Purification and properties of hemoprotein 559 in bovine heart microsomes

ARCHIVES

OF

BIOCHEMISTRY

Purification

and

AND

BIOPHYSICS

682-692

118,

Properties

(1967)

of Hemoprotein

559

in Bovine

Heart

Microsomes HITOSHI Laboratory

of Biochemistry,

SHICHI

Department

AND YOICHIRO

of Agricultural Received

Chemistry,

November

KURODA University

of Tokyo,

Tokyo,

Japan

14, 1966

From the bovine heart microsomes, a CO-binding hemoprotein has been solubilieed with the aid of trypsin and extensively purified by fractionation with ammonium sulfate, chromatography on Sephadex, and electrophoresis on starch. The purified sample that migrates as a single band in electrophoresis at pH 9.5 and 12 possesses an extremely low peroxidase activity and contains a small amount of phosphorus. The pigment, when reduced, absorbs light at 423, 530, and 559 rnp at pH 9.5. The reduced pigment which combines with carbon monoxide is rapidly autoxidizable. Based on the or-absorption band of the ferrous spectrum at pH 9.5, the pigment was named hemoprotein 559. At pH 12, the a-band shifts to 557 rnF with a concomitant increase in absorbancy. The pa-dependent hyperchromism has been studied and interpreted as arising from the increased electronic transition probability in the porphyrin T-system.

In a preliminary communication (l), we have briefly described that bovine heart microsomes contain, as hemoprotein component’s, lipid-bound myoglobin and a hemoprotein-absorbing light at 559 rnp in the reduced form (hemoprotein 559). The myoglobin is extracted with deoxycholate from the microsomes, and when the residue thus treated is digested with trypsin, hemoprotein 559 can be obtained in the soluble form. The hemoprotein, when reduced, combines with CO and resembles in various properties a CO-binding pigment isolated from liver microsomes (2). In the present report, extensive purification and properties of hemoprotein 559 are described. The intensity of the a-band of the pigment in t’he reduced form increases markedly in alkaline medium, and a possible mechanism for this hyperchromism is discussed. MATERIALS

AND

METHODS

ISOLATION AND PURIFICATION HEMOPROTEIN 559

OF

Step i . Preparation of trypsin digest. The bovine heart microsomal particles prepared by the acid682

sedimenting method described previously (3) were suspended in 0.3yo sodium deoxycholate (Tokyo Kasei Kogyo Co.) dissolved in 0.1 M phosphate buffer, pH 7.5, so that the protein concent,ration of the resultant suspension became 2-3 mg/ml; the suspension homogenized by hand in a PotterElvejhem homogenizer and incubated with gentle shaking at 30” for 20 minutes. The suspension was then centrifuged at 20,OOOg for 20 minutes and the supernatant fluid was discarded. The deoxycholate treatment effectively removes the microsome-bound myoglobin. The sediment was treated likewise with deoxycholate three more times and finally washed with 0.1 M phosphate buffer, pH 7.5. The residue was then suspended in the buffer and digest,ed with crystalline trypsin (1 mg trypsin/l&15 mg protein; Nutritional Biochemicals) with shaking for 3 hour at 30”. The solution was centrifuged at 20,OOOg for 20 minutes and the supernatant portion containing hemoprotein 559 was stored at 3”. Step 2. Ammonium sulfate fractionation. The tryspin digest was adjusted to pH 9.5 with 1 N NaOH and brought to 30% saturation by the addition of solid ammonium sulfate; throughout the procedure of salting-out, the pH of the solution was kept at 9.5 by the addition of aqueous ammonia. The solution was allowed to stand for 20 minutes at 3” and spun at 15,000g for 10 minutes. The precipitate was then dissolved in 0.1 M

glycine buffer, pH 9.5 (pH 9.5 butl’er) and the solution was again brought to 304;) ammonium sulfate saturation. After 4 cycles of the precipitation procedure with ammonium sulfate, the sediment was finally dissolved in aborti. 10 ml of pH 9.5 buffer (10-15 mg protein/ml). Step 5. Purification on Sephadex columns. The sample from Step 2 was loaded on a colurJJrl (3.0 X 25 cm) of Sephadex gel (G200, Pharmacia) previously equilibrated with pH 9.5 buffer. When the column was elated with the same buffer, a reddishbrown band moved gradually down and left a greenish-brown band above it. The fast-moving fraction was collected and concentrated in a dialysis tubing OJL pellets of polyethylene glycol 4NKl (Dainippon Seiyaku). The concentrated solution (about 5 ml; 10-15 mg protein/ml) was rechromatographed on a Sephadex (G-200) column (1.5 X 25 cm) freshly prepared, and the main eluate was concentrated in the same manuer. Step 4. F?rrther purijication. Abollt 350 gm of potato starch (Wako Chemicals) was suspended in 700 ml of pH 9.5 buffer and allowed to stand for 20 minutes, and the sllpernatant port.ion was removed by decantation. Starch gralrldes washed several times in this manner were suspended in bufier and poured into a vessel to form a rectangular slab (5 X 3 X 47 cm). The starch block containing approximately 5OC;, moisture was left) to stand for 3 hours at, 3” in order to thoroltghly equilibrate it with pH 9.5 buffer. The eluate from the second Hephadex chromatography was packed in the center of t,he slab, and electrophoresis (300 V, 10 mA) was carried ollt at 3” for 20 hours in pH 9.5 buffer. After electrophoresis, the pigment which had migrated as a single band about, 10 cm toward the anode was scraped out and eluted with pH 9.5 buffer. Since the purified sample often carried a small amount. of st,arch which resisted sedimentation by centriftlgation at 2O,OOOg, it was passed through a Sephadex (G-200) column and concentrated in a dialysis tubing placed on polyethylene glycol.

Spectral examinations. A Cary recording spectrophotometer model 14 was used for recording visible and ult,raviolet spectra. Cells with an optical path of 1.0 cm were used. A Kohken infrared spectrophotometer model DS 301 was used for recording infrared spect,ra in KBr pellets. Measurement of peroxidase activity. Peroxidase activity was measured by the increase in absorbancy at 460 rnp with o-dianisidine as substrate (4). Protein was assayed by means of the FolinCiocalteu reagent (5) and heme was determined by preparing the pyridine hemochrome (6) and on the basis of brnM, jj7 = 33.1. The activity was

expressed in terms of the increase of ahsorbancy at 460 m~./minut~e/~mole of hemc. Determination of phosphorus. The prlrified sample (1-2 mg protein) was digested with GO?;, potassium perchlorate, and the determination of phosphorus was made by the method of Allen modified by xakamura (7). ~‘ltt~~centr~~~uyal ann1ysi.s. Irltr:rceJltrifllg:~tion was carried out with Schlieren optics at, 51,200 rpm at 13” in a Hitachi analytical ltlt,racentrifrlge, model UCA-1. The protein concentration was 3040 mg/ml . -1mino acid nnal!/sis. Amino acid analysis was made on the acid hydrolyzate of purified sample which had been hydrolyzed in 6 x HCl for 24 hours; a rapid amino arid analyzer (Shibata Kagakll Kikai, Tokyo) was used. The tryptophan content was estimated l)y a calorimetric method (8) and cysteine content was determined by the niethod of Boyer (9) essentially in the same maimer as described previollsly (10). The dislllfide linkages were redllced with xaBH4 (11) and t,it rated with p-cahloromercuribenzoate. Electrophoresis on poluacrylnmide gel. A 546 polyacrylamide gel was prepared according to t,he method described by Iteisfeld ct al. 112) in 0.1 M glycine bufler, pH 9.5, ill a long rectangular tubing (20 X 2 X 0.3 cm). When the effect of urea ou hemoprotein 559 was stlldied, acrylamide was mixed with 4 M llrea prior to polymerization. The trlbing was fixed in a vert.ical position and a small amolmt of sample was placed on the gel, on which 0.05 M glycind blltfer, pH 9.5, was gently imposed. The top of the tubing was connected to the cathode and the bottom to the anode, as the hemoprotein was known to migrate t,oward t,he anode. Electrophoresis (200 V, 15 JILkj was effected for 3 hollrs at, 3”. Measurement of optical rotutor,y dispersion. A JASCO optical rotatory dispersion recorder, model OIID/UV-S, was used for recording optical rotatory dispersion. Most. measllrements were made in optical quartz cells with an optical path of 1.00 cm at room temperat,ure (15-20”) under a dim light to avoid stray light. Cells with an optical path of 0.10, 0.20, or 0.50 cm were also used. The results obtained with the recording instrument were checked with the more accurate mannual unit,. The hemoprotein concentration was 4.1 ELM at t,he wavelength limits of 300-600 mp, and it was 0.7 PM in the 19&300 rnp region. The data were analyzed by the Moffitt eqllation (13): [m’]~ = [3M&OO(n2 + 2)].[a]h = a&?/(h2 - X0*) + bo~04(~Z - ho2)2. Due to the strong Cotton effect in the visible region, the graphic determination of the iClo!fit,t constant b. was limited to the wavelength range of 240-300 mr. The value of 216 was used for the parameter X,1 (14). For the calculation

SHICHI

684

AND

KURODA

TABLE PURIFICATION

Purification

OF HEMOPROTEIN

559

9GO GM OF BOVINE

Total volume (ml)

steo

Microsomal residue after deoxycholate ment Trypsin digest (t30% Ammonium sulfate saturation 1st precipitate 2nd precipitate 3rd precipitate 4th precipitate Eluate from 1st Sephadex column 2nd Sephadex column After starch zone eleotrophoresis at pH 9.5 at pH 12

I

FROM

treat-

HEART

MUSCLE

Total &:;,

AL581 at pH 9.5

Am

200

11.04

450

1.27

40.8

0.04

114

8.72

446

7.94

51.2

0.05

52.8 53.9 55.1 10.6

7.63 G.63 6.36 6.14

191 164 150 133

6.17 5.08 4.77 3.36

25.0 24.9 23.6 21.6

0.11 0.12 0.14 0.13

17.4 10.0

2.40 1.44

2.50 2.34

24.4 20.3

0.14 0.15

17.1 5.1

0.402 0.087

2.12 2.07

23.8 22.5

0.17

of the effective residual rotation [m’]~ values, the molecular weight of the hemoprotein was set to be 21,000, which is actually a minimum molecular weight per heme, and the mean residue molecular weight M. was assumed to be 105 from the amino acid composition per mole of heme. The pH of the sample was adjusted as desired with 0.1 N NaOH.

58.7 29.2 9.42 2.82

3ocI-

RESULTS

For the purpose of solubiliz559 from the heart microsomespretreated with deoxycholate, trypsin is the only effective agent so far available; snake Habu (Trimeresures riukiuanua) venom, wheat germ lipase, steapsin, and pronase gave little or no solubilization. The effectiveness of trypsin was unaffected by the presence or absenceof air in the incubation flask. Table I shows the typical results of stepwise purification of hemoprotein 559. The high peroxidase activity found in the crude digest decreased markedly after further purification. Repeated precipitation of hemoprotein 559 with ammonium sulfate at 30 % saturation and subsequent Sephadex chromatography were particularly effective in removing the activity (Table I and Fig. 1). When a major fraction of the enzymic activity was precipitated from the trypsin digest at 3CMO % ammonium sulfate saturation, purified on a DEAE-cellulose column (equilibrated with 0.1 M phosphate buffer, Pur$kation. ing hemoprotein

100

Fraction

Number

FIG. 1. Separation of hemoprotein 559 from the contaminating peroxidase fraction on Sephadex.

pH 7.5) and examined spectrally as the pyridine hemochrome, it was found that denatured cytochrome c, probably derived from contaminating mitochondrial fragments, was largely responsible for the peroxidase activity. The highly purified sample migrated in electrophoresis as a single band toward the anode at pH 9.5 as well as pH 12. When such a preparation was spun at 51,200 rpm in an ultracentrifuge at’ pH 9.5, it sedimented apparently as a single boundary, but the peak diminished rather quickly due to diffusion. The reason for the high diffusion rate of the protein is not clear.

HE41:T

RIICILOSO~IAL

Abswption spectra. IJltraviolet and visible absorption spectra (at 20”) of the most purified sample at] pH 9.5 and 12 are shown in Fig. 2. The oxidized form shows a Soret absorption at arou1~1 400 rnp and a shoulder at 360 m/l. The weak band at about 600~GO mp in t)hc pH 9.5 spectrum may indicate that the ionic stntc prcdominnt)es in the ferric form (15). The height, and posit,ion of the O(,p, and Sorct bands are not significantly aff&cd by pH. However, the decrease in intensity of the band beyond 600 mp together wit#h the rise in absorbawy at NO600 mp which is observed at pH 12 suggests increased amounts of the covalent state at high pH. In the reduced form, absorption bands are found at’ 423, 530, and 559 mp. The wband, seen at 359 mp at, pH 9..i, shifts t,o 557 mp with a concomitant increase in molar absorbancy at alkaline pH values. The absence of a weak band at 600-650 mp in the reduced form spect,ra suggests that the ferrous complex is largely in t,he covalent, st,ate regardless of pH. E’igurc 13shows low temperature (-190’) spec+ra of the pigment in 0.1 nr glycine buffer, pH 9.5. The reduced a-band shows no sign of splitting

685

HEhIOPROTEIK 0.6423 “I

05-

----

’I : I :

> 0.4

Oxidized Reduced

I

;

4 $A-_,; I I I I I / 1 ( I / I I I / 1, I I / I / I I / / 400

450

500 Wavelength

550

600

Cm/i)

FIG. 3. Absorption

spectra of huemoprotein The spect,ra were taken in 0.1 M glycine bu8er, pH 9.5. 559

at -190”.

418 ri

I

a

150-

OxidizedtKCN

(PH9.5)

--------

Oxidized+KCN (pH Reduced+CO

I

2 2 -

---

12)

(PH9.5)

---

ReducedtCO (PH

121

120 a x

---

Reduced

(pH 12 1

Wavelength

(m/l)

FIG. 4. Absorption spectra of CO- and cyanidecomplexes at 20”. To obtain the CO-spectrum, a stream of commercial CO was bubbled through the reduced sample. For the cyanide spectrum, a few grains of KCS were added to the oxidized form of the pigment.

300

400 Wavelength

FIG. 2. Absorption at 20”. The hemoprotein hydrosulfite.

500

600 (my)

spectra of haemoprotein 559 was reduced with sodium

at low temperature but it shifts to 556 nip. This hemoprotein combines wit’11 carbon monoxide in the reduced form and wit.h potassium cyanide in t#he oxidized form, showing characteristic absorption bands, respectively (Fig. 4). The pa-dependent hyperchromism is again observed at the Soret band of the CO spect,rum at, pH 12. The molar extinct,ion coefficients of the

686

SHICHI

AND

main absorption bands of these spectra are shown in Table II; the molar extinction coefficients are based on the absorbancies exhibit’ed by 21,000 gm of protein/l, which is a minimum molecular weight per heme (see below). The hemoprotein is reduced by hydrosulfite or borohydride but not by NADH, ascorbat#e, L-adrenaline, hydroquinone, or cysteine. The pigment was not reduced by heart microsomal particles and NADH under either aerobic or anaerobic conditions. The reduced pigment is rapidly

KURODA

oxidized by air or ferricyanide. The infrared spectrum of the pigment is shown in Pig. t5. The presence of amide linkages (1650 and 1546 cm-l) and covalent phosphate groups (1230 and 1070 cm-l) are detected. The most purified sample contained, on the average of three determinations, 173 pg of phosphorus/pmole of heme. The presence of phosphorus in the purified sample may be due t,o phospholipid or nucleic acid possibly associated with bhe protein. Amino acid composition and molecular

TABLE MYLAR

EXTINCTION

COEFFICENTS DERIVATIVES

OF MAIN AT DIFFERENT

BANDS (I =

OF HEMOPROTEIN

1.0 cm,

pH 9.5 -

Form

Oxidized

pH

abs. max. (m,~)

aM (X 103)

610a 5308 400 360*

5.64 9.47 G9.5 59.1

266

88.4 __-___-

_____ Oxidized

II

ABSORPTION PH VALUES

+ KCN

Reduced

abs. max.

(m/4

559

AND

Ratio

of CM.,,” ‘M.*A 9.6

12 rid (X

109

560a 394 360* 290* 266*

9.09 69.4 59.1 68.8 97.7 ____

0.96 1.00 1.00 1.10

543 419 361

15.7 82.8 52.5

540 419 361

11.7 75.6 49.7

0.75 0.91 0.95

559 530 423

13.7 9.58 89.5

557 528 422

25.0 14.8 133

1.83 1.55 1.49

567 537 418

14.1 15.1 168

Reduced

+ CO

568 538 424

11.8 12.5 135

Q Hump * Shoulder

Wavenumber FIG.

5. Infrared

spectrum

(cm-l)

of haemoprotein

ITS

20”)

559 at 20”

1.20 1.21 1.25

LZ/

687 TABLE AMINO ACILI COMPORITIOS Amino acid

Asp Thr Ser c:1n I’ro Gl?; Ah \‘.a1 Met Ileu Lf?U TS’r Phe His LYS Arg Try* Cys SH* Cysh Total

III

,mole foundU

0.0884G O.OflG7 0.06849 0.08990 O.O-t562 0.09988 0.09160 0.08321 0.0272G O.OGG3G 0.12408 0.04524 0. oG394 0.02977 0.01484 0.03115

Mole ratio

16 11 13 16 8 18 17 15 5 12 23 8 12 5 8 0 2 4 1 200

c -

559

OF HEMOPROTEIN

w

1.6

Before reduction After reduction

3.vo,dg = 6.3

1.4 1.2 -

I mmoles

2

3

p-chloromercuribenzoate

4 ~104

6. Titration of sulfhydryl groups in hemoprot,ein 559 with p-chloromercuribenzoate. FIG.

was not based on extrapolation to zero time hydrolysis and the sample hydrolyzed contained the heme group which is known to o Analysis was made on 0.05 ml of acid hydrolycat,alyze the destruction of sensitive amino zate (0.14 mg protein). The numbers are average acids like serine and threonine. When the values of two experiments. heme content was determined on three h See text for determinat,ion of trypt,ophan, cystein, and cystine. samplespreviously dried to constant weight at 110”, 0.045-0.050 pmole of heme was weight. Table III shows t,he results of the found per milligram of the dried sample, quant)itative amino acid analyses. The indicating a minimum molecular weight, of amounts of the different amino acids, ex- 20,000~22,000. pressedas mole rat,ios, are estimat’ed assumStability. The pigment is stable at 15” for ing 5 moles of methionine per mole of heme. weeks as long as bact’erial growth is preThe tryptophan content, was calorimetrically vented. Freezing and thawing has no effect assessedt,o be 1.9 mole/mole of heme. Figure on spectra but leads to structural deteriora6 shows t,he results of the titration of the tion; a sample thus treated no longer sedipigment) with p-chloromercuribenzoate; 3.9 ments in ultracentrifuge at pH 9.5 as a moles of reactive sulfhydryl groups per single boundary. Because of the unstability mole of heme could be detected in the urea- lyophilization could not be employed for t,he denatured sample, while 6.3 sulfhydryl purpose of concentrating or storing the groups could be tit,rated in the urea-de- hemoprotein. Prolonged exposure of the natured and reduced pigment. The results pigment, to extremely acidic pH values results in irreversible aggregat’ion. Warmindicate that there are four sulfhydryl linkages and one disulfide linkage in the ing at 70” for 10 minutes neither affects the native protein. The heme group which can spect,ral propert,ies nor increases t,he peroxbe split off from t’he apoprotein by acid idase activit’y appreciably. acet,one was previously identified as protoEfect oj pH aggregation and hyperA solution of hemoprotein 559 heme IX (1). From t,heseresults, a minimum chromism. molecular weight of 22,600 was calculated becomes turbid in aridic medium. If the per mole of heme. This value should be increase in turbidity at various pH values taken only as an approximat’e, inasmuch as is followed by the increase in light, scattering the caalculation of amino acid composibion at G10 rnp, a wavelength at which light, 1.08245

SHICHI

688

AND

absorption of the oxidized pigment is least, affected by lowering the pH, a sigmoidal curve (pK’ = 6.1) is obtained (Fig. 7). The turbid solution clears as a rise in pH. The turbidity of the solut’ion causes considerable light scattering at all wavelengths which is responsible for the deformation of absorption spectra. However, the position of the 01- and p-bands was little affected: the Soret band appeared to shift a few millimicrons toward the shorter wavelengt’h side as the pH was lowered. Although evidence is still limited, it is very probable that aggregation of the protein is largely responsible for the optical changes in acidic medium. Figure 8 represents difference (reduced vs. oxidized) spectra recorded at alkaline pH values. The intensities of the (Y-and P-bands of the spectra gradually increase as t’he rise in pH, unt’il maximum values are reached near pH 12. Lowering the pH results in an immediate decrease in intensity, and it can be reversibly repeated many times. Isosbestic points were found at 503 and 620 rnp in the difference spectra. Molar extinction coefficients of the a-bands in the spectra are 10.4 X lo3 at pH 9.5 and 15.9 X lo3 at pH 12. As the pH is raised from 9.5 to 12.2, the CYabsorption band shifts gradually from 559 to 557 rnp. It is noted that bhe band has a symmetrical shape at all pH values, and the change in the band-width is lessmarked 0.40

r

03 h

P” FIG.

acidic M.

7. Aggregation pH. The protein

of haemoprotein 559 at concentration was 2 X 10-S

KURODA

I

I

I

,

I

500

I

I

FIG.

8. Difference pH values

I

I

I

600

Wavelength

at various

1 I

550

(m/A

spectra at 20”.

of haemoprotein

559

compared with the height. The same pH effect was observed, irrespective of the use of either dilute aqueous NaOH or various buffers (glycine, Tris, Versene, phosphate) for adjusting the pH. If the degree of ionization (CZ)is assumed to be zero at pH 8 and 1.0 at pH 12, and if it is related to the relative absorbancy of the a-band in such a way that (A, - AE&,~IZ - (Aa - AdVH8 = 1.0, the ionization constant (pK,) of the ionizable group will be given by the following equation: pH = pK, + log a/(1 - a). By substituting experimentally found values of o( at different pH values in the equation, the pK, was calculated to be 10.5; the theoretical curve with this value of pK, is drawn on t’he pH-muplots in Fig. 8. In the range of pH 7-12, there was litt,le or no increase in peroxidase activity. For comparison, similar experiments were performed with microsomal suspensions. Microsomes prepared by sedimentation at 100,000 g in sucrose-phosphat,e (3) did not demon&ate the hyperchromism. However, the microsomes which had been exposed to acid (3) or treated with deoxycholate (1) showed the phenomenon. Purified horse-radish peroxidase (Boeringer-Mannheim) demonstrates a similar hyperchromism at t’he reduced a-band at extremely alkaline pH values (Fig. 9). In this case, however, little difference is observed in sedimentation patterns between pH 7 and 12. In order to examine semiquantitatively

HEART

RIICIZOSOMAL

pH 9.5 s_-

689

FIEMOl’I:OTEIN

0.08

pH I2

- -

0.07

0.6

0.5

0.4

t 2

0.3

ii d 0

ul 2

0.1

aI

I

.

.

400

.

.

.

.

.

.

.

.

450

I

500

9. Absorption

spectra

of horse-radish

t,he hyperchromism in hemoprot’ein 559, hypothetical a-bands of absolute spectra at pH 9.5 and 12 were separated (Fig. lo), assuming that a band has a bell shape with a center of symmet’ry at its maximum wavelength. Equat,ions for these curves, g = 13.7 x lo3 e-[(z - 559)/143]2 at pH 9.5 and y = 25 X lo3 e-[(z - 557)/154]2 at pH 12, were obtained by the met’hod of Drabkin (16). From these curves and f = 4.32 X 1O-gJ G dv, where v is wavenumber in cm-’ (17), oscillator strengths (f) were calculated to be 0.041 at pH 9.5 and 0.054 at pH 12. Rj’ect of urea. Urea (up to 8 RI) added to the pigment at pH 7 did not, affect the differ-

peroxidase

2

.

.

,

600

(my)

WAVELENOTH

FIG.

.

550

at, different

1.9 I95 I.9218 1.76 ,111,

pH

values

at 20”.

Wove number (crn~ll I7 1.67 XIV 2 1.9 185 ISZlS I,

III,,

I.16

1.7 1.67 XIO’ ,I

pH 9.5

x103

y~,3.7r,03~w)i

~~5x,ole-~~~

3-

EMT-

500

tion

550

600

FIG. 10. The graphic bands of the reduced

pH values.

500

550

Wovelength (“i’)

600

analysis of t,he or-absorphemoprotein at different

690

SHICHI

AND

ence spectrum, while at’ pH lo-12 it caused a diminution (by about 30%,) of the magnitude of hyperchromism at the reduced a-band. Urea alone failed to cause the hyperchromism at pH 9.5. During the purification procedure, the trypsin digest was sat,urated at 30% with ammonium sulfate to precipitate hemoprotein 559. When the precipitate was incubated in pH 9.5 buffer containing 8 M urea overnight at 3”, a major fraction of the original precipitate no longer became sedimentable at 30 % ammonium sulfate saturation. If the urea-treated protein was subjected to electrophoresis on a polyacrylamide gel, a number of diffuse bands were detected. When a purified sample was exposed to urea only during the period of electrophoresis on a polyacrylamide gel containing 4 M urea, a new fast-moving band start -Urea

A

1

I -6.Ocm -6.3cm

I -

+Urea

I

I

-9.8cmt band B e

electrophoresis FIG.

havior

11. Elect of urea on electrophoretic of hemoprotein 559 on polyacrylamide

200 FIG.

12. Optical

rotatory

250

begel.

300

dispersion

KURODA

band was detected (Band B in Fig. 11)) in addition to the band A that was observed also on a control gel devoid of urea. The alterations in solubility to ammonium sulfate and in electrophoretic mobility, both caused by urea, may be attributed to chemic*al modificat8ions of the protein molecule. Alternatively, it is also plausible that the protein is electrostatically cleaved by urea into smaller fragments. Treatments of the protein with sodium chloride (up to 1 N), heating (70”, 10 minutes), sonication (5 Kc, 3 minutes), or freeze-thawing, all failed t,o cause t’he hyperchromism at pH 9.5. pChloromercuribenzoate (added up to 200 moles/mole of heme) did not inhibit the pH-dependent increase of the a-absorption band. Optical rotatory dispersion. The opt,ical rotatory dispersion curves of oxidized hemoprotein Tj59 at) pH 9.5 and 12 are shown in Fig. 12. In the visible region, in addition to the marked Cotton effect at 400 rnp, anomalous dispersion is observed at 480 rnp, which increasesconsiderably at pH 12. The Cotton effect in the ultraviolet’ region appears to be relatively unaffected by pH. Figure 13 shows the dispersion plots of the pigment at the two pH values, calculated from the curves in Fig. 12. The dispersion plots are linear over the range between 240 and 300 mp. From t’he plots, the Moffitt constants b. were estimated to be -80 at bot,h pH

350 400 Wavelength

450 (rnr)

of hemoprotein

500

550

559 at pH

600 9.5

and

12.

MW < 200,000) at’ pH 9.5 suggests t,hat, the actual molecular weight of the protein will be several times greater than the minimum molecular weight. It, is therefore unlikely that the rather prolonged tryptic digestion of microsomes proceeded so exz I tensively as to produce a mixture of heme -q -200 peptides. The heart microsomnl particles G-400 lack an enzyme capable of catalyzing the E reduction of the pigment with SADH. This “-600 is in contrast t’o a CO-binding pigment’ and -800 cytochrome bg in liver microsomes, both of I I I I I I I I I, which are reduced by the microsomes with 0.2 0.4 0.6 0.8 1.0 Nt1l)H (2). Spectrally similar pigments are x10-4 (,l-J.\)-’ found in the microsomal fractions from FIG. 13. The dispersion plots of hemoprotein bovine adrenals (20, 21) and rabbit liver (2), 559 at pH 9.5 arld 12. and their parbicipation in hydroxylat,ion reactions in these microsomes has been sugvalues. If a mean bo value of -567 is as- gest,ed. The pigment that can bc solubilized sumed for a perfect helix (18), the fraction only by tryptic digestion of i‘n~icrosomes” of helicity H was calculated as 14%, from is probably derived from t’he membraneous H = ho/-567. structure of heart sarcoplasmic reticulum; attempts to identify a comparable pigment DISCUSSION in the soluble fraction were without success. A highly purified preparation shows only Of all visible bands of the reduced pigment, a trace of peroxidase activity (approxithe pH-dependent increase in intensity is mately l/lo4 of the a,ctivity of crystalline the most remarkable at the a-band. A highly horse-radish peroxidase). It is therefore not purified preparation of RHP (Rhorlospirillum justifiable to call t’his pigment a peroxidase rubrum hemoprotein) (22) and a CO-binding (1). It has been suggested that b-type cytopigment in liver microsomes (2) are known chromes are the hemoproteins which possess to demonstrat’c a similar pH-dependent acid-acetone removable protoheme IX as hyperchromism, but no explanat,ion has been the prosthet,ic group, absorb light at around provided in these cases. The bacterial hemo430, 530, and 560 rnp in the reduced form, protein appears to be disaggregated at, pH are autoxidizable, and show no peroxidase 11. A question is raised as to whether the acbivity (19). The spectral features and the hyperchromism is related to the molecular association of only an extremely low peroxfragmentation or not’. Urea treatment or idsse activity with this pigment, the presfreeze-thawing of hcmoprotein 559, which ence of acid-acetone extractable protoheme apparently cause Wurtural modifications IX as the prosthetic group, and the rapid of the pigment molecule, fail to induce the autoxidizability of the reduced pigment all hyperchromism. indicate that the hemoprotein is a b-type Horse-radish peroxidase, whose sedimentacytochrome. However, it differs from b-type tion patterns are unaffected at neutral and cytochromes in its capability of binding alkaline pH values, shows a similar pHwith carbon monoxide or cyanide. Because dependent increase in the reduced ol-absorpof t,his difference, we prefer to designate this tion. From these findings, we were inclined pigment hemoprotein 559 on the basis of t’he to consider t,he hyperchromism in hemoposition of the reduced a-band at pH 9.5. protein 559 not as the result of molecular h minimum molecular weight of 21,000 per modification on the whole, e.g., fragmentaheme was found for this protein. The fact tion, but rather as related t’o pa-dependent that the hemoprotein passes rather quickly changes in interactions between voordinatthrough a molecular sieve for large molecaules (Sephadex G-200; for molecules of ing groups from the protein and the heme

692

SHICHI

AND

iron at the prosthetic group. The increase in oscillator strength from 0.041 to 0.054 in going from pH 9.5 to 12 indicates an increase in transition probability of r-r* transition in the porphyrin system; the a-absorption band is known to arise from a second r-a* transition (15). The n-bonding property of the ferrous ion in the porphyrin ring will be strongly affected by the basicity (electronrichness) of extra ligands at t(he fifth or sixth position. It is predicted that, in hemoprotein 559, high hydroxy ion concentrations will increase the basicity of the extra ligands and enhance an overlap between d,(Fe++) and n(porphyrin) orbitals, consequently lowering the s-ground state and increasing the probability of a-a* transition. Hence, the increase in the a-absorption ensues. The 2-rnp shift of the band to a shorter wavelength at high pH is accounted for by the lowering of the energy level of the r-ground state (i.e., Nhc(l/X5~7 l/&J = 0.18 kcal/mole). Whether the increase in the orbital overlap is caused by ionization of a ligand or, alternatively, by an exchange of ligand remains yet to be studied. As for the ligand group(s) involved, amino acid residues such as lysine and tyrosine are suggested from the value of pK, = 10.5. Optical rotatory dispersion data indicate that the hemoprotein has a low helical content at alkaline pH values. REFERENCES 1. SHICHI, H., KAMIRYO, T., AND FUNAHASAI, S., Biochim. Biophys. dcta 99, 381 (1965). 2. OMURA, T., AND SATO, R., J. Biol. Chem. 239, 2370 (1964). 3. SHICHI, H., SUGIMURA, Y., AND FUNAHASHI, S., Biochim. Biophys. dcta 97, 483 (1965). 4. SHICHI, H., AND HACKETT, D. P., J. Biol. Chem. !Z37, 2955 (1962). 5. LOWRY, 0. H., ROSENBROUGH, N. J., FARR, A. L., AND RANDALL, R. J., J. Biol. Chem. 193, 265 (1951).

KURODA 6. SHICHI, H., AND HACKETT, D. P., J. Biol. Chem. 237, 2959 (1962). 7. NAKAMURA, M., Nippon Nogeikagaku Kaishi 24, 1 (1950). 8. SPIES, J. R., AND CHAMBERS, D. C., Anal. Chem. 21, 1249 (1949). 9. BOYER, P. D., J. Am. Chem. Sot. 76, 4331 (1954). 10. SHICHI, H., HACKETT, D. P., AND FUNATSU, G., J. Biol. Chem. !Z38, 1156 (1963). 11. MOORE, S., COLE, D., GUNDLACH, H. J., AND STEIN, W. H., in “Proceedings of the Fourth International Congress of Biochemistry” (0. Hoffmann-Ostenhof, ed.), Vol. 8, p. 52. Pergamon Press, London (1960). 12. REISFELD, R. A., LEWIS, U. J., AND WILLIAMS, D. E., Nature 196, 281 (1962). 13. MOFFITT, W., AND YOUNG, J. T., Proc. iVatl. dcad. Sci. li.S. 42, 596 (1956). 14. URNES, P. J., IMAHORI, K., AND DOTY, P., Proc. Natl. dcad. Sci. U. S. 47, 1635 (1961). 15. WILLIAMS, R. J. P., Chem. Rev. 66, 299 (1956). 16. DRABKIN, D. L., in “Haematin Enzymes” (J. E. Falk and R. Lemberg, eds.), p. 142. Pergamon Press, London (1961). 17. KOIZUMI, M., “Kokagaku Gairon (Photochemistry),” p. 92. Asakura, Tokyo (1963). 18. JIRGENSONS, B., J. Biol. Chem. 240, 1064 (1965). 19. PAUL, K. G., in “The Enzymes” (P. D. Boyer, H. Lardy and K. Myrbgck, eds.), second edition, Vol. III, Part B, p. 277. Academic Press, New York (1960). 20. RYAN, K., AND ENGEL, L. L., J. Biol. Chem. 226, 103 (1957). 21. ICHIKAWA, Y., AND YAMANO, T., Biochem. Biophys. Res. Commun. 20, 263 (1965). T., AND KAMEN, M. D., Biochim. 22. HORIO, Biophys. dcta 48, 266 (1961). ACKNOWLEDGMENTS We are gratefully indebted to Professor S. Funahashi for his interest in this work and to Professor T. Yamano of Osaka University for recording the low-temperature absorption spectra. We also wish to thank Mr. N. Nishizawa for carrying out amino acid analyses.