Inhibition of sickle cell hemoglobin gelation by some aromatic compounds

Inhibition of sickle cell hemoglobin gelation by some aromatic compounds


332KB Sizes 0 Downloads 1 Views

Vol. 77, No. 4, 1977


INHIBITION OF SICKLE CELL HEMOGLOBIN GELATION BY SOME AROMATIC COMPOUNDS Philip D. Ross and S. Subramanian Laboratory of Molecular Biology, NIAMDD National Institutes of Health Bethesda, Maryland 20014 Received



SUMMARY: Some simple organic molecules have been studied for their effects on the solubility of deoxyhemoglobin S. Compounds having an aromatic group with a pendant aliphatic polar moiety are the most effective in increasing the solubility. The results are rationalized in terms of a simple mechanism of non-covalent inhibition of gelation. Since polymerization of hemoglobin S (lib S) into ordered bundles of fibers (gelation) is responsible for sickling in erythrocytes, the prevention of this process should be accomplished by the disruption of forces in the localized areas of contact between hemoglobin molecules.

Such disruption would increase

the solubility of deoxyhemoglobin S and even small changes in solubility will greatly retard the kinetics of polymer formation (i). Our approach has been to try to supplant the protein-protein interaction with a protein-small molecule interaction having the same precise complementarity.

In this co~unica-

tion we report on a class of compounds which increase the solubility of deoxyhemoglobin S and propose a simple mechanism of their inhibitory action which is also capable of accounting for most observations on the non-covalent inhibition of hemoglobin S gelation.

MATERIALS AND METHODS Hemoglobin S was purified by chromatography as previously described (i). All solutions contained 0.05 M sodium dithionite and were at pH 7.15. Solubilities were determined at 25°C by spectrophotometric determination of hemoglobin concentration after ultracentrifugation as described elsewhere (2). The kinetics of turbidity change accompanying gelation were studied on the same samples (2) after a rapid temperature jump from 0 to 20°C in 0.i M bistris buffer. The turbidity remained constant during the delay period, td, then increased sigmoidally until 2 td, as found previously (1,2). All chemicals except phenylalanine (Sigma) were obtained from Aldrich Chemical Co. in the best grades available.

Copyright © 1977 by Academic Press, Inc. All rights o/reproduction in any ]orm reserved.

1217 ISSN 0006-291X

Vol. 77, No. 4, 1977



A 1.2









0.16 0.0






llO 0.0

0 02






Fig. IA - Solubility of deoxyhemoglobin S in the presence of various alcohols. Potassium phosphate buffer, 0.15 M, pH = 7.15, 25°: O , t-butanol;

[ ] , i-propanol; V , ethanol; A 0 ' benzyl alcohol.

, n - p r o p a n o l ; O , n-butanol;

IB - Ratio of the deoxyhemoglobin S solubility in the presence and absence of some aromatic inhibitors. Bis-tris buffer, 0.i M, pH - 7.15, 25°: O , ~ C H 2 O H ; O , ~CH2CH2OH; [ ] , ~ C H 2 C O O N a ; m , +CHmeHmeOONa; O ' L-phenylalanine; A , ~CH2NH3CI; A , ~ C H 2 C H 2 N H B C I ; V , m-I~CH2NHBCI.

RESULTS AND DISCUSSION The effect of various alcohols in increasing the solubility of deoxyhemoglobin S is shown in Fig. IA.

Increasing the length of the aliphatic

chain causes a proportional increase in the solubility i.e., n-butanol> n-propanol>ethanol.

Increasing the degree of branching, however, results only

in a minimal effect upon the solubility increment i.e., EtOH~-i-PrOH~t-BuOH. Benzyl alcohol, in contrast, produces a marked increase in the solubility pointing to the greater effectiveness of the aromatic ring compared to the aliphatic group, i.e., phenyl>>methyl. The effect upon the solubility of deoxyhemoglobin S by arylalkyl compounds with different functional groups and of varying side chain length is shown in Fig. 1B.

These materials were more extensively examined at lower concentra-

tions where the kinetics of gelation could be studied in the absence of generalized medium effects.

These compounds containing a phenyl residue and


Vol. 77, No. 4, 1977


a pendant side chain terminating in a functional group with hydrogen bonding capacity i.e., NH~, COO-, OH, all produced, approximately

at comparable concentrations,

the same increase in the solubility of deoxyhemoglobin



effect of side chain length could not be discerned at the low solute concentrations


The delay time, td, of the kinetics of gelation was proportional

to a high

power of the hemoglobin solubility and followed the empirical supersaturation law; td-l~ Sn (1,2,3) where n is a constant and S is the ratio of the initial hemoglobin concentration experiment.

to the solubility at the temperature of the kinetic

For the control and for the eight substances

in Fig. IB at con-

centrations between 0.01-0.04 M, the value of the exponent, n, was found to be 34.2 ± 2.5.

This value agrees well with those found previously in phosphate

buffer at similar hemoglobin concentrations

for the effect of; urea, n = 36;

CO, n = 33; pH, n = 40 (2,3) and two amino acids, n = 32 (4). results indicate that the basic mechanism of polymerization

These kinetic

(1,3) is not

greatly affected in the presence of these inhibitors. The outstanding are as follows:

features that evolve from the results of Fig. IA and IB

(i) increasing

solubility correspondingly than aliphatic groups.

the length of the nonpolar group increases the

and (ii) aromatic residues have a greater effect

The importance of the role of the aromatic residue is

also borne out by the work of Noguchi & Schechter

(4) who found that of seven-

teen amino acids tested only the aromatic amino acids produced cant increase in the deoxyhemoglobin examined several oligopeptides

S solubility.

(5) any signifi-

Recently, Votano et al.


and found that the most effective inhibitors of

gelation were the ones which had both a hydrophobic residue such as phenylalanine and a hydrophilic

residue such as arginine or lysine.

known to increase the solubility of deoxyhemoglobin and hydrophilic residues,

e.g. alkylureas


S, also have hydrophobic

The combination of an aromatic

residue with a functional polar group alone is not sufficient increase in solubility,


for Noguchi and Schechter


Other compounds

to cause an

(5) have found

Vol. 77, No. 4, 1977


that phenol, aniline and salicylic acid do not increase the deoxyhemoglobin S solubility and we have found that 5-iodosalicylate decreases the solubility. It thus appears that the hydrophilic group must be located on an aliphatic side chain attached to the aromatic ring which provides the proper flexibility or distance to effectively interact with deoxyhemoglobin S. In order to account for these findings in molecular terms it is first necessary to examine the interactions at the contact region of the deoxyhemoglobin S polymer.

The 3A X-ray diffraction study (8) of deoxy-Hb S crystals

indicate there is essentially only one kind of side-to-side contact between two strands of hemoglobin molecules, and this contact involves AspE17(73), PheFl(85) and LeuF4(88~ on one B chain of strand 1 with ThrAl(4) and ValA3(6) of another B chain on strand 2.

The Asp residue is capable of H-bonding with

Thr while the Phe and Leu residues appear to be involved in a hydrophobie interaction with Val.

Double strands involving these contacts have been

incorporated into a model of the fiber (8) that is consistent with the results from electron microscopic studies of Hb S fibers (9). Studies with mutant hemoglobins (i0) indicate that the B73 Asp residue is important in promoting gel formation in Hb S, initiated probably by the substitution of B6 Val for B6 Glu.

An examination of the deoxy Hb A quaternary

structure reveals that Thr 84, Phe 85 and Leu 88 in each B chain are clustered together and exposed to the solvent.

When the B6 Val of another strand B-chain

interacts (during gelation) with this region there would be a mutual stabilization of the hydrophobic groups in the contact region. Using the crystal contacts (8), reinforced by the mutant hemoglobin results (I0), a schematic diagram of the lateral contact region between two B chains in the deoxy Hb S polymer is presented in Fig. 2A.

Sites I and III are in hydro-

phobic contact and sites II and IV are hydrogen bonded.

A typical ligand such

as benzyl alcohol could mimic the B4-B6 segment of the 2~ 2 strand and bind to the i~ 1 chain as shown in Fig. 2B.

The details of the interaction may not be

unique but in general an aromatic ring or other highly polarizable aliphatic


Vol. 77, No. 4, 1 9 7 7


..~..i ~ ~


....,,,!!iiiiiiiiiiiiiiiiiiiiiiii ,




2//, 1#,

lf3, (lf3, -- 2~,)

(lp,--2pz) + (I)CH20H ---..~(1/3,-- eCH20H ) -t-2/~2

Fig. 2A - A schematic description of the contact region in the deoxyhemoglobin S polymer, iS 1 and 282 are two ~ chains from Hb S tetramers in adjacent strands (8). Sites I and II are in I~ 1 chain and III and IV in 282 chain. For the sake of clarity, only segments of the 8 chains are shown. The critical contacts for gel formation are indicated as I ÷ ÷ III (hydrophobic) and II + ÷ IV (hydrogen bond). 28 - Benzyl alcohol (~CH20H) is shown here as binding to the 181 chain at sites I and II thereby separating the 282 chain; this would interfere with polymer formation and thereby increase the solubility of deoxyhemoglobin S.

nonpolar group linked to a H-donor group would interact with sites I and II in the i~ 1 chain provided the distance requirement met.

for optimal interaction is

This description will also apply to the amines.

The carboxylates are

incapable of intereacting with site II since they are not H-donors.


they can interact with site I and some other H-donor site on the 181 chain or they could interact with sites III and IV on the 282 chain. This model can also explain the inhibitory action of phenylalanine alkyl ureas (7), lysolecithin,



(ii) and other similar compounds.


alkyl or aryl group can interact with the hydrophobic patch on the i~ 1 strand shown in Fig. 2A while the hydroxy or amino group can H-bond to the 873 Asp. The chain length effect observed in the alcohols as well as the ureas

(7) is

an indication of a critical distance requirement between the polar head group



Vol. 77, No. 4, 1977


and the end of the nonpolar tail.

In general, an effective inhibitor should

contain both hydrophobic and hydrophilic groups (at specified distances) that match the original comp!ementarity of the contact ' regions in the Hb S polymer. The stoichiometric interaction between inhibitor and hemoglobin presented above would require an inhibitor concentration comparable to that of hemoglobin in the case of tight binding.

If rate factors are decisive in controlling

sickling (1,3) somewhat lower inhibitor concentrations may prove effective. Lower inhibitor concentrations would be desirable from a toxicological viewpoint. The model proposed above suggests that a more efficient inhibitor might be designed by enhancing the polarizability of the aromatic ring by proper substitution with heavy halogen, alkyl or aryl groups.

The model also makes

some interesting predictions and suggests additional experiments that might be carried out to test the postulated mode of inhibitor action.

It follows from

the model that Hb A should bind all of the aromatic inhibitors (alcohols and amines) and the alkylureas; it may or may not bind the aromatic acids depending upon the locus of their binding site. by optical and NMR spectroscopy.

Such interactions could be detected

Direct X-ray studies of crystals grown with

some of these inhibitors should provide detailed visualization of the mode of binding and thereby test the validity of the model. While the effects of the inhibitors used in this study in increasing the solubility of deoxy Hb S are not sufficiently great to be of immediate therapeutic significance, the model presented above to account for their behavior should provide a rational chemical and molecular basis for finding more efficient non-covalent inhibitors.


The authors wish to thank L. Leive and A. N. Schechter for

suggestions concerning the preparation of this manuscript and C. T. Noguchi and A. N. Schechter for their kind permission to quote their unpublished work.


Vol. 77, No. 4, 1977


REFERENCES: i. 2. 3.

4. 5. 6. 7. 8.


i0. ii.

Hofrichter, J., Ross, P. D. and Eaton, W . A . (1974) Proc. Nat. Acad. Sci. U.S.A. 71, 4864-4868. Hofrichter, J., Ross, P. D. and Eaton, W . A . (1976) Proc. Nat. Acad. Sci. U.S.A. 73, 3035-3039. Hofrichter, J., Ross, P. D. and Eaton, W . A . (1976) in Proceedings of the Symposium onMolecular & Cellular Aspects of Sickle Cell Disease, Hercules, J. I., Cottam, G. L., Waterman, M. R. and Schechter, A. N., eds. (DHEW) Pub. No. (NIH) 76-1007) pp. 185-222. Noguchi, C. T. and Schechter, A. N. (1977) Biochem. Biophys. Res. Comm. 74, 637-642. Noguchi, C. T. and Schechter, A. N., private communication. Votano, J. R., Gorecki, M. and Rich, A. (1977) Science 196, 1216-1219. Elbaum, D., Nagel, R. L., Bookchin, R. M. and Herskovits, T. T. (1974) Proc. Nat. Acad. Sci. U.S.A. 4718-4722. Wishner, B. C., Hanson, J. C., Ringle, W. M. and Love, W. E. (1976) in proceedings of the Symposium on Molecular & Cellular Aspects of Sickle Cell Disease, Hercules, J. I., Cottam, G. L., Waterman, M. R. and Schechter, A. N., eds. (DHEW Pub. No. (NIH) 76-1007) pp. 1-29. Edelstein, S. J., Josephs, R., Jarosch, R. H., Telford, J. N. and Dykes, G. (1976) in Proceedings of the Symposium on Molecular & Cellular Aspects of Sickle Cell Disease, Hercules, J. I., Cottam, G. L., Waterman, M. R. and Scheehter, A. N., eds. (DHEW Pub. No. (NIH) 76-1007) pp. 33-59. Bookchin, R. M. and Nagel, R . L . (1974) Seminars in Hematology ii, 577-595. Freedman, M. L., Weissmann, G., Gorman, B. D. and Cunningham-Rundles, W. (1973) Biochem. Pharmacol. 22, 667-674.