Transport of Vitamin BI2 in Escherichia coli SOME OBSERVATIONS ON THE ROLES OF THE GENE PRODUCTS OF BtuC AND TonB’

The transport of vitamin BIZ (cyanocobalamin, CN- Cbl) in Escherichia coli requires the proton motive force and functional products from the genes BtuB, TonB, and BtuC. E. coli cells that contain the btuC mutation are defective in CN-Cbl transport across the inner membrane and accumulate large concentrations of CN-Cbl in the periplasmic space. This uptake of CN-Cbl into the periplasmic space requires the TonB gene prod- uct, the proton motive force, and two functional sites on the outer membrane cobalamin (Cbl) receptor pro- tein (the Cbl-binding site and a second site which is necessary for transfer of Cbl from the receptor towards the interior of the cell). We propose that the TonB gene product, the proton motive force, and the second do- main on the receptor interact to increase the rate of dissociation of the receptor-Cbl complex, thus releas-ing Cbl into the periplasmic space. We also suggest that a reasonable mechanism for this catalyzed release of Cbl is via the interaction of the second functional do- main on the receptor with a diffusible messenger that is generated in the periplasmic space by the combined action of the proton motive force and the inner membrane TonB protein. a 3- to 12-fold of CN-Cbl transport across the outer membrane, was seen in btuC cells, membrane transport of CN-Cbl in wild type cells. resulted the of CN-Cbl and

Transport of Vitamin BI2 in Escherichia coli SOME  The transport of vitamin B I Z (cyanocobalamin, CN-Cbl) in Escherichia coli requires the proton motive force and functional products from the genes BtuB, TonB, and BtuC. E. coli cells that contain the btuC mutation are defective in CN-Cbl transport across the inner membrane and accumulate large concentrations of CN-Cbl in the periplasmic space. This uptake of CN-Cbl into the periplasmic space requires the TonB gene product, the proton motive force, and two functional sites on the outer membrane cobalamin (Cbl) receptor protein (the Cbl-binding site and a second site which is necessary for transfer of Cbl from the receptor towards the interior of the cell). We propose that the TonB gene product, the proton motive force, and the second domain on the receptor interact to increase the rate of dissociation of the receptor-Cbl complex, thus releasing Cbl into the periplasmic space. We also suggest that a reasonable mechanism for this catalyzed release of Cbl is via the interaction of the second functional domain on the receptor with a diffusible messenger that is generated in the periplasmic space by the combined action of the proton motive force and the inner membrane TonB protein.
Arsenate was found to have two distinct effects upon CN-Cbl uptake, a 3-to 12-fold stimulation of CN-Cbl transport across the outer membrane, which was seen in btuC cells, and inhibition of inner membrane transport of CN-Cbl in wild type cells. The latter effect resulted in the periplasmic accumulation of CN-Cbl and may indicate a requirement for phosphate bond energy in Cbl transport across the inner membrane. A variety of other reagents, including arsenite, diamide, N-ethylmaleimide, and sucrose, gave small stimulations (usually <2-fold) of CN-Cbl uptake in btuC cells, possibly as a result of increasing the size of the periplasmic space.
Most of the genes involved in the transport of cyanocobalamin (CN-Cbl) across the cell envelope of Escherichia coli are shared with other membrane processes. The first step in this transport process is the binding of Cbl' to the outer membrane Cbl receptor, the protein component of which is the product of the BtuB gene and also functions as the receptor for bacteriophage BF23 and the E colicins (1,2). The * This investigation was supported by United States Public Health Service National Institutes of Health Research Grant AM12653 from the National Institute of Arthritis, Metabolism, and Digestive Diseases. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "adoertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom requests for reprints should be addressed.
' The abbreviations used are: Cbl, cobalamin; Ado, adenosyl-; CN, cyano-; Me, methyl-. subsequent energy-dependent transfer of Cbl from the receptor to the interior of the cell requires the proton motive force and the products of the genes TonB and BtuC (3-6). The proton motive force participates in many different transport processes in E. coli  In contrast to the genes BtuB and TonB, BtuC has no known function apart from its involvement in Cbl transport. Mutants in this locus were first isolated and characterized by Di Girolamo et al. (5,8) who showed that such strains, which were also metE, required increased concentrations of Cbl in the medium for growth in the absence of added methionine and displayed several changes in Cbl uptake. These changes included greater exchangeability of cellular Cbl with exogenous Cbl, a reduced energy-dependent phase of Cbl transport, and reduced conversion into coenzyme forms of the Cbl taken UP.
The results presented in this paper were obtained primarily from btuC mutants and provide some clarification of the roles of the BtuC and TonB gene products in Cbl transport. We have concluded that the proton motive force and the TonB gene product are involved in the release of Cbl from the outer membrane receptor and that the BtuC gene product is necessary for Cbl transport across the inner membrane.  As KBTOO1, but also btuA4l BtuA-KBT069

KBTlO3
As KBT001, but also btuC103 BtuC-1485F-AtonB-trp AtonB-trp TonB-str, lac, tonA metH, TonA' acid. In some experiments with the low phosphate medium, nitrilotriacetic acid at a final concentration ofO.1 mM was added to minimize the concentration of iron available to the cells. The cells were harvested during the mid-log phase of growth, at a density between 5 X 10' and 1 X 10' cells/ml. After washing three times with minimal Hepes, the cells were resuspended in minimal Hepes at a density of about 1 X lo1" cells/ml. Minimal Hepes contained 90 mM Hepes, 15 mM potassium chloride, 0.4 m~ magnesium sulfate, and 20 p~ calcium nitrate, and was adjusted to pH 6.6 with potassium hydroxide. The washed cell suspensions could be stored at 0-4°C for several days without loss of appreciable Cbl uptake activity.

CN-Cb1
Uptake-The methods were essentially the same as those described previously (3) and consisted of incubating cells of E. coli with CN['H]CbI or CN["7Co]Cbl, followed by separation of the cells from the reaction mixtures by filtration through Millipore filters. The reaction mixtures usually contained about 3 X lo8 cells/ml, 10 mM labeled CN-Cbl, 40 mM energy source (D-lactate, glucose, or potassium succinate), and were made up to volume with the buffer, either minimal Hepes or 0.1 M potassium phosphate at pH 6.6. The components of the reaction mixtures were also made up in the appropriate buffer solution. The cells were preincubated aerobically at 37°C with the energy source for 5 min before addition of the labeled CN-Cbl. Samples (1 m l ) were removed at different times and were filtered through Millipore membrane fdters (0.45-pm pore size, 25-mm diameter). The filters were washed twice with 10 ml of 0.2 M lithium chloride, dried, placed in 10-ml Filmware tubes (Sybron/Nalge Corp.) with 1.5 ml of toluene containing 0.4% 2,5-diphenyloxazole (PPO) and 0.01% 1,4-bis[2-(4-methyl-5-phenyloxazolyl)]benzene (dimethyl-PO-POP) and were counted in a Beckman LS-1OOC liquid scintillation counter. The results are expressed as picomoles of CN-Cbl taken up per 10' cells.
Osmotic Shock Treatment ofthe Cells-In experiments where the effects of an osmotic shock upon the retention of the Cbl taken up were examined, the shock procedure of Nosqal and Heppel (10) was modified as follows. Prior to the shock, Cbl uptake was measured by removing 0.1-ml samples from a reaction mixture that contained about 3 X loy cells/ml. The osmotic shock procedure was initiated by the addition of solid sucrose to such reaction mixtures to give a final concentration of 10%. After a further 4-min incubation at 37°C to allow the cells to equilibrate with the sucrose, potassium ethylenediaminetetraacetate was added to a final concentration of 2 mM and the suspension was then immediately diluted 100-fold with distilled water at 37'C. Samples (5 ml) were then taken at different times to measure the amount of labeled Cbl still associated with the cells.
Conversion of CN-Cbl into Other Cobalamins-The extent of conversion of CN["Co]Cbl into other cobalamins was examined by means of the following procedure which was performed in dim light or darkness. The cells were grown on media containing CN['Co]Cbl and were harvested during the late log phase of growth. After suspension in distilled water and addition of 3 volumes of absolute ethanol and 25 (11 of a Cbl standard mixture (containing -120 pg each of unlabeled AquoCbl, AdoCbl, CN-Cbl, and MeCbl), the cells were heated at 70-80" for 15 min. Precipitated protein and other insoluble matter was removed by centrifugation and the ethanolic solution was concentrated on a warm water bath under a stream of air to less than 1 ml and was then diluted with deionized water to about 5 ml. The corrinoid compounds were purified by phenol extraction, reextracted into water, concentrated to small volume, and electrophoresed on Schleicher and Schuell 507c paper in 0.5 M acetic acid. Autoradi-Kodak XR-5 film.
ographs were prepared after exposure of the electrophoretograms to Detection of Arsenate and Arsenite-The paper chromatographic procedure of Yamaguchi (11). with Whatman No. 1 paper and a mixture of 3 volumes of methanol and 1 volume of 1 M ammonium hydroxide as solvent, was used to check the purity of our supplies of arsenite and arsenate, and to examine the ability of E. coli cells to interconvert these compounds. Arsenate, arsenite, and phosphate were detected on the chromatograms by spraying with ammoniacal silver nitrate.

RESULTS
Cbl Uptake Patterns-The strains of E. coli used in this study show four distinct patterns of CN-Cbl uptake (Fig. 1). Normal uptake by wild type cells can result in a proton motive force-dependent accumulation of more than 25 pmol of CN-Cbl/109 cells/h, equal to about 15,000 CN-Cbl molecules/cell or an intracellular Cbl concentration of about 25 PM. Two of the uptake patterns are products of defects in the gene BtuB which is the structural gene for the outer membrane Cbl receptor protein. Strain KBT069, which possesses an average of 0.5 Cbl receptors/cell(2), took up about 100 Cbl molecules/ cell/h. There was virtually no measurable Cbl uptake by cells of strain RK4102-74-3, which is completely lacking Cbl receptors (results not shown). The btuA mutation is located within the BtuB cistron (6) and results in a modified Cbl receptor in which the Cbl-binding site is normal but the receptor is unable Note that the scale of CN-Cbl uptake in B is different from that in A. pattern of uptake was also characteristic of tonB cells (Fig. 1) and of wild type and btuC cells incubated with CN-Cbl in the presence of 2 m~ 2.4-dinitrophenol (results not shown). Cells containing the btuC lesion show a proton motive force and TonB-dependent accumulation of Cbl which has an initial rate approximately equal to that of the wild type, but which reaches a plateau after taking up about 3 to 5 pmol/109 cells. [6'Co]Cbl. Cells of each strain were grown aerobically at 37'C on minimal media without added methionine but containing 1 and 10 nM CN['%o]Cbl. After harvesting the cells in the mid-log phase of growth, the cobalamin compounds associated with the cells were extracted as described under "Materials and Methods," and then electrophoresed on Schleicher and Schuell507c paper in 0.5 M acetic acid. Following electrophoresis, the paper was exposed to x-ray film.
The figure shows a photograph of the developed x-ray film. The darkened areas represent the images generated on the film by the "Co contained in the cobalamin extracts. The patterns obtained with the extracts from strain KBTOOl are shown on the left while those from KBT103 are shown on the right. The central track is the control pattern given by some of the original CN["Co]Cbl which had not been incubated with living cells but which had been put through the same extraction and electrophoresis procedures. The positions to which the standard unlabeled cobalamins migrated during the electrophoresis are shown on the extreme right. The cells were grown on the low phosphate medium with glucose as the main carbon and energy source, and the experimental procedures were the same as described in the legend to Fig. 1 metE cells is consistent with the ability of these cells to grow on media containing CN-Cbl but not methionine. Accordingly, we believe that btuC is unlikely to be a defect in one of the enzymes responsible for synthesis of the Cbl coenzymes. This conclusion is also supported by our earlier observation that there is no obligate coupling between Cbl uptake and ita chemical conversion into other forms (12).
If the enzymes for the conversion of CN-Cbl into AdoCbl and MeCbl are present, the most likely explanation for the lack of conversion of most of the CN-Cbl taken up by btuC cells is inaccessibility of the CN-Cbl to the enzymes. We suggest that this is indeed the case and that most of the CN-Cbl taken up by btuC cells is located in the periplasmic space.
The results shown in Fig. 3 are consistent with this suggestion. Panel A shows the effects of an osmotic shock upon the ability of wild type and btuC cells to retain CN[3H]Cbl. Only about 12% of the Cbl taken up by the wild type strain was released by the osmotic shock, compared with about a 70% release from the btuC strain. The CN[3H]Cbl taken up by these two strains also responds differently to the addition of EDTA (Fig.  3B). The CN-Cbl accumulated by btuC cells is released by EDTA, and the rate of the release, compared with rates measured in an earlier study (13), indicates that it is receptormediated, rather than the result of a nonspecific increase in the permeability of the outer membrane. Although EDTA is known to have several distinct effects upon the cell envelope of E. coli, its ability to release Cbl from btuC cells is evidently dominated by its ability to act as a competitive inhibitor of Cbl binding to the Cbl receptor. EDTA caused only a small transient loss of Cbl from wild type cells, followed by continued Cbl uptake at a slower rate than that seen before the EDTA addition.
The differences between btuC and wild type cells in their responses to EDTA and unlabeled CN-Cbl, and in the formation of Cbl coenzymes, are indications that the Cbl taken up is located in different cell compartments in these two strains and that the Cbl-containing compartment of btuC cells is more accessible to the medium. The release of most of the Cbl from btuC cells by an osmotic shock is strong evidence that the Cbl is located predominantly in the periplasmic space in this strain, and indicates that the BtuC gene product functions in Cbl transport across the inner membrane. If we assume that the periplasmic space occupies about 20% of the cell volume interior to the outer membrane, 3 to 5 pmol of Cbl taken up per IO9 cells represents a concentration of about 15 to 25 ~L M in the periplasmic space. It is evident from earlier studies (3) that this accumulation of Cbl by btuC cells is dependent upon both TonB and the proton motive force. Under "Discussion," we consider some possible mechanisms of coupling the proton motive force and the TonB gene product to transport of Cbl across the outer membrane.
Arsenate Effects upon Cbl Uptake-In previous experiments, in which Cbl transport in E. coli was shown to be dependent upon the proton motive force, arsenate had little effect upon the system (3). Since those experiments measured the early rates of energy-dependent Cbl uptake, they would not have shown any appreciable difference between wild type and btuC cells. The suggestion above that the proton motive force may be primarily concerned with the movement of Cbl across the outer membrane, together with some earlier observations that the uptake of some Cbl analogs appeared to be distinctly triphasic (13), prompted us to take another look at the energy dependence of Cbl uptake. The results obtained with arsenate were of particular interest in that it modified Cbl uptake in wild type and btuC cells differently. Arsenate had little, if any, effect upon the rate of Cbl uptake in either cell strain but altered the final plateau level of Cbl taken up. With wild type cells, the plateau value was usually lowered by the inclusion of arsenate in the reaction mixture, although this is not shown well in Fig. 4A since the unlabeled CN-Cbl was added before the plateaus were reached. In contrast, arsenate caused a substantial increase in the plateau levels of Cbl taken up by btuC cells (Fig. 4 B ) . The magnitude of this effect was variable but was rarely less than 3 times, and was w'  6 (rzght). Effects of arsenate upon CNC3H]Cbl uptake in E. coli strains KBTOOl and KBT103. The cells were grown in the normal high phosphate minimal medium, but after harvest, were washed with and resuspended in minimal Hepes, pH 6.6. The experimental procedures were essentially the same as those described in the legend to Fig. 1, except that the reaction mixtures contained 40 mM lithium D-(-)lactate as energy source and minimal Hepes, pH 6.6, as buffer. Potassium arsenate, when present, was added prior to the preincubation at a fmal concentration of 10 mM. At 60 min, unlabeled CN-Cbl, to a final concentration of 2 p~, was added to each reaction mixture (arrows). Results  sometimes as great as 12 times, the level of Cbl uptake in the absence of arsenate. The labeled CN-Cbl taken up by both wild type and btuC cells in the presence of arsenate was readily released following the addition of a large excess of unlabeled CN-Cbl to the reaction mixtures. This indication that the Cbl taken up by both strains in the presence of arsenate accumulated in the periplasmic space was contirmed by showing that most of the Cbl taken up under these conditions was released by an osmotic shock (Fig. 5). Similarly, the Cbl accumulated by arsenate-treated wild type and btuC cells was released by added EDTA (results not shown) with essentially the same kinetics as shown for btuC cells without arsenate in Fig. 3B. Since the Cbl taken up by arsenatetreated btuC cells still accumulates in the periplasmic space, it is evident that the effect of arsenate has not been to correct the btuC lesion. On the other hand, the periplasmic Cbl accumulation in arsenate-treated wild type cells suggests that arsenate has generated a btuC-like defect in the normal Cbl transport process. A further observation that must be considered in any explanation of the arsenate effect is that its addition to btuC cells during the course of Cbl uptake results in an immediate stimulation (Fig. 4B).
During the course of this work, we repeatedly gained the impression that arsenate reduced the level of Cbl uptake in wild type cells to about the same level to which Cbl uptake in btuC cells was stimulated. Fig. 6 shows the results of such an experiment, in which both cell types had been grown under as close to identical conditions as possible. In the presence of arsenate, the two uptake plots were nearly superimposable. We are inclined to believe that this is more than coincidence and is consistent with our view that arsenate has two distinct effects upon Cbl transport; an inhibition at, or after, the BtuC-dependent step, and a stimulation prior to BtuC involvement.
If arsenate were acting as a phosphate analog, its effects should be either inhibited or mimicked by added phosphate. It should be noted that the arsenate experiments were done with cells that were grown in the low phosphate medium and that the reaction mixtures for Cbl uptake were without added phosphate. As shown in Fig. 7B, phosphate acted as an antagonist of arsenate-stimulated Cbl uptake in btuC cells. In contrast, phosphate seemed to potentiate the arsenate inhibition of Cbl uptake in wild type cells (Fig. 7 A ) . One possible explanation of these results, in terms of the two possible effects of arsenate, is that phosphate is capable of antagonizing the arsenate stimulation of Cbl transport prior to the btuC step but has no effect on the arsenate inhibition of the later step. Phosphate alone had no discernible effect upon Cbl transport by either cell strain.
Some bacteria can reduce arsenate to arsenite (14). Accordingly, we examined the effects of arsenite on Cbl uptake in btuC cells. Fig. 8 shows that arsenite gave some stimulation, which was seen consistently, but which was usually much less than that given by arsenate. The concentration of arsenite (-0.7 mM) necessary to give half-maximal stimulation was appreciably greater that that of arsenate (-50 p~) .
By means of paper chromatography to separate arsenate and arsenite, we have found that the E. coli cells did not convert arsenate into arsenite, or vice versa, and that our supplies of arsenate and arsenite were not contaminated to a detectable extent by each other. The stimulation by arsenite prompted us to examine the effects of other reagents that might be expected to react with disulfide or sulfhydryl groups. Some of these results are also shown in Fig. 8. Diamide and N-ethylmaleimide were stimulatory but were much less effective than arsenate. Unlike arsenate, when these reagents were added during Cbl transport, there was usually a 10-to 20-min lag before stimulation was observed. Although these reagents inhibited Cbl uptake in wild type cells, with the possible exception of arsenite, they did not cause appreciable periplasmic accumulation of Cbl. One possible explanation for the increased plateau levels of Cbl taken up in arsenate-treated btuC cells was expansion of the periplasmic space. We have not measured the size of this cellular compartment directly, but have found that the inclusion of 5 and 10% sucrose in the reaction mixtures resulted in about a 2-fold increase in Cbl uptake by btuC cells. It therefore seems probable that a change in the size of the periplasmic space could exert an effect upon the uptake of Cbl in cells of this strain. It should be noted, however, that Stock et al. (15) found that arsenate had no appreciable effect upon the size of the periplasmic space in cells of either E. coli or Salmonella typhimurium. Accordingly, we believe it is unlikely that arsenate could induce changes in the size of the periplasmic space rapidly enough, and of sufficient magnitude, to account for the effects on Cbl uptake in btuC cell shown in Fig. 4. The smaller, more slowly developing stimulations of Cbl uptake by the disulfide and sulfhydryl reagents are perhaps more likely to be results of increases in the periplasmic space. Arsenate had no detectable effect upon Cbl uptake in tong or btuA cells or in cells of any strain incubated in the presence of 5 mM 2,4-dinitrophenol.

DISCUSSION
We have known for some time that the transfer of Cbl from the outer membrane receptors into the interior of E. coli cells required functional gene products from TonB and BtuC as well as the proton motive force and a domain on the receptor itself that was separate from the Cbl-binding site (3-6). This second functional domain on the receptor is defective in btuA mutants and we shall refer to it as the B t d domain. The evidence presented in this paper that almost all of the Cbl taken up by btuC cells accumulated in the periplasmic space has allowed us to distinguish those functions that are necessary for Cbl movement across the outer membrane from those that participate in transport across the inner membrane. The BtuC gene product is evidently necessary for Cbl transport across the inner membrane. The proton motive force, TonB gene product, and BtuA domain on the receptor clearly all participate after Cbl binding to the receptor and before the participation of the BtuC gene product. Earlier, we showed that the rate constant for dissociation of the receptor-Cbl X represents a hypothetical messenger that is generated in the penplasmic space by a TonB and proton motive forcedependent reaction and which promotes the dissociation of the receptor. Cbl complex. complex in outer membrane particles was appreciably slower than the observed rates of sustained Cbl transport in intact cells, and concluded that during normal Cbl transport, the rate of dissociation of the Cbl-receptor complex is increased by a proton motive force-dependent interaction within the cell envelope (13). On the basis of our present results, we would like to extend this conclusion and propose that the interaction that catalyzes the release of Cbl from the receptor requires the participation of the TonB gene product, the BtuA domain on the receptor, and the proton motive force. The nature of this interaction is at present unknown, but the scheme shown in Fig. 9 and outlined below represents our current hypothesis concerning the mechanism of Cbl transport in E. coli.
Cobalamin is too big to pass readily through the porin pores in the outer membrane and utilizes its own receptor protein (the BtuB gene product). Previous studies with the purified solubilized receptor protein indicated that lipopolysaccharide was required for fully active binding of Cbl (16). It is also possible that the receptor has some preferred interactions with other outer membrane components, such as phospholipid or specific proteins. We assume that the receptor protein most likely spans the membrane and, by analogy with some other outer membrane systems, may well be oligomeric. There is some experimental evidence that upon binding Cbl, the receptor undergoes a conformational change such that the Cblbinding site is no longer freely exposed to the exterior of the cell (13). Central to our proposal shown in Fig. 9 is the suggestion that the proton motive force and the TonB protein in the inner membrane interact to generate a diffusible periplasmic messenger ( X ) which then reacts with the receptor at a site within the BtuA domain and catalyzes the release of Cbl into the periplasmic space. We propose that the interaction with the receptor results in the loss of X from the periplasmic space, possibly either by a covalent change or by an antiport exchange with Cbl across the outer membrane.
Similarly, although the scheme shows the TonB protein acting as a transport protein that catalyzes a proton motive forcedependent efflux of X into the periplasmic space, we are also considering the possibility that the TonB protein generates X by means of a covalent change in some molecule already present in the periplasmic space. One possiblity is that, in the periplasmic space, X undergoes a cyclic oxidation-reduction in which the reduced form is generated by a combination of the proton motive force and the TonB protein and is subsequently reoxidized during its interaction with the Cbl receptor. By either mechanism, the net result of these interactions would be the proton motive force-dependent accumulation of Cbl within the periplasmic space. There is evidence from other laboratories that the TonB gene product is an inner membrane protein (17) and is involved in coupling the proton motive force to some other functions of the outer membrane (18).
The suggestion (Fig. 9) that Cbl is released from the receptor directly into the periplasmic space is a reasonable extrapolation of the evidence that Cbl accumulates in the periplasmic space in btuC cells and in arsenate-treated wild type cells, but a direct transfer of Cbl from the receptor to an inner membrane carrier during normal Cbl transport in wild type cells cannot be ruled out at present. Although the BtuC gene product is evidently involved in Cbl transport across the inner membrane, we have no direct evidence that it serves as the Cbl carrier. The accumulation of Cbl in the periplasmic space of btuC cells provides an explanation for Kadner's observation (19) that high concentrations of exogenous Cbl gave less repression of receptor synthesis in btuC cells than in wild type cells.
Both Taylor et al. (20) and Di Girolamo et al. (5) have previously noted the presence of small amounts of a Cblbinding protein in the periplasmic space of E. coli cells, and the scheme shown in Fig. 9 includes a role for this protein in Cbl transport. Although there is a considerable amount of circumstantial evidence in support of such a role, including the sensitivity of the system towards osmotic shock and the close similarity between the corrinoid specifcities of the periplasmic protein and Cbl transport (13), there is no conclusive direct evidence. Hong et al. (21) have recently suggested that acetyl phosphate is the energy source for transport systems that are dependent upon periplasmic binding proteins. This suggestion is the basis for indicating a possible involvement of acetyl phosphate in Cbl transport and is consistent with the arsenate inhibition of Cbl transport across the inner membrane. However, our model does not provide an explanation for the stimulation by arsenate of Cbl transport across the outer membrane shown in btuC cells. The most plausible suggestion we can make at present is that arsenate increases the concentration of the diffusible messenger X in the periplasmic space and that this effect is quite distinct from the arsenate effect on Cbl transport across the inner membrane. It should be evident from the above discussion that the mechanism of Cbl transport represented in Fig. 9 is mainly hypothesis and, although it is consistent with the available experimental data, there is little or no direct evidence for some of the proposed steps. Nevertheless, it is proving useful to us in planning further experimentation. E. coli cells that contain the lesions btuC and metE are able to grow on media containing CN-Cbl without methionine. If BtuC is necessary for Cbl transport across the inner membrane, how does enough Cbl for methionine synthesis get into the interior of the cell when the BtuC gene product is absent? MetE cells grown in the absence of methionine need a minimum Cbl concentration in the cytoplasm of about 30 nM for growth (5). As shown above, btuC cells can accumulate 3 to 5 pmol of CN-Cbl/109 cells in the periplasmic space, which would give a periplasmic CN-Cbl concentration of about 15 to 25 PM, close to about lo00 times the minimum concentration required in the cytoplasm. We suggest that Cbl may be sufficiently soluble in phospholipid that, in the presence of such a large periplasmic Cbl concentration, the rate of Cbl