Comparative Structural and Immunochemical ... - Wiley Online Library

Eur. J. Biochem. 66, 389- 399 (1976)

Comparative Structural and Immunochemical Properties of Leghaemoglobins John G. R. HURRELL, Nicos A. NICOLA, William J. BROUGHTON, Michael J. DILWORTH, Elizabeth MINASIAN, and Sydney J. LEACH Department of Biochemistry, University of Melbourne, Victoria School of Biological Sciences, University of Malaya, Kuala Lumpur and the School of Environment and Life Sciences, Murdoch University, Western Australia (Received December 29, 1975/March 29, 1976)

Circular dichroism studies on leghaemoglobins from snake bean, lupin, serradella and other plants show that, in common with soybean (reported earlier) they have a similar overall polypeptide chain conformation and haem environment and orientation. Immunochemical studies, on the other hand, suggest that the antigenic determinants on the surface of the leghaemoglobins vary considerably. Thus, firstly the a-helix content of the leghaemoglobins as a class is very similar (60- 65 and approaches that of the myoglobins, secondly, the sign, magnitude and shape of their circular dichroism spectra in the near ultraviolet, Soret and visible regions suggest close similarities in the environment and orientation of a structurally important tryptophan residue and of the haem moiety, and thirdly, there is comparatively weak haem-protein interaction. The extent of immuno cross-reactivity was found to be best demonstrated using the Farr radioimmunoassay procedure. The results were (a) 5 leghaemoglobins from one plant (soybean) crossreacted completely but with varying affinities. (b) The extent of cross reactivity between leghaemoglobins from different plants was compared to that within a single plant; the reaction of antiserum to a soybean leghaemoglobin with a serradella leghaemoglobin was weak, with a snake bean leghaemoglobin still weaker (and incomplete) while lupin leghaemoglobins showed no cross reactivity at all. (c) The “rapid” attenuation of cross reactivity among different plant leghaemoglobins is explicable in terms of the extensive amino acid substitutions which have been demonstrated in the literature and in the present studies. (d) In view of this rapid divergence it is not surprising that sperm whale and horse heart myoglobins showed no cross reactivity with soybean leghaemoglobins. In summary, amino acid substitutions in the leghaemoglobin family are conformationally but not immunochemically conservative.

x)

The evolution and genealogy of haemoglobins continues to interest molecular biologists with respect to changes in amino acid sequence, conformation and function. Comparisons of 55 contemporary globins ranging from plant to human haemoglobins show that only two amino acids have remained unchanged during more than 700 million years of evolution. By contrast, the changes in chain conformation have been very small over the same period (see for example, Vainshtein et al. [l]). The rates of evolution of the globin genes have, however, varied considerably, being rapid in the earliest studied stages (up to the ____ Abbreviation. CD, circular dichroism.

invertebrate-vertebrate ancestor to the chicken-mamma1 ancestor) and slower from that point to the present time (see Goodman et al. [ 2 ] ) . Little is known about the nature and rates of change of globin sequences and conformations prior to the insect and annelid divergences. The availability of many leghaemoglobins from modern plant descendants, however, should permit a study of the earliest ancestry of the globin family. Dozens of such plant species are available and the nodules of a single leguminous plant are known to produce as many as 5 or 6 different leghaemoglobins. Already, the amino acid sequences of leghaemoglobins from the root

390

nodules of soybean [3], kidney bean [4] and broadbean [5] are known and others are being studied in our laboratory and elsewhere. At the same time, the first X-ray crystal structure determination of a leghaemoglobin has been carried out at a resolution of 5 A, namely that of lupin, while circular dichroism studies have compared the chain folding and the detailed haem environment of a soybean leghaemoglobin with that of sperm whale myoglobin [6,7]. It has been generally observed that genetic mutations affecting the surface residues of proteins are more readily accepted than those affecting the interior of the molecule and that this effect will therefore be most noticeable for small monomeric proteins [8]. We have commenced an examination of the evolutionary changes occurring within the Leguminosae plant group using three experimental approaches. Some preliminary results on amino acid sequence variations within the leghaemoglobins of a single plant species (soybean Glycine max. cr. Lincoln) will be reported very briefly here. The main purpose of the present paper, however, will be to describe the results of studies designed to contrast the changes occurring on the outside of leghaemoglobin molecules with those occurring within the molecules and in their haem crevices. Surface changes are discussed in terms of immunochemical cross reactivities, while conformational comparisons are made by measuring circular dichroism spectra between 190 and 650 nm. Comparisons of these two sets of properties are made between leghaemoglobins of the same plant, between leghaemoglobins of plants of differing phylogeny and finally between plant leghaemoglobins and mammalian myoglobins. MATERIALS AND METHODS Preparation of Leghaemoglobins

Experimental procedures for extracting, fractionating and analysing soybean leghaemoglobins are described in an earlier publication [9]. The five fractions obtained by this procedure are referred to as leghaemoglobins a b c1 c2 and d in the order in which they were eluted from a DEAEcellulose column. The major components are leghaemoglobins a c1 and c2. Leghaemoglobins from the nodules of snake bean [long bean or Vigna sinensis (L.)] were extracted as follows. The homogenate from 200 g of nodules in 4 volumes of phosphate buffer (0.1 M, pH 6.9, 4 "C) was clarified by centrifugation and the supernatant was fractionated using ammonium sulphate. The fraction precipitating between 50 and 80% saturation (at 4 "C) was redissolved in phosphate buffer and dialysed overnight against EDTA (0.001 M, 30 1). After centrifugation, the supernatant was fractionated on Whatman DE-52

Circular Dichroism and Immunology of Leghaemoglobins

cellulose with sodium acetate buffer (pH 5.2, 0.013 M) - see also [lo]. The first fraction eluted at this pH was concentrated to about 10 mg/ml by dialysis against Aquacide II-A (Calbiochem, La Jolla, U.S.A.) and then dialysed against phosphate buffer (pH 6.8, 0.05 M) and converted to the aquoferric form as follows. A Sephadex G-25 column (30 x 3.5 cm) was equilibrated with the above phosphate buffer at 4 "C and a solution of ferricyanide (0.08 g for each 0.05 g of leghaemoglobin) containing sucrose (20 %) was run onto the column. The leghaemoglobin solution was then loaded and allowed to run through the ferricyanide layer with quantitative conversion to the aquoferric form. The leghaemoglobin fractions were pooled and concentrated (3 - 5 mg/ml) by dialysis against Aquacide, transferred to plastic tubes and frozen to - 196 "C until required. Leghaemoglobins from lupin [Lupinus luteus (L.)] and serradella (Ornithopussativus Brot.) were prepared in a similar manner but not subjected to the final ferricyanide oxidation before use [ll]. Further analytical details and the methods for converting leghaemoglobins to their various liganded states are described elsewhere [7]. Circular Dichroism Measurements

'

The equipment and procedures used have been described earlier [7]. It was found most important to carry out both CD and concentration measurements in replicate on freshly fractionated leghaemoglobin samples. When necessary, samples were stored as frozen solutions and then for the minimum possible time at the lowest possible temperature. Radioactive Iodination of Leghaemoglobins

Soybean leghaemoglobin a and c1 were iodinated using a modification of the chloramine-T method [12]. 1 mCi of carrier-free Na ''1 (Radiochemical Centre, Amersham, England) of specific activity > 14 mCi/pg was added to the protein (200 pg) in phosphate-saline buffer (pH 7.4, 50 pl) at 0 "C. Labelling was initiated by adding chloramine-T (2.5 pg). After 15 min the reaction was terminated by adding sodium metabisulphite (5.0 pg) and the unreacted iodine separated from the iodinated protein by passage down a Sephadex G-25 column pre-equilibrated with phosphate-saline buffer containing 0.5 % bovine serum albumin. The radioactivity was found to be 95 - 96 % trichloroacetic-acid-precipitable and the efficiency of '''1 incorporation was 55 % of the total added. Preparation of Antisera

Antisera to soybean leghaemoglobin a and c1 were prepared in young castrated male sheep by intra-

J. G. R. Hurrell, N. A. Nicola, W. J. Broughton, M. J. Dilworth, E. Minasian, and S. J. Leach

muscular injections of protein (1 mg) emulsified in Freund's complete adjuvant (1 ml), a quarter of each dose into each flank. Six weeks after the priming injection, the first booster injection was given and this was repeated at 3-week intervals for a total of 9 weeks. Trial bleedings at weekly intervals showed that 15 weeks after the primary injection, the animals were hyperimmune as judged by conventional quantitative precipitin assays. One litre of blood was collected from each animal per the jugular vein at weeks 16 and 18. The blood was allowed to stand at room temperature for 6 h and then at 4 "C overnight to shrink the clots. The sera were decanted from the clot and centrifuged to remove residual cells. The sera (8001000 ml per animal) were stored at -20 "C in 100 ml lots. Prior to use, a 100 ml lot was thawed at room temperature and complement-inactivated by incubation at 56 "C for 20 min.

Xmmunochemical Cross Reactivities of Leghaemoglobins Method I . Leghaemoglobin (0 - 150 pg) in Trissaline buffer (pH 8.1, 40 mM Tris-HC1, 90 mM NaCI, 200 pl) was added to inactivated antiserum (apropriately diluted to 200 pl). After incubation (30 min, 37 "C), the assay tubes were left for 48 h at 4 "C. Precipitates were centrifuged down, washed twice with buffer (pH 8.1, 0.1 M Tris-HCl, 0.15 M NaCl) and dissolved in NaOH (0.1 M, 1 ml). Optical absorbancies at 280 nm were recorded and plotted against the amount of leghaemoglobin in each tube. Each assay was carried out in triplicate. Method 2. Double diffusion assays were carried out in Petri dishes containing agar (1 % Difco, dissolved in 20 ml borate-saline buffer, 0.2 M, pH 8.5 containing 0.02 Thiomersal as preservative). A centre well was used for the undiluted antiserum and five peripheral wells for leghaemoglobin solutions (1 mgiml). Plates were developed at room temperature for 48 - 60 h in a humid atmosphere. Method 3. In this method, ammonium sulphate was used to precipitate soluble as well as insoluble immune complexes [13]. Aliquots of iodinated leghaemoglobin (30 ng in 500 pl, containing approximately 20000 counts/min) were transferred by Eppendorf pipette into polystyrene tubes (Camalec V.H.A., Melbourne). Varying amounts (1 ng to 1 mg) of competing unlabelled leghaemoglobin, either homologous or heterologous (in quadruplicate) were added by microsyringe. Antiserum (500 1-11 previously diluted 1 in 20000 using borate-saline buffer, pH 8.4 and containing 10% normal sheep serum) was added to each tube. This antiserum dilution had previously been found to bind 7076 of the total radioactivity added. The stoppered tubes were mixed and incubated for 24 h at 37 " C , then cooled to 4 "C. Saturated ammo-

391

nium sulphate (1 ml) was added to each tube at 4 "C while mixing on a vortex mixer. After standing for 30 min at 4 "C, the tubes were spun at 4000 rev./min for 10min in an MSE bench centrifuge, the supernatant decanted and the precipitate washed twice with 50 % saturated ammonium sulphate; care was taken to completely re-suspend the precipitate between each aash. The radioactivity in the precipitate was counted in a Packard auto-gamma counter. The percentage of the total radioactivity bound to immunoglobulin was calculated by a method [14] which corrects for nonspecific binding, this being proportional to the amount of unbound labelled leghaemoglobin in the supernatant of each tube. A normal serum control was included in each series of tubes to provide the data for this correction.

Amino Acid Analysis and Peptide Mapping The basic methods used were described earlier [9]. The detailed identification of peptide fragments and sequencing of soybean leghaemoglobin CI and cz is incomplete and will be published later. RESULTS AND DISCUSSION

Circular Dichroism Spectra The far ultraviolet CD spectra of snake bean, lupin and serradella leghaemoglobins were very similar in the shape and magnitude of their negative extrema at 220- 222 nm (Table 1). Variations between the different oxidation states and complexes are within experimental error. The values reported here agree well with those recently observed for leghaemoglobins from Trifolium subterranaem (Thulborn, unpublished). Previously published a-helix estimates [6,7] are lower and all of the values reported here correspond to 60 and 65 %. One may conclude that leghaemoglobins as a class have a similar high a-helix content, approaching that of the mammalian my oglobins. In the near ultraviolet, the CD spectra of soybean leghaemoglobin a, lupin leghaeinoglobin a, serradella leghaemoglobin I1 and snakebean leghaemoglobin a all have qualitative features in common. The major aromatic bands appear to be positive (in contrast to those in myoglobin [7] where they are negative) and they are superimposed in most cases, on a broad haem CD band centred around 260 nm which, compared with myoglobin [7], is only weakly positive. The sharp positive peak at 291-292 nm in all spectra and the ill-defined broader negative one at 295 - 297 nm in Fig. 1A and B probably arise from Lb and La tryptophan transitions respectively, in keeping with the presence of a common tryptophan residue which is rigidly held in a non-polar environ-

392


Table 1. Mean residue ellipticities ([0]222/degree x emZx dmol-‘) for some leghaemoglobins and their complexes The proteins (0.1 to 0.2%) were dissolved in phosphate buffer (pH 6.8, 0.05 M) Leghaemoglobin Source

[01222

type

ferric aquo

ferrous cyano

nicotino

OXY

deoxy

degree x cm2 x dmol-’ Snake bean Lupin Serradella

I a I1

- 21 400 - 22700

- 20 500 - 22200

- 21 000 -

- 21 000

- 20 500 -

- 22200

- 20300

- 21 300

- 21 000

- 21 100

-

0

A* ?

-20

0

- 30

260 Wavelength 280(nm) 300 - 40

Wavelength (nm)

Fig. 1. Near ultraviolet C D spectra of Ieghaemoglobins. (A) (-) Lupin leghaemoglobin a , ferric; (. . ’ . .) serradella leghaemoglobin 11, ferric; ( ---) lupin leghaemoglobin a, cyanoferric. Conditions : phosphate buffer (pH 6.8, 0.01 M), concentrations 0.3 - 0.4 %. (B) Snake bean leghaemoglobin a (-) oxyferrous; (...-.) deoxyferrous; (-----) cyanoferric. Conditions: as (A) except for oxyferrous and ferrous which were in phosphate buffer (pH 7.05, 0.02 M)

ment as in myoglobin and several other haem and nonhaem proteins [7]. This tryptophan is doubtless the one which, in thermal perturbation difference spectroscopy, also behaves as though it were in a non-polar situation (see [6], Nicola et al. unpublished). The rigidity is a pre-requisite for a well-defined C D pattern and a non-polar environment would explain the red shift as well as the sharpness and resolution of the aromatic peaks. The weakness of the broad haem envelope around 260 nm is indicative of a weaker interaction between the haem moiety and the protein than in myoglobin. Of the leghaemoglobins examined, snake bean leghaemoglobin a has the most positive ellipticity at 250 nm (Fig. 1 B), especially in the oxyferrous form and one might therefore anticipate that snake bean leghaemoglobin has a somewhat “tighter” haem pocket than the other species. We have observed that this leghaemoglobin is some-

what more stable than other leghaemoglobins to autoxidation on storage in the frozen state. A study of the relationship between C D characteristics in this spectral region, and the 0 2 and other ligand-binding affinities of the leghaemoglobins should be rewarding. The Soret C D spectra of leghaemoglobins from soybean, lupin and serradella are not only qualitatively but also quantitatively alike (see Fig. 2), with closely similar negative extrema at 403 - 404 nm, and therefore opposite in sign to those of mammalian myoglobin [7]. Fig. 2A shows, however, that on conversion to the cyanoferric derivatives, the Soret CD bands for all three species diminish to about one third of their intensity and shift to 412-413 nm. Fig. 2 B shows that oxygenation of snake bean ferrous leghaemoglobin shifts its Soret C D band from 422 nm back to 404 nm without a change in intensity. The net sign and magnitude of the Soret C D band in haem

J. G. R. Hurrell, N. A. Nicola, W. J. Broughton, M . J. Dilworth, E. Minasidn, and S. J. Leach

A

393

,.

--50

--

B

4

l

l

I

1

1

1

1

1

1

1

350 400 Wavelength (nrn)

1

1

1

)

4 50

500 550 Wavelength (nm)

600

650

Fig. 3. Yisible-region C D specit.rr of snake hcan leghaemogiohzn a. (-) Ferrous; (. . . . .) oxyferrous. Conditions: phosphate buffer (pH 7.05, 0.02 M), concentrations 0.2-0.4%

1

450

Fig. 2. Soret-region C D spectra of Ieghaemoglohins. A : (-) soybean leghaemoglobin a, ferric, pH 7.03; (.... .) lupin leghaemoglobin a, ferric, pH 6.8; (----) serrddella leghaemoglobin 11, ferric, p H 6.8 ; (-O-) soybean leghaemoglobin a, cyanoferric, pH 7.05 ; (. . . .) lupin leghaemoglobin a, cyanoferric, pH 6.8 ; (- .-. -) snake bean leghaemoglobin a, cyanoferric, pH 6.8. B: (- -) snake bean leghaemoglobin a , ferrous, pH 7.05; (....-) snake bean leghaemoglobin a, oxyferrous, pH 7.05. All at concenin sodium phosphate buffer (0.02 M) trations 0.2-0.8

proteins is a summation of ellipticity contributions from many coupled electronic transitions involving the haem and its aromatic side-chain contacts with the protein [15]. It therefore seems the more remarkable that this group of leghaemoglobins from different plants, differing (as we shall see) in amino acid sequence, should all show qualitatively and quantitatively, such closely similar Soret CD spectra to each other, but all opposite in sign and magnitude to those of myoglobin [7]. CD spectra in the visible are summarised in Fig.3 and 4. The spectra for the ferrous and oxyferrous derivatives of snake bean leghaemoglobin a (Fig. 3) are similar to those of soybean leghaemoglobin a [7]. There are four optically active transitions, two positive and two negative, a resolution which is not apparent in the corresponding visible absorption spectra. The cyanoferric leghaemoglobin a derivatives from soybean, lupin and snake bean (all low-spin, like oxyferrous) have remarkably similar visible CD spectra (Fig.4) their three negative ex-

4

Wavelength (nrn)

Fig. 4. Visible-region C D speclru of the cyunqferric dcrivarive.sof'myoglohin and leghaemoglobins. (--- ) Sperm whale myoglobin; (--) soybean leghaemoglobin a ; (. . . ..) lupin leghaemoglobin a ; (-.-. -) snake bean leghaemoglobin a ; (---) snake bean leghaemoglobin a nicotino-ferric. Conditions: phosphate buffer (pH 6.8, 0.02 M) except for myoglobin which was at pH 7.05; concentrations 0.2- 0.6

394

Circular Dichroism and Immunology of Leghaernoglobins

Fig. 5. Double difsuon on agarplates of’ieghaemoglobinsagainst sheep antiserap~eparedagalgainstthem. (A) Two upper plates : antisera to soybean leghaernoglobin a (centre well, top left) and soybean leghaemoglobin c1 (centre well, top right) versus leghaemoglobins a b c, c2 and d. (B) Lower plate: antiserum to soybean leghaemoglobin c1 (centre well) versus soybean leghaemoglobin a and c, and snake bean leghaemoglobin a (sb)

trema being almost mirror-images of the three positive extrema of myoglobin. Only their fourth extremum at 470 nm is similar in size and sign to that of myoglobin. Again, the similarity in haem orientation and environment among the various leghaemoglobins, and the difference between them and myoglobin is emphasized.

Immunoclzemical Properties of’Legghaemoglohins

Antibody titres for sheep immunised with soybean leghaemoglobins for 16 weeks were (for single animals) : 3.2 (leghaemoglobin a), 15.2 (leghaemoglobin CI) and 8.0 (leghaemoglobin CI) mg/ml respectively. In contrast to earlier reports on the immunogenicity of myoglobins in rabbits [16], we have experienced little difficulty in obtaining high antibody titres from sheep or goats using myoglobin or leghaemoglobins with our immunisation schedule. Antisera to soybean leghaemoglobin a and c1 both gave precipitin lines of identity with no spurs,

when each was tested against soybean leghaemoglobin fractions a b CI cz and d (see Fig. 5 A) suggesting that components a and c1 have common determinants with each other and with b cz and d and also that there are no determinants peculiar to a and cl. Spur formation was, however, observed when antiserum to soybean leghaemoglobin c1 diffused against snake bean leghaemoglobin a (Fig. 5 B). This result suggests that soybean leghaemoglobin CI , while sharing some common determinants, has one or more determinants which are not common to snake bean Ieghaemoglobin a and which produce and antibody population which does not cross-react with the snake bean protein. Other double diffusion experiments (not shown in Fig. 5 ) showed no cross-reactivity between soybean leghaemoglobin CI and lupin or serradella leghaemoglobins. Precipitin curves (Fig.6A, B and C) bear out the conclusion that the five soybean leghaemoglobin proteins all cross-react with one another. These curves give a better indication than the diffusion plates of the extent of cross-reactivity. Antisera to leghaemo-


0 Leghaemog 10 bin

(bg)

I"

50 100 Leg haemoglobin

395

150

(vg

I

Leghaemoglobin (pg)

Fig. 6 . Precipitin curves comparing the extent of cross-reactivity benveen the five leghaemoglobins from soybean. Antisera were titratedwith soybean leghaemoglobins a b CI c2 and d. (A, B) Antisera to leghaemoglobin CI (two different sheep). (C) Antiserum to leghaemoglobin a (one sheep). The different leghaemoglobins are; (0)a ; ( 0 )b; (0)C I ; (A) C Z ; (A) d.

globin c1 (Fig. 6A and B) fail to differentiate between their homologous antigen leghaemoglobin c1 and any of the other soybean leghaemoglobins ; this result was obtained qualitatively by double diffusion and here quantitatively for antisera from two different animals. On the other hand, antiserum to soybean leghaemoglobin a (Fig. 6C) showed marked preferences; thus, while the antiserum reacts equally well with leghaemoglobin b and its homologous antigen leghaemoglobin a, the reaction with leghaemoglobin c2 and leghaemoglobin d is markedly less, and the cross-reaction with leghaemoglobin c1 is very small indeed. In summary, for anti-(leghaemoglobin a), a = b > d = c2 > cl. The lack of reciprocity between the two sets of curves (Fig.6A, B and C ) points to subtle differences between determinants - particularly between leghaemoglobin a and leghaemoglobin c1. All antibodies to leghaemoglobin c1 react equally well with leghaemoglobin cI and leghaemoglobin a, but antibodies to leghaemoglobin a react poorly with leghaemoglobin c1. One might postulate that the antibody combining site for a common leghaemoglobin a/leghaemoglobin CI determi-

nant differs sterically or electrostatically in such a way that the anti-(leghaemoglobin cl) site can accommodate the leghaemoglobin a determinant but the anti-(leghaemoglobin a) site can accommodate the leghaemoglobin c1 determinant only with difficulty and therefore displays a low binding affinity. The precipitin test is critically dependent on affinity constants [I71 and these constants can be modified by changes in amino acid sequence either at or near the antigenic determinant [18]. Radioimmunoassay by the Farr technique [13,14] corroborates the above findings but gives a clearer and more quantitative estimate of relative antigenantibody affinities. Fig. 7A demonstrates conclusively that all the soybean leghaemoglobin components a b c1 c2 and d share common determinants inasmuch as they are all able to completely inhibit the binding of 1251-labelledleghaemoglobin a to its homologous anti-(leghaemoglobin a) (ordinate) if they are added in sufficient concentration (abscissa) : that is, all four curves return to the baseline. The relative efficiencies of inhibition already noted with the precipitin tests,

396

Circular Dichroism and Immunology of Leghaemoglobins 120

40 20 -

log Competing antigen (tog ng) log Competing antigen

(log ng)

log Competing antigen (log ng)

Fig. 7. Radioimmunoassuys by the Furr method /13,14/. The curves compare the ability of increasing amounts of unlabelled competing leghaemoglobins to inhibit the binding of 1251-labelledsoybean leghaemoglobins to their homologous antibodies. (A) Antiserum to soybean leghaemoglobin a. (B) Antiserum to soybean leghaemoglobin a : left-hand curve, soybean holo-leghaemoglobin a; right-hand curve, soybean apo-leghaemoglobin a. (C) Antiserum to soybean leghaemoglobin CI.For (A) and (C) the symbols for the different leghaemoglobins are; (0)soybean a ; (0)soybean b; (0)soybean c1 ; (A) soybean CZ;(A)soybean d ; ( x ) serradella 11; (W) snake bean a ; (a)lupin (unfractionated); and for the myoglobins are (+) sperm whale myoglobin; (+) horse heart myoglobin. Error bars are shown only for (3).Standard deviations were never larger than k 5 and usually k 3

however, are now more clearly seen. Again, leghaemoglobin b competes equally well with the homologous antigen leghaemoglobin a suggesting a high degree of homology. It is now clear, however, that leghaemoglobin d ranks next in affinity, followed by leghaemoglobin cz . Finally, leghaemoglobin CI , which appeared to barely cross-react at all in the precipitin test, is now clearly shown to cross-react if used at sufficiently high concentrations. Thus, to achieve say 50 % inhibition of the 1251-labelledleghaemoglobin a/ anti-(leghaemoglobin a) reaction one needs 20 times as much unlabelled leghaemoglobin c1 as unlabelled leghaemoglobin a, suggesting a somewhat greater attenuation of affinity than for leghaemoglobin d and leghaemoglobin c2. In summary, a = b > d > c2 > c1.

It is probably significant that the curves for which attenuation of affinity is apparent become progressively less sigmoidal in shape. Venning has suggested that such curves should be anticipated when the antigenicity of most but not all determinants is weakened; this is in contrast to curves which are laterally displaced but are of unchanged sigmoidal shape in which case all determinants are weakened [19]. Two additional radioimmunoassay curves were determined. In one (not shown in Fig.7), the homologous anti-(leghaemoglobin a) versus 1251-labelled leghaemoglobin a reaction was carried out in the presence of increasing titres of cyanoferric leghaemoglobin a ; in the other, the competing antigen was apo-leghaemoglobin a. The purpose of the first experiment was to determine whether the powerful


cyanide ligand induced any tightening up of the leghaemoglobin conformation, possibly resulting in a changed affinity; there was no evidence of such a change since the curve fell exactly on that for the homologous reaction. The experiment with apoIeghaemoglobin a (Fig. 7 B), however, showed an attenuation of binding for the latter, the curve being displaced to the right of the homologous reaction. We believe that this is consistent with a “loosening up” of the leghaemoglobin conformation on removal of haem. The binding affinity of an antigen to an antibody is a function not only of Kass (the association constant) but also of Kconf the equilibrium constant for the folding/unfolding equilibrium of the antigen [20]. In the present instance, Kconf was presumably shifted towards the unfolded state by removal of haem and we are investigating the use of such data for evaluating the Kcollfof haem proteins. Complete cross-reactivity, that is, 100 % inhibition for all soybean leghaemoglobins was noted also when anti-(leghaemoglobin CI j serum was used (Fig. 7C). The three curves of Fig. 7 C were obtained using antiserum from a single animal; an almost identical set of curves was obtained with a second animal (not shown). Unlike the results using anti-(leghaemoglobin a) serum, leghaemoglobin cz and leghaemoglobin d were as effective as the homologous leghaemoglobin c1 antigen in inhibiting the ‘251-labelledleghaemoglobin c~/anti-(leghaemoglobin c~j reaction. Leghaemoglobin b and leghaemoglobin a were less effective, in that order. In summary, CI = c2 = d > b a. The lack of complete reciprocity between Fig.7A and C may be explained in the terms discussed earlier; the ability of anti-(leghaemoglobin cl) antibodies to accommodate small changes in determinant conformation which are not acceptable to anti-(leghaemoglobin a) antibodies. Turning from soybean leghaemoglobins to leghaemoglobins from other plants, it is noted that serradella leghaemoglobin I1 is able to completely inhibit the homologous soybean leghaemoglobin a/ anti-(leghaemoglobin a) reaction, showing that all determinants are still recognisable and therefore cross-react (Fig. 7A). The attenuation in affinity, however, is now so severe that it requires nearly a 100-fold excess of serradella leghaemoglobin I1 to achieve the same (50 inhibition as with soybean leghaemoglobin a. With snake bean leghaemoglobins cross reactivity is weakened still further and 50% inhibition requires some 3000-fold excess of snake bean leghaemoglobin. Although the curve is incomplete, there is some suggestion that it may not achieve complete inhibition; if so, this suggests not only a considerable decrease in antigenic affinity towards the soybean leghaemoglobin a antibodies but also that not all leghaemoglobin a determinants are now present. This is in accord with the findings by double diffusion

x)

391

(Fig. 5 B) in which partial cross-reactivity was noted. Using Venning’s hypothetical curves [19] one might also suggest that since the serradella and snake bean inhibition curves are fully sigmoidal (cf. the left-hand curve of Fig. 7A), all of the cross-reacting determinants in these more distantly related leghaemoglobins are now attenuated. There is an apparent contradiction between the Farr radioimmunoassay results and the double diffusion results with respect to the relative crossreactivities of serradella and snake bean leghaemoglobins with anti-(soybean leghaemoglobin cl). While the former procedure suggests a much weaker affinity for snake bean leghaemoglobin, the latter is still able to form a visible precipitation line with soybean leghaemoglobin cl in the diffusion procedure; this is in contrast to serradella leghaemoglobin which, in spite of repeated attempts at various concentrations, produces no such line of precipitation. We have no explanation for this anomaly but feel that the radioimmunoassay result is the more informative. Lupin leghaemoglobins fail to show any sign of cross-reactivity with soybean leghaemoglobin a, in agreement with the double diffusion result. In view of this rapid divergence within the plant leghaemoglobins it is hardly surprising that the mammalian myoglobins (sperm whale and horse heart) also fail to show any inhibition (Fig. 7A). Amino Acid Sequences

The complete amino acid sequence of soybean leghaemoglobin a has been determined by Ellfolk and Sievers [3]. A tentative sequence has been assigned to soybean leghaemoglobin CI and a partial sequence to leghaemoglobin c2 (Nicola et al., unpublished). Of the 142 residues in leghaemoglobin a there appear to be about 10 substitutions in leghaemoglobin CI and a comparable number, though at different places, in leghaemoglobin c2. In the latter case there is also evidence for some deletions. In all, there are at least 19 mutation loci among the 3 soybean leghaemoglobin proteins, some involving 2-base changes. Conclusions While the circular dichroism studies stress the overall conformational similarities among the plant leghaemoglobins, the immunochemical studies stress rather their surface amino acid sequence differences and divergences (cf. [21]). Amino acid sequence studies, as far as they have gone, rationalise both sets of properties in terms of the extent and location of residue substitutions. Thus, the a-helical content of the leghaemoglobins is very similar (see also [22]j and approaches that of myoglobin. From the alignments of the soybean [3],

398

kidney bean [4] and broad bean [5] leghaemoglobins with other globins [23] one would anticipate that the D-helix of myoglobin is deleted and that a somewhat hower a-helical content (say 5 %) is present. Certainly, preliminary X-ray data suggest a very similar overall folding for lupin leghaemoglobin and the mammalian myoglobins [I]. The Soret and visible-region circular dichroism spectra are very sensitive indicators of haem orientation and environment [7,15] and the remarkable similarity in sign, magnitude and shape of the leghaemoglobin spectra in these regions suggests great similarities in the haem pocket stereo-chemistry among the leghaemoglobins. This stereochemistry differs significantly from that of the myoglobins judged by the opposite sense of the spectra. A notable feature of all leghaemoglobin near ultraviolet CD spectra is the evidence for a weaker haem-protein interaction than in myoglobins. A systematic study of immunochemical crossreactivities among a group of closely related proteins from different species should, if there is sufficient structural and sequence hemology, provide information about their phylogeny. This has proved true, for example, for cytochrome c [24] and the bird lysozymes [25] where phylogeny based on cross-reactivity studies agrees with that based on known sequences. Even where sequences are not known or only partially known, as in the lactate dehydrogenases [26], reasonable schemes may be postulated for the evolution of enzymes from a common ancestral gene. The comparatively large number of amino acid substitutions noted even within the leghaemoglobins from a single plant cultivar (some 19 residues differ out of 142 between soybean leghaemoglobin a, leghaemoglobin CI and leghaemoglobin cz - Nicola et al., unpublished) are mostly chemically and conformationally conservative. Using the alignments of Dayhoff et al. [23] and the data of Atassi [27] on the the identity of the antigenic determinants of myoglobin, one would conclude that most of the substitutions occur at the surface of the molecule even when they form parts of a-helices and are at ends of antigenic loops. These substitutions are insufficient to prevent complete cross-reactivity among the leghaemoglobins of a single plant (soybean) biit do affect their relative affinities for the same soybean leghaemoglobin specific antibody population. To this extent the substitutions are not immunochemically conservative. Cross-reactivities are further attenuated when one compares leghaemoglobins from different plants until, with lupin leghaemoglobin versus anti-(soybean leghaemoglobin a) there is no reaction even with the sensitive Farr radioimmunoassay. The precipitin reaction, depending as it does on multivalent crossreactivity and high association constants [17] does not provide adequate information and indeed can be misleading in a study of this type.


It is likely that the rapid falling off in crossreactivity is a reflection of the extent of amino acid substitution between plant species, though one must always bear in mind the large disparities which have sometimes been observed between single substitutions in a “key” antigenic position [28].One may also expect non-reciprocal cross-reactivities when using several antisera (as in Fig. 7 A and C); the antibody directed towards a determinant containing a bulky amino acid may be able to accommodate a smaller amino acid at the antibody combining site with little or no change in affinity, while the reverse substitution might decrease or destroy affinity [29]. With these reservations in mind, one might usefully attempt to construct phylogenetic charts of the Leguminosae or check on existing ones [30] with a view to tracing the early history of the globin family. A preliminary account of this study was presented at the 19th Annual Meeting of the Australian Biochemical Society in May 1975 (Proc. Aust. Bioch. Soc. 8, 1, 1975). The authors wish to acknowledge financial support from the Australian Research Grants Committee (S. J. L.), postgraduate research awards from the Commonwealth Government (J. G. R. H.) and the Commonwealth Scientific and Industrial Research Organization (N. A. N.). The Commonwealth Serum Laboratories are also thanked for their support (J. G . R.H. and S. J. L.) and Mr P. Lewis and Mr E. Liefman for their interest and cooperation in immunising animals. Soybean leghaemoglobin supplies from Dr C. A. Appleby of the Division of Plant Industry, C. S. I. R. O., Canberra are gratefully acknowledged.

REFERENCES 1. Vainshtein, B. K., Harutyunyan, E. H., Kuranova, I. P., Borisov, V. V., Sosfenov, N. I., Pavlovsky, A . G., Grebenko, A. I. & Konareva, N. V. (1975) Nature (Lond.) 254, 163164. 2. Goodman, M., Moore, G. W. & Matsuda, G . (1975) Nature (Lond.) 253,603 - 608. 3. Ellfolk, N. & Severs, G. (1971) Acta Chem. Scand. 25, 3532 - 3534. 4. Lehtovaara, P. & Ellfolk, N. (1974) FEBS Lett. 43, 239-240 5. Richardson, M., Dilworth, M. J . & Scawen, M. D. (1975) FEBS Lett. 51, 33-37. 6 . Nicola, N. A., Appleby, C. A., Leach, S. J. & Minasian, E. (1974) Proc. Aust. Biochem. Soc. 7 , 21. 7. Nicola, N. A,, Minasian, E., Appleby, C. A. & Leach, S. J. (1975) Biochemistry, 14, 5141 -5149. 8. Dickerson, R. E. (1971) J. Mol. Evolution, I , 26-45. 9. Appleby, C. A . , Nicola, N. A., Hurrell, J. G. R. & Leach, S. J. (1975) Biochemistry, 14, 4444-4450. 10. Broughton, W. J . & Dilworth, M. J. (1971) Biochem. .I. 125, 1075- 1080. 11. Broughton, W. J., Dilworth, M. J. & Godfrey, C. A. (1972) Biochem. J . 127,309-314. 12. Greenwood, F. C., Hunter, W. M. & Glover, J. S. (1963) Biochem. J . 89, 114- 123. 13. Farr, R . S. (1958) J . Infect. Dis. 103, 239-262. 14. Minden, P. & Farr, R . S. (2973) in Hundhook of E.wpcrirnenta1 Immunology (Weir, D. M., ed.) 2nd edn. vol. 1. p. 15.4, Blackwell Scientific Publications, London. 15. Hsu, M.-C. & Woody, R. W. (1971) J . Am. Chem. Soc. 93, 3515-3525.

J. G. R. Hurrell, N. A. Nicola, W. J . Broughton, M. J. Dilworth, E. Minasian, and S. J. Leach 16. Crumpton, M. J. & Wilkinson, J. M. (1965) Biochem. J . 94, 545 - 556. 17. Reichlin, M. (1974) Immunochernistry, 11, 21 -27. 18. Prager, E. M., Fainaru, M., Wilson, A. C. & Arnon, R. (1974) Immunochemistry, 11, 153- 156. 19. Venning, M. M. (1975) Immunochemistry, 12, 365-372. 20. Sachs, D. H., Schechter, A., Eastlake, A. & Anfinsen, C. B. (1972) Proc. Nail. Acad. Sci. U.S.A. 69, 3790-3794. 21. Margoliash, E. & Schejter, A. (1966) Adv. Protein Chem. 21, 113-286. 22. Ellfolk, N. & Sievers, G. (1975) Biochim. 5ioph.y~.Actu, 405, 213-227. 23. Dayhoff, M. O., Hunt, L. T., McLaughlin, P. J. & Jones, D. D . (1972) in Atlas of Protein Sequence and Structure

24. 25. 26. 27. 28. 29. 30.

J. G. R. Hurrell, N. A. Nicola, E. Minasian, and S. J. Leach*, Russell Grimwade School of Biochemistry, University of Melbourne Parkville, Victoria, Australia 3052 W. J. Broughton School of Biological Sciences, University of Malaya, Kuala Lumpur, Malaya M. J. Dilworth School of Environment and Life Sciences, Murdoch University, Murdoch, Western Australia

* To whom correspondence should be addressed

399

(Dayhoff, M. O., ed.) vol. 5, p. D369, National Biomedical Res. Found., Washington. Margoliash, E., Nisonoff, A. & Reichlin, M. (1970) J . Biol. Chem. 245,931 -939. Prager, E. M. & Wilson, A. C. (1971) J . Biol. Chem. 246, 5978 - 5989. Holmes, R. S. & Scopes, R. K. (1974) Eur. J . Biochem. 43, 167- 177. Atassi, M. Z. (1975) Immunochemistry, 12, 423-438. Nisonoff, A,, Reichlin, M. & Margoliash, E. (1970) J . Biol. Chem. 245,940-946. Margoliash, E., Reichlin, M. & Nisonoff, A. (1967) in Conformation of Biopolymers (Ramachandran, G. N., ed.) vol. 1, pp. 253 -277, Academic Press. Senn, H. A. (1938) Bihliogr. Genet. 12, 175-337.