Structure and Methylation of Coenzyme M (HSCH,CH,SO,)*

involved in methyl reactions in Methanobacterium.

Information derived from infrared, proton NMR, and ultraviolet spectroscopy as well as from chemical tests and quantitative elemental analysis reveals that the coenzyme is 2,2'-dithiodiethanesulfonic acid. Verification of this structure resides in the comparison of authentic with chemically synthesized 2,2'-dithiodiethanesulfonic acid. Evidence indicates that an active form of this cofactor is Z-mercaptoethanesulfonic acid which is methylated producing Z-(methylthio)ethanesulfonic acid; this derivative is subsequently reductively demethylated, yielding methane. Since Peak B contained from 70 to 88% of the total activity, this fraction was used for further purification. Ammonium acetate was removed by flash evaporation at 40". The CoM-containing eluate from the equivalent of 4 kg of wet cells was dissolved in 8 to 10 ml of water and fractionated on a Sephadex G-10 column (2.5 x 90 cm) that had been equilibrated with water.
The elution fluid was water. As shown in Fig. 2, CoM eluted in a single peak and was separated from much contaminating material that absorbed at 260 nm, the bulk of which was coenzyme Fd2,,. The pooled active fractions from the G-10 eluate were reduced to 10 ml by flash evaporation.
The sample was applied to a continuous electrophoresis apparatus, and separation was effected on a paper curtain at 400 volts (8 to 10 ma) in a pH 4,0.02 M ammonium acetate buffer. CoM migrated 10 to 13 cm toward the positive pole while descending 7 cm and was separated from fluorescent compounds.
The active fractions were flash-evaporated at 40" to remove the ammonium acetate buffer. The solid residue was taken up in a small amount of water and recrystallized three times from aqueous acetone, yielding colorless platelets with a melting point greater than 250".
The recovery of CoM from this purification scheme was 65 mg of crystalline CoM per kg of wet packed cells. Accurate determination of the recovery from this scheme was not possible as inhibitors of the methyltransferase reaction were present in both the initial, lyophilized water extract and the methanolsoluble fraction.
For example, the total activity found in Peak B of the anion exchange eluate (Table I)  isolated is the ammonium salt of 2,2'-dithiodiethanesulfonic acid (S-COM)~.
This compound possesses the following structure : +H,N-03SCH,CH2SSCH,CH,SOa-NH,+ Fig. 3 shows that the infrared spectra of authentic and synthetic (S-CoM)* are identical.
Each spectrum is largely dominated by the absorptions of the sulfonate and ammonium ions. The strong absorptions at Peaks d and e (1170 and 1035 cm-r) represent, respectively, the asymmetric and symmetric SOa stretching modes characteristic of sulfonic acid salts. The shoulders on the left of Peak d are probably the wag motion of the methylene groups attached to sulfur.
The ammonium ion stretch and deformation modes appear at a and c, respectively (3200, 3060, and 1435 cm-l); these absorptions disappear revealing the methylene stretch (2970 and 2930 cm-r) and deformation (1420 cm-l) modes when the ammonium ion is replaced by sodium ion as shown in the inset of Fig. 3. The fact that untreated aqueous solutions of both authentic and synthetic (S-CoM), gave a positive reaction with Nessler's reagent provides additional chemical evidence for the presence of the ammonium ion. Evidence suggests that the peaks located at g (590 and 530 cm-l) may involve the sulfonate group; Palmer (14), working with potassium dithionate, assigned the sharp peaks observed at 577 and 516 cm-r to be fundamental absorptions of the SOa deformation modes.
The absorptions observed with a variety of CoM derivatives which possess an intact sulfonate group as well as with sodium ethanesulfonate suggest that the peaks located at f and g involve the sulfonate moiety.
All absorptions in this region as well as in the SO, stretch region were lost in the spectrum of ammonium 3,3'-dithiodipropionate where the sulfonate group is replaced by a carboxyl group.
Data provided by Bellamy (15)  The possibility of having an AA'BB' system of the configuration of X(CH&H&Y is eliminated on the basis of quantitative elemental analysis, and particularly the synthesis of (S-CoM)t. The assignments made were based on observing the chemical shifts of the ethylene resonances of a variety of aliphatic sulfonate, sulfide, and disulfide compounds (17, 18). The presence of both a sulfonate and disulfide group in (S-CoM)s displaces the resonances of both methylene groups approximately 0.4 ppm downfield relative to the resonances observed in aliphatic compounds which contain only one of the functional groups of interest. Resonances of methylene groups next to the sulfonate functional group were in general located downfield of t.hose methylene groups next to sulfide or disulfide moieties.
Ultraviolet spectra of authentic and synthetic (S-COM)~ exhibit absorption maxima at 191 to 193 and 245 nm. The extinction coefficients which were determined for the short wavelength band for both preparations ranged between 5800 and 6400 liters mole-l cm-i. Because these measurements were made at the extreme lower end of the useful wavelengt,h range of the Cary model 14 instrument, attention should be drawn only to the order of magnitude. The extinction coefficient at 245 nm, however, may be determined with accuracy and for synthetic (S-CoM)z was found to be 380 liters mole-1 cm-i. The absorption maxima and extinction coefficients for (S-CoM)t are close to those obtained for straight chain aliphatic disulfides.  Table I; neither differs from that calculated by greater than 0.46%.
Authentic and synthetic (S-CoM)z were identical in their biological behavior. When 1.5 pmoles of synthetic (S-CoM)t were added to a standard reaction mixture which contained 4.9 pmoles of [methyl-r%]methylcobalamin and loo-fold purified methylcobalamin-CoM methyltransferase (8 pg of protein), a perfectly linear reaction rate was produced, with 400 nmoles of r4CHa-S-CoM being produced in 10 min. The reaction rate of authentic (S-COM)~ was identical. In a separate experiment where the reaction was allowed to proceed to completion, the reaction mixture contained 52 pg of loo-fold purified methyltransferase and 2.5 pmoles of [me&&14C]methylcobalamin. Two levels of synthetic (S-COM)~, 152 and 304 nmoles, and two levels of authentic (S-COM)~, 148 and 296 nmoles, were tested. For each mole of (S-CoM)t added 1.98 ZIZ 0.10 moles of methyl groups were bound. The ratio remained within the above described limits when either loo-fold purified methyltransferase was used with sodium borohydride as the electron donor or when crude extracts were used with sodium borohydride or hydrogen as the electron source.
Reduction of (S-CoM)L-The function of the electron requirement in the methyltransferase reaction was found to be that of reducing the disulfide bond of (S-CoM)z prior to methylation. FIQ. 4. NMR spectra of authentic and synthetic (S-CoM)s. The NMR spectra of (S-CoM)* in deuterium oxide (99.8% isotopic purity) were obtained with a Varian HA-109 NMR spectrometer. A NMR tube (2-mm inside diameter X 5-mm outside diameter X 17.7 mm) was used. The spectra were obtained under the following conditions (where the conditions are different, the value for authentic (S-CoM)2 precedes that for synthetic (S-CoM)t: concentration, 9 mg per 0.07 ml and 9 mg per 0.06 ml; temperature, '23"; frequency response, 2 and 20 Hz; radio frequency attenuator, 26 db; sweep time, 500 s; sweep width, 506 Hz (inset, 250 Hz); sweep offset, 0 and 472 Hz; spectrum amplitude, 10,009; lock signal, sodium 3-trimethylsilyl-propionate-2,2,3,3-dr (!Z'SP) and hydrogendeuterium oxide (HOD) ; field milligauss (manual oscillator frequency), 0.1 mG; field milligauss (sweep frequency), 0.04 and 0.1 mG; reference compound, TSP.
cobalamin. HS-CoM was stoichiometrically methylated when the methyltransferase-catalyzed reaction was allowed to proceed to completion.
B-(MethyZthio)ethanesulfonic Acid-The results presented above indicate that two methyl groups are bound for each mole of (S-CoM)a. The methyl derivative of the coenzyme retained a strong negative charge, suggesting that methyl-CoM could be 2-(methylthio)ethanesulfonic acid. Results presented in Fig. 6 show that the methylated coenzyme which was isolated from a large scale reaction mixture (see "Experimental Procedure") possesses an NMR spectrum identical with that of ammonium 2-(methylthio)ethanesulfonate which was chemically synthesized. The resonances of protons a and b appear, respectively, at 3.15 ppm (relative intensity, 1.8) and 2.88 ppm (relative intensity, 2.0). The large singlet labeled c at 2.17 ppm (relative intensity, 3.1) is typical of the proton resonances of a methyl group attached to sulfide. The small singlet at 2.24 ppm resulted from a contaminant in the solvent used for crystallization.
The presence of the methyl group shifts resonances a and b upfield, with that of b being shifted to a greater extent. This lends additional support to the proton assignments made. No evidence was obtained to support the possibility that the methylated derivative of CoM is the methylsulfonate ester. Such a CoM derivative would be a neutral molecule and possess physical properties considerably  6. Comparison of the NMR spectra of authentic and synthetic CHQ-S-CoM. The NMR spectra of CH&-CoM in deuterium oxide (99.8% isotopic purity) were obtained with a Varian HA-100 NMR spectrometer. A NMR tube (2-mm inside diameter X 5-mm outside diameter X 17.7 mm) was used. The spectra were obtained under the following conditions (where the conditions are different, the value for the enzymically synthesized, authentic CH&-CoM precedes that for synthetic CH&-CoM): concentration, 3 mg per 0.05 ml and 10 mg per 0.07 ml; temperature ,28'; frequency response 1 and 5 Hz; radio frequency attenuator, 20 db; sweep time, 1000s and 5OOs, sweep width, 509 Hz; sweep offset, -473 and 0 Hz; spectrum amplitude, 10,090 (inset, 10 X gain), and 6,ooO, lock signal, sodium 3-trimethylsilylpropionate-2,2,3,3-d, (TSP) and hydrogen deuterium oxide (HOD) ; field milligauss (manual oscillator frequency), 0.1 and 0.05 mG; field miliigauss (sweep frequency), 0.05 and 0.1 mG; reference, compound, TSP. The spectrum was obtained in deuterium oxide (99.8% isotopic pn;ity) with a Varian A-66 NMR spectrometer.
An NMR sample tube (Z-mm inside diameter X 5-mm outside diameter x 17.7 mm) was used. The spectrum was obtained under the following conditions: concentration of (CH&-S-CoM, 9 mg per 0.07 ml; temperature, 44'; different than those observed.
The methyl proton resonances of an aliphatic methylsulfonate ester lie very close to 3.9 ppm, a position very easily distinguishable from the resonances observed.
We examined the possibility that (S-CoM)t was methylated to yield 2-(dimethylsulfonium)ethanesulfonate ((CH&-S-CoM) which decomposed to CH&S-CoM prior to analysis. The NMR spectrum of chemically synthesized (14CH3)2-S-CoM is presented in Fig. 7. The spectrum in deuterium oxide revealed symmetrical ethylene proton resonances of the type AzBz at 3.65 and 3.46 ppm (cumulative relative intensity, 4.0). A singlet corresponding to the methyl proton resonances was located at 3.00 ppm (relative intensity, 5.5). CH,-S-CoM and (CH&-S-CoM were readily separable by anion exchange chromatography and thin layer electrophoresis. (CH8)2-S-CoM forms an internal salt between the sulfonium and sulfonate moieties. This imparts a charge considerably different from that of CHS-S-CoM.
(CH3'2-S-CoM was found to be completely stable under the conditions used for analysis. Reaction mixtures which contained methyltransferase were directly fractionated by anion exchange chromatography (Fig. 8A).
When limiting amounts of [methyZ-14C]methylcobalamin were allowed to react to completion, 99% of the radioactivity added was found in a single peak which possessed electrophoretic properties identical with that of chemically synthesized CH3-S-CoM (Fig. 8B). No radioactivity above background was found in the void volume (V,) where (CH,)2-S-CoM and methylcobalamin elute. Incubation of (14CHp)2-S-CoM in a methyltransferase reaction for 30 min at room temperature and for 30 min at 40" did not result in decomposition to i4CH3-S-CoM. Biological Activity of CH&X'oM--Chemically synthesized and authentic CH3-S-CoM exhibited identical biological activity. When 2.9 pmoles of authentic or synthetic CHS-S-COM were added to a reaction mixture (see "Experimental Procedure") which contained crude cell extract (3.3 mg of protein), a perfectly linear rate of methane formation was observed in each reaction vessel, the rates being identical; 22 nmoles of methane were produced in 25 min. In a separate experiment, the reaction filter band width 4 Hz; radio frequency field, 0.15 mG; sweep time 500 s; sweep width, 500 Hz; sweep offset, 0 Hz; spectrum amplitude, 32; integral amplitude, 56; reference compound, sodium 3-trimethylsilyl-propionate-2,2,3,3-d4 (TSP).
(a-a) indicates only that the cluster of resonances is from protons a and b; assignments have not been made. HOD, hydrogen deuterium oxide.
was allowed to proceed to completion with the methyl group from 101 and 202 nmoles of synthetic CH,-S-CoM being completely converted to methane.
As shown in Table II, (CH8)2-S-CoM was found to be completely inert as a methyl donor in the methylreductase-catalyzed reaction in incubation periods up to 70 min. No methane production was observed when (CH&-S-CoM and HS-CoM were present together in the reaction mixture; (CH&-S-CoM would not methylate HS-CoM to yield biologically active CH&CoM. Results from a separate experiment showed that CHa-S-CoM could not be further methylated by methyltransferase. DISCUSSION The fact that CoM was isolated as the disulfide possibly reflects an artifact of isolation. Thiol derivatives such as HS-CoM are, under neutral or alkaline conditions, easily oxidized to disulfides. Anaerobic precautions were not taken when CoM was isolated. Early work on the isolation of CoA revealed that much of it was in the form of a disulfide (20). Mild hydrolysis of partially purified CoA released 2,2'-dithiodiethylamine.
Disulfide formation may likely explain the presence of more than one active derivative of CoM in crude preparations.
The component of

Peak A in the QAE-Sephadex
A-25 eluate may be a heterodisulfide possessing only one sulfonate moiety RSSCH&H$O,-(R-S-S-CoM).
The sulfonate group will significantly govern the behavior of such a CoM derivative on an anion exchange column.
Thus, (S-COM)~, possessing two sulfonate moieties would be expected to elute at a higher ionic strength than a derivative possessing only one sulfonate group. This is in fact the case. (S-COM)~ elutes from a Q,AE-Sephadex A-25 column at approximately 2.5 M ammonium acetate whereas CH,S-CoM and HS-CoM both elute at a concentration less than 2 M. The observation that the ratios of CoM activity in Peaks A and B vary in a reciprocal manner from preparation to preparation may simply reflect the proportions of (S-COM)~ and R-S-S-CoM present.  (specific activity, 1730 cpm per nmole). The reaction was initiated by injecting an anaerobic solution of enzyme (104 pg of protein) with a syringe. The reaction was allowed to proceed until all of the [methyZ-i%]methylcobalamin had reacted. A portion of the reaction mixture (0.35 ml) was applied to a QAE-Sephadex A-25 (acetate form) column (1.2 X 15 cm) that had been equilibrated with water. Elution was effected at 5.4 ml per hour with 200 ml of a 0 to 2 M ammonium acetate linear gradient.
The profile shown was obtained by measuring the absorbance at 400 nm. (C&)2-S-CoM if present would elute with the void volume, VO. Cobalamin derivatives are indicated by B12. The per cent recovery was determined by comparing the total counts applied to the column to that recovered in the radioactive peak. B, the material in the radioactive peak (Fig. 8A) was lyophilized to dryness and dissolved in 0.5 ml of water; 5 ~1 were applied to a Brinkmann MN 300 cellulose thin layer plate. Standard (CH3)2-S-CoM and CH,-S-CoM also were applied.
Electrophoresis was performed in 0.4% pyridine-0.8yo acetic acid buffer at pH 4 (400 volts, 10 ma) for 30 min. The known samples were located by staining with iodine vapors. Radioactivity was determined by counting sections (1.0 X 1.5 cm) of the thin layer in Bray's solution.
The origin is indicated by X.
HS-CoM, 2-mercaptoethanesulfonate, has unique properties. It is one of the smallest and simplest of the known coenzymes, is highly acidic, and has an unusually high concentration of sulfur for its size. When the mercapto group is methylated, the coenzyme does not employ an "onium" ion as the active site of methyl transfer. This latter property stands in contrast to methyltransfer compounds such as S-adenosylmethionine, choline, betaine, dimethylthetin, or dimethyl propiothetin. The coenzyme has similarities to CoA and lipoic acid but so far apnears to have a comnletelv different role in biochemistrv. It I Each reaction mixture (2.0 ml) contained 100 pmoles of potassium phosphate buffer at pH 7.1, 10 rmoles of magnesium sulfate (in the appropriate flask 4.9 /rmoles of CH&CoM, 5.0 pmoles of HS-CoM, and 5.0 pmoles of (CHZ)&COM), 10 rmoles of ATP, and dialyzed crude extract (33.8 mg of protein).
The reaction rates of the positive control reactions were linear for 20 min. 125 CHa-S-CoM + (CHa)l-S-CoM.
. 80 will be interesting to see if the coenzyme can handle a Ci moiety more oxidized than a methyl group.