Isolation of the Subunits of Transcarboxylase and Reconstitution of the Active Enzyme from the Subunits

Abstract The separation of the 12 SH central subunit, the 5 SE peripheral metallo subunit, and the 1.3 SE biotinyl carboxyl carrier protein which are formed on the dissociation of transcarboxylase has been accomplished by molecular sieving on Bio-Gel. The 12 SH and 5 SE subunits have been obtained in nearly homogeneous form as judged by the sedimentation velocity profiles and by acrylamide gel electrophoresis. The 1.3 SE carboxyl carrier protein has given less consistent results; sometimes a single band of molecular weight of approximately 12,000 is obtained on gel electrophoresis in sodium dodecyl sulfate but sometimes there is an additional band of lower molecular weight of approximately 11,000. This lower molecular weight component may result from limited proteolytic degradation in spite of efforts to prevent it. Two or more bands are obtained in the absence of dodecyl sulfate. This heterogeneity may result from aggregation. subunits. The most effective reconstitution is accomplished by a two-step process. First, the 1.3 SE carboxyl carrier protein and 5 SE metallo subunit are combined to form the 6 SE complex; then this product is combined with the 12 SH subunit to yield active enzyme. With a limiting amount of the 6 SE complex and an excess of the 12 SH subunit the resulting enzymatic activity is proportional to the concentration of 6 SE complex. Likewise, with an excess of the 6 SF, complex and the 12 SH subunit limiting, the enzymatic activity is proportional to the concentration of the 12 SH subunit. The maximum specific activities of the reconstituted 6 SE complex and the 12 SH subunit were approximately 60% and 50%, respectively of the value estimated for their specific activities in the native 18 S form of the enzyme with three peripheral 6 SE subunits. Because assay of the 6 SE complex is done using an excess of the 12 SH subunit, this may yield enzyme with a single peripheral subunit. The 6 SE subunit may be less effective in this form. In the case of the 12 SH subunit, the activity is most likely low because the reconstituted 6 SE complex does not contain the full complement of 1.3 SE biotinyl carboxyl carrier proteins. The carboxyl carrier protein provides the groups which link together the 5 SE peripheral subunits with the central subunit. Evidence is presented that the 1.3 SE biotinyl carboxyl carrier must first combine with the 5 SE subunit to assume a form which effectively provides the combining groups for association with the 12 SH subunit.

This heterogeneity may result from aggregation.
Active enzyme is readily reconstituted from the isolated subunits.
The most effective reconstitution is accomplished by a two-step process. First, the 1.3 SE carboxyl carrier protein and 5 SE metallo subunit are combined to form the 6 SE complex; then this product is combined with the 12 SH subunit to yield active enzyme. With a limiting amount of the 6 SE complex and an excess of the 12 SH subunit the resulting enzymatic activity is proportional to the concentration of 6 SE complex. Likewise, with an excess of the 6 SF, complex and the 12 SH subunit limiting, the enzymatic activity is proportional to the concentration of the 12 SH subunit.
The maximum specific activities of the reconstituted 6 SE complex and the 12 Sn subunit were approximately 60% and SO%, respectively of the value estimated for their specific activities in the native 18  Cleveland, Ohio 44166. assay of the 6 SE complex is done using an excess of the 12 SH subunit, this may yield enzyme with a single peripheral subunit. The 6 SE subunit may be less effective in this form.
In the case of the 12 SH subunit, the activity is most likely low because the reconstituted 6 SE complex does not contain the full complement of 1.3 SE biotinyl carboxyl carrier proteins.
The carboxyl carrier protein provides the groups which link together the 5 SE peripheral subunits with the central subunit.
Evidence is presented that the 1.3 SE biotinyl carboxyl carrier must first combine with the 5 SE subunit to assume a form which effectively provides the combining groups for association with the 12 SH subunit.
It catalyzes the following reaction: CH&H (COO-)COSCoA + CHICOCOO-F! CH&H&OSCoA + -OOCCH&OCOO-Transcarboxylase dissociates to subunits which can be reconstituted forming an active enzyme as diagrammed in Fig. 1. The accumulated evidence (l-8) including electron microscopy (1) shows that three peripheral subunits (designated 6 SE) are attached loosely to one face of a cylindrical shaped central subunit (designated 12 S,).
The enzyme has a molecular weight of 790,000 and ~~0,~ = 18 S and dissociates at pH 8 and low ionic strength with sequential loss of the peripheral 6 SE subunits (1,6,7). The isolated enzyme frequently contains a mixture of the 18 S form with three peripheral subunits together with a 16 S form of the enzyme with two peripheral subunits (1,6).
The 6 SE subunit of molecular weight 144,000 dissociates slowly at pH 8 and rapidly at pH 9 to a dimeric metallo subunit of molecular weight 120,000 (designated 5 SE) (1,6) and to two biotinyl subunits of molecular weight of approximately 12,000 (designated 1.3 SE) (2,3). In the presence of denaturing agents, the dimeric 5 SE subunit dissociates to its constituent peptides (designated 2.5 S,) (1). The  1. Subunits of transcarboxylase are formed at low ionic strength and alkaline pH and they reassociate at high ionic strength in phosphate buffer, pH 6.5, or in acetate buffer, pH 5. The sedimentation coefficients are given with subscripts H indicating the "head" or central subunit or E indicating the "ear" or peripheral subunit (1). SDS, sodium dodecyl sulfate. slowly at pH 8 and more rapidly at pH 9 to dimeric subunits of molecular weight 120,000 (designated 6 Sn). The 12 Sn subunit does not contain either metals (Co2+ or Zn*) or biotin (1,3,5).
In the presence of sodium dodecyl sulfate or urea, the dimeric 6 Sn subunits dissociate to the constituent peptides (designated 2.5 Sn) (1). The 18 S form of the enzyme thus is made up of 18 peptides of three different types; 6 contain the metals (cobalt and zinc) ,I 6 are the 1.3 SE biotinyl carboxyl carrier proteins, and 6 make up the central 12 Sn subunit.
Glycerol (207,) retards the dissociation of the 12 Sn subunit to 6 Sn dimers and of the 6 SE subunit to 5 SE and 1.3 SE subunits. Furthermore, glycerol appears to maintain the native configuration of the 12 Sn and 6 SE subunits because they readily reconstitute to the active enzyme under appropriate conditions. Subunits formed in the absence of glycerol do not reconstitute as readily and give rise to an active enzyme2 with a sedimentation coefficient of approximately 24 S with the peripheral subunits attached to both faces of the central subunit (1,6).
The 12 Sn subunit has been isolated by glycerol gradient centrifugation and recombined with the 6 SE biotinyl metallo subunit in 0.25 M acetate at pH 5.0 to 5.2 or in 0.75 M phosphate at pH 6.5 to 6.8 to form active enzyme (1,3 All of the Bio-Gel columns were maintained in 0.02% sodium azide to inhibit bacterial growth. The other methods and reagents were as described previously (1,5,6,12). It then had a specific activity of 1.5 and Fig. 2B shows that it consisted of some ~16 S material and -12 S material but the major peak was -6 S material. This dissociated enzyme was placed on a Bio-Gel A-1.5m column (4.5 x 172 cm) which had been equilibrated with 0.05 M phosphate buffer (pH 7.0) and was eluted with the same buffer over a period of about 40 hours at 4'. Three protein peaks which were not completely resolved were obtained and were separately  Because the 2.5 Sn and 2.5 SE peptides are separated by this technique (7), this material contained little or no 6 Sn subunit.

Isolation of Subunits
It was, however, a mixture of 6 SE and 5 SE subunits; the latter being formed by partial dissociation of the 6 SE subunit.
That this had occurred was evident because there was a small peak of radioactivity in the eluate from the Rio-Gel column where the 1.3 SE subunit occurs. It is evident that there was considerable reconstitution to active enzyme during the chromatography on the Rio-Gel column because the dissociated enzyme (480 mg) had a specific enzymatic activity of 1.5, equivalent to 816 units whereas there were -4320 units in the 140 mg of protein in the pool from the first peak and it did not include all of the active enzyme eluted from the column.
Isolation of 12 SH Central Subunit and 5 SE Metallo Subunit from "Dead" Transcarboxylase-Our first successful large scale isolation of both the 12 Sn and 5 SE subunits was made possible by a fortuitous result during the preparation of a large batch of transcarboxylase.
On this occasion, all of the enzymatic activity was lost during the purification of transcarboxylase on a cellulose phosphate column.
The protein peak of the eluate with 0.3 M phosphate buffer (pH 6.8) which usually contains the transcarboxylase had a sedimentation coefficient of approximately 16 S, but there was no radioactivity associated with the protein, which indicated that the tritiated biotin had somehow been removed from the enzyme. This inactive protein, designated dead transcarboxylase, was stored at -20" for 2 years, until we learned that mild treatment with trypsin rapidly removes biotinyl peptides from normal transcarboxylase leaving the remaining portion of the molecule intact but inactive (8). The dead transcarboxylase was then examined and found to have a distinct advantage because dissociated dead transcarboxylase did not reassociate during Rio-Gel chromatography.
The dissociation was done as described above for the native enzyme and 420 mg of the dissociated dead transcarboxylase was fractionated on the Rio-Gel A-1.5m column.
Two protein peaks which were not completely resolved were obtained.
The peak fractions were carboxylase, the dissociated protein, and fractions after chromatography of the dissociated protein on a Bio-Gel A-1.5m column.
A, the original dead transcarboxylase, 15.7 S; B, the dissociated dead transcarboxylase, 13.0 S, and 6.9 S; C (top), first nrotein neak eluted from the Bio-Gel column. 11.8 S; C (bottom). protein irom fractions between the two peaks, 11.8 S and 5.5 St and D, protein from the second protein peak after dialysis at pH 9 and chromatography to remove any residual portion of the 1.3 SE subunit which might remain attached to the 5 Sx subunit, 5.8 S. Sedimentation was at 52,000 rpm at 4" for 56 min from right to left in a 30-mm double sector cell in the experiments of A, B, and D and a double sector wedge cell for C. The milligrams per ml of protein were: A, 2.5 mg per ml; B, 2.1 mg per ml; C (lop), 2 mg per ml ; C (bottom) 2 mg per ml ; and D, 2.2 mg per ml.
pooled separately and also the valley fractions between the two peaks. Sedimentation velocity profiles of the original dead transcarboxylase, of the dissociated dead transcarboxylase, and of the protein fractions from the Bio-Gel A-1.5m column are shown in Fig. 3. The material from the first peak ( Fig. 3C, lop) had an s20,W = 11.8 S. That from the second peak was used for isolation of 5 SE subunit as described below.
The protein from the valley fractions had values of 11.8 S and 5.5 S. It is apparent from the comparison of Fig. 2, C and D with Fig. 3C that unlike the subunits from normal transcarboxylase, those from dead transcarboxylasa did not reconstitute during the chromatography on Bio-Gel A-1.5m. The protein from the second peak was combined with another similar preparation from dead transcarboxylase for isolation of the 5 SE subunit.
The material (171 mg of protein in 10 ml) was dialyzed at room temperature for 23 hours with three changes of 1 liter each at 12, 17, and 23 hours against 0.05 M Tris-HCl (pH 9.0) containing 10m4 M phenylmethylsulfonylfluoride. This treatment dissociates from the 5 SE subunit any residual portion of 1.3 SE subunit that might remain attached to the 5 SE subunit. The material was then chromatographed on a Bio-Gel A-1.5m column (4.7 X 171 cm) using 0.05 M phosphate buffer (pH 7.0) containing low4 M phenylmethylsulfonylfluoride for the elution. A single symmetrical protein peak was obtained from the column. The pooled protein fractions were precipitated by 80% saturation with ammonium sulfate and then subjected to ultracentrifugation. The protein had an ~~0,~ = 5.8 S (Fig. 3D).5 Polyacrylamide gel electrophoresis of this protein and that from the first peak (Fig. 3C, top) in gels containing 8 M urea and sodium dodecyl sulfate (7) gave single bands showing that there was little cross contamination of the 5 SE subunit by the 12 Sn or 6 Sn subunits or of the 12 Sn subunit by the 5 SE subunit.
These two preparations of subunits in combination with the normal 1.3 SE subunit were found to be effective in reconstituting active transcarboxylase (see "Reconstitution of Transcarboxylase from Subunits").
Isolation of ld SH and 5 SE S&units by Other Methods-Mild treatment of transcarboxylase wit.h trypsin removes biotinyl peptides leaving the remaining portion of the enzyme intact but 6 The sedimentation coefficient of the 5 SE subunit which arises from the 6 SE subunit had not been determined previously because it had not been isolated.
It has been designated 5 SE for convenience in differentiating it from other subunits of similar sedimentation coefficients.
However, the yields have not been as good as with dead transcarboxylase because there is some recombination of the dissociated subunits even though a portion of the 1.3 SE subunit is removed by the trypsin treatment.
These results are described in detail by Ahmad et al. (8).
A method for isolation of 5 SE and 12 Sn subunits from transcarboxylase by complexing native enzyme with avidin-Sepharose has been developed by Berger and Wood (17). The complex is dissociated at pH 8 to liberate the 12 Sn subunit and then at pH 9 to obtain the 5 SE subunit. The 1.3 SE subunit remains in a complex with the avidin, thus the problem of reconstitution is eliminated and the eluated subunits may be collected under conditions which prevent their dissociation.
This procedure is useful in obtaining the 5 SE and 12 Sn subunit but is of no use in obtaining the 6 SE or 1.3 SE subunits.
Isolation of 1.3 SE Biotinyl Carboxyl Carrier Protein-We have not had consistent results in isolating a homogeneous form of this subunit.
Isolation of the 1.3 S, subunit as described previously (2) involved dissociation of transcarboxylase at 4" in 0.05 M Tris-HCl (pH 8.8) for 72 hours followed by a stepwise elution with an increasing concentration of KC1 from a DEAE-Sephadex A-50 column equilibrated with 0.05 M Tris-HCl (pH 8.8). The 1.3 SE subunit was obtained in a small breakthrough peak as well as in the 0.1 M KC1 eluate and gave a single band on polyacrylamide gel electrophoresis at pH 9 and had a molecular weight of approximately 12,000. Subsequently, this and other procedures have not given consistent results because more than one band is often observed on polyacrylamide gel electrophoresis. One method which we have adopted involves denaturation of transcarboxylase at 100" in 6 M urea plus 10ea M dithiothreitol and then chromatography on I&Gel A-1.5m in 6 M urea plus low4 M dithiothreitol.
As shown in Fig. 4, the resulting 2.5 S, and 2.5 SE peptides in Fractions 23 to 28 are easily observed by ultraviolet absorbtion at 280 nm and are well separated from the radioactive 1.3 SE subunit in Fractions 47 to 57. The 1.3 SE subunit, by virtue of its low content of aromatic amino acids, has a low ultraviolet absorbtion and is observed only by the presence of the aH label. The results of analytical polyacrylamide gel electrophoresis of the 2.5 Sn and 2.5 SE preparation in urea and of the 1.3 SE preparation in dodecyl sulfate are shown in the insets of Fig. 4. Two major bands were observed in the fractions containing the 2.5 SE and 2.5 Sn peptides and these had no radioactivity.
Fractions 47 to 56 containing the 1.3 SE subunit gave one major band which was highly radioactive and contained about 85y0 of the total recovered radioactivity.
There was a nearby smaller band with about 15y0 of the radioactivity. Some preparations obtained by this method give two bands in these positions almost equally stained and radioactive.
Both types of preparations are effective for the reconstitution of active enzyme in combination with 12 Sn and 5 SE subunits.
A second method for isolation of the 1.3 SF: subunit involves dissociation of transcarboxylase to the 6 Sn, 5 SE, and 1.3 SE subunits in 0.1 M Tris-HCl, pH 9, plus 10% glycerol. This mixture is then chromatographed on Bio-Gel A-1.5m (Fig. 5). The 6 Sn and 5 SE subunits are eluted in Fractions 60 to 76 which is consistent with their molecular weights (approximately 120,000) but the 1. The fractions were monitored for 280-nm absorbance and for radioactivity. Fractions 23 through 28 and 47 through 56 were pooled separately and each was dialyzed against 2000 ml of 0.02 M sodium phosphate buffer (pH 7) with three changes of buffer. Pool 23 to 28 was concentrated by rotary evaporation to 7.2 ml (~55 mg per ml) and Pool 47 to 56 tb 2.2 ml (some turbiditv develoned in this nool). The 1.3 SF nrenaration'was subjected to gel eiectrophoresis in dodecyl suliiie 6) and the 2.5 SE and 2.5 SH preparation to gel electrophoresis in urea (17). The results are shown in the insets. and/or (c) dissociation and reassociation of the 5 SE and 1.3 S3 subunit during the chromatography.
The protein in Fractions 60 to 76 was precipitated with ammonium sulfate and the protein in Fractions 80 to 102 was concentrated by lyophilization. Gel electrophoresis in the absence of denaturing agents of the preparation containing 6 Sn and 5 SH subunits gave two major bands and gel electrophoresis in sodium dodecyl sulfate of the 1.3 SE preparation gave a single band aside from a small band near the origin which may be an aggregate.
Some preparations of 1.3 SE have given two radioactive bands by this procedure.
A similar procedure has been used to obtain the 1.3 SE and 5 Sn subunits from the 6 SE subunit shown in Fig. 20, bottom.
Gel electrophoresis of preparations of the 1.3 SE subunit in the absence of dodecyl sulfate often yields multiple bands. This observation is considered under "Discussion." The concentration of the 1.3 SE subunit has been calculated from its biotin content on the basis that it contains 1 mol of biotin per mol of molecular weight of approximately 12,000 (2). For this purpose, the specific radioactivity (counts per min per nmol of biotin) of a given batch of transcarboxylase is determined from its total radioactivity and biotin content. Then knowing the total radioactivity of the 1.3 SE preparation, its biotin content is calculated from the specific radioactivity of the biotin. Reconstitution of Transcarboxylase from PurifLed 12 SH, 5 SE, and 1 .S XB Subunits-The reconstitution of transcarboxylase has been accomplished previously with the unresolved subunits of dissociated transcarboxylase either by adjustment to pH 5 with acetate buffer or by addition of a high concentration of phosphate buffer, pH 6.5 to 6.8 (3, 7) A similar experiment in which the 12 Sn subunit was limiting isolated 1.3 SE, 5 SE, and 12 Sn subunits does combine to form the and the reconstituted "6 SE" subunit was in excess is shown in active enzyme, but better yields are achieved when the conversion Fig. 7. Four concentrations of 12 Sn subunit were used with an is done by a two-step process. First, the 5 SE and 1.3 S, subunits excess of the reconstituted "6 SE" complex which was prepared in are combined to form the "6 SE" subunit,4 then the resulting acetate buffer at pH 5. The reconstitution was done in 0.75 M "6 SE" subunit is combined with the 12 Sn subunit to yield active phosphate buffer (pH 6.8) and the enzymatic activity was transcarboxylase.
The extent of the conversion is determined by assayed in O.Ol-ml portions. The results of Fig. 7 show that at measurement of the enzymatic activity of the reconstituted 27 hours, the enzymatic activity increased linearly with the enzyme by the usual spectrophotometric assay which measures concentration of the 12 Sn subunit up to 3.2 pg. the oxalacetate formed from pyruvate and methylmalonyl-CoA Determination of Specijic Activity of Subunits-It is evident using malate dehydrogenase.
from the experiments of Figs. 6 and 7 that the specific activity of Linearity of Enzymatic Activity with "6 SE" Subunit Limiting-a given subunit in forming active transcarboxylase could be The results of two experiments are shown in Fig. 6 in which the determined by making the one subunit limiting and the other 5 SE subunit was combined with the 1 3 SE subunit using phos-subunit in excess, much as is usually done in assaying an enzyme. phate buffer at pH 6.5 in one case and acetate buffer at pH 5 A comparison between the activities of the buffer (pH 6.5) which were held at 0" and assayed over a period of reconstituted subunit with that of the native enzyme, therefore, several days. The mixtures contained 0.205 nmol of the 12 Sn requires an estimation of the number of 1.3 S, subunits bound to subunit and 0.029, 0.058, or 0.116 nmol (3.5, 7, and 14 pg) of each 5 SE subunit.
For this purpose, the [3H]biotinyl 1.3 SE the 5 SE subunit, which had been converted in part to the "6 SE" subunit and 5 SE subunit were reconstituted and then the unsubunits.
Thus, the 12 Sn subunit was in large excess. bound 1.3 SE subunit was separated by glycerol gradient cen- Fig. 6 shows that the enzymatic activity (AA per min per 0.01 trifugation from the 5 SE subunits containing bound 1.3 SE ml) was linear at 96 hours with 3.5 and 7 pg of 5 SE subunit when subunits.
The results of such an experiment are shown in Fig. 8. either phosphate at pH 6.5 or acetate at pH 5 was used in the Two peaks of radioactivity were obtained; the first coinciding reconstitution.
The Pool 21 to 28 was used without further purification for the reconstitutions of the experiment of Fig. 9.

923
Pools 21 to 28 from the glycerol gradient were used directly for reconstitution with the 12 Sn subunit in 0.75 M phosphate buffer (pH 6.5). The results are shown in Fig. 9. The proportion of 12 Sn subunits to "6 SE" subunits was varied so as to obtain ratios of sites which varied from 17 to 0.13 assuming a 12 Sn subunit has six sites for binding a 1.3 SE subunit and that one 1.3 SE subunit bound to one 5 SE subunit yields one "6 Se" site. The specific activity of the "6 SE" subunit was calculated from the observed enzymatic activity of the reconstituted mixture and it was assumed that 1 nmol of 1.3 SE subunit bound to a 5 SE subunit is equivalent to 0.072 mg of "6 SE" protein (this follows because 1 nmol of 6 SE subunit containing 2 nmol of 1.3 SE subunits is equal to 0.144 mg). If the enzymatic activity is the same for the 1.3 SE subunit, whether one or two are bound to the 5 SE subunit, then this calculation is correct.  924 activity of 27 for the 18 S form of the enzyme with a full complement of 3 complete 6 SE subunits (50 X 3 X 1.44 X 105/7.90 x lo5 where 1.44 x lo5 is the molecular weight of the 6 SE subunit and 7.90 x lo5 of the 18 S form of the enzyme). The observed specific activity for the 18 S form of the enzyme is approximately 45, thus, the "6 SE" subunit had about 60 y0 of the activity in the reconstituted enzyme as the same amount of this subunit has in the native enzyme. Fig. 9B gives the specific activities based on the content of the 12 Sn subunit.
A maximum specific activity of approximately 40 was reached when the ratios of the "6 SE" subunit sites to 12 Sn subunit sites were 7.6 and 3.0. A specific activity of 40 for the 12 Sn subunit is equivalent to an activity of 18.2 for the 18 S form of transcarboxylase (40 x 3.6 x 105/7.9 X 10" where 3.6 x lo5 is the molecular weight of the 12 Sn subunit and 7.9 x lo5 of the 18 S transcarboxylase).
Thus, the 12 Sn subunit had about 40% of the activity in the reconstituted enzyme as the same amount of this subunit has in the native enzyme.
In other experiments this value has been as high as 50%.
Competition of Subunits with "6 SB" Subunit for Binding with 12 SH Subunit-The 1.3 SE subunit is required to form a complex between the 5 SE and 12 Sn subunits in transcarboxylase (8) and apparently provides the groups that promote this linkage. We have noted that more complete reconstitution with the 12 Sn subunit is obtained if the reconstitution is done in two steps in which the 5 SE plus 1.3 SE subunits are first combined to "6 SE" subunits which are then combined with the 12 Sn subunit. These results suggested that the 1.3 SE subunit, when combined with the 5 SE subunit, might undergo a conformational change which would facilitate combmation with the 12 Sn subunit. If this were so, the free 1.3 SE subunit might not compete effectively with the "6 SE" subunit in forming a complex with the 12 Sn subunit during reconstitution.
To test this hypothesis, the ability of the 1.3 SE subunit to compete with the "6 SE" subunit for 12 Sn sites was investigated in experiments in which the amount of "6 SE" subunit and 12 Sn subunit were maintained about equal on the basis of "sites" and a large excess (approximately go-fold) of 1.3 SE subunit was added during the reconstitution in 0.75 M phosphate buffer.
In addition, the effect of the non-biotinyl peptide was tested. This is the portion of the 1.3 SE subunit which remains combined with transcarboxylase when the biotinyl peptides are cleaved from it by trypsin. This peptide has been isolated and contains the portion of the 1.3 SE subunit which provides the combining groups for linkage of the 5 SE and 12 Sn subunits (8). The effect of an excess of 5 SE subunits was also tested.
The results of these experiments are presented in Table I, which are expressed as enzymatic activity of the reconstituted mixture on the basis of the specific activity of the 12 Sn subunit. It is seen that neither the 1.3 SE subunit nor the nonbiotinyl peptide had a significant influence on the amount of enzyme reconstituted. This lack of inhibition by a large excess of the 1.3 SE subunit compared to the "6 SE" and 12 Sn subunits makes it very likely that the 1.3 SE subunit has a much stronger affinity for the 12 Sn subunit when if forms a complex with the 5 SE subunit than does the free 1.3 SE subunit or the nonbiotinyl pcptide.
Most likely, the 1.3 SE subunit undergoes a conformational change during the formation of the complex which increases its capacity to combine with the 12 Sn subunit.
The 5 SE subunit, as expected, had no influence on the reconstitution of active enzyme.

DISCUSSION
Transcarboxylase is a monofunctional enzyme catalyzing only a single metabolic reaction but in many respects it resembles the multi-enzyme complexes with components capable of catalyzing two or more metabolic reactions.
Ciassical examples of the multiple enzyme complexes are the cr-keto acid dehydrogenases (18) and the fatty acid synthetases (19-21) in which the lipoyl moiety serves as an arm to link the various enzymes of the a-keto acid dehydrogenases and the 4'-phosphopantetheine to link the enzymes of the fatty acid synthetascs.
Transcarboxylase is similar because it involves two partial reactions (shown below) which in combination make up the over-all reaction. 12  complex in which the lipoyl or the phosphopantetheine group serves as the link. There are numerous biotinyl enzymes in which the biotinyl group serves as a carboxyl carrier (see review by Moss and Lane (23)) but transcarboxylase is the only one which has been isolated as an intact complex, dissociated to its subunits, the subunits isolated, the partial reactions studied with the subunits, and the active enzyme complex reassembled from the isolated subunits.

925
Similar studies have been done with acetyl-CoA carboxylase of E. coli (24) but in this case the complex has not been isolated nor have the constituent subunits been reassembled into a complex. In this regard, acetyl-CoA carboxylase of E. coli resembles the fatty acid synthetase from E. coli, which has not as yet been obtained as a multi-enzyme complex (24, 25). Thus far, we have devoted most of our efforts to isolating the 12 Sn, 6 SE, 5 SE, and 1.3 SE subunits by chromatography on Bio-Gel; separation of the 6 Sn, 6 SE, and 5 SE subunits by this technique is not feasible because of their nearly identical size. Although studies with the 6 Sn subunit have been very limited, it seems likely that the 6 Sn subunit does reassociate to form the 12 Sn subunit.
For example, a major portion of the dissociated enzyme of the experiment of Fig. 2 had a sedimentation coefficient of approximately 6 S (Fig. 2B). When this material was passed through a 13io-Gel column much of it was recovered in the form of the reconstituted -18 S enzyme and the 12 Sn subunit (Fig. 2C). Clearly, the 6 Sn subunits were reconverted to the 12 Sn subunit during the chromatography and subsequently combined with 6 SE subunits to form the -18 8 enzyme.
The 12 Sn subunit has been isolated from dead, normal (17), and trypsinizcd transcarboxylasc (8). Each preparation has been found to be active with 6 SE subunits in yielding reconstitutcd active enzyme.
The 12 Sn subunit from trypsinizcd transcarboxylase has had a somewhat lower specific enzymatic activity on reeonstitution than the other 12 Sn subunits. The 5 SE subunits from all three sources have been found to be active but the activities of the different preparations have not been compared extensively.
Both the 12 Sn and 5 SE subunits have consistently been obtained in nearly homogeneous form as judged by sedimentation velocity and by gel electrophoresis.
The same has not been true of the 1.3 SE subunit.
Gel electrophoresis in sodium dodecyl sulfate has yielded a single band with a number of preparations but some preparations have given two major bands. Gel electronhoresis in the absence of dodecvl sulfate but in the presence of mercaptoethanol frequently yields multiple bands which may result from aggregation.
Likewise, there is evidence of a self-associating system during sedimentation equilibrium studies of the 1.3 SE subunit6 A further difficulty with the 1.3 SE subunit is that with storage it frequently loses its capacity to promote formation of active enzyme in combination with the 5 SE and 12 Sn subunits.
Most of the studies reported here were done with preparations that had been stored frozen no more than 2 or 3 weeks (usually less).
Vagelos and co-workers in a series of publications (26-30) have described multiple forms of the biotinyl carboxyl carrier protein from E. coli. Molecular weights of 9,065, 10,267, -22,500, and -45,000 have been observed and prolonged dialysis of the homogeneous protein resulted in polydisperse mixtures with molecular weights ranging from 20,000 to greater than 200,000 (29). They propose that the protein of molecular weight 45,000 is a dimer and is the native form of the carboxyl carrier protein and that the forms smaller than 22,500 arise by proteolysis which occurs during the isolation of the protein (30).
We have attempted to prevent proteolysis by use of benzamidine-HCl as employed by Fujikawa et al. (31) and also of phenylmethylsulfonyl fluoride (32). The cells were broken in the presence of these inhibit.ors and they also were included at each stage of purification of the enzyme. The dissociation was done by addition of the enzyme to boiling 6 M urea plus lop3 M 6 F. Ahmad, unpublished observations. dithiothreitol. Gel electrophoresis in the presence of dodecyl sulfate gave a major single band of the 1.3 SE subunit.
Our preparations have molecular weights of 11,000 to 13,000.
The present study is the first that the authors are aware of in which it has been possible to determine the activity of a subunit in forming the active enzyme by making the given subunit limiting and adding the other subunits to it in cxccss. The comparison of the activity of the given subunit with its activity in the intact enzyme is made somewhat uncertain, however, because various preparations of transcarboxylase, albeit pure as judged by their sedimentation behavior, possess varying specific activities.
There are two factors which may influence the specific activity.
The first factor is the number of peripheral biotinyl subunits that are attached to the central subunit.
The 18 S form of the enzyme has three peripheral subunits and has a specific activity of approximately 45 but values in the 50s have been observed occasionally.
The 16 S form has only two peripheral subunits and has a correspondingly lower activity. In addition, there is a -24 S form of the enzyme2 (1, 6) which may have six peripheral subunits attached to the central 12 Sn subunit. This form has not been isolated and its specific activity is not known but may be greater than that of the 18 S form. The second factor influencing specific activity is manifest by a loss of enzymatic activity (sometimes quite rapidly) which is not accompanied by a change in the sedimentation coefficient or the dissociation to subunits (6,12). Sometimes this loss in activity can be restored by incubation in 1.5 M (NH&S04 at 25" (6). The above factors make it difficult to evaluate the efficiency of reconstitution of active enzyme from the isolated subunits. We have chosen as our standard of comparison a specific activity of approximately 45 for the 18 S form of the enzyme and have compared the observed specific activity of the subunits with this value. Thus, the theoretical maximum specific activity for the 12 Sn subunit is 98.8 (45 X 7.9 x 105/3.6 X 105) and for the 6 SE subunit is 82.3 (45 X 7.9 x 105/3 X 1.44 X 105) where 7.9 x lo5 is the molecular weight of the 18 S form of the enzyme, 3.6 x 10" of the 12 Sn subunit and 1.44 X lo5 of the 6 SE subunit (1). On this basis, when the 12 Sn subunit was made limiting and the reconstituted "6 SE" subunit4 was added in excess, the specific activity of the 12 Sn subunit had about 507, of its potential activity.
There are several factors which may cause the specific activity to be lower than that calculated from the standard.
One is the uncertainty of the value to be used for the standard as explained above. A more important factor is the fact that the "6 SE" subunit formed by reconstitution from the 5 SE and 1.3 SE subunits only carried about 50 y0 (Fig. 8) of the full complement of 1.3 SE subunits which is two per 5 SE subunit. Thus, when the 12 Sn subunit is made limiting, three peripheral "6 SE" subunits which are deficient in the 1.3 SE subunits may form a complex with the 12 Sn subunit.
In this case, some of the 12 Sn sites would be ineffective in the transcarboxylation reaction because they would lack the biotinyl carboxyl carrier protein which is essential for the activity of that site to become evident. It is not clear why we have been unsuccessful in obtaining a complete conversion of the 5 SE subunit to a 6 SE subunit with its full complement of 1.3 SE subunits.
Perhaps a larger excess of the 1.3 SE subunit is required than was used in the reconstitution of the "6 S n" subunit.
The situation with regard to the specific activity of the "6 SE" subunit is somewhat different.
Here, the calculation of its specific activity has been done on the basis of each nanomole of complexed 1.3 SE subunit being equivalent to 0.072 mg of "6 SE" subunit (0.144 mg = 1 nmol of 6 SE with 2 nmol of 1.3 SE).

926
Thus, if there is only one 1.3 SE subunit in combination with a 5 SE subunit, only one-half of the weight of 6 SE is considered in calculating the specific activity and as noted above, a theoretical maximum specific activity of 82 would be anticipated for the 6 SE subunit.
The observed value was about 50 (Fig. 9) when the 12 SH subunit was in large excess. Under these conditions, the reconstituted enzyme may consist of forms with only one peripheral "6 SE" subunit combined with a 12 SH subunit.
Possibly there is cooperativity and a single "6 SE" subunit bound to a 12 SH subunit is less effective than when two or more are bound on the same 12 SH subunit.
A more detailed study is required to obtain information about this possibility.