Magnetic resonance studies of metal-enzymes

~u~~~~~~ Chemistry Reviews, I2 (1974) 407430. 0 ELsetier Scientific Pub& mg Company, Amsterdam - Printed iinThe

Netherlands

CONTENT

..................... A. ~~troduct~o~ ................. B. Nuclear ~~~~at~~~ fates (i) Nuclei with I = 4 in a paramagnetic (ii)

Effect

of chemical

407

environment

1 408 409

...........

exchange on relaxation rates

....

(iii) Relaxation sate enhancement in prtramagnetic systems ............ 1’ (iv) Nuclei with I > 5’ in a diamagnetic environment .......... systems ... c. Applications of nuclear magnetic relaxation data from metal-enzyme ii) ~~terrn~~ati~~ of ass~c~t~u~ c~~$t~ts ............ (ii] ~ua~~t~ve ~deu~~~~t~o~ of binding sites ............ (iii) Determination of some structural parameters ........... (iv) Sonx studies of the role*of metal-enzymes in biochemical reactions .... ................... D. Concluding remarks References ._ I. . . . . . . . . . . . . . . . . . . . .

410

4lj 414

415 416 416 417 418

-

427 428

A. INTRO;DU~ION

Enzymesgre prtiteins which, due to their particular structures, ac’t as vex-y specific biochemical datalysts. The oxid&ve’enzymes containing heme iron’ were the first ~e~o~~~sed ex pies of metal ion ~art~~~~a~ion in b~o~o~~a~systA=. Since then a kg@ number of e~yrnat~~ reactiotis requiring a metal ion as an essential

component

have been documented’

-‘.

In this reiriew we consider a metal-enzyme

to be one in which the met;tl io’n is an essential participant in the catalytic process. An unequivocal demonstration of such participation is usuzitty extremely difficult to achieve and various type’s df evidence are req~jred' t It is ~~~~~~~e~~ to divide Mets-e~ymes into two classes: ( 1) Enzymes with stro bound metal ions. These are called metallo-enzymes4; (2) enzymes

with disso le metal ions. These are referred to as metal-ian-activatedEnzymes in these two classes show distinct properties. Quantitatively the two classes may be distinguished by means of their stability constants. For

enzymes’.

metallo-enzymes6

the stability constants are 2 IO* M-

* To whcrm correspondence

!ihould be addressed.

”

wkife: for rn~t~-~o~*a~t~vat~d-

P.J_ QUILLEY,

408

G.A. WEBB

enzymes they are < IO8 ikl-l.MetalIoc-enzymes can be isolated with the metal attached t3 the protein of the enzyme whereas the the metal ion often dissociates during the ~~rj~cat~on of the metal-ion-activated-enzymes. This leads to a loss of activity which is usually only regained upon the addition of metal ions of the COTrect size 4. In general the metal ions in metallo-enzymes have unusual stereochemistries which most probably contribute to their catalytic behaviour. Each enzyme has an Enzyme Council classification, e.g. pyruvate kinase (E.C.2.7.1.40) which indicates the substrate upon which the enzyme acts, the products of the ~~~~rne-&at~ys~d reaction and which metal ions or other co-factors are required for the specific activity of the enzyme. To date ~fte~~ metal cations have been reported to activate one or more etzzymes each. Since enzymes are proteins their stability is pH and temperature

dependent, consequently these parameters have to be carefully controlled during experimental studies, This can often be achieved much more readily in magnetic resonance exan in tfrose employing many other physj~~ t~~~~q~es. personas This review is not intended to be fuffy rigorous, its main aim is to present a view of the present state of development of the field and attempt to suggest further experiments. It is hoped that the reader will approach the original literature on this subject with added interest.

3, NUCLEAR

RELAXATION RATES

In diamae;netic environments the relaxation of nuclei with I = i is usual Iy controlled by nuclear dipole-dipole interactions, whereas in the case of nuclei wi& I> $ quadru~u~ar relaxation is usually dominant. The effect of a p~~agR~~~~ centse on the relaxation rates of a n~~~e~s with I = 4 depends upon the distance between the two interacting centres and tie motional freedom that they experience. Consequently changes in nuc1ea.r relaxation rates concomitant with the introduction of ii.paramagnetic centre permit the estiourhood of the active site and mation of internuclear distances in the ne5 ~h~a~ter~e the local mo~e~u~~ ~ut~~~ cf to this site9. Nuclear ~ua~upo~~ rel ation depends ~~0~ the j~te~act~on between the nudear electric quadrupofe moment and the electric field gradients within its close proximity7 _The electric field gradients are very sensitive to changes in electron distribution around the nucleus such as occur in bonding. Consequently IocaXzed molecular motions can also be studied by means of the NMR spectra of quadrupolar nucler-8 . In addition to lot geometrical and rnot~~n~ parameters nuclear relaxation rates formation relating to the rate of substrate exchange with an enzyme (ref. 9, lo), Consequently substrate exchan may also be studied by means of nuclear relaxation rate measurements.

in which it is assumed that the g tensor is isotropic; modification is required for g 15. The first term in each of these equations represents the ditensor anisotropy polar interaction between nuclei and unpaired electrons whilst the second term represents the ~o~taet interaction= It is widely assumed that these expressions are also applicable to rnacrorn~~~~ul~ complexes. The correlation times for the dipolar and contact interactions are related to the correlation times characterising the rotational motion of the separation vector between the paramagnetic centre and the nuekus being studied, or, the electron spin relaxation time TVand the residence time of the observed nucfeus in the furst ~~ordi~ati~~ sphere of the ~~r~agneti~ centre, rM, by eqns. (5) and (c;)_

l l -2-+L_ -=52 5 1 -=-

%

51

(5)

l+l

52 %

7M

Equation (5) shows that the magnitude af ~c is determined by whichever of is the shorter. For first row transition metal ions TVvaries from 1O- 8 Lg ~6-““%, for unbound aquo-ions r is of the order of 1W 11 set and T~$ is usually several orders of rna~~~tude larg& than either or or 7s. For Fetf, Cou and bitt, f < -r and for GI~~I M# and Cu” r < T . ~~nsequ~~t~y proton relaxation data cktaik information’abaut the molecfrlar dynamics of these ions in solution (ref. 16, 17). At high temperatures 7M < rS for metal ions; thus re becomes controIled by TM(ref. 18). In the majority of reported cases of interaction between a paramagnetic centre and a ma~romo~e~uIe, re is ~~f~~ie~t~y farge such that wf T: 9 1; ~~~seq~~~t~y the contest contribution to T,, becomes ne~~~b~y smafI whilst remaining sig~~~ant for & _Thus by measuring TIM as a function of the concentration of paramagnetic centrer in solution a value of r” can be obtained from eqn. (3) by assuming that uN = 0 and that a value for 7, is known for each proron in the macromolecule. This technique has been applied to some complexes of Cr*I1, Mnir, FellI, Con, Nirr and CuII wiiEr some amino acids and A~~.~r~~. 38).

(ii) Effecr

ofchemical exchange on relaxarirsrt rates

If the moltlcuIe containing the nucleus under observation &changes very rapidly between the bufk solvent and the first coordination sphere of the pammagnetic metaI im, tfien the rate of ~~chan~~~ Y$, ciln b~~~rn~ fast ~Qrn~~~~d with relaxation rate in the first coordination sphere, Tm:, i.e. TIN S Tag. Under these circumstances the paramagnetic contribution to the observed relaxation rate, Tc$t is the weighted average of the nuclear relaxation rates when the molecule is in the bulk solvent and when it is in the first coordination sphere afthe metal. WV is the

number of &and molecules in the coordination the met4: l&and concentration ratio, then

sphere of the metal

ion and jr is

1 -=Nn

IiF IP

Gbe

The conditions of fast exchange implicit in eqn. (7) apply to transition metal ions in aqueous solution; if the metal is bound to a macromolecule they may no longer be applicable. Since hgand molecules may 1: side on the metal longer than the minimum time required for relaxatian to occur, ~~~f~~~e~t re~~~ti~n of the bull solvent molecules results, The p~r~agnet~c contribution to the relaxation time becomes correspondingly longer, The residence time, rhI, of the nucIeus in the first coordina, tion sphere is taken into account in eqn, (8) (refs. 19, 20)

(8) where Tl(u)-’ is the outer sphere contribution arising from interactions the paramagnetic centre and all of the nuclei beyond its fist coordination For fast exchange T,, 3 rlclXand eqn. (8) becomes ’

if Tl(o)-’ exchange

then eqn. (9) reduces is small compared with T$ rM S T,, and eqn. (8) becomes

and if 7’l(o)-1

is

small compared

between sphere.

to eqn (7). For slow

with T&‘,

Equations similar to (7) - (11) hold for 7’c.t of sornc1: transition metal ions 19_ As can be seen from eqn. (8) the observed value of TIP depends uponN, fl, T!, T,, and possibly T*(o). Equation (3) shows that T,, is a function of r and T, whrch could be dominated by TV*T~ or TV_ Hence in order to interpret experimental values of TIP we require irrdependent estimates of limits of the range of vahres t&en by N, r, T=, rS and 7M (ref, 91. To obtain these it is often necessary to consider the relaxation rates a~ a function of temperature and frequency. If the temperature coefficient of TrJ is negative it is unlikely that chemical exchange is the rate-limiting process since the rates of chemical reactions have positive ‘I

P-3. QUILLEY,

412

GA, WEBB

temperaturecoeffxients. Consequently for fast exchange 3”,, is thus controlled by TIM as shown by eqx~(7), where T,, cm be determined by either rr or 7,. The value of 5rrusuidly decreises with increasin temperature whilst rs may either be ~~~~d

or decreased in mag

of Tip’ is positive then further data are required in order to dMinguish between three possibl situations: (a) fast exchange is taking If the temperaturecoefficient

place as desmiid

by eqn. (7) and TIM is controlled by 7s which has a positive temperature coeffkient. This may be verified by obtaining an estimate of the lower l&nit of ‘~1fm EPR dab if wailable, If rC is found to be ~~~t~~ than thq tower f TSthen ~~a~~~ (a) does not occur. ~~~~er,if 7s ir; ~nobt~~ab~e from n&nt data or if rs < -rcthen the GNation is less clear since (b), the chetical exchange rate could be sufficiently slow for T$ to be dominated by ri’ as indic&ed by eqn. (8). Measurement of Tzp could clariiy this situation, since if TIP is limited T. must also be exchange limited _ Frequency dependence men@ of T$ may be used to d~s~~~~ between si~a~ons (a) and (bj, if no frequency dg~ndence is observed then fb) is the relevant situation. Finally, (c) ifr,>

w=9

set then W:T~ > 1 and the dipolar term in eqn. (3) becomes a

*I. Thus an increase in temperature will lead to smaller values of rc function of rG -' This situation can be distinguished from and a posith temperature coefficient T,,. (b) by fn=qucncy dependence studies. ~~~e~~y, r~~~ati~~ times may be obtained from either continuums wave or pulsed NMR techrl.iques27. Both have been used to provide estimates oFN and r for metakmzyme systems’ _The type of interpretational problem that can arise is illustratedti FQ. 1 for solutions ofE’. c~lr’ ribosomal RNA containing Mn’* ions3’. It is considered that the chemical exchange rate is fast over the entire temperature ms ~o~~e~~ and that rs is -the dominant correlation time at the lower end of this r~~~~‘. ~~~~q~e~~y, the data trisected in Fig, 1. can be accounted far OR

!?&

1. A

eon~~

ofkIT against reeiproca~ temperature for solutions M# ions at 20 MHz (ref. 39).

plot

ofE. mli

ribo~~d

RNA

the basis af rapid chemical exchange provided that 5 for Mn” has a positive temperature coefficient. However, there are conflicting views on this point in the EPR data could decide the si of the temliterature 13%40.Viably-t~mpera~re f~~jent of rS and ~o~~~q~~~t~y whether peratu e reported ’ Yr~~t~rpr~~t~~~ of the data is correct.

The effect of CP , ~~~~ and Curi on the rel~at~on rate of water pr~t~~~ is int~b~~o~ to rC arises creased in presage of nucleic acids’t _ bi~d~g to a macrofrom TVin aqu~c~rn~l~x~s of these me molecule, 3-r shows a significant increase, c quently nuclear spin relaxation becomes more efficient and the proton relaxation rate (PRR) is enhanc The presence of nucleic acids has Little effect on the PRR of the aquo-complexes of FeU, Co”” and Ni’ since the dorn~~t co~~but~o~ to ~~ for these ions arises from TS which. most probab macromolecules. A large PRR enhancemen rS > 1O- ’ ’ set, organic free radicals may may provide structural changes of unknown magnitude2” _ t correlation time for a system showin FRR e~~cern~~t could b&i ~~~~~, abbe-met and terser ~~~y~e-met~ I3 7 >7 thus7 istfiedo ant ~orrel~tiun time in Much of de edly WC& on PRR enhancement due to Mn” bound to macromolecules has been interpretedg‘I on tie assumption that T, > i,, this work requires reappraisal25. The ~~~~cerne~t

factor,

f Ez far the l~~~tud~~~

relaxation

time is defined*’

as

r,(obs)

-T,

where T p is the contribution of the pammagnetic ion to the observed relaxation time Ty I!obs) and T, is the relaxation time in the absence of a paramagnetic ion. The asterisk indicates data obtained in the presence of a macromolecule. A similar enhancement factor, ‘2 has been defined for the transverse relaxation time” _ The observed e~an~ern~~t factor el is a wei~t~d ~v~~~ge of the e~~~~e~~n~ due to p~~ag~eti~ ions blind to rna~~~rno~ecules~ elbT and that free paramagnetic ions in solution, elf, In the absence of macromo of elf is unity from eqn. (12). However, since the structure of bulk water can be modified close to macromolecules, T1 and T1 * are not necessarily equ Elf may differ slightly from unity. By measuring Ed for a series of macromolecule

414

netic centresit is p~~si~je to ugly Jif to be ante. At ~~~t~ adorns

eIb

froma plot

rna~r~rn~~~~~~~c~n~~~~~ti~~~ magnetic centres are bpund to the macromalecule and eIb tends to Ed asMBf tends to zero. Some of the applications of elb measurements to the determination of metal-enzyme parameters are discussed i6ktion C.

with a salutiun contti

ions which bind to specific protein sites and subsequently with halide ions from the solutisn. In the majority of halide ion probe studies on proteins the number and availability of metal ion binding sites have been determined from continuous wave 35 CI NMR line width measurements. Abragam 28 has described an extreme narrowing approximation which occurs when

(14)

From eqns. [J4) and ( IS) it follows that

electric field ~~d~~nt at the nucleus, is assumed to have cylindrical the 35C1 nucleus, eqn. (16) be~~m~s~”

A = 22 diPi

i

where A and A(o) are ahe line wid the metal JOE-IS, the asterisk indicates d to macromolecules the lin

chelattig

agent, For chloride

ions of similar size are able to re lam each other in enzymes, usually without loss of activity. Consequt=ntly M t which is the natural activator fcm ymes, can be replaced by MnIl which produ for the ~ep~dases, Wx”‘, Ca[‘, Nix’ and 2%” ng g~~~~rn~nt factors can provide infor

416

PJ_

Qt.HLLEY, GA_ WEBB

If a hydrated metal ion b ds to a macromulecute the resultin tke water protons, depends upan phicai treatment of the titration data ob ined from solutions state. Therefore concentrations of appropriate metal ion containing suita d enzymes provides es of ~~~~~ a~o~~a~~~ CC3stants and the ~~rnb~r of ~~nd~~ sites per: macrloes m le? A number of ternary complexes of the type have been ~~v~~t d, A~so~~at~~~ ~on~~a~~~ reported following boti E with MS have the ~xp~r~e~ta~ data9. In the case of adenyi & talined from these two forms of analysis iterative solutions sf the equations des~~b~~~ the system are supported by ather data, thus a r~~~v~~t~~ation of the earlier recuts obt~ned by ~ra~~~~ extrapofation methods is c

nt of Water protons in the presence of the b, with the e~a~~~ment of the ternary come classified 26 metal-enzymes accordin three coordination schemes’ . Type enzymes produce little or no enhancement of II”; ~ow~v~r~ when a substrate is added a large tfte fJRR in bj~a~ ~ornp~~x~~ Eltt < flrc” Tfris is ~~~~~~t~~~With the e~~~~~~e~~ 5s observe$ and d complex, E-S-M, in which the substrate is reformation of a substrate bri d anal ion into the e~v~o~~~t of the rn~~~~ ~o~~b~e for br~~~~~~the h molecule. The relatively srn s of Ela and eElb indicate weak binding betwe enzyme or the substrate in binary complexes. Hence type I taken to be me tal-ion-ac tivat~d* prudish sj~~~~~ant e

Type III enzymes probably form a ternary complex of the type M-E-S in involved in any bon s which may occur when and CUhn WaS pub Since tkz t~b~l~ti~~ s, their ~~ordi~at~~n ave been ~~~~i~~d ac and fructose 1#Ldipfro~-

~A~~ETi~ RESONANCE STUDIES OF

METAL-ENZYMES

417

phatase4’ are reported to be type II enzymes. 13C relaxation rate studies on the quaternary complex pyruvate kinase - M# - phosphate - pyruvate show that the M# ion is ap~r~x~ately the and carbony carbon atoms of py~vate4~ _The Same pittance from the carbo run, are consistent with the presence of pyruvate estimated distances, 0.7 1 in the second caardination sphere of MI?‘. Proton NMR data are consistent with carbonyl and phosphoryl coordination in the type II terna complexes yeast ase - MnU - acetai phosphate and te48e ‘IP NMR mea~rements on the yeast aIdolase - Mnr ’ - fructose dipho t~r~~~~ornp~~x fructose 1,6-~~~~ospba~~ - MntZ - fructose - 1 - phosphate and 13C data on the binary Mn” - fructose - 1 - phasphate have led to the propoxd Mn’ environment shown in Fig, 2, (ref. 49)_

As indicated by eqn. (31, (4) and (7) - (1 I) metrical ~f~~ation concerning metal-enzyme systems may be obtained from lear relaxation rate measurements. By assuming that 7, is dominated by or for srrmll rigid complexes of MnI’ and lo-” set, the dipolar term in eqn. (3) provides an estimate of metalthat rC =3x flu&us separationg*‘? In this manner, distance values have been re orted are Gom~~~ble 50 those obtund by X-ray diffraction t~~~~ique~~~ s O+ However, when the relaxation rate of ‘the coordinated ligand is too rapid or when it is not possible to obtain a high concentration of the complex in solution, as occurs with many complexes involving macromolecules, then eqn. (8) has to be considered. In general, the calculated distances are not obtained with such accuracy as they are for small

hdl and 31P and 13C nuclei in a fructose-l-phosphate COW from NMR data. For clarity, protons are omitted49. {A) The 1-C8_ 4x3mplex; 03) the twist g-fxuctcrfuranose complex-

2. Some distances

between

obttied

fiuetopyramse

rigid molecules. Errors of f 30% are present in some distance values 9,47,49,50 . These are generally due to uncertainty in the value of rC. If T, is evaluated independently from a frequency-dependent study of TIP, errors of less than 5% may be found? approach is tQ employ a metals ion which is less effective than nuclear relaxation. This technique has been used in the case of carbonic anhydrase where the natural activator, nH, is replaced by Co'!The Car’ proton separation in the binary complex is reported to be 0.25 - 0.29 nm (ref, 5 1), By rn~~~ ofeqn.. (7) -(I I) it is possible to est~ate the number of water mofecules, M, in the first coordination sphere of a metal ion in a ry or ternary comprerequisite is the plex with a _macromolecule from relaxation studies. A neces availability of acceptable values for 7, and the metal-nuclear separation, t. A value of I= has been obtained for the protons of water mo!ecuies coordinated to Mn”’ from frequency-dependent relaxation dataz4. By taking this together ~4th r = 0.29 rnn a value of 3 has been esttiatedz4 for N for the binas-y complex facet between pyruvate kinase and hydrated Mn”. In a similar investigation on the quaternary complex, creatine klnase-MnlI - ADP-creatine, when r is taken to be 0.286n.m N is found to be less than one half25. Frequency and temperature dependent measurements of T,, shaw that this low value of N cannot be attributed to slow e~~h~g~_ It seems likely that outer spher scribed in eqns. ~~~~~lU~, plays an ~porta~t role in this system, the mew-proton separation for protons residing beyond the first coordination sphere can only be crudely estimated, it is not at present possible to decide how important outer sphere contributions to the relaxation data are25,

Information relating to the electronic and geometric structures of the active sites of an enzyme as well as the chemical proees s involved in the catalytic reztction may be obtained from NMR data. We consider here the results of some paramagnetic and quadrupolar probe studies for a number of metal+nzymes treated under six c~ass~~c~tio~s. 1. a”rartsferases

These enzymes catalyse the transfer of a functional group, e.g. carbonyl, amino, b~t~veen ~o~ecul~s~ Included amongst them are it the kinases, most of which em$oy Mg”” as an essential activator. ~r~at~~e kinase and pyruvate kinase have been studied extensively, the results obtained show that there is:no general mechanism for the action of enzymes in this group.

419

Creatirre kiatrse (E.C.2_7.3.2)

catalyses the reaction

ATP + creatinc - ADP + phospho-creatine G2s3. The results and has a d~m~ric structure with one essential kinase and variof paramagnetic probe NMR studies on the i ous substrates have been comprehensively reviewed’* I’_ Recently a nitroxide radical, IV-( I-oxyl-2,2, 5, 5-tetramethyl-3-pyrrolidinyi) iodoacetamide has been edified enzyme lacks used in a study on the active site of creatine kinasse 54_ T ph~sph~~~nsferase and ATPase activities but retains the to bind nucleotf substrates. The PRR e~anc~~e~t for water in ternary complexes enzyme- MR nucleotide are found to be 13 & 2 fur the.ADP complex and 1 I 5 2 for the ATP complex. These enhancements are lower by 38% and 15% respectively than those for the unmodified enzyme. From EPR data the dipolar interaction between Mn" aration of 0.75 -C0.15 nm for fhe and the nitraxide radical corresponds to a. ADP compfex and I _I4 + OXI6 m-n for the This suggests different e~nf~~at~u~s for the twlr, complexes an us to that reported for the corresponding Mgrr complexes 56 . NMR and EPR experiments have shown that the addition crf c;atine kinase to solutions of creatine kinase-M#-ADP and creatine kinasc-Mn -ATP produces a rearrangement at the active site which results in a less symmetrical environment for the M# ion57 _The addition of nitrate or chloride ions produces further changes in the EPR spectrum and in the PRR of water in the complex c~~at~~~ L kinase-Mn’* -ADP-creatine “. These observatians have been accounted for by the binding of creatine with M# whilst the anions are assumed to bind at the vacant phosphoryl site in the quaternary complex. Mpdification of the ihiol oups of the enzyme reduces its ability to under c~~fo~~tio~~ changes whew creatine is added to the ternary complex creatine kinase-M# --ADP. This loss of c~nfo~at~~~a~ adaptability of the enzyme appears to be refated to the inability of the modified enzyme to catalyse the phosphoryl transfer reactior+. Fyruvczre kinme (E.C.2.7.1.40) requires both Mg” and K* as natural activators. It catalyses the reaction ATP f pyntvate =z=ADP f phosphoenol

pyruvate

Pyruvate kinase is reported to show type 11 behaviour towards all substrates so far studied ‘. Rabbit muscle pyruvate kinase consists of four sub-units of the same molecular weight. It has four bindin sites for Mn”’ (refs. IO, 6 1). The addition of urea r‘sreported to transform the tetramer into an active dimer whilst the number of Mn” binding sites remains unchanged 61 _ The presence of four bound 2~” ions per mole of enzyme has been established

e compkxes of pyruvate k&use Fig. 3. Comp>ison of a compositemodel of the (upper figure) with the termiry complex formed with P”-enolpyruvate Ciawer figlue)59.

o,‘I.studies on the interactions between muscle pyruvate kinase and e analogs reveal that the ~rnob~~at~o~ oached? This immoas the reaction centre is at tie r~ac~~o~ centre co Bow ~r~e~~at~o~~~effects of the cereal analog to be upera~ive in ~~~yrn~ cat~ys~s~9- A composite rno een derived from NMR ~ata6~ . TMs is g&en in Fig. 3 where it is ed with the ternary complex formed with P-enolpyruvates9. A corresponding model of the phosphoenofpyruvate analogs at the active site of pyfuvate kinase inciuding that nuclear correlation times is p sented in Fig, 4 (ref. 58 . =C nuclear rnaiy ~urnp~ex py~vate ~~ase-~~ i ‘boned py~vate is oriented ~~~‘-py~vat~ camplex”? A pa&b1 for the enolization reaction vate kinase which is consistent with I36 and ” P NMR data, is shown in ses the reaction

421

Fig. 5. Mechanism of the enolizatian of pyruvate kintlsc consistent with 13C and “P

NMR data4’.

35Cl NMR data has been re orted for beef liver rhodanesegO. AIthough the enzyme binds one equivalent of Zn R it can also be prepared zinc-free in which case it retains its activity. the activity is found Rot to be e~an~ed by the addition of Zn” ionsgu. 1t.is thus bluely that Z# takes part in the e~yme-ca~~ysed cleavage of the sulphur-sulphur bond of the substrate. Obviousiy the function of the strongly bound zinc atom in this enzyme requires further investigation.

These e~y~~~ catalyse r~~~t~on~ in which a complete group of atoms is removed from a molecule, e-g. decarboxyiation, deamination. Early NMR work OR enolase, histidine deaminase and citrate lyase has recently been reviewed ‘. AI&Ease (E.C.4.1.2.6.) catalyses the reaction

Yeast aldolase contains ZnU which can be removed to give an apoenzyme

which is reactivated by M# ,Co" or Zn” (ref. 9). PRR enhancement, EPR and kinetic data on the apoenzyme have revealed two catalytically active sites which bind Mn’” as some we~er binding sites48* From PRR e ancement measurements it has been reported that e~~e*bo~nd Mn” co~rdinat~s with acetaf phosphate dibydroxyacetone phosphate, dihydroxyacetone pho~hate hydrate and fructose diphosphate48. Some probable structures and mechanisms for yeast aldolase-substrate complexes based upon these $servations are given in Fig_ 6. The high dissociation constant for the binary -fructose diphosphate complex su sts ~~depende~~ bi~d~~~ of Mn”’ to e Fho~ho~~ groups. As ~d~catcd in Fig. 6, caxbonyl group coordination to the enzyme-bound Mn’” ion reveals that the metal ion plays the role of electrophiIe6*. Aconifase (ILC.4.2.1.3.) catalyses the dehydration of citric and isocitric acids to form cis-aconitic acid&e reverse reactions and the interconversion of citric and

422

rather than a ~~t~ytic role for Fe”‘. r its activity it ako binds wiffi this does not activate it. There ark retorted to be two ti together with five ta Seven wetier tines in aconitase4? Relaxation time measurements have dern~ns~r~t~d that MnIf induces ty ~av~~~r in acanit rmation of ternary metal bridged ccsmplexes terser complex ae9nitasc-Nnil-citrate is The ~n~~-t~-~~~t U.~~ nrn and the ~~~~-t~-~r~t~~ distastes in reported from NMR data to be 0. the aconitase--Felf-monsdeutlerocitrate G to be 633 zk fl.62 n.Tn (-my--_) cement with X-ray diffractive md 0.31 rf 0.02 nm (-CHD-) which are in spectrum of citrate ions when data4’. A contact interaction is found in tw’f?t?ll e~yrne*~~~~d Fe” is present. This is consistent with direct ~~~rd~nat~~ Feff ions l

e ~~~sist~nt with the NbiR data, are presented in 4 (ref. 45). Following e of water to farm the Aaconitate combs&ate may ehan from a ~itrat~-~~~~ ta an ~s~cit~~te-1~~ csgnforma3 tian by routes I, ILOXIII. RANit cmmprise~ the ferrous wheel mechanism twist mechroute If the ~~~~~~~ferruus whed rn~~~la~~srnand route III ~~~~2. ~i~~y water is added to form isseitrate. ligands such as sulphur, coordinate to Fe” and undergo the Bailar twist mechanism which is associated with a rapid change from hi low-spin Fe” (refs, X2,64,65). The fact that FeI’ irsa unique a&ivatcrr together with the f~~at~~~ of 42% low-spin Fe” aeon ~um~i~~tiun between and the r~~u~~rn~~t of a red~~~~ a ~~yrn~ and FG? (ref. re on the ~nzyrn~ favours ioute III. which c~uId ~r~d~~e 8 sulphur d~nat~g c However, none of the routes d~s~~~ed in .7 can be ignored at present4’. tZk&~ic a;r&pkz~t? (E.C.4.2.1.1.) cat~ys~s the hydration Of clarbon dioxide at high pH. It c~~ta~~s ane duly bound zinc atam per molecule which is ~~~~~

MAGNETIC

RESONANCE

STUDIES OF METAL-ENZYMES

Fig. 7. Three aitemativc mechanisms for which are consistent with NMR data4’.

423

the conversion of citrate to isoc-itrate by aconitasc

sary for its activity.. Variation in the 35C1 NMR line width during titration demon-

strates the presence of one coordination site which is available fur chloride or cyanide binding, but which can be inhibited by acetazolamide 33. Bovine carbonic anhydrase is found to be a thousand times’more effective than hydrated 2%” ions in broadening the 35C1 signal whereas the zinc-free apoenzyme is twenty five times less effective 33 &‘. These obs~~a tions have been in terpretcd

in favour of a chloride ion exchange process and a moiecular rotational interaction with correlation times of ” lo-* set for carbonic anhydrase and 2 IO-” set for hydrated Zn’* ions 33*67. PRR enhancement data have been obtained by replacing Znrl by Co” fief, 5 1). The Co--proton separation is estimated to be b~t~v~e~ 0.25 an KS run by as-L1 suming that the correlation time for the Co” enzyme binary complex is ” X0 set and that only one water molecule is present in the first Cof’ coordination sphere 5 t . However, if the Co” ion has only one bound water molecule then it is most probably four coordinate, the electronic spectrum does not provide unambiguous information on this point 7r *72. Some figand binding studies have been recently reported” on carbonic a~ydrase with encouraging results. 3. Lygases

These comprise a small group of Iess than fifty known enzymes which catalyse

424

ILLEY, GA, WEBB

condensation reactions. Over half of the lygases require Mgrr as the natural activator and are thus amenable to paramagnetic probe studies using M#. fymvare carbcrxylase (E.C.6,4, I: 1.) from chicken liver was the f?rst naturally ~~cu~n~ ~~~~r~~t~~~~y~~ ta be discovered. Cunseq~e~t~y it has been the focus . Pyruvate carboxylase catalyses t~~~s~arboxy~at~on of much research a~~~v~t~~~~? reactions_ When purified from calf liver it is found to contain both Mg” and Mn” bound in a combined stoichiometry equivalent to the biotin content of the enzyme wlrich implies a different structure to that found in the enzyme obtained from chicken liver. However, it is possible to substitute Mg” for M# with retention of ~at~y~~ activity in the ~y~va~e carboxy~a~ purified from ~~~~k~~ liver. Further studies are required in order to differentiate between a number of possible models which have been proposed to describe the mechanism of this enzyrneg*47*68,6g_ l

ese enzymes catalyse ~x~da~o~-~e~~~ tion reactions. As impfied by the title, The results of NMR studies on cytochrome C, some nonheme-iron proteins and sume copper enzymes belonging to this category have been recently reviewed9*r ‘. Liver alcohol &$ry&ogefrase (IX. 1.1.1.1.). The role of zinc in the binding of the co-enzyme NADH and in aintaining the structural integrity of horse liver has been the subject of ~~rn~r~us ~~v~st~~a~~~~s~$7g_ LADH consists of two sub-u each contains two zinc atoms, it is rendered inactive if zinc is either removed or chelated 73. The presence of LADH in sodium chloride solutions produces an increase in the width of the 3”C1 NMR signa.l which increases with protein concentration74$75. A similar study involving ” Br NMR has not revealed any line broadening in the ADH76. Howev~~~ the Addison of NADH to a solution containing eous sodium chloride produces a decrease in the 35CI Nap signal nstrated in Fig. 8. The differences in the titratiorr curves couId arise from aheat denaturation step empIoyed in the preparation of the enzyme7’. Lindman et al.75 report the NADH : LADH staichiometry ratio to be 1 : 1 at the end point of the titration, whereas Ward et a.l.74 report a-ratio of 2 : I from specata taken under ~de~t~ca~~0~~~~0~s to those used in the 35C1 From the data available it is not possible to imply that NADH binds to zinc in LADH but merely that the presence of NADH prevents the zincchloride interaction from taking place in aqueous sodium chloride solutions. Since the 35 Cl Iine broadening produced b LADH could arise from specific Zn” - J5 Cl 1‘Cl interactions or possibty from viscosity efinteractions, non-specific pro&’ f~cts33~~’ the role of zinc in L remains ~~dec~ded* Other dehydrogenases which should be amenable to study by means of the paramagnetic ion probe technique include the MnU activated enzymes, malic enzyme (EC. 1-I. 1.40) and isocitrate dehydrosenase (E-G. 1.1. I-4 1). XantIzine oxi&rse (EC.. I .2.3.X). Although molybdenum metallo-flavoproteins

MAGNETIC RESONANCE STl.KWS

Nalar

OF METAL-ENZYMES

425

Kltio f++$

Fig. 8. Plots af “Cl NMR line width against the molar ratio of NADH ts LADH (refs. 74 (bottom), 75 (top)).

are not readily amenable to paramagnetic

probe studies, due to long electronic times, they may be investigated by EPR. With the exception of milk xanthine oxidase, moIybdenum enzymes have not been frequently investigated by magnetic resonance techniques. It appears that the mcrlybdenum ion plays a direct role in the redox catalysis. In both xanthine oxidase and aldehyde oxidase (E.C.1.2.3.1.) the metal is involved in the int~~a&tion of the enzymes with reducing substrates, while in nitrate reductase (EC 1.6.6.X) it apparently interacts with the oxidising substrate7’. relaxation

Tfiese enzymes catalyse hydrolysis reactions. Carboxypeptidase A (E.C.3.4.2. I.) is a zinc metalIo-enzyme. The results of paramagnetic probe studies on this enzyme have been recently reviewed’. It contains two half-cystine residues whose function in the native enzyme has been the subject of much discussion ‘I. Protein thiol groups have been invest@ted by 3SC1 and recently this technique has been used to assay carboxypeptidase A for active thiol groups by titration with I-IgClz (ref. 86). Although Kg does bind to the enzyme it is not remove6d by dialysis indrcating that the binding is unlikely to be through a suIphur atom . This implies the absence of free thiol groups and lends support to the proposal of a disulphide bond linking the two half cystine residues8?.

lgF NMR data have been reported for salutims containing fluoride ions and carbox eptidase A in which Znri has been replaced by Mn’ (ref, 8%) An increase in the YBF ~~~a~ati~~ rate implies b~~d~~~between fluoride ions and Mn” which is snate *‘. From a ~~~sideratj~~ of the PRR e ted by ~-phe~yI~~ ted em ~~d X-ray d~ffr~~~~~~d~~~‘~ 5 on the Mn” sub r,ative enzyme it has been reported in both cases the ~urnbe~ of me bonds is five QPmares8_ ydrolysis of orthophosphork

s shown in Fig. 9. The dramatic

increase

in

Senee of four zinc atoms per male sf protein in th m speciftc XSivity of the

line not being a ~~rn~~sit~

12345678 Zinc:

ape-alkatine

phosphotase

9.

{molt rotid

Data sbtained fxorn the titration of apw3kaline phosph;rtise with ZrF ians84.

former is correct then the primary function of zinc in alkaline phosphatase appears ?o be to hold the protein in the active conformation rather than in promoting the

formation of a bridged enzyme-metal-substrate complex. However, the possibility remains that the substrate could displace one or more metal-enzyme Iinkages to @. PRR enhance ent data on binary and ternary Mnzl ~~kal~n~ phosphatase complexes could provide valuable information on this matter, Studies on the rate of chloride exchange with the metallo-enzyme would also help to clarify the function of zinc in alkaline phosphatase. There are a number of metallo-enzymes in this class which are candidates for future NMR ~nv~stjg~ti~~s using parama tic probes, these include the Ca” activated enzymes lipase (E.C_3_4_4_13 d myosin (ATPase) (E.G.36 1.3). Some further hydroiases are activated by Mg1I9 M#, Fe”‘, Co” and Z#. 6. Isomer~ses is small group of enzymes catalyse is~merisati~~ reactions. The members of D-xyiase isomerase is a Mg” activated enzyme which catalyses the interconversion

of D-xylose and D-xylulose, the results sf NMR studies on this enzyme have recently been reviewed’ . Some other M&‘-activated enzymes in this class which could lend themselves to parmagnetic probe studies using M# include arabinose isomerase (EC5.3.1.3.) and ~so~e~ty~ pyro~hosp~~te ~somer~se (E.CS~XSLL).

D. CONCLUDING

REMARKS

It has been our aim in this review to report on magnetic resonance a from interns which have extended or revised ~~~~~~~~ and c XUSiCMlS reached earlier9? ’ 0T50_The results obtained from PRR e~a~cerne~t and EPR data are largely complementary. YRR enhancement experiments are best suited to the determination of equilibrium data for ternary and quaternary complexes, while some structural information may be obtained if sufficient knowledge of the system On the other hand EPR measurements using spin labels can be more innoising subtle structural and dynanCc ~~rt~r~at~u~s in the immediate of the metal ion activators, Until recently little progress has been reported in the utilisation of alkaline or alkaline earth metal cations as NMR probes. The increasingly widespread utilisation of pulse and Fourier transform NMR techniques2’ will doubtless lead to studies on these and other quadrupoi~ nucfei in ~~~chemic~iy interested systemsgl. Bryant and his ~~~~rk~rs have reported tie results of some initial i~v~st~~ati~~s on %g, 3gK and 43Ca NMR in biochemical systemsg”-94. The addition of pyruvate’kinase to a 2 M solution of potassium chloride results roximate doubling of the 3%CNMR line width. Control experiments with rum albumin signify that direct binding occurs between potassium and metal-e~yme

428

P.J. QUILLEY,

G;A. WEBB

pyruvate kinase. The results of a 205Tl NMR investigation” indicate that MnH and Tl’ occupy positions in close proximity to one another on the enzyme. Obviously further information of this nature would be of great interest. Of ‘the non-metallic quadrupolar nuclei so far used as probes in enzyme systems, 35C1 has proved to be the most fruitful. It is particularly useful in investigating the number and strength of metal ion binding sites in zinc-enzymes and as a reporter in the study of conformationai changes occurring during the formation of binary and ternary complexes. Recently the 7gBr and 81Br nuclei have been used as NMR e/robes of protein meaconformation using pulsed NMR with promising resultsg6 and ‘Br relaxation surements have been used to study the Interaction of DNA with the Zn” metailoloo. I’0 and 23Na are other quadrupolar nuclei of bioenzyme DNA polymerase chemical interestg7-gy. It is to be hoped that further advances in NMR instrumentasystion will permit the omni resent 14N nucleus to be studied in macromolecular tems in the near future lop.

REFERENCES 1 B.L. Vallee, in P.D. Boyer, H. Lardy and K. Myrbach (Eds.), The Enzymes, Vol. 3, Academic Press, New York, 1959, p_ 225. 2 J.E. Coleman,Progr. Bio-org. Gem. 1 (1971) 159. 3 0. Lehrdnger, PhysioL Rev., 30 (1950) 393. 4 B.L. Valiee, Advart Protein Chem. 10 (1955) 317. 5 B. Maimstram and K. Rosenberg, Advan. Enzymol., 21 (1959) 131. 6 A.S. Mildvan, in P.D. Boyer (Ed.) The Enzymes, VoL 2, 3rd Edn., Academic Press, New ‘Lark, 1970, p. 447. 7 M. Witanowski and G.A. Webb, in E.F. Mooney (Ed.) Annual Reports on NMR Spectroscopy, VoL 5A, Academic Press, London, 1972, p. 395. 8 0. Jardetzky and J.E. Wertz,J. Amer. Chern Sot.. 82 (1960) 813. 9 AS. Mlldvan ‘md M. Cohn, Advan EnzymoL. 33 (1970) 1. 10 M. Cohn, Quart. Rev. Biophys., 3 (1970) 61. 11 G.A. Webb, in E.F. Mooney (Ed.), Annual Reports on NMR Spectroscopy, VoL 3, Academic Press, London, 1970, p. 211. 12 G.A. Webb in E.F. Mooney (Ed.), Annual Reports on NMR Spectroscopy, Vol. 6, Academic

-Press, London, in press. 13 14 15 16

N. Bioembergen and L-0. Morgan, J. Chem Phyr, 34 (1961) 842. 1. Solomon, Phys. Rev.. 99 (1955) 559. H. Sternlicht, J. Chem Phys., 42 (1965) 2250. H.G. Hertz, in J.W. Emsley, J. Feeney and L.H. Sutcliffe (Eds.), Progress in NMR Specfroscopy, Vol. 3, Pergamon Press, Oxford, 1967, p. 159. 17 G.H. Reed, J.S. Leigh and J.E. Pearson, J. Chem Phyr. 55 (1971) 3311. 18 N. Bloembergen, Nuclear Magneric Relaxation. Benjamin, NewYork. 1961, p. 90. -19.T.J- Swiftand R.E. Connick,J. Chem Phys.. 37 (1962) 307. 20 Z. Luz and S. Meiboom, J. Chenr Plays., 40 (1964) 2686. 21 J..Elsinger, R.C. Shulman and B.M. Szymanski,J. Cilem Pl?ys., 36 (1962) 1721. 22 M. Cohn, in E.B. Chance, C.P. &ek and J.K. Blaise (Eds.), Probes of S{rwture and Fqcrion of hiacromolecules and Membranes,. Vol. 1, Academic Press, New York, 1971, p. 97.

MAGNETIC 23 24 25 26 27

~E~~~AN~E

STUDIES

OF ~fETAL-ENZYMES

429

B.T. Hooton;Bioclremisf~, 7 (1968) 2063. J_ Reuben and M. Cohn, 1. Bio& Chem, 245 (1970) 6539. G.H. Reed, H. Diefenbach and M. Cohn, J. Biol. Chem, 247 (1972) 3066. B.B. Garrett and L.0, Morgn,J. CIIem PAYS.,44 (1966) 8% TX. Farrar and E.D. Becker, Pulse and Fourier ~ra~~s~~~~ CAR, Academic P~CSS,NW York, 1971.

Plenum Press, London, 19 73, p. 79. 30 A. AbraE;rm and R-V. Pound, P&s. Rev_. 92 (1953) 943. 31 T-R. Stem& and J.D_ Baldeschwciter, plot. /Vat_ Acad. S&i U_S,A+, 55 f 1966) 1020. 32 J.A. Happc and R.L. Ward, f, A~fec CIrem Sot_, 9 1 ( 1969) 4906. 33 R.L. Ward, ~~~~~~~~~~~r~~8 t 2969) 1879. 34 T.R. Stem@ and J.D. B~~d~~~~w~~~~r, J_ Amer. Ckrn Sue.. 89 ( 1967) 3045. 35 A-G. Marshall, Biochemisfry, 7 ( 1968) 2450. 36 R.G. Bryant, 3. Amer. C&m Sot., 91 (1969) 976. 37 J-L, Sudmeier and 3.3. Pcsek, AMZ. Bitxhem, 41 (1971) 39. 38 D.F.S. Natusch,J. Arner_ G’/2em Sex., 95 (1973) 1688. 39 A-R, Peacocke, R.E. Richards and B. Sheard, &&?I. Pltys., I6 ( X969) Jr77. 40 M.C. Scrutton and A.S. hfildvsn, ~~~c~!e~?t~s~?y~ 7 ( 1968) 1490. 41 B. Sheard and E.M. Bradbury, Progr. ~~~~~y~ IblclT.Biol.. 20 ( 1970) f 87. 42 G-H, Reed, hf. Cohn and W-J* 0’SuIiivan.J. .&‘ol, CItem. 245 (1970) 6547. 43 J-J. Villafranca and AS. Mildvan,X SioL Cilent, 246 (1971) 772. 44 J-J_ Villafiranca and AS_ bfildvan, 3. 8X CJle,n. 246 (1971) 5791, 45 J.J. Villafranea and AS. Mildvan, J. Biol C1lern. 247 (1972) 345446 J.P. Siater,T. Tamir, L.A. Loeb and AS. ~fildvan, .& Dial, Clef*, 247 (1972) 6784. 47 C-H. Fung. AS. MiIdvan, A, Allcrhand, R_ KQm~noski and MC, Scrutton. ~~~c~e?~~~r~, 22 (1973) 620. 48 A.S. Mildvan, R.D. Kobes and W.J. Ruttcr, Bfochemisrry, 10 (1971) 1191. 49 S.J. Bcnkovic, J.J. Villafranca and J.J. Kicinschustcr, A&#. Biochcm Bioph_vx, 155 ( 1973) 458. 50 A.% hfildvan, in E.B, Chance, C.P. Lee and J.K. Blaise (Eds.), plrcr&g ~fSttucr~~ and

S I hf,E, Fabry, S.H. Kijnig and W.E. Shiltinger, J. Biuf: Chem, 245 (1970) 4256. 52 D.M. Dawson, H.M. Eppcnberge and N.O. Kaplan, Biochem Biuohys. Res. Cornm~tf, 21 (1965) 346. 53 R.E. Benesch, H.A. Lordy and R. Bcnesch,J. Biol. Clreln, 216 (1955) 663. 54 J.S. Taylor, J.S. I-e&hand hf. Cohn,Proc. Nat. Aead. ScE U.S.A., 64 (1969) 219, 55 bf, Cufin, W, ~icfe~bach and J.S. Taylor, & B&l, C#~em, 246 (1972) 6037. 56 J.S_ Taylor, A. McC~ou~l~ and M. Cuhn, 2 BioL Cknr. 246 f f 97 1f 6029. 57 C,H_ Reed and M, Cohn, J. Biol, Cirem.. 247 (1972) 3073. 58 G.L. Cottam and R.L. Ward, Arcjr. Biochern. Biuphys., 132 (1969) 308. 59 T. Nowak and A.S. Mildvan, BiOdJE??2kfr)*, 11 ( 1972) 28 13. 60 T. Nowak and A.S. Mildvan, Bioc/zemisrry, 11 ( 1972) 28 19. 61 G.L. Cottam and AS, Mildvan,& Viol, C&m., 246 (1971) 4363. 62 J.C. Bailar,J. Inorg. NucL C&em, 8 11958) 165. 63 J.P_ Glusker,J, icfbf. SicrZ., 38 (1968) 149. 64 E.L. hfuetterties, Accuunfs Cbem Rtx, 3 t 1970) 266. 65 L.H. Pignolet, R.A. Lewis and R.H. Helm., J. Aner. Chem Sot., 93 (1971) 360.

66 A, Lanir and G. Navon, Biochemirfry,

10 (1971)

1024.

P.J. QUlLLEY, G.A. W-EBB

430

67 R.L. Ward, Biochemistry, 9 (1969) 2447. 68 M.C. Sm~iton, P. Griminger and J.C. Wallace, & BioZ. Cilem.. 247 (1972) 3305. 69 M,C. Smutton and A.S. Mildv~, A&. Sioc&?nriSfry BiQp!zys.., id0 (1970) I3 f .

70 RX_ Bray md J.C. Swarm, SIruct ~~~~~~g ~~e~~i~~* 11 f 1972) 107. md Kf. Fritz, Biochem ~~u~~~s. Res. Gommun.. 39 ( 1970) 707. 72 A.E. Dcnnaxd and R.J.P. WUiams, in R.L. Carlin (Ed.), Transition Metal Chemistry, Vat 2, Arnold, London, 1967, p. 116. 73 DE. Drum and B.L. Vallee, Bisclzemistry, 9 (1970) 4078. 14 R.L. Ward 9nd J.A. Happe, Biuchem, Biophys. Res. Commun, 43 (1971) !444. 75 B, ~~d~a~, hi, Zeppezauer and A. Akeson, Bischim Btbphys. Acra 257 (1972) X73. 76 M. Zeppezauer, B. ~~drn~~ S. Forsen and 1. indqvist, Bi0c&enr. Biq&ys Res Commun, 37 (1969) 137. 7 1 F&t. Wsd

77 M. Zeppczauer, personal communication. 78 R.L. Ward, personal communication. ‘79 I. Iweibo and H, Weiner, Bioclzemistry, 11 (1972) 1003. 80 R.T. Simpson and B.L. Vallce, Biochenzistr_v, 7 (1968) 4343. 83 J.C. Coleman and B.L. VaUee,J, Bicrl. CFzem,, 236 C1961) 2244. 82 C. Lazdunski, C. PetticIerc and M, Lazdunski, &‘ZK.I. ~i~~~ze~, 8 C1969) 5 10.

83 R.C. Bryant, personal communication. 84 G.L. Cottam and R.L. Ward, Arcl~. Biodzem. Biophys., 141 (1970) 768. 85 H. Csopak, B, Lindman and H, Lilja, FEBS Let&., 9 ( 1970) 189.

86 R.G. Bryant, Biochem Biophys. Res. Commun., 40 (1970)

1162. 87 W.N. Lipscomb, J.A. Hartsuck, G.N. Peeke, F.A. Quiocho, P.H. Bethge, M.L. Ludwig, T.A. S teitz, K Muirhead and J.C. Coppoia, ~~~~kh~~e~ Symposium, No. 2 1t 1968, p. 24. 88 C. Navon, R.G. 5~~~rn~, B.J. ~~y~~da and T. Yamaha, J, 4-M. BlioL. 5 f C1970) 15. 89 G. Navon, R.G. Schuiman, B.J. Wyluda and T, Yamane, Proc. Nat. Acad. ScL U.S.A.. 60 (1968) 86. ochem Biuphys. Res. Commurz., 45 (1971) 532. 90 R.G. Bryant and S. Rajender 91 T. Axenrod and G.A. Webb ( .)., NMR of Nuclei orher rhan Protons, fohn Wiley, New York, 1974. 92 R.G. Bryant,S. Meg. Res.,

93 R.G. Bryant,

6 (1972) 159. .J.C. Yeh and T.R. Stengle, ~i~c~~~ 3~~~~~~~ Rex. ~~rnrn~~, 37 Cl9693

603. 94 R.G. Bryan”&.F.Amer. Chem Sac. 91 (1969) 1870. 95 F.J. Kayne zmd 3. Reuben, .I, Amer. Chem Sot., 92 (1970) 220. 96 T.R. Collins, 2. Starcuk, A.M. Burr and E.J. Wells, J. Amer. Chem. Sot.. 95 (1973) 1649. 97 J. Reuben, h J.W. Emsley, 3. Feeney and t.H, Sutcliffe (Eds.), &~gress in NMR Spectioscopy, Vd 9, F~r~rn~~, Oxford, f 973, p. 1. 98 T.L. James and J,H. Noggfc, &UC. Naf, Acad Sci, OXA., 62 ~1969) 644. 99 D.H. Haynes, B.C. Pressman A. Kowalsky, Bziwhemisstty, 10 (197 1) 352. 100 C.F. Sprin Abramson, J.L. Engle and L.A. Lomb, .I. BioL Ch@&

248 (1973) S987. 101 M. Witanoweki, L. Stefaniak and H. Januszewski, in M. Witanowski and G.A. Webb (Eds.), Niirogen fVMR, Plenum Press, London, 1973, p. 163.