The Coenzymic and Chemical Properties of a Carbocyclic Analogue of Vitamin B,, Coenzyme*

SUMMARY An analogue of vitamin Blz coenzyme in which the ribosyl oxygen has been replaced by -CH2-has been synthesized. This compound has spectral properties very similar to vitamin Blz coenzyme. The chemical properties of the analogue differ from those of vitamin B1e coenzyme in that reaction with CN- and BH4-does not result in cleavage of the carbon-cobalt bond. The analogue functions as coenzyme in the reaction catalyzed by dioldehydrase. The reaction in the presence of analogue is one-third as fast as with the coenzyme. Other properties of the enzymic reaction in the presence of the analogue are similar to those observed with the coenzyme: spectral changes in the presence of substrate and glycolaldehyde, hydrogen exchange between substrate and coenzyme, and over-all deuterium isotope effect. It was concluded that the ribosyl oxygen of the coenzyme cannot play an essential role in the catalytic process. coenzyme

An analogue of vitamin Blz coenzyme in which the ribosyl oxygen has been replaced by -CH2-has been synthesized. This compound has spectral properties very similar to vitamin Blz coenzyme.
The chemical properties of the analogue differ from those of vitamin B1e coenzyme in that reaction with CN-and BH4-does not result in cleavage of the carboncobalt bond.
The analogue functions as coenzyme in the reaction catalyzed by dioldehydrase. The reaction in the presence of analogue is one-third as fast as with the coenzyme.
Other properties of the enzymic reaction in the presence of the analogue are similar to those observed with the coenzyme: spectral changes in the presence of substrate and glycolaldehyde, hydrogen exchange between substrate and coenzyme, and over-all deuterium isotope effect. It was concluded that the ribosyl oxygen of the coenzyme cannot play an essential role in the catalytic process.

Investigations
of the mechanism of action of Blz coenzyme have led to the hypothesis that the carbon-cobalt bond of the coenzyme undergoes chemical modification, most likely dissociation, during the course of the reaction.
Several proposals have been made concerning the mechanism and nature of this modification (l-5).
One possibility, which we have considered, is suggested by nonenzymic reaction of cobalamin and cobaloximes containing oxygen 0 to the carbon-cobalt bond. In these compounds, cleavage of the carbon-cobalt bond can occur under relatively mild acid conditions (6,7). For instance, hydroxyethyl cobalamin is converted to ethylene at 25" at pH 3 with t+ of 2 h0urs.l These reactions have led us to consider the possibility that interaction of the enzyme and coenzyme leads to cleavage of the carbon-cobalt bond through protonation of the ribosyl oxygen of the coenzyme by an acidic group on the enzyme. This results in the formation of a complex, consisting of an electron deficient cobalt, which interacts, possibly through a a com-plex, with the double bond of the deoxyribosyl moiety. This complex could function as an intermediate hydride acceptor in the conversion of propanediol to propionaldehyde. The formation of the activated complex is shown in Fig. 1. A similar proposal has been made for the reaction catalyzed by methylmalonly coenzyme A isomerase (1). The validity of this type of mechanism was tested by preparing a coenzyme analogue in which the oxygen atom of the ribosyl moiety of 5,6-dimethylbenzimidozolylcobamide 5'-deoxyadenosine is replaced by -CH,.
Partial structures of the coenzyme analogue, which will be called carbocyclic DBCC,z are shown in Fig. 2. If this analogue has coenzyme activity, the mechanism cannot be valid.
The method of synthesis, coenzymic, and chemical properties of carbocyclic DBCC are reported here.

MATERIALS AND METHODS
Partial Synthesis of Curbocyclic DBCC-The carbocyclic analogue of adenosine was a gift from Dr. Y. F. Shealy, Southern Research Institute, Birmingham, Alabama (8). 2', 3'-Isopropylidene carbocyclic adenosine was prepared from 10 mg of carbocyclic adenosine by the method of Hampton (9). The crude 2', 3'-isopropylidene derivative was brought to dryness under reduced pressure and dried over PzOs for 12 hours. To the flask containing the dried isopropylidene derivative, 0.4 ml of anhydrous pyridine (dried and distilled over KOH) was added. After all the solid material had dissolved, 10 mg of p-toluenesulfonyl chloride was added and the solution was incubated at room temperature for 6 hours with the exclusion of moisture. The resulting 2', 3'-isopropylidene 5-p-toluenesulfonyl carbocyclic adenosine was then reacted with B1zs by procedures previously used for the synthesis of DBCC (10, 11). The reaction product was isolated by phenol extraction and purified by chromatography on Dowex 50 (10). 2', 3'-Isopropylidene carbocyclic DBCC was extracted into phenol and back extracted into water. The aqueous solution was concentrated to 80 ml in a rotary evaporator and 10 ml of 2 N HCI were added.
The solution was incubated at room temperature for 24 hours and carbocyclic DBCC was isolated by chromatography on Dowex 50 (IO). The coenzyme analogue was then subjected to paper electrophoresis on Whatman No. 3MM paper with 0.5 N NH,OH for 3 hours at 10 volts per cm. Under these conditions, the analogue migrated 2  The solution was incubated at 37'. After 30 min, 2 ml of 2 N HCl were added and the solution passed through a Dowex 50-H+ column (0.5 X 4 cm). The column was washed with 10 ml of water and carbocyclic DBCC eluted with 10 ml of 1 N NH40H.
The spectrum of the recovered product was identical with that prior to cyanide treatment.
Dioldehydrase was prepared and assayed as described (12). Protein determinations were by the method of Lowry et al. (13). Radiochemical assays were performed by liquid scintillation counting in an Ansitron scintillation counter by the method of Bray (14).

Chemical
Properties of Carbocyclic DBCC-The spectra of Fractions A and B of carbocyclic DBCC are identical with that of DBCC ( rnp (not shown in Fig. 3). Exposure to light after addition of cyanide yields the spectrum of dicyanocobalamin.
Addition of cyanide without light exposure causes a slight spectral change which does not undergo further changes after 1 hour. We tentatively attribute this spectral change to the formation of an adduct in which cyanide has displaced the benzimidazole base and is coordinated to the bottom position of the corrin ring. Since carbocyclic DBCC can be reisolated after cyanide treatment (see "Materials and Methods"), no covalent changes take place. Treatment of DBCC under similar conditions leads to complete conversion to dicyanocobalamin.
The relative susceptibility of DBCC and carbocyclic DBCC (Fractions A and B) to NaBH4 reduction was determined.
A solution (0.4 ml) containing 0.039 M carbocyclic DBCC or DBCC, 0.1 M Tris-Cl, pH 8.0, and 0.3 M NaBH4 was incubated in the dark at room temperature for 3 min. The reaction was stopped by the addition of 0.1 ml of acetone and the extent of decomposition was determined from the absorbance at 352 rnp. It was found that more than 90% of DBCC was converted to hydroxy Blz whereas carbocyclic DBCC (Fraction A or B) remained intact.
The rates of acid decomposition of carbocyclic DBCC and DBCC were compared.
Carbocyclic DBCC (Fraction B), 5.1 X low6 M, and DBCC, 7.0 x lOA M, were heated in 0.1 N HzS04 at 100" in the dark.
At 0, 10,20,30, and 50 min, 0.2-ml aliquots were withdrawn and diluted to 25 ml with 8 X 10e4 RI NaOH. Aliquots of that solution were assayed enzymically for coenzyme activity (16). The rates of disappearance of coenzyme activity for both coenzymes followed first order kinetics.
Coenzymic activity was lost in both at nearly identical rates with a t+ of 10.5 to 12.0 min. It is known that DBCC under acid conditions is converted to Blz(,), adenine, 2,3-dihydroxy-4-penten-al (17). To test whether Blz(,) was formed from carbocyclic DBCC, the compound was heated at 100" in 0.1 N HzS04 and spectra were taken at 0, 15, and 50 min. A similar experiment was carried  out with DBCC. Spectra were taken at 0 and 50 min. In both cases, it was shown by enzymic analysis that at the end of the reaction, no coenzymic activity remained.
As expected, after 50 min, the spectrum of DBCC was that of Blzca) and no further spectral change occurred upon exposure to light.
The spectrum of carbocyclic DBCC (Fig. 4) was also changed, but the compound was not completely converted to B 12~~) since after light exposure, a further spectral change occurred.
This spectrum now resembles that of B12ca). The spectrum in Fig. 4 also shows that the spectral change was complete after 15 min. At that point not all coenzymic activity is lost. It was concluded that carbocyclic DBCC undergoes a chemical modification in acid, which causes loss of coenzyme activity.
Unlike DBCC, the product of the acid reaction may not be B12ca) and certainly is not exclusively B Iz(a). The significance of the spectral change observed upon heating carbocyclic DBCC in acid is not clear. Insufficient material was available to further explore this reaction.
In the reactions with CN-, NaBHa, and possibly with acid, the carbon-cobalt bond of DBCC is more susceptible to cleavage than that of carbocyclic DBCC.
This difference in reactivity provides evidence for the contribution of the ribosyl oxygen in reactions in which the carbon-cobalt bond dissociates so that the electrons of that bond remain with the adenosyl moiety.
&enzyme Activity of Carbocyclic DBCC-The coenzyme activity of the two fractions of carbocyclic DBCC was tested with dioldehydrase and compared to that of DBCC. Saturating amounts of both coenzymes were used. The results are summarized in Table I. Carbocyclic DBCC can function as a coenzyme and therefore the ribosyl oxygen of the adenosyl moiety is not essential for coenzymic activity.
Exposure of carbocyclic DBCC to cyanide and subsequent purification did not alter its coenzyme activity when determined under saturating and nonsaturating conditions. This eliminates the unlikely possibility that the coenzyme activity observed with carbocyclic DBCC was caused by contamination by DBCC. Carbocyclic DBCC (Fraction B) shows appreciably more coenzyme activity than Fraction A. It is possible that the small amount of activity of Fraction A is caused by contamination by Fraction B, and that Fraction B may also contain some Fraction A so that the activity reported here is a minimal activity.
Insufficient carbocyclic DBCC is available at this time to further define the conditions for separation of Fractions A and B.
Fraction A inhibits the conversion of propanediol to propionaldehyde as illustrated by the results in Table II. The results show that Fraction A is as inhibitory as hydroxy B12, an irreversible inhibitor of dioldehydrase. Fig. 5 shows double reciprocal plots of the initial velocities against DBCC or carbocyclic DBCC concentrations. The apparent K, obtained for the two coenzymes is quite similar (6.1 X lo-' M for DBCC and 8.7 X 10V7 M for carbocyclic DBCC). It is difficult to assign a mechanistic significance to these apparent K, values since the combination between enzyme and coenzyme is essentially irreversible. One of the characteristics of the reaction catalyzed by dioldehydrase is that tritium from the C-l position of the substrate is incorporated, during the course of the catalytic process, into the C-5' position of the coenzyme (4). To establish whether this tritium exchange also occurs with carbocyclic DBCC, 60 pg of carbocyclic DBCC, Fraction B, were incubated at 10" with 242 units of dioldehydrase, 20 /Imoles of potassium phosphate buffer, Carbocyclic Vitamin Blz Coenxyme Vol. 245,No. 5  6. Isolation of dioldehydrase-carbocyclic DBCC complex. In a total volume of 0.845 ml 70 units of dioldehydrase, 7.1 X 10-6~ carbocyclic DBCC (tritium-labeled, specific activity = 1.16 X 106 cpm per pmole), 5.9 X 1W2 M KzHPO, were incubated at 37" for 10 min. The reaction mixture was layered onto a Sephadex G-75 column, 30 X 2.5 cm, equilibrated with 1 X 1W3 M K2HPOa. At a flow rate of 2 ml per min, 3-ml fractions were collected.
Each fraction was analyzed for protein concentration (X), radioactivity (01, and enzyme activity (u). Tritium-labeled carbocyclic DBCC was prepared with dioldehydrase and DL-1, 2-propanediol-1JH as described in the text. pH 8.0, and 16 pmoles of DL-1 , 2-propanediol-l-3H (specific activity 1.5 x 10' cpm per pmole) in a total volume of 2.8 ml. The reaction was allowed to proceed for 58 set and was stopped by the addition of 0.4 ml of 20% trichloracetic acid. Carbocyclic DBCC was then isolated and purified by procedures previously used with DBCC (4). The specific activity of the re-isolated carbocyclic DBCC was 1.1 x lo5 cpm per pmole.
Qualitatively, carbocyclic DBCC resembles DBCC in that hydrogen exchange occurs during the catalytic process between substrate and carbocyclic DBCC.
Insufficient material was available to carry out quantitative comparisons. To establish whether a different rate-limiting step occurs with carbocyclic DBCC than with DBCC, the isotope effect obtained and assayed for aldehyde as in the assay procedure (12). Inset shows the kinetics of aldehyde formation when DBCC and 0.1 unit per ml of dioldehydrase is used (12).
with C-1-deutero-DL-propanediol in the presence of carbocyclic DBCC was determined.
Under standard assay conditions the nondeuterated substrate reacted 8-to lo-fold faster than the deuterated substrate.
It has been previously shown (18) that with DBCC an isotope effect of 10 to 12 is obtained under these conditions. It is therefore highly probable that the same step, the breaking of the substrate C-H bond, is rate-limit.ing with both coenzymes.
The dioldehydrase-DBCC complex reacts with oxygen in the absence of substrate to give a catalytically inactive complex (19). This oxygen inactivation involves breaking of the carbon-cobalt bond. The stability of the dioldehydrase-carbocyclic DBCC complex was examined.
The results are summarized in Table  III.
The enzyme-carbocyclic DBCC complex is completely stable under conditions where the corresponding DBCC complex is largely inactivated.
The failure to observe loss of catalytic activity in the presence of carbocyclic DBCC might be attributed to the stability of the carbocyclic DBCC-enzyme complex or to the inability of the analogue to form a stable complex with the apoenzyme in the absence of substrate.
This latter interpretation is made unlikely by the demonstration that the carbocyclic analogue protects against inactivation by DBCC.
To test more directly whether a stable complex is formed, dioldehydrase was allowed to react with tritiated carbocyclic DBCC in the absence of substrate and then passed through a Sephadex column.
The experimental conditions and the elution pattern from the Sephadex column are shown in Fig. 6. Fractions 14,15,and 16 contained,respectively,6.6,9.0,and 6.7  The reaction mixture contained enzyme, 0.6 unit; potassium phosphate buffer, pH 8.0, 0.04 M; carbocyclic DBCC (FractionB), 1.2 X 10-e M; bovine serum albumin, 1 mg. The reaction mixture was previously incubated at 4" for 5 min to allow for the formation of enzyme-carbocyclic DBCC complex. Glycolaldehyde (0.2 pmole) was then added (final volume, 1 ml) and the tubes incnbated at 37". At times indicated, 0.1.ml aliquots were withdrawn and assayed for enzymatic activity under standard assay conditions (12). has been shown that 7 to 10 units of enzyme bind I pg of DBCC (11). The enzyme in the column effluent was completely saturated with coenzyme, since further addition of carbocyclic DBCC or DBCC did not increase the catalytic activity. This experiment establishes that the apoenzyme has high affinity for carbocyclic DBCC in the absence of substrate and further illustrates the stability of the carbocyclic DBCC-enzyme complex. It is not possible t.o isolate an enzymatically active enzyme-DBCC complex under the conditions employed for the isolation of the enzyme-carbocyclic DBCC complex. There are two additional processes which lead to the inactivation of the enzyme-Dl3CC complex.
(a) When ethylene glycol is converted t,o acetaldehyde, t.he enzyme-coenzyme complex becomes catalytically inactive (12). This inactivation does not occur when DL-I ,2-propanediol is the substrate.
(b) Addition of glycolaldehyde to dioldehydrase and DBCC leads to the formation of a catalytically inactive enzyme-DBCC-glycolaldehyde complex.
Formation of this complex is accompanied by a spectral change identical with that observed in the presence of substrate (12). Carbocyclic DBCC was examined with respect to both these points.
The time course of the reaction in the presence of ethylene glycol and DL-1 ,2-propanediol is shown in Fig. 7 Both CN-and NaBH43 are known to bring about heterolytic cleavage of the carbon---cobalt bond of cobalamin in which the electrons of the bond remain with the leaving group.
Presumably, this reaction is facilitated by the electronegative oxygen of DBCC which can stabilize the incipient negative charge at C-5'. No corresponding stabilization can occur with carbocyclic DBCC, where the oxygen is replaced by -CH2-.
The acid-catalyzed cleavage of the carbon-cobalt bond of cobalamin resembles the reaction with CN-and BH4 in that it also is a heterolytic cleavage in which the electrons remain with the leaving group. It differs from the reaction with CN-and BH,-in that it probably does not involve attack of nucleophile on the carbon-cobalt bond. It was therefore of interest to compare the susceptibility of the carbon-cobalt bond of carbocyclic DBCC and DBCC to acid. Our results, pertaining to this point, are ambiguous, since we have not identified the products of the acid reaction. Although, upon exposure of carbocyclic DBCC to acid, coenzymatic activity was lost, a light sensitive compound was still present. This suggests that loss of coenzyme activity may not be caused by cleavage of the carbon--cobalt bond, but by modification of another part of the molecule, and it is possible that carbon-cobalt bond of DBCC is less susceptible to acid cleavage than that of DBCC.
The dioldehydrase-DBCC complex and dioldehydrase-carbocyclic DBCC complex have similar properties in many respects. V max with the carbocyclic coenzyme for the conversion of DL-I ,2-propanediol to propionaldehyde is one-third that with DBCC.
With both coenzymes, tritium exchange between substrate and coenzyme is observed during the course of the catalytic process. Identical spectral changes are obtained when either substrate or glycolaldehyde are added to the complex. Glycolaldehyde inactivates both complexes.
In addition, the K, and over-all kinetic isotope effects are quite similar.
The mechanisms of the reactions with the carbocyclic coenzyme and with DBCC are probably identical and it can be concluded that the ribosyl oxygen does not play an important part in the catalytic process. Therefore, any mechanism involving carbon-cobalt bond cleavage in which the stabilization of the incipient negative charge at C-5' is dependent on oxygen must be excluded for the reactions in which B12 coenzymes participate.
A specific example of this type of mechanism was discussed in the introduction of this paper. This conclusion does not imply that carbon-cobalt bond cleavage does not occur by other mechanisms.
We believe that it is very likely that such a process occurs during the course of the reaction.
An important difference between the two coenzymes is the greater stability of the carbocyclic DBCC-dioldehydrase complex than the corresponding DBCC-enzyme complex, in the absence of substrate.
Although it is known that oxygen is involved in the inactivation of DBCC-enzyme complex, the the mechanism of the inactivation process is unknown.