PORPHYRIN BIOSYNTHESIS-IMMOBILIZED ENZYMES AND LIGANDS-X. A NOVEL APPROACH TO THE STUDY OF THE RELATIONSHIP BETWEEN THE QUATERNARY STRUCTURE OF AMINOLAEVULINATE DEHYDRATASE AND ITS ACTIVITY

Abstract 1. 1. Evidence for dissociation, renaturation, re-association and re-hybridization of bovine liver aminolaevulinate dehydratase attached to Sepharose 4B is reported. 2. 2. When insolubilized enzyme was treated with 3 and 6 M urea, non covalently bound subunits were dissociated and detected in the eluate; these subunits can be re-associated into a soluble functioning enzyme with a specific activity close to that of the original pure soluble dehydratase preparation. 3. 3. After being washed with a renaturing buffer mixture, the matrix-bound subunits recovered a level of enzymatic activity equal to 50 and 20% of that of the immobilized native aminolaevulinate dehydratase. 4. 4. The reversibility of the dissociation process was investigated. Bound-subunits dehydratase can associate with nascent soluble bovine liver aminolaevulinate dehydratase subunits in situ. The product of such treatment, bound-re-associated enzyme, has the same activity as that of the original bound-dehydratase. The matrix-bound-dissociated bovine liver enzyme was also re-hybridized with soluble dehydratase subunits from E. gracilis . 5. 5. The apparent K m and optimum pH of the immobilized subunits were the same as those of the bound-octameric enzyme. 6. 6. A scheme is proposed, explaining the sequence of reactions leading from the bound-octameric dehydratase to the possible different derivatives, formed during the dissociation and re-association experiments.


INTRODUCTION
The importance of quaternary interactions is indicated by the fact that most soluble globular proteins exist as oligomers (Klotz et ul., 1970). The reversibility of dcnaturation and dissociation of these proteins has largely been investigated and can be used to advantage for studying the mechanisms by which a polypeptide chain folds into a specific conformation: also the assumption is often implicitly made that the assembly of subunits, giving rise to the quaternary protein structure, involves binding domains which are unique for every set of subunits. lmmob~iization of enzymes has proved useful for elucidating the problem of subunit assembly, inter-subunit interactions and structure-function relationships in oligomeric prorcins.
A novel approach to the study of matrix-bound subunits was first introduced by Chan (1970). who dem[~nstrated the existence of catalytically active aldolase subunits bound to Sepharose. This valuable procedure was then applied to the investigation of several other multi-subunit enzymes (Chan & Mawer, 1972;Chan, 1973; Nagradova t't LII., 1974; Feldman et cd., 1976;Bruch et al., 1976;Carvajal rt nl., 1977).
In studying the relationship between the quaternary structure of an oligomeric enzyme and its activity, an interesting point is to know whether the individual subunits are active or whether association is necessary for enzymic activity. For this purpose, the intact enzyme is attached to a matrix via a single or at least two subunits, and then, those not covalently bound are removed; so that the properties of the isolated immobilized subunits can be studied under conditions where re-association is prevented. Further, this method may also be used to investigate re-association and hybridization of the bound subunits with soluble subunits.
This approach appeared particularly attractive for studying aminolaevulinate dehydratase (ALA.D) (5-aminolaevulate-hydrolase, EC 4.2.1.24), the enzyme catalysing the condensation of two molecules of aminolaevulinate (ALA) into the monopyrrol porphobilinogen (PHG): which has been extensively investigated in our laboratory (Stelia & Batlle, 1977 and refs therein) and many others. Bovine liver ALA.D was used for this study, because we can easily obtain relatively large amounts of highly purified enzyme (Batlle et al., 1967(Batlle et al., , 1970Stella & Batlle, 1977) and we already know the properties of ALA.D attached to Sepharose (Stella et nl.. 1977).
It has been shown that the enzyme has a molecular weight of about 2X0.000 and is composed of eight similar subunits. Substantial work has been done on ALA.D. and an excellent review on its structure. function and mechanism of action has been recently published by Shemin (1976).
PrcCous work with the K/I. spl~e~oitle~ ALA.D (Nandi. 1971) has shown that this enzyme, like few others. undergoes reversible inhibition by urea. guanidine and methyl derivatives of urea. and that this inhibition is competitive in nature. However. it was simultaneously demonstrated that the inactivation ol ALA.D by urea is also produced by its dissociation into subunits. Thcrcfore. we may expect that bovine liver ALA.D can be dissociated by urea into individual polypeptides. and that by removing or diluting this reagent. the subunits would then reassociate to form the active octameric oligomcr.
WC report here evidence for dissociation. renaturation. re-association and rchybridiration of bo\ ine li\er ALA.D attached to Sepharose and we propose a scheme for describing the dill'crent matrix-bound derivatives obtained.

MATERIALS AiX;D METHODS
Immobilized bovine liver ALA.D wa\ prepared following the procedure reported by Stella PI trl. (1977). cwcept that Sepharose 4B was activated using X-50 mg of cyanogen cnzymc (or 120 units of soluble E. y,ncilis ALA.D) per ml of gel, through a packed column: and finally washing with renaturing butfer mixture until no further protein and activity was eluted.
From preliminary comparative experiments performed by using a packed column and a batch procedure.
it was found that both dissociation and rc-association of the bound enzyme was superior when employing the former system.
The protcln cluates obtained after treatment of the bound-enzyme with urea and samples of soluble ALA.D both native and urea treated. were subjected to electrophoresis in polyacrylamide-gels.

RESl LTS AND DISCUSSION
Taking into account data reported by Nandi (1971). preliminary experiments were carried out to study the in r>ir~o effect of different concentrations of urea on the activity of soluble bovine liver ALA.D (Fig. 1). Inhibition rapidly increases with increasing concentrations of urea; 1 M urea brings about 80". inhibition which is practically IOO",, at 2 M. It is also shown that the inhibition by 1 M and 2-3 M urea is KY,, and 70",, reversible. respectively. This work seems to be in line with that of Nandi (1971), therefore supporting his hypothesis about the dual effect of urea on the activity of ALA.D.
On the basis of these results. it was feasible to apply Ghan's method (1970) for dissociating the Sepharose-ALA.D octamer into subunits, and characterize some of the properties of both the bound and solubilized subunits. by using urea as the dissociating agent.
The study of Sepharose-ALA.D subunits, implies the binding of intact octamer enzyme to the gel. followed by treatment with dissociating agents, generally protein denaturants, to remove those subunits not covalently attached and finally. renaturation of both products.
The specific activity of Sepharose-ALA.D was approximately 30",, of that of the soluble pure enzyme used in its preparation.
In a previous paper (Stella & Rattle. 1977) we obtained matrix-bound ALA.D with specific activity as high as 7@75",, that of the enzyme in solution. but it might be recalled that both the amount of cyanogen bromide used in the activation of the gel and that of the protein added were higher than those used in these experiments. We have already discussed the reasons why the activity of the bound enzyme is lower than that of the native soluble enzyme. These amounts were deliberately diminished, to reduce the degree of multi-point attachment of ALA.D.
When a column of immobilized bovine liver ALA.D was treated with 3 and 6 M urea, as is seen in Table 1, enzymic activity was detected in the eluate; samples of the gel removed from the column and washed with the renaturing buffer mixture, also showed activity in all cases, although the percentage of the original activity regained was dependent on the concentration of urea used. After elution with 3 M urea the matrix-bound subunit recovered a level of enzymic activity equal to 50:;, of the immobilized native ALA.D, treatment with 6 M resulted in a derivative with only 20",, remaining activity: while specific activities were close to 4 and 2 respectively.
It is noteworthy that the specific activity of the solubilized subunits was twice that of the insolubilized ALA.D and near that of the original pure soluble preparation.
The fact that after treatment with urea, 50",, and close to 25:,, of the activity bound to the matrix remained. suggests a change of the bound enzyme from an octamer into a tetrameric and dimeric state. We do not actually know at present the number of sites of attachment of the ALA.D to the gel; however these findings indicate that no more than two subunits are involved in the binding to the support, although wc cannot exclude either, that under variable experimental conditions. the coupling could occur through a single or even through more than two subunits. Moreover, different concentrations of urea appears to specifically affect the inter-dimeric contact in the octamer.

Rt,-a.s.socicrrion e.xpcritwwt,s
The reversibility of the dissociation process was investigated.
to see if the matrix-bound derivatives would pick up subunits added in solution. The Sepharose-ALA.D dissociated. packed in a column was eluted with a decreasing discontinuous gradient of 2-O M urea, in the presence of soluble ALA.D as described above. After extensively washing with the renaturing buffer mixture until the eluate was free of ALA.D activity; the gel was assayed to detect whether bound subunits had re-associated or re-hybridized with subunits in solution to re-form the matrix-bound octamer. The data of Table 2 show that the above treatment increased the activity of the bound subunits derivatives up to a value equal to that of the bound native ALA.D. These results indicate that the immobilized subunits are able to re-associate with the  in situ, to yield a product resembling the original bound-enzyme.
These effects were specific, since control experiments in which columns were washed, either with the urea gradient in the absence of soluble bovine liver ALA.D. or with soluble native ALA.D in the same buffer mixture but without the urea gradient, showed practically no increase in the activity of the insolubilized dissociated enzyme.
The specific activity of the re-associated boundenzyme also approached that of the insoluble ALA.D. confirming restoration of the native octamer structure. It was also found that the bound species picked up an amount 05 protein and enzymic activity almost equal to that initially lost by the insolubilized octamer.
In one experiment, the bound-ALA.D dissociated was re-hybridized with soluble E. graciks ALA.D; in this instance, restoration of 75",, activity also occurred. This suggests that it might be possible to form active hybrids of the bovine liver and the E. grucilis enzymes, and it further shows, that if there are any differences in the aminoacid sequence between these two dehydratases. they seem not to significantly modify the inter-subunit contact sites: so that, domains involved in subunit interaction appear to have been preserved in the evolution of thcsc species.
Therefore, this technique proves useful for hybridization studies, between dehydratases from different species or between native and chemically modified subunits, and complementation investigations including restoration of activity by non-covalent interaction of enzyme subunits. To obtain some insight on the enzymic properties of the insolubilized derivatives here prepared, their optimum pH and K, values were calculated. From data listed in Table 3, as already found (Stella ef trl.. 1977), the kinetic properties of ALA.D are altered by its binding to Sepharose; a decrease in the K, of the matrix-bound enzyme has been observed, while there is practically no difference in the optimum pH between free enzyme and its insoluble analogue. Further, the same optimum pH and K,s values were calculated for the matrix-bound ALA.D subunit derivative, indicating a similarity in the catalytic properties of the different oligomeric forms of the enzyme. and raising the question of why the octameric structure is actually necessary. Different samples of bovine liver ALA.D. either native or treated with urea and solubilized subunits. were subjected to polyacrylamide gel-electrophoresis (Fig. 2). Molecular weights of the resulting protein bands were calculated. It was found that in the presence of 3 M urea, a mixture of octamer and tetramer enzyme is present. while in 6 M urea, octamers, tetramers and dimers are visualized. Recombination of the solubilized subunits could also be detected by this procedure.
where again the main species were octamers and tetramers.

CONC'LCISIONS
The aim of this work was to investigate the role of the octameric structure of bovine liver ALA.D on the conformation and function of this enzyme. Many oligomeric enzymes are formed by identical or closely similar subunits. containing each an active site: so the significance of a quaternary structure is not very clear. However. it appears that evolution has chosen, for the sake of regulation.
that the potential for cooperativity must be preserved, so that many proteins might become active in regulated catalysis as an oligomer.
The results reported here. tempted us to propose a scheme that, although rather speculative, would describe the sequence of events which lead from the bound octameric ALA.D to the different subunits derivatives (Fig. 3).
The immobilized native octameric ALA.D (1) seems to be covalently linked to the gel by an average of two subunits. Treatment with less concentrated urea would remove a tetramer (2)

structure,
re-assembling into normally soluble functionlng ALA.D (4); while the bound tetramer is still enzymically active (5) it can be dissociated further by higher concentrated urea into a bound-dimeric derivative (6); suggesting that there should exist different types of binding between dimers and tetramers (shown as u and h in the insoluble derivative number 1). Finally, both of these bound-dissociated species (5 and 7). are able to pick up soluble subunits and restore the original octameric structure (8); confirming that urea has depleted the immobilized ALA.D of subunits, but that these can be replaced under convenicnt re-association conditions. Since the immobilized subunit derivative still exhibit catalytic properties, it would appear that each catalytic site could function somehow independently of the other, but only within a dimer formation. On the other hand, it has already been established that only 4 of the 8 subunits of ALA.D form a Schiff base with one molecule of substrate. indicating that the enzyme exhibits the phenomenon of half-site reactivity, and according to the mechanism postulated by Nandi rt a/. (1968) only one of the two molecules of ALA in the reaction, forms a covalent bond with the enzyme: it is therefore possible, that we are dealing with a minimal functional dimer formed by two kinds of subunits, that, although having similar composition, play a different role in PBG synthesis. We propose that one subunit might be involved in the formation of the Schiff base with one molecule of ALA and the other in the non-covalent binding with the second molecule of substrate. Nonetheless, these are just assumptions, for the available experimental data cannot yet define which is the minimal structure necessary for activity. Experiments on this line are in progress in our laboratory and further results are eagerly awaited.