Pseudoisoenzymes of Rabbit Muscle Phosphoglucose Isomerase*

Abstract Rabbit muscle phosphoglucose isomerase, either in crude muscle extract or after isolation and crystallization, has been found to occur as three different species when chromatographed on carboxymethyl Sephadex. The three forms are indistinguishable with respect to over-all amino acid composition except that they differ in accessibility of their sulfhydryl groups to p-mercuribenzoate. They can be interconverted both by treatment with dithiothreitol and by exposure to oxidative conditions, indicating that they represent different states of the same protein rather than genuinely dissimilar protein species as claimed in another report (Yoshida, A., and Carter, N. D. (1969) Biochim. Biophys. Acta, 194, 151). When subjected to isoelectric focusing, these "pseudoisoenzymes" yield characteristic isoelectric points which correspond to their elution positions on cation exchange columns.

It:M,it skeletal muscle l~l~osl~hoglr~coze isonierasc (I)-glucosc-(i-~)lios~)hnte ketol-isomeruse, ICC 5.8. 1 .I)) Ii:ls it1 i,ecent yeutx beeu the object of detailed studies in t.his l;ll)or>ttory (l-6). Work on the structure and mechanism of this enzyme 11ns in the list bee11 carried out with five times crystallized prep:trationn (7,8) which had been shown to be homogeneous by various criteria (6). The extent to which t.he data obtained represent meilllitlgful analytical information, however, w;ts 1,cndered doubtfnl w1w11 Yoshida and Carter (9) reported that they could resolve t.lirce isoeuzymes from II commercial l,rel):iratioli of oryst,:tlline mbbit muscle phosphogluoo~e isomerase. On tho basis ot' difcreut amino acid compositions aud different. l)eptide maps 01 ~tainecl for two of their chromatographic species, these :lllthors $  suggested (9) that their isomerase isoenzymcs n$$t h;\ve tlifferent primary structures aud thus originate frnm different geues in the rabbit. We will show in this paper that ~~l~o~l~l~oglucose isomerase from skeletal tnuscle of the rabbit, wheu c:llronlatx,gr:1I,l1etl uudel the conditions given by Yoshida and Cuter (9), does not resolve into the isoenzyme species described by these xuthors; that under conditions entirely different from those of Yoshida aud Carter, partially purified as well as cryat;rlliue I)llosl~llo~lr~cose isomerase can be separated iuto :L number of ('11r(~1tt:lfoFr:II)hiC forms the relative amouuts of which depentl OIL the o\.itl:&oli state of available sulfhydryl groups; that these cliroin;ilo~r:i~~~~ic enzyme species, in contrast to the isoenzyrnes tlcsoibctl bJ Yoshida and Carter (9), do not differ siguifkxntly iu trl.er-all amino acid composition and that they are, ill fxct, interctrnvertible; and that these multiple chlomato~l:~.~~llir species (~amlot, therefore, be isoenzymes, but rather represent, different htates of the same molecular species of rabbit muscle phosphoglucose isomernse.
Itssays-I'tlosphoglucose isoulernse activity was nle:lsllretl tit 30" I)y coupled spectrophotornetri(~ assay :is previously dcioribetl (7,8). I3uxyme units are defiued as micromoles of fructose Ol)hoslA&e converted to glucose O-1~llosl~h:~t.e per mill :lt 30'. I'rotein concentrations were determined by measuring the aI)sorbauce at 280 nm in 0.01 M sodium l~hosl~h:rte of 1111 '7.0; uudel these couditious, an absorbance of 1.0 for ;I l&mm light path coi~res~mrds to 0.76 mg of enzyme per ml (Gj. Erqme Isolation-Detailed descriptions of the isolatiou a11t1 c:rystullization of phosphoglucose isomerase from rabbit skeletal muscle have been provided earlier (7,8). In the present work, these procedures were followed as described when prel):uiug the crystalline enzyme; crystallizations were allowed to occur very slowly (6). When only 1)artially 1)urified enzyme was desired as stwrtiug material for further chrolnatof;i,aphic studies, the origiuxl method was followed through completion of Fraction III, i.e. the collection of the l)r'ecil)ittlte at 0.M :m~mouium sulfate saturation, which was then dialyzed x&nat the .*t,artiiig buffer of the subsequent c:llr(~rnt~to~ra~)llic column. Other changes of the standard purification procedure, necesxit:lted by the conditions required for a particular experiment, are described in the text or in the legends t,o the corresponding figures. The rabbits were obtained from several differeut local rabbit farms without selection or specification as to origin.  to 24 mar in Nn+, in a total volume of 1200 ml. At the end of the gradient the column was \vashed with 1 I\I N&l. These standard conditions2 were used for the column measuring 2.5 X 20 cm. For columns of different size, concentrations and volume,s of the gradient buffers lvere adjusted accordingly to obtain maximum resolution.
isoelectric Ic'ocusing-Isoelectric points of the different enzyme species were measured at 4' by isoelectric focusing in an LKB model 8101 electrofocusing column (capacity, 110 ml). Sucrose density gradients were prepared with an LKB model 8121 gradient mixer and pH gradients from pH 7 to 9 were established in the column with LKB Ampholine ampholyte solution kept for 96 to 120 hours at a constant voltage of 600 volts. At the end of the focusing period the column contents were removed at a flow rate of 30 ml per hour with the aid of a Harvard model 1201 peristaltic pump.
Fractions of approximately 1 ml were collected and their protein content, enzymatic activity, and pH were determined.
Amino Acid Analyses-Aliquots of the various chromatographically separated fractions were hydrolyzed for 20 hours in 6 N HCl and analyzed with the aid of a BeckmanSpinco automatic amino acid analyzer model 12OC as previously described in detail (6).
Determination of Xuljhydryl Groups-Sulfhydryl assays were performed at 30" in 0.05 M sodium phosphate, pH 7.1, by Boyer's p-mercuribenzoate method (10) with care not to exceed a ratio of 1.5 eq of PMB per titrable ---SK group (1). The increase in absorbance due to mercaptide formation was followed spectrophotometrically at 255 nm with use of a molar absorptivity COefficient of 5.62 x lo3 (1, 10). Matched sets of 2.0-ml cuvettes were used throughout. Initial enzyme concentrations ranged from 1.5 to 4.5 p&f, initial PMB concentrations ranged accordingly from 25 to 75 PA<, in a total volume of 2.0 ml. Contents of the blank cuvette were identical with those of the sample cuvette except that phosphate buffer was substituted for the enzyme solution.
The reaction was allowed to proceed for 40 mm; the absorbance difference was extrapolated back to zero time and this value was taken as representative of the immediately available -SH groups (Type I, cf. Reference 1). Then 0.20 ml of 5% sodium dodecyl sulfate (in the same phosphate buffer) was added to both blank and sample cuvette to obtain the absorbance change reflecting the tot.al sulfhydryl content (1). To obt;ain the number of -SH groups per molecule, before and after addition of sodium dodecyl sulfate, the measured absorbance values were corrected for self-absorption at 255 nm by both enzyme and PMB and for volume changes caused by addition of the denaturant.
Determination of Total fIal$cystine Content-The total halfcystine content was measured as cysteic acid after performic acid oxidation according to the procedure of Moore (11). The oxidized enzyme samples were hydrolyzed in 6 N HCl under the usual conditions and the cysteic acid was quantitated by chromatography on the automatic amino acid analyzer.

Earlier work in this laboratory
on the chromatographic behavior of rabbit muscle phosphoglucose isomerase on both DEAE-and cnrboxymethyl celluloses had given strong indication that multiple peaks of enzymatic activity obtained by these RUN No. 57 5. Separation of five times crystallized enzyme into Fractions G-IA, G-IB, and G-II by chromatography on CWSephadex in the absence of DTT.
Enzyme (105 mg, 938 units per mg), dialyzed against the gradient starting buffer, was applied to the CM-Sephadex column (2.5 X 20 cm) and eluted with 1300 ml of a sodium phosphate gradient, pH 6.90 (8 to 16 m31 in phosphate, 12 to 24 rnM in Na+). Symbols as for Fig. 1. methods vcere most likely artifacts and not real isoenzymes. 3 It had neither been possible to reproduce identical elution profiles nor had conditions been found at that. time under which a previously single peak would not elute as multiple species on rechromatography.
The project was therefore discontinued. However, diverging results have recently been published by Yoshida and Carter (9), who concluded from their data that phosphoglucose isomerase isolated from rabbit skeletal muscle can be resolved into what they feel are genuine inoenzymes with different primary structures.
These authors reported (9) that they could resolve a commercial preparation4 of crystalline rabbit muscle phosphoglucose isomerase into three active enzyme peaks (and one inactive protein peak) by chromatography on carboxymethyl Sephadex with elution of the isoenzymes by means of a NaCl gradient from 0 to 0.1 M in 0.01 JZ phosphate buffer of pH 6.8. Their elution profile is redrawn in the top portion of Fig. 1. They also state that both their major components ("Isomerase I" and "Isomerase II") on rechromatography under the same conditions eluted in the same positions as observed for the first chromatography (9). Numerous attempts in this laboratory to reproduce their separation pattern have failed with either two or five times  . 6. Chromatograph>-of partially purified rabbit muscle phosphoglucose isomerase on CM-Sephadex. A. partially purified enzyme (5.9 g) was prepared as described under "Experimental Procedure," dialyzed against gradient starting buffer, applied to a CM-Sephadex column (5 X 41 cm), and elutcd with 5000 ml of a linear sodium phosphate gradient, pTI 6.90 (8 to 24 mM in phosphate, 12 to 36 mu in Na+); B, rechromatography of the combined G-I fractions without DTT; C, rechromatography in the presence of 1 mM DTT of the combined G-I fractions kept in 10 rnxf DTT for G days; D, rechromatography in the presence of 1 rnM DTT of the G-II fraction (pooled as shown in A) kept in 1 ma< DTT for 12 days. Column dimensions for B, C, and D were 2.5 X 20 cm; elation was carried out with 1200 ml of a sodirlm phosphate gradient, pll 6.90 (8 to 16 mu in phosphate, 12 to 24 mM in Nat). Symbols as for Fig. 1 Examples of these attempts are Phosphate Gradient-Aft,er it leas found impossible to reproduce presented in the middle anal botlonz portion of Fig. 1, showing the cluomatographic patterns of Yoshida and Carter, :I syschromatography of twice crystallized enzyme and recluomatof;-tematic study of the elution behavior of rabbit muscle phosphorapliy of the peak with enzymatic activity, respectively, all runs glucose isomerase was begun in which vei'y shallow sodium being made esnctly under the conditions of column size, buffer, phosphate gradients of different pH were esplorctl. l'his led to pH, and NaCl glndient :I': described by Yoshidn and Carter (9). the conditions illustrated in Fig. 2, i.e. equilibration of the col-It is :iplxirent tht chromatography under these particular conumn, application of the protein samljle, and subsequent xrshing ditions failed to 1)roduce ali> si~~nific:rnt resolution of activit,y of the column with 6 rmv sodium l~l~ospl~nte (9 rnRI in iYa+), pH peaks.
6.90, followed by a sodium phosphate gradient from 6 mar (9 minx in Na+) to 12 mx (18 mx in Na+) of the same PIT. 'l'he system was found to be extremely sensitire to pI-1. \Vherens pH values of 7.0 or higlirr n-ould not allo\v the enzyme to l)e ilritially adsorbed, values of pII 6.8 or lower would yield progres-ively poorer resolution of the activity peaks. The upper portion of the figure shows the elution profile of twice crystallized enzyme, prepared by t'he standard isolation procedure (8). Except for small am0ullt.s of :ict,i\?ty \vhich were elut,ed \vlienever larger quantitiw of itlcrt protein appeared in t,he effluent (ill the i&al "~v:lsl~~tl~l.oUgll" nlld in the final NaCl wash t,o stl,il) the column), I~hosl)hoglucox iwnerase activity appeared ill two peaks about half~\T-:ly tlll~ollgll t11e gradiellt, designated :IS (2-I and G-II. On l,echrolll:~togr~ll~ll~ of the pooled fractions of l'eak G-II, enz>-rile acti\.it\; was agahl resol\-ed ilrto the two nxtjor peaks, G-I a11d (i-11, \vith some indication of splitting of (1-T into separate cwnlwl~ent5 'Thus;, enzyme form G-11 \&en pooled, coilcentrated, and rechromatogrsphecl &hout additional protective measures, will be partially converted to form GT. ('Uris is in contra-;t to the findings of Yohhida and Carter (0) who state that on reclll,onl:ttogr~~l)hy l,heir "isoenzymes" relnaill sirlgle peaks.) Elut~ion p:rtterns analogous to those shown iii Fig. 2 were also obtainetl when rabbit muscle estract, prepared according to the standard iwlation procedure (6)) was chrornatographed in the same system. Except for some phosphoglucohe isomerase which eluted wit11 :I huge amount of inert protein in the initial "washthrough" l)enk (for esplanation, see below), the majority of the activity appewed in t,he elution profile where the G-I and G-II peaks are observed for the crystalline ellzyxe.
Chromatogrnph~y of crystalline Phospl~oglwc~sc~ Isomernse Kept in Presence 0s l~ithiotlzreilol-~~~lieil tl\-ice cry5tallized enq'me was extensively dialyzed against buffer contanling 1 MM DT'i', the elution profile from ClWSephades (~p/xr porlion of Fig. 3) was quite similar to that shown in Fig. 2, excel)t, that the ratio of the total activities recovered in Peaks (i-1 :nitl G-II \~a< wmcwhat more in favor of form G-II compared \yit,ll c:hromat,ography in the absence of the thiol. However, on l~ecllromntoglnl)l~g of the G-IT fl,action-continuously kept in 1 nisr 1)1'T--the pat tern differs dwnatically in that no conversion of for111 G-II illto G-I took place, and all of the enzyme actirit\-:ll!pe:Lred ill :I yingle peak in the xrme location on the chronlutogr:rl)Eiic profile (boltom portion of Fig. 3).  (6). Thc,y wprtw~nt, the best data from a stud,v of the recovery of the illdiviclrl:ll :\rnino acids as a fnllction of h~tlrr)l,vsis t imc. The of Fig. 4. The uppermost pro$le represents the chromatogram of the original extract prepared in the presence of 5 m;M DTT and dialyzed and chromatographed at a 1 ml2 concentration of the thiol. This figure also provides an example of the experimental problems encountered in chromatographing enzyme species comprising only a small fraction of the total protein applied to the column (in this case, CM-Sephadex).
Part of the enzyme protein, either unspecifically bound to or merely swept along with the mass of protein not adsorbed to the ion exchanger but eluted from the column in the initial "wash-through," appeared as an artificial activity peak. On rechromatography, however, with substantially less total protein present on the column, the enzyme activity was located in the normal Peak G-II position, preceded by some activity in the G-I region (midportion of Fig. 4). This may be interpreted as indicating that. there wa,s still a small amount of the enzyme present in the original extract, that had not been reduced.
The bottom portion of Fig. 4 finally shows that the G-II form of the enzyme remains a single discrete species if it is kept in contact with the thiol reagent.
Chromatography of Five Times Crystallized Rabbit Muscle Phosphoglucose Isomerase--ilfter the studies described in the preceding sections had shown that variable mixtures of chromatographically distinct enzyme species can be obtained from both crude muscle extract and two times crystallized phosphoglucose isomerase, it was of paramount interest to investigate the chromatographic pattern of the five times crystallized rabbit muscle enzyme which had served as the standard source material for all of the chemical and physical studies previously reported from this laboratory.
To achieve maximal resolution, buffer volumes and column dimensions were modified (for details, see legends to figures) to further decrease the Na+ gradient. As a result, Peak G-I resolved into two subfractions, G-IA and G-IB, well separated from the main peak, G-II (Fig. 5). The chromatogram shown in Fig. 5 was obtained without DTT and reflects the distribution of the various species in a routine preparation of fifth crystals of phosphoglucose isomerase, i.e. 17, 21, and 62% for Fractions G-IA, G-IB, and G-II, respectively. Approximate average contributions of the three forms estimated for the various analyzed preparations ranged from 7 to 17y0 for Fraction G-IA, 15 to 25% for Fraction G-IB, and 60 to 75% for Fraction G-II.
Resolution and Interconversion of Fractions G-IA, G-IB, and G-II from Partially Purified n/luscle Extract-To provide larger quantities especially of Fractions G-IA and G-IB for more detailed chemical analysis, the standard phosphoglucose isomerase preparation (8) was carried through the first two steps, i.e. to the end of the ammonium sulfate fractionation, and then chromatographed on a column (5 x 41 cm) of CM-Sephadex with a sodium phosphate gradient of pH 6.90, 12 to 36 mM in Na+ (Fig. 6A). The effluent fractions were pooled as indicated in the figure and concentrated by membrane ultrafiltration.
The combined (G-IA)/(G-IB) pool was divided into two equal portions, one of which was kept in 10 mM DTT for 6 days, the other of which was rechromatographed without further treatment and without DTT present during chromatography.
The thioltreated sample was dialyzed and chromatographed under the same conditions except for the presence of 1 m&l DTT.
The untreated, combined G-I pool yielded the elution profile shown in Fir 6B, which completely lacks a G-II peak. In contrast, m. exposure of the ~me material to DTT resulted in approximately 307, co111 ersion of (i-1 species to the reduced G-II form (Fig.   6C).
The G-11 pool w:rs made 1 mhr wil,h DTT, dialyzed, and kept in tIllat medium for 72 days, then recllrom:ltogra1,~led in the pi'e-;ence of the thiol.
It yielded a single peak iii the G-II position (Fig. 6L>), indicating that the reductive conditions had prevented any conversion into faster eluting species.

Amino
Acid Compoeition of Cl~ro~~zatograplzically Separated Enzyntc! *S'pecies-~-~ltliough the conversion experiments left little doubt t,hat the various chromntographic species could not differ subst~antially in their primary structure, it was felt necessary to perform ammo acid analysis on the individual enzyme species to counteract Yoshidn :uld Carter's claim (9) that the resolved phosphoglucose isomerase "Isoenzymes" have different amino acid compositions.
Numerous amino acid analyses were carried out under rigorously controlled conditions (6) ; the results are summarized in Table I. The data show remarkably small deviations both among each other and when compared with the amino acid analysis of the five times crystallized enzyme previously reported (6). Certainly, the differences are not large enough to warrant the postulate that the various chromatographic forms possess different primary structures. Suljhyclryl and Total IIalJ-Cystine Content-The pronounced effect of dithiothreitol on the distribution pattern of the chromatographic fractions as well as its ability to convert the G-I chromatographic species to the G-II form suggested that the varying extents of oxidation of sulfhydryl groups might be responsible for the different chromatographic behavior. Numerous sulfl~yclryl analyses were therefore performed on the three chromatographit forms isolated from a substantial number of different preparations aiid the results are sholvn in Table II. In addit,ion, ali9uots of the three forms from several chromatographic experiments were also subjected to total half-cystine analysis to ascertain whether a, priori differences exist in the total cysteine-cystine content. These data are summarized in Table ITI. The result-; of both typeti of analysis are presented in their entirety, rather t,han only as averages Tvith tlreir deviations, lo illustrat,e the difficulty of obtaining rel)resentati\-e nie:L,suremcnts.
:llthouglr the final averages demoiistratc con vincingly a differeiicc iii the number of titrable sulfhgdryl groups, it is also apparent that this conclusion can only be drawn from the suni of the csl)crinlctits, whereas :I 1)articular single analysis might ilot aln:r)--conform to the pattern.
Even though all possible 1)recautioiis were taken to perform the sulfliydi~yl analyse; under conditions \I-hich would minimize osidation after clll~om:~iogra~)llic kolatioil :uid during tlic manipulations required for the assay, a disconcerting scatter of the data could not be avoided.
Ke:-erthcle;s, it is clear from Tables II and III that (a) the various cllrom:rtoglaphic species do not differ in their total h:llf-cystine content, (b) the number of -SH groups accessible to 1'XU iii tlie iiative state increasei by about 1 per molecule in the order (i-I-1, G-IU, G-II, and (c) the number of -SI-I groups that can be titrated in the denatured enzyme species increase; in t,he s:rine ortler by about 2 per molecule.
Notably, for the G-II fraction the number of total titrable --SII groups is equal to the total half-cystine content.
From an analytical point 01 view, it i3 gratifying to note that this detailed study of the sulfhydryl and total half-cystine con tent confirmed the routine analyses previously made on five times crystallized rabbit muscle 1)hosphoglucose isomerase, which contained variable amount.+ of t'he three chromatographic species (depending upon the extent of sulfhydryl oxidation inadvertently produced in the course of isolation).
A total cysteine content of 12.5 residues per molecule by PMB assay and of 11.5 (12.2 after correction) after performic acid oxidation had previously been reported for the five times crystallized enzyme (6).

Isoelectric
Focusing Experiments--To characterize the individual enzyme species in terms of their over-all net charge, pooled mixtures as well as individual chromatographic species were subjected to isoelectric focusing. Fig. 7 represents a typical clution profile obtained on focusing a mixture of the three species in an ampholyte pII gradient from pH 7 to 9. The isoelectric points measured by this technique vc-ere within the ranges of 8.05 =t 0.10 for Fraction G-IA, 8.25 f 0.15 for Fraction GIB, and 8.50 f 0.05 for Fraction G-II.
For comparison, the isoelectric point of five times crystallized rabbit muscle enzyme had previously been determined to be 8.5 (extrapolated to zero ionic strength) by free boundary electrophorcsis in various buffers as a function of ionic strength (4). h notable observation made during isoelectric focusing was that the experimental conditions a,pparently induced oxidation of the enzyme species. In the absence of added DTT, a larger percentage of the G-I species KLS always found on elution from the ampholyte column than would be expected from the known composition of the mixture applied to the column.
On the ot.her hand, addition of DTT to the am-