Biochimica et Biophysica Acta, 1080 (1991) 68-77 , 1991 Elsevier Science Publishers B.V. All rights reserved 0167-4838/91/$03.50 ADONIS 016748389100305G

68

BBAPRO 34018

Structural and functional features of Pseudomonas cytochrome c peroxidase Nils Ellfolk *, Marjaana R6nnberg * and Kaj Osterlund Department of Biochemistry, UniL'ersity o[ ttelsinki, Helsinki (Finland) (Received 28 March 1991)

Key words: Cytoehrome c peroxidase: Two-fragment complex: Heine-heine interaction; Enzyme mechanism; (P ae,,uginosa 9

The secondary structure of Pseudomonas cytochrome c peroxidase (ferrocytochrome c: hydrogen-peroxide oxidoreductase, EC 1.11.1.5) has been predicted from the established amino acid sequence of the enzyme using a Chou-Fasman-type algorithm. The amount of a-helicity thus obtained is in agreement with previously obtained results based on circular dichroic measurements at far IN. The two heme c moieties of the enzyme have earlier been shown to have widely different characteristics, e.g., the redox potentials of the heroes differ with about 600 mV, and carry out different functions in the enzyme molecule. The structural comparisons made in this study enlighten the observed functional differences. The first heme in the polypeptide chain, heine 1, has in its environment a folding pattern generally encountered in cytochromes. In the region of the sixth ligand, however, profound differences are noted. The cytochromal meth~onine has been replaced by a lysine with a concomitant lowering of redox-potential thus making peroxidatic activity possible. Around heme 2, extra amino acid residues have been added to the peroxidase as compared with Rhodospirillum molischianum cytochrome c 2 core structure in the 20's loop. After completion of the cytochromal fold around heine 2 an additional tail consisting of 25 residues is linked. This tail shows no stabilizing elements of secondary structure, but contains a strongly hydrophobic segment which suggests a possible membrane contact site of this extrinsic membrane protein. Heme 2 is concluded to have a cytochromal function in the molecule. To further elucidate the functional properties of the enzyme, a noncovalent two-fragment complex was produced by specific cleavage of the peroxidase by Pseudomonas elastase. The complex was studied with respect to its properties to the native enzyme. The two-fraknnent complex of Pseudomonas peroxidase retains the overall conformation of the native enzyme showing, however, no heme-heme interaction. Thus, a comparison of the properties of the native enzyme with those of the two-fragment complex permitted some conclusions to be drawn on the structure of the enzyme as well as the mechanism of heme-heme interaction. From the present results we conclude that the two distal heme surfaces in the peroxidase are oriented toward each other. This structural arrangement allows an inter-heme communication in the enzyme molecule and it also forms the structural basis for the enzyme mechanism. Tho ~tructural comparisons also give insight into the evolution of an ancestral cytochn, me c into an efficient peroxida~e that has a versatile control m e c h a n i s m interaction.

Introduction

Pseudomonas cytochrome c peroxidase (ferrocytochrome c-551:H20: oxidoreductase, EC 1.11.1.5) is an extrinsic membrane protein that catalyzes the per-

* Deceased on 12 October. 1990. * Present address: United Laboratories Ltd. P.O.B. 222, SE 00381, Helsinki,Finland. Correspondence: M. R6nnberg, United Laboratories Ltd., P.O.B. 222, SF-00381 Heisinki, Finland.

in h e m e - h e m e

oxidatic oxidation o1 reduced Pseudomonas c-type cytochromes and azurin. The enzyme contains two heine c moieties attached to a single polypeptide chain by covalent bonds [1,2]. The properties of the two heroes are different, the apparent midpoint-potentials at ptI 6 being +320 mV and - 3 3 0 mV, respectively [3]. Sequence studies have shown that the first heine (heme 1) is attached to the peptide chain by cysteine residues 51 and 54, histidine-55 being the proximal histidine. The second heme q.heme 2) is attached to the peptide chain by cysteines 177 and 180, histidine-181 bc!ng the proximal histidine in this case [4]. One of the hemes

69 has been concluded to have a histidine-residue c l o ~ to the home iron on its distal side [5] whereas the other has a methionine residue as the sixth ligand [5-7]. The enzyme as prepared (i.e. the resting enzyme) has both heroes in the ferric form and has no catalytic activity. The half-reduced enzyme, with one heine reduced and the other heine oxidized, is catalytically active [3,8]. A heine-heine interaction has been observed to occur between the two heroes and to be essential to the peroxidatic activity of the enzyme [5]. Heine-heine interaction has also been observed in the cytochrome c peroxidase isolated from P. stutzeri [9]. Somewhat different characteristics of Pseudomonas peroxidase have been reported, however, with a preparation showing no heme-heme interaction [10]. In th~s study, the secondary structure of Pseudornonas cytochrome c peroxidase was predicted from the established sequence. Assignments were derived from the results of sequence comparisons with the known tertiary structures of Pseudomonas cytochrome c-551 [11], tuna cytochrome c [12] and that of yeast cytochrome c peroxidase [13]. Of particular interest are regions around the "~,,o home groups which could explain their cytochromai or pe[oxidatic properties, the positions of possible insertions, the surface properties that could explain the substrate specificity and regions involved in the protein-membrane attachment. Additional information on the structure-function relationship of the enzyme was obtained from studies of its two-fragment complex [14,]5]. Complexes consisting of two fragments have been effectively used to elucidate the correlation between protein structure and biological activity. The classical studies of the complexes of pancreatic ribonuclease and staphylococcal nuclease have provided valuable infi)rmation on the function of mese enzymes [16-20[. Similarly informative are the complex studies on cytochrome c [21-24].

acetone-dried cells of P. aeruginosa according to Ambler [27] and Ambler and Brown [28], respectively. Reduced electron donors were obtained by reduction with solid dithionite and excess reductant was removed by dialysis. Analytical SDS-polyacrylamide gel electrophoresis was performed according to Weber and Osborn in 10% gels [29]. Standard proteins were used for molecular weight estimation as previously described [14]. The cyanogen bromide fragment (CB 4) containing the high-potential home c was prepared and isolated as recently described [15]. It was dissolved in 0.1 M acetic acid for spectroscopic measurements. Reaction velocities and absorption spectra were recorded on a CaD' 15 spectrophotometer thermostated at 25°C. Circular diehroism measurements were recorded on a Cary 61 spectropolarimeter thermostated at 25°C. The primary structure of Pseudomonas cytochrome c peroxidase [4] and different cytochromes c [30] were compared using the FASTA algorithm of the GCG program package of the Genetics Computer Group of the University of Wisconsin. Predictions of secondary structures were carried out by the method of Chou and Fasman [31]. Assignments in ratio to the conformation of Pseudomonas cytochrome c peroxidase are derived from the results of the sequence comparisons with the known structure of tuna cytochrome c [12], Paracoccus cytochrome c-550 [32] and PsaMomonas cytochrome c-551 [11]. Hydrophilicity plots were computed employing the method of Kyte and Doolittle using 5-residue spans [33]. Results and Discussion

From the established sequence of Pseudomonas cytochrome c peroxidasc the secondary structure has been predicted. The two homes of the peroxidase are numbered 1 and 2 ;:ccording to their position in the polypeptide chain.

Materials and Methods Pseudomonas cytochrome c peroxidase w.'~s purified from acetone-dried cells of P. aerugmosa as previously described [1,25] The specific activity of the enzyme preparation was 120 U / m g measured as reported before [26]. The. two-fragment complex of an elastase diges~ of Pseudomonas cytochrome c pcroxidase was isolated as previously described [14,15]. The preparation conststed of fragments a and b formed after cleavage of the peptide chain at the Ser (202)-Val (203) bond and showed a peroxidatic activity of 6 U/rag. The half-reduced forms of the native o n , ' m e and the two-fragment complex were obtained using 1 mM .sodium ascorbate (Sigr~,a) as electron donor and 5 tzM phenazine methosulphate (PMS. Sigma) as a meuiamr. Cytochrome c-551 and azurin were purified from

The predicted s~conda~" structure of Pseudomonas ~'tochrorne c peroxidase Strblclltre around hcme I. A comparison was made with the sequence around heine 1 of the peroxidase and c31ochrome c from tuna as shown in Fig. 1. A c3"tochl'omal folding pattern around heine 1 seems dearly recognizable. An a-helix conformation is predicted in the regions of residues 2-12, 16-24, 26-40 and 45-50, Beyond the heine attachment site a-helical conformation is deduced to be present in the regions of residues 79-87, 100-108, 111-125 and 132-147. Detailed compacisons are hard to make starting from secondary structure predictions. Some general observations can. however, usually be made. With this in mind we tD' to correlate the secondary structure of Pseudomonas ~3'ro~chrome c peroxidase to that of cytochromcs c with established tertiary structure.

70

40

50

PaCCP hemc J

Tuna cyt

GDVA

K GKKTF

PaCCP heme 1

T[-~'],_N[~G T

Tuna cyt

Q

20 T__

Q K

Dt

E_N

A G

P 1"

K Jl

(;

30 F

70

~

80

PaCCP heine 1 Tuna cyt

LF

N LW-

RKTGQAE_

I

PaCCP heme 1 Tuna cyt

90 E Q A_ " _G P501 Q i ~ V GY S Y T_ D A N oL l

Q M H S K G I V W l

I--

110

100

PaCCP heme I

Tuna cyt

NNDTLME

ENP o~ 120

K !

heine 1 Tuna cyt

PaCCP heine I

I P _G_T K °_MOIIFIA G I KG o4 ! 130 140 S F D N MA L A I E A Y E A 90

100

Tuna cyt E R Q D L V A Y L K S A T S Fig. 1. Comparison of the sequence of heine 1 of Pseudomonas cytochrome c peroxidase(PaCCP) with that of tuna cytochrome c. Alignmentis made by heine bindingsites. Identicalsequences are boxed. Underlinedresidues refer to conservativesubstitutions.Predicted a-helicesare denotedby a.

Two of the predicted helices are in good agreement with those observed in X-ray crystallography of tuna cytochrome c. On the left hand side of tuna cytochrome c the proximal histidine-18 corresponds to histidine-55 in the peroxidase and proline-30 to Pro-66, respectively. In the cytochrome molecule the evolutionary conserved glycines involved in 3 m bends, Gly-23, Gly-34 and Gly-37, correspond to Gly-60, Gly-72 and Gly-75 in the sequence around heine 1 of the peroxidase. In the position corresponding to the sixth ligand of the heine iron, methionine-80, in tuna cytochrome, there is a iysine residue (Lys-ll8) in the sequence around heine 1. This position was confirmed by phenyl-

alanine 120 corresponding to the invariant Phe-82 in tuna, tyrosine at 112 corresponding to the invariant Tyr-74 in tuna and Lys-125 corresponding to Lys-87 in tuna. Methionine can be replaced by lysine as a consequence of a one-point mutation. Such a change in the vicinity of the home is bound to cause alterations in its function. In yeast cytochrome c perorddase, an arginine residue (Arg-48), situated near the heme pocket with side chain above the distal face of the heme, has a charge stabilizing property which seems to be essential for the peroxidatic function [34,35]. In Pseudomonas peroxidase, a corresponding arginine residue is situated in position 117 on the distal side of heme 1. This arginine residue could have a charge stabilizing property such as has been proposed in :he yeast enzyme. In the Pseudomonas enzym=, Lys-ll8 could assist in a similar function. A distinct differen, c in the heme I structure of the peroxidase with respect to cytochrome c is the lack of tryptophan which in high-potential cytochromes stabilizes the heme crevice structure. EPR otudies show that a distal histidine residue is involved with the low-potential heme [5]. From the sequence it can be concluded that only two histidine residues (240 and 276) can be considered for this role. By structural analogy with cytochromes c from different Pseudomonads it is suggested that the imidazol¢ ring of histidine 240 is held in a fixed position through hydrogen bonding to the inner propionic acid of the high-potential heme (heme 2, see below). NMR studies indicate that the two heroes are physically close to each other [36], and it can therefore be speculated that the E-nitrogen of His-240 could perhaps function as the sixth ligand to the low-potential heme, thereby providing the electronic linkage suggested by NMR data. On the other hand, a comparison with yeast cytochrome c peroxidase suggests that His-276 could also be in stericai proximity in this region and, therefore, cannot be excluded as a ligmd. Resonance Roman [6], magnetic susceptibility and NMR measurements [36] show that the low-potential heme exists in a Iow-~pin/high-spin equilibrium. A ferric high-spin heme is either pentacoordinated or coordinated with a weak field iigand like water or a carboxyl group. Resonance Raman spectra indicate a carboxyl group to be involved with the high-spin form of the low-potential heme, a water molecule being less likely according to the magnetic measurements. As Lys-I18, according to sequence comparison (Fig. 1), is in the position of the sixth ligand, the closest carboxyl group along the sequence is offered by aspartic acid in position 114, other dicarboxylic acids being more distant. The prediction of the secondary structure indicates Asp-ll4 to be situated in an a-helical structure and thus should be close to the heme iron having the

71 same position as Glu4367 in MHb Milwaukee [37] and Asp-f167 in MHb Bristol [38]. X-ray crystallography studies show that in MHb Milwaukee Glu-/]67 is linked directly to the heme iron forming an ion pair which stabilizes the ferric state. As in MHb Milwaukee, the iron of the high-spin form of the low-potential heme is suggested to be in an out-of-plane position. In circular dichroic (CD) measurements the oxidized enzyme shows a positive maximum at 400 nm. This ellipticity is doubled in the half-reduced enzyme which indicates the occurrence of an out-of-plane iron [39]. Thus, the out-of-plane high-spin iron of the low-potential heme is concluded to form the peroxidatically active state of the enzyme. Effects of ionizable amino acids on the reduction potential of recombinant myoglobin has recently been studied [40]. The substitution in the fourth residue from the distal histidine by Glu or Asp decreases the redox-potential about 200 inV. Such an amino acid substitution near the heme group causes the ferric form to be unusually stable preventing a reduction to ferrous heme. In addition to such a change, the heme crevice of heine ! of the peroxidase contains six aromatic amino acid residues less than the heine crevice of mitochondrial cytochrome c (Table !) also indicating a decrease in the redox-potential [41]. From these comparisons it seems well established that heme 1 is the low-potential heme of the peroxidase. Structure around heme 2. A sequential comparison was made with cytochrome c 2 of Rhodospiriilum rnolischianum [30] and the sequence around heme 2 (Fig. 2). A noticeable difference in the two sequences is on the proximal side of the heme where extra amino acid

t 70 PaCCP herae 2 R.

iso-1

hemePaCCP2

190 C['I] L _G20~Q A Y F P F G L V K

I~. iso-1

- [2]] p A~(;I. . . . . . . . . . --

k,,=.~

I.--200

R.

iso-I

R.

- - K N G

G

L S 22O

PaCCP heme

210

_A V T K T Q S D E ~ V 2

iso-1

40 C R k__' g C L A P N o(

A Y

!- R A ~

W

K Y _S _P

t

_~230

240

PaCCP herae 2

P

R N V A L T A P Y F tl S Q

R.

Htda]A S" g M T I D ~3 a M l. T F

i 50

iso-1

O0

,~

q 250

heine

2

R. iso-1

11 Y L A N P

1!] E voT

I P G S K M G

260

270

berne 2

R.

iso-!

. ~A 8A0 __

L K

.

.

.

AtD V~A V A ___ _

TABLE I Hydrophohw and aromatic amino aod r,,~idues aro*:nd heine crevice in cTtochrome~ and their counterparts around hernc 1 and heine 2 in Pseudomonaa cytochrome c pc.rorlda,~e

Tuna

Phe- 10 Pro-30 Phe-36 Tyr-46 Tyr-48 Tyr-68 Leu-69 Trp-59 Phe-74 Met-80 Phe-82 Ile-85 ile-94 Tyr-97 Leu-98

Pieidommma c-551

Heme I

Phe-7 Pro-25 Phe-34 * * 11c-48 L~,~s-49 Trp-56 Val-55 MeI-60 Pro-63 Val-66 Leu-74 Trp-77 Val-78

His .47 Pro-66 Lys- 74 Gin-g4 Lys-86 Thr- 105 Leu- 106 Set- 108 Tyr- 112 Lys- 118 Phe- 121) Ala- 123 Met-132 Ala-135 lie- 136

* Deleted region in cyl(x:hromc c-551.

Heine 2

Phe-: 72 Pro- 21)5 Phe-212 Tyr-'r22 Pbe.224 Fyr-23~ Phe- 239 Trp-245 Leu-247 Met-254 Asn-256 Leu -259 11e-271 Phe-274 I.cu-275

280

p~ccp

v~_F,,

IlL-q!.,. ~ L- s ~; ~ Q P R v t~

R.

_I

U

heine 2 iso-1

i

i

Y

100 K T V K

Fig. 2. (Mmparison of the sequence of heine 2 (rc'stdues lh8- 2,'e~6)of Pseudem,zonas cytochrome c peroxidase IPaCCPJ with that of Rhodosplrillum molisctuanum c3,1ochrome c: i~o-I. Alignment is made by heme binding sites. Identical sequences are boxed. Underlined re~dues refer to conservaliv¢ substitutions. Predicted a-helices are denoted I-~,a

residues (188-199) have been added to the peroxidase as compared with the cytochrome c 2 core structure in the 20"s loop. Such an insertion was first observed in c3~tochrome c-550 of Paracoccus denitnficans [32] and later in cytochrome c 2 of Rhodopseudomonas sphaeroides and Rhodopseudomonas capsulata [30]. Because c 2 of Rhodospirilhan molischianum does not contain tryptophan : comparison was also made with

72 Pa('CP h~,mo 2 R.

tcnuJs

Pat(iP here 2

R. t e n u i s

22O b? F A V "i K I q S E~ I: ¥ V F R ~. ~ (; i' 4 ~

(; h

V :\ k

~ Y k

(i

A ~ P L R 5, \ A L T A P ~¢ F 'tt i ;0 30 (,', G A 5 A k L V A K V ),l A G

9c hemePaCCPT' -S~2ili L K I)A_\~A~- ~ k. ten' !~

~

IQ__ oo C, v w_a K ~ I__c, -3 i2

p

260 heine

2

70

-,

i ' i

-

0[

R. re,~,,is

P ; ~ '._."',_"T',, _L g

i I~ L M K

PaCCP

270 280 N i V A F L HS L S G KQ P R

heme 2 R. tenuis

9O ~' %i L S L K Q I D Y I