactor analogue. The measured activity was low and no
formation of cob(II)alamin could be detected in the steady-state of the reaction for either enzyme. Comparative
interactions of AdoCbl and 3′,4′-anAdoCbl with diol dehydrase and ethanolamine ammonia-lyase suggest that
cleavage of the Co-C bond is facilitated by enzyme-coenzyme binding contacts that are remote from the
Co-C bond.
Introduction
Most coenzyme B12-dependent enzymes catalyze a 1,2 hydrogen
shift, where the other migratory substituent is either a hetero-
atom, as in the cases of diol dehydrase, ethanolamine ammonia-
lyase, and lysine 5,6-aminomutase, or a carbon atom as for
Enzymes that use coenzyme B12, facilitate the cleavage of
the Co-C bond of the cofactor to generate a 5′-deoxyadenosyl
radical, which acts as an initiator in radical rearrangements.
2
3
methylmalonyl CoA mutase and glutamate mutase. One class
(
1) Abbreviations: AdoCbl, 5′-deoxyadenosylcobalamin (coenzyme B12);
of ribonucleotide reductases use AdoCbl to generate a transient
3
′,4′-anAdoCbl, 5′-deoxy-3′,4′-anhydroadenosylcobalamin; B12s, cob(I)-
4
thiyl radical in their catalytic mechanism.
alamin; B12r, cob(II)alamin; B12a, cob(III)alamin; DMB, dimethylbenzimi-
dazole; DDH, diol dehydrase; EAL, ethanolamine ammonia-lyase; TEMPO,
The Co-C bond of adenosylcobalamin is weak relative to
most covalent bonds and displays a bond dissociation energy
2
,2,6,6,-tetramethylpiperidinyl-1-oxy; HPLC, high performance liquid chro-
matography; yADH, yeast alcohol dehydrogenase; ESI-MS, electrospray
ionization mass spectrometry.
(2) Frey, P. A. Chem. ReV. 1990, 90, 1343-1357.
1
0.1021/ja0013780 CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/01/2000

8808 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Magnusson and Frey
of about 30 kcal/mol.5,6 Enzymes that use coenzyme B12 must
facilitate the cleavage of the Co-C bond by as much as 12
orders of magnitude, which corresponds to a destabilization of
Scheme 1
7
the Co-C bond by ∼15 kcal/mol. How this large rate
enhancement is achieved has remained somewhat of an enigma,
but a number of hypotheses have been put forward. These
include a steric distortion by an upward flexing of the corrin
ring that would weaken the Co-C bond,8 the so-called trans
steric effect, and the trans electronic effect which entails the
importance of the lower axial ligand, dimethylbenzimidazole
-12
13,14
(DMB) in modulating the strength of the Co-C bond.
Both
of these hypotheses are mainly based on model and structural
studies of the coenzyme and different corrinoids. The importance
of the lower axial ligand has been drawn into question by FT-
Raman15 and resonance Raman studies on alkylcobalamins
that show that the strength of the Co-C bond is affected very
little by the lower ligand, and its role appears more complex,
especially since some of the B12-dependent enzymes replace
16
1
7-19
the DMB moiety with a histidine residue of the enzyme.
20,21
Theoretical work,
enzymatic studies with cofactor ana-
logues and spectroscopic studies23 of enzyme-bound AdoCbl
have postulated direct bending or other distortion of the adenosyl
moiety to cleave the bond.
22
The weak Co-C bond of AdoCbl is photolabile, and the
homolytic cleavage of the bond generates the very unstable 5′-
deoxyadenosyl radical that is quenched rapidly in solution.
Hogenkamp and co-workers2
4,25
have shown that anaerobic
Scheme 2
photolysis of AdoCbl yields cob(II)alamin (B12r) and 5′-deoxy-
,5′-cycloadenosine (2) (Scheme 1). In the presence of oxygen,
8
B12r is oxidized to cob(III)alamin (B12a), and the adenosyl radical
is converted to (2) and adenosine-5′-carboxaldehyde (1) (Scheme
1
). Here we report on the synthesis and characterization of an
oxygen-sensitive AdoCbl analogue that displays an unusual
(3) For recent reviews on B12-dependent enzymes, see: Chemistry and
Biochemistry of B12; Banerjee, R., Ed.; Wiley-Interscience: New York, 1999.
(
(
(
(
(
4) Licht, S.; Gerfen, G. J.; Stubbe, J. Science 1996, 271, 477-481.
5) Finke, R. G.; Hay, B. P. Inorg. Chem. 1984, 23, 3041-4043.
6) Hay, B. P.; Finke, R. G. J. Am. Chem. Soc. 1986, 108, 4820-4829.
7) Hay, B. P.; Finke, R. G. J. Am. Chem. Soc. 1987, 109, 8012-8018.
8) Grate, J. H.; Schrauzer, G. N. J. Am. Chem. Soc. 1979, 101, 4601-
4
611.
(
9) Chemaly, S. M.; Pratt, J. M.J. Chem. Soc., Dalton. Trans. 1980,
2
274-2281.
(
10) Marzilli, L. G.; Toscano, J.; Randaccio, L.; Breschiani-Pahor, N.
Calligaris, M. J. Am. Chem. Soc. 1979, 101, 6754-6756.
11) Halpern, J.; Sook-Hui, K.; Leung, T. W. J. Am. Chem. Soc. 1984,
06, 8317-8118.
12) Krautler, B.; Konrat, R.; Stupperich, E.; Gerald, F.; Gruber, K.;
Kratky, C. Inorg. Chem. 1994, 33, 4128-4139.
13) Ng, F. T. T.; Rempel, G. L.; Halpern, J. J. Am. Chem. Soc. 1982,
04, 621-623.
14) Marzilli, L. G.; Summers, M. F.; Bresciani-Pahor, N.; Zangrando,
resistance to photolysis under anaerobic conditions. The com-
pound, 3′,4′-anhydroadenosylcobalamin (Scheme 2) activates
both diol dehydrase (DDH) and ethanolamine ammonia-lyase
(
1
(
(EAL) and owing to allylic stabilization of the 5′-deoxyadenosyl
(
radical should be very useful in studying the first step in the
catalytic cycle of coenzyme B12-dependent enzymes, the cleav-
age of the Co-C bond.
1
(
E.; Charland, J. P.; Randaccio, L. J. Am. Chem. Soc. 1985, 107, 6880-
6
888.
Experimental Section
(15) Puckett, J. M.; Mitchell, M. B.; Hirota, S., Marzilli, L.Inorg. Chem.
1
996, 35, 4656-4662.
4
Materials. AdoCbl, hydroxocobalamin (B12a), KBH , alcohol de-
(
16) Dong, S.; Padmakumar, R.; Banerjee, R.; Spiro, T. G. J. Am. Chem.
Soc. 1996, 118, 9182-9183.
17) Padmakumar, R.; Taoka, S.; Padmakumar, R.; Banerjee, R. J. Am.
Chem. Soc. 1995, 117, 7033-7034.
18) Zelder, O.; Beatrix, B.; Kroll, F.; Buckel, W. FEBS Lett. 1995, 369,
hydrogenase, and NADH were from Sigma. TEMPO and ethylene
glycol were obtained from Aldrich. All other chemicals and reagents
were commercial products of the highest purity.
(
(
Synthesis of 3′,4′-anhydroATP. This compound was synthesized
by enzymatic phosphorylation of 3′,4′-anhydroadenosine as described
in our previous communication.26 MALDI-MS, calculated mass for
2
52-254.
(
19) Chang, C. H.; Frey. P. A. J. Biol. Chem. 2000, 275, 106-114
(20) Christianson, D. W.; Lipscomb, W. N. J. Am. Chem. Soc. 1985,
+
1
1
07, 2682-2686.
C
H
N
O
2
P
) 489.0, found m/e M(H ) ) 490.0. H NMR (200
O) δ 8.17 (s, 1 H, H-2 or H-8), δ 8.13 (s, 1 H, H-2 or H-8),
δ 6.42 (d, 1 H, H-1′); δ 5.54 (d, 1 H, H-3′); δ 5.15 (t, 1 H, H-2′); δ
10
14
5
12
3
(
21) Marques, H. M.; Brown, K. L. Inorg. Chem. 1995, 34, 3733-3740.
MHz in D
(22) Toraya, T.; Matsumoto, T.; Ichikawa, M.; Itoh, T.; Sugawara, T.;
Mizuna, Y. J. Biol. Chem. 1986, 261, 9289-9293.
31
4
.57 (d, 2H, H-5′
2 3 4
). P NMR (81 MHz, referenced to 80% H PO .) δ
(23) Dong, S.; Padmakumar, R.; Banerjee, R.; Spiro, T. G. J. Am. Chem.
-6.5 (γ-P); δ -10.2 (R-P); δ -21.5 (â-P).
Soc. 1999, 121, 7063-7070.
(
24) Hogenkamp, H. P. C. J. Biol. Chem. 1963, 238, 477-480.
(25) Hogenkamp, H. P. C.; Ladd, J. N.; Barker, H. A. J. Biol. Chem.
(26) Magnusson, O. Th.; Reed, G. H.; Frey, P. A. J. Am. Chem. Soc.
1999, 121, 9764-9765.
1
962, 237, 1950-1952.

Synthesis of 3′,4′-Anhydroadenosylcobalamin
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8809
using the molar extinction coefficient of adenosine, 14,700 M 1cm-1
-
.
Expression and Purification of ATP:Corrinoid Adenosyltrans-
ferase. An Escherichia coli overproducing strain JE2875 from a pT7-7
plasmid in BL21(DE3) cells was a generous gift of Prof. J. C. Escalante-
Semerena, University of Wisconsin-Madison. Cells were grown in LB
media, induced by 0.5 mM IPTG and harvested 6 h after induction.
CobA was purified as described elsewhere. The purified enzyme was
precipitated by ammonium sulfate (70% saturation), resuspended in
anaerobic Tris buffer (50 mM, pH 8), frozen and stored at -70 C.
Synthesis of 3′,4′-anAdoCbl. A 50 mL solution containing 0.2 mM
1
The total yield was g90%. H NMR (400 MHz in d
6
-DMSO) for 5′-
TEMPO-3′,4′-anhydroadeosine: δ 8.16 (s, 1 H, H-2 or H-8); δ 8.12
(s, 1 H, H-2 or H-8), δ 7.33 (s, 2 H, NH ), δ 6.26 (d, 1 H, H-1′, J )
2.5 Hz), δ 5.36 (d, 1 H, H-3′, J ) 2.5 Hz), δ 5.15 (t, 1 H, H-2′, J )
2.5 Hz), δ 4.34 (ABq, 2 H, H-5′ , J ) 13 Hz, the inner lines are almost
fully collapsed, and the outer lines are only ∼5% of the intensity of
the middle line), δ 1.25-1.53 (m, 6 H, TEMPO CH ’s), δ 1.13 (s, 6
H, TEMPO CH ’s), δ 1.05 (s, 3 H, TEMPO CH ), δ 1.03 (s, TEMPO
). H NMR for 3′-TEMPO-4′,5′-anhydroadenosine: δ 8.27 (s, 1
H, H-2 or H-8), δ 8.17 (s, 1 H, H-2 or H-8), δ 7.32 (s, 2 H, NH ), δ
6.04 (d, 1 H, H-1′, J ) 4.5 Hz), δ 4.95 (t, 1 H, H-2′, J ) 4.5 Hz), δ
4.73 (d, 1 H, H-3′, J ) 4.5 Hz), δ 4.53 (s, 1 H, H-5′ ), δ 4.42 (s, 1 H,
H-5′ ), δ 1.42 (s, 6 H, TEMPO CH ’s), δ 1.22 (s, 3 H, TEMPO CH ),
δ 1.17 (s, 3 H, TEMPO CH ), δ 0.97-1.27 (m, 6 H, TEMPO CH ’s).
2
27
2
2
3
3
1
3
′,4′-anhydroATP, 0.17 mM OH-cobalamin, 1.5 mM MgCl2, 0.5 mM
CoCl , and 50 mM Tris‚HCl at pH 8 was deoxygenated under a stream
of argon for 1 h. The sealed flask was transferred to a Coy anaerobic
chamber, and 20 mg of KBH was added to reduce cob(III)alamin to
CH
3
2
2
4
a
cob(I)alamin. The reaction was started by the addition of 20 mg of
CobA, and its progress was monitored spectrophotometrically at 530
nm. After a 2 h incubation, the reaction mixture was cooled on ice and
transferred to an Amicon pressure cell with a YM10 membrane. The
protein was separated from the cobalamins and smaller molecules by
this ultrafiltration method, and the filtrate was collected in a sealed
flask under argon. The filtrate was subjected to a second ultrafiltration
step using a YM1 membrane. This removed salts, and the retentate
containing the product was collected, frozen, and stored in liquid
nitrogen.
b
3
3
3
2
Enzyme Assays. Samples of EAL and DDH from Salmonella
typhimurium were gifts from Dr. Vahe Bandarian and Dr. Andreas
Abend, respectively. The enzymes were purified as described else-
29,30
where.
Since the product of both reactions is acetaldehyde, a coupled
assay using yeast alcohol dehydrogenase (yADH) was employed.
Assays with EAL contained 50 mM Hepes, pH 7.5, 10 mM ethano-
lamine, 0.15 mM NADH, 45 IU yADH, 20 µM AdoCbl or 3′,4′-
anAdoCbl and the appropriate amount of EAL. Assays with DDH
i
, pH 8, 125 mM propanediol, 23 mM Na-cholate,
.15 mM NADH, 45 IU yADH, 20 µM AdoCbl or 3′,4′-anAdoCbl
Photolysis. Anaerobic samples of AdoCbl and 3′,4′-anAdoCbl were
prepared in the anaerobic chamber using solutions that were deoxy-
genated with O free Argon. Aerobic samples were prepared using
2
contained 100 mM KP
0
and the appropriate amount of DDH. Assays were performed at 25 °C.
distilled water. Concentration of cobalamins was approximately 50 µM
in each case. Samples in quartz cuvettes, fitted with Teflon stoppers,
were kept on ice and irradiated with a 150 W tungsten lamp at a distance
of 50 cm. At designated time points UV-vis spectra were aquired on
a Hewlett-Packard, model 8452, diode array spectrophotometer.
Anaerobic samples were monitored for cob(II)alamin formation by the
decrease in absorbance at 525 nm, while aerobic samples were
monitored for cob(III)alamin formation by increase in absorbance at
Results and Discussion
(A) Synthesis and Purification of 3′,4′-anAdoCbl. The
ATP:corrinoid adenosyltransferase (cobA) catalyzes the alkyl-
ation of the strong nucleophile, cob(I) alamin (B12s) by the 5′-
carbon af ATP to produce AdoCbl and tripolyphosphate.27 We
used this enzymatic reaction with 3′,4′-anhydroATP and B12s
as substrates to make 3′,4′-anAdoCbl. The reaction was carried
out under strict anaerobic conditions in an anaerobic chamber
since B12s formed by reduction of cob(III)alamin is highly
susceptible to oxygen. With a slight excess of 3′,4′-anATP the
reaction proceeded to completion, as measured by the increase
3
55 nm. Samples with TEMPO contained 20-fold excess (1 mM) of
the radical trap. First-order rate constants (kobs) where determined by
fitting the data to a single-exponential equation (eq 1) using Kaleida-
graph, where A is the absorbance at time t, A is the initial absorbance,
t o
and ∆A is the difference between the final and initial absorbance.
3
1
in absorbance at 530 nm. It became apparent that the AdoCbl
analogue was very labile and sensitive toward oxygen, which
made purification of the compound a daunting process. All
conventional chromatographic procedures were unsuccessful due
to degradation of the product (also see ref 31). It appears that
interaction of the compound with any resin leads to decom-
positon through cleavage of the Co-C bond, since cob(II)alamin
is the only corrinoid species detected. With this restriction in
mind we used ultrafiltration to partially purify the compound.
Use of a YM 10 membrane separates the protein catalyst from
the rest of the reaction mixture. Partial separation of small
molecules from the product was done using a 1000 molecular
weight cutoff membrane, which serves to concentrate the
product while removing salts and smaller molecules. Care has
At ) ∆A[(1 - exp(-kobst))] + Ao
(1)
Thermolysis Kinetics. Samples for kinetic runs were prepared on
ice in the anaerobic chamber and put into Teflon-stoppered quartz
cuvettes. Samples contained 55 µM 3′,4′-anAdoCbl and 2 mM TEMPO
g30 equiv) in anaerobic ethylene glycol. Samples were transferred to
a circulating water bath (Fisher Scientific), thermally equilibrated for
min and incubated at the desired temperature. At designated time
(
2
points, the cuvettes were quickly removed from the water bath and
spectra acquired. Thermal equilibration was tested in control samples,
and data points were excluded if they did not meet the criteria suggested
2
8
by Brown and Evans. Kinetic runs were measured at 4 °C intervals
between 12 and 32 °C ((0.1 °C) by monitoring the decrease in
absorbance at 530 nm. First-order rate constants were determined by
single-exponential fits as described above. All fits gave correlation
2
coefficients of R ) 0.9998 or better.
Nucleoside Products. Approximately 7 µmol of 3′,4′-anAdoCbl and
5 µmol of TEMPO were incubated at 25 °C in 20 mL of Argon flushed
water for 2 h. The solution was concentrated by rotary evaporation
and analyzed by HPLC using a C18 column (Phenomenex, 250 × 10
mm). A linear gradient of 40% to 100% MeOH in 40 min was
employed. Two compounds eluting at 31 and 33 min were characterized
by NMR spectroscopy as 3′-TEMPO-4′,5′-anhydroadenosine (1.1 µmol)
and 5′-TEMPO-3′,4′-anhydroadenosine (5.3 µmol), respectively. The
yield of each compound was estimated by the absorbance at 260 nm
(31) Formation of 3′,4′-anAdoCbl can also be detected by HPLC using
a C18 column with a H2O/MeOH gradient containing 0.02% TFA. The
analogue has a similar retention time as authentic AdoCbl (18 min) and
can easily be separated from other corrinoid species (B12a, etc.). However,
partial degradation of the compound occurs on the column, since a
substantial amount of B12a is also eluted (11 min). Under the acidic elution
conditions, 3′,4′-anAdoCbl is in the base-off form due to protonation of
the lower axial DMB ligand. By using an HPLC system with a diode-array
detector, we were able to show that the peak eluting in ∼18 min is that of
a base-off 3′,4′-anAdoCbl by comparison to a base-off spectrum of authentic
AdoCbl (data not shown). The base-off cobalamin appears to be very labile
as well as the base-on species. Absorption spectra acquired of the eluted
peak revealed partial degradation to a mixture of aquo-cob(III)alamin and
base-off 3′,4′-anAdoCbl (data not shown). This observation implies that
purification of the base-off species could be very difficult, a suggestion
that came up during the review of this paper.
2
(
25.
(
(
(
27) Suh, S. J.; Escalante-Semerena, J. C. J. Bacteriol. 1995, 177, 921-
9
28) Brown, K. L.; Evans, D. R. Inorg. Chem. 1994, 33, 6380-6387.
29) Bandarian, V.; Reed, G. H. Biochemistry 1999, 38, 12394-12402.
30) Abend, A.; Nitsczhe, R.; Bandarian, V.; Stupperich, E.; Retey, J.
(32) Giannotti, C. In B12; Dolphin, D., Ed.; Wiley-Interscience: New
York, 1982; Vol. 1, pp 393-430.
Angew. Chem., Int. Ed. 1998, 37, 625-627.

8810 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Magnusson and Frey
Figure 2. Photolysis of cobalamins under anaerobic conditions. 3′,4′-
Figure 1. Absorption spectrum of 3′,4′-anAdoCbl (solid line) in
comparison with AdoCbl (dotted line). Spectra were aquired at neutral
pH under anaerobic conditions.
anAdoCbl in H
in H
2
O (0) and ethylene glycol (]). Photolysis of AdoCbl
-1
2
O (O) yields a first-order rate constant of 0.5 min .
A mass spectral analysis was obtained to further verify the
chemical structure of the product. ESI-MS spectra revealed a
small peak for 3′,4′-anAdoCbl plus K at m/z 1600.5 (calcd
to be taken to run the reaction to near completion since
separation of any contaminating corrinoids is not feasible by
this method.
+
1600.7), while the largest peaks of the spectrum were due to
The UV-vis spectrum of partially purified 3′,4′-anAdoCbl
is shown in Figure 1 in comparison with AdoCbl. The spectra
look similar, except for an intensity difference of the γ-band at
around 350 nm. This can partially be attributed to impurity in
the preparation from B12a (see part C). The difference can also
be rationalized by the comparative properties of the axial ligands
to the cobalt. It has been shown that there is a correlation
between the difference in the σ-donor strength of the upper and
lower axial ligands and the intensity of the γ-band. The closer
fragmentation at the Co-C bond, giving [(M + K) - 3′,4′-
+
anAdo] at m/z ) 1368.5 (calcd 1368.5) and [(M + K + H) -
2+
3′,4′-anAdo] at m/z ) 684.4 (calcd 684.5). These spectra were
acquired using low-fragmentation voltage (100 V). At higher-
fragmentation voltage the peak for the molecular ion disap-
peared, and only the fragments were observed. Spectra of
AdoCbl acquired with a 100 V fragmentation voltage gave a
strong signal for the molecular ion, while the peak for
fragmentation at the Co-C bond was absent. This is an
interesting observation and bears on results regarding the
strength of the Co-C bond in these compounds as discussed
later in the paper.
32
matching of the σ-donors, the more intense the band becomes.
The lower ligand is in both cases a nitrogen donor (DMB);
however, the allyl group of the 3′,4′-anhydroadenosyl moiety
is a weaker σ-donor than the alkyl moiety of the adenosyl group
and therefore a closer match to the N-donor.
(B) Photolysis Experiments. The homolytic cleavage of the
Co-C bond of AdoCbl can be induced by light. The products
are B12r and 5′-deoxy-8,5′-cycloadenosine, if the reaction is done
1
Characterization of 3′,4′-anAdoCbl by 500 MHz H NMR
24
anaerobically, or B12a and adenosine-5′-carboxaldehyde, when
(D2O) was done by comparing the aromatic region of the spectra
2
5
the reaction is conducted aerobically (Scheme 1). Similar
cyclic nucleosides and nucleoside aldehydes are also obtained
upon photolysis of most purine-containing analogues of adeno-
to AdoCbl, an approach that has been used to access the purity
3
3
of cobalamins. Peaks in the high-field region of the spectra
were difficult to interpret and compare to authentic AdoCbl,
especially since the excess buffer components carried over from
the synthesis are overlapping a part of the spectrum in that
3
5
sylcobalamin.
Photolysis experiments were conducted to examine the
photolability of 3′,4′-anAdoCbl in comparison with AdoCbl.
The experiments were done on ice using a tungsten lamp as a
light source. Aerobic samples were monitored for cob(III)alamin
formation at 355 nm. The analogue and AdoCbl yielded rate
1
region. The abbreviations used for H NMR assignments are:
A, 5′-deoxyadenosyl; B, dimethylbenzimidazole base; R, ribose;
C, corrin. The following chemical shifts were measured in the
aromatic region: δ 8.17 (s, A2), δ 7.88 (s, A8), δ 7.16 (s, B7),
δ 6.94 (s, B2), δ 6.28 (d, R1′), δ 6.21 (s, B4), δ 5.95 (s, C10),
-
1
-1
constants of 2.8 min and 0.5 min , respectively. Faster rate
of photolysis for 3′,4′-anAdoCbl relative to AdoCbl is compat-
ible with the expected weakness of the Co-C bond of the
analogue. In fact, thermal cleavage of the Co-C bond of 3′,4′-
anAdoCbl is readily observed at room temperature in the
absence of light (see part C). Under anaerobic conditions, the
rate constant for cob(II)alamin formation, as measured by a
δ 5.88, (d, A1′). These values can be compared to the aromatic
region of AdoCbl:34 δ 8.19 (s, A2), δ 8.00 (s, A8), δ 7.16 (s,
B7) δ 6.95 (s, B2), δ 6.26 (d, R1′), δ 6.24 (s, B4), δ 5.93 (s,
C10), δ 5.56 (d, A1′). The key differences in the spectra are
observed in the altered upper axial ligand where there is a shift
from δ 8.00 to δ 7.88 for one of the adenine protons, and the
anomeric proton on the ribose ring is shifted downfield by 0.32
ppm, from δ 5.56 to δ 5.88. Contamination by other corrinoid
species was estimated to be e15% by comparison of the
integrals for all the resonances in the aromatic region.
-
1
decrease in absorbance at 525 nm was 0.5 min for AdoCbl;
however, no formation of cob(II)alamin could be detected for
3
′,4′-anAdoCbl (Figure 2). The failure to detect cob(II)alamin
was tentatively attributed to the stability of the 3′,4′-anhydro-
′-deoxyadenosyl radical which apparently reacted with cob-
5
(
(
33) Brasch, N. E.; Finke, R. G. J. Inorg. Biochem. 1999, 73, 215-219.
34) Summers, M. F.; Marzilli, L. G.; Bax, A. J. Am. Chem. Soc. 1986,
(II)alamin to regenerate 3′,4′-anAdoCbl instead of undergoing
cyclization to 3′,4′-anhydro-5′,8-cycloadenosine.
To examine whether 3′,4′-anAdoCbl would be photolyzed,
the experiment was conducted in the presence of 1 mM nitroxide
1
08, 4285-4294.
35) Hogenkamp, H. In B12; Dolphin, D. Ed.; Wiley-Interscience: New
York, 1982; Vol. 1, pp 295-323.
(

Synthesis of 3′,4′-Anhydroadenosylcobalamin
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8811
-d[AdoB ]
spin trap (TEMPO) to probe the reactivity of the allylic radical.
Efficient trapping of the radical was observed concomitant with
the formation of B12r. The observed rate constants for the process
1
2
)
{[Koff-on/(Koff-on + 1)]kh,on
}
dt
k + k [HOCH CH OH] + k {TEMPO]
-
1
-1
c
a
2
2
T
were 0.5 min and 1.5 min for AdoCbl and 3′,4′-anAdoCbl,
respectively. These results show that light induces homolysis
of the Co-C bond, but it does not lead to a secondary reaction
of the incipient allylic radical. Indeed, recombination of B12r
and the organic radical appears to be very efficient because no
net formation of B12r can be detected spectrophotometrically in
the absence of dioxygen or TEMPO. Only in the presence of a
radical scavanger is the allylic radical quenched. In addition,
homolysis of 3′,4′-anAdoCbl in anaerobic ethylene glycol does
not lead to B12r formation, showing that the allylic radical does
not abstract a hydrogen atom from the glycol (Figure 2). Both
{kc
a
2
2
T
}
+ k
[HOCH
CH
OH] + k
{TEMPO] + k [Co(II)B12]
r
[AdoB ] ) k [AdoB ] (2a)
1
2
obs
12
where kc is the rate constant for cyclization of the adenosyl
radical, ka refers to quenching by the solvent, ethylene glycol,
kr is the second-order rate constant for recombination of the
radical with cob(II)alamin and kT the rate constant for scaveng-
ing by the radical trap. Koff-on describes the equilibrium between
the two forms of adenosylcobalamin where the lower axial,
dimethylbenzimidazole ligand is either coordinated to the cobalt
35
6
photolysis and thermolysis of AdoCbl in ethylene glycol lead
to appreciable formation of 5′-deoxyadenosine due to hydrogen
abstraction from the glycol. The allylic 3′,4′-anhydro-5′-
deoxyadenosyl radical is however more stable than the primary
(base-on) or where the cobalt is five-coordinate (base-off). kh,on
is the rate constant for homolysis of the base-on form only.
The base-off form is more stable toward homolysis than the
base-on AdoCbl. At ambient temperatures, the compound is
predominantly in the base-on form, but at the elevated temper-
5′-deoxyadenosyl radical and does not abstract a hydrogen atom
from ethylene glycol. Instead, it recombines with B12r to reform
the Co-C bond. The results can be rationalized on thermody-
namic grounds by considering the estimated difference in bond
dissociation energies of these compounds. The BDE’s are
approximately 97, 91, and 85 kcal/mol for RCH2-H, HOC-
5,38
atures required to induce homolysis of AdoCbl, a correction
for this equilibrium needs to be employed due to the substantial
amount of base-off AdoCbl. This kinetic equation can be
simplified by considering that the allylic radical neither cyclizes
nor is quenched by ethylene glycol as shown previously by the
photolysis experiments, and more importantly by assuming that
36
(
R)2-H, and CH2dCHCH2-H, respectively. This predicts that
while abstraction of a hydrogen atom from ethylene glycol by
the primary adenosyl radical is feasible, the abstraction by the
allylic radical is unfavorable.
3
′,4′-anAdoCbl is exclusively in the base-on form at the
relatively low temperatures required to induce homolysis, that
is, Koff-on . 1. This latter assumption is verified by our
observation that the absorption spectra of 3′,4′-anAdoCbl taken
in ethylene glycol does not change between 12 and 32 °C, and
thus the base-off form which has a different absorption spectra
is not formed in this temperature range (data not shown). The
kinetic equation (eq 2b) for thermolysis of 3′,4′-anAdoCbl can
therefore be expressed in much simpler terms.
The apparent inability of the allylic radical to cyclize can
also be attributed in part to steric constraints. The vinyl group
on the ribose causes flattening of the five-membered ring which
may prevent the 5′-carbon from attacking the 8-postion of the
adenine base. Spin delocalization is, of course, also contributing
to the decrease in reactivity at the 5′-carbon. Taken together,
the results are an example of the persistent radical effect. The
kinetic paths for cyclization and H-atom abstraction have been
removed, and the two paramagnets recombine effectively,
thereby suppressing any side reactions that could occur, such
as dimerization of the allylic radical.
37
-
d[anAdoB ]
kT{TEMPO]
1
2
) kh,on
{
}
dt
k {TEMPO] + k [Co(II)B ]
T
r
12
[
anAdoB ] ) k [anAdoB ] (2b)
12
obs
12
(C) Thermolysis of 3′,4′-anAdoCbl. Since oxygen is an
efficient radical scavenger, sensitivity of 3′,4′-anAdoCbl in air
is a clear indication of the lability of the Co-C bond. As
mentioned above, the compound breaks down within minutes
at room temperature in the presence of oxygen, as measured
by the formation of B12a. To get a quantitative assessment of
the strength of the Co-C bond, we used the radical trapping
Conditions were established where the rate of homolysis was
zero-order in TEMPO, making Co-C bond cleavage the rate-
determining step (data not shown). In subsequent experiments,
excess TEMPO was used (g30 equiv) to determine the first-
order rate constants for homolysis in ethylene glycol at
temperatures ranging from 12 to 32 °C. A typical reaction profile
is shown in Figure 3. The reaction is clean and shows the
conversion to B12r with isosbestic points at 327, 396, 492, and
597 nm, respectively. Notice the small shoulder at the end of
the reaction. This represents the B12a impurity in the preparation
5,6,38
technique developed by Finke and co-workers
to measure
activation parameters and determine the bond dissociation
energy of the compound. TEMPO has been shown to be a good
spin trap and was successfully used by Hay and Finke to
determine the bond dissociation energy of AdoCbl in ethylene
glycol.5 Equation 2a describes the kinetics for thermolysis of
-
1
-1
as discussed earlier. Using the ∆ꢀ520 nm of ∼4 mM cm for
,38
39
AdoCbl and B12r, we can estimate that the preparation contains
3
8
adenosylcobalamin using the TEMPO trapping technique
at least 85% of 3′,4′-anAdoCbl, an independent measurement
that corroborates the NMR results discussed before. Kinetic data
were fitted to a single-exponential equation to determine the
(
36) Carey, F. A.; Sundberg, R. J. AdVanced Organic Chemistry: Part
A, 3rd ed.; Plenum Press: New York, 1990; pp 683-684.
37) Daikh, B. E.; Finke, R. G. J. Am. Chem. Soc. 1992, 114, 2938-
943.
rate constants (kobs) at different temperatures. An Eyring plot
(
q
(
Figure 4) yielded the activation parameters ∆H ) 29.5 ( 0.4
2
1
q
40
(
(
38) Hay, B. P.; Finke, R. G. Polyhedron 1988, 7, 1469-1481.
kcal/mol and ∆S ) 27 ( 2 kcal/mol, respectively.
39) Marsh, E. N. G.; Ballou, D. P. Biochemistry 1998, 37, 11864-
To determine the bond dissociation energy of 3′,4′-anAdoCbl,
we adopted the radical cage formalism developed for the
determination of BDE of organotransition metal compounds in
1872.
40) The reported values and errors were obtained by nonlinear regression
(
∆
Sq/R -∆Hq/RT
fitting to the Eyring equation: k ) (kT/h)e
shown for clarity, and the parameters obtained by that method were identical
within error to the ones obtained by the direct fitting procedure.
e
. The linear plot is
41
solution by Koenig et al. This approach takes into account
the formation of a caged radical pair in solution as described in
(
41) Koenig, T. W.; Hay, B. P.; Finke, R. G. Polyhedron 1988,7, 1499-
4
1
1
516.
general form in eq 3.

8812 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Magnusson and Frey
Table 1. Activation Parameters for Co-C Bond Homolysis in
Ethylene Glycol by the Nitroxide Spin Trapping Method
k
h
(s-1)
∆H
(kcal/mol)
q
h
∆S
(eu)
q
h
Co-C BDE
(kcal/mol)
a
compound
AdoCblb
NpCblc
at 25 °C
-10
-4
4 × 10
3 × 10
34.5 ( 0.8 14 ( 1
32.2 ( 0.6 33 ( 2
29.5 ( 0.4 27 ( 2
30 ( 2
28 ( 2
24 ( 2
-
3
3
′,4′-anAdoCbl 1 × 10
Calculated from ∆H
a
q
q
values. b From ref 38. c From ref
h
and ∆S
h
4
4.
these corrections in mind, the BDE of 3′,4′-anAdoCbl can be
estimated as 24 ( 2 kcal/mol. On the basis of these data, it is
clear that the Co-C bond of the analogue is weaker than in
AdoCbl (BDE ) 30 ( 2 kcal/mol) by approximately 6 kcal/
mol.
In Table 1, a comparison is given for the activation parameters
obtained in this study to those measured by Finke and
co-workers by the same radical trapping method for AdoCbl
Figure 3. Thermolysis of 3′,4′-anAdoCbl at 24 °C. The experiment
was performed in ethylene glycol in the presence of 2 mM TEMPO.
Spectra were aquired at the designated time points: 0 min, 5 min, 14,
3
8,44
and neopentylcobalamin (NpCbl).
NpCbl provides a good
comparison to a sterically strained cobalamin with a weak Co-C
bond which has been studied extensively as a model for Co-C
2
8, and 70 min.
8
,9,28,44
bond cleavage.
In fact, NpCbl displays some similar
behavior as 3′,4′-anAdoCbl, it is oxygen-sensitive and resistant
to homolysis under strict anaerobic conditions in aqueous
solution. However, NpCbl is quenched in ethylene glycol,
whereas 3′,4′-anAdoCbl is quite stable (see part B). By
comparing the data in Table 1, one can see that the rate of
7
homolysis at 25 °C has increased by almost 10 compared to
AdoCbl and is even 3-fold higher than NpCbl. The t1/2 for 3′,4′-
anAdoCbl is only 11 min at 25 °C. This incredible rate
enhancement can be attributed to allylic stabilization of the
radical product. It is worth noting that the free energy of
q
q
activation is ∆G 25°C ) 30.3 kcal/mol and ∆G 25°C ) 21.5 kcal/
mol for AdoCbl and the analogue, respectively. The difference
of 8.8 kcal/mol approaches a value of ∼11 kcal/mol estimated
for stabilization of an allylic radical relative to a primary alkyl
3
6
radical.
The difference in entropy of activation between AdoCbl and
3
′,4′-anAdoCbl is of some interest. Restricted rotation of the
Figure 4. Eyring plot of kobs for the thermolysis of 3′,4′-anAdoCbl.
-
1
The values for kobs (s , (3%) and temperatures (°C), respectively,
allyl group as allylic resonance is set up in the transition state
would be expected to cause a decrease in the entropy of
activation compared to AdoCbl. The opposite is actually
-
5
-4
-4
are as follows: 9.93 × 10 , 12; 2.16 × 10 , 16; 4.55 × 10 , 20;
.14 × 10 , 24; 1.83 × 10 , 28; 3.61 × 10 , 32.
-
4
-3
-3
45
9
q
observed, where ∆S is 13 eu higher for the analogue than
k1(soln)
kd
T•
•
•
•
•
AdoCbl. The reason for this is not known. However, a possible
explanation for this unexpected finding can be provided by
considering that the upper axial ligand may be interacting
unfavorably with the corrin. Relief of ground-state strain in the
transition state could therefore be more important than the
opposing effect, the stiffening of the allyl group.
The products of the thermolytic reaction are B12r as docu-
mented above and a trapped radical species (Scheme 3). To
determine the nature of the nucleoside product, thermolysis of
3′,4′-anAdoCbl at 25 °C was performed in the presence of excess
spin trap. The reaction mixture was separated by HPLC and
two main nucleoside products isolated. Structural analysis by
NMR shows that these products are 5′-TEMPO-3′,4′-anhy-
droadenosine (1) and 3′-TEMPO-4′,5′-anhydroadenosine (2),
formed in a 4:1 ratio, respectively. The total yield was estimated
to be >90% showing that trapping by TEMPO is very efficient.
It is interesting to note that these results emphasize the allylic
nature of the radical formed upon homolysis of the Co-C bond;
M-L y kc z [M DN L ] y z [M ] + [L ]
98 L-T (3)
k1(soln)
kT
cage pair
free radicals
The formalism describes the correction for solution deter-
q
mined activation parameters as follows: BDECo-C = ∆H obs(soln)
.
q
-
Fc ∆H η, where Fc is the fractional-cage efficiency (i.e.,
describes the efficiency of recombination of caged radicals) and
q
∆
H η is the activation enthalphy for viscous flow of the solvent
q
q
used. This treatment neglects ∆H c and sets ∆H d equal to
q
q 41
∆
H -1(soln), which can be considered equal to ∆H η. The
enthalphy for viscous flow of ethylene glycol in the temperature
range employed is 6.5 kcal/mol.42 Transient photolysis studies
on AdoCbl have estimated a fractional-cage efficiency of 0.7-
43
0
.8. Given the structural similarities between AdoCbl and the
analogue in question, a value of 0.8 ( 0.2 is used for Fc. With
q
(
42) This value is obtained from Andrade’s law, ln η ) Aη + (∆H η +
RT)/RT. The viscosity of ethylene glycol at different temperatures was
obtained from: Viswanath, D. S.; Natarajan, G. Data book on the Viscosity
of liquids; Hemisphere Publishing Corporation: New York, 1989.
(
43) (a) Chen, E.; Chance, M. R. J. Biol. Chem. 1990, 265, 12987-
(44) Waddington, M. D.; Finke, R. G. J. Am. Chem. Soc. 1993, 115,
4629-4640.
(45) Benson, S. W. Thermochemical Kinetics, 2nd ed.; John Wiley and
Sons: New York, 1976; p 100.
1
1
2994. (b) Chagovetz, A. M.; Grissom, C. B. J. Am. Chem. Soc. 1993,
15, 12152-12157. (c) Walker, L. A.; Shiang, J. J.; Anderson, N. A.; Pullen,
S. H.; Sension, R. J. J. Am. Chem. Soc. 1998, 120, 7286-7292.

Synthesis of 3′,4′-Anhydroadenosylcobalamin
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8813
Scheme 3
reaction of DDH activated by AdoCbl is B12r.46 Therefore,
despite the weaker Co-C bond in 3′,4′-anAdoCbl, it is not
cleaved in the steady-state in the way that the stronger Co-C
bond of AdoCbl is cleaved. This suggests that the enzymes
cannot properly facilitate the cleavage of the Co-C bond and
separate the incipient radical from the cobalt, as has to occur
for catalysis to take place. One would expect the allylic radical
to be stable enough to be detected, but efficient separation from
cob(II)alamin is required to prevent recombination of the
paramagnets. Also, hydrogen abstraction from the substrate
should be slow, especially considering that ethylene glycol
which is a substrate for EAL does not quench the allylic radical
in solution (see part B). In view of the fact that the Co-C bond
of the analogue is much weaker than for AdoCbl and the rate
enhancement for its cleavage is very low, the data suggest that
the upper axial ligand is not oriented properly in the active site
of these enzymes, stressing the probable importance of a steric
effect from the ligand. The 5′-deoxy-3′,4′-anhydroadenosyl
moiety incorporates trigonal carbons at positions 3′and 4′and
lacks the 3′-OH group of the 5′-deoxyadenosyl moiety. Its ring
conformation and binding specificity therefore differ signifi-
cantly from 5′-deoxyadenosyl.
Conclusions
We have synthesized and characterized a novel analogue of
coenzyme B12, designed to stabilize the 5′-deoxyadenosyl radical
formed upon homolysis of the Co-C bond. This is an adaptation
of our earlier strategy for radical stabilization of the 5′-
deoxyadenosyl radical generated by reductive cleavage of
S-adenosylmethionine in the reaction of lysine 2,3-aminomu-
however, the product ratio probably does not reflect the spin
density at the 3′- and 5′-carbons since trapping at the 3′-position
is the sterically less favored reaction.
(D) Activation of DDH and EAL. The coenzyme analogue
2
6
was tested for activity with two B12-dependent enzymes, diol
dehydrase (DDH) and ethanolamine ammonia-lyase (EAL).
Activity could be detected with both enzymes using a coupled
assay with alcohol dehydrogenase by monitoring the oxidation
of NADH spectrophotometrically. The specific activity of DDH
with 3′,4′-anAdoCbl is 0.020 IU/mg compared to 90 IU/mg with
AdoCbl, or 0.020%. Similarly, the activity of EAL is 0.024 IU/
mg compared to 45 IU/mg with AdoCbl, or 0.053%. On the
basis of these data, kcat with 3′,4′-anAdoCbl as cofactor can be
tase. A substantial weakening of the Co-C bond of the
analogue is observed which almost surely can be rationalized
by the stability of the allylic radical formed by the homolytic
cleavage of the bond. The small rate enhancement of Co-C
bond cleavage in DDH and EAL along with the lack of
formation of any radical species is suggestive of the adenosyl
moiety being the key determinant for the mechanism by which
B12-dependent enzymes labilize the bond. The results also
emphasize the fact that B12-dependent enzymes need not only
to displace the equilibrium of Co-C bond cleavage toward
homolysis, but must also physically separate the fragments to
prevent recombination. The coenzyme analogue should be an
excellent tool for the study of other B12-dependent enzymes,
especially in determining the factors that facilitate the homolysis
of the novel organometallic bond.
-
1
-1
calculated to be 0.07 s and 0.03 s for DDH and EAL,
respectively. The assays were done at 25 °C, and it is interesting
to note that kobs for thermolysis of 3′,4′-anAdoCbl at that
-
1
temperature is ∼0.001 s . This means that the enzymes do
not have to provide more than about 30-70-fold rate enhance-
ment for the cleavage of the Co-C bond. This is little compared
12
to the ∼10 rate enhancement required for AdoCbl.
Acknowledgment. This work was supported by Grant
DK28607 from the National Institute of Diabetes and Digestive
and Kidney Diseases (P.A.F.) and The Sam. C. Smith predoc-
toral fellowship, administered by the Department of Biochem-
istry, University of Wisconsin-Madison (O.Th.M). We thank
Dr. Padma Marwah for help with aquiring ESI-MS data. This
study made use of the National Magnetic Resonance Facility
at Madison, which is supported by NIH grant RR02301 from
the Biomedical Research Technology Program, National Center
for Research Resources. Equipment in the facility was purchased
with funds from the University of Wisconsin, the NSF Biologi-
cal Instrumentation Program (grant DMB-8415048), the NIH
Biomedical Research Technology Program (grant RR02301),
the NIH Shared Instrumentation Program (grant RR02781), and
the U.S. Department of Agriculture.
Since our preparation of 3′,4′-anAdoCbl is not perfectly pure,
a point of concern is the possibility that the low activity is due
to AdoCbl as a trace contaminant. This issue was clarified by
incubating a sample of 3′,4′-anAdoCbl in the dark at room
temperature after exposure to air for 3 h. Authentic AdoCbl
and other known active analogues are stable under these
conditions, whereas 3′,4′-anAdoCbl undergoes thermal degrada-
tion due to irreversible cleavage of the Co-C bond as described
in this paper. No enzymatic activity could be detected after this
treatment, providing compelling evidence that the observed
activity is not due to AdoCbl but to the action of 3′,4′-anAdoCbl
as a cofactor.
The formation of B12r was tested in enzymatic samples
prepared anaerobically. No change in the UV-vis spectrum can
be detected under catalytic conditions for either enzyme, that
is, no build up of B12r is observed in the steady-state of the
reaction. The dominant chromophore in the steady-state of the
JA0013780
(46) Wagner, O. W.; Lee, H. A. Jr.; Frey, P. A.; Abeles, R. H. J. Biol.
Chem. 1966, 241, 1751-1762.