Differential Hydration of Protein

Analogs in D20 and H20.

By Olga Likhodi

A thesis submitted in conformity with the requirements

for the degree of Master of Science

Graduate Department of Pharmaceutical Sciences

University of Toronto

National Library I*I ofCanada Bibliothèque nationale du Canada

Acquisitions and Acquisitions et Bibliographie Services services bibliographiques

395 Wellington Street 395, rue Wellington OttawaON K1AON4 Ottawa ON K1A ON4 Canada Canada

Your rsrs Vonar6fdnnncs

Our & Notre reMnmce

The author has granted a non- L'auteur a accordé une licence non exclusive licence allowing the exclusive permettant à la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or sel1 reproduire, prêter, distribuer ou copies of this thesis in microform, vendre des copies de cette thèse sous paper or electronic formats. la forme de rnicrofiche/~, de

reproduction sur papier ou sur format électronique.

The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts f?om it Ni la thèse ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation.

@Copyright by Olga Likhodi 2000

Abstract

Differential Hydration Of Protein Analogs In D20 and H20.

MSc., August 2000 Olga Likhodi Department of Pharmaceutical Sciences University of Toronto

The partial molar volumes, p, and adiabatic cornpressibilities, Ks1 of five

a-amino acids and five oligoglycines have been determined in D20 at 25'. In

addition, the partial molar volumes, VO, expansibilities, EO. and adiabatic

compressibilities, KQs, for a homologous series of eight a,o-aminocarboxylic

acids in D20 solution have been detenined within the temperature range 18 to

55 OC. The resulting data have been used to estimate the volume, expansibility,

and compressibility contributions of the component aliphatic, polar and charged

groups. We used these data to characterize quantitatively the differential

hydration properties of charged and hydrophobie groups in DzO and HzO. Our

results suggest that the hydration properties of component groups in D20 are

measurably distinct frorn those in H20 in their absolute values and in their

temperature dependences. These results represent a step in building up an

ernpirical database of differential volumetric parameten of protein functional

groups in D20 and HzO. Such a database is required for developing a

methodology in which insights into the number and chemical nature of solvent-

exposed protein groups can be gained from difFerential volumetric measurements

in the absence of structural information.

Acknowledgements

I would like ta take this opportunity to thank my supervisor, Prof. Tigran

Chalikian, for his support and guidance during the course of this project. I would

also like to thank my cornmittee members, Prof. Robert Macgregor Jr. and Prof.

Peter Pennefather for their insights and opinions provided. Also, the help offered

by my lab colleagues, has been immeasurable. Finally, the generous funding

from the Ontario Graduate Scholarships in Science and Technology (OGSST) is

gratefully acknowledged.

Publications from this Thesis

1. Partial molar volumes and adiabatic compressibilities of a series of aliphatic

amino acids and oligoglycines in D20.

Likhodi O., Chalikian T. V.. J. Am. Chem. Soc.. 121, 11 56. 1999.

2. Differential hydration of a,o- aminocarboxylic acids in D20 and H20.

Likhodi O., Chalikian T.V.. J. Am. Chem. Soc., 122. 7860, 2000.

Table of contents

To pic

Abstract

Acknowledgments

Publications frorn this thesis

Table of Contents

List of tables

List of figures

Chapter 1 : Introduction

1.1 Role of water in protein folding and recognition.

1.2 Model compounds

1.3 Heavy water, its properties and research trends.

1.4 Thesis outline

1.5 References

Chapter 2: Partial molar volumes and adiabatic

Compressibilities of a series of aliphatic aminoacids

and oligoglycines in D20

Chapter 3: Differential Hydration of

a,o-Aminocarboxylic Acids in D20 and H20

Chapter 4: Conclusion

4.1 Summary of findings

4.2 Conclusions

Appendix

Page

i i

iii

iv

v

vi

vi i

List of tables

Chapter 2

Table 1. The molecular weights, M, relative molar sound velocity increments, [U],

partial molar volumes, Vol and partial molar adiabatic cornpressibilities, KSI of

the arninoacids in D20 at 25 C.

Table 2. The molecular weights, M, relative molar sound velocity increments, [U],

partial molar volumes, Vol and partial molar adiabatic compressibilities, Kos, of

the oligoglycines in D20 at 25 C.

Chapter 3

fable 1. Molecular weights, M. and relative molar increments of sound velocity,

[U], as a function of temperature for the a,w-aminocarboxylic acids in D20.

Table 2. Apparent molar volumes, 4V, as a function of temperature for the a,o-

aminocarboxylic acids in D20 and H20a.

Table 3. Apparent molar adiabatic compressibilities, @Cs, as a function of

temperature for the a,o-aminocarboxylic acids in D20 and H20a.

Table 4. Partial molar expansibilities, EO, as a function of temperature for the

a,o-aminocarboxylic acids in D20 end H20a.

Table 5

The interaction volume, VI. for the independently hydrated amino and carboxyl

termini as a function of temperature for the a,o-aminocarboxylic acids in D20 and

H20a.

List of Figures

Chapter 2

Figure 1. The thickness, A, of the thermal volume as a function of the diameter

of a solute. dZ, in D20 and H20.

Figure 2. (a) The dependence of the partial molar volume, Vc, of the cavity

enclosing a spherical solute on its diameter, dz, in D20 (A) and H20 (V); (b) The

difference, AVc, in the partial molar volume, Vc, of the cavity enclosing a

spherical solute in D20 and H20 as a function of Vc.

Figure 3. The difference between the partial molar volume, Va. and the van der

Waals volume, VW, for the amino acids as a function of the number of carbon

atoms in the side chain.

Figure 4. The difference between the partial molar volume, VO, and the van der

Waals volume. VW. for the oligoglycines as a function of the nurnber of peptide

bonds,

Figure 5. The difference, No, between the partial molar volumes, V", of the

oligoglycines in D20 and H20 as a function of the number of peptide bonds. The

partial molar volumes, Vo, of the oligoglycines in H20 are from ref. [28].

Figure 6. The partial molar adiabatic cornpressibility, Kos, of the amino acids in

020 (0) and Hz0 (@) as a function of the number of carbon atoms in the side

chain. The partial molar adiabatic cornpressibilities, Kas, of the amino acids in

H20 are from ref. [25].

Figure 7. The partial molar adiabatic compressibility, Kos, of the oligoglycines in

D20 (0) and H20 (@) as a function of the number of peptide bands. The partial

vii

rnolar adiabatic compressibilities, KS, of the oligoglycines in H20 are from ref.

WI *

Figure 8. The difference, AKOs, between the partial rnolar volumes, KS, of the

oligoglycines in D20 and H20 as a function of the number of peptide bonds. The

partial rnolar adiabatic compressibilities, Ks, of the oligoglycines in H20 are from

ref. [28].

Chapter 3

Figure 1. The difference between the partial molar volume of the a,o-

aminocarboxylic acids, VO, in D20 and their van der Waals volume, Vw, as a

function of the van der Waals surface area, Sw. at 18 OC (e), 25 OC (O), 40 OC

(i), and 55 O C (O).

Figure 2. The temperature dependences of the contributions of the

independently hydrated -CH2- group to the partial molar volume of the a,w-

aminocarboxylic acids in H20 (@) (from ref. 11) and D20 (0).

Figure 3. The dependence of the interaction volume, VI, on the number of

rnethylene groups in the a,o-aminocarboxylic acids in D20 at 18 OC (a), 25 OC

(O), 40 OC (i), and 55 "C (O).

Figure 4. The difference between the interaction volumes, VI, of the

independently hydrated amino and carboxyl temini of the a,o-aminocarboxylic

acids in D20 and H20 as a function of temperature.

Figure 5. The dependence of the partial molar expansibility of the a,o-

arninocarboxylic acids in D20 on the number of methylene groups at 18 OC (*),

25 OC (O), 40 OC (i), and 55 OC (a).

viii

Figure 6. The temperature dependences of the contributions of the

independently hydrated -CH2- group to the partial molar expansibility of the a,o-

aminocarboxylic acids in H20 (a) (from ref. I l ) and D20 (0).

Figure 7. The temperature dependences of the contributions of the

independently hydrated amino and carboxyl termini to the partial molar

expansibility of the a,o-aminocarboxylic acids in H20 ( O ) (from ref. 1 1 ) and 020

(0)-

Figure 8. The dependence of the partial molar adiabatic compressibility of the

a,w-aminocarboxylic acids in DzO on the number of methylene groups at 18 OC

(a). 25 O C (O), 40 "C (i). and 55 OC (O).

Figure 9. (a) The temperature dependences of the contributions of the

independently hydrated -CH2- group to the partial molar adiabatic compressibility

of the a.o-aminocarboxylic acids in H20 (e) (from ref. 1 1 ) and D20 (O); (b) The

difference between the contributions of the -CH2- group to the partial molar

adiabatic compressibility of the a,o-aminocarboxylic acids in D20 and H20 as a

function of temperature.

Figure 10. (a) The temperature dependences of the contributions of the

independently hydrated amino and carboxyl termini to the partial molar adiabatic

compressibility of the a,o-aminocarboxylic acids in H20 (e) (from ref. 1 1) and

D20 (O); (b) The difference between the contributions of the independently

hydrated amino and carboxyl termini to the partial molar adiabatic compressibility

of the a,w-aminocarboxylc acids in 90 and H20 as a function of temperature.

Chapter 1. General introduction.

1 .l Role of water in protein folding and binding.

Phamaceutical significance of hydraüon studies

Rational drug design consists of modifying a molecule in order to improve

its biological activity in a predictable manner and is based on the analysis and

engineering of a single reactant pair'*2. A successful rational design strategy

implies that it is possible to predict how a biologically active molecule can be

obtained by chemical synthesisl, and to identify the optimal structure of a ligand

for interaction with target protein2.

Water is known to play an important rote in recognition and stabilization of

the interaction between a ligand and its target macromolecule. The presence of

water molecules. site-specifically bound to polar groups on the surfaces of

macr~molecules and the interfaces of macromolecular complexes3, has been

detected by the techniques of X-ray, neutron diffraction and NMR. However, the

contributions of the displacement and bunal of these water molecules to the

thermodynamics of rnacromolecular assembly processes have remained

obscure.

Biomolecular assernbly processes occumng in aqueous media are

influenced by the differential interactions of water molecules with unassemb led

and assembled components of the system. The water molecules at binding sites

bridge the protein and ligand, and interact with other water molecules to fom a

complex network of interconnecting hydrogen bonds4.

Water can confer a high level of adaptability to interacting surfaces3.

allowing undiscriminating binding, because it imposes few steric constrains on

bond formation and can take part in multiple hydrogen bonds, at the same time

being as structurally important as the other protein-site atoms? Also, it can

provide exquisite specificity and increased affinity to an interaction3. The role of

water is enigmatic in that it can confer a high level of adaptability to a surface,

allowing a biomolecule to be undiscriminating in binding and yet it is capable of

providing exquisite specificity and increased affmity to an interaction2. Water is

highly versatile interactive species at the interfaces of biomolecular complexes

and can act as both a hydrogen bond donor and acceptor. Water occupies less

space than polar side chains of a protein but can participate in multiple hydrogen

bonds. In addition, the diminutive size of water molecule puts few steric

constrains on its ability to adopt a suitable bonding orientation4.

Structure - based drug design strategies largely ignore the effects of

water, because the structural and themodynamic effects of water's inclusion in

binding interfaces are hard to determine or model15. This is unfortunate since

water can be used to improve the efficacy of a drug by engineering it into the

binding interface5. The conventional assumption is that, since the entropic cost

of trapping highly mobile water molecules is large, an interface that leaves no

space between the interacting protein and the ligand will give a higher binding

afin@ than one in which the interface contains gaps filled by water. But in some

cases6, the enthalpy gain that results from making extra water-rnediated

hydrogen bonds is greater than the entropic penalty that must be paid for

immobilizing the water involved. The binding sites are known where the

presence of water contributes to the overall stability of the cornplexS. This

improvement in binding can be used by applying the structural and

themodynamic principles deduced from ihese observations of how to induce

water in structure-based ligand design strategies.

An accurate description of the effects of water on protein-ligand

interactions requires cornbining high resolution structural detail with accurate

thermodynamic data. The balance between the enthalpic and entropic

contributions is a fine one: for a water molecule to contribute to increasing the

binding afinity, it has to be in a binding site which provides the maximum number

of hydrogen bonding partners at the right proximity and orientation5. Before any

conclusions can be drawn on the effects of the inclusion of water molecules in a

protein-ligand interface, the positions and orientations of the bound waters have

to be ascettained. Inclusion of a water molecule into an interface c m provide a

significant contribution to the free energy of an interaction. If the presence of a

water molecule can result in enhanced binding, an awareness of the uses of

water molecule could assist in ligand design. The conditions have to be carefully

chosen for a water rnolecule to improve the binding characteristics of an

interaction5. Water-mediated ligand interactions are essential to biological

processes, yet the structural chemistry influencing whether bound water is

displaced in ligand binding is not well characterized7. Which water molecules are

likely to be displaced dun'ng ligand binding, which of them have to be considered

while designing a ligand, and do these water molecules have any structural

significance? This questions have important implications for drug d e ~ i g n . ~

Given the structure of a biological receptor, it should be possible to design

or diswver molecules that will bind to it. Putative ligands are chosen for their

complementarity to the structure of the recepto?. Failure to consider ligand

solvation during screening for complernentarity to receptors of known structure

offen leads to ligands that are too highly charged or too largea. lncluding ligand

hydration in molecular docking dramatically changes the relative ranking of

compounds in database screens.

In short, a greater understanding of the part that water plays in

biomolecular interactions could be exploited in the design of new ligands. The

inclusion of water into the binding site can potentially improve the affinity of an

interaction and helps to rank compounds in database screensM.

Properties of water of hydration

One significant issue in thermodynamics of protein recognition is to

identify and characterize the relative importance of various interactions stabilizing

proteins and their complexesg- 'O. Among these interactions, hydration is of

particular importance since al1 protein recognition events occur in aqueous milieu

and, consequently, solute - solvent interactions represent a major determinant of

the confornational preferences of polypeptide chains". In general, the

thermodynamic properties of water of protein hydration are distinct from those of

water in the bulk state. Therefore, redistribution between the bulk water and

water of hyd ration associated with protein recognition (e.g . , folding and binding )

should have impact on the overall energetics of the process.

Microscopie structural studies on proteins provide information on the

solvent accessible functional groups and the amount of water molecules

iocaiized on the proiein"? i3y wmplementing these studies by paraiiei

macroscopic thermodynamic studies, it is possible to describe quantitatively the

differential properties of bulk water and water of protein hydratiod3.

Thermodynamic data on proteins are generally pmvided by calorirnetric and

11-14 volumetric methods . Application of calorimetry is based on the use of two

conjugate and extensive variables, C, and T, on the basis of which enthalpy,

entropy. and free Gibbs energy are c a l c ~ l a t e d ~ * ' ~ ~ ~ ~ . The voiurnetric approach is

based on evaluation of the partial molar volumes, expansibilities and adiabatic

compressibilities of protein systems. For comprehensive thermodynamic

characterization of protein hydration, complementary employment of both

10.13 approaches is essential .

Volumetric derivation of thermodynamic data

In general, the hydration of a protein or DNA is predominantly determined

by the accessibility of the polymer chain to the solvent and the nature of the

exposed atomic groups. Therrnodynamic characterizations of the extent and

magnitude of hydration forces between functional grou ps of biolog ical molecules

and water are required for a better understanding of the role of solute-solvent

interactions in protein folding and binding. In recognition this fact, volumetric and

calorimetric techniques have been widely used to investigate the hydration

properties of biopolymers in so~ution'~.

Volumetric obsewables, such as partial molar volume and adiabatic

cornpressibility, are known to be sensitive to solute hydrati~n"*'~. They are highly

nonselective, enabling one to sample the whole population of water of hydration.

In addition, volumetric characteristics can be used for discriminating between

waters solvating charged, polar and nonpolar atomic gr~ups. '~ The hydration

contribution of various protein groups to the partial molar volume and adiabatic

compressibility can be estimated by performing correlation analysis of the

volumetric properties of proteins or their low molecular weight analogs (such as

aminoacids or short peptides) with the accessible surface areas of specific

classes of solvent exposed gr~ups '~ . Compressibility and volume contributions

of a specific group can be expressed as the contributions of to Va and K~',

respect ive1 y.

The partial molar volume, p, of a solute in solution can be defined as the

apparent volume occupied by 1 mole of a solute at infinite dilution. A traditional

approach for qualitatively interpreting partial molar volume data in ternis of

hydration uses the following relationship16:

O = VM + A V ~ = VU + nh(vho - vo ), (1)

Where VM is the intrinsic volume, which is the geometric volume occupied by a

solute molecule itself; AVh is the hydration-induced effect on the solute volume;

VOO and vh0are the partial molar volumes of water in the bulk state and in the

hydration shell of solute, respectively; and nh is the "hydration number" which

refers to the nurnber of water molecules in the hydration shell of a solutel?

The partial molar adiabatic compressibility, K:, of a solute is a linear

function of the first pressure derivative of the partial molar vol~rne'~. It

represents the apparent compressibility of 1 mole of a solute at infinite dilution 15.

Analogous to partial molar volume, the partial molar adiabatic compressibility.

K:, can be presented as the sum of the intrinsic, KM, and hydration, &,

contributions. The intrinsic compressibility, KM, of a solute is a measure of

intramolecular interactions, while the hydration induced change in the solvent

compressibility, K:, reflects the influence of the solute molecules on the

soivent 15.

The partial specific adiabatic compressibility of a globular protein can be

considered to be the surn of positive and negative c~ntributions'~.

K0s = + nh(b - k)

The positive contribution represents the intrinsic compressibility, KM, which

arises from the irnperfect packing of the arnino acid residues within the solvent

inaccessible cure of the protein. The negative contribution, Kh. represents a

decrease in the solvent compressibility due to the interactions of water molecules

10.17 with the solvent - exposed groups of the protein .

The partial molar adiabatic expansibility, E', of a solute is the first

temperature derivative of the partial molar volume, p, and represents the

apparent thermal expansibility of 1 mole of a solute at infinite dilution j5. It also

c m be represented by the sum of the intrinsic, EM, and hydration. Eh,

contributions 14. For small molecules, such as low molecular compounds studied

in this work, intrinsic contributions to expansibility, €O, and compressibility, K:,

are predominantly detemined by properties of chemical bonds and extemal

electron shells within the molecules. These are small and can be neglected. By

contrast, the hydration-induced changes in the solvent compressibility and

expansibility, refiecting the 'quaiity* and the "quantity" of solute-solvent

interactions are large in magnitude j5.

The concentration dependences of the apparent molar volumes and

compressibilities are known to be negligible in the range of concentrations used

in this thesis'! Within the limits of experirnental error, the apparent molar

volumes and adiabatic compressibilities coincide with the values of the partial

molar volume and adiabatic cornpressibility, obtained by extrapolation to infinite

dilutioni6.

Relationships between macroscopic volurnetric properties of biological

substances and the hydration properties of constituent groups depend on the

chemical nature of these groups. Charged groups influence adjacent water

molecule dipoles by causing strong electrostatic contraction. This contraction

causes a partial loss in the mobility of hydrating waters, which, in tum, results in

a diminution in the partial molar volume and compressibility of these waterss2'.

Polar groups in a solute generally influence adjacent water molecules via

hydrogen bond fornation. The properties of water around a polar group strongly

depend on the distance between the group and other polar rnoietîes. If polar

groups within a solute molecule are situated sufficiently close to each other, each

adjacent water molecule can form simultaneously two hydrogen bonds with the

neig hboring polar groups. This leads to partial irnmobilization of water molecules

within the solute hydration shell and consequently, to a decrease In their

cornpressibility and volume. The influence of closely situated polar groups on the

adjacent water molecules qualitatively resembles that of charged groups. Water

molecules in the hydration shell of an isolated (single) polar groups fom a single

hydrogen bond with a solute molecule. The mobility of the water molecules

bound to single polar groups and their ability to f o n hydrogen bonds with other

solvent molecules are similar to those of water molecules in the bulk stateP20.

The contribution of intermediate polar groups to the partial molar adiabatic

compressibility and partial molar volume of a solute is intemediate between the

contribution of a single polar group and that of closely positioned polar gro~ps'~.

Water molecules sunounding nonpolar groups do not interact directly with

the solute molecules. Therefore. the contribution of nonpolar groups to

interaction volume is negligible. These water molecules are forced to f o n their

complement of solvent-solvent hydrogen bonds within a more limited space.

thereby becoming more ordered than bulk v*3er1'. Water molecules that solvate

aliphatic groups are less compressible than bulk water at low temperatures,

while, at higher temperatures, they become more compressible relative to bulk

water. At temperatures below 30 OC the compressibility of water near non-polar

groups is less than that of bulk waterlg.

1.2 Model compounds

Use of low molecular weight compounds

There are several approaches that are currently used to interpret

macroscopic volumetric data or proteins in tems of hydratiod7. In one approach,

the protein hydration is characterized based on the volumetric contributions of

the surface atomic groups. To obtain the requisite volurnetric characterizations

of the atomic groups that comprise the solvent-accessible surface of proteins,

low molecular weight model compounds containing the same groups can be

studied?

Such model system studies have at least two advantages. First.

microscopic interpretation of experimental data denved from studies on low

molecular weight compounds is relatively easy. Second, model studies readily

allow one to systernatically alter the structure, by using homologous series, so

that the contribution of a chosen atomic group can be assessed '?

Previous investigations have characterized the volumetric properties of

linear, branched and cyclic r n o l e c ~ l e s ' ~ * ~ ~ ? In particular, hydration properties

of charged, nonpolar and polar groups have been evaluated. These

investigations have revealed that volurnetic contributions of component atomic

groups of a çolute strongly depend on the proximity and chernical nature of

neighboring groups. These observations can be interpreted in tems of

interactions between the hydration shells of cornponent atomic groups. Based

on these results, it has been concluded that hydration shells of the charged

amino and carboxyl termini involve only 1 -1.5 effective layers of water

16.17 molecules .

To establish the database required to understand and to interpret at a

molecular level the volumetric properties of charged, polar and nonpolar groups,

mode1 compounds have been widely studied in light wate?? Among most

suitabie srnail moiecuies that mimic solvent-exposed protein groups are

oligopeptides, monosaccharides. and aminoacids.

Oligopeptides

In recent years'5, the volumetric properties of oligopeptides containing

amino acids with various sidechains have been invest~gated~'~~. Contributions

of amino acid sidechains strongly depend on the proximity and the chernical

nature of neighboring atomic groups. These observations suggest interactions

between the hydration shells of amino acids side chahs and the neighboring

groups. These data lead to the important conclusion that, due to the complexity

of the hydration patterns of even these small molecules, care must be exercised

when the global hydration of such molecules is modeled as the sum of the

hydration contributions of the constituent atomic groups.

The simplest oligopeptide homologous series is represented by

oligoglycines. At neutral pH, oligoglycines contain oppositely charged arnino and

carboxyl terminal groups, one or more peptide groups, and aliphatic -CH2-

groups. The hydration properties of these cornponent groups can be obtained

from volurnetric studies of oligoglycines. The partial molar volumes,

expansibilities, and compressibilities of short oligoglycines have been evaluated

O 17.20 in light water over the temperature range 18 - 55 C.

To interpret the hydration properties of the charged and uncharged termini

as a function of the proximity and chemical nature of the adjacent non-polar side

21.24 chains, diglycyl tripeptides and glycil dipeptides2' with aromatic and aliphatic

side chains have been studied volumetrically. These results have been

interpreted in ternis of hydration of the constituent atomic groups, thermodynamic

contributions of these groups, and mutual interactions between them. 29-30

Monosacchatides

Simple sugars, consisting of an aliphatic moiety and polar hydroxyl

groups, are appropriate models for describing hydration properties of polar

groups. Volumetric investigations have revealed strong stereochemical

dependence of the hydration properties of mono sac char ide^^^. Their closely

related stereoisomers exhibit different values of partial molar volume and partial

molar adiabatic compressibilities. For example, polar groups of 2-deoxyribose,

2-deoxyglucose, and 2-deoxygalactose exhibit less negative compressibility

contributions and srnaller contraction of water relative to polar groups of

pentoses, 6-deoxyhexoses and hexoses. To explain this observation. it has

been suggested that some hydroxyl groups (e.g., the hydroxyl group in the 2-

position) play important role in cooperative enhancement of sugars hydration.

while the role of other hydroxyl groups (e.g., the hydroxyl group in the 6- position)

is modest?

Amino acids with various side chains

The volumetric propeities of many amino acids with various sidechains

were investigated v~lurnetrically'~*' 7-21*24-27*33 . Again, it was found that the

volumetric contributions of amino acid sidechains strongly depend on the

proximity and the chemical nature of neighboring atornic groups. These

obsewations are explained in terms of the interactions between the hydration

shells of amino acids side chains and adjacent charged tem~ini'~? Also, it was

suggested that the region of influence of a polar center does not extend beyond

1 1.21 the p carbon atom . Microscopie interpretation of the volumetrk properties of

the naturally occurring amino acids is often based on cornparison of nonpolar

17.33 amino acids and amino acids containing hydrophilic groups .

One way to estimate the extent of solute-induced solvent perturbation

(hydration) is to determine the minimum distance within a single molecule at

which two groups no longer interact with each other via solvent. which is the

distance at which hydration shells of two atomic groups no longer ~ver lap'~. This

approach requires a system in which the distance between the two groups can

be systematically varied: by studying homologous series13. Such an approach

was employed in the study of a,o-amino carboxylic acids3'.

a,oAmino carboxylic acids

Partial molar volumes, expansibilities. and adiabatic compressibilities of a

homologous series of eight a,o-aminocarboxylic acids in aqueous solutions were

determined16 within the temperature range 18 - 55 OC. These substances exist

at pH 7 as mitterions containing two oppositely charged carboxyl and amino

terminal groups separated by a nonbranched chah of CH2 groups. a,o-

Arninocarboxylic acids differ from oligoglycines by the fact that the charged

amino and carboxy termini are separated by an aliphatic (-CH2-)" chain rather

than a (-CH2CONH-), chain. For this reason, a comparison of the data obtained

on these homologous series of compounds may provide insight into the hydration

features of the polar peptide group.

To explore solute-solvent interactions in the vicinity of the polar peptide

group, one should compare the partial molar volumes of a,o-aminocarboxylic

acids with the corresponding values for the ~l igogl~cines'~. For such

comparisons, one must choose pairs of oligoglycines and a,o-aminocarboxylic

acids that have similar van der Waals volumes and accessible surface areas,

and the same nurnber of covalent bonds between the charged termini, so that the

intercharge distances are approximately the same. Using these criteria, the

appropriate cornparisons are diglycine and 5 - aminopentanoic acid, triglycine

and 8 - aminooctanoic acid, and tetraglycine and 1 1 - aminoundecanoic acid17.

Significantly, the van der Waals volumes of these "pairedn oligoglycines and a,o-

aminocarboxylic acids differ by less than 1 %. With respect to the number of

covalent bonds, the van der Waals volume and the accessible surface area, each

-CH2CONH- group in the oligoglycines corresponds to three aliphatic -CH2-

groups in the a.o-aminocarboxylic acids17.

Also, a,o-aminocarboxylic acids are convenient model systems for

studying hydration effects of charged and hydrophobic atomic groups, as they do

not contain other atomic groups3'. The distance between the carboxyl and amino

groups c m be varied systernatically by changing the number of aliphatic CH2

groups in the molecule, allowing to study the interaction of the charged ends as a

function of the distance between them. The results are consistent with a model

in which the oppositely charged amino and carboxyl groups cease to interfere

with their individual hydration shells when they are çeparated by at least four CH2

groups. Thus, 5-aminopentanoic acid can be considered as a zwitterionic

rnolecule with a distance between the charged atomic groups corresponding to

the geometric sizes of the amino and carboxyl groups.

This distance is estimated as a value of 6 angstrom, which corresponds to

2-2.5 water molecule diameters3'. This result suggeçt that even for such strong

solute-solvent interactions, like those between water molecules and the charged

terminal groups in the a, w-aminocarboxylic acids, the hydrating water molecules

appear localized mostly within the first coordination sphere.

1.3 Heavy water, its prcperties and trends in research

Themodynamic solvent isotope effect

One approach to characterize a thenodynarnic systern is to slightly

perturb the equilibrium of the system and to monitor themodynamic

consequences of the perturbation. An ideal perturbation of this sort for aqueous

solutions is to substitute the H20 with D20. None of the specific molecular

interactions are significantly altered in this case, even though heavy water affects

the themodynamics in a complex way. Themodynamic solvent isotope effect is

manifested in al1 thermodynamic parameters: free energies, enthalpy, entropy,

neat capacig4? The thennodynamic solvent isotope effect (TSlEj can be used

as an experimental tool for characterizhg the hydration properties of various

atomic groups and exploring the solvent-dependent conformational preferences

of proteins and nucleic acids?

Several models explain the molecular basis of the TSIE. Mixture models

consider water as a mixture of interconverting species with different

arrangements of hydrogen bondsJ4. Solutes act in this model as structure

rnakers (e.g., nonpolar groups) or breakers (e.g., e~ectrolytes)~.~~. Water in the

vicinity of a solute is arranged in the hydration shell and exhibits structural and

thermodynamic properties distinct frorn those of the bulk water. TSIE thus arises

from differential hydrogen bonding of water of hydration relative to bulk water in

H20 versus D@. Models based on the scaled particle theory ernphasize the

differences in the work of cavity formation in Iight versus heavy water due to

differential density, expansibility and cornpres~ibil ity~~*~?

Differential thermodynamics of light and heavy water

The two solvents are composed of molecules of almost identical dipole

moment, shape, size, bond lengths3'. The rnolecular dimensions and the

hydrogen bond lengths in ice of light water and heavy water are identical? The

viscosity, the melting point, the temperature of maximum density, and heat

capacity are higher in liquid neavy ~ a t e ? ' * ~ ~ . To account for the obsewed

differences between the properties of light and heavy water. it has been

proposed that higher structural order exists in D20 relative to H20. which implies

the higher degree of hydrogen bondingJ7.

Formation and breaking of hydrogen bonds in liquid water is a cooperative

process: several bonds form and break simultaneously. Hence, highly hydrogen

bonded short-lived clusters are embedded in the liquid among non-bonded water

molecules. The clusters are compact in shape, since this should produce

maximum stabilityJ8. The cluster size is considerably higher for heavy than for

light water at low temperatures, while with increasing temperature it drops

fasteFg. At higher ternperatures. both liquids appear to have similar degrees of

order.

It is reasonable to expect that the differential energetics of hydrogen and

deuterium bonds are related to the differences in vibrational energy. Since

hydrogen and deuteriurn are electronically identical, the electronic part of the

total energies of the related complexes should be equal. Consequently, any

differences in the energetics of hydrogen bonding should be associated with the

masses, and, primarily, with the vibrational energies. The intemolecular

vibrational mode of highest frequency can be characterized as a bending motion

of the proton donor molecule, which distorts the linearity of hydrogen bond. The

heavier mass of deuteron causes a lowering of the frequency and energy of this

vibrational mode3? A large difference in zero point vibrational frequencies is

another factor determining the differential properties of the two isotopic forms of

wate?: the greater the degree of structure, the higher the temperature of

maximum density, viscosity, heat of vaporization, and the heat of sublimation of

heavy water.

Differential solvent properties of light and heavy water.

In general, proteins exhibit enhanced thermal stability in heavy water

relative to light ~ a t e r . ~ ' Changes in desolvation energies, effects upon protein-

protein and protein-solvent hydrogen bonds are factors to be consideredJg. The

isotopic substitution of hydrogen by deuterium affects the strength of solvent-

solvent, solute-solute and solvent - solute hydrogen bonds4'. It is reasonable to

expect that the changes of stability and folding rates of proteins in heavy water,

at least, partially, result from this combination". The change in hydrogen bonding

between water molecules will influence hydrophobic interactions. while the

change in amide-carbonyl, amide-water of carbonyl-water hydrogen bonding will

influence the energetics of solvation of polar groups of the polypeptide backbone

in proteins? Stronger hydrogen bonds in heavy water perhaps account for

negative free energies of transfer of non-polar groups into heavy water, as

37,39 opposed to the positive values observed in H20 .

The therrnodynamics of transfer of solutes from H20 to 40 may be

presented as the sum of four contributions: (1) hydrogen exchange (replacement

of exchangeable hydrogen atoms by deuterium atoms); (2) the change in the

state of protonation of the solute molecule due differential affinity of ionizable

groups for H' and D'; (3) differential hydration properties of charged, polar and

nonpolar groups in light and heavy water; and (4) structural changes of the solute

molecule that may occur as a consequence of items 1 to 339.

1.4 Thesis outline

Comparative studies of solution thermodynarnics carried out in light and

heavy water provide additional insights into hydration properties of various

functional groups of proteins and DNA. In particular, there have been attempts to

elucidate water isotope effects in hydrophobic hydration in terms of the partial

molar heat capacities, enthalpies, entropies and Gibbs free energies 36.39.41

Because of the exquisite sensitivity of volumetric parameten with respect tu

solute hydration, the differential solvent-isotope approach should prove

especially viable when employed in volumetric investigations. In general. the

partial molar volume and adiabatic compressibility of an atomic group in H20 are

distinct from those in D20, which corresponds to differential hydration properties

of these groups in H20 and D20. This work is aimed at understanding the

differential hydration of charged, polar and nonpolar groups in light and heavy

water as a function of temperature. To this end, we have begun to establish a

systematic library of the hydration properties of functional groups of different

11 .l4 chernical nature (charged, polar, and nonpolar), in light and heavy water .

More specifically, in this work, homologous series of oligoglycines, a -

amino acids with aliphatic side chainsl' and a,o-aminocarboxylic acidst4 have

been studied volumetrically. Chernical structures of these substances are shown

in the Appendix.

These relatively simple organic molecules are appropriate models for

biopolymers since they consist of groups identical to constituent groups of

proteins and nucleic acids. Consequently, volumetric characteristics of these

simple compounds can be used to rlucidate basic pnnciples governing patterns

of hydration of more complex biochemical

1517 systems .

The first part of this work (Chapter 2) is devoted to studying the

homologous series of oligoglycines (from glycine to pentaglycine) and a - amino

acids with nonbranched aliphatic side chains (from glycine to norleucine) in

hoavy water at a single temperature of 25 OC.

The second part is devoted (Chapter 3) to studying the homologous series

of a,w - aminocarboxylic acids in heavy water over the wide temperature range

of 18 to 55 OC.

At neutral pH, al1 three homologous series exist as mitterions containing

oppositely charged carboxyl and arnino groups. In a,w - aminocarboxylic acids

the charged termini are se~antsd by a nonbranched chain of aliphatic -CH2 -

groups. In oligoglycines, the charged temini are separated by glycil units

consisting of a polar peptide group plus an aliphatic -CHz - group. a - amino

acids contain the closely located arnino and carboxyl termini and a nonbranched

aliphatic side hai in'*-'^.

The main objectives of this study are:

To determine the volurnetnc characteristics of biological substances in heavy

water.

0 To compare these values to those previously reported for the same

compounds in H20.

To quantitatively characterize the differential hydration properties of

component charged, polar and nonpolar groups in light versus heavy water.

To examine temperature dependencies of these properties.

References

P. R. Connely at al., Probing hydration contributions to the thermodynamics of

ligand binding by proteins. Biochemistry, 32, 5583, 1993.

Van Regenmortel M.H.V., Molecular design versus empirical discovery in

peptide-based vaccines. Vaccine, 18,216.2000.

Connelly P.R. The wst of releasing site-specific, bound water molecules

from proteins toward a quantitative guide for structure-based drug design.

In: Structure-based drug design. Editors: J.E. Ladbury, P.R. Connelly.

Springer, 1 997.

Poomima C.S.. Dean P.M., Hydration in drug design. 1. Multiple hydrogen-

bonding features of water molecules in mediating protein-ligand interactions.

J. Cornput-Aided Moi. Design., 9,500, 1995.

Ladbury J.E. Just add water! The effect of water on the specificity of protein-

ligand binding sites and its potential application to drug design.

Chernistry and biology, 3, 973, 1 996.

Renzony DA., Zvellebil M.J.J.M., Lundback T., Ladbury J.E. Explon'ng

uncharted waters: water molecules in drug design strategies. In: Structure-

based drug design. Editors: J.E. Ladbury, P.R. Connelly. Springer,1997.

Raymer M.L et al. Predicting conserved water-mediated and polar ligand

interactions in proteins using a K-nearest neighbors genetic algorithm.

J. Mol. Bioi. 265,445, 1997

Shoichet B.K. at al., Ligand solvation in molecular docking.

Proteins. 34,4, 1999.

Connelly P.R., Thomson J.A., Fitzgibbon M.J., Bruuese F.J.,

Probing hydration contributions to the thennodynarnics of ligand binding by

proteins.

Biochernistry, 32, 5583, 1993.

10.Sorenson J. at al. Determining the role of hydration in protein folding.

J. Chem. Phys. B., 103,26, 5413.1 999.

1 1. Likhodi O., Chalikian T. V.,

Partial molar volumes and adiabatic compressibilities of a series of aliphatic

amino acids and oligoglycines in D20.

J. Am. Chem. Soc, 121.1 156,1999.

12.T. V. Chalikian, M. Totrov, R. Abagyan and K. J. Breslauer

The hydration of globular proteins as derived frorn volume and

compressibility measurements: cross correlating themiodynamic and

structural data. J. Mol. Biol., 260, 588, 1996.

13.T. V. Chalikian. A. P. Sarvazyan, K. J. Breslauer.

Hydration and partial compressibility of biological compounds.

Biophys.Chern., 51.89, 1994.

14. Likhodi O., Chalikian T.V.,

Differentiai hydration of a,o- arninocarboxyiic acids in D20 and H20.

J. Am. Chem. Soc. ,122,7860,2000.

15. Chalikian T.V., Breslauer K.J.,

Thermodynamic analysis of biomolecules: a volumetric approach.

Cun. Opin. Sbuct. Biol.. 8:657, 1998.

16.Chalikian T.V., Satvazyan A.P., Breslauer K.J. Partial molar volumes,

expansibilities, and compressibilities of a.o-aminocarboxylic acids in aqueous

solutions between 18 and 55 OC. J. Phys. Chem., 97: 1301 7-1 3026, 1993.

17. Chalikian T., Satvazyan A., Funk T., Breslayer K.

Partial rnolar volumes, expansibilities, and cornpressibilities of oligoglycines

in aqueous solutions at 18-55'~. Biopolymers, 34,541,1994.

18. Kharakoz D.P., Volumetric properties of proteins and their analogues in

diluted water solutions. 2. Partial adiabatic compressibilities of aminoacids at

15-70'~. J. Phys. Chem. 95,5634, 1991.

19.Kharakoz D.P. Volumetric properties of proteins and their analogues in diluted

water solutions. J. Phys. Chem., 95: 5634, 1991.

20. N. Azizov. Effect of hydrophobic hydration on volume characteristics of

solution. Russian Journal of General chemistry, 67.4. 527,1997.

21. Lepori L., Gianni P., Partial molar volumes of ionic and nonionic organic

solutes in water: a simple additivity scheme based on the intrinsic volume

approach. J. Soi. Chem., 29,5,405.2000.

22.Gianni P., Lepori L.. Group contributions to the partial molar volume of ionic

organic solutes in aqueous solution. J. Sol. Chem., 25, 1, 1 - 19, 1996.

23.Schwitzer M.A., Hedwig G.R., Thermodynamic properties of peptide

solutions. 16. Partial molar heat capacities and volumes of some tripeptides

of sequence Gly-X-Gly in aqueous solution at 25 OC.

J. Chem. Eng. Data, 43,477, 1998.

24. Hedwig GR., Hoiland H., Hogseth E., Thermodynamic properties of peptide

solutions. 15. Partial molar isentropic compressibilities of sorne glycyl

dipeptides in aqueous solutions at 15 O and 35 OC.

J. Sol. Chem., 25, 11, 1041, 1996.

25.Hakin A.W., Hedwig G.R., The partial molar heat capacities and volumes of

some N-acetyl amino acid amides in aqueous solution over the temperature

range 288.15 to 328.15 K.

Phys. Chem. Chem. Phys., 2.8 ,1795,2000.

26.Hackel M., Hinz H.J., Hedwig GR., Model compounds for the evaluation of

the partial rnolar heat capacities of amino acid side chain in proteins.

Thenochim. Acta, 308, 23, 1998.

27. Hedwig G.R., Partial molar isothermal and isentropic campressibilities of

glycine, alanine, and glycylg lycine is aq ueous solution at 25'~.

J. Phys. Chem., 99,31,12063,1995.

28. Mishra A.K., Ahluwalia J.C., Apparent molar volumes of amino acids, N-

acetoamino acids, and peptides in aqueous solutions.

J. Phys. Chem., 88,86, 1984.

28. Chalikian T.V., Gindikin V.S., Breslauer K.J., Hydration of diglycyl tripeptides

with non-polar side chains: a volumetric study.

Biophys. Chem. ,75, 57, 1 998

29. Hedwig G.R. Partial molar heat-capacities volumes and compressibilities of

aqueous solutions of some peptides that model sidechains of proteins.

Pure and Applied Chemistry, 66 ,3 , 387-392, 1 994.

30. Hedwig G.R.. Hastie J.D.. Hoiland H., Thermodynamic properties of peptide

solutions. 1 4. Partial rnolar expansibilities and isothermal compressibilities of

some glycyl dipeptides in aqueous solution. J. Sol. Chem., 25, 7, 61 5, 1996.

31. Chalikian T.V. Ultrasonic and densimetric characterizations of the hydration

properties of polar groups in rnonosaccharides.

J. Phys. Chem., 102: 6921 -6926, 1998.

32. Kharakoz D. P. Volumetric properties of proteins and their analogs in diluted

water solutions. 1. Partial volumes of amino acids at 1 5-50 CO.

Biophys. Chem., 34, 1 15-1 25, 1989.

33.0as TG., Toone E.J., Thermodynamic solvent isotope effects and molecular

hydropho bicity. Adv. Biophys. Chem. 6:1-52,2OOO.

34. N. Muller. Search for a realistic view of hydrophobic effects.

Acc. Chem. Res., 23,23,1990.

35. Scheiner S., Cuma M ., Relative stabiiity of hydrogen and deuterium bonds.

J. Am. Chern. Soc., 118:1511,1999.

36. Conway R. E., lonic Hydration in Chernisby and Biophysics (Elsevier,

Amsterdam, 1981 ), p.551.

37. Nemethy G., Sheraga H.A., Structure of water and hydrophobic bonding in

proteins. IV. The thenodynamic properties of liquid deuterium oxide.

J. Chern. Phys., 41, 3 , 680, 1964.

38. Kuhlman B., Raleigh D.P., Global analysis of the thermal and chernical

denaturation of the N-terminal domain of the ribosomal protein L9 in H20 and

D20. Protein Sci., 7,11,2405-2412.

39. Parker M.J., Clarke A.R., Amide backbone and water-related HID isotope

effects on the dynamics of a protein folding reaction.

Biochemistry, 35,5786, 1999.

40. Lopez M. M., Makhatadze G.I., Solvent isotope effect on thermodynamic of

hydration. Biophys. Chem.,74, 1 17-1 25, 1998.

Chapter 2. Partial molar volumes and adiabatic

compressibilities of a series of alipha tic aminoacids and

oligoglycines in DzO

Reproduced with permission from Journal of the Amencan Chemicai Society,

1999,121,1156-1 163. Copyright 1999, American Chernical Society.

Dr. Tigran Chalikian did some of the calculations and experiments on

oligoglycines and aliphatic amino acids.

Abstract: This paper reports the first characterization of the hydration properties

of some amino acids and oligoglycines, low rnolecular weight analogs of proteins

in 40. Specifically, the partial molar volumes, VO, and adiabatic

compressibilities, KoS, of 5 a-amino acids and 5 oligoglycines have been

determined in D20 at 25 OC. The resulting data have been used to estimate the

volume and compressibility contributions of the component nonpolar (methylene

group), polar (peptide group), and charged (oppositely charged amino and

carboxyl terminal groups) chemical groups. It was found that the volume and

compressibiI*ty contributions of these charged, polar, and nonpolar groups in D20

are "measurablyn distinct from those in H20. This distinction, in principle, may

allow one to develop a method by which differential volumetric measurements of

proteins in D2O and Hz0 can be used to gain insight into the nature of the sol)--nt

exposed protein groups in the absence of detailed structural information.

1. Introduction

Hydration is widely acknowledged to be one of the major forces driving pmtein

recognition events, in particular, protein foldinglunfolding transitions 11-31.

Consequently, hydration properties of globular proteins in their native and

denatured states have been under intensive scnitiny using various experimental

and cornputational techniques [4-71. In this connection, the volumetric

charactenstics of substances (e.g., the partial molar volume and adiabatic

compressibility) have proven to be reflective of and sensitive to solute-solvent

interactions [8-IO]. Consequently, volumetrk properties represent useful

observables for studying the hydration properties of proteins. In recognition of

this fact, several laboratories have investigated the volumetric properties of

proteins in aqueous solutions and have proposed different approaches for

interpreting these macroscopic data in terms of protein hydration (1 1-1 91. Such

interpretations are not straightfonnrard and always model-dependent. However,

despite the difficulties in interpreting the volumetric data for systems as complex

as proteins, experiments of this type has begun to provide important data against

which difTerent models of protein hydration can be evaluated [20, 211.

The microscopic interpretation of the measured volumetric properties in t e n s

of protein hydration is usually performed in conjunction with structural data on the

surface atomic groups [l 1 , 13, 181. Unfortunately, such structural data are not

always available, especially for the denatured states of proteins that include

molten globule and unfolded states. Consequently, the microscopic interpretati~n

of the volumetric properties of denatured protein states remains highly

speculative and subjective in nature [14-17, 191. This limitation is serious and

prevents published results from being used with confidence for analysis of the

hydration features of proteins as a function of their conformational states.

One potential way to denve information on the solvent-exposed atomic groups

of a protein in solution in the absence of structura! data is to conduct differential

measurernents of the partial molar volume, VO. and partial molar adiabatic

compressibility, Ks, in Iight (H20) and heavy (D20) water. Note that the physico-

chemical properties of D20 are not very different from those of H20.

Consequently, the substitution of H20 with 90 should cause only a very small

perturbation of the structural preferences of a solute.

A differential solvent isotope approach rnay prove viable if the volurnetric

contributions of the water-accessible atomic groups of different chemical nature

(charged, polar, and nonpolar) in D20 are distinct from those in H20. To this end,

a systematic library of the hydration properties of charged, polar, and nonpolar

groups in H20 and D20 should be established. As a first step towards this goal,

this paper reports the partial rnolar volumes, VO, and adiabatic compressibilities,

Kos, for two homologous series in D20 at 25 OC: arnino acids with aliphatic side

chains (glycine, alanine, a-arninobutiric acid, norvaline, and norleucine) and

oligoglycines (glycine, diglycine, triglycine, tetraglycine, and pentaglycine). The

corresponding values of the partial molar volume and adiabatic compressibility of

the amino acids and oligoglycines in H20 have been reported previously [22-281.

Both homologous series consist of component chemical groups identical to

those that can be found in proteins. Specifically, at neutral pH, the amino acids

studied here consist of a zwitterionic skeleton of oppositely charged amino and

carboxyl termini and a nonbranched chah of aliphatic groups. Oligog lycines

contain oppositely charged amino and carboxyl terminal groups, one or more

polar peptide groups, and aliphatic -CHr groups. The main objective of this study

is to characterize the hydration properties of the component chemical groups of

amino acids and oligoglycines in D20 and compare these values to those

previously reported for the same groups in H20.

2. Materials and Methods.

All chemicals used in this study were of the highest purity commercially

available and were used without further purification. Specifically, the

oligoglycines studied here (diglycine, triglycine, tetraglycine, and pentaglycine)

were purchased from Fluka (Switzerland), while the a-amino acids (glycine,

alanine, waminobutiric acid, norvaline. and norleucine) and D20 were obtained

from Sigma Chemical (USA). All the amino acids were of the L-stereisomenc

form.

Prior to the densirnetnc and ultrasonic velocimetric experiments, the amino

acids and ohgoglycines were dissolved in D20 and lyophilized to exchange labile

protons for deuterons. The concentrations of the sarnples were determined by

weighing 10-20 mg of solute with a precision a.03 mg, and then dissolving the

matenal in a known amount of 40. Before weighing, amino acids glycine,

alanine, a-aminobutiric acid, norvaline. and norieucine were dried at 110 OC for

24 hours, while oligoglycines diglycine, triglycine, tetraglycine, and pentaglycine

were dried for 72 hours under vacuum at room temperature in the presence of

phosphorus pentoxide.

Densities were rneaçured at 25 O C with a precision -+1 .5-1W6 g cmœ3 using a

vibrating tube densimeter (DMA-60, Anton Paar, Austria). These density values

were used to calculate the apparent molar volume, 4V, using the well-known

equation [29]:

ov = M ~ P - (P - PO) 1 (popml (1

where M is the solute molecular weight; rn is the molal concentration; and p and

PO are the densities of the solution and solvent, respectively. The requisite value

for the density of DzO, po, equal to 1 .IO4449 gcm'3 at 25 O C was taken from Kell

[301.

The apparent molar adiabatic cornpressibility, $KS, was calculated from the

densimetrk and ultrasonic data using the expression [31, 321:

4Ks = Pso (2W - 2WI - Mbo) (2)

where [U] = (U - Uo)/(UaC) is the relative molar sound velocity increment; U and

Uo are the sound velocities in the solution and solvent, respectively; and Pso is

the coefficient of adiabatic compressibility of water. The value of Bso, equal to

46 .2~10 '~ bai1 at 25 OC was calculated from data on the dençity 1301 and sound

velocity [33] of 40 using the expression pso = ( p ~ ~ ~ 2 ) - ' .

Using a differential device [34] based on the resonator method [34-381,

ultrasonic velocities in the solutions of the amino acids and oligoglycines were

measured at a frequency of about 7.5 MHz with a relative accuracy kloœ4 % ai

25 O C . These sound velocity measurements were perfomed as previously

described [28, 391.

For each evaluation of +V or $Kç, three to five independent rneasurements

were carried out within the concentration range of 3-4 mglrnl for each of the

amino acids and oligoglycines studied, with the exception of pentaglycine. For

this longer homologue, concentrations of about 1 mglml were used because of its

low solubility.

3. Results

The relative molar sound velocity increments, [U], apparent molar volumes.

@VI and apparent molar adiabatic cornpressibilities, $Ks, at 25 OC for the amino

acids and oligoglycines are shown in Tables 1 and 2, respectively. The reported

errors include contributions frorn the solute concentration and apparatus

limitation, as well as any temperature variability in the measuring cells. Previous

studies have shown that, for amino acids and oligoglycines in H20, the apparent

molar volumes and the apparent molar adiabatic compressibilities do not depend

strongly on concentration [22, 23, 26, 271. By extension, one may assume that, in

D2O too, the concentration dependences of (IV and t)G are weak. In other words,

the apparent molar volumes, (IV, and adiabatic compressibilities, $Kç, of the

amino acids and oligoglycines determined in the concentration range of 1-4

mg/ml can be assumed to wincide with the partial molar volumes, VO, and

adiabatic compressibilities, Kos, obtained by extrapolation to infinite dilution.

Consequently. below, the apparent molar and partial molar characteristics of the

amino acids and oligoglycines will be treated as equivalent.

4. Discussion

4.1 Partial Molar Volume

Theoretical Considerations It is convenient to consider the dissolution of a

solute as consisting of two steps: (i) the creation of a cavity in the solvent large

enough to enclose the solute rnolecule; and (ii) the introduction into the cavity of

a solute molecule that can interact with the solvent. Consequently. the partial

molar volume of a solute at infinite dilution, Vol represents contributions from

each step [40-431:

V" = VM + VT + VI + pTORT (3)

where VM is the geometric volume occupied by the solute molecule itself; VT is

the "thermal" volume (the volume of the void space surrounding the solute

molecule), which is due to the thermal motion of solute and solvent molecules; VI

is the volume change associated with solute-solvent interactions, the "interaction

volume": PTO is the coefficient of isothermal compressibility of the solvent; R is the

universal gas constant; and T is the absolute temperature. The terni PTORT

describes the volume effect related to the kinetic contribution to the pressure of a

solute rnolecule due to translational degrees of freedom [Ml. For D20 at 25 OC,

the ideal terni, BToRT, in eq. (3) is equai to 1.15 cm3moi-' wmpared to 1.12

cm3mor1 for H20. Note that the sum VM + VT in eq. (3) represents the partial

molar volume of the cavity, Vc, enclosing the solute.

Scaled particle theory [40-42, 441, by ernploying simple statistical mechanical

and geornetric arguments to describe the dissolution of a solute, allows one to

evaluate the intnnsic and thermal contributions of a spherical solute to the partial

moiar volume, V". A characteristic feature of scaied particie Lheory is that it

considers dissolution themodynamics in t e n s of geometnc properties of the

solute and thermod ynamic pro perties of the solvent. Scaled particle t heory has

been often used to calculate the partial molar volumes, VO, of nonpolar solutes

(such as gases and hydrocarbons) in aqueous solutions [40.42]. It is worthy to

note that rernarkably good quantitative agreement has been observed between

the calculated and experimental values of VO [40, 421.

Based on the concepts of scaled particle theory, the cavity volume, Vc, of a

spherical solute is given by the expression [40,42]:

Vc = 82.054 PToHd~RT + K ~ ~ J N A / G (4)

where a0 is the volume coefficient of thermal expansion of the solvent; NA is

Avogadro's number; d2 is the effective diameter of the solute molecule; and Hc is

the partial molar enthalpy of the cavity formation that is given by the relationship:

Hc = aa l3~* [6~~ / (1 -y)* + 36~?/(1 -y)3 + y/( 1 -y)] (5)

where y = nd f3~d (6~ "o ) is the packing density of the solvent; dl is the effective

hard-sphere diameter of the solvent molecules; Vo0 is the partial molar volume of

the solvent; B = 2(d12/d1)2 - d12/di; C = (di2/di)2 - d12/d1 + 0.25; and dl2 = (dl +

d 2)/2.

Edward and Farrell [45] have proposed a model in which the thermal volume,

VT, is considered as consisting of a layer of "empty" volume of a thickness A

surrounding a solute molecule. Based on this model, the following expression for

the partial molar volume, Vc, of the cavity enclosing the solute can be derived: Vc

= nNA(dz + 2 ~ ) ~ 1 6 . Consequently. for a spherical solute, the value of A can be

calculated as follows:

A = O . ~ [ ( ~ V & C N ~ ) " ~ - d2] (6)

lmplicit in this model is the assumption that the thickness, A, of the thermal

volume depends primarily on the themodynamic and structural characteristics of

the solvent (such as the coefficient of isothermal compressibility, PTO, and the

packing density, y) while only secondarily depending on the diameter, d2, of the

solute molecule. This assumption can be used to evaluate the effective hard

sphere diameter, dl, of the solvent molecule from eqs. (4) - (6). More specifically,

if the optimum value of dl is used, then the calculations perfomed using eqs. (4)

- (6) should result in the weakest dependence of A on the solute diameter, dz.

Such calculations revealed that, for both H20 and D20, the optimum value of dl

is 2.74 A, in good agreement with previous estimates [43,46.471. Fig. 1 shows

the dependencies of A on dz in H20 and D20, calculated from eqs. (4) -(6) by

using a dl value of 2.74 A. As can be seen in Fig. 1, for any value of dz, the

thickness, A, in D20 is slightly larger than in H20. For solutes with a diameter, d2,

of 2 A and larger. the values of A in D20 and H20 are equal to 0.56 A and 0.55 A,

res pectively .

Fig. 2a shows how the partial molar volume of the cavity, Vc, enclosing a

spherical solute depends on the solute diameter, d2, in D20 (A) and H20 (V),

respectively. Inspection of Fig. 2a reveals that the values of Vc in D20 and H20

are quite similar. To illustrate this point at a greater resolution, Fig. 2b shows the

difference. AVc, between the values of Vc for a sphen'cal solute in 90 and H20

as a function of Vc. As can be seen in Fig. 2b. Vc in D20 is slightly larger than in

HzO. More specifically, when Vc increases from O to 500 cm3mol-'. the difference.

AVc, increases from O to 2 cm3mol*'. Note that, since the intrinsic volume, Vu, of

a solute is the same in D20 and H20, the value of AVc is detenined solely by

the differential value of VT in D20 and H20. Consequently, for solutes with a

value of Vc between 50 and 200 cm3mof' (which roughly corresponds to the

range of values of Vc of the amino acids and oligoglycines studied here). the

thermal volume, Vr, in D20 is larger than in H20 by 0.3 to 1 .O cm3mof'.

Amino Acids. The intrinsic volume, VM, of a simple solute molecule can be

approxirnated by its van der Waals volume, Vw. Thus, by setting VM=Vw in eq.

(3), the difference between the partial molar volume of a solute and its van der

Waals volume, (VO - Vw), becomes equal to the sum (VT + VI + PmRT), which

reflects the volume effect of solute-solvent interactions. The requisite values of

Vw for the amino acids studied here can be calculated using the additive

approach of Bondi [48]. However, it should be noted that the choice of a specific

scheme to divide the solution volume into the solute and solvent components (e.

g., if one uses excluded volume instead of van der Waals volume) may affect the

interpretation of the origin of volume changes associated with various

biomolecular processes [445 11.

Fig. 3 shows how the difference between the partial molar volume, VO, and

van der Waals volume, Vw, of the amino acids depends on the nurnber of carbon

atoms in the side chain. Note that the plot in Fig. 3 is linear. This observation can

be interpreted to rnean that each methylene group added to the moiecule causes

the same alteration in the volume effect of hydration. Thus, the partial molar

volume observable does not detect interactions, if any, between the aliphatic side

chain and the zwitterionic skeleton of the amino acid molecules.

Based on the above discussion, the volume contribution of an independently

hydrated methylene group, V(CH2), can be estimated from the data presented in

the fourth column of Table 1 as the average incrernent of the partial molar

volume per -CHT group in alanine, a-aminobutiric acid, norvaline, and

norieucine. This estirnate yields a value of V(CH2) in D20 equal to 16.720.3

cm3mof' which is sornewhat higher than the value of V(CH2) in H20, 15.9M.4

cm3mof1 [23, 24, 391. Recall that, for nonpolar methylene groups, the volume

contribution, V(CH2), is the sum of its intrinsic volume, VM, and the thermal

volume. VT, with the interaction volume, VI. being negligibly small. The intrinsic

volume. VM, of the -CH2- group is the same in 40 and H20. Consequently, a

small positive difference in the V(CH2) values in 90 and H20 suggests that the

thermal volume, VT, in D20 may be slightly higher than that in H20. This

experiment-based conclusion is in qualitative agreement with the theoretical

results presented in Fig. 2b. For the methylene group with the value of Vc on the

order of 10-20 cm3mol~', the calculated thermal volume, VTi in D20 is about 0.2

cm3rnol-' larger than in H20 (see Fig. 2b).

Differentiating eq. (3) results in the following relationship for the difference,

No, between the partial molar volumes of the amino acids in 90 and H20:

No = AVM + NT + NI + A(PTORT) (7)

The Vw and proRT components of the partial molar volume of the amino acids

are similar in D20 and H20. Inspection of the data in Table 1 reveals that the

partial molar volume, Va, of glycine in D20 practically coincides with that in H20

(No = -0.3k0.5 cm3rnol-'). Furthemore, the thermal volume, VT, of glycine in

D20 is only 0.3 cm3mol-' larger than in H20 (see Fig . 2b). Hence, as is seen from

eq. (7), within our experimental error, AVa I AVi z O. We conclude that, within

t0.5 cm3mol*', the interaction volume, VI, of glycine in D20 is similar to that in

HzO. The interaction volume, VI, of glycine refiects the water contraction due to

electrostatic solute-solvent interactions in the vicinity of the zwitterionic skeleton

of the molecule [24]. Such a contraction of water in the vicinity of charged groups

iç usually referred to as electrostriction. Thus, in D20, the oppositely charged

amino and carboxyl termini of the amino acids cause a contraction of water which

is very close to that in H20.

In a-amino acids, the closely located charged temini interact with each other

via overlapping hydration shells [22,24,25, 28, 391. Therefore, the value of hVi

determined above may not represent the true difference in the electrostriction of

independently hydrated oppositely charged amino and carboxyl groups in 020

and H20. In this respect, long oligoglycines represent a better rnodel for

determining the electrostriction of independently hydrated amino and carboxyl

termini.

Oligoglycines. Fig. 4 shows how the difference between the partial molar

volume, VO, and the van der Waals volume, Vw, of the oligoglycines depends on

the number of peptide bonds in the molecules. Note that two pronounced breaks

are O bserved at points corresponding to diglycine and triglycine. ARer triglycine,

the plot in Fig. 4 becomes linear. Similar observation has been made for the

partial molar volumes, VO, of oligoglycines in HzO [28]. This obsewation has been

interpreted as indicating that, in glycine and diglycine, the oppositely charged

amino and carboxyl termini interact via overlapping hydration shells [28].

Significantly, the degree of overlap should be smaller in diglycine than in glycine,

which accounts for the break point corresponding to diglycine. The amino and

carboxyl tenini in triglycine, tetraglycine, and pentaglycine cease to interact and

become independently hydrated as reflected by the constant dope of the line in

Fig . 4 after triglycine. This interpretation implies that, in "longn oligoglycines, the

glycil units. -CH2COND-, begin to exert an independent and significant influence

on the characteristics of the adjacent water molecules. Hence, any incremental

increase in the partial molar volumes. VO, of these "longn oligoglycines reflects

the contribution of independently hydrated -CH2COND- atomic group.

The wlume contribution of a single -CH2COND- group can be determined

from the data presented in the fourth wlumn of Table 2 as the incrernent of the

partial molar volume, VO, per glycil unit in tnglycine, tetraglycine, and

pentaglycine. This estimate yields a value for the volume contribution of the glycil

unit, V(CH2COND), in D20 equal to 39.8M.8 cm3mol*'. The volume contribution

of the peptide group, V(COND), can be calculated by subtracting the value of

V(CH2) from that of V(CH2COND): V(C0ND) = 39.8 - 16.7 = 23.1I1.1 cm3mol-'.

This value is 1.5 cm3mol-' larger than 21.6 cm3rnol-', the estirnate for the volume

contribution of the peptide group, V(CONH), in H20 [28]. The difference between

V(C0ND) and V(C0NH) can be ascribed solely to the difference between the

interaction volumes, VI, of the polar peptide group in D20 and H20.

Peptide groups contain two polar entities -CO- and -NH- (or -ND-). These

polar entities are capable of forming hydrogen bonds with adjacent water

molecules thereby causing solvent contraction (which is reflected in the value of

VI). In our previous paper [28], the value of VI per peptide group in H20 was

estimated to be equal to -10.5M.5 crn3rnol~', with this value being temperature-

independent. On average. each of the polar groups (-CO- or -NH-) of a peptide

moiety in H20, is characterized by an interaction volume of VI = -5.310.3 cm3rnol-

' [28]. In DzO, the value of VI of a peptide group is equal to -10.5 + 1.5 = -9.Otl.l

cm3mof'. On average, each of the polar groups (-CO- or -ND-) of a peptide

moiety in D20 is characterized by an interaction volume of VI = 0.5x(-9.0k1.1) = -

4.5k0.5 cm3mol-'. Note that the contraction of water caused by polar groups in

D20, as reflected in the values of VI, iç about 15 % smaller than in H20.

Fig. 5 shows the difference. in partial molar volumes, No, of the oligoglycines

in D20 and H20 plotted against the number of peptide bonds in the molecule. In

fact. the values of AVO represent the volume change of transferring the

oligoglycines from H20 to D20. These data can be discussed in the framework of

eq. (7). The intrinsic volume, VM, of the oligoglycines and the ideal term, PToRT,

in D20 and H20 are similar. Furthermore, for oligoglycines, the thermal volume,

VT, in D20 is higher than in H20 by less than 1 cm3mol-' (see Fig. 2b) which is

within the experimental uncertainty of the data in Fig. 5. Consequently, for the

oligoglycines the value of AVO predominantly reflects the difference in the

interaction volume, VI, in D20 and H20: AVO = VI [see eq. (7)].

Inspection of Fig. 5 reveals a pronounced break at the point corresponding to

triglycine. Before triglycine, AVO practically remains constant and close to zero as

the number of peptide bonds increases. By contrast, after triglycine, AVO

increases proportional to the number of peptide bonds. The break point in Fig. 5

indicates the different character of the solute-solvent interactions "before and

beyondn triglycine. Specifically, as discussed above, in glycine, diglycine, and

triglycine, the hydration shell is predominantly determined by electrostatic solute-

solvent interactions. Furthemore, due to a decrease of the overfapping hydration

shells of the amino and carboxyl termini from glycine to triglycine. the effective

hydration of the charged groups increases. However, since charged groups in

D20 and Hz0 exhibe similar VI (as observed above in the case of glycine), the

value of AV' for the oligoglycines does not change and remains close to zero

when the intercharge distance increases from glycine to triglycine. As noted

above, in triglycine and longer homologues, the charged terrnini cease to interact

with each other, and the hydration of the peptide groups begins to influence the

partial molar volume, VO. As the VI of a glycil unit (-CH2COND-) in D20 is higher

(less negative) than in H20, after triglycine, A V O increases as the number of

peptide bonds increases.

To a first approximation, the value of AVO corresponding to triglycine can be

viewed as a measure of the difference in the electrostriction of an independently

hydrated pair of the oppositely charged amino and carboxyl groups in 40 and

H20. As is seen from Fig. 5, this difference is negligibly srnaIl(-0.220.8 cm3mol-

1 ). Therefore, one may reasonably conclude that the electrostriction of an

independently hydrated pair of the oppositely charged amino and carboxyl

groups in D20 is, within f 0.8 crn3mol*', similar to that in HzO. In H20, the

electrostriction of an independently hydrated pair of the charged amino and

carboxyl groups has been estirnated to be -26.0 cm3mol*' [39]. Consequently,

electrostriction of the independently hydrated amino and carboxyl termini in D20

is -26.0I0.8 cm3mol-' .

4.2 Partial Molar Adiaba tic Compressibilities

The partial molar adiabatic compressibility, Kos, of a solute can be

represented as the sum of intrinsic and hydration contributions (1 O]:

K O s = KM + A& = KM + nh(KOh - KoO) (8)

where KM is the intrinsic compressibility of the solute molecules; AKh is the

compressibility effect of hydration; Koo and KOh are the partial molar adiabatic

compressibilities of water in the bulk state and in the hydration shell of the solute,

respectively; and nh, the "hydration number", is the number of water molecules

within the hydration shell of a solute.

For low molecular weight substances, the intrinsic compressibility term, KM, in

eq. (8), is negligibly small since 1 is determined by the compressibilities of

covalent bonds and external electron shells, which are close to zero [IO]. Thus,

the partial molar adiabatic compressibility, KOs, of low molecular weight

substances primarily refiects solvent hydration changes as reflected by the

reduced form of eq. (8):

Cs = nh(Koh - Ko) (9)

Amino Acids. Fig. 6 shows the partial rnolar adiabatic cornpressibility, KoS, of the

amino acids in D20 (0) and H20 (O) plotted against the number of aliphatic

carbons in the side chain. As can be seen in Fig. 6, there is a pronounced break

at the point corresponding to alanine in both plots. After alanine, the plot in Fig. 6

becomes linear. This observation c m be interpreted to mean that only the alkyl

group in the P-position interacts with the zwitterionic skeleton of the molecule,

perhaps, via overlapping hydration shells. After alanine (in a-aminobutiric acid,

norvaline, and norleucine), each rnethylene group added to the molecule is

independently hydrated and, consequently, contributes to the same extent to the

amino acid compressibility. This result should be viewed with caution when

applied to the amino acids in proteins that are branched at the fkarbon, such as

valine, isoleucine, and threonine. Such P-branched compounds, whose bulk side

chains beyond the p-carbon are quite close to the main chain, and therefore their

hydration shells might well overlap those of the adjacent backbone regions.

The compressibility contribution of an independently hydrated methylene

group, K(CH2), can be detemined as the increment of the partial molar adiabatic

compressibility per -CHT group in alanine, a-aminobutiric acid, norvaline, and

norieucine. This estimate yields a value of K(CH2) in D20 equal to -(3.2k0.4)~10~

cm3rnof'bar-' which is significantly smaller (more negative) than -1 .9x104

cm3moî1bai1, the value of K(CH2) in H20 [25]. Thus, at 25 OC. aliphatic groups in

D20 cause a decrease in the solvent compreçsibility, which is about 60 % larger

than in H20.

The compressibility contribution of a methylene group in H20 is known to be

strongly dependent on temperature: it is negative at low temperatures but

becomes positive at high temperatures passing through zero at 30 OC. It is

reasonable to expect that K(CH2) in D20 exhibits an equally strong temperature

dependence. Therefore, one cannot nile out the possibility that the relationship

between the values of K(CH2) in D20 and H20 may be qualitatively and

quantitatively different from that obsewed at 25 O C . Clearly. further temperature-

dependent studies are required to address this important issue.

Further inspection of Fig. 6 reveals that the partial molar adiabatic

compressibility, Kos, of glycine in D20 is (Z.?f 0 .8 )~ 1 O+ cm3moi-' bar" more

negative than in H20. Since the hydration properties of glycine are dominated by

electrostatic solute-solvent interactions, one may reasonably suggest that the

compressibility contribution of charged groups in D20 is lower (more negative)

than in H20. This is in agreement with the previous data on the partial molar

adiabatic compressibility, Kos, of salts in H20 and D20 [471. Specifically,

Desrosiers and Lucas (471 found that the partial molar adiabatic compressibilities,

KSI of 1-1 electrolytes NaCI, KCI, CsCI, and NaF in D20 are srnaller than those

in H20 by 4.4x104, 2.8x104, 1 .7x104, and 5 . 5 ~ 1 0 ~ cm3mo~~'bai', respectively.

Oligoglycines. Fig. 7 shows how the partial molar adiabatic compressibility, KSI

of the oligoglycines depends in D20 (0) and H20 (m) on the number of peptide

bonds in the molecule. As discussed above, the chsrged termini do not interact

with each other in triglycine and longer homologues. Consequently, the

compressibility contribution of an independently hydrated -CH2COND- group can

be determined as the increment of the partial molar adiabatic compressibility,

Ks, per glycol unit in trig lycine, tetrag lycine, and pentag lycine. The estimated

value of K(CH2COND) in D20 is equal to (-2.210.5)xI cm3mol-'bar-' which is

somewhat smaller (more negative) than (-1 .l t0.5)xq o4 ~ r n ~ r n o l - ~ bar-' , a value of

K(CH2CONH) in H20 [28]. The compressibility contribution of the peptide group,

K(COND), in D20 can be calculated by subtracting K(CH2) from K(CH2COND):

K(C0ND) = - 2 . 2 ~ 1 0 ~ + 3 . 2 ~ 1 0 ~ = (1.0t1.0)~10~ cm3mol-'bai1. This value is

somewhat larger than ( 0 . 5 ~ . 8 ) ~ 1 0 ~ cm3mor1bar-', Our previous estimate for the

compressibility contribution of the peptide group, K(CONH), in H20 [28].

Unfortunately, the large experimentaVcomputational uncertainty does not allow

us to detemine whether the observed differential compressibility contribution of

the peptide group is statistically significant.

Fig. 8 shows the dependence of the difference in partial molar adiabatic

compressibility, Ars, of the oligoglycines in D20 and H20 on the number of

peptide bonds in the molecule. In fact, the values of A K ' ~ represent the adiabatic

compressibility change of transferring the oligoglycines from Hz0 to D20.

Inspection of Fig. 8 reveals a pronounced break at the point corresponding to

triglycine. In the context of the foregoing discussion, this break is related to the

fact that, in triglycine and longer homologues, the charged temini cease to

interact with each other. As a first approximation, the value of AKs

corresponding to triglycine, -(3.4t1 .O) cm3mol*'bai', can be cunsidered to be the

difference between the compressibility contributions of an independently

hydrated pair of the oppositely charged amino and carboxyl groups in D20 and

H20. In H20, the compressibility contribution of an independently hydrated pair of

the charged amino and carboxyl groups has been estimated to be -34.0 cm3mol-

1 bai ' [39]. Consequently, the compressibility contribution of the independently

hydrated amino and carboxyl termini in D20 can be calculated to be -34.0 - 3.4 =

-(37.411 .O) crn3mol-'bar-'. Note that charged groups in D20 cause a decrease in

solvent compressibility. which is by about 10 % larger than in H20.

5. Concluding Remarks

This paper presents the first study of the hydration properties of amino acids

and oligoglycines in D20. Specifically. the partial rnolar volumes, VO, and partial

molar adiabatic cornpressibilities, Ks, of 5 amino acids and 5 oligoglycines have

been determined in D20 at 25 OC. The resulting data have been used to estimate

the volume and compressibiliKy contributions of the component nonpolar

jmethylene group). polar (peptide group), and charged (oppositely charged

amino and carboxyl temini) chernical groups.

The volume contribution of a nonpolar methylene group, V(CH2), in 9 0 is

equal to 16.7m.3 crn3mof1. This value is slightly higher than 15.9k0.4 cm3mor1,

the volume contribution of a methylene group in H20. The compressibility

contribution of a methylene group. K(CH2), in D20 is equal to -(3.2*0.4)x104

cm3rnol'"bar-' which is significantly smaller (more negative) than -1 .9x104

crn3mol~'bai', the value of K(CH2) in H20.

The volume contribution of a polar peptide group, V(COND), in D20 is equal

to 23.1 tl.l cm3rnol-'. This value is 1.5 cm3rnol-' larger than 21.6 cm3mol", the

volume contribution, V(CONH), of a peptide group in H20. This difference

suggests that hydrogen bonding between polar groups of a solute and solvent

molecules brings about a slightly weaker contraction of water in D20 than in H20.

The compressibility contribution of the peptide group. K(C0ND). in D20 has been

calculated to be (1 .OI1.0)x1 O+ cm3mol-' bafl which is somewhat higher than

(0.5f0.8)xI oa cm3mol-'bar-', the compressibility contribution, K(CONH), of the

peptide group in H20.

The electrostriction of an independently hydrated pair of amino and carboxyl

termini in D20 practically coincides with the corresponding value in H20 and has

been estimated to be -26k0.8 cm3mol-'. The compressibility contribution of an

independently hydrated pair of amino and carboxyl termini in D20 is equal to -

(37.4~1.0)xl o4 cm3mol-'bai1 which is 10 % smaller (more negative) than the

corresponding value in HzO.

In the aggregate, the volume and compressibility contributions of charged,

polar, and non polar groups in D20 are "measurably" distinct from those in H20.

This distinction, in principle, may allow one to apply differential volurnetric

measurements to protein solutions in 9 0 and H20 to gain insight into the nature

of the solvent exposed protein groups in the absence of structural information. A

prerequisite for such applications of volumetric measurements is the

accumulation of an empirical data base large enough to permit reliable

interpretation of the differential D201Hfl volumetric results in terrns of solvent-

exposed atomic groups. To achieve this goal, more systematic investigations of

low molecular weight model compounds and proteins including temperature-

dependent studies should be conducted. Such work is in progress.

Finally, this work raises fundamental questions about the molecular nature of

the differential hydration properties of charged, polar, and nonpolar groups in

020 and H20. Further experimental and theoretical studies are required to

answer these questions.

Acknowledgement. The authors would like to thank Mr. Andras Nagy for his

expert technical assistance in developing and assembling pieces of the

experimental set up, Ms. Lilit Chalikian for performing some of the density and

sound velocity measurements at the initial stages of this work, and Dr. Jens

Volker for many stimulating discussions and critical suggestions. This work was

supported by an operating grant from the Natural Sciences and Engineering

Research Council of Canada.

References

(1) Kauzmann, W. Adv. Protein Chem. 1959, 74, 1-63.

(2) Rashin, A. A. Prog. Biophys. Molec. Biol. 1993, 60, 73-200.

(3) Ladbury, J. E. Chernistry and Biololgy 1996, 3, 973-980.

(4) Pethig, R. Annu. Rev. Phys. Chem. 1992, 43, 177-205.

(5) Belton, P. S. Prog. Biophys. Molec. Biol. 1994, 67, 61 -79.

(6) Makhatadze, G. 1.; Privalov, P. L. Adv. Protein Chem. 1995, 47, 307-

425.

(7) Levy, R. M.; Gallicchio, E. Annu. Rev. Phys. Chem. 1998, 49, 531-567.

(8) Zamyatnin, A. A. Annu. Rev. Biophys. Bioeng. 1984, 13. 145-1 65.

(9) Sarvazyan, A. P. Annu. Rev. Biophys. Biophys. Chem. 1991, 20,321 -

341.

(10) Chalikian T. V.; Sarvazyan A. P.; Breslauer K. J. Biophys. Chem.

1994, 57,89-109.

(1 1 ) Gekko, K.; Hasegawa Y . Biochemistry 1986, 73, 145-1 65.

(12) Iqbai, M.; Verrall, R. E. J. Phys. Chem. 1987, 91, 967-971.

(13) Kharakoz, D. P.; Sarvazyan A. P . Biopolymers 1993, 33, 11-26.

(14) Nolting, B.; Sligar, S. G. Biochemisby 1993, 32, 12319-12323.

(15) Tamura, Y.; Gekko, K. Biochernistry 1995, 34, 1878-1 884.

(16) Chalikian, T. V.; Gindikin, V. S.; Breslauer, K. J. J. Mol. Biol. 1995,

250,291 -306.

(17) Chalikian, T. V., Gindikin, V. S.; Breslauer, K. J. FASEB J. 1996, 10,

164-1 70.

(18) Chalikian, T. V., Totrov, M., Abagyan, R.; Breslauer, K. J. J. Mol. Biol.

1996,260, 588-603.

(19) Chalikian, T. V.; Volker J.; Anafi, D.; Breslauer, K. J. J. Mol. Biol.

1997, 274, 237-252

(20) Chalikian, T. V.; Breslauer, K. J. Proc. Natl. Acad. Sci USA 1996, 93,

1012-1014.

(21) Chalikian, T. V.; Breslauer, K. J. Biopolymers 1996, 39,619-626.

(22) Millero, F. J.; Lo Surdo, A.; Shin, C. J. Phys. Chem. 1978, 82, 784-

792.

(23) Mishra, A. K.; Ahluwalia, J. C. J. Phys. Chem. 1984, 88, 86-92.

(24) Kharakoz, D. P. Biophys. Chem. 1989, 34, 115-125.

(25) Kharakoz. D. P. Phys. Chem. 1991, 95,5634-5642.

(26) Iqbal, M.; Verrall, R. E. J. Phys. Chem. 1987, 91, 967-971.

(27) Hedwig, G. R.; Hoiland, H. J. Chem. Themodynamics 1991,23,

1029-1 035.

(28) Chalikian, T. V.; Sarvazyan, A. P.; Funck, Th.; Breslauer, K. J.

Biopolymers 1 994, 34, 54 1 -553.

(29) Millero, F. J. In Water and Aqueous Solutions, Home, R. A., Ed.; New

York, John Wiley & Sons, Inc.: 1972, pp. 51 9-595.

(30) Kell, G. S. In Water. A Comprehensive Treatmeni; Franks, F., Ed.;

New York, London, Plenum Press: 1972; vol. 1. pp. 363-41 2.

(31) Barnatt, S. J. Chem. Phys. 1952, 20,278-279.

(32) Owen, B.; Simons, H. L. J. Phys. Chem. 1957, 61,479482.

(33) Conde, O.; Teixeira, J.; Papon, P. J. Chem. Phys. 1982, 76,3747-

3753.

(34) Sarvazyan, A. P. Uhasonics 1 982, 20, 1 51 -1 54.

(35) Eggers, F.; Funck, Th. Rev. Sci. lnstrvm. 1973, 44, 969-978.

(36) Eggers, F. Acustica 1992, 76,231-240.

(37) Sarvazyan, A. P.; Selkov, E. E.; Chalikian, T. V. Sov. Phys. Acoust.

1988, 34, 631-634.

(38) Sarvazyan, A. P.; Chalikian, T. V . Ultrasonics 1991, 29, 11 9-1 24.

(39) Chalikian, T. V.; Sarvazyan, A. P.; Breslauer, K. J. J. Phys. Chem.

1993, 97, 1301 7-1 3026.

[40] Pierotti, R. A. J. Phys. Chem. 1965, 69, 281-288.

[41] Stillinger, F. H. J. Solution Chem. 1973, 2, 141 -1 58.

[42] Pierotti, R. A. Chem. Rev. 1976, 76, il 7-726.

[43] Kharakoz, D. P. J. Sohtion Chem. 1992, 27.569495.

[44] Reiss, H. Adv. Chem. Phys. 1965, 9, 1-84.

[45] Edward, J. T.; Farrell, P. G. Cm. J. Chem. 1975, 53, 2965-2970.

[46] Philip, P. R.; Jolicouer C. J. Solution Chem. 19f5, 4, 105-125.

[47] Desrosiers, N.; Lucas, M. J. Phys. Chem. 1974, 78, 2367-2369.

(481 Bondi, A. JO Phys. Chem. 1964, 68.441451.

[49] Paci E.; Marchi M. P m . Nafl. Acad. Sci. USA 1996. 93.1 1609-11614.

[50] Paci E.; Velikson B. Biopolymers 1997, 47,785797.

[51] Murphy L.R.; Matubayasi N.; Payne V.A.; Levy R.M. Folding and

Design 1998, 3,105-1 18.

Table 1

The molecular weights. M, relative molar sound velocity incrernents, [U], partial

molar volumes, VO. and partial molar adiabatic compressibilities, Ks. of the

amino acids in D20 at 25 OC.

M [u] VO K O ~

Da ~ r n ~ r n o l - ~ crn3rnol-l 1 o4 crn3molo1 bar''

Glycine 78.09 39.3H.3 42.9k0.4 -29.3*0.6

(43 .2)a (-26 .6)b

Alanine 92.1 1 48.3I0.3 60.2I0.4 -27.510.6

(60.4)= -(25.1 lb

Aminobutiric 106.12 63.0k0.3 77.2k0.4 -31.340.6

Acid (75. 5)a(-27. 1 lb

Nowaiine 120.12 76.8k0.3 93.210 .4 -35.1 a . 6

(91 .7)a (-29.0)~

---

Norleucine 1 34.22 89.7&0,3 1 10.3'0.4 -37.2kO .6

(1 07.9)a (-31 .llb

a the partial molar volumes, V", of the amino acids in H20 at 25 O C are from ref.

P31-

the partial molar adiabatic compressibilities, rs, of the amino acids in H20 at

25 OC are from ref. (251.

Table 2

The molecular weights, Ml relative molar sound velocity increments, [U], partial

molar volumes, VO, and partial molar adiabatic compressibilities, KoS, of the

oligoglycines in D20 at 25 OC.

M [u] VO K O ~

Glycine 78.09 39.3k0.3 42.9k0.4 -29.310.6

(43.2)= (-26 .6)a

Diglycine 136.1 62.2k0.3 76.610.4 -43.6k0.6

(76.2)a (-40.8)a

Triglycine 194.2 76.420.6 11 1.940.5 -48.4'1 .O

(1 12.1)a (-45.0)a

Tetraglycine 252.2 92.1f0.7 152.Ok0.6 -50.1 21 -2

(149.6)= (-45.

Pentaglycine 310.2 108.2t1 .O 191.5k1.0 -52.9k1.8

(1 86.9)a (-47.2)"

a the partial molar volumes, VO, and partial molar adiabatic compressibilities, KOs,

of the oligoglycines in H20 at 25 O C are from ref. [28].

Figure Legends

Figure 1

The thickness, A, of the thermal volume as a function of the diameter of a solute.

dZ, in D20 and H20.

Figure 2

(a) The dependence of the partial molar volume, Vc, of the cavity enclosing a

spherical solute on its diameter, dZ, in D20 (A) and H20 (V); (b) The difference,

AVc, in the partial molar volume, Vc, of the cavity enclosing a spherical solute in

D20 and H20 as a function of Vc.

Figure 3

The difference between the partial molar volume, VO, and the van der Waals

volume, VW, for the amino acids as a function of the nurnber of carbon atoms in

the side chain.

Figure 4

The difference between the partial molar volume, V', and the van der Waals

volume, VW, for the oligoglycines as a function of the number of peptide bonds.

Figure 5

The difference, AVO, between the partial molar volumes, VO, of the oligoglycines

in D20 and H20 as a function of the number of peptide bonds. The partial molar

volumes, VO, of the oligoglycines in H20 are fkom ref. [28].

Figure 6

The partial molar adiabatic compressibil*Q, Kos, of the amino acids in D20 (0)

and H20 (O) as a function of the number of carbon atoms in the side chain. The

partial molar adiabatic compressibilities, KoS, of the amino acids in H20 are from

ref. [%].

Figure 7

The partial molar adiabatic compressibility, KoS, of the oligoglycines in D20 (0)

and H20 ( 0 ) as a function of the number of peptide bonds. The partial molar

adiabatic compressibilities, Vs, of the oligoglycines in H20 are from ref. [28].

Figure 8

The difference, AKoç, between the partial molar volumes, Kç, of the oligoglycines

in D20 and H20 as a function of the nurnber of peptide bonds. The partial molar

adiabatic compressibilities, Ks, of the oligoglycines in H20 are from ref. [28].

Figure 1. The sickness of the thermal volum as a function of the diameter of a solute, d,, in D O and H O 2 2

Figure 2(a). The dependence of the partial molar volume, 1 V C , of the cavity enclosing a spherical solute as a function 1

of its diameter, d, in D,O and H,O

Figure 4. The difference between the partial molar volume, V", and the van der Waals volume, V , as a function of the number of peptidebonds

O 1 2 3 4 5 Number of Peptide Bonds

Figure 5. The difference between the partial molar volumes of the oligoglycines

in D,0 and H,O s a function of the number of peptide bonds.

I I I

G ~ Y Ala Abu

1

Nva

I

Nle

Number of Carbon Atoms in Side Chains

Figure 6. The partial molar adiabatic compressibility, K,of the amino

acids in D,O and H,O as a function of carbon atoms in the side chain

Digly Trigly Tetragly Pentagly

1 2 3 4

Number of Peptide Bonds

Figure 7. The partial molar adiabatic compressibility, Kso, of the

oligoglycines in D,O and H,O as a function of the number

of the peptide bonds

Chapfer 3. Differen tial Hydration of a,o-Aminocarboxylic

Acids in DzO and HzO

Reproduced with permission from

Journal of the Amencan Chernical Society, in press.

Unpublished work copyright 2000. American Chemicai Society.

Olga Likhodi and Tigran V. Chalikian

Differential Hydration of a,o-Aminocarboxylic Acids in D20 and H20.

Dr. Tigran Chalikian assisted me in some of the calculaiions and experiments.

Abstract

We report the relative molar sound velocity increments, [U], partial molar

volumes, VO, expansibilities, EO, and adiabatic cornpressibilities, KoS, for a

homologous series of eight a,o-aminocarboxylic acids in D20 solution within the

temperature range of 18 to 55 OC. We use the resulting data to estirnate the

volume, expansibility, and adiabatic compressibility contributions of the

comportent aliphatic (methylene groups) and charged (oppositely charged amino

and carboxyl termini) chernical groups. We compare these group contributions

with similar group contributions for the same set of a,o-aminocarboxylic acids in

Hz0 (Chalikian. T. V.; Sarvazyan, A. P.; Breslauer. K. J. J. Phys. Chem. 1993,

97, 1301 7-13026). We use these data to characterize quantitatively the

differential hydration properties of charged and hydrophobic groups in D20 and

H20, Taken together, our results suggest that the hydration properties of

hydrophobic and charged groups in D20, as reflected in their volume,

expansibility, and compressibility contributions are measurably distinct from

those in H20. Significantly, these volumetric characteristics of the solute

hydration differ not only in their absolute values but also in their temperature

dependences. Such characteristics should prove useful in developing a better

understanding of the role of differential DPOlH20 hydration in modulating thermal

and therrnodynamic stability of proteins. In addition, these results represent a

further step in building up an empirical datebase of differential volumetric

parameters of protein functional groups in D20 and H20. Such a database is

required for develo ping a methodology in which differential volumetric

measurements in D20 and H20 can be employed to gain insight into the arnount

and chemical nature of solvent-exposed protein groups in the absence of

structural information.

Introduction

Investigations of solvent-induced perturbations of protein stability provide one

approach for studying the role of hydration in dictating the conformational

preferences of a polypeptide chain at given experimental conditions. In this

respect, differential studies of protein stability in light (H20) and heavy (D20)

water are especially promising and have recently attracted considerable attention

[1 - 41. For al1 proteins studied so far, the thermal stability (TM) increases in the

presence of D20 [ M l . However, the thermodynamic stability (AG") in D20 may

either slightly decrease or slightly increase relative to H20 as a result of entropy-

enthalpy compensation that has been attributed to the differential hydration

properties of protein groups in DzO and H20 (1, 21.

It should be noted that H20 and D20 represent an ideal pair for studying

solvent-induced perturbations of protein stability, since they are chemically

identical yet their physical properties differ significantly [5]. The differences

between physical properties of H20 and D20 have been traditionally explained by

the differential energetics of intermolecular hydrogen bonds. The lengths of the

hydrogen bonds in 020 (2.766 A) and H20 (2.765 A) are essentially the same,

while the energy of a hydrogen bond in D20 is 0.24 kcaVmol (-5 %) higher than

in Hz0 [5]. In addition, a larger degree of structural order exists in D20, although

this order breaks down faster with increasing temperature [5]. These and other

differences in physical properties of light and heavy water render them unequal

as solvents. Consequently, the thenodynamics of solvation of various atomic

groups in light and heavy water is considerably different, which is reflected in

corresponding changes in Gibbs free energy, enthalpy, entropy. heat capacity,

volume, compressibility, and other thermodynamic characteristics upon transfer

of various substances frorn H20 to D20 [6 - 101.

Volumetric measurements have proven useful for the quantitative

characterization of hydration properties of biopolymers (proteins and nucleic

acids) as well as their low molecular weight analogs [Il-151. Differential D20

versus H20 volumetric studies, performed on a systematic basis, will help us gain

insight into the hydration properties of charged, polar, and nonpolar atomic

groups. This knowledge is important for a better understanding of the role of

water in determining the structural characteristics of proteins and nucleic acids as

well as origins of increased thermal stability of proteins in heavy water. In

addition, as discussed in Our previous work, if the volumetric contributions of

charged. polar, and nonpolar groups in D20 are measurably different from those

in H20, differential volumetric measurements in light and heavy water may offer

one potential way to derive information on the amount and chemical nature of

solvent-exposed protein groups in the absence of structural data [9].

In our previous work, we have ernployed volumetric measurements to

investigate the hydration properties of charged (oppositely charged amino and

carboxyl temini), polar (a peptide group), and nonpolar (a methylene group)

groups of a-amino acids and oligoglycines in D20 at a single temperature of 25

OC [9]. At room temperature, judging by the volume and compressibility

contributions, the hydration properties of these groups in D20 are slightly yet

"measura blyn distinct from those in H20. However, since solute-solvent

interactions are strongly temperature dependent, the information content of such

singletemperature studies is rather limited. Clearly, further temperature-

dependent investigations are required to better understand the difFerential

hydration properties of various atomic groups in light and heavy water. In

recognition of this need, we now expand Our studies and report on the partial

molar volume, VO, expansibility, EO, and adiabatic compressibility, Kas, of a

homologous series of a,o-aminocarboxylic acids in D20 over the temperature

range 18 to 55 OC. At neutral pH, a,a-aminocarboxylic acids are zwitterions,

consisting of oppositely charged amino -ND3' and carboxyl -CO0 teminal

groups separated by an unbranched chain of rnethylene -CHT groups. Thus,

only charged and nonpolar aliphatic groups contribute to the measured

volumetric parameters of these solutes. We find that the dependences of VO, EO,

and Kos of the a,o-aminocarboxylic acids on the nurnber of their constituent -

CH2- groups are qualitatively similar to the same dependences previously

reported for the same set of a,w-aminocarboxylic acids in H20 [Il]. However, on

a quantitative level, the VO, E", and Ks contributions per -CH2- or a pair of

charged groups in D20 are considerably different from those in H20.

Significantly, these volumetric characteristics of the solute hydration differ not

only in their absolute values but also in their temperature dependences. In

general, we compare the volumetric results of this work with Our previous data on

the same set of a,o-aminocarboxylic acids in H20 [Il]. We interpret the

combined set of experimental data in terms of differential hydration properties of

charged termini and aliphatic -CHr groups in H20 and D20. We also discuss the

implications of our results for developing an understanding of the role of solvent

in rnodulating conformational stability of proteins.

Materials and Methods

The a,o-arninocarboxylic acids and D20 (99.9 %) used in our studies were

purchased from Sigma-Aldrich Canada, Ltd. (Oakville, Ontario, Canada). The

a,o-arninocarboxylic acids were of the highest purity commercially available and

used without further purification. Prior to the densimetric and ultrasonic

velocimetric experiments, the a,o-aminocarboxylic acids were dissolved in D20

and lyophilized to exchange labile protons for deuterons. The concentrations of

the samples were determined by weighing 10-20 mg of solute material with a

precision f0.03 mg, and then dissolving the material in a known amount of heavy

water. Glycine, p-alanine, and 4-aminobutanoic acid were dried at 110 OC for 12

hours pnor to weighing. All other a,w-aminocarboxylic acids were dried under

vacuum in the presence of phosphorus pentoxide for 72 hours prior to weighing.

To prevent formation of air bubbles, al1 solutions were preheated in sealed

Eppendorf tubes to 5 OC above the measurement temperature before filling the

ultrasonic or densimetric cells.

Solution densities were measured using a vibrating tube densimeter (DMA-

601602, Anton Paar, Gratz, Austria) with a precision of I 1.5~10" gcmJ at 18, 25,

40 and 55 OC. The apparent molar volume, OV, was calculated from these density

values using the relationship [16]:

w = M 1 P - (P - PO) f POP^) (1

Where M is the solute molecular weig ht; m is the molal concentration of a solute;

and p and are the densities of the solution and solvent, respectively. The

requisite values for the density of D20, PO, were taken from Kell [IA. The values

of po are equal to 1.105599, I.lO4449, 1.099958, and 1.093251 gcm" at 18.25,

40, and 55 OC, respectively.

The solution sound velocities required to calculate the apparent molar

adiabatic compressibility, +Ks, of a solute were measured at 18, 25, 40 and 55 OC

using the resonator method [18-201 at a frequency of about 7.5 MHz. The

sample and reference resonator cells with minimum volumes of 0.8 cm3 were

thermostated with an accuracy of I 0.01 OC, and a previously described

differential technique was employed for al1 measurements [19]. Theoretical

analyses [21, 221 have shown that, for the type of ultrasonic cells used in Our

studies, the accuracy of the sound velocrty rneasurements is about t104 % at

frequencies between 6 and 8 MHz. The analyses of the frequency

characteristics of the resonator cells were perfomed by a Hewlett Packard Model

HP41 95A networklspectrum analyzer (Mississauga, Ontario. Canada).

The key characteristic of a solute directly derived from ultrasonic velocimetric

measurements is the relative molar sound velocity increment, [U]:

[UI = (U - Uo) 1 (UoC) (2)

Where U and UO are the sound velocities in solution and solvent, respectively;

and C is the molar concentration of a solute.

The apparent molar adiabatic compressibility, 4 Kç, was calculated from the

densimetric and ultrasonic data using the expression [23, 241:

+Ks = Pso (20V - 2[Ul - Mlpo) (3)

Where Pso is the coefficient of adiabatic wmpressibility of 40. The requisite

values of Pso were calculated from data on the density, po. [17] and sound

velocity, Uo, [25] of D20 using the expression Bso = (p&2)! At 18, 25.40, and

55 OC, the values of Uo are equal to 1382, 1400, 1435, and 1443 ms-',

respectively (251. The calculated values of Pso are equai to 47.36x106,

46.20~1 06, 44.1 5x106, and 43.93~1 bafl at 18,25,40. and 55 OC,

respectively.

For each evaluation of +V or $Kç, three to five independent measurements

were carried out within a concentration range of 2 - 4 mglrnL for each of the a,o-

aminocarboxylic acids, with the exception of 1 1 -aminoundecanoic acid. For this

long homologue, concentrations of 1 mg/mL or less were used due to its low

solubility.

Results

Tables 1, 2, and 3 show the relative molar sound velocity increments, [U],

apparent molar volumes, +V, and apparent molar adiabatic compressibilities,

bKsl of eight a,o-arninocarboxylic acids at 18, 25,40 and 55 OC, respectively.

Due to space limitations, we intentionally omit our primary experimental data on

solution densities, p, sound velocities, U. and solute concentrations required for

calculation of the apparent molar volumes. bV, and relative molar sound velocity

increments, [U], from eqs. (1) and (2), respectively. These primary data are

relatively uninfonative and, if needed, can be easily obtained from the values of

+V and [U] using eqs. (1) and (2). Errors reported represent maximum

uncertainties due to the concentration determination. temperature drifts, and

apparatus limitations. Previous studies have shown that. for the atm-

aminocarboxylic acids in H20, the apparent molar volumes and the apparent

molar adiabatic cornpressibilities do not depend strongly on concentration [26,

2T]. By extension. one rnay plausibly assume that, in D20 too. the concentration

dependences of +V and +Ks are weak. In other words, within experimental error,

the apparent molar volumes, +V, and adiabatic cornpressibilities, 4&, of the a,o-

aminocarboxylic acids determined in the concentration range of 1-4 rnglml can

be assumed to coincide with the partial molar volumes, VO, and adiabatic

compressibilities, Kos, obtained by extrapolation to infinite dilution. Therefore,

below, we do not discriminate between the apparent molar and partial molar

characteristics of the a.o-aminocarboxylic acids.

We have approxirnated Our measured temperature dependences of the partial

molar volumes, Vo, by second order polynomials. The temperature derivatives of

V0 then were determined analytically by differentiation of the approximating

functions at each of the temperatures studied. Table 4 presents the resulting

data as the temperature slopes of the partial molar volume [equal to the partial

molar expansibility, EO, since EO = (dVOlûT)p] at 18, 25,40, and 55 O C .

Discussion

Partial Molar Volume

The partial molar volume of a solute at infinite dilution, Vo, can be interpreted

in t e n s of hydration based on the following relationship:

V0 = VM + A V ~ = VM + nh(Vh - VO) (4)

where AVh is the volume effect of hydration, that is the solute-induced change in

the solvent volume; Vo and Vh are the partial molar volumes of water in the bulk

state and in the hydration shell of a solute, respectively; and nt, is the "hydration

numbef, that is, the number of water molecules in the hydration shell of a solute.

Scaled particle theory (SPT), originally formulated for a system of hard

spheres, has been subsequently extended with great success to description of

aqueous solvation of both polar and nonpolar solutes [28-331. Based on SPT

theory, the hydration contribution, AVh, in eq. (4) can be presented as the sum of

three terms [29-341:

AVh = VT + VI + BTORT (5)

where VM is the intrinsic volume of the solute molecule; Vr is the "thermal"

volume, which results from the mutual thermal motion of solute and solvent

molecules; VI is the interaction volume. which accounts for solvent contraction

under the influence of polar (hydrogen bonding) and charge (electrostriction)

groups of the solute; PT^ is the coefficient of isothermal cornpressibility of the

solvent; R is the universal gas constant; and T is the absolute temperature. The

ideal terni pmRT is small and does strongly depend on temperature. For D20,

the value of PsaRT is equal to 1.1 5 cm3rnol-' at 18 OC and increaçeç to 1.23

c r n 3 m d at 55 OC.

For low molecular weight substances. the value of VM c m be approximated by

the van der Waals volume, Vw. We have calculated Vw for the eight a,oW

aminocarboxylic acids based on the group contribution data of Bondi [35]. To

derive the interaction volume, VI, which represents the electrostriction of

oppositely charged amino and carboxyl termini of the am-aminocarboxylic acids.

one needs to estimate the thermal volume, VT. AS previously discussed, the

thermal volume, VT, is a linear function of Sw, the van der Waals surface area [9,

1 1, 34, 361:

VT=ASw+B (6)

where the coefficients A and 6 are the same for a homologous series of solutes.

The coefficient B in eq. (6) represents the volume of the cavity containing a

point particle of zero radius and c m be obtained readily from SPT theory [28-311.

For D20 and H20 based solutions, the value of B is 0.6 cm3mol" and practically

does not depend on temperature within the 18 to 55 OC range. The coefficient A

in eq. (6) can be determined using a previously described approach [ l l ] in which

the contribution of nonpolar groups to the interaction volume, VI, is assumed to

be negligible [ I l , 34, 361. With this assumption, inspection of eqs. (4) - (6)

reveals that any two homologues which are distinct with respect to the size of

their nonpolar moieties will have (VO - Vw) values which differ only by the

difference between the thermal volumes, VT. Hence, the coefficient A in eq. (6)

can be deterrnined as the slope of a A(VO - Vw) versus plot. Figure 1 shows

such a plot for the a,o-aminocarboxylic acids studied here.

Cornparison of data presented in Figure 1 with similar data obtained earlier for

the same set of a,o-aminocarboxylic acids in H20 [Il] reveals a great deal of

similarity. In particular, in both H20 and D20, the dependences of (Vu - Vw) on

Sw at each temperature exhibit slight yet noticeable breaks at the point

corresponding to 5-aminopentanoic acid. For the atm-aminocarboxylic acids in

HzO, we have rationalized these breaks by proposing that the character of

solute-solvent interactions in long and short homologues (before and beyond 5-

aminopentanoic acid) is qualitatively different [Il]. In the short homologues

(glycine, p-alanine, and 4-aminobutanoic acids), the hydration shell is

electrostatic in nature and is predominantly determined by the mutual interacting

oppositely charged terminal groups. For 5-aminopentanoic acid and longer

homologues, the charged terrnini stop interacting with each other and each

added -CHT Iink becomes independently hydrated. Thus, any incremental

change in the partial molar volume, VO, of the "long" atm-aminocarboxylic acids is

determined by the volume contribution of an independently hydrated -CHT

group, V(CH2).

Figure 2 presents the temperature dependences of V(CH2) in H20 (e) and

O20 (O). Inspection of Figure 2 reveals that the values of V(CH2) in the two

solvents practically coincide within k0.3 crn3rnol", although the temperature

dependence of V(CH2) may appear to be somewhat steeper in H20 relative to

020. In 90 at 25 O C , the volume contribution of an independently hydrated

methylene group, V(CHp), in u.o-aminocarboxylic acids equals 15.8 I 0.3

cm3mol-' which is somewhat lower than 16.7k0.3 cm3mol-'. our estimate of

V(CH2) in a-amino acids with nonbranched aliphatic side chain [9]. It is difficult to

assess if this discrepancy is statistically significant or to provide a reliable

explanation for its origins. At present, we just make notice of this fact for our

future investigations.

Since the interaction volume, VI, of aliphatic groups is negligible, an inmase

in the value of (VO - Vw) with increasing solvent accessible surface area, SW, of 5-

aminopentanoic acid and longer homologues can be ascribed solely to an

increase in the thermal volume, VT. AS is seen from Figure 1, the dependence

(VO - Vw) on Sw beyond 5-aminopentanoic acid is linear. As mentioned above,

the value of the coefficient A in eq. (6) can be evaluated as the dope, A(VO - VW) I

ASw, of this straight line. We estimate the coefficient A in eq. (6) equal to

3.98~10-', 4.1 1 XIO". 4.18~10-~, and 4.35~10.' cm at 18. 25,40. and 55 O C ,

respectively. These results are very close to our previous estimates of the

coefficient A for a.o-aminocarboxylic acids in H20: 3.87~1 O-'. 4.05~1 O-', 4.1 9x 10'

', and 4.54~10" cm at 18, 25. 40. and 55 OC. respectively [ I l ] . This similarity

suggests that the thermal volumes, VTi for the a,o-aminocarboxylic acids in 40

do not differ strongly from those in H20. This conclusion is in agreement with our

SPT-based calculations for the partial molar volume, Vc, of cavities enclosing

sphen'cal solutes in D20 and H20 [9]. Recall that, at 25 OC, the cavity volumes,

Vc, for solutes with diameters up to 10 A were estimated to be similar in D20 and

in H20 (the difference was on the order of 1 cm3mol") [9]. In the present study,

we repeat these calculations for l8,4O, and 55 OC and find that, at these

temperatures too, the cavity volumes are similar to within tl cm3mor1. Note that

the cavity volume, Vc, represents the intnnsic volume, VM, of a solute plus its

thermal volume, VT [34, 361. Importantly, since the intrinsic volume. VM, of a

solute in D20 is equal to that in H20, the differential value of AVc is detemined

solely by the differential value of VT. Thus, within the 18 to 55-T-temperature

range, the thermal volume, VT, for solutes approximately of the size of the eight

a,o-aminocarboxylic acids studied here are very close in D20 and H20.

Armed with the value of A, we can now use eqs. (5) and (6) to calculate the

interaction volumes, VI, for the a,o-aminocarboxylic acids. The resulting data are

plotted in Figure 3 against the number of -CHT groups. Inspection of Figure 3

reveals that the interaction volume, VI, decreases (becomes more negative)

going frorn glycine to 5-aminopentanoic acid and becomes constant for the

longer a.o-aminocarboxylic acids, when four or more -CH2- groups separate the

charged end groups. As shown in Figure 3, the value of VI at 55 OC is somewhat

more negative than at lower temperatures. We have made a similar observation

of VI for the a,o-aminocarboxylic acids in H20 [Il]. These observations simply

reflect nonlinear, parabolic (concave downwards) temperature dependences of VI

for charged groups in H20 and D20. The average value of VI for the "longer"

a.o-arninocarboxylic acids can be taken as a measure of the total electrostriction

of independently hydrated ND; and COO' groups.

Table 5 compares the values of VI for the independently hydrated arnino and

carboxyl temini of the a,o-aminocarboxylic acids in D20 with those in H20.

Figure 4 presents the difference between the interaction volumes, VI, of the

independently hydrated amino and carboxyl temini of the a,o-aminocarboxylic

acids in D20 and H20 as a function of temperature. It is ternpting to separate

individual volumetnc contributions of the amino and carboxyl temini. It should be

noted that there are some indications in the literature that the positively charged

amino group may be solvated significantly less than the negatively charged

carboxyl group (e. g.. see references 37, 38). However, any attempt to separate

the hydration contributions of the amino and carboxyl termini would be highly

speculative and require the use of non-thermodynamic assurnption(s).

Therefore. in our analysis below, we treat the pair of charged termini of a,o-

arninocarboxylic acid as a single thermod ynamic entity .

Inspection of Figure 4 reveals that the electrostriction of oppositely charged

amino and carboxyl groups is stronger (more negative) in D20 at low

temperatures. However, above 35 O C , the situation is opposite: the

electrostriction of charged groups becomes stronger in HzO. For example, at 18

OC, NI is negative and equals -1 -021 .O cm3mor' (or -5 % of VI), while, at 55 O C ,

AV, becornes positive and equals 2.7I1 .O cm3mol-' (or 10 % of VI). Note that. at

25 OC. AVi is small and equals -0.4t1 .O cm3mol". This result is in qualitative

agreement with our previous evaluation of the interaction volume ,VI, of glycine

[9]. Specifically, we found that. at 25 OC. the interaction volumes, VI. of the

charged temini of glycine in DzO and H20 coincide within M.5 cm3mol" [9].

Partial Molar Expansibility

Differentiating eq. (4) with respect to temperature and assuming that the

hydration number, nt,, does not strongly depend on temperature, one obtains the

following relationship for the partial molar expansibility:

EO = EM + AEh = Eu + nh(Eh - Eo) (7)

where EM is the intrinsic expansibility of a solute molecule; rlEh is the

expansibility effect of hydration; Eo and Eh are the partial molar expansibilities of

water in the bulk state and in the hydration shell of a solute, respectively.

For small molecules, such as a,o-aminocarboxylic acids, the intrinsic

expansibility, Eu, in eq. (7) is small and can be neglected. Consequently. only

the hydration changes contribute to the partial molar expansibility, EO, of low

molecular weight substances: EO = AEh.

Figure 5 shows the dependence on the number of methylene groups of the

partial molar expansibility, EO, of the a,w-aminocarûoxylic acids at 18, 25, 40, and

55 OC. Even though the experimental errors in EO are relatively high, pronounced

breaks can be observed at a point corresponding to 5-arninopentanoic acid. It

should be noted that similar breakpoints have been observed for the dependence

of the partial molar expansibility, EO, of the a,w-aminocarboxylic acids on the

number of methylene groups in HzO [Il]. Anabgous to volume, the breaks

observed in Figure 5 suggest that the nature of hydration of the short a,o-

aminocarboxylic acids qualitatively differs from that of the longer homologues.

The incremental change in the partial molar expansibility, EO, of 5-

aminopentanoic acid and the longer homologues represents the expansibility

contribution per independently hydrated -CHr group, €(CH2). Figure 6 shows

how the value of E(CH2) depends on ternperature in H20 (a) (calculated frorn Our

previous data presented in ref. I l ) and D20 (O). Inspection of Figure 6 reveals

that the ternperature dependences of E(CH2) in H20 and D20 are significantly

different. In H20, the value of E(CH2) is positive and linearly increases with

temperature from 0.01 3k0.003 at 18 O C to 0.026M.003 at 55 O C . By contrast. in

D20, the value of €(CH2) is still positive but lineariy decreases with temperature

from 0.01310.005 at 18 OC to 0.008kû.002 at 55 OC. This striking discrepancy

reflects the differential hydration of aliphatic groups in H20 and D20. Based on

eq. (7), the molar expansibility of water solvating aliphatic groups, Eh, is greater

than that of bulk solvent. Eo. However, in H20, Eh increases faster with

temperature Vian Eo, while. in D20, Eo increases faster than Eh.

The partial molar expansibiiity, EO, of 5-aminopentanoic acid can be assumed

to be roughly equal to the contribution of noninteracting amino and carboxyl

groups. Based on this assumption. the differential expansibility contribution of

charged amino and carboxyl groups in D20 and H20 can be obtained by

comparing the partial molar expansibility, EO, of 5-aminopentanoic acid in light

and heavy water. Figure 7 shows how the expansibility contributions of charged

amino and carboxyl groups depends on ternperature in H20 [E(NH3' + COO')]

(@) and D20 [E(ND~' + C00-)] (0). Inspection of Figure 7 reveals that the

expansibility contributions of the charged temini E(NH3' + COO-) and E(ND; +

COO') in both solvents are positive at low temperatures but become negative at

higher ternperatures: at -40 OC in H20 and -55 "C in D20. Based on eq. (7). this

observation suggests that, at low temperatures, the molar expansibility of water

solvating charged groups, Eh, is larger than that of bulk H20 or D20. However, at

higher ternperatures, Eo becomes larger than Eh. Further inspection of Figure 7

reveals that, in H20, the value of E(NH3 + COU) diminishes with temperature

faster than E(ND3* + CO03 in D20. As a result, at low temperatures, the

expansibility contribution of charged groups is smaller in D20, while at high

temperatures (above -35 O C ) , the contribution of charged groups in D20

becomes larger than that in H20.

Partial Molar Adiabatic Compressibility

Analogous to volume and expansibility, the partial molar adiabatic

compressibility, Kos, of a solute can be described as the sum of intrinsic and

hydration contributions:

Kos = KM + A& = KM + nh(& - KO) (8)

where KM is the intrhsic compressibility of a solute molecule; AKh is the

compressibility effect of hydration; & and Kh are the partial molar adiabatic

compressibilities of water in the bulk state and in the hydration shell of a solute,

respectively; and nh, the hydration number, has the same meaning as in eqs. (4)

and (7).

For low molecular weight substances, the intrinsic compressibility, KM, is

predominantly detemined by the srnall compressibility of covalent bonds and

extemal electron shells, and, therefore, can be neglected [12-151. Hence, only

the hydration changes contribute to the partial molar adiabatic compressibility of

low molecular weight substances:

Koç = A & = nh(k - KQ) (9)

Figure 8 shows the dependences on the nurnber of -CHT groups of the partial

molar adiabatic compressibilities, Kç, of the a,w-aminocarboxylic acids.

Compared to volume and expansibility (see Figures 1 and 5), the dependence of

the partial molar adiabatic compressibility, Kos, on the number of rnethylene

groups is more cornplex, as reflected in Figure 8. Inspection of Figure 8 reveals

that, at 40 and 55 OC, the interaction between the termini continues to influence

the value of Kos up to a point corresponding to 6-aminohexanoic acid when the

charged groups become separated by five -CHT links. This observation

indicates that compressibility may be a more sensitive parameter with respect to

subtle features of solute solvation as compared to volume and expansibility. We

have made similar observation earlier for the partial molar adiabatic

compressibility, KOs, of the a,o-aminocarboxylic acids in H20 [ll]. Note that the

intercharge distance in 6-aminohexanoic acid in H20 determined from dielectric

constant rneasurements is between 6.3 and 7.1 A [371, which corresponds to 2 to

2.5 diameters of water molecule. Thus, based on our compressibility data, the

hydration shell of the charged group in a,o-aminocarboxylic acids in D20 roughly

involves 1 -1.5 effective layers of water molecules, which coincides with earlier

estimations of the "thicknessn of hydration shell of charged groups in H20 [l 1 -1 4,

381.

Figure 9 (panel a) shows the temperature dependences of the compressibility

contributions of independently hydrated -CHT groups, &(CH2), of the a,o-

aminocarboxylic acids in H20 (a) (from ref. 11) and D20 (O). Inspection of

Figure 9a reveals that, in both solvents, the value of Kç(CH2) is negative at low

temperatures but becomes positive roughly above 35 OC. Based on eq. (9), this

observation suggests that, at low temperatures. water solvating aliphatic groups

is less compressible than bulk H20 or D20 but becomes more compressible at

higher ternperatures. Note that, at 25 OC, the compressibility contribution of an

independently hydrated methyiene group, Ks(CH2), in the a,o-aminocarboxylic

acid in D20 equals -(2.810.3)x104 cm3rnol-'bai' which is in good agreement with

-(3.&0.4)xl o4 cm3mol-'bar-', Our previous estimate for Kç(CH2) in a-amino acids

with nonbranched aliphatic side chains [9]. Further inspection of Figure 9a

reveals that, below 40 OC, the value of KS(CH2) in D20 is smaller than that in

H20. However, at higher temperatures, the value of Kç(CH2) in 90 becornes

larger than in HzO. This trend is more clearly shown in Figure 9 (panel b) which

depicts the temperature dependence of the difference between the

compressibility contributions of the independently hydrated rnethylene groups,

A&(CH2), in D20 and H20. In fact, the data shown in Figure 9b represent the

temperature dependence of the adiabatic cornpressibility change accompanying

the transfer of an independently hydrated methylene group from H20 to D20.

Analogous to volume and expansibility, the compressibility contribution of the

independently hyd rated amino and carboxyl terminal groups, &(ND3' + COO').

can be considered to be roughly equal to the partial molar adiabatic

cornpressibility, Kos, of 5-aminopentanoic acid. Figure 10 (panel a) shows the

temperature dependences of Ks(NH3+ + COO') in H20 (a) (from ref. 1 1 ) and

Ks(ND; + COO-) in D20 (O). Inspection of Figure 10a reveals that, in both

solvents, the compressibility contribution of charged groups is negative within the

whole temperature range studied although R becomes less negative with

temperature increasing. In agreement with converitionai wisdom, this

observation suggests that, within the entire temperature range studied, water

dipoles experiencing strong electrostatic influence of charged groups exhibit

reduced compressibility relative to buik H20 or D20. Further inspection of Figure

10a reveals that, over the whole temperature range studied, the value of Ks(ND3'

+ CO03 in D20 is smaller (more negative) than the value of Ks(NH3+ + COO') in

H20. Figure 10 (panel b) depicts the temperature dependence of the difference

between the compressibility contributions of an independently hydrated pair of

amino and carboxyl termini in D20 and H20. It should be noted that the data

presented in Figure 1 Ob represent the temperature dependence of the

compressibility change accompanying the transfer of charged groups from H20

to D20. At 18 OC, the compressibility contribution of a pair of independently

hydrated charged amino and carboxyl groups in D20 is (6.6t1 .4)x104 cm3rnol-

1 bar*' [-15 % of K ~ ( N H ~ + + COO-)] smaller (more negative) than that in H20. At

55 OC, this difierence decreases to (1.911 .9)x4o4 cm3mof'bai' [-IO % of

Kç(NHB + C00-)].

Conclusion

We have determined the relative molar sound velocity increments, [U], partial

molar volumes, VO, expansibilities, EO, and adiabatic compressibilities, Kas, for

eight a,o-aminocarboxylic acids in D20 solution within the temperature range 18

to 55 OC. We have used the resulting data to estimate the volume, expansibility,

and adiabatic compressibility contributions of the component aliphatic (rnethylene

groups) and charged (oppositely charged amino and carboxyl termini) chernical

groups.

The volume contributions of an independently hydrated methylene group,

V(CH2), in D20 and H20 are sirniiar b within 10.3 crn3mol-', although the

temperature dependence of V(CH2) may be sornewhat steeper in Hz0 relative to

D20. In D20, the electrostnction, VI, of an independently hydrated pair of

charged amino and carboxyl terminal groups decreases (becomes more

negative) with temperature increasing. Specifically, the value of VI is equal to

-26.2N.5 crn3mol" at 18 OC and decreases to -27.510.5 crn3rnol-' at 55 OC.

Compared to H20, the electrostriction of the charged termini is stronger (more

negative) in 40 below 35 OC. At higher temperatures, the situation is opposite:

the solvent contraction in the vicinity of the charged termini is stronger in H20.

In D20, the expansibility contribution of rnethylene groups, E(CH2), linearly

decreases with temperature from 0.013I0.005 at 18 OC to 0.008k0.002 at 55 OC.

By contrast, in H20, €(CH2) increases with temperature from 0.013k0.003 at 18

OC to 0.02610.003 at 55 OC. This discrepancy refiects the fact that, in H20, the

molar expansibility of water solvaüng aliphatic groups, Eh increases faster with

temperature than the expansibiI*ity of the bulk solvent, Eo, while, in D20, Eo

increases faster than Eh. ln D201 the expansibility contributions of the charged

termini. E(ND3 + COO-). decreases from (9.713.5)~10" cm3mol-'K-' at 18 O C to

(0.4*3.5)~10'~ cm3mol-'K-' at 55 O C . Compared to H20, the expansibility

contribution of charged groups is smaller in D20 at low temperatures, while

above -35 OC, the contribution of charged groups in D20 becomes larger than

that in H20.

In D20. the compressibility contribution of an independently hydrated

methylene group, KS(CH2), increases from (-4.0k0.3)xl O* cm3mol~'bar*' at 18 OC

to (3.3i0.3)~10" cmJmol"bar" at 55 'C. Compared io H20, the value of &(GHz)

in D20 is smaller below 40 OC. However, at higher temperatures, the value of

&(CH2) in D20 becomes larger than in H20. In D20, the compressibility

contribution of charged groups, Ks(ND3' + COO'), is negative within the whole

temperature range studied although it becomes less negative with temperature

increasing: it increases from (-48k0.9)xl o4 crn3mof'ba~' at 18 OC to

(-24.M .2)x1 o4 cm3mof1bar-' at 55 OC. Over the whole temperature range

studied, the value of &(ND3' + COO') in D20 is smaller (more negative) than the

value of Ks(NH3' + COO-) in H20, but the difference diminishes when the

temperature increases.

Taken together, our data suggest that. in D20, the hydration properties of

hydrophobic and charged groups, as reflected in their volume. expansibility, and

compressibility contributions, are measurably distinct from those in H20.

Significantly, these volumetric characteristics of solute hydration differ not only in

their absolute values but also in their temperature dependences. These results

provide a detailed quantitative description of the hydration of nonpolar and

charged protein groups in D20. Such characteristics should prove useful in

developing a better understanding of the role that differential D20/Hfl hydration

of atomic groups plays in modulating thermal and themiodynamic stability of

proteins. In addition, these results represent one further step in building an

empirical database of differential volumetric parameters of protein functional

groups in D20 and H20. Such a database is required for a possible application

of differential volumetric measurements to protein solutions in D20 and H20 to

gain insighi into the amount and chernical nature of solvent-exposed protein

groups in the absence of structural information.

Acknowledgement. The authors would like to thank Mr. Andras Nagy for his

constant technical support in developing and assernbling pieces of the

experimental set up, Dr. Jens Volker for his critical comrnents and many

stimulating discussions, and Mr. Arno G. Siraki for bis careful reading of the

manuscript. This work was supported by operating gants from the Natural

Sciences and Engineering Research Council of Canada ( W C ) and Ontario

Research and Development Challenge Fund (TVC).

References

(1) Makhatadze, G. 1.; Clore, G. M.; Gronenbom, A. M. Nature Stnict. 8M

1995, 2, 852-855.

(2) Kuhlman, B.; Raleigh, D. P. Protein Sci. 1998, 7,2405-2412.

(3) Verheul, M.; Roefs, S. P. F. M.; de Kruf, K. G. FEBS Lett 1998,427,

273-276.

(4) Gough, C. A.; Bhatnagar, R. S. J. Biomol. Stnrct. Dyn. 4999, 77,481-

491.

(5) Nemethy, G.; Scheraga, H. A. J. Chem. Phys. 1964, 41, 680-689.

(6) Desrosiers, N.; Lucas, M. J. Phys. Chem. 1974, 78,2367-2369.

(7) Ben-Naim, A.; Marcus, Y. J. Chem. Phys. 1984, 87,2016-.

(8) Lopez, M. M.; Makhatadze, G. 1. Biophys. Chem. 1998, 74, 11 7-1 25.

(9) Likhodi, O.; Chalikian, T. V. J. Am. Chem. Soc. 1999, 127, 1 156-1 163.

(10) Desrosiers, N.; Lucas, M. J. Phys. Chem. 1974, 78, 236702369,

(1 1) Chalikian, T. V.; Sarvazyan, A. P.; Breslauer, K. J. J. Phys. Chem.

1993, 97, 13017-13026.

(1 2) Sarvazyan, A. P. Annu. Rev. Biophys. Biophys. Chem. 1 991,20,321-

341.

(1 3) Chalikian, T. V.; Sarvazyan, A. P.; Breslauer, K. J. Biophys. Chem.

1994, 57, 89-1 09.

(14) Chalikian, T. V.; Breslauer, K. J. Cura Opin. Struct Biol. 1999, 8, 657-

664.

(1 5) Chalikian, T. V.; Breslauer, K. J. Biopolymers (Nucl. Acid. Sci.) 1998,

48, 264-280.

(1 6) Millero, F. J. In Water and Aqueous Solutions, Home, R. A.. Ed.; New

York, John Wiley â Sons, Inc.: 1972, pp. 519-595.

(17) Kell, G. S. In Water. A Comprehensive Treatnent; Franks, F., Ed.; New

York, London, Plenum Press: 1972; vol. 1, pp. 363-412.

(18) Eggers, F.; Funck, Th. Rev. Sci. lnstrum. 1973, 44,969-978.

( 1 9) Sarvazyan, A. P. Ultrasonics 1982, 20, 151-154.

(20) Eggers, F. Acustica 1992, 76,231-240.

(21) Sarvazyan, A. P.; Selkov, E. E.; Chalikian, T. V. Sov. Phys. Acoust

1988, 34, 631-634.

(22) Sarvazyan, A. P.; Chalikian, T. V. Ultrasonics 1991, 29, 11 9-1 24.

(23) Barrtatt, S. J. Chem. Phys. 1952, 20, 278-279.

(24) Owen, B.; Simons, H. L. J. Phys. Chem. 1957, 67.479-482.

(25) Conde, O.; Teixeira, J.; Papon, P. J. Chem. Phys. 1982, 76, 3747-

3753.

(26) Cabani, S.; Conti, G.; Matteoli, E.; Tine, M. R. J. Chem. Soc., Faraday

Trans. 7 1981,77,2385-2394.

(27) Shahidi, F.; Farrell, P. G. J. Chem. Soc. Faraday Soc. 1. 1978. 74,858-

868.

(28) Reiss, H. Adv. Chem. Phys. 1965, 9, 1-84.

(29) Pierotti, R. A. J. Phys. Chem. 1965, 69,281-288.

(30) Stillinger, F. H. J. Solution Chem. 1973, 2, 141-158.

(31 ) Hirata, F.; Arakawa. K. Bull. Chem. Soc. Jpn. 1973. 46.3367-3369.

(32) Pierotti, R. A. Chem. Rev. 1976, 76,717-726.

(33) Hirata, F.; Imai, T.; Irisa, M. Rev. High Pres. Sci. Tech. 1998, 8, 96-103.

(34) Kharakoz, D. P. J. Solution Chem. 1992, 27, 569-595.

(35) Bondi, A. J. Phys. Chem. 1964, 68,441 -451.

(36) Kharakoz, 5. P. Biophys. Chern. 1989, 34, 1 15-1 25.

(37) Perrin, C. L.; Gipe, R. K. Science 1989, 238, 1393-1394.

(38) Adya, A. K.; Neilson, G. W. J. Chem. Soc. Faraday Tram. 1991, 87,

279-286.

(39) Kirchnerova, J.; Farrell, P. G.; Edward, J. T. J. Phys. Chem. 1976, 80,

1974-1 980.

(40) Chalikian, T. V.; Sarvazyan, A. P.; Funck, Th.; Breslauer, K. J.

Biopolymers 1994, 34, 54 1-553.

Table 1

Molecular weights, M, and relative molar increments of sound velocity, [U], as a

function of temperature for the a,o-aminocarboxylic acids in 40.

glycine 78.09 41.5H.3 38.5k0.3 37.5k0.4 32.3k0.5

4-amino butanoic 106.14 65.7k0.4 59.740.4 55.3I0.5 52.0k0.7

acid --

acid --- 6-aminohexanoic 134.19 96.4k0.5 87.6'0.5 80.2I0.6 73.9k0.8

acid

7-amino heptanoic 148.22 110.1I0.6 103.1I0.5 88.9'0.7 81.9k0.8

acid

acid

- - 1 1-aminoundecanoic 204.33 164.1 20.9 150.5M.9 l2O.M .O lO4.Otl.2

acid

Table 2

Apparent molar volumes, $V, as a function of temperature for the a,o-

aminocarboxylic acids in D20 and H20a.

- -

glycine

4-aminobutanoic acid

5-aminopentanoic acid

6-amino hexanoic acid

7-aminoheptanoic acid

8-aminooctanoic acid

-1 1 -aminoundecanoic 181.3t0.9 183.510.9 184.611.1 186.4+,1.3 acid (1 80.1)" (1 82.6)" (1 84.5)= (1 86.9)a

a the partial molar volumes, V", of the a,o-aminocarboxylic acids in H20 are from ref. [Il].

Table 3

Apparent molar adiabatic compressibilities, 4Kç, as a function of temperature for

the a,o-arninocarboxylic acids in D20 and H20a.

glycine

4-aminobutanoic acid

5-aminopentanoic acid - 6-aminohexanoic -50.820.9 -40.4tl.O -31.8k1.1 -25.7I1.4 acid (045.6)~ (-37.1 )a (-27.1 )a (-2 1 .4)a

7-aminoheptanoic acid

8-aminooctanoic -60.3I1 .O -48.811.1 -29.6k1.5 -20.311.6 acid (-52.1 )a ( 4 1 .l)= (024.9)~ (-1 6.1 la

1 l -aminoundecanoic -71.1k1.6 -55.0t1.6 -25.1I1.9 -9.612.2 acid (-6 1 (-45.1 )a (-2 1 .2)a (-6.9)=

a the partial molar adiabatic compressibilties, Kas, of the a,o-aminocarboxylic acids in H20 are from ref. [ l 11.

Table 4

Partial molar expansibilities, EO, as a function of temperature for the a,o-

aminocarboxylic acids in D20 and H20a.

-- glycine 8.0k1.5 6.811 .5 4.4I1.4 1.9k1.5

(8.5Ia (6 .SIa (3 .2Ia (O)a

4-aminobutanoic acid

5-aminopentanoic 9.7-3.5 7.9I3.5 4.1 k3.5 0.4k3.5 acid (14.9P (1 0.6)a (1 .5)a (-7.6)=

6-aminohexanoic 1 1 . l +".O 9.324.0 5.5k4.0 1.7f4.0 acid (1 5.8la (1 1 .9)a (3.4)= (-5

7-amino heptanoic acid - 8-aminooctanoic 17.8S.O 14.716.0 8.026.0 1 -3k6.0 acid (1 7.2)= (14.1)a (7.5)" (1 - 1 1 -aminoundecanoic 19.2I7.5 16.7273 11.3k7.5 6.027.5 acid (24.6)a (21 .8)a (1 5.9)" (9.9)"

a the partial molar expansibilities, EO, of the a,o-aminocarboxylic acids in Hz0 are from ref. [ I l ] .

Table 5

The interaction volume, VI, for the independently hydrated amino and carboxyl

temini as a function of temperature for the a,o-aminocarboxylic acids in D20 and

H20a.

a The interaction volumes, VI, for the independently hydrated amino and carboxyl

termini of the a.w-aminocarboxylic acids in H20 are from ref. (1 1).

FIGURE LEGENDS

Figure 1

The difference between the partial molar volume of the a,o-aminocarboxylic

acids, VO, in D20 and their van der Waals volume, Vw, as a function of the van

der Waals surface area, SW, at 18 O C (O). 25 OC (O), 40 OC (a), and 55 OC (0).

Figure 2

The temperature dependences of the contributions of the independently hydrated

-CH2- group to the partial molar volume of the a,o-aminocarboxylic acids in Hz0

( a ) (frorn ref. 1 1 ) and D20 (0).

Figure 3

The dependence of the interaction volume, VI, on the number of methylene

groups in the a,o-aminocarboxylic acids in D20 at 18 OC (a), 25 OC (O), 40 OC

(i), and 55 O C (O).

Figure 4

The difference between the interaction volumes, VI, of the independently

hydrated amino and carboxyl terrnini of the a,o-aminocarboxylic acids in D20

and Hz0 as a function of temperature.

Figure 5

The dependence of the partial molar expansibility of the a,o-aminocarboxylic

acids in D20 on the number of methylene groups at 18 OC (a), 25 O C (O), 40 OC

(i), and 55 OC (O).

Figure 6

The temperature dependences of the contributions of the independently hydrated

-CHT group to the partial molar expansibility of the a.o-aminocarboxylic acids in

H20 (O) (from ref. I I ) and D20 (O).

Figure 7

The temperature dependences of the contributions of the independently hydrated

arnino and carboxyl termini ta the partial rnolar expansibility of the u,a-

arninocarboxylic acids in H20 ( O ) (frorn ref. 11) and 9 0 (0).

Figure 8

The dependence of the partial molar adiabatic compressibility of the a,on

aminocarboxylic acids in 90 on the number of methylene groups at 18 OC (*),

25 OC (O), 40 OC (i), and 55 OC (O).

Figure 9

(a) The temperature dependences of the contributions of the independently

hydrated -CHT group to the partial rnolar adiabatic cornpressibility of the a,o-

aminocarboxylic acids in H20 (a) (from ref. 11) and D20 (0); (b) The difference

between the contributions of the -CH2- group to the partial molar adiabatic

compressibility of the a,w-aminocarboxylic acids in D20 and H20 as a function of

temperature.

Figure 10

(a) The temperature dependences of the contributions of the independently

hydrated amino and carboxyl termini to the partial rnolar adiabatic compressibility

of the a,o-aminocarboxylic acids in H20 (a) (from ref. 11) and D20 (O); (b) The

difiference between the contributions of the independently hydrated amino and

carboxyl termini to the partial molar adiabatic compressibility of the a.o-

aminocarboxylic acids in D20 and H20 as a function of temperature.

S,, IO" cm2mol" Figure 1. The difference between the partial molar volume p, of the aminocarboxylic acids in D,O and their van der Waals volume, V W , as a function of the van der Waals surface area, S W , at 18, 25,40, 55 OC

Figure 2. The temperature dependences of the contributions of the independently hydrated -CH,-group to the partial molr volume of the aminocarboxylic acids in H,O

Figure 3. The dependence of the interaction volume, V., I on the number of methylene group in the aminocarboxylic acids in D,O at 18, 25, 40, and 55 OC

Figure 4. The difference between the interaction volumes, V., I of the independently hydrated amino and carboxyl termini of the amino acids in D,O and H,O as a function of temperature

Number of -CH,- Groups

Figure 5. The dependence of the partial molar expancibility of the aminocarboxylic acids in D,O on the number of methylene groups at 18,25,40, and 55 OC

Figure 6. The temperature dependences of the contributions of the independently hydrated -CH,- group to the partial molar expancibility of the aminocarboxylic acids in H20 and D20

Figure 7. The temperature dependences of the contributions of the independently hydrated amino and carboxyl termini to the partial molar exspanibility of the aminocarboxylic acids in H,O and D,O

Number of -CH,- Groups

Figure 8. The dependence of the partial molar adiabatic compressibility of the aminocarboxylic acids in D,O on the number of methylene groups at 18,25,40 and 55 OC

Figure 9(b). The difference between the contributions of the -CH,- group to the partial molar adiabatic compressibility of the aminocarboxylic acids in D,O and H,O as a function of temperature

Figure 1 O(a). The temperature dependences of the contributions of the independently hydrated amino and carboxyl termini to the partial adiabatic conpressibility of the amincarboxylic acids in H,O and D,O

Chapter 4. Conclusions

4.1 Summary of findings

The partial molar volumes and adiabatic compressibilities of five a - amino

acidç with aliphatic side chain at 25 OC, five oligoglycines at 25 OC, and eight a,o

- arninocarboxylic acids at temperatures l8,25,4O and 55 OC have been

detemined in D20. Based on these data, contributions of constituent groups ta

the volume and compressibility of solutions have been evaluated.

At 25 OC, the volume contribution of a nonpolar methylene group in D20 is equal

to 16.7 r 0.3 cm3mol". This value is slig htly higher than 15.9 i 0.4 cm3mol~', the

volume contribution of a methylene group in H20. At 25 OC, the compressibility

contribution of a methylene group in 90 is equal to - (3.2 I 0.4) x 1 o4

cm3mol-' bar-' which is significantly smaller (more negative) than -1.9~ 1 o4

cm3mo1~'bai', the value of K(CH2) in H20.

At 25 OC, the volume contribution of a polar peptide group, V(COND), in

DzO is equal to 23.1 t 1.1 cm3mol-'. This value is 1.5 cm3rnol-' larger than 21.6

cm3rnol*', the volume contribution V(CONH), of a peptide group in H20.This

difference suggests that hydrogen bonding between polar groups of a solute and

solvent molecules brings about a slig htly weaker contraction of water in D20 than

in H20.

At 25 OC, the compressibility contribution of the peptide group, K(COND),

in D20 is (1 .O I 1 .O) ~ 1 0 ~ cm3mol-'bar-', which is somewhat higher than

(0.5 t 0.8)x104 cm3mol"bai', the compressibility contribution, K(CONH), of the

peptide group in H20.

At 25 OC, the electrostriction of an independently hydrated pair of amino

and cahoxyl temini in D20 practically coincides with the corresponding value in

H20 and has been estimated to be -26 + 0.8 cm3mol-'. At 25 OC, the

compressibility contribution of an independently hydrated pair of amino and

cahoxyl temini in 90 iç equal to 437.4 + 1 . 0 ) ~ 1 0 ~ crn3m0r1bar-', which is 10%

more negative than the corresponding vaiue in H20.

The volume contributions of an independently hydrated methylene group.

V(CH2), in D20 and H20 are similar within -10.3 cm3rnoi-', although the

temperature dependence of V(CH2) may be somewhat steeper in H20 relative to

D20. In D20, the electrostriction, VI, of an independently hydrated pair of

charged amino and carboxyl terminal groups decreases (becornes more

negative) with temperature increasing. Specifically. the value of VI is equal to

-26.2 k 0.5 cm3mol-' at 18 "C and decreases to -27.5t0.5 cm3mol-' at 55 OC.

Compared to H20, the electrostriction of the charged tenini is stronger (more

negative) in D20 below 35 OC. At higher temperatures, the situation is opposite:

the solvent contraction in the vicinity of the charged temini is stronger in H20.

In D20, the expansibility contribution of methylene groups, €(CH2), linearly

decreases with ternperature from 0.013 I 0.005 at 18 O C to 0.008 I 0.002

at 55 OC. By contrast, in H20, E(CH2) increases with ternperature from 0.01 3 I

0.003 at 18 OC to 0.026 I 0.003 at 55 OC. This difference is statisticaliy

signficant. This discrepancy reflects the fact that, in H20, the molar expansibility

of water solvating aliphatic groups, Et, increases faster with ternperature than the

expansibility of the bulk solvent, Eo, while, in D20, Eo increases faster than Eh. In

D20, the expansibility contributions of the charged temini, E(ND~+ + COO-),

1 1 decreases from (9.713.5)~10" cm3mor K- at 1 8 OC to (0.413.5)~ 1 0" ~ r n ~ r n o l - ' ~ '

at 55 O C . Compared to H20, the expansibility contribution of charged groups is

smaller in D20 at low temperatures, while above -35 OC, the contribution of

charged groups in D20 becomes substantially larger than that in H20.

in D20, the compressibiiity contribution of an independently hydrated

methylene group, &(CH2), increases from (-4.0 t 0.3)~ 1 o4 cm3mof' bar*' at 1 8

OC to (3.3 t 0.3)x104 cm3mol-'bai' at 55 OC. Compared to H20, the value of

&(CH2) in D20 is smaller below 40 OC. However, at higher temperatures, the

value of &(CH2) in D20 becomes larger than in H20. It should be noted that,

within the whole temperature range studied, the difference between the values of

Ks (CH2) is statistically significant. In D20, the compressibility contribution of

charged groups, &(bJD3* + COO-), is negative within the whole temperature

range studied although it becomes less negative with temperature increasing: it

increases from (-48 & 0.9)xl o4 cm3mof'bai' at 18 OC to (-24.1 t

1 .2)xlo4 cm3rnol-'bar-' at 55 OC. Over the whole temperature range studied, the

value of KS(ND~' + COD) in D20 is smaller (more negative) than the value of

K S ( N H ~ + COO-) in H20. but this statistically significant difference diminishes

when the temperature increases.

4.2 Concluding remarks

Taken together, Our data suggest that, in D20, the hydration properties of

hydrophobic. polar and charged groups, as reflected in their volume and

cornpressibility contributions, are measurably distinct from those in H20.

Significantly, these volumetric characteristics of solute hydration differ not only in

their absolute values but also in their temperature dependences. These results

provide a detailed quantitative description of the hydration of nonpolar, polar and

charged protein groups in D20. Such characteristics should prove useful in

developing a belter understanding of the rule that difkrentlal D201H20 hydration

of atomic groups plays in modulating thermal and thermodynamic stability of

proteins. In addition, these results represent a step in building an empirical

database of differential volumetric parameters of protein functional groups in D20

and H20. Such a database is required for a possible application of differential

volumetric measurements to protein solutions in D20 and H20 to gain insight into

the amount and chernical nature of solvent-exposed protein groups in the

absence of structural information.

Appendix a-Amino acids

Glycine

Alanine

Aminobutiric acid

Norvaline

Norleucine

Oligog lycines

Glycine

Diglycine

Triglycine

Tetraglycine

a,o-aminocarboxylic acids

Glycine

acid

5-ami nopentanoic

acid

acid

7-aminoheptanoic

acid

8-aminooctanoic

acid

Y acid