Cooperative Cyanide Dissociation from Ferrous ... - Semantic Scholar

THE

J O U H N A L O F BIOLOGICAL CHEMISTRY hy The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 267, No. 4, Issue of February 5 , pp. 2258-2263, 1992 Prrnted in U.S.A.

c 11992

Cooperative Cyanide Dissociation from Ferrous Hemoglobin* (Received for publication, July 29, 1991)

Maurizio BrunoriSB, Giovanni Antoninin,Massimo Castagnolall , and AndreaBellelliS From the SDipartimento di Scienze Biochimiche e Centro Biologia di Molecolare del Consiglio Nazionale delle Ricerche, Uniuersita’ La Sapienza, Piazza Aldo Mor05, 00185 Roma, the BDipartimento di Medicina Sperimentale e Scienze Biochimiche, Uniuersita’ Tor Vergata, Roma, and the11 Istituto di Chimica, Uniuersita’ Cattolica del Sacro Cuore, Roma, Italy

Rapid reduction of cyano-met hemoglobin (Hb’CN-) upon rapid reduction of the cyano-met derivative (Cox and leads to the formation of a n intermediate species, the Hollaway, 1977; Olivas et al., 1977; Bellelli et al., 1990). cyanide derivative of ferrous hemoglobin, which disThe spectroscopic and functional propertiesof the cyanide sociates to unliganded hemoglobin because of the ex- derivative of several ferrous myoglobins, including two sperm tremely low affinity of the ligand for the ferrous heme whale mutants obtained by site-directed mutagenesis, have iron. The properties of the intermediate were studied been recently published (Bellelli et al., 1990). We have shown by transient spectroscopy in humanhemoglobin and its that the nature of the residue on the distal side controls the isolated a and B chains, in the presence and absence of CO. When mixing with dithionite, the time courses of optical spectrum and thekinetic properties of the intermedireduction of the heme iron anddissociation of cyanide ate. In that paperwe suggested that CN- is the species bound overlap considerably; addition to the reaction mixture to theferrous iron, and thata proton can be transferred from of the redox indicator methyl viologen considerably the distal imidazolate,where present, to cyanide, to form increases the rate of reduction and allowsunequivocal hydrocyanic acid which rapidly dissociates from the complex. determination of the spectroscopic and kinetic prop- This mechanism,which is analogous to thatproposed for the dissociation of OH- from met myoglobin (Ilgenfritz and erties of the intermediate. The results show that(i) the dissociation of cyanide Schuster, 1971, Giacometti et al., 1975), may account for the fromthe isolated a and B chains (as well as the effect of pH on the rate of dissociation of cyanide in mono(aCO),(B’CN-), hybrid) is a simple process; (ii) the two meric hemoproteins which possess a His residue at position chains display similarrate parameters, but showspec- E7 (the distal His). troscopic inequivalence, bothin the Soret and the visThat work providedevidencefor the interaction of the ible regions; (iii) cooperative effectsare shown to con- charged ligand with the distal histidine butgave no explanatrol the rate of dissociation of cyanide from hemoglo- tion for the chemistryof the bindingof cyanide to theferrous bin, similarly to what happens for oxygen; and (iv) heme. Nonetheless, it is of general interest to find out if allosteric effectors (typically inositol hexaphosphate) cyanide is able to promote the quaternary conformational rate of dissociation by stabilization change which constitutes the physical basis of the allosteric increase the overall of the T state. We have, therefore, shown for the first time that the dissociation of cyanide from ferrous he- behavior of hemoglobin. One might indeed suspect that the weak binding of cyanide to the heme iron is governed by moglobin is controlled by the quaternary state, thereby adding one more ligand to the analysis of the structure- substantially different mechanisms as compared to oxygen and other ligands of ferrous Hb. On this point the “ancient” function relationships inhemoglobin. finding of Stitt andCoryell (1939), who reported that binding of this ligand appears tobe non-hyperbolic, is the only result available to date. It was demonstrated long ago that cyanide, besides binding As demonstrated in this paper, the rate of cyanide dissociato ferric hemoglobin, also binds to ferroushemoglobin, yield- tion from human hemoglobin indeed depends on its quatering a diamagnetic complex (Balthazard and Philippe, 1926; naryconformationand is sensitiveto allosteric effectors. Stitt and Coryell, 1939). As a ligand of ferrous hemoglobin, When thereduction was driven by dithionite alone, this result however, cyanide has unique properties, since it is the only was not immediately apparent because the properties of the known charged ligand, and its affinity for the ferrous heme tetramer are concealed by the spectroscopic and chemical iron is indeed very low (Kd= -1 M). It was suggested (Reise- inequivalence of a and p chains (which are reduced by dithiberg and Olson, 1980a) that itschemical state, once bound to onite at different rates). When the redox indicator methyl the heme iron, might be that of hydroisocyanic acid. viologen was used to obtain a faster reduction ofHb’CN-, The reaction of reduced myoglobin with cyanidewas studied the time overlap between the reduction of the heme iron and at equilibrium, using very high ligand concentration (0.5-1 the dissociation of cyanide was minimized (Cox, 1977), M ) and very alkaline pH (>9; see Keilin and Hartree, 1955), thereby allowing a totally unequivocal kinetic analysis. As or kinetically, under milder conditions, analyzing the time shown below, thetransient kinetic datacan be described course of cyanide release fromthe intermediatespecies formed quantitatively, and it is demonstrated that dissociation the of cyanide from ferrous hemoglobin is a cooperative process. * This work was partially supported by a grant from the Minister0 dell’universita’ e della Ricerca Scientifica e Tecnologica (40% “Proteins”)and Consiglio Nazionale delle Ricercbe Grant P.S. n. 91.02436.CT14. The costsof publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate thisfact. 0 To whom correspondence should be addressed.

MATERIALSANDMETHODS

Human hemoglobin, as the carbon monoxide derivative, was prepared using standard procedures (Antonini and Brunori, 1971). The 01 and /3 polypeptide chains were purified after reactionwith phydroxy-mercuribenzoate, as described by Bucci and Fronticelli (1965); separation wasachieved by ion-exchange chromatography

2258

Cooperative Cyanide Dissociation using a DEAE-Sepharose Fast Flow (Pharmacia, Sweden) column equilibrated with 0.02 M Tris buffer, pH 8.1, and eluted with a linear 0-0.3 M gradient of NaCl in the same buffer. The (I and p chains were freed from p-hydroxymercuribenzoateby 4h of incubation with 5 mM 6-mercaptoethanol under 1 atm of carbon monoxide andthe number of titratable -SH groups was checked by spectrophotometric titration (Boyer, 1954). Hemoglobin and its chains were converted into the cyano-met derivative by overnight incubation with three stoichiometric excesses of potassium ferricyanide and 1 mM KCN. The valency hybrid (aCO),(p'CN-), was prepared by mixing the desired derivatives of the two chains in stoichiometric amounts and chromatographic purification. The equilibrium of cyanide with reduced hemoglobin was studied spectroscopically using a gas tight sealed cuvette in the absence of gaseous phase; a few grains of sodium dithionite were added to keep the protein reduced and deoxygenated. Buffer was 0.1 M glycine, pH 9.6; temperature was 20 "C. The kinetics was followed using either a Gibson Durrum single wavelength stopped flow or a rapid scanning photodiode array spectrophotometer (Tracor NorthernTN6500) coupled to the observation chamber of a Gibson Durrum stopped flow apparatus; most experiments were carried out in 0.1 M phosphate buffer at pH 7 containing 1mM KCN. In a typical experiment, a 1-20 PM (in heme) hemoglobin solution was mixed with 25-200 mM dithionite, and optical spectra were collected at time intervals ranging from 10 ms to 120 s over a wavelength range of 150 or 300 nm. When the experiment is carried out in the presence of 0.5 mM carbon monoxide, the final species is a liganded derivative and, due to the relatively high rate ofCO combination (see Antonini and Brunori, 1971),hemoglobin remains liganded throughout the reaction time course; thus no allosteric transition is expected to occur, assuming that thecyanide derivative of ferrous Hb is in the R state. The two-step scheme (see also Bellelli et al., 1990) Hb' CN- + Hb CN-

"+

Hb

+ HCN

SCHEME 1 was used to give an approximate description of the observed time courses; a more refined kinetic scheme is presented under "Results" (Scheme 2). The spectra acquired in 10msover a logarithmic time scale spanning from 10 ms to 120 s were deconvoluted using the singular value decomposition algorithm (Press et al., 1986; Hofrichtr et al., 1983) and thesignificant columns of the matrix V analyzed according to Scheme 1, using the equations developed by Bateman (1910): Hb+ CN- = e l - k , . t )

Hb = 1 - ([Hb' CN-]

+ [Hb CN-I).

Transient spectra obtained under all experimental conditions were simultaneously fitted, imposing the same extinction coefficients and allowing the rate constants to vary. The reduction of cyano-met hemoglobin by dithionite is heterogeneous and not sufficiently faster than thedecay of the ferrous cyanide intermediate, yielding a considerable overlap between the two kinetic phases. The intermediate species is usually populated to less than 40% of the total pigment, and in some cases considerably less. The reliability of the rate constants obtained in the experiments carried out in the absence of methyl viologenwaslow, due to the high correlation with the extinction coefficients of the intermediate; this is a direct consequence of the low population of the intermediate. A much greater population of the intermediate, approaching 100% was obtained in the experiments carried out in the presence of dithionite and methyl viologen, a reducing agent capable of faster reaction with the heme iron (see Cox, 1977);under these conditions an unambiguous fit was obtained. As described by Keilin and Hartree (1955), the cyanide derivative of ferrous myoglobin is highly photosensitive. Therefore, caution must be applied that therate of cyanide dissociation is not affected by the monitoring beam. We have shown that a conventional single wavelength stopped flow apparatus yields arateconstant for cyanide dissociation approximately 2-2.5 times smaller than that recorded under full illumination of the sample (as intrinsic in the operation of the TN6500 photodiode array spectrophotometer). The conclusions

2259

1.2

I

wavelength FIG. 1. SVD kinetic analysisof (I chains (panel A), fl chains (panel B ) , the hybrid ((ICO),(fl'CN-)2 (panel C),a n d HbA (panel D ) . Experimental conditions: T = 20 "C, 0.1 M phosphate buffer, pH 7, containing 1 mM KCN; dithionite concentration = 50 mM, heme = 10 ~ L M(after mixing). The first two U and V columns, each multiplied by its weight S are reported, together with the spectra for the initial species, the final species and the mixture corresponding to the highest concentration of the intermediate. obtained using the photodiode array system are, however, unaffected, since it was shownindependently that theeffect of CO, pH, andIHP' is present and equivalent in magnitude also with the conventional single wavelength stopped flow. The high quantum yield of this derivative, confirmed by this finding, could not be explored with higher precision, due to thevery highcyanide concentration required to saturateferrous hemoglobin, even in the dark. RESULTSANDDISCUSSION

Spectroscopic Propertiesof Ferrous Hemoglobin Cyanide and Ligand Binding at Equilibrium-The affinity of ferrous human hemoglobin for cyanide is so low that the protein is modified before saturation is achieved because of the comand possibly bined effects of high pH, high ionic strength, heme extraction. At cyanide concentrations lower than 1.0 M and p H between 9.6 and 10, an incomplete titration is observed with isosbestic points at 575, 560, and 548 nm; the spectroscopic transition recorded is completely reversible upon dilution. These results do not warrant further analysis. Transient Spectroscopy of Cyanide Dissociation-The time course of dithionite reduction of the cyano-met derivative of the isolated CY and p chains and the valency hybrid ((YCO)~(P+CN-)~ is described by two significant components in t h e SVD analysis (see Fig. l),which could be fitted to Scheme 1. The two chains display significant spectral and kinetic differences, the overall rate of reaction being greater by a factor of 4 or more for the p chains. While the time course of t h e /3 chains and t h e (aCO)2(/3+CN-)2hybrid can be described by Scheme1with reasonable confidence, in the case of t h e CY chains the reduction is so slow that the population of the intermediate is very smalland the fittingis unreliable, due to the high correlation between the extinction coefficients of the intermediate and the two rate constants. When cyano-met hemoglobin is mixed with dithionite a complex time course is recorded at all wavelengths explored; indeed the reaction does not conform to Scheme 1. Under these conditions the SVD algorithm generates at least four significant componentsand requires three to four independent exponentials for a satisfactory fit. The results of t h e SVD The abbreviation used is: IHP, inositol hexaphosphate.

2260

Cooperative Cyanide Dissociation

chains (Fig. 4). It is of some relevance to recall that a small analysis of a typical set of experiments is reported in Fig. 1. If the experiment is carried out in thepresence of 0.5 mM blue shift in the spectrum of the CY chains has alsobeen CO, hemoglobin remains liganded (and presumably R state) reported forseveralisocyanidederivatives (Talbotet al., 1971). throughout the reaction time course, with no change in the The experimental data for hemoglobin and both chains allosteric state (like the isolated chains). The timecourse of the reaction is still complex but can be described using the recorded at selected wavelengths in the presence and in the same spectroscopic and kinetic parameters for the a chains, absence of methyl viologen were all fitted to Scheme 1. As the p chains, hemoglobin, and the hybrid, and assuming the shown by the parameters reported in Table I, the two data chains in the tetramer to react independently. By contrast sets differ only in the value of k l ; this demonstrates that no the same setof rate constants does not describe the reaction specific effect is tobe attributed to methylviologen per se on recorded in the absenceof CO, where in thecase of hemoglo- the rateof cyanide dissociation.More interestingly, the analybin the final species is deoxyHb in the T quaternary state. sis shows that in thepresence of CO, i.e. when the quaternary Unfortunately, the spectroscopic properties and reaction pa- state of hemoglobin does not change during the release of rameters for the ferrous cyanide derivative of the CY chains cyanide, the rate constants obtained with the isolated a and remained still undetermined, in spite of the constraints im- p chains are strictly comparable to those found from the fit posed on the fit by the simultaneous minimization of all the of hemoglobin data, whereas in theabsence of CO hemoglobin transient data. Therefore, the marked difference between the releases cyanide at a faster rate. Although the difference in the rate constants correspondsonly to a factor of 2, it is CY and p chains in the rate of reaction with dithionite causes the reaction between cyano-met hemoglobin and dithionite to diagnostic for cooperativity since the maximal effect observof appear complex and heterogeneous. The heterogeneity of able under these conditions is smaller than the number reduction, coupled with the spectroscopic difference between interactingsites (seeHopfield et al., 1971; Antoniniand the CY and fl chains (see below), fully explains the complex Brunori, 1971). Kinetic Evidences for Cooperativity-As reported above, a SVD pattern outlined above. The time course is considerably simplified by addition to first evidence for kinetic cooperativity lies in the observation the reaction mixture of methyl viologen, which accelerates that in the presence of CO the dissociation of cyanide is considerably the reduction of the heme iron (see above and slower (Fig. 3 and Table I). Met hemoglobin is alsoveryrapidlyreduced by methyl Cox, 1977). Experiments were therefore run in the presence viologen. Therefore, it ispossible to obtain partiallyliganded of 10 PM methyl viologen. As shown in Figs. 2 and 3, the intermediate is very rapidly populated, the difference in rate cyanide derivatives of ferrous hemoglobin simply by mixing between the two phases being greater than a factor of 50. with methyl viologen and dithionite a solution of met hemoUnder these conditions the ambiguities of the fit are elimi- globin incompletely saturated with cyanide. In this experinated and the correlation between the rate constants and the ment, which is analogous to theoxygen pulse as described by Gibson (1973), all the ferric hemes arerapidly reduced, and a extinction of the intermediate is very small. From the results obtained in thepresence of methyl violo- statistical distribution of cyanide liganded intermediates is gen, the spectrumof the pure ferrous heme-cyanide derivative obtained. The experiment reported in Fig. 5 shows that disis easily obtained and canbe compared with that of myoglo- sociation of the ligand from partially saturated Hb is faster bin(s) (Bellelli et al., 1990). A remarkable property of this than the sameprocess for fully saturated Hbby a factor of 5derivative lies in the spectroscopic differences between the 10. This experiment demonstrates thatcyanide binding stabihemoglobin chains, the peaks of the CY chains being consistently blue-shifted (by 3 nm) with respect to those of the (3 lizes the R state of hemoglobin, but, a t a closer look, the data

FIG. 2. Three-dimensional spectra of the reaction of cyano-met human hemoglobin with dithionite (panel A ) , or reduced methylviologen (panel8 B-D), in the presence of IHP (panel C) or IHP and CO (panel D ) . Experimental conditions: methyl viologen concentration = 10 @M (panels B-D); IHP concentration = 2

/

I Q.0

wlcrmth

100.0

400.0

wlcrmth

ioo.0

-

C

mM (panels c and D);co concentration = 0.5 mM (panel D);other conditions are as in Fig. 1.

I

4M.O

/

wlmth

100.0

am.O

wlmth

100.0

zide Dissociation Cooperative Cya~

2261

0.B 0.E

Y 0.4 0.2

time (s)

0.0 0.0

FIG. 3. Time courses at wavelength = 566 and 535 nm from the experiments of Fig. 2. Rate constants are as in Table I.

2.a

t,!I

L O

8.0

10.0

lime (s) FIG. 5. Time course of cyanide dissociation as observed after rapid reduction of partially cyanide-saturated metHb. Experimental conditions are as in Figs. 1 and 2, monitoring wavelength = 566 nm. The absorbancechanges were normalized to obtain the fraction of cyanide bound ferrous Hb ( Y ) as a function of time. Contrary to the experiments reported in Figs. 1 and 2, in this case

160

the conventional single wavelengthGibsonDurrumstopped flow apparatus was used and, due to the low intensity of the observing light beam, the rate constants are approximatively two times smaller than tose reported in Table I. The rate constants were as follows: Y = 0.16 k = 1.41s; Y = 0.34 k = 1.151s; Y = 0.75 k = 0.31; Y = 0.96 to 1 k , = 0.15/s, k2 = 0.351s. All samples contained approximately25% of a slowly reacting component ( k = 0.04 to 0.071s) attributed to the dimers (see text).

1M

E 80

JD

400

IU

520

580

€40

iw

uvaveiength FIG. 4. Absolute spectra of the cyanide derivative of reduced a and j3 chains (the a chainsare blue-shifted with respect to the fl chains by 3 nm). The extinction of the visible region was multiplied by 10. TABLEI Rate constants for the a chains, 0 chains, and hemoglobin as calculated from non-linearleast squares regression of the data presented in Figs. 1-4 Experimentalconditions: T = 20 "C, Na2S20450 mM, methyl viologen 10 yM, 0.1 M phosphate buffer, pH7, containing 1 mM

slowly reacting materialwas present even at high hemoglobin concentration (50 wM/heme, data not shown). The autocatalytic behavior, very clearly observed in the experiment reported in Fig. 5, is ascribed to the allosteric transition taking place during ligand release fromtetrameric hemoglobin. This effect resembles the time course typically observed during thecombination of deoxyHb with CO (Antonini and Brunori,1971), which has been taken as an unequivocal evidence for kinetic cooperativity. A better fit to the autocatalytic time course observed when the reaction is started from fully saturated cyanide ferrous Hb requiresa modified kinetic scheme(Scheme 2), which (disregarding the fast reduction phase) introduces the two simultaneous processes

cyanide (all concentrations are after mixing). Values in parentheses are the apparent constants calculated without taking into account the heterogeneity of the reaction (see text). It is important to remark that the values of the rate constants for the methyl viologen driven reduction of cyanomet hemoglobin approachthe time resolution limit and of the instrument. Sample

kl

(dithionite)

chains 10 yM 0.171s 351s 0.0451s (3 chains 10 p M 0.41s HbA 1 p M (1.2019) HbA 10 p M (0.451s) HbA 10 y M + IHP HbA 10 UM + CO

k , (methyl viologen)

Hb,(CN-), -+ Hb,(CN-),

Hb2(CN-)1or

+ Hb4

+ Hb,.

kp

SCHEME 2

a

351s 321s 251s 25/s 30/s

0.151s 0.261s 0.241s 0.42/s 0.16/s

reported in Fig. 5 reveal some details whose interpretation demandsadditionalconsiderations.First of all, thetime courses are heterogeneous, approximatelyone-fifth of the overall absorbance change being slow. Furthermore, cyanide dissociation from Hb largely (>80%) or fully saturated with cyanide accelerates as the reaction proceeds, with a clear cut autocatalytic behavior. The former featuremay be attributed, to a first approximation, to the presence of high affinity, R state ab dimers in equilibrium with hemoglobin tetramers in the initial sample of Hb'CN-. This was supported by the dependence of the fraction of slowly reacting species on hemoglobin concentration,although a very smallamount of

In this scheme, Hb4 refers to the tetramer and Hb, to the dimer. The two step dissociation process for tetrameric hemoglobin accounts (empirically) for cooperativitybecause the second rate constant is higher than the first, while ligand dissociationfrom the dimer is slow and corresponds to a single exponential process. This model is unnecessary when the initial cyanide saturation is lower than -50%, in which case the whole reaction is described by two independent exponentialdecays. Effect of p H and InositolHenaphosphate on the Rate of Cyanide Dissociation-The effects of pH and theallosteric effector inositol hexaphosphate (IHP)were also explored. An effect of p H ontherate of cyanidedissociation(already demonstrated for sperm whale andhorse myoglobins, see Bellelli et al., 1990) is also observed in the case of the a and p chains of hemoglobin. The titrationof the rate constantfor cyanide dissociation from the isolated a and the p chains, obtained with stopped flow measurements at different pH,

2262

Dissociation Cyanide

Cooperative

conforms to the reaction of a single ionizable group with pK = 7.6 for the a chains, and pK = 7.3 for the p chains; the overall difference intherateconstants between the two extreme pH values does not exceed a factor of 3, being faster at acidic pH. Since the binding of other ligands to theferrous form of most monomeric hemoproteins (such as the a and p chains) is pH-independent (Antonini and Brunori, 1971), a mechanism of cyanide release involving protonation of the bound anion may be invoked to account for the phenomenon (Bellelli et al., 1990). In thecase of hemoglobin, where the Bohr effect may come into play, the protonation of both the distal histidine and the Bohr residues affects the rate of cyanide release. In this case the pH dependence of the reaction rate constant is much larger than in the case of the (Y and p chains (being expressed by a factor of =IO), and thelow asymptote of the titrationis shifted toward more alkaline pH values (data not shown). When the reaction is carried out in the presence of IHP, which is known to stabilize the T state of Hb, the ligand dissociation is faster (see Fig. 3 and Table I), whereas in the presence ofCO (where only the R state is populated at all times) the reaction is slow, even in the presence of 2 mM IHP.

bin cyanide is bound as the anion and stabilized by a salt bridge with Arg E10. Human hemoglobin, the isolated a and 17 chains, horse, and sperm whale myoglobins, all dissociate cyanide at a rate higher than Aplysia myoglobin, and all display a pH dependence. In these proteinsthe anion and/or thehydroisocyanic acid might be the ligand to the ferrous iron in the intermediate species. Indeed, it may be argued that, since hydroisocyanic acid is chemically similar to alkyl isocyanides, it might be a better ligand than theanion. However, in theabsence of independent evidence we favor the hypothesis that the preferred ligand is CN-, in analogy to thecase of Aplysiu myoglobin. Comparison of the properties of the hydroxy met and ferrous cyanide derivatives of Aplysia and sperm whale myoglobin supports this hypothesis and suggests a possible mechanism for dissociation of CN-. Ferrous Aplysia myoglobin dissociates oxygen and the ligands of the ferric iron at rate higher than myoglobins which possess a distal His (reviewed by Mattevi et al., 1991). The most notable exceptions to this trend are the (relatively) slow rate of dissociation of the hydroxyl ion from the ferric heme (Giacometti et al., 1975) and of cyanide from the ferrous heme (Bellelli et al., 1990). The mechanism of dissociation of the hydroxyl ion inmet sperm whale myoglobin CONCLUDING REMARKS was investigated by Ilgenfritz and Schuster (1971) who proThe three-dimensional structure of the oxygen and carbon posed that the distal His E7 acts as a relay, transferring a monoxide derivatives of ferrous hemoglobin have been re- proton to thebound ligand; obviously this mechanism cannot ported by Heidner et al. (1976) and Shaanan (1983) and by operate inAplysia myoglobin. If CN- is the species ligated by Baldwin (1980), respectively. The structure of the cyano-met all ferrous hemoproteins, the same mechanism may apply. In derivative (Deatherage et al., 1976) is also available, although this case the effect of pH on the dissociation rate constantof at lower resolution. It is accepted that the iron-oxygen bond cyanide observed for the proteins possessing His E7 may be in HbO, is bent, the Fe-0-0 angle being an estimated 160” attributed to the titration of this residue interacting with the (Perutz, 1979; Perutz et al., 1987).The bent position of oxygen bound cyanide (with a pK 2 neutrality; Bellelli et al., 1990). is characteristic of synthetic metalloorganic complexes mimIn contrastwith Aplysia myoglobin, the distal side mutants icking the heme proteins (the picket-fence model, Collman et of sperm whale myoglobin lacking His E7 dissociate cyanide al., 1975). The bound oxygen is stabilized by formation of an at a rate higher than the wild type (Bellelli et al., 1990). In H bond with the distal HisE7, and itsorientation is sterically view of the high autoxidation rate of these proteins (Springer hindered by residues Phe CD1 and Val E l l (Shaanan 1983). et al., 1989),we have suggested that in thesecases protonation The effect of H bonding on the rate of dissociation in myo- of the bound CN- depends on water molecules diffusing from globin has been proven by site-directed mutagenesis and the bulk into thepocket, which is more open than in the wild kinetic experiments (Olson et al., 1988). type protein. By contrast the Fez+-CO and the Fe”+-CN- bonds, which We are aware that alternative mechanisms may account for are suggested to be linear in metallorganic compounds (Coll- the effect of pH on the rateof dissociation of cyanide, but in man et al., 1975; see also Heidner et al., 1976), display a so far as the anion is the likely state for the bound ligand in significant deviation from linearity inMb and Hbimposed by the case of Aplysia myoglobin (see above), we believe that distal side residues (Baldwin, 1980). Moreover, it should be ligation of sperm whale myoglobin and human hemoglobin recalled that CO and CN- display superimposable electron may be interpreted along the same lines. density maps (Deatherage et al., 1976). Comparison of the “off” rate constants for the reaction of Cyanide, although isoelectronic and isosteric with CO, several derivatives of ferrous hemoglobin is summarized in which makes no H-bond to His E7, may nevertheless interact Table 11. For 0, and NO cooperative effects are manifested with distal side residues in the heme pocket because it may largely in the dissociation rate constants, while the effect is be negatively charged. Belowwe shall briefly review some predominantly in the combination rate constantsfor CO and evidence for interactions between cyanide and the protein (to a lesser extent) isocyanides. This type of information has moiety in the complex with ferrous hemoproteins. been analyzed by Szabo (1978) in a paper which provided a As shown by Bellelli et ul. (1990), ferrous Aplysia limacina correlation between the kinetic manifestations of cooperativmyoglobin displays a very slow and almost pH-independent ity for different ligands and the structure of the transition rate constantfor cyanide dissociation, at variance with sperm state in hemoglobin. Within this framework cyanide appears whale and horse myoglobins. Aplysia myoglobin lacks a distal to group with CO and the isocyanides, rather than NO and His, having a Val at position E7 (Tentori et al., 1973; Bolog- On, although the latter is stabilized by a H bond to His E7. nesi et al., 1989). However, the side chain of Arg E10 is The slow cyanide release from reduced hemoglobin, unexsufficiently long and flexible to swing back into the heme pected for such a low affinity complex, may bediscussed with pocket and interact with fluoride, azide, and cyanide bound to the ferric iron, as shown by x-ray crystallography and reference to distal and/or proximal effects, as exemplified in NMR spectroscopy (Bolognesi et al., 1990; Mattevi et al., the cases of alkyl isocyanides and CO, respectively. Bulky alkyl isocyanides (but notmethyl isocyanide) display 1991).’ This strongly suggests that in ferrousAplysia myogloslowoff kinetics (Reiseberg and Olson, 1980b), andtheir Bin, J., La Mar, G . N., Ascoli Marchetti, F., Bolognesi, M., and reaction with hemoglobin is characterized by relatively low apparent quantum yields and relevant geminate rebinding Brunori, M., (1992) J. Mol. Biol., in press.

Dissociation Cyanide

Cooperative

2263

REFERENCES Antonini, E., and Brunori, M. (1971) Hemoglobin and Myoglobin in Their Reactions withLigands, North Holland, Amsterdam Baldwin, J. M. (1980) J. Mol. Biol. 1 3 6 , 103-128 Ligand State k,tt Refs. Balthazard, V., and Philippe, M. (1926) Ann. Med. Legale (Paris) 6 , S" 137-145 NO T >8 X Moore and Gibson Bateman, H. (1910) Proc. Cambridge Phyl. SOC.15,423-427 1x (1976) R Bellelli, A., Antonini, G., Brunori, M., Springer, B. A., and Sligar, S. G. (1990) J. Biol. Chem. 265,18898-18901 0 2 T 200-1100" Gibson (1973) Bolognesi, M., Onesti, S., Gatti, G., Coda, A., Ascenzi, P., and Brunori, R 12 M. (1989) J. Mol. Biol. 2 0 5 , 529-544 Bolognesi, M., Coda, A., Frigerio, F., Gatti, C., Ascenzi, P., and co T 9 X lo-* Sharma etal. (1980) Brunori, M. (1990) J. Mol. Biol. 213,621-625 R 9 X 10-3 Boyer, P. D. (1954) J. Am. Chem. SOC.76,4331-4337 Brunori, M., Talbot, B., Colosimo, A., Antonini, E., and Wyman, J. Cyanide T 1.5 This work (1972) J. Mol. Biol. 65, 423-434 R 0.12 Bucci, E., and Fronticelli, C. (1965) J . Biol.Chem. 2 4 0 , PC551PC553 Ethyl isocyanide T 11-21" Reiseberg and Olson Collman, J. P., Gagne, R. R., Reed, C.A., Robinson,W. T., and 0.09-0.4" (1980b) R Rodley, G. A. (1975) Proc. Natl. Acad. Sci. U. S. A. 7 1 , 1326-1329 Cox, R. P. (1977) Biochem. J. 167,493-495 Methyl isocyanide T 83-140" Reiseberg and Olson Cox, R. P., and Hollaway, M. R. (1977) Eur. J. Biochem. 7 4 , 5753.6-7" (1980b) R 587 'The cr and p chains show a significant kinetic difference for the Deatherage, J. F., Loe, R. S., Anderson, C. M., and Moffat, K. (1976) J. Mol. Biol. 1 0 4 , 687-706 reaction with this ligand. Giacometti, G . M., Da Ros, A., Antonini, E., and Brunori, M. (1975) Biochemistry 14,1584-1588 phenomena, due to their slow diffusive motion inside the Gibson, Q. H. (1959) Biochem. J. 7 1 , 293-303 protein matrix (Gibson et al., 1986). On the other hand the Gibson, Q. H. (1973) Proc. Natl. Acad. Sci. U. S. A . 70, 1-4 slow off rate of CO is due to thehigh activation energy of the Gibson, Q. H., Olson, J. S., McKinnie, R. E., and Rohlfs, R. J. (1986) J. Biol. Chem. 2 6 1 , 10228-10239 Fe-C bond and is associated with high apparent quantum E. J., Ladner, R. C., and Perutz, M. F. (1976) J. Mol. Biol. yield and scarce geminate phenomena (Gibson et al., 1986). Heidner, 1 0 4 , 707-722 Similarly, the slow dissociation of cyanide from ferrous Hb Hofrichter, J., Sommer, J. H., Henry, E. R., and Eaton, W. A. (1983) cannot be attributed toa large geminate recombination, given Proc. Natl. Acad. Sci. U. S. A. 80, 2235-2239 that the overall quantum yield is high (Keilin and Hartree, Hopfield, J . J., Schulman, R. G., and Ogawa, S. (1971) J . Mol. Biol. 61,425-443 1955), and the diffusion of this small ligand in the protein Ilgenfritz, G., and Schuster, T. M. (1971) Probes of Structure and matrix should not constitute a barrier. Function of Macromolecules and Membranes, Vol. 2, pp. 399-406, According to these considerations, the similarities between Academic Press, New York cyanide and CO as ligands of ferrous hemoproteins are high, Keilin, D., and Hartree, E. F. (1955) Biochem. J. 6 1 , 153-171 and following Szabo (1978) we expect that they may becarried Mattevi, A., Gatti, G., Coda, A., Rizzi, M., Ascenzi, P., Brunori, M., and Bolognesi, M. (1991) J . Mol. Recognition, 4, 1-6 further, predicting that cooperativity should be expressed also Moore, E. G., and Gibson, Q . H. (1976) J . Biol. Chem. 251, 2788(and perhaps more significantly) in the "on" constants being 2794 different for the R and T quaternary states. Experiments to Olivas, E., De Waal, D. J. A., and Wilkins, R. G. (1977) J. Biol. Chem. 252,4038-4042 directly determine these rate constants are not easy because of the very low affinity, but will be pursued in an attempt to Olson, J. S., Mathews, A. J., Rohlfs, R. J., Springer, B. A., Egeberg, K. D., Sligar, S. G., Tame, J., Renaud, J. P., and Nagai, K. (1988) extend the batteryof ligands available to the study of hemoNature 3 3 6 , 265-266 globin function. Perutz, M. F. (1979) Annu. Reu. Biochem. 48, 327-386 A final point worth specific mention is the possible use of Perutz, M. F., Fermi, G., Luisi, B., Shaanan, B., and Liddington, R. C. (1987) Acc. Chem. Res. 20, 309-321 this reaction for the analysis of cooperative effects in hemoglobin samples which may have been stored as the (very Press, W. H., Flannery, B. P., Teukolski, S. A., and Vetterling, W. T. (1986) Numerical Recipes,Cambridge UniversityPress, Cambridge, stable) cyano-met derivative because of autoxidation, or which United Kingdom may be available in small amounts. It should be emphasized Reiseberg, P., and Olson, J. S. (1980a) J. Biol.Chem. 2 5 5 , 41444150 that the rateof dissociation starting from partially saturated metHb CN- is a very simple, yet informative experiment Reiseberg, P., and Olson, J. S. (1980b) J. Biol.Chem. 256, 41594169 which may be successfully carried out even with a normal Shaanan B. (1983) J. Mol. Biol. 1 7 1 , 31-59 spectrophotometer. Therefore,an experiment like that shown Sharma, V. S., Vedvick, T. S., Magde, D., Luth, R., Friedman, D., in Fig. 5 may become a useful test for the presence of coopSchmidt, M. R., and Ranney, H. M. (1980) J. Biol. Chem. 2 5 6 , 5879-5884 erativity in mutanthemoglobins which, in view of higher rates of autoxidation and/or lower intrinsic stability, may be diffi- Springer, B. A., Egeberg, K. D., Sligar, S. G., Rohlfs, R. J., Mathews, A. J., and Olson, J. S. (1989) J. Biol. Chem. 264,3057-3060 cult to examine by oxygen equilibrium. Stitt, F., and Coryell, C. D. (1939) J . Am. Chem. SOC.61, 1263-1266 Szabo, A. (1978) Proc. Natl. Acad. Sci. U. S. A. 75, 2108-2111 Acknowledgments-The skillful help of Dr. B. Vallone (Rome) and Talbot, B., Brunori, M., Antonini, E., and Wyman, J. (1971) J . Mol. V. Cavalli (Geneve) is gratefully acknowledged. We are also indebted Biol. 5 8 , 261-276 to Dr. E. Henry (Bethesda)for introducing us to themysteries of the Tentori, L., Vivaldi, G., Carta, S., Marinucci, M., Massa, A,, Antonini, SVD analysis. E., and Brunori, M. (1973) Znt. J . Pept. Protein Res. 5, 187-200

TABLEI1 Kinetic parameters for the dissociation of the ligands of ferrous hemoglobin, at pH 7.0, and 20 "C