I952

Attempted Isolation of Haem a and Porphyrin a from Heart Muscle

BY J. E. FALK AND C. RIMINGTONDepartment of Chemical Pathology, University College Hospital Medical School,

University Street, London, W.C. 1

(Received 1 August 1951)

Haem aisthe namewhichhasbeengiven (Rawlinson& Hale, 1949) to the haem which is apparently theprosthetic group ofat least some ofthe cytochromesa. Studies of the visual spectrum of this haem, andof its iron-free derivative, porphyrin a, in com-parison with the spectra of other porphyrins andhaems of known structure (Lemberg & Falk, 1951)have led tothe postulation oftwo possiblestructures,consistent with all the data available.In addition to visual spectra an extensive study

has been made of the infrared spectra of mostnatural and many synthetic porphyrins and haems(Falk & Willis, 1951) in the hope that these mightprovide an analytical tool through which furtherdetails of the structure ofhaem or porphyrin a maybe made clear. The infrared spectra of porphyrinsand haems provide a useful means of identification,not only ofindividual porphyrins, but also ofcertainside chains on porphyrins. When the infraredspectrum of pure haem or porphyrin a can be ob-tained, some further light may be thrown on itsstructure by this means. Unfortunately, the isola-tion ofpure material has not yet been achieved. Thestudy reported below concerns further work on theisolation of the compound, for the purpose of ob-taining pure material for measurement of its infra-red spectrum and, if it could be obtained in sufficientquantity, for direct chemical study. Haem a ofdoubtful purity has been obtained in very smallquantities from horse heart (Negelein, 1933; Roche& B6n6vent, 1936) and from both ox heart and thecells of Corynebacterium diphtheriae (Rawlinson &Hale, 1949). The haemochromogen band of thishaem, at about 587 m,u., is easily observed with thehand spectroscope in pyridine extracts of manytissues (heart muscle, pigeon-breast muscle, insectthoracic muscle) after dilution with water andreduction with Na2S204. Attempts to isolate thehaem are complicated, however, by four factors:(1) The lability ofthe haem itself, particularly in thepresence of tissue components such as cysteine(Rawlinson & Hale, 1949). (2) The concurrentextraction of much protohaem (from haemoglobin,myoglobin, catalase, etc.). (3) The concurrentextraction of a lipid material which it is extremelydifficult to remove completely. (4) The relativelyminute amount of the haem present in the tissues.

Negelein's (1933) method depended on the ex-traction of the total haems from water-washed,minced muscle with acetone acidified with hydro-chloric acid. The yield ofmaterial with 'a very weakprotohaemochromogen band at 557 mu.' was12 mg. from 5 kg. offresh mince. The Soret band wasat about 430 m,u.

Negelein considered that the haemochromogen ofhaemahad a single visual band (at 587 mL.). Roche& Ben6vent (1936), on repeating Negelein's pro-cedure, obtained acompoundwithhaemochromogenbands at 587 and 530 m,. By a modification ofNegelein's process they obtained a compound,completely free from protohaem, with only a singleband in the visual region, at 587 m,.; the Soretband was at 425 m,. Roche & B6n6vent wereunable to crystallize this compound satisfactorily;they presented evidence which led them to believethat the compound with the two-banded haemo-chromogen (587 and 530 m,.) was the true haem a,and the compound with a single visual haemo-chromogen band an artifact.In 1949 Rawlinson & Hale developed a new

method for the separation of haem a from proto-haem. After extraction ofthe haems from the tissuesby acetone-hydrochloric acid, they were trans-ferred to ether, adsorbed on a column of aluminiumoxide, some lipid removed by washing the columnwith ether, the haems eluted with hot glacial aceticacid, and tranferred again to ether. On extractingthis ether solution several times with an aqueouspyridine-hydrochloric acid buffer the protohaemwas completely removed, leaving haem a in theether phase. The haem so obtained was contami-nated with lipids, but its haemochromogen hadonly a single absorption band in the visual region(at 587 m,u.). Rawlinson & Hale found that thehaem could react with compounds such as cysteineto yield a substance which gave a haemochromogenwith visual bands at 553 and 525 m,u., and that suchreactions could occur during isolation by unsatis-factory procedures. They considered, and it nowappears acceptable, that the natural haem a is thecompound which has a haemochromogen with thesingle visual band (at 587 mg.). Though Rawlinson& Hale's process was a great improvement onNegelein's, involving far simpler and fewer manipu-

36

HAEM a FROM HEART MUSCLElations, the yield was very small, and the product,which was contaminated by lipids, was unstable.Rawlinson & Hale (1949) prepared a porphyrin fromthis haem; the visual spectrum of the porphyrinprepared in this manner was used as a basis for someof the work of Lemberg & Falk (1951).

Negelein had earlier (1932a) reported the isola-tion ofa porphyrin from pigeon-breast muscle whichhe called 'cryptoporphyrin'. The haem prepared bythe introduction of iron into this porphyrin gave ahaemochromogen with bands (about 582 and531 m,.) recalling those of some other early haem apreparations, and at first he thought that this waspossibly the porphyrin of the prosthetic group ofthe cytochromes a. Shortly afterwards, however,Negelein (1932b) reported evidence which led him tobelieve that this porphyrin was an artifact arisingfrom protoporphyrin through the action of hydro-chloric acid during the isolation; indeed, in theoriginal paper he reported that the porphyrin couldbe obtained from crystallized, but not recrystal-lized, haemin from blood. No cytochrome a hasever been identified in blood, and there is thusgood evidence that the porphyrin was an artifact.This was further discussed by Lemberg & Falk(1951).

It was now sought, after extraction of the haemsfrom ox heart and conversion ofthese to porphyrins,,to prepare porphyrin a in greater quantity and in apure state. A process was indeed found by whichrelatively large amounts of porphyrin, free fromprotoporphyrin, can be prepared conveniently inordinary laboratory apparatus. It has been shown,however, that porphyrin a prepared by this method,and presumably by any method so far available, is amixture of closely similar substances. Evidence ispresented which shows that these substances arise,during the isolation, from one, or at most relativelyfew precursors.The cause of the degradation of the original sub-

stance(s) has been found to be the action of acid andno process has been found in which this can beavoided entirely. Until such a process is devised,the problem appears to be insoluble.

MATERIALS AND METHODS

Absorption spectra were measured with a Beckman photo-electric spectrophotometer.

Ether was treated to remove peroxides.Hydrochloric acid concentrations. Because the familiar

HCl number (Willstatter no.) widely used in the purifica-tion of porphyrins is stated in terms of % (w/v) HCI, thisform is used instead of normality.

Reaction with hydroxylamine. To a solution of theporphyrin in pyridine, excess of a mixture of equivalentamounts of solid hydroxylamine hydrochloride andNa2C03 was added, the mixture refluxed gently for 5 min.,cooled and filtered.

Preparation ofporphyrin a. Method A

(1) Extraction of haems. Fresh ox heart (4-6 kg.), dis-sected free of macroscopic fat, yielded 3-2 kg. of mincedmuscle; this was washed twice with acetone at 00, pressingout each time, and air' dried (800 g.). Of this dried mince,500 g. were extracted at 3° for 2 hr. with 21. acetone con-taining 40 ml. conc. HCI. The extract was filtered from thetissue residue; so little haem remained in the tissue that asecond extraction was not profitable. The filtrate was mixedwith an equal volume of ether, and the acetone and HCIwashed out with 2% NaCl to minimize emulsions.

(2) Preliminary defatting. The ether solution was now runthrough a column (10 x 3 cm.) of MgO grade III (Nicholas,1951) packed in ether; the haems were adsorbed as a verydeeply coloured layer at the top of the column. The columnwas then washed with ether (about 21.) until the etherrunning throughno longer left afattyresidue onevaporation.The dark zone containing the haems was separated from thecolumn, and the haems eluted with glacial acetic acid. SinceMgO dissolves in glacial acetic acid, the elution was quanti-tative and could be done at the melting point of acetic acid.The acetic acid solution was mixed with about 2 1. ether, andthe acetic acid and the magnesium acetate washed out with2% NaCl; the ether was then removed in vacuo.

(3) Removal of ironfrom haems. The residue was dissolvedin 100 ml. hot glacial acetic acid, and this solution treated in20 ml. portions as follows. The haem solution was broughtquickly to the boil, and while refluxing gently, about 5 ml. ofa boiling saturated solution of ferrous acetate in acetic acid(prepared under C02) and 2 ml. conc. HCI were added. Theresulting porphyrinsolutionwas cooled as quickly as possibleunder the tap. This is the process of Warburg & Negelein(1932), modified so as to use the least possible amount ofheat.The several lots of porphyrin solution so prepared were

combined, mixed with 21. ether, the acetic acid neutralizedwith sodium acetate, andthe ether solution ofthe porphyrinswashed several times with 2% NaCl.

(4) Removal ofprotoporphyrinfrom the porphyrin mixture.The ether solution was shaken with 500 ml. portions of 4%(w/v) HCI until no more protoporphyrin was removed. Thiswas usually achieved in six or seven extractions; verylittle porphyrin a was extracted at the same time, but mostofit remained in the ether phase, where its absorption bandscould be seen with the hand spectroscope at 648, 582, 560,518 m,u. approx. Emulsions were broken when necessary bycentrifuging.

(5) Removal of lipids by treatment with 25% hydrochloricacid. The ether solution, besides the porphyrin a, still con-tained much lipid material. It was found that this could bequantitatively removed as follows. The ether was removedin vacuo, and the residue shaken with 25% HlC1 at - 10°.After standing for about an hour at this temperature theporphyrin was in solution and the lipids which had solidifiedwere easily separated by gravity filtration at -10° (What-man no. 54 paper). The filtrate was clear and olive-green incolour. Ether was added, the mixture diluted with water,and on neutralization with sodium acetate the porphyrinwas transferred to ether. The ether solution could now bewashed with water; indeed, after this treatment, no moreemulsions occurred at all. No more fatty material could beremoved by repeating the 25% HCI treatment. The por-phyrin now appeared to be stable ifkept in ether or pyridine

Vol. 5I 37

J. E. FALK AND C. RIMINGTONsolution. No change could be detected by spectrophoto-metric measurements in material stored for several monthsat 30From 500g. dried mince (equivalent to 2000g. fresh

muscle) yields of 18-20 mg. of this porphyrin were regularlyobtained in the course of a working day. Its spectroscopicproperties were very close to those ofthe porphyrin preparedby Rawlinson & Hale (1949) (cf. Table 2).

Preparation of porphyrin a. Method BThe haems were extracted from the acetone-dried tissue

as in method A, step 1, and gross fat removed as in step 2,except that a column of AI,60 (Savory & Moore) was usedinstead of MgO, and the haems eluted by several lots of hotglacial acetic acid.

After the elution the haems were again taken into ether,and the ether solution shaken repeatedly with an equalvolume of pyridine-HCl buffer (30 vol. pyridine, 0-15N-HCIto 100 vol.; cf. Rawlinson & Hale, 1949) until no moreprotohaem remained in the ether phase. The ether solutionof crude haem a was then evaporated to dryness in vacuo.A portion now dissolved in pyridine, diluted with 2 vol. ofwater, and reduced with Na,S304 gave a haemochromogencurve identical with the curve published by Rawlinson &Hale (with a single visual band at 587 mp.).The haem wasnow dissolved in glacial acetic acid, and the

iron removed as in step 3 above; the porphyrin obtained wastreated with 25% HCl as in step 5.

RESULTS

Porphyrin prepared by Method A. Spectrophoto-metric curves of the material before and after thetreatment with 25% HCI, and of the fatty residue,are shown in Fig. 1. The ratios of the intensities ofthe absorption bands I-III to that of band IVprovide a useful means of comparing such curves(Table 1). As may be seen from Fig. 1, with theremoval of the strong absorption in the blue regiondue to the fat, the intensities ofbands I-III increaserelative to IV, though the positions of the maximaare hardly changed.

Porphyrin prepared by Method B. This process isessentially the same as that used by Rawlinson &Hale (1949) for the preparation oftheir porphyrin a;the main difference is that instead of the treatmentwith 25% HC1 they removed some fatty materialfrom the porphyrin by repeated transfers betweenHCl and ether.

Absorption curves of the porphyrin before andafter the 25% HCI treatment were similar to thoseshown in Fig. 1. Indeed, the material obtained by

this method behaved in all respects like that fromMethod A. The manipulations were much moretroublesome, however, and the yields much smallerand for most of the experiments reported belowmaterial prepared by Method A was used.

600Wavelength (mIL)

Fig. 1. Visual absorption spectrum of the porphyrin:a, before, and b, after the treatment with 25% HCI(stage 5, Method A); - - -, the fatty residue. Solvent,pyridine.

Preliminary ether -HCl fractionation of porphyrin aRawlinson & Hale observed (personal communi-

cation) that the HC1 number of their porphyrinapparently became lower as transfers between HC1and ether were repeated. We made similar observa-tions. Thus before the 25% HC1 treatment (step 5) itwas possible to remove the protoporphyrin with4% HC (step 4) without appreciable loas of por-phyrin a. After the treatment, however, even 1 %HCl extracted significant amounts of porphyrin a-like material from ether.We found, on preliminary fractionation of our

material, that 6% HC1 removed a considerablefraction, and when no more porphyrin was removedby acid of this strength, a further fraction at least aslarge could be extracted by 15% HC. Absorptiondata (in pyridine) for typical 6 and 15% fractionsare shown in Table 2, where the measurements ofRawlinson & Hale (1949), calculated to the sameform, are included for comparison. The positions ofthe bands in all the materials were very similar, butband I (about 650 m,.) in the 15% fraction wasmore intense than in the fractions extracted byweaker HCI solutions.

It was evident that our porphyrin a, which spec-trophotometrically was virtually identical with thatdescribed by Rawlinson & Hale, was a mixture. It

Table 1. Positions of absorption maxima, and ratios of intensities of the porphyrin (in pyridine solution)a, before, and b, after the treatment with 25% HCI (cf. Fig. 1)

Positions of maxima(me.)

A

... IV III II Ia 517 560 582 647b 517 560 583 648

Intensities, relative to band IV

IV1.01*0

III1*1971-525

II0*8521-148

Band I0*2950*328

38 I952

HAEM a FROM HEART MUSCLE

Table 2. Po8ition8 of maxima and ratio8 of intensitime of ab8orption bandsof porphyrin preparations (see text); solvent, pyridine

Positions of maxima(mjL.)

Band . ... ... ... ... ... IVMaterial

Porphyrin of Rawlinson & Hale (1949) 516Porphyrin prepared by Method A 517Fraction extracted by:

6 % HCI 51815 % HCI 517

was at first thought that the material with higherHCI number and increased intensity of band Imight be an artifact which had arisen during themanipulations. Artifacts with such characteristicsare not uncommon in porphyrin chemistry.Coiitrolled experiments showed that the proportionof this material obtained was not influenced by:

(1) The length oftime for which the acetone-driedmince was stored (at 30) before extraction. There didnot appear to be any significant spectrophotometricdifference between the product obtained from one

half of a batch of acetone-dried ox-heart mincewhich was extracted at once, and the products fromthe other half, which was extracted after it had beenstored for 24 days at 3°.

(2) The length of time the mince stood with ace-

tone-HCl for extraction of the haems. A batch ofacetone-dried ox-heart mince was halved. One halfwas extracted with acetone-HCl for 1.5 hr. and theporphyrin a prepared by Method A immediately.The other half was extracted for 18 hr. and theporphyrin prepared in the same way. There was no

significant difference in the spectrophotometricproperties, nor in the relative amounts of theporphyrins extractable by 6 and by 15% HC1 ineach experiment.

(3) The use of the magnesium oxide or aluminiumoxide columns. A batch of the porphyrin was pre-pared essentially by Method A, the preliminarydefatting on the column (step 2) simply beingomitted. The procedure was rendered rather more

difficult by emulsion formation, but the 25% HCltreatment removed the fat completely. The productwas extracted exhaustively with 6%, and then with15% HCI; two crude fractions were again obtained,their spectrophotometric properties being similarto those reported above (Table 2).

There remained as a possible cause of degradationthe treatment with aqueous HC1-at the stages ofether-HCl fractionation, and the 25% HCI treat-ment for the removal offat. Most known porphyrinsare quite stable to such treatment, and in additionthe strong band at about 650 mp. was observedwith the hand spectroscope in preparations whichhad never been treated with aqueous HCI. (proto-haemin having been removed by Rawlinson &

III II

Intensities,relative to band IVK A

I IV III II I

559 582 647 1.0 1-512 1-097 0-301560 583 648 1.0 1-525 1*148 0*328

562 583 648 1.0 1-530 1-162 0 346562 584 650 1.0 1*510 1-085 0-523

Hale's pyridine buffer treatment). Further, theabsorption curves (cf. Fig. 1) before and after the25% HC treatment suggested only that this treat-ment caused a fall in the absorption at the region of500-520 m,u. relative to that at about 650 mpA., andnot a specific increase in intensity of the band at650 mI,.The hypothesis that the substance of higher HCI

number was an artifact was thus not directly provedor disproved.A possible altemative hypothesis was that it was

the substance with lower HCInumber and band I oflower intensity which was the artifact. If this werethe case, the experiments described above shouldhave provided evidence about it just as well as theoriginal hypothesis they were designed to test. Theone process not yet tested was the action ofaqueousHC. The reason which made it appear unlikely thatsome effect of HOC could have been the cause of theappearance of the substance of higher HC1 number(observation of the banxd before any aqueous HC1had been used) argues not against, but for the possi-bility that the converse process was taking place,namely, some change was caused by aqueous HClas a result of which the substance of lower HCInumber was derived from the substance of higherHOC number.Evidence about this was sought by careful

fractionation, with HC1, of an ether solution ofporphyrin prepared by Method A, and refractiona-tion in the same way of the fractions so obtained.

Full fractionation and refractionationof the porphyrin preparation

An ether solution of the porphyrin (free fromprotoporphyrin) prepared by Method A was ex-

tracted with successive portions of HC1 as shown inFig. 2a. The volume of the ether phase was keptconstant by the addition of fresh ether as required,and the volume of HC1 used at each extraction wasequal to that ofthe ether. The position and intensityof maximum absorption in the Soret region (about410 m,.) was determined in each HCl extract.Beer's law was obeyed at the concentrations used,and the density readings were proportional to theporphyrin content ofeach fraction. For this purpose

Vol. 51 39

J. E. FALK AND C. RIMINGTON

it was not necessary to adjust the HCI concentrationto the same value in every fraction.The distribution of porphyrin in the successive

fractious is shown in Fig. 2a. It is obvious thatperfect separation into true fractions was notachieved, but there is good evidence for the presenceof several (at least four) components. For the re-fractionation, the combined 2, 3, 4 and 6% extracts(b), the 8 and 10% extracts (c), and 12, 15 and 20%extracts (d) were transferred to ether, and the ether

(c) on refractionation yielded significant fractions to1, to 8 and to 10% HC1, but virtually no porphyrinremained after this (fluorescence under ultravioletlight hardly visible). This meant that the fractionswhich passed from the original into 15 and 20%HC1 were not present in fraction (c), nor did theyarise during the refractionation. The 1% fraction ofthe original should equally have been excluded fromfraction (c), but apparently more of this materialarose during the refractionation. This direction of

aLhLLLLLL~~~~~~~~~~~~~~~I

a ~~~~~~~~~~~~~~~~~~~~I

b I

l~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

_-_12~~~~~~~~~~'C~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

c

L~~~~~~~h km; LL

IIIIIIIIIIIIIIIIII S.-

1 2 3 4 6 8 10 12 15 20Concn. HCI (%)

Fig. 2. a, the relative yield ofporphyrin (determined at the Soret maximum, see text) at successive extractions with HCI.Each step on the histogram represents one extraction with HCI of the concentration shown on the abscissa. b-d, therelative yields on refractionation, in the same way, of b, the combined 2, 3, 4 and 6 %; c, the combined 8 and 10 %(readings x 4); and d, the combined 12, 15 and 20 % (readings x 8) fractions from the fractionation shown in 2a.

solutions (b-d) extracted with successive lots ofHCI in the same way as in the original fractionation.The results of these refractionations (Fig. 2 b-d)were very interesting. It was evident that shakingthe ether solutions with aqueous HC1 caused a

degradation. The direction of this degradation wasfrom products with higher, to products with lower,HCI numbers.Thus fraction (d) could hardlyhave contained any

of the original material which was extracted by HC1concentrations lower than 8 %, yet on refraction-ation it yielded a pattern of fractions very like thatfrom the original porphyrin solution. Again fraction

the degradation, towards products with lower HCOnumbers, is strikingly confirmed by the results ofrefractionation of fraction (b).

Absorption spectra of fractions. The first HCIextract at each HC1 concentration (cf. Fig. 2a), as

the fraction least likely to be contaminated withmaterial of higher HCI number, was transferred toether, washed well, the ether removed in vacuo, andthe porphyrin dried and dissolved in pyridine.Absorption curves in the visual region were taken oneach of these pyridine solutions; the positions ofthe maxima and the ratios of the intensities of thebands are shown in Table 3. It was found useful to

40 I952

HAEM a FROM HEART MUSCLE

Table 3. Maximxa and ratios (in pyridine solution) of the first fraction extracted by HCIat each of the concentrations indicated, in the fractionation shown in Fig. 2a

First fractionextracted by HCI

(%)123468

10121520

Positions of maxima(mIA.)

A

Soret413417417417417418417418418418

Intensities, relative to band IV

IV III II I IV III512 555 581 645 1x0 1*230517 560 583 643 1.0 1-660517 560 582 642 1*0 1-530516 559 583 645 1*0 1-380516 561 582 644 1.0 1-338517 561 583 644 1*0 1*500518 560 583 645 1.0 1.555517 561 583 646 1-0 1*560517 560 584 646 1.0 1-445517 561 583 649 1.0 1*390

introduce a graphical method for comparison of thespectroscopic properties of these rather similarmaterials. The changes in position of the maximaare readily appreciated from a consideration of thetable, but the pattern in the ratios of intensities ofthe bands in the different fractions is made muchclearer by plotting them as in Fig. 3.

1-5 _

%f_

t!. I -@ _

._1

v I_~

X

'CI \

I \ /I ...

XI / '.ol: X,, *s. .. o

I 'X * * ,Jt--XsI"X

XI' X%/ -X, -AK 11. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~. It

0-4 'C '* )Ix

1 2 3 4 6 8 10 12 15 20.Concn. HCI (%)

Fig. 3. Diagram showing the change of intensities of bandsI-III relative to band IV; data from Table 3. The relativeintensities after reaction with hydroxylamine are shownas-O-.

The similar data from the refractionation offractions (b) and (d) are shown in Fig. 4, in which thepattern of the ratios of intensities of the bands inthe original fractions is strikingly reproduced. Thissimilarity leaves no doubt that really differentmaterials are contained in these different fractions.

It is noteworthy that all the compounds fromrefractionation of fraction (d) (Fig. 4 a) have therelative intensities of bands II and III depressed incomparison with the corresponding original frac-tions (Fig. 3). For example, the porphyrin extractedby 6% H10 (Fig. 3) has a true oxorhodo typespectrum, while the 6% fraction (Fig. 4 a) has onlya weak rhodo type spectrum. The accuracy of the

II0.9001*2201*1601-0781-0321*1001*1501.1501*0701*050

I0-6150-2780-3160-3130-3060*3100-3120-3470-3450-420

determination of the ratios relative to band IV wasapproximately ± 0-02 unit for bands II and III and+ 0-05 unit for the much weaker band I. Band Iwas significantly more intense in the 15 and 20%fractions, and this increased intensity was paralleledby a slight shift to longer wavelengths. This was theonly consistent marked change in the positions ofthe maxima.

1-5 I %\

- X ~I 'X)( I X

1-0IV.~~~~~~~~ _(.' ..l o\

: .

O---O .O.._ - *- ° ... . 10- 11CY .b

'C II

1 2 3 4 6 8 10 12 15a Concn. HCI (%)

r x\-xxl

X-' \x

II

1234

1 2 3 4 6

Fig. 4. The changes in relative intensities in the materialsfrom: a, stage d, Fig. 2; b, stage b, Fig. 2. Relative in-tensities after reaction with hydroxylamine are shown as-0-.

Fractionations in the above manner were carriedout on several samples of porphyri, prepared byMethod A from different batches of ox hearts, withconsistent results.

Effect of heating the porphyrin withaqueous hydrochloric acid

Since it appeared that the changes were caused insome way by aqueous HC1, its effect was studied inmore detail. It was hoped that the material mightbe degraded by HC1 under more drastic conditionsto a useful, single degradation product.

In a preliminary experiment some material prepared byMethod A was fractionated as in Fig. 2a above. The com-bined 8 and 10 % fractions, and the combined 12, 15 and

Vol. 5I 41

V)vt!cvr-

v

.41m.vw

J. E. FALK AND C. RIMINGTON20% fractions were transferred to ether, the solutionwashed, the ether removed, and the residues dissolved in25% HCI. These HCI solutions were each divided into fourportions. Portion 1 was at once transferred to ether, washed,

pected to shift towards shorter wavelengths; there were notany significant changes in the positions of the other bands.The changes in the relative intensities of the bands were,however, quite marked and these are shown in Fig. 5.

1-5 -

43

.;

0-4

X\Band III- K- X

\, Band IV

- Band 11

Band Ix--X--X--X

a b c dA

'CX

111

X--X--X--X I

a b c dB

Fig. 5. The change in intensities of bands I-Ill relative toband IV, when: A, the combined 8 and 10% fractions andB, the combined 12, 15 and 20% fractions (Fig. 2a) stoodin the dark in 25% HCI solution a, at zero time; b, after60hr. at 100; c, after 30hr. at 300; and d, after 3 hr. at 700.

1*5

W_

*4c

43

V 0-8

0-4

\X.,""-----X--X__

_ ~~ ~_X%

X~~~~~%X.,

-"X--x__x_ -y V,

x xIII

IvVI)0.,

L

X--X--X--X___ d-.---XIX4XX .o --X--- -

, . , ., ,-X,.. .

I I Ic 1 Ifa b cde f g h i j

Fig. 6. The change in relative intensities of the bands ofporphyrin prepared by Method A, after standing in thedark in 25% HCI solution a, at zero time; b, after 50 hr.at -10°; c-h, after 1, 2, 3, 4, 5, 6 hr. at 570; i, j, after1-25 and 5 hr. at 956. The relative intensities after reactionwith hydroxylamine are shown as -0-.

the ether removed and an absorption curve measured on theresidue dissolved in pyridine. Portion 2 was allowed tostand in the 25% HC1 at - 10° for 50 hr.; portion 3 stood inthe dark at 300 for 30 hr. and portion 4 stood in the dark at700 for 3 hr. At the end of these times each sample wastransferred to pyridine and its absorption curve measured.In the small samples used, band I was too weak for accuratedetermination of its position, which might have been ex.

It is seen that the 25% HCl, even at - 100,caused a depression of the intensities of bands IIand III relative to that of band IV. The depressionof these bands was increased and hastened as thetemperature increased. These measurements wereon the whole samples, without fractionation. Thedepression of the bands in the separate fractions(Figs. 3 and 4a) was paralleled in the present ex-periment by the depression of the bands in theunfractionated material.

In another experiment, porphyrin prepared byMethod A was treated, in HC1 solution, as indicatedin Fig. 6. The material, which in this experimenthadbeen treated at 570 for 6 hr., was transferred to etherand fractionated with HCL. The pattern of relativeintensities of the absorption bands of the fractionswas similar to that shown in Fig. 4a. As in theprevious experiment, the depression of the ratios ofbands II and III increased both with time andtemperature, though the band positions hardlychanged. Even treatment ofa 25% HCI solution at950 for 5 hr., however, led to a change only fromoxorhodo- to rhodo-type spectrum, band I re-mainiing at 644 m,i.

Reactions with hydroxylamineVarious fractions were treated with hydroxyl-

amine, and the spectroscopic properties of the pro-ducts are indicated in Figs. 3, 4a and 6. It may beseen that in all fractions so treated, irrespective ofthe character of the spectrum before the'treatment,oxime formation had taken place, demonstratingthat the -CHO group was still intact. This is dis-cussed below.

Evidencefor degradation during removal of iron(8tep 3, Method A)

There is no doubt that HCI, during the HCl-etherfractionations, and also during the treatment with25% HCI causes changes in the material. Itappeared likely that similar changes would occurduring the removal of iron.

Porphyrin mixtures after removal of iron fromthe haems were transferred immediately to etherand esterified with diazomethane. Chromatographyof the esters on columns of aluminium oxide gradeIV and of magnesium oxide grades III and IV (cf.Nicholas, 1951) showed the presence, apart fromprotoporphyrin, of a variety of porphyrin a-likematerials which could not be satisfactorily resolved,It is only necessary to report briefly that materialswith spectra similar to most of the fractions shown

t

42 I952

HAEM a FROM HEART MUSCLEin Fig. 3 were obtained from the chromatograms. Itthus appeared that, during the removal of iron,materials similar to those which arise during treat-ment with HCI had appeared.

Attempts to purify haem a

There is no evidence yet that the haem, as such, islabile in acid conditions. This was bome out to someextent when, instead of acetone-HCl, cold pyridinewas used to extract the haems from the tissue. Thehaems were transferred to ether, and without anyattempt at removal of the large amount of lipidmaterial the protohaem was removed by theaqueous pyridine-HCl buffer method of Rawlinson& Hale (1949). The haem a so obtained was grosslycontaminated with fat, but its haemochromogenhad only a single visual absorption band, at 587 m,u.,apparently the same as the material prepared afteracetone-HCl extraction of the tissue followed byelution from alumina by boiling acetic acid. Thishaem was eventually converted to porphyrin, andthis treated by Method A. On fractionation of anether solution of this porphyrin with HC1, fractionsidentical with those shown in Fig. 2a were obtained.

Chromatography of haems

The haems were extracted from the tissue anddefatted as in steps 1 and 2 of Method A. The re-sulting ether solution of the haems was evaporatedto dryness in vacuo. Neither light petroleum norbenzene extracted any fatty material from the dryresidue. The dry material was soluble in butanol,which also dissolved the haem when shaken witha suspension of it in water brought to pH 4.The dried haems, or these, after esterificationwith diazomethane, were used for the followingexperiments.The haem esters were chromatographed on

columns of the following absorbents: Aluminagrades II and IV, magnesium oxide grades IIand III (cf. Nicholas, 1951); talc and kieselguhr(Hyflo supercel). The following solvents were used,singly and in pairs, in varying proportions: chloro-form, methanol, ether, benzene, pyridine. Separa-tion ofboth the haems and the haem esters was alsosought on partition columns. The following solidswere tried as supports: filter paper powder (What-man, standard grade) and kieselguhr (Hyflo super-cel). For 'reverse phase' chromatography, Hyflotreated with dichlorodimethylsilane was used (cf.Martin, 1949; Howard & Martin, 1950). The solventsystems tried for all three supports were lutidine-water, equilibrated and used at both 210 and 30 andether-pyridine buffer (pyridine 30 vol., 0-15N-HC1to 100 vol.), at room temperature (18°).

There was no indication in any of these experi-

ments that the haem a might be separated from theother materials. The coloured material in every casemoved slowly down the column with the solventfront. Material which eventually ran through thecolumns had the same proportion of haem a toprotohaem as the starting material.

DISCUSSION

From the experiments reported, it became clearthat once the iron is removed from haem a, theporphyrin is very unstable in the presence of acid.The spectrum of the product obtained was the sumof the spectra of the degradation products, and it isclear that the proportions of these, and the re-sulting mixed spectrum, vary with both time andtemperature during manipulations with acid.The tendency of the material with higher hydro-

chloric acid number and band I at longer wave-lengths to be changed to material with lower hydro-chloric acid number and band I at shorter wave-lengths, as well as the constant downward trend inthe intensities of bands II and III-relative to bandIV, are consistent with the hypothesis that the

CHO Me CHO Me

Me Et MeEt

Me o M Me Me Me

2{ HC==CH YH2C O~H CH2COOH COOH ICH 0

(I) (II)

changes are due to the destruction of 'rhodofying'groups (Lemberg & Falk, 1951). Among such groupsare the -CHO group, the -CH:CH.COOH group(Formula I), and the unsaturated isocyclic ring(Formula II). These were suggested by Lemberg &Falk (1951) as possible structures for porphyrin a,consistent with the visual spectroscopic propertiesof haem a and porphyrin a in comparison with theproperties of compounds of known structure. As abasis for this study, the spectra of haem a andporphyrin a described by Rawlinson & Hale (1949)and Rimington, Hale, Rawlinson, Lemberg & Falk,(1949) were used. The presence of a -CHO groupwas confirmed (Lemberg &; Falk, 1951), but theexact nature of the other 'rhodofying' group is notknown. These studies are considered not to beinvalidated by the present work, since it seemslikely that the undegraded, natural porphyrin a hasspectroscopic properties very close to those ofRawlinson & Hale's (1949) material.

VoI. 5I A43

J. E. FALK AND C. RIMINGTONSince it is not clear how a-HO side chain could

be changed by HC1 in such away as to yield productswith the characters described above, it was con-sidered more likely that it was the other 'rhodo-fying' group which was being changed. That thiswas so became apparent when it was found that thematerial which had been heated to 950 for 5 hr. in25% hydrochloric acid solution was still able toreact with hydroxylamine, the spectrum changingfrom rhodo to aetio type and band I shifting from647 to 636 mit.These findings were parallel with those of Raw-

linson & Hale (1949) and Rimington et at. (1949),who found that treatment of their porphyrin a withdiazoacetic ester or HI (double bonds in side chains)or hydroxylamine (carbonyl groups in side chains)led to a change from oxorhodo- to rhodo-typespectrum, band I hardly shifting in position. Theaction of both these types of reagent in succession,however, led to a product with aetio-type spectrumand band I at 625 mp. The parallelism between thespectra of numerous fractions before and afterreaction with hydroxylamine (Figs. 3, 4a, 6)showed, moreover, that the effect of the -CHOgroup on the spectrum was approximately equiva-lent in all the materials. Thus the differencesbetween the fractions must be due to a series ofchanges in the other rhodofying group. It is evidentthat this group is gradually changed by acid towardsan end state in which its rhodofying effect is com-pletely lost. It is not possible, however, to postulateintermediate steps in this degradation which couldaccount for the many apparent stages in the change.

Acid apparently caused degradation of the por-phyrin even at step 3 (removal ofiron). It is possiblethat the numerous products revealed by chromato-graphy directly after this step were original com-ponents, but their similarity to the materials whichwere shown to be produced by hydrochloric acid atlater stages makes it more likely that they arose inthe same way. All methods for removing metalsfrom metalloporphyrins (including the relativelymild method of Paul, 1950), feature strongly acidconditions except the sodium amalgam method(Fischer & Hilger, 1924). The latter was unsuitable,however, because the -CHO side chain and theside chain with an ethylenic double bond would bereduced. Thus no suitable altemative process isavailable for this step, nor could any be found for theother steps involving the use of acid. Fractionationofether solutions ofthe haems with aqueous NaOH,Na2CO3 or Na2HPO4 was ineffective. Until suitabletechniques are developed for all these steps,attempts at purification through the porphyrinsmust be unsuccessful.

It appeared more profitable to turn again to theseparation and purification of the haem as such.There is no direct evidence that the haem is unstable

in acid conditions, though this possibility cannot beignored. It should be pointed out that the haemo-chromogen band at 587 m,. in haem a prepared byRawlinson & Hale's (1949) method, and, indeed, indirect pyridine extracts oftissues, is broad, andmayinclude the bands of several similar compounds. Itis quite possible that there exists more than onenatural haem a, perhaps corresponding to differentcytochromes a. Examination at very low tempera-tures (cf. Keilin & Hartree, 1949) of this haemo-chromogen band in pyridine extracts of tissuesmight allow the detection of such components,though the lipids extracted concurrently by pyri-dine would make such a study difficult.There is little doubt that the failure to separate

the haems by chromatography was due to the pre-sence of lipids, which might be expected to changetheir partitioning properties. The gross fat can beremoved without much trouble, but the lipid whichis encountered in smaller but appreciable quantitiesin attempts to purify haem or porphyrin a is still,perhaps, the greatest single factor hindering itsisolation.

It appears now that the most fruitful approach tothe problem might be to abandon the efforts topurify the natural compounds and instead toattempt to prepare a stable derivative of the haemthrough, for example, catalytic hydrogenation,fusion in resorcinol or the action ofdiazoacetic ester.Material obtainedby the procedures ofsteps 1 and 2,Method A, may be sufficiently free of gross contami-nation for this purpose. Unfortunately, it is notcertain that the haem even at this stage has notalready suffered some change.

SUMMARY1. During attempts to find a method for the

preparation, from ox-heart muscle, of porphyrin ain quantities sufficient for direct chemical study, itwas found that the porphyrin is very unstable inacid media.

2. Evidence is presented that the -CHO sidechain of the porphyrin is not changed during thisdegradation. It appears that the other 'rhodofying'group is modified in a gradual manner, leading to aseries ofporphyrins with rather similar spectra. Thedegradation is hastened at raised temperatures, andin the end-state the 'rhodofying' properties of thegroup are lost completely. It was suggested(Lemberg & Falk, 1951) that this group may be anacrylic acid side chain. It is not yet possible, how-ever, to interpret in terms of chemical structure thechanges which occur during the degradation.

3. No procedure has been found by which theuse of acid can be avoided entirely during theisolation of the porphyrin.

4. Haem a, as such, may not be unstable to acids,though there is no direct evidence on this point.

44

Vol. 51 HAEM a FROM HEART MUSCLE 45

Attempts were made by the use of both absorptionand partition chromatography to find a method forthe isolation of relatively large amounts of haem a,but without success.

The authors gratefully acknowledge helpful discussionswith Dr R. Lemberg and Prof. W. A. Rawlinson, who in

Sydney and Melbourne respectively are collaborating in thisprogramme, and with Prof. D. Keilin, F.R.S.The work was done during the tenure by J. E. F. of a

Fellowship from the Nuffield Foundation.The Beckman spectrophotometer was purchased out of

a grant toC. R. fromthe Central Research Funds Committeeof the University of London.

REFERENCES

Falk, J. E. & Willis, J. B. (1951). Au8t. J. Sci. Re8. (in thePress).

Fischer, H. & Hilger, J. (1924). Hoppe-Seyl. Z. 138, 49.Howard, C. A. & Martin, A. J. P. (1950). Biochem. J. 46,

532.Keilin, D. & Hartree, E. F. (1949). Nature, Lond., 164,254.Lemberg, R. & Falk, J. E. (1951). Biochem. J. 49, 674.Martin, A. J. P. (1949). Biochem. Soc. Symp. 3, 12.Negelein, E. (1932a). Biochem. Z. 248, 243.Negelein, E. (1932 b). Biochem. Z. 250, 577.

Negelein, E. (1933). Biochem. Z. 266, 412.Nicholas, R. E. H. (1951). Biochem. J. 48, 309.Paul, K-G. (1950). Acta chem. 8cand. 4, 1221.Rawlinson, W. A. & Hale, J. H. (1949). Biochem. J. 45,247.Rimington, C., Hale, J. H., Rawlinson, W. A., Lemberg, R.& Falk, J. E. (1949). 18t Int. Congr. Biochem. (Abstr.),p. 379.

Roche, J. & B6n6vent, M. T. (1936). Bull. Soc. Chim. biol.,Pari8, 18, 1650.

Warburg, 0. & Negelein, E. (1932). Biochem. Z. 244, 9.

Some Properties of the Glutaminase of Clostridium welchii

BY D. E. HUGHES ANi D. H. WILLIAMSONMedical Re8earch Council Unitfor Re8earch in CeU Metaboli8m, Department of Biochemi8try,

the Univer8ity, Sheffield 10

(Received 26 July 1951)

In previous papers from this laboratory (Krebs,1948; Hughes, 1949, 1950) it was reportedthat cetyltrimethylanmmonium bromide (cetavlon)accelerates the decarboxylation of glutamine andglutamic acid in intact cells and extracts of Clostri-dium welchii. The present paper is a study of themechanism of this effect. The glutaminase has beenpurified and the effect of cetavlon upon purifiedenzyme preparations has been investigated. Whilstcetavlon accelerates the rate of deamination ofglutamine in intact cells and crude extracts itinhibits it in the purified extracts. These findingsand the result of kinetic studies support the viewput forward previously (Hughes, 1949) that theaccelerating effect of cetavlon is due to the removalofan intracellular inhibitor normally accompanyingthe enzyme.

METHODS

Organisms. Three strains of C(lotridium welchii (strainsSR 12 and 1490 of the National Collection ofType Cultures,and a locally isolated strain) were maintained in Robertson'smeatmedium. Through the courtesy ofDrB. C. J. G. Knight,two batches of about 500 g. (wet wt.) of strain 1490 weremade available from the Wellcome Physiological Labora-tories, Beckenham, Kent. These cells had been collectedafter 5 hr. growth on the papain digest described below.

Growth medium. The usual medium consisted of caseinhydrolysate, meat, yeast extract and glucose (see Krebs,

1948). The papain digest meat medium was preparedaccording to Ainsworth, Brown, Marsden, Smith & Spils-bury (1947). A semi-synthetic medium was prepared fromhydrolysed casein (McIlwain & Hughes, 1944) as describedby Boyd, Logan & Tytell (1947).

Mea8ureqment ofenzyme activity. In general, the activity ofthe glutaminase was estimated by determination of the rateof ammonia formation. A fresh solution (0-5 ml.) of glut-amine (0-02m in 0-25M-sodium acetate buffer containing0-025m-KCl) was placed in one arm of a branched test tubemade from 20 mm. diameter Pyrex tubing in the form of aninverted Y. The other arm contained 1-5 ml. of the enzymesolution in acetate buffer (final concn. 0-2M) and KCI(final concn. 0-025m). A series ofparallel tubes was placed ina water bath maintained at 40.00, and after 1 min. equilibra-tion the contents of the two arms were mixed without re-moving the tubes from the bath. At 5 min. intervals thetubes were removed from the bath and the reaction stoppedby placing in ice water and adding 0-5 ml. N-H,SO4. NH3was determined according to Parnas. Blank NH, determi-nations were made on all reagents. In this way a time curveof the glutaminase activity was obtained. The initial rate ofNH3formation was linear in intact cells and in crude extractsof Cl. welchii until approximately 50-60% of the added224 pl. of glutamine was decomposed, except where theglutaminase activity was low, i.e. where less than 5% of thesubstrate was decomposed in 15 min. The enzyme dilutionwas therefore adjusted so that not more than 60% and notless than 10% of the glutamine was decomposed in 15 min.Under these conditions the initial rate of reaction wasproportional to the dilution of the enzyme. Duplicate