Proc. Nat. Acad. Sci. USAVol. 72, No. 1, pp. 23-27, January 1975

A Cycle of Deprotonation and Reprotonation Energizing Amino-Acid Transport?(diamino acids/H+ cotransport/proton gradients/plasma membrane/Ehrlich cell)

HALVOR N. CHRISTENSEN AND MARY E. HANDLOGTEN

Department of Biological Chemistry, University of Michigan, Ann Arbor, Mich. 48104

Communicated by J. L. Oncley, September 23, 1974

ABSTRACT Although lowering the pK2 of neutralamino acids only weakens their concentrative uptake byEhrlich cells, the same change greatly enhances uptake ofdiamino acids. This effect does not arise merely fromputting the distal amino group in its uncharged form, butdepends on an enhanced deprotonation of the a-aminogroup. Parallel effects are seen for the transport systemfor basic amino acids, for which the assignment of pKvalues within the membrane is less ambiguous. To explainthe paradoxical advantages of having the a-amino groupprotonated yet readily deprotonated, we propose that aproton withdrawn from that group is pumped over anintramembrane interval to energize amino-acid transport.

Although a driving of amino-acid uptake by cotransport withNa+ has long been known (1-4), this has recently been shownnot to be the obligatory mode of energization (5). Further-more, Na+-independent systems can unambiguously produceuphill transport (5). Therefore, we have long given attentionto suggestive effects of modifying the H+-dissociation of theamino acids as clues to a more fundamental mode of energiza-tion (ref. 5; ref. 6, pp. 91-92; refs. 7-9).

Table 1 illustrates that a systematic lowering of pK2' ofneutral amino acids (in this case by introducing one, two, orthree fluorine atoms) decreases both the initial rate of uptakeand the steady-state distribution attained with Ehrlichascites tumor cells. Fig. 1 shows a corresponding effect of thepresence of two fluorine atoms in the amino-acid molecule:the pH optimum for uptake is lowered sharply. These resultsindicate that the amino acid is accepted for uptake by thesystem in question with the amino group in the form RNH3+rather than RNH2.

This result did not prove, however, that the recognitionsite at the interior surface of the membrane has the samepreference. In 1958 we considered, using Fig. 2, whether adifference in the state of protonation preferred at the twosurfaces might produce co- or countertransport of H+ withthe amino acid (3). The figure shows a hydrogen ion leftbehind when the amino acid combines with the externalreceptor site, and a new hydrogen ion taken up by the aminoacid on entering the cytoplasm. This scheme produces acountertransport between H+ and the amino acid, which wasin 1958 an arbitrary choice over cotransport. Possibilities ofthis kind have subsequently become more important throughthe development by Mitchell of his chemiosmotic hypothesisfor the intermediation of hydrogen ion gradients in the processof energy transduction by the mitochondrial and chloroplastmembrane (11, 12). The observation of cotransport of H+with amino acids (see for example refs. 8 and 13) and withsugars (14-16) has accelerated the extension of these ideas tothe plasma membrane (see ref. 17).

Uptake of the lower homologs of lysine and ornithine

By chance we encountered in 1952 behavior revealing thatthe data of Table 1 and Fig. 1 do not tell the whole story of therole of dissociation in amino-acid transport (18). The struc-tural feature responsible for this behavior was a nearness of asecond amino group to the a-amino group. In the homologousseries of a,-diamino acids composed of lysine, ornithine,2,4-diaminobutyric acid (A2bu), and 2,3-diaminopropionicacid, we observed an abrupt increase in the rate of uptakeand intensity of accumulation as the chain is shortened tofour carbons or less (Fig. 3). The most intensely accumulatedis A2bu, both for Ehrlich ascites tumor cells and for rat liver(18). Almost all of the endogenous levels of K+ and Na+could be displaced from the Ehrlich cell by this organic cation,its uptake along with Cl- having in the meantime caused thecells to swell to two to four times their normal volume duringseveral hours. Gradients as high as 180 mAM could be obtainedwhen uptake occurred from hypertonic solution, thus mini-mizing simultaneous water transfer.The principal difference among these diamino acids lies not

in the rates at which they are accumulated by the cationicamino-acid system, for which their Km values are relativelylow, but in their reactivities at high Km values withsystems for neutral amino acids. When we contrasted theconcentration dependence of the uptake of lysine and A2bu byEhrlich cells, we saw similar hyperbolic portions for the curvesat low levels, representing uptake by the cationic system,inhibitable by homoarginine. At higher levels we noted acomponent that was difficult to saturate, much larger forA2bu than for lysine. An inhibition analysis showed that thiscomponent for A2bu and for diaminopropionate occurs bySystem A and is Na+-dependent (19). This was a surprisingresult, because that system ordinarily has neutral aminoacids for its substrates (20).We begin in this way to encounter a series of paradoxes.

Even though A2bu, pK2 = 8.4, should be cationic to theextent of 91% at pH 7.4, most of its uptake at high levelsoccurs by a neutral amino-acid system. Of the 9% that ispresent in dipolar form in water solution at that pH, a majorpart takes the form of the a,-y-dipolar ion, by analogy not asubstrate for a transport system for a-amino acids. Thevalues for pK2' for the homologous series of Fig. 3 in order ofdecreasing chain length, are 9.5, 8.7, 8.4, and 6.8. The initialrates of uptake increase as pK2' decreases, just the reverseof the order shown in Table 1 for monoamnino acids in the sametransport system.The latter paradox becomes not quite so implausible when

one considers that for transport these diamino acids need tolose a proton to convert their cationic form to a dipolar species.One might expect that A2bu could become as good a substrate

23

Abbreviation: A2bu, L-2,4-diaminobutyric acid.

24 Biochemistry: Christensen and Handlogten

? 1.6

x

D 1.2

co

75-0 0.800)

a 0.4

EE 0~:: 6

0.2

0.1

7 8

PHex

FIG. 1. Effect of the presence of two fluorine atoms in 2-aminoisobutyric acid on the pH optimum for the rate of uptakeby Ehrlich ascites tumor cells. Uptake observed during 2 min at

370 from Krebs-Ringer phosphate solution, adjusted to give thefinal pH shown. Presumably, the increasing quantity of thisamino acid in its dipolar form compensates in part for the de-crease in rate that would otherwise occur below pH 7.4. *, 0.2mM difluoroaminoisobutyric acid; 0, 0.2 mM aminoisobutyricacid.

for System A as homoserine, once it has lost a proton, givenonly that this proton be lost from the distal amino group.

But now the paradox reappears: A2bu is far more stronglyaccumulated by the Ehrlich cell than is homoserine (Fig. 3).Why should these diamino acids act as "super-substrates,"given their special difficulty in assuming the preferred struc-ture for uptake? Even though A2bu can assume this structuremore readily than ornithine, which can assume it more readilythan lysine, all of these amino acids should be at a handicapin comparison with homoserine or other ordinary neutralamino acids. Clearly, the ability of the amino acids to assume

statically their a,a-dipolar form by no means provides a fullexplanation of the effects of changes in the pK values.

Let us examine, then, the possibilities introduced along witha second amino group. We may propose, first of all, that thereceptor site for entry provides a microenvironment for theside chain such that the distal amino group of the diaminoacid is stabilized in its deprotonated form, given that itstendency to retain the proton is not too great. The largegradients of A2bu that can be generated correspond to a largeasymmetry between influx and efflux, retained even at highloads. We may seek provisionally to account for this phe-nomenon by asserting that the behavior shown by the recep-

tor site for entry, whereby the acceptable species of the aminoacid presumably is stabilized, is not shared by the internalreceptor site serving for exodus; or alternatively, that a

R-CR'-COO-

R"-NH2+

Outside

R-CR'-C=O+ X+ + H+

R"-N 0

x

Inside IR-CR'-COO H 0 R-CR'- C=O

OH- + + X+ 2R"-N

x

FIG. 2. Scheme to show the amino acid accepted in its de-protonated form for entry into the membrane; and subsequentlytaking up a hydrogen ion from the cytoplasm on its release intothe cell. Reproduced with permission from Riggs, T. R., Walker,L. M. & Christensen, H. N. (1958) J. Biol. Chem. 233, 1479-

1484.

TABLE 1. Effect of 3-fluoro substitution on transport of2-aminoisobutyric acid into the Ehrlich cell

Distribution ratio

pK'/ for Ehrlich cell

Substrate (250) at 1 min at 30 min

Aminoisobutyric acid, 1 mM 10.21 0.85 27.7Fluoroaminoisobutyric acid,

1 mM 8.58 0.95 13.0Difluoroaminoisobutyric acid,3mM 7.40 0.17 3.6

Trifluoroaminoisobutyric acid,1 mM 5.94 0.27 0.74

Difluoroaminoisobutyric acid is the symmetrical difluoroderivative, prepared from 1,5-difluoro-3-pentanone by the sameprocedure used for the monofluoro derivative. At 30 min, itsdistribution ratio was at the maximum value. In trifluoroiso-butyric acid, the fluorine atoms are all on the same carbon atom.The transport disadvantage of the fluoro analogs applies also torenal resorption and hepatic uptake (10).

stabilization of a species unsuitable for transport by theinner receptor site is not shared by the outer receptor site.To show that the two opposed transport receptor sites

place the substrate in dissimilar electrostatic environmentswould not provide a sufficient explanation of the flux asym-metry, however, unless we add an energy input to maintain theenvironmental difference, whatever it is. Under the simplemodel implied here, energy would have to be applied steadilyto keep changing the environment of the site for entry into themembrane to an environment suitable for release into thecytoplasm.We early began to attribute the contrast in transport

among the diamino acids to the inductive interactions be-tween the two amino groups. In lysine, the separation betweenthem is great enough so that the presence of each aminogroup has little effect on the pK of the other, so that the ob-served pK' values lie close to the intrinsic pK values, namely,at 9.53 for the a- and at 10.94 for the C-amino group (21).As the two groups are brought nearer together, this interac-tion is increasingly transmitted through the carbon chain.The result is that both pK2 and pK3 are lowered, and thedisparity in their internal environments in the moleculedecreases somewhat. It becomes easier then for the order oftheir dissociations to be reversed, given only a modest changein the external environment. The lowering of the values ofpK2 and pK3 has thus made the dissociation of the aminogroup particularly sensitive to its immediate environment.Numerous cases are known where the sequence of dissocia-tion of a substance can be inverted by the addition of a mis-cible organic solvent, e.g., dioxane, to an aqueous solution.As a result of these interactions, whether we focus our atten-tion on one or the other of the two amino groups of A2bu,that group may be seen as environmentally sensitive to aheightened degree. The question may then be asked, is thetransport system particularly responsive to the environmentalsensitivity of the a- or of the distal amino group, or of both?

Other diamino acids showing strong flux asymmetry

Before facing this question, let us note that the phenomenonis not restricted to the transport system (A) implicated up to

(scaleatright

- (scale at left)

Proc. Nat. Acad. Sci. USA 72 (1975)

De- and Re-protonation Cycle Energizing Transport 25

this point. By placing two methyl groups on the distal aminogroup of diaminopropionic acid, the branched-chain aminoacid, azaleucine, pK = 6.8, is obtained. This diamino acidalso proved to undergo transport largely as a neutral aminoacid, the reactivity decreasing as the pH is lowered, aroundan apparent midpoint of pH 5.8 even though pK2 in freesolution at 250 is 6.8 (22). As a predictable consequence of thebranching (19), this transport occurs largely by Na+-indepen-dent System L, for which the amino acid shows a pronouncedflux asymmetry.The replacement of a methylene group in the side chain of

lysine with a sulfur atom to yield thialysine, pK2 = 8.4, alsoleads to a "gradient-sensing" transport substrate (5, 9). Inthis case uptake is especially fast both by the Na+-dependentand the Na+-independent routes. Other diamino acidsstrongly accumulated by Na+-independent System L are4-amino-4-carboxyl-1-methylpiperidine, pK2 = 7.2, and cis-1,4-diaminocyclohexanet arboxylic acid, pK2 = 8.7. Theordinary substrates of System L are concentrated only ratherweakly. Hence, these analogs serv3 to show that the presenceof a second amino group with a low pK2 increases trapping ofthe energy accessible to the Na +-independent system.The observation that the trans isomer of the last-named

amino acid is much more weakly accumulated than the cisisomer (8) brought attention back to an idea entertained forsome years. In all the diamino acids showing the phenomenondiscussed so far, the two amino groups in each molecule cancome together, for example, to share a proton (as in Fig. 14 ofref. 19). This behavior would not be expected in water solution(23), but it could be facilitated in some membrane environ-ment.We have recently found, however, that 2,6-diamino4-

hexynoic acid (24), pK' = 8.4, can also generate rather highgradients across the plasma membrane of the Ehrlich cell,principally by System A. In this compound the two aminogroups can scarcely approach each other. Nevertheless, di-aminohexynoic acid in the DL form generates gradients as highas 100 to one, taking into account near failure of the D isomer tobe taken up (Fig. 4). Rather strong gradients are also generatedfor trans-4,5-dehydrolysine. These results appear to mean thatthe amino groups do not need to come together. This conclu-sion probably applies to both Systems A and L, since bothsystems participate substantially in the strong accumulationof diaminohexynoate. If one of these systems were only weaklyconcentrative, exodus by it would deplete the gradientgenerated by the other. Apparently, therefore, we need tolook for another explanation in the topography provided bySystem L for the discrimination between the cis and transisomers of diaminocyclohexanecarboxylic acid. The highreactivity and high stereospecificity for dehydrolysine andsimilar diamino acids also point to a preference as to the posi-tion taken at the receptor site by the second amino group.

Does more specific facilitation of deprotonation of thedistal amino group enhance uptake?

If the neutral transport systems merely need to direct thedeprotonation inherent in a low pK2 toward the distal aminogroup, then we should be able to assist the process by intro-ducing structural features more specifically enhancing thatdissociation. Note that the structural changes selected so farfor lowering pK2 have all acted on both amino groups. That is,

40 A

.co

20

20-

30 60 90 120

Minutes

FIG. 3. Comparison of time course of uptake of the homol-ogous diamino acids (1 mM) in the presence of 10 mM\I homo-arginine. Contrast between uptake of A2bu and homoserine (each10 mM) in media in which Li+ replaces Na+. A2pr, 2,3-diamino-propionic acid. The four longer curves show the abrupt intensi-fication of uptake of the diamino acids by neutral systems (notinhibited by homoarginine) when the chain length falls below 5carbon atoms. The shorter curves contrast accumulation of A2buand homoserine, both somewhat handicapped by substitutionof Li+ for Na+ to restrict homoserine uptake to System A (seetext).

chain has mutual effects on the probability of the dissociationof each. Similarly, the sulfur atom in thialysine lies midwaybetween the amino groups, and the unsaturated links intro-duced into lysine will affect both dissociations. Therefore, wetested next the effects of structural features designed specifi-cally to favor the deprotonation of the distal amino group. Forexample, the pK2 = 6.0 of histidine is well known to pertainlargely to the imidazole group. This amino acid, at 3 mM, wasconcentrated only 1.45-fold from Na+-free medium containingexcess homoarginine to block the cationic system. Canavanineand canaline, pK2 = 7.4 and 4.3, respectively, were concen-trated only 3- and 6-fold, respectively, via System A, andm-aminophenylglycine, pK2 = 3.6, was not concentrated atall from a Na+-free medium. These tests need extension todiamino acids for which pK2 applies mainly to the distal aminogroup and yet falls in the range 7-8.4. Several of the stronglyaccumulated diamino acids show through chemical shifts byproton magnetic resonance that pK2 in D20 solution concernsthe alpha more than the distal amino group; but a small pref-erence in the other direction may well serve also. The resultsso far offer no encouragement, however, for the idea that whatis needed for strong uptake is merely to get the side chain into

4) 20

S 158

1< 5 1

0

EE

10

05

0.2 mM

o 0 30 60 90C

E Minutes

FIG. 4. Strong accumulation of 2,6-diamino-4-hexynoic acidby Ehrlich cells. Partition coefficients as high as 100 could becalculated for the L isomer. At 10 m-M, a gradient of 84 mMI wasobtained at 120 min. Of the uptake from 3 mMI solution, 55%

bringing the two amino groups closer together on the carbon

Proc. Nat. Acad. Sci. USA 72 (1975)

was blocked by 2-(methylamino)-isobutyric acid in excess.

26 Biochemistry: Christensen and Handlogten

1 100

0

0 8

-S 6

cc0 2.E

U 2 4 6 8[Amino acid],,, mM

FIG. 5. Comparison of the extent of accumulation by Ehrlichcells in 3 hr of model substrates of the cationic amino-acid systemLy +, L-homoarginine (0) and e-N-methylhomoarginine (0),with that of 4-amino-1-guanylpiperidine-4-carboxylic acid (A)and ,3-guanidino-L-alanine (star). Uptake was at 370 from Krebs-Ringer bicarbonate medium at pH 7.4. Lysine and arginine ac-cumulation resembles that of homoarginine.

an uncharged form. On the contrary, they draw attention backto the probability that the a-amino group needs to participatein a deprotonation for optimal energy trapping.

Evidence from the cationic amino-acid transport system

So far we have dealt with the paradoxical transport of diaminoacids by neutral systems. Let us now consider transport bythe cationic system. If we make the pK2 of the distal aminogroup high enough, as in lysine, we can minimize transport byneutral systems and maximize that by System Ly+, the basicsystem. Better yet, by using a guanidinium group for the distalbasic structure, the selectivity for System Ly+ can be madeunequivocal and nearly complete.

In preparing 4-amino-1-guanylpiperidine-4-carboxylic acidas a model, metabolism-resisting substrate for System Ly+(25), we unintentionally obtained a low value for pK2 of 8.0,

and high degrees of accumulation (8). Arginine, pK2' = 9.04,and homoarginine are accumulated by only 2- or 3-fold, exceptat relatively low substrate levels (Fig. 5). Their uptake may infact be entirely passive, since the transmembrane potentialgradient is sufficient to account for a 2-fold accumulation. Theanalogous 1-amino-4-guanidine-cyclohexanecarboxylic acid,pK2 = 9.3, also showed only weak accumulation, whereasg-guanidinium-L-alanine, pK2 = 7.8, was concentrated muchmore strongly under similar conditions. Incidentally, thesecompounds, along with others already discussed, showed nomeasurable binding to undialyzable components of an un-diluted cytolysate prepared by freezing and thawing Ehrlichcells.These results show that the association between low values

for pK2 and high degrees of accumulation extends also to thistransport system. What is especially significant is that it ishighly unlikely that pK2 for these compounds applies to theguanidinium group, to which we assign instead the pK3higher than 12 noted on titrating. Even in special environ-ments of the membrane, pK2 probably applies almost exclu-sively to the a-amino group. Attention is thus directed evenmore emphatically to the indications that a ready deprotona-tion of the a-amino group gives a special energy-trappingability to amino acids for transport, provided that a secondbasic group is present on the side chain. Note also that theordinary substrates of the neutral transport systems carry noamino group on their side chains. Hence, the only amino groupon those molecules whose traffic in protons can figure in theenergization of transport is the a-amino group.

DiscussionTable 1 and Fig. 1 indicate that a proton should be present onthe a-amino group if a neutral amino acid is to be accepted fortransport. Also for the cationic system, it is clear that the

7NH+3R-CH

Coo

externalsolution

r

H

H

I NH

< =: R-CH / 3I N~~coo~

Barrier#1

mediatedpassage ofzwitterion

Zone of Hpumping;

mediated passagemay favor

RCH (NH2) COOwith H+ impelled

separately

Barrier#1

mediatedpassage ofzwitterion

FIG. 6. Scheme to show one form of our proposal as to how pumping of a separated proton may produce a structure-dependent

energization of amino-acid transport of neutral or cationic amino acids. The amino acid is shown being accepted from the externalsolution in an a, a-zwitterionic form, net charge zero, or if we are discussing cationic System Ly +, net charge plus one. The mediate-dpassage of the amino acid through the membrane brings it to a zone of internal H + pumping (represented as barrier 2) across whichpassage of the amino-acid molecule is largely limited to the form with the a-proton absent. This zone could lie within an oligomeric pro-

tein assembly. The gradient attributed here to a proton pump could instead be a gradient of polarization or of electron distributionsufficient to cause the a-amino group to yield a proton over this spatial interval. Facilitation of the reversible removal of the proton by

side chain structures plays a role in determining the flux asymmetry.

0 o

RCH ,-yopls

cytoplasm

Proc. Nat. Acad. Sci. USA 72 (1975)

De- and Re-protonation Cycle Energizing Transport 27

a-amino group needs to be protonated; the initial rates oftransport of the two arginine analogs of low pK2 showed nodecrease as the pH was lowered from 8 to 5.5, whether forentry or exodus. Nevertheless, we have considered evidencehere that the readiness of departure of the proton from thatgroup (perhaps to another position on the same molecule)gives transport an intense directional asymmetry. The situa-tion seems to say that the uphill process finds advantages bothin the protonated and in the deprotonated state of the a-amino group. Since the amino-acid molecule can scarcelypresent these two states simultaneously, our interpretation(Fig. 6) is that it probably does so sequentially. First theamino acid must be accepted in the proton-bearing form, inadvance of the rate-limiting event, to yield the kineticsreflected by Table 1 and Fig. 1. Ultimately, the amino-acidmolecule must again be released into the cytoplasm carryingthe same proton load, since we can detect so far no cotransportof H+ with these diamino acids across the plasma membraneof the Ehrlich cells. Furthermore, this cell shows only smalltransmembrane gradients of H + or the electrical potential (4).[It is an important additional circumstance that another classof unusual substrates for System L, illustrated by a,a-diethylglycine, does cause H+ uptake during its unusuallyconcentrative entry into the cell, and that lowering the ex-ternal pH stimulates the accumulation of ordinary substratesof System L (5, 8, 9).] Somewhere between the binding andrelease events, we suggest that a proton is reversibly disso-ciated from the a-amino group, and subjected to protonpumping (Fig. 6). We suggest further that this process repre-sents the fundamental mode of coupling of the driving forcein the membrane to uphill amino-acid transport.This idea gives the selected diaminio acids importance not

merely as probes that may sense and report differences inhydrogen ion availability at separated points along the trans-port pathway through the plasma membrane, as suggestedearlier (5). The concept implies, in addition, that the structureof these diamino acids intensifies the availability of H+ fromthe amino acid itself, thus pointing to the way in which thedirectional thrust may be given to the amino-acid moleculewithin the membrane.The observed similarity in the responses of Systems A and L

to a lowering of pK2 indicates that both of these systems, aswell as System Ly+, probably participate in an intramem-brane linkage of amino acid and H+ movements. Although, incontrast to System L, we have seen no evidence for linkedmovements of H+ and amino acids in System A, neither havewe sufficient evidence for any other way in which the thrust isuniversally applied to amino acid flow through that system.

We do not suggest that this mode of coupling between protonflows and the flows of organic metabolites is universal sinceno chemical basis is obvious for a parallel mode of coupling offlows between simple sugars and the hydrogen ion.

This research received support from Grant HD01233, NationalInstitute for Child Health and Development, U.S. Public HealthService.

1. Christensen, H. N., Riggs, T. R. & Ray, N. E. (1952) J.Biol. Chem. 194, 41-51.

2. Christensen, H. N. & Riggs, T. R. (1952) J. Bidl. Chem.194, 57-68.

3. Riggs, T., R., Walker, L. M. & Christensen, H. N. (1958)J. Biol. Chem. 233, 1479-1484.

4. Heinz, E. (Od.) (1972) Na-linked Transport of Organic Solutes(Springer, Berlin), pp. 1-98.

5. Christensen, H. N., de Cespedes, C., Handlogten, A. E. &Ronquist, G. (1974) Ann. N.Y. Acad. Sci. 227, 355-379.

6. Christensen, H. N. (1955) in Amino Acid Metabolism, eds.McElroy, W. D., & Glass, B. (Johns Hopkins Press, Balti-more, Md.), pp. 63-106.

7. Christensen, H. N. (1974) Biological Transport (W. A.Benjamin, New York), 2nd ed., pp. 176-196.

8. Christensen, H. N. (1973) J. Bioenerg. 4, 31-61.9. Christensen, H. N., de Cespedes, C., Handlogten, M. E. &

Ronquist, G. (1973) "Reviews on biomembranes," Biochim.Biophys Acta 300, 487-522.

10. Christensen, H. N. & Oxender, D. L. (1963) Biochim.Biophys. Acta 74, 386-391.

11. Mitchell, P. (1961) Nature 191, 144-148.12. Mitchell, P. (1970) in Membrane and Ion Transport, ed.

Bittar, E. E. (Wiley-Interscience, London), Vol. 1, pp. 192-256.

13. Eddy, A. A. & Nowacki, J. A. (1971) Biochem. J. 122, 701-711.

14. West, I. C. (1970) Biochem. Biophys. Res. Commun. 41, 655-661.

15. West, I. C. .& Mitchell; P. (1972) J. Bioenerg. 3, 445-462.16. Kashket, K. R. & Wilson, T. H. (1972) Biochem. Biophys.

Res. Commun. 49, 615-620.17. Harold, F. M. (1972) Bacteriol. Rev. 36, 172-230.18. Christensen, H..N., Riggs, T. R., Fischer,; H. & Palatine,

I. M. (1952) J. Biol. Chem. 198, 1-15, 17-22.19. Christensen, H. N. & Liang, M. (1966) J. Biol. Chem. 241,

5542-5551.20. Oxender, D. L. & Christensen, H. N. (1963) J; Biol. Chem.

238, 3686-3699.21. Bradbury, J. N. & Brown, L. R. (1973) Eur. J. Biochem.

40, 565-576.22. Christensen, H. N. (1964) J. Biol. Chem. 239, 3584-3589.23. Hine, J., Via, F. H. & Jensen, J. H. (1971) J. Org. Chem.

36, 2626-2629.24. Davis, A. L., Rodney, L., Maul, S., Cook, D. E. & McCord,

T. J. (1964) Arch. Biochem. Biophys. 104, 238-240.25. Christensen, H. N. & Cullen, A. M. (1973) Biochim. Biophys.

Acta 298, 932-950.

Proc. Nat. Acad. Sci. USA 72 (1975)