UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A comparative study of the swelling and mechanical properties of vertebrate elastins Chalmers, Gavin William Geddes 1988

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-UBC_1988_A6_7 C42.pdf [ 3.99MB ]
JSON: 831-1.0097666.json
JSON-LD: 831-1.0097666-ld.json
RDF/XML (Pretty): 831-1.0097666-rdf.xml
RDF/JSON: 831-1.0097666-rdf.json
Turtle: 831-1.0097666-turtle.txt
N-Triples: 831-1.0097666-rdf-ntriples.txt
Original Record: 831-1.0097666-source.json
Full Text

Full Text

A COMPARATIVE STUDY OF THE SWELLING AND MECHANICAL PROPERTIES OF VERTEBRATE ELASTINS BY GAVIN WILLIAM GEDDES CHALMERS B.Sc, The University of British Columbia, 1984 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Zoology) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA July 1988 ©Gavin William Geddes Chalmers In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date DE-6(3/81) i i ABSTRACT The swelling-temperature compensation hypothesis as proposed by Gosline and French (1979) is examined by investigating the physical and mechanical properties of an evolutionary series of vertebrate elastins. Temperature-dependent swelling, low water contents and thermodynamics typical of hydrophobic systems were observed for all elastins except salmon. Salmon elastin, on the other hand, showed temperature-independent swelling and a high water content. Thermodynamic analysis showed that salmon elastin still contained a hydrophobic component. The swelling-temperature compensation hypothesis suggests that the extreme hydrophobic nature of elastin evolved in order to provide the proteins with temperature-dependent swelling and thus, maintaining elastic efficiency over a wide temperature range. All elastins should then be very hydrophobic systems which is inconsistent with the physical chemical results. All vertebrate elastins are not necessarily hydrophobic systems since salmon elastin shows no temperature-dependence to its swelling. The efficiency of a series of vertebrate elastins was measured over 0 to 60°C temperature range using a forced vibration technique. Over a wide frequency range, both lower and higher vertebrate elastins were capable of efficient spring-like behaviour. Both a higher vertebrate, with temperature-dependent swelling, and a lower vertebrate, with no temperaturate-dependence to its swelling, showed elastic efficiency. The swelling-temperature compensation hypothesis must be rejected. i i i T A B L E OF CONTENTS ABSTRACT ii T A B L E OF CONTENTS iii LIST OF TABLES v LIST OF FIGURES vi ACKNOWLEDGEMENTS vii I. G E N E R A L INTRODUCTION 1 II. PHYSICAL CHEMISTRY OF V E R T E B R A T E ELASTINS A. INTRODUCTION 8 B. MATERIALS A N D METHODS 10 1. Isolation and Purification 10 2. Amino Acid Composition 11 3. Hydrophobic Index 12 4. Water Contents 13 5. Swelling Tests 14 6. Calculation of Volume Fractions 16 7. Thermodynamics 18 C. RESULTS 22 1. Amino Acid Compositions 22 2. Water Contents 25 3. Swelling Tests 28 4. Thermodynamics 31 D. DISCUSSION 33 IR. M E C H A N I C A L PROPERTIES OF V E R T E B R A T E ELASTINS A. INTRODUCTION 42 B. MATERIALS A N D METHODS 45 i v 1. Sample Preparation 45 2. Dynamic Tests 45 3. Time-Temperature Superposition 47 C. RESULTS 49 1. Closed System 49 2. Open System 53 3. Effiency of Salmon and Turkey Elastin 53 D. DISCUSSION 58 IV. G E N E R A L CONCLUSIONS 70 REFERENCES 72 APPENDIX I. 79 V LIST OF TABLES Table 2.1 Amino acid composition data 23 Table 2.2 Amino acid composition summary 24 Table 2.3 Thermodynamics data 32 Table 2.4 Literature survey of arnino acid compositions for pig and cow ligament elastin 34 Table 3.1 Vertebrate hydrophobic indices and mean systolic pressures 67 V I L I S T O F F I G U R E S F igure 2.1 Swe l l i n g test apparatus. 15 F igure 2.2 Compar i son o f c o w l igament swel l ing w i th literature. 19 F igure 2.3 Absorpt ion isotherms o f a l l measured elastins. 26 F igure 2.4 The correlat ion between water content and hydrophobic index. 27 F igure 2.5 Swe l l i n g curves o f a l l measured elastins. 29 F igure 2.6 The correlation between swel l ing index and hydrophobic index. 30 F igure 2.7 Shark elastin network schematic. 40 F igure 3.1 T y p i c a l modulus-frequency and tan d frequency data. 50 F igure 3.2 Time-temperature superposition pr inc ip le. 51 F igure 3.3 C lo sed system L o g aT-temperature curve for turkey elastin. 52 F igure 3.4 T y p i c a l data for sa lmon and turkey elastin in swe l l i ng equ i l ib r ium. 54 F igure 3.5 Open system L o g aT-temperature curves fo r p ig , sa lmon and turkey elastin. 55 F igure 3.6 Tan d-temperature curves for salmon and turkey elastin. 56 F igure 3.7 Modulus-temperature regions fo r amorphous po l ymer 60 F igure 3.8 Open and c losed system turkey elastin 100 H z tan d-temperature curves. 62 F igure 3.9 Corre lat ion o f mean systol ic pressure w i th hydrophobic index. 65 ACKNOWLEDGEMENTS It is with great pleasure that I thank my research supervisor, Dr. John M . Gosline, for the privilege of joining his laboratory. This study could not have been completed without his thoughtful guidance, generous support and rare ability of making difficult concepts easy. I would also like to thank the other members of the lab, Dr. Margo Lillie, Cynthia Nichols and Steve Katz, for their helpful comments and friendship. This research was supported by a Natural Sciences and Engineering Research Council (N.S.E.R.C.) grant to Dr. John Gosline. 1 CHAPTER I: G E N E R A L INTRODUCTION Elastin is a vertebrate protein (Sage and Gray, 1979, 1980, 1981) which must function like a rubber band in a great many tissues of the body such as lung, cartilage, skin, ligament and arterial walls. Regardless of the location, elastin must possess two essential rubber-like properties. First, elastin must be extensible over the long range. Ligament elastin and arterial wall elastin may be subjected to 100% and 40% extensions respectively. Second, elastin must be able to store elastic energy efficiently. That is, the application of an external force causes the tissue to deform. The energy used to deform the tissue is stored and released to power the elastic recoil and recovery to its original dimensions. These two essential properties of elastin are crucial to proper cardiovascular function. As an extensible rubber, elastin smooths the pressure pulse originating at the heart and thus, reduces the pressure gradient the heart must pump against (Berne and Levy, 1988). By releasing elastic energy stored during the circumferential stretches of systole, elastin helps to propel blood through the circulatory network during diastole (McDonald, 1974). Elastin is one of a large group of polymeric materials which may have diverse chemical natures but which all show rubber-like properties. Moreover, it is the molecular structure of these polymeric networks which provide the rubber-like properties. Three criteria are essential if a polymer is to possess rubber-like properties. First, polymeric networks are composed of random-coil chains; molecules which have no set conformation but are constantly moving through Brownian motion. These random-coil chains are long, polymeric molecules in which the monomer units are linked by bonds that permit free rotation giving the chains flexibility (Flory, 1953). Carbon-13 NMR has been used to investigate the backbone chain mobility of elastin (Lyerla and Torchia, 1975; Fleming et a l , 1980). These workers have found that the elastin 2 backbone has a rotational correlation time, a measure of carbon mobility, between 10-7 and 10-8 seconds, characteristic of isotropic, rotational diffusion. In contrast, the correlation time for a fixed structured protein such as collagen is greater than 10-3 seconds (Gosline, 1976). Second, the polymer network must be isotropic, having similar properties in all dimensions. For elastin, polarized light microscopy studies by Aaron and Gosline (1979; 1980) show that cow ligament elastin has no intrinsic (molecular) or form (submicroscopic-3 to 5nm fiber range) birefringence. Since birefringence is essentially a measure of order, elastin fibers are optically isotropic in agreement with a random-coil network. Further, from X-ray diffraction studies, no apparent order exists when elastin is in the hydrated state (Gosline, 1976). For individual chains, each is isotropic, in both long and short range deformation (Gosline and Rosenbloom, 1984), so that there are few regions conducive to stable, intra- or inter- chain association. Third, the molecular chains are crosslinked into a macroscopic meshwork preventing them from slipping past one another during deformation. Elastin is made up of peptide chains of about 72,000 daltons (Gosline and Rosenbloom, 1984) crosslinked by tetrafunctional lysine derivatives, called desmosine and isodesmosine (Eyre et al., 1984), into a molecular meshwork. The elastic mechanism of rubber-like polymers is resident within the random-coil chains. In the undeformed state, the chains, by having their conformations subject to constant change through Brownian motion, have maximum entropy (Queslel and Mark, 1986). However, when deformed, the chains are pulled out in the direction of elongation (Flory, 1953) reducing the number of possible conformations. The molecules become ordered and conformational entropy is decreased. Since all systems spontaneously proceed toward a state of maximum entropy, this decrease in entropy can be used to store elastic energy. This type of elasticity is known as conformational entropy elasticity and is responsible for the restoring force of all known rubber-like polymers (Queslel and 3 Mark, 1987). The elastic mechanism of elastin is also entropic, however, it has two contributions. The first is called the hydrophobic mechanism (Gosline, 1978a). Elastin is made up primarily of non-polar amino acid residues (Ross and Bornstein, 1969; Stevens et al., 1974; Starcher and Galione, 1976; Serafini-Fracassini et al., 1985; Guantieri et al., 1987). Elastin, under physiological conditions, is immersed in an aqueous environment so that most of the hydrophobic amino acid side chains are only partially hydrated as compared to polar amino acid side chains and peptide backbones. When elastin is stretched, the molecular chains are forced to increase their average hydration. Since non-polar groups want to minimize their contacts with water, this unfavourable situation has a positive free energy change, thus, free energy can be stored and used to power the elastic recoil. The unfavourable free energy for the process is driven by large decreases in entropy due to the organization of water molecules around non-polar amino acid side chains. Gosline (1978a) found that this mechanism is only important up to about 70% extension. Beyond 70% extension, the second contribution to the elastic mechanism, the conformational entropy mechanism, dominates the elastic process. Like all amorphous polymers, elastin is a viscoelastic material capable of showing both elastic and viscous properties. Under physiological conditions, viscous properties are minimal and elastin behaves elastically. That is, virtually all of the strain energy imparted to elastin is recovered when elastin recoils elastically. However, even under optimum physiological conditions, total recovery is not possible since weak intermolecular interactions exist which resist the changes in network chain conformation. Mechanical energy is required to overcome these intermolecular interactions, which in effect cause frictional forces. Thus, of the total energy put into the system, only part is available to contribute to the elastic recoil while the rest is dissipated as heat. Elastin, like all polymers which will have intermolecular interactions, can show 4 large viscous properties under three conditions. First, if the elastin is deformed at sufficiently high rates, the molecular chains, driven solely by random thermal motion, will not be able to alter their conformation as rapidly as required by the applied deformation, and large amounts of energy will be required to overcome the frictional forces. So, for a given energy input, only a fraction will be available for elastic recoil, and, the polymer will not be an efficient spring. Second, if elastin is cooled to low temperatures, even at low strain rates, viscous energy losses can be significant since the molecular chains will have insufficient thermal energy to overcome frictional forces. Similarly, high temperatures provide the molecular chains with sufficient thermal energy to give elastin high efficiency at high strain rates, at least until degradation of the network occurs. This strain rate-temperature correspondence is known as the time-temperature superposition principle and will be explained in greater detail in chapter 3. Since elastin is immersed in an aqueous environment and is largely composed of hydrophobic amino acid residues, much of the water associated with the network is loosely bound and easy to remove. Decreasing the water content of elastin can have dramatic effects and represents the third condition where elastin can suffer viscous losses. Gotte et al. (1968) showed that the elastic modulus, a measure of rigidity, is dependent on water content. At low water contents, the elastin is like a rigid glass, but with increasing amounts of water up to saturation, elastin becomes a rubbery material. If the water content is reduced, less water molecules will be available to the elastin network to disrupt frictional forces. The interplay between these variables, strain rates, temperature and water content, allows elastin to be rubber-like under a wide range of conditions. For example, at low strain rates and near zero water content, a temperature of about 200OC is required for elastin to be rubbery (Kakivaya and Hoeve, 1975). Moreover, at high strain rates (100 Hz) and low temperatures (5oC), large increases in water content allow elastin to remain 0 5 rubbery (Gos l ine and French, 1979). These three variables are capable o f affecting the ef f i c iency o f elastin as a spring and must be dealt w i th i n order for it to per form phys io log ica l ly . Gos l i ne and F rench (1979), us ing a forced v ibrat ion testing protocol , found that elastin was capable o f storing elastic energy e f f i c iendy over both a w ide temperature range ( 0 - 7OC ) and a w ide v ibrat ional frequency range (0-200 Hz ) . However , elastin on ly demonstrated a h igh ef f ic iency when it was tested wh i l e immersed in an aqueous environment. W h e n removed f r om the aqueous environment or tested at a f i x ed water content, elastin becomes ineff ic ient at l o w temperatures. Gos l i ne and F rench attributed this h igh ly temperature-independent e f f i c iency to the temperature-dependent water content o f elastin. The water content o f elastin increases by approximately 6 5 % as the temperature is decreased f rom 70 to 0°C, w i th most o f the increase occurr ing between 30 and 0°C (Gos l ine, 1977; Gos l i ne , 1978b). S ince the so lubi l i ty o f non-polar groups increases w i th decreasing temperature (Tanford, 1980), Gos l i ne (1977; 1978b) correlated the increas ing water content or swe l l i ng o f elastin w i th the weaken ing o f hydrophobic interactions due to changes i n non-polar group so lubi l i ty as temperature decreases. A s the temperature decreases, the thermal energy o f the chains also decreases unt i l a point is reached where the chains are no longer able to overcome the intermolecular attractive forces. The material becomes a glass and is def in itely not suitable to funct ion ef fect ive ly w i th in arterial wal l s . F o r elastin, water is an important p last ic izer (Gos l ine, 1976) and can interrupt intermolecular interactions and delay the onset o f glass- l ike behaviour. Swe l l i n g offsets the energy losses w h i c h w o u l d occur due to f r i c t ion and thus, the e f f i c iency o f elastin does not decrease w i th decreasing temperature. In fact, the overa l l mechanica l ef f ic iency o f elastin is preserved over a wide temperature range due to swel l ing. Moreover , swe l l ing is d i rect ly dependent on the hydrophobic nature o f elastin. W i thout a hydrophobic nature, temperature-dependent 6 swelling would not occur. Based on the relationship between swelling and hydrophobicity, Gosline and French (1979) proposed what shall be called the swelling-temperature compensation hypothesis: the hydrophobic nature of elastin evolved in order to provide a protein which is capable of storing elastic energy over a wide temperature range. This hypothesis implies that elastin must increase its water content with decreasing temperature in order to be efficient mechanically, particularly at low temperatures. Sage and Gray (1979, 1980, 1981) found that the hydrophobic nature of elastin varies amongst the vertebrates. The most primitive vertebrates, the Agnathans, do not have elastin by histological (Sage and Gray, 1980; Wright, 1984) and biochemical (Sage and Gray, 1979) criteria. However, considering vertebrates which have elastin, trends are evident. Higher vertebrates (birds and mammals) have more hydrophobic elastins than the lower vertebrates (fish), with reptiles and amphibians having elastins of intermediate hydrophobicity. Given that these differences in elastin biochemistry exist, how do they affect temperature-dependent swelling, and subsequently, mechanical efficiency? Further, in their studies, Gosline and French used cow ligament elastin. Cows are homeotherms and require that their elastin only be efficient at about 37°C. Other vertebrates have body temperatures ranging from -2<>C in arctic fish (Wilson, 1979) to 42oC in birds (Sturkie, 1976). If the swelling-temperature compensation hypothesis is to be accepted, a study of elastins from vertebrates other than homeothermic mammals seems warranted. In particular, the fish elastins must be studied since, of all vertebrates, they are exposed to the lowest temperatures where mechanical efficiency may become limiting. To evaluate the swelling-temperature compensation hypothesis, a series of elastins from each of the vertebrate classes must be exarnined. The purpose of this study is to evaluate the swelling-temperature compensation hypothesis proposed by Gosline and French (1979) by examining an evolutionary series of vertebrates. Chapter 2 will investigate the physical-chemical differences, swelling and 7 water contents, which occur between elastins from fish to mammals and their relationship to amino acid composition. Chapter 3 will use a forced vibration testing protocol, similar to Gosline and French (1979), to assess the mechanical behaviour of elastins from animals with the extremes of physical chemical properties. In this study, the physical and mechanical properties of bird and fish elastin will be tested for the first time. 8 C H A P T E R n: T H E P H Y S I C A L C H E M I S T R Y O F V E R T E B R A T E E L A S T I N S A . I N T R O D U C T I O N Water is extremely important to the rubber- l ike protein elastin. In fact, when removed f r o m water, elastin becomes a r i g id , glassy so l id (Gosl ine, 1976). Further, part o f the elastic mechan i sm o f elastin requires the exposure o f non-polar groups to water (Gos l ine, 1978a). D u r i n g extension, there is an unfavourable increase i n average hydrat ion o f non-polar groups, thus, free energy can be stored and used to power the elastic reco i l . E la s t in has been shown to have a temperature-dependent water content (Grut and M c C r u m , 1974; Hoeve and F lo ry , 1974; Gos l ine, 1977; Gos l ine, 1978b). Th i s temperature-dependent water content is referred to as swel l ing, an increase i n the vo lume o f the protein network brought about by the uptake o f solvent. Spec i f i ca l l y , elastin increases its water content or swells by approximately 6 0 % as the temperature decreases f r o m 60°C to 1°C. A s the so lubi l i ty o f non-polar groups i n water increases as temperature decreases (Tanford, 1980), Go s l i ne (1978b) correlated the temperature-dependent swe l l i ng o f elastin w i th decreasing strength o f hydrophobic interactions due to changes i n non-polar amino ac id side chain solubi l i ty. A l l o f the above studies have focused on c o w or ox l igament elastins, since they represent the sources w i th the highest proportions o f elastin (Ayers, 1964; M i n n s and Stevens, 1978). However , elastin is present i n a l l vertebrates except Agnathans (Sage and Gray, 1979, 1980, 1981; Wr ight , 1984), and a l l o f these elastins are not the same. Sage and G r a y (1979, 1980, 1981), i n an extensive evolut ionary survey o f elastin us ing h i s tochemica l and b iochemica l procedures, found that higher vertebrates have elastins w i th a higher percentage o f hydrophobic amino acids than l ower vertebrates. They used a phys ica l -chemica l parameter, ca l led hydrophobic index (HI), to assess 9 biochemical differences. Essentially, the HI is the ratio of non-polar amino acids to polar amino acids. One of the main attributes of elastin is the ability to store energy efficiently. Further, when considering ectotherms, the ability to store energy at low temperatures becomes the crucial requirement. Efficient low temperature elastic energy storage has been attributed to the temperature-dependent swelling of elastin (Gosline and French, 1979). Further, it has been suggested by Gosline and French (1979) that the extreme hydrophobic nature of elastin evolved in order to provide vertebrates with proteins that could store energy efficiently at low temperatures. To evaluate what could be termed the swelling-temperature compensation hypothesis of Gosline and French (1979), the elastins from an evolutionary series of vertebrates have been examined in two different respects. First, the swelling and water content of elastin have been measured and correlated with the HI of Sage and Gray (1981) in order to establish the presence or absence of a hydrophobic component to the network. Secondly, these elastins have been tested mechanically to estimate their elastic efficiency as a function of temperature and this data will be dealt with in the following chapter. 10 B. MATERIALS AND METHODS 1. Isolation and Purification. Bovine ligamentum nuchae (Bos sp.) and pig (Sus scrofa) and turkey (Meleagris sp.) aortic arches were obtained from local slaughter houses. Salmon (Onchorhynchus keta) bulbus arteriosae and shark (Squalus acanthias) ventral aortae were obtained from local fish plants. Frog (Rana catesbeiana) and turtle (Chrysemys picta) aortic arches were obtained from healthy animals from university laboratories. In all cases except bovine ligamentum nuchae, the criterion for elastin selection was proximity to the heart. Hence, only ascending aortae or equivalent were used. Elastin was isolated following a purification protocol modified from Sage and Gray (1979). The modification involved the exclusion of organic solvent extraction. I tested the effects of organic solvent extraction and found no significant change in weight following a 24 hour extraction in acetone. Therefore, I decided to omit this extraction step. Samples were stripped of fat, small vessels and blood clots and washed several times with distilled, de-ionized water. The samples were then autoclaved at 15 psi and 120°C six times for periods of 30 to 45 minutes to remove collagen. Samples were then extracted with 6M guanidine hydrochloride for 24 hours to remove soluble proteins (Sage and Gray, 1979). Following several rinses with distilled water, the elastin samples were soaked overnight in distilled water and then autoclaved in distilled water twice for 30 to 45 minute periods. Elastin samples were then stored in distilled water in sterile containers for subsequent use. 11 2. Amino Acid Composition. Purified elastin samples from turtle and shark were hydrolysed following a standard protocol. Briefly, approximately 10 mg of protein was hydrolysed with 1 ml of 6M HC1 and flushed with liquid C02 and sealed under vacuum in hydrolysis tubes. The sealed tubes were then placed on a heating block for 22 hours at 105OC. Following overnight hydrolysis, the tubes were cracked and the hydrolysates centrifuged for 10 minutes at 15,000 rpm. The supernatant was transferred to a new centrifuge tube which was placed over NaOH pellets in a vacuum desiccator for 2 or 3 days, allowing the amino acids to dry to a powder. The dried amino acids were shipped to a commercial lab and analysed using standard techniques. The first amino acid hydrolysates were contaminated with carbohydrate, since the powders were very brown in colour, and the amino acid compositions of turtle and shark showed abnormally large peaks for polar amino acids compared to literature values for similar animals (Sage and Gray, 1981). Carbohydrate impurities typically co-elute with polar amino acids. Consequently, the elastins were prepared for amino acid composition by further purification by immersion in 0.1 M sodium hydroxide at 7 0 ° C for 25 minutes (Lansing et al, 1952). Turtle and shark elastin amino acid compositions were examined in order to determine whether any elastin proteolysis had occurred during the NaOH step. The ratios of glycine, alanine, proline and valine, the four dominant amino acids in elastin, were compared before and after NaOH hydrolysis. In the event of elastin proteolysis, these ratios should decrease since they would be removed prior to amino acid hydrolysis. The ratios increased after the hydrolysis, suggesting that the contaminants were removed and the integrity of the elastin network was preserved. The amino acid compositions for pig, turkey, salmon and frog were taken from Sage and Gray (1979). 12 3. Hydrophobic Index. To quantify the differences between amino acid compositions, the hydrophobic index (HI) (Sage and Gray, 1981) was used. The HI is made up of two components: (1) the fractional charge (FC) (Welscher, 1969), and (2) the average hydrophobicity (Havg) (Bigelow, 1967). Average Hydrophobicity HI= (1) Fractional charge The fractional charge is simply the sum of the mole fraction of polar amino acids (Glx, Asx, His, Lys and Arg). The average hydrophobicity is represented as follows: Havg =Zm x nO (2) The average hydrophobicity is the sum of the free energies of transfer (/CLFO multiplied by the mole fraction of each amino acid (nj). It represents the average residue hydrophobicity of the protein. The free energy of transfer, as given by Tanford (1962) provides a hydrophobic ranking system for individual amino acids, represents the change in free energy when an amino acid is transferred from an organic solvent, ethanol or dioxane, to water. In general, when a non-polar molecule, such as a non-polar amino acid side chain, is placed in water, water molecules are organized into cage-like hydration shells (Cantor and Schimmel, 1980). This organization involves two thermodynamic events: one, a decrease in enthalpy reflecting an energetically favourable increase in bonding between water molecules (Tanford, 1980), and two, a decrease in entropy reflecting the unfavourable increase in organization of the water molecules in the hydration shell. Since this entropy term dominates, the overall free energy for the process is positive. 13 4. Water Contents. Water contents (g water/g protein) were obtained using the method of Bull (1944). This method involves the dehydration of wet tissue in a sealed chamber at reduced relative humidity (RH) using a saline solution. For example, 0.15M NaCl solution gives a 99.5% relative humidity at 37°C (Scatchard et al, 1938). I am using 99.5% RH as saturation level since it is not possible to obtain 100% RH with this method. If one were to use 100% RH, water droplets would fall on the sample. For pig elastin, the absorption isotherm data were fitted to a second order polynomial by least squares regression (r2=0.995). Extrapolating to 100% RH, I estimate that by using 99.5% RH water content as the saturation value, I am underestimating the saturation value by about 5%. Only one absorption isotherm point was obtained for turtle and frog elastin, thus, this extrapolation technique could not be used. Purified elastin samples were placed into sterilized glass canning jars containing an inverted plastic beaker. Sterilized forceps were used to transfer elastin from a sterilized petri dish to a sterile gauze pad used to blott excess water from the sample. The bottom of the beaker was cut out and replaced with a polymer coated steel mesh. Elastin samples were placed on the mesh, suspended above a sodium chloride salt solution (350 ml), sealed, placed in a constant temperature chamber (37°C + 0.5°C) and equilibrated for 10 days. Equilibrated elastin samples were weighed on a Mettler balance. This weight measurement was called the wet weight. The elastin samples were then placed in a 105°C + 1.0°C oven for two days. Following this drying period, a dry weight was obtained. Thus, water content (g water/g protein) was defined as: wet weight - dry weight W37,99.5RH= (3) dry weight 14 5. Swelling Tests. Purified elastin samples (15x5mm) were allowed to air dry for about 30 minutes in an open petri dish. For frog and turtle elastin, whole arteries were used because the vessels were so small. The drying period allowed the tissue to gain enough rigidity for gluing. The elastin samples were potted with araldite epoxy into modified stainless steel machine screws with grooves cut into the ends. The elastin was rehydrated after potting by autoclaving, causing the tissue to expand against the epoxy. No difference in colour or texture was noted between sample before and after the air drying episode. The samples were stored in sterile, distilled water until use. The elastin samples were impaled with two short segments of 30 gauge hypodermic tubing and suspended in a 70 ml plastic culture flask with a glass face. The container had a thermistor (T) epoxied through the wall and placed within about 2 mm from the test sample so that temperature could be monitored (see figure 2.1). The thermistor was calibrated with a mercury thermometer to within + 0.5°C by measuring resistances over a 0° to 70°C temperature range. The calibration data gave a non-linear curve and a polynomial regression provided a poor fit to the data. For this reason, the data were sub-divided into seven segments and linear regressions were fit to the data. This procedure allowed me to predict temperature from thermistor resistance to within + 0.1°C over the 0 to 70°C temperature range using a Digital PDP 11-23 computer. Swelling data, consisting of length changes with temperature, were collected using a video measuring system (DeMont and Gosline, 1988; Biedka et al., 1987; Bush et al., 1982) as illustrated in figure 2.1. A video camera was focused on the centre of the elastin sample, giving about 20x magnification on the television monitor. A bright light was place behind the temperature bath so that the sample appeared as a black object on a light background. The video dimension analyzer (VDA, model 303, Instruments for Physiology and Medicine, San Diego, CA.) produces an Camera VDA Monitor Compute Fig. 2.1. Swelling test apparatus. 16 electrical signal proportional to the distance between two contrast boundaries in the video image. Since the whole sample image is black, the VDA can sample the distance between the two hypodermic needles. The analogue output from the VDA was digitized by an A/D converter on a Digital Mine 11-23 computer. Length changes were converted into volumes changes by assuming that elastin swells equally in all dimensions. Thus, length changes were raised to the third power giving volume changes. This assumption is supported by two lines of evidence. First, using polarized light microscopy, two studies have shown that elastin birefringence, a measure of structure, was essentially zero (Aaron and Gosline, 1980, 1981). Second, by measuring the dimensional changes which occur when elastin is transferred from water to DMSO, a solvent which swells elastin dramatically at room temperature (Mistrali et al., 1971), I concluded that elastin swells isotropically. The details are not presented here but the null hypothesis of an one-way ANOVA could not be rejected at the 95% confidence level (F= 1.823, df = 2, 15). For the purposes of comparing swelling profiles for different samples from different sources, the swelling test volumes were normalized to 1.0 at 37°C, therefore, they are expressed in arbitrary units and are called relative volumes, V37 r e l . 6. Calculation of Volume Fractions. A volume of a piece of elastin contains two components: the protein volume and the solvent volume (water). Thus, (4) Dividing by the total volume converts these volumes into volume fractions: 1 = v 2 + Vi (5) 17 where v 2 is the vo lume fract ion o f po lymer and v i is the vo lume fract ion o f solvent. B y assuming the protein vo lume w i l l not change over the 0 to 60°C temperature range, any changes i n the total vo lume o f elastin are due entirely to changes in solvent vo lume. Thus, v i can be thought o f as a measure o f solvent content. In this study, the solvent is water. The calculat ion o f vo lume fractions is out l ined below. The vo lume fractions o f protein and water at 37°C can be calculated f r om the measured water contents at 37°C. U s i n g an elastin density o f 1.23 g/cm3 fo r l igament elastin (Scandola and Pezz i n , 1980), the part ial specif ic vo lume o f protein w i l l be 0.81 cm3/g for a l l elastins tested. The part ial specif ic vo lume o f water was assumed to be 1.0 at a l l temperatures. Therefore, by mu l t i p l y i ng the water content at 37°C (W37,99.5 RH, grams water/gram elastin) by the above part ial specif ic volumes, mass, measurements are converted to volumes. Further, by assuming that the volumes are addit ive, the addit ion o f Vprotein and Vwater, 37°c gives Vt0tai, 37°c- The vo lume fract ion o f protein at 37°C can be calculated f r om the f o l l o w i n g equation: Vprotein v2, 37°C = (6) Vtotal, 37°C U s i n g the assumptions o f addit ive vo lumes and constant protein vo lume, the vo lume fract ion o f protein at any temperature, v 2 , T°C» can be calculated. F r o m swe l l i ng tests, the relat ive vo lume at any temperature, V37 r e l (T), is used i n the f o l l o w i n g equation to calculate the vo lume fract ion o f protein at any temperature, v 2 i T°C-v2,37°C V 2 , T ° C = (7) V37 r e l (T) The vo lume fract ion o f water at any temperature ( v 0 can be calculated by subtracting the vo lume fract ion o f protein (v 2 ) f r o m 1.0. U s i n g c o w l igament elastin, I was able to compare my data w i th Ha rmon and 18 Gosline (unpublished results) and Mistrali et al. (1971). Figure 2.2 shows a plot of the volume fraction of water (vO plotted as a function of temperature. The data sets are all very similar, with the means within the overlapping standard error bars. Estimates of the error associated with the data of Mistrali et al. (1971) were not available, however, their values lie within the error bars. To compare the overall swelling of elastin, a parameter called the swelling index (SI), defined as the ratio of relative volumes at the two extreme test temperatures, 1°C and 60°C, is used and calculated as follows, V 3 7 re l ( l °C) SI = (8) V37 r e l (60C) 7. Thermodynamics. According to the Flory-Rehner theory for the swelling of a crosslinked polymer network, the total free energy for dissolving solvent in a polymer network is given by the sum of the changes in the conformational free energy and the mixing free energy, as expressed, -[ln(l-V2)+V2+S(v2)2] = DV/Mc[(v20)2/3(v2)l/3-V2/2] (9) where v2 is the volume fraction of polymer, v2o is the volume fraction point at which crosslinks were introduced, D is the density of dry polymer, V is the molar volume of solvent, M c is the molecular weight between crosslinks and 6 is the solvent elastomer interaction parameter. For all elastins tested with water as the solvent, the molar volume and density were assumed to be 0.018 L/mol (Gosline, 1978b) and 1.23x103 kg/m3 (Scandola and Pezzin, 1980) respectively. By measuring stiffness via stress/strain experiments, Aaron and Gosline (1981) calculated Mc to be about 6000 for cow ligament elastin using the relationship between 19 Fig. 2.2. Comparison of swelling profiles for cow ligament elastin for Chalmers (•, n = 10), Harmon and Gosline (• , n = 7) and Mistrali et al (1971)(o). Error bars represent standard errors about the means. 20 stiffness and network characteristics for network polymers. Equation 9 is insensitive to variation in Mc when elastin is tested in water. Using Mc values ranging from 4.0 kg/m3 to 15.0 kg/m3, the calculated interaction parameters differed by only 3% or less. The term on the right-hand-side of equation 9 represents the conformational free energy, which can be equated with entropy changes associated with the expansion of the polymer network. As the network expands, the molecular chains are forced into less favourable conformations, and this creates a force resisting deformation. The mixing free energy given on the left-hand-side of equation 9 contains two components. The first two terms on the left represent the ideal entropy of dilution of the polymer by the solvent without considering the chemical interaction between them. The B parameter is proportional to the free energy of interaction between polymer and solvent, which will have enthalpy and entropy contributions. From swelling tests, volume fractions (v2) are used to calculate B parameters for each temperature (equation 9). Flory (1953) showed that B could be equated with the standard free energy of mixing, associated with the transfer of one mole of solvent from the bulk phase into the swollen network. According to Flory (1953), A G m i x = B R T / z x (10) where R is the gas constant (cal/mol K), T is the absolute temperature and zX is the lattice coordination number, which was calculated by Gosline (1978b) to be 4.2 for cow ligament elastin. Thus, in theory, AGm^ represents the change in free energy to transfer one mole of water into the elastic network. The free energies were calculated from equation 10 over the 0 to 60°C temperature range for all elastins studied. Following the protocol by Edelhoch and Osborne (1976), the free energy versus temperature data were fitted to third order polynomials by least squares regression. The worst case fit had a correlation coefficient 21 of 0.998. The equation has the form: A G ^ = A+ BT+ C T 2 + DT3 (11) The coefficients derived from equation 11 are used to calculate the relevant thermodynamic quantities: enthalpy change (AH), entropy change (AS) and heat capacity change (ACp). Thus, A H AS ACp -B - 2CT - 3 D T 2 A - C P - 2DT3 -2CT - 6 D T 2 (12) (13) (14) 22 C. RESULTS 1. Amino Acid Compositions. The results of the analysis of turtle and shark amino acid compositions are presented in table 2.1 along with pig, salmon, frog and turkey elastin amino acid compositions, taken from Sage and Gray (1979, 1981). The bovine ligament composition was taken from Stevens etal. (1974). At the bottom of table 2.1, the calculated hydrophobic indices and their component parameters, average hydrophobicities (Havg) and fractional charges (FC) are given. In general, from lower vertebrates to higher vertebrates, the HI increases while the fractional charge decreases. Average hydrophobicity does not correlate with any trend from the vertebrates studied and is approximately equal in all elastins except salmon and shark. Interestingly, salmon elastin has the lowest average hydrophobicity whereas shark elastin has the highest, although there are not enough data to add statistical significance to this statement. Table 2.2 summarizes the relative percentages of polar and non-polar amino acid residues. Consistent with the trends observed for HI and fractional charge, the fraction of non-polar residues increases and the fraction of polar residues decreases as one moves from lower, towards higher vertebrates. Note that the percentage values in table 2.2 do not include glycine, which makes up between 28 and 40% of the total elastin compositions. Elastins with higher hydrophobic indices, like turkey and pig, have both a higher fraction of non-polar residues and lower fraction of polar residues than elastins with lower hydrophobic indices, such as the fish elastins. 23 Table 2.1. Amino acid compositions of vertebrate elastins. Values in residues per 1000 residues. AMINO T U R K E Y PIG FROG T U R T L E SHARK S A L M O N COW ACID LIGAMENT Lys 3 5 7 9 13.5 8 2 His 0.5 1 2 2 3.5 6 -Arg 7 8 9 7 17 26 4 Asx 6 6.5 12 11 23 17 5 Thr 13 15 23 20 17 29 7 Ser 7 12 15 13 19 30 8 Glx 15 19 29 29 53.5 36 16 Hyp 7 9 5 n/a n/a 6 6.5 Pro 128 113 104 124 129 86 124 Gly 353 313 402 353 284 421 337 Ala 187 244 154 189 153 151 245 Val • 154 128 83 124 136 46 116 De 24 18 33 15 24 10 21 Leu 56 54 60 64 43 63 64 Tyr 14 19 42 35 51 49 8 Phe 22 33 15 11 27 11 29 Cys <1 <1 3 - - <1 -Met 1.2 <1 4 2 7 6 -Ide 1.2 2 1 n/a n/a <1 6.3 Des 1 1.3 1 n/a n/a <1 6.3 A v g H 1.05 1.04 1.06 1.02 1.12 0.82 1.03 F C 32 40 59 58 111 94 28 HI 33 26 16 17 10 9 37 24 Table 2.2. Summary of elastin amino acid compositions. Vertebrate HI % Non-polar % Polar Turkey 33 59 4 Pig 26 57 5 Turtle 17 52 7 Frog 16 46 7 Shark 10 52 13 Salmon 9 40 12 Cow Ligament 37 60 3.5 Note that glycine, which makes up between 30 and 40% of the elastin composition, is not included the percentages. 25 2. Water Contents. Figure 2.3A shows the 3 7 0 C absorption isotherms for salmon, pig, turkey, cow ligament and shark elastin. Turtle and frog elastins are included as single points at 99.5% RH. In general, all the curves indicate that a large proportion of the water associated with the elastin is loosely bound and possibly associated with non-polar groups and can be removed at very high RH. For example, the elastins tested lose between 25 and 50% of their bound water when subjected to small saline gradients between 99.5 and 97% relative humidity, respectively. Further, the slopes of the curves diminish considerably below 97% RH, suggesting that the water is more tightly bound. The cow ligament results in panel A are similar to Gosline (1987), who went to relative humidities as low as 90%. Thus, more tightly bound water, such as that associated with the polar backbone and ionic groups, will require relative humidities of 90% and lower before it is removed, as cow ligament data from Green (1948) shows in panel B. The data from Green (1948) at 90% RH are similar to Gosline (1987) at 90% RH. Figure 2.4 shows the mean saturation water contents (g H20/g elastin) at 3 7 ° C plotted as a function of hydrophobic index (HI). A linear regression of individual water content data excluding shark elastin, indicated that water content is inversely related to HI. This correlation is significant to the P=0.001 level (r=0.5, df=54). So, a more hydrophobic elastin, with a higher HI, will have a lower average hydration than a less hydrophobic elastin, with a lower HI. Specifically, the water content of salmon elastin, with the lowest HI, is significantly higher than any other measured elastin. Note, however, that shark elastin does not fit this pattern. It has a significantly lower water content than any other elastin, and yet its HI is not significantly different from salmon elastin. In fact, if shark elastin is included in the above regression analysis, the correlation is not significant at the P=0.05 level (r=0.22, df=63). Clearly, the water content for shark elastin is anomalous. 25a Fig. 2.3A. Absorption isotherm data for salmon (•, n= 10), pig (•, n= 12), turkey (A, n= 12), cow ligament (O, n= 12) and shark elastin (v, n = 12). Frog and turtle elastin are represented as a single point at 99.5% relative humidity (• , n = 10 for both). Error bars represent standard errors about the means. Fig. 2.3B. Absorption isotherm data for cow ligament elastin taken from Green (1948)(A) and this study (O). 27 C 0.3 * — 1 i 1 r -0 10 20 30 40 Hydrophobic Index Fig. 2.4. The relationship between water content and hydrophobic index (mean ± S.E., n>.l0) as indicated by linear regression of the data excluding shark elastin (r*=0.75). With shark elastin included in the regression, the r 2 drops to 0.16. The symbols represent salmon (o, n = 7), pig (•, n = 12), turkey (•, n = ll), cow ligament (1, n=9), frog (•, n = 6), turtle (a, n = ll) and shark elastin (•, n=9). 28 3. Swelling Studies. Figure 2.5 shows the temperature-dependent swelling profiles for all elastins, with volume fraction of water (yi) as the y-axis. v i is useful since it incorporates both the relative volume changes and the absolute water contents. Generally, all of the proteins studied, except salmon elastin, showed a typical temperature-dependent swelling profile associated with mammalian elastin, where volume increases as temperature decreases. On the other hand, salmon elastin seems to be nearly temperature independent, but it has a higher v i at all temperatures above 20OC. At 20OC, there is a crossover point where the other elastins, through swelling, attain higher water contents than salmon elastin. Near 0°C , salmon elastin has the lowest water content of all elastins studied. Figure 2.5 also shows the temperature-dependent swelling profile of shark elastin. It has the lowest water content at all temperatures except IOC where it was essentially the same as salmon elastin. Again, shark elastin appears anomalous since its hydrophobic index (HI) was shown to be similar to salmon elastin. This anomaly is shown graphically in figure 2.6, in which the means of a semi-quantitative measure of the temperature dependence of elastin swelling, the swelling index, are plotted against HI. A linear regression of individual swelling indices, including shark elastin, was significant (r=0.8, df=59, P<0.001). However, excluding shark elastin, the regression was more robust (r=0.96, df=50, P<0.001). When the mean swelling indices, as shown in figure 2.6, are regressed against hydrophobic index, including shark elastin, the correlation is not significant at the P=0.01 level (r=0.83, df=5). When shark elastin is excluded from the regression, the correlation is significant at the P=0.001 level (r=0.97, df=4). Again, shark elastin does not fit the general trend. 29 0.6-c o o 03 E o > 0.5 0.4 0.3 10 20 30 Temperature (°C) 40 50 60 Fig. 2.5. Swelling curves. The volume fraction of water as a function of temperature. Each line represents swelling results from at least 7 different samples. The symbols on this graph represent: salmon (A), turkey (•), pig (o), shark (v), frog (•), turtle (A) and cow ligament elastin (•). Means and standard errors are given in appendix I. 30 Fig. 2.6. The relationship between swelling and hydrophobic index for salmon (o), turkey (•), pig (•), shark (•), frog (•), turtle (A) and cow ligament elastin (A). The line represents a linear regression of the data exchiding shark elastin (mean ± S.E., n>7; 1^  = 0.91). With shark elastin included, the r2 drops to 0.71. 31 4. Thermodynamics. Table 2.3 shows the calculated free energy, enthalpy, entropy and heat capacity changes associated with the mixing process for all the elastins studied. In all cases, the free energy of mixing is positive and becomes increasingly positive as temperature increases, indicating that the process for transferring water into the elastin network (ie increasing the average exposure of non-polar groups to water) becomes even more unfavourable as temperature is increased. Following an opposite trend to that of free energy, both enthalpy and entropy changes become less negative with increasing temperature and in all cases, the heat capacity change is positive at all temperatures. The significance of these results will be considered in the discussion. 32 Table 2.3. Elastin Thermodynamics Temperature Free Energy Enthalpy Entropy Heal Capacity (°C) (caVmol) (cal/mol) (cal/mol K) (cal/mol K) Pig 1 95 -243 -1.23 25 10 106 -219 -1.15 2.8 20 117 -190 -1.05 3.1 30 127 -157 -0.94 3.4 40 135 -121 -0.82 3.8 50 143 -82 -0.7 4.1 60 150 -38 -0.6 4.5 Turkey 1 94 -295 -1.42 2.7 10 107 -270 -133 2.8 20 119 -241 -1.23 3.0 30 131 -210 -1.13 3.2 40 142 -178 -1.02 3.3 50 152 -144 •0.92 3.5 60 160 -108 -0.81 3.7 Frog 1 100 -195 -1.08 2.6 10 109 -173 •0.99 2.4 20 119 -150 -0.92 22 30 127 -129 -0.85 1.9 40 136 -111 •0.79 1.6 50 143 -96 -0.74 1.3 60 150 -84 -0.71 1.0 Turtle 1 101 -176 -1.01 22 10 110 -156 -0.94 22 20 119 -134 -0.86 2.3 30 127 -110 -0.78 2.3 40 134 -87 -0.71 2.4 50 141 -62 -0.63 2.5 60 147 -37 -0.55 2.5 Salmon 1 103 -64 -0.61 1.5 10 109 -51 -0.56 1.4 20 114 -37 -0.52 13 30 119 -25 -0.47 12 40 123 -13 -0.44 1.1 50 128 -3 -0.40 1.0 60 132 +6 -0.38 0.8 Shark 1 102 -241 -1.25 0.3 10 113 -237 -1.24 0.6 20 125 -228 -1.21 1.0 30 137 •216 -1.17 1.4 40 148 -200 •1.11 1.9 50 159 -179 -1.05 2.3 60 169 -153 -0.97 2.8 CowLig 1 89 -362 -1.65 4.7 10 103 -318 -1.49 5.0 20 117 -267 - U l 53 30 129 -211 -1.12 5.7 40 139 -153 -0.93 6.0 50 148 -91 •0.74 6.4 60 154 -25 -0.54 6.8 33 D. DISCUSSION Although the amino acid compositions of a large number of vertebrate elastins are known, most of the data is for fish and mammalian elastins. In general, fish elastins have larger proportions of polar amino acids, and consequently, lower hydrophobic indices than bird and mammal elastins (Serafini-Fracassini et al, 1978; Spina et al, 1979; Sage and Gray, 1979; Guantieri et al, 1987). Less is known about amphibian and reptile elastin; however, what is known suggests that these elastins have intermediate hydrophobic indices (Sage and Gray, 1979, present study). To lend credence to the present study, it is necessary to establish the accuracy of the purification method. Table 2.4 shows a literature survey of amino acid compositions, calculated hydrophobic indices and purification methods for pig aortic elastin and cow ligament elastin. These sources contain the greatest proportion of elastin known, and thus, they have been studied most extensively. For both elastins, the hydrophobic index (HI) does not change very much even though different purification methods were used. Therefore, I estimate that the maximum errors associated with the calculation of HI should not be more than 2 HI units. This error is sufficiently small that by making the maximum error, the correlations in this study remain robust. The swelling-temperature compensation hypothesis (Gosline and French, 1979) predicts that all elastins should show temperature-dependent swelling. Further, Gosline and French (1979) proposed that the ability to store elastic energy efficiently over a wide temperature range depended on an increase in water content as temperature decreases, a physical property unique to hydrophobic systems. Swelling has been correlated with the weakening of hydrophobic interactions due to the increase in solubility of non-polar groups (Gosline, 1977; Gosline, 1978b). Non-polar groups can be partially hydrated since they tend to minimize their contacts with water. For a network structure, this means that on average, non-polar groups spend less time bonded to 34 Table 2.4. Literature survey of amino acid compositions. Pig Aortic Elastin HI Method of Purification Reference 26.2 Autoclaving/GuHCl Sage and Gray (1981) 26.6 Formic acid/CNBr Rasmussen et al (1975) 27.1 Autoclaving/Formic acid/CNBr Starcher and Galione (1976) 27.6 NaCl/GuHCl/enzymes Starcher et al (1973) x=26.9+/-0.6 Cow Ligament Elastin 37.0 NaOH Stevens and Minns (1974) 34.8 Enzymes Stevens and Minns (1974) 30.1 GuHCl/enzymes Ross and Bomstein (1969) 32.3 Autoclaving/formic acid Starcher and Galione (1976) 33.3 NaCl/GuHCl Guantieri et al (1987) 35.3 NaCl/GuHCl/enzymes Serafini-Fracassini et al (1985) x=33.8+/-2.4 35 water molecules than do polar groups. Tanford (1980) describes the hydrogen bonds between polar groups and water as less labile than hydrogen bonds in pure water. Polar groups should be fully hydrated over the 0 to 60°C temperature range, and therefore, will contribute little to the temperature-dependent hydration changes. On the other hand, interactions between non-polar groups and water are more labile than hydrogen bonds in pure water. Consequently, changes in temperature can strongly affect the hydration of elastin because it has a very high proportion of non-polar groups. Consistent with a network structure of fully mobile molecular chains, the more hydrophobic elastins generally tend to have lower average hydrations than the less hydrophobic elastins (see figure 2.4). In addition, absorption isotherms showed that much of the water is loosely bound and can be removed at high RH. And again, more hydrophobic elastins lose a greater proportion of bound water than do less hydrophobic elastins. At relative humidities below 97%, the slope of the absorption isotherms must level out, suggesting that the rest of the water is probably associated with the peptide backbone and polar amino acid side chains and requires much lower RH to remove it. At water contents lower than 0.35 grams water/gram elastin, Kakivaya and Hoeve (1975) and Ceccorulli et al. (1977) found that water within the elastin network does not freeze at temperature below 0°C, indicating that it is tightly bound. Gosline (1987) found similar results with cow ligament elastin. Further, the correlation between hydration level and HI suggests that one of the main determinants of hydration is the ratio of non-polar groups to polar groups, as described by the hydrophobic index. The temperature-dependence of elastin swelling, as measured by swelling index, correlates well with the arnino acid compositions, as measured by HI (see figure 2.6). The data support the swelling mechanism described by Gosline (1978b). That is, more hydrophobic networks should have a larger swelling index than less hydrophobic networks simply because they are more susceptible to temperature-dependent solubility 36 changes associated with hydrophobic interactions. A low HI means that the network will have a high water content and a low temperature dependence. On the other hand, a high HI means a low water content and a high temperature-dependence. The differences in swelling profiles can also be analysed through the thermodynamics of the mixing processes that are responsible for the swelling (see table 2.3). The standard free energy change for introducing water into an elastin network, AGmix. is positive for all elastins at all temperatures, indicating that the interaction between elastin and water is unfavourable. In other words, water is a poor solvent for elastin (Gosline, 1977, 1978b) and becomes worse as the temperature increases, characteristic of hydrophobic interactions (Nemethy and Scheraga, 1962). Also, the magnitude of the free energy change is greatest for elastins with higher His, indicating that it requires more energy to add a mole of water to the cow ligament or turkey elastin network than to salmon elastin. Interestingly, salmon elastin also has a positive AGm^, which increases with temperature. This implies that salmon elastin has a substantial hydrophobic component to its swelling process, even though it contains a higher relative proportion of polar groups. The entropy of mixing (TASmix) is always negative but becomes less negative with increasing temperature (see table 2.3). These data indicate that water is being organized and that the organization is greatest at low temperatures consistent with a non-polar network increasing its average hydration. Enthalpy changes represent the changes in bonding (Marshall, 1978; Gosline and Rosenbloom, 1984). Negative enthalpy changes mean increased bonding either in number or strength. As with entropy, enthalpy changes become less negative with temperature, consistent with less exposure of the network to water. In fact, Grut and McCrum (1974) showed that swelling, the change in volume with temperature, is proportional to the enthalpy of mixing non-polar groups with water. The temperature dependence of these enthalpy changes indicate that there is a positive heat capacity change and a positive heat capacity change is perhaps the most 37 distinctive thermodynamic feature of hydrophobic interactions (Tanford, 1980; Edelhoch and Osborne, 1976; Tanford, 1970). All the elastins in this study had positive heat capacity changes, suggesting that non-polar group hydration is a dominant process in temperature-dependent swelling. Using these thermodynamic data, one can qualitatively explain why the different elastins show different temperature-dependent swelling profiles. The heat capacity changes for all elastins are positive implying that they should be subject to swelling behaviour typical of hydrophobic systems. Further, amino acid composition data show that non-polar amino acids make up between 40 and 60% of the total composition, excluding glycine which alone makes up between 28 and 37% of the total composition. Thus, in accordance with Gosline's swelling mechanism, all elastins should show temperature-dependent swelling. In fact, all elastins do show strong temperature-dependent swelling, except salmon. Salmon elastin shows very limited temperature-dependence to its swelling. Amino acid composition data show that while there may be only a 20% difference in non-polar amino acid proportion among the elastins, there may be a 300% difference in polar amino acid proportion (see table 2.2). These large differences in polar amino acid content may contribute to unusual swelling characteristics for a network system. Typically, polar group interactions with water have negative heat capacity changes (Tanford, 1980) meaning that the free energy change of mixing will get less positive with increasing temperature. Thus, the swelling tendency for the polar component of the network will be to increase volume with increasing temperature rather than decreasing volume with increasing temperature as hydrophobic systems do. By combining the polar and non-polar components, the opposing heat capacity changes can cancel each other. However, because the relative magnitudes of the polar and non-polar components may not be equal, the network need not show total temperature-independent swelling. Salmon elastin has a small positive heat capacity change and as a consequence shows a very 38 limited temperature-dependence to its swelling. This study has shown that the physical properties of elastin (water contents, swelling behaviour and thermodynamics), with the exception of shark elastin, are related to its biochemical composition, as measured by hydrophobic index (HI). The hydration of salmon elastin shows virtually no temperature-dependence. Since the teleosts for which amino acid composition data are available (Spina et al., 1979; Sage and Gray, 1981; Guantieri et al., 1987) seem to have similar hydrophobic indices, 10+2 for 14 different species, and because salmon elastin fits the general trends observed for the other elastins, it may be that the physical properties of teleost elastin are similar to that measured for salmon elastin. However, the data collected for shark elastin showed two significant anomalies. First, the absolute water content of shark elastin is significantly lower than salmon (see figure 2.4); in fact, it is the lowest of all measured elastins. Second, the swelling index, an independent measure from water content, for shark elastin was similar to pig elastin (see figure 2.6). Based on the water contents and swelling indices, shark elastin should have a HI of about 25 to 40 respectively. However, from its amino acid composition, the HI was similar to salmon elastin, at about 10. Sage and Gray (1979, 1980, 1981) also found similar hydrophobic indices for two other species of sharks, suggesting that the composition data are reasonable. These rather perplexing data suggest that the molecular organization of the shark elastin is somewhat different than that of the other elastins. In fact, the results for shark elastin suggest that the network systems for cartilaginous fish may be different from that of the other elastins. Firstly, there may be a polar impurity lodged in the network but not contributing to the swelling of the shark elastin. This impurity may be occupying space within the network preventing contribution to the swelling. I believe the existence of an impurity is unlikely since other shark elastin amino acid compositions also show high degrees of polarity and have hydrophobic indices of about 10 (Sage and Gray, 1979,1981). 39 Secondly, this anomaly may reflect a difference in the structure of the network chains in the shark elastin network. The theory used in this study assumes that elastin has a structure consisting of kinetically-free chains periodically crosslinked into a molecular network. Further, it assumes that interchain interactions are very weak and essentially negligible, and therefore, every part of the network has the potential of interacting with water equally. However, an elastin with a high proportion of ionizable residues may be subject to stronger, stable interchain interactions, creating areas which are no longer kinetically-free. Such stable, structured regions would reduce the proportion of the network capable of associating reversibly with water and contributing to the swelling process. Figure 2.7 shows the suggested schematic for shark elastin. Structured regions of ionizable polar residues, as represented by hatched lines in figure 2.7, would not contribute to the swelling process because they would be involved in internal hydrogen bonding, and therefore, not available to interact with solvent. Only the random-coil chains, as represented by solid lines, can contribute to the swelling process. The significance of these postulated globular regions is uncertain. It may be that these regions are residual components from an ancestral, elastin-like molecule. Wright (1984) showed that the elastic tissue in Agnathans, the most primitive of the vertebrates, has a filamentous structure. Perhaps the structured regions in shark elastin represent the remnants of this filamentous precursor. If shark elastin has regions of stable, globular structure, chemical perturbations of the network should bring about measurable changes in physical properties. Other workers have shown that hydrophobic polyelectrolytes, aliphatic chains with ionizable functionalities, undergo transition from a compact globular structure to a random-coil network upon ionization in aqueous media (Martin and Strauss, 1980; Sugai and Ohno, 1980). The existence of polar regions of structure within the shark elastin network could be verified by conducting swelling tests over a pH range. Katz et al. (1973) showed that bovine and human hemoglobin are capable of volume changes when the pH was 40 Fig. 2.7. A schematic diagram for a shark elastin network. The circles represent the crosslinking points of the peptide chains, represented by lines. Interchain interactions are indicated by hatched lines. 41 varied under isothermal conditions. The ionization of basic and acidic amino acids are sensitive to pH and I would expect changes in shark elastin water content with changes in pH if these regions of structure exist. Finally, the differences in physical properties seen in this chapter should have significant effects on the mechanics of elastin if the swelling-temperature compensation hypothesis of Gosline and French (1979) is correct. In the next chapter, the mechanics will be tested in order to assess the elastic energy storage efficiency of the physically different network systems described in this chapter. 42 CHAPTER HI.: T H E MECHANICAL PROPERTIES OF V E R T E B R A T E ELASTINS A. INTRODUCTION In a wide variety of tissues, such as skin, ligament and arterial walls, the rubber-like protein, elastin, is subjected to extensions of varying magnitude and rate. In dynamically loaded tissues such as ligaments and arterial walls, elastin must function like an efficient rubber-band. Historically, it has been the ligamentum nuchae of large mammals with big heads and mammalian major arteries which have been studied because these represent sources with the greatest proportion of elastin. For example, elastin makes up about 80% by weight of bovine ligamentum nuchae (Ayers, 1964), and the aortae of pigs may be 60% elastin by weight (Minns and Steven, 1978). However, elastin is present in all vertebrates except cyclostomes (Sage and Gray, 1979) and though it may be more difficult to work with in these lower vertebrates because of relatively lower elastin content and tissue size, elastins from lower vertebrates should be studied in order to gain a complete understanding of this protein's function and evolution. In the circulatory system of vertebrates, elastin must serve two essential functions. One, by acting as an extensible protein rubber, elastin smooths the pressure pulse originating at the heart and thus, reduces the pressure gradient the heart must pump against. Two, by releasing elastic energy stored during the circumferential strain of systole, elastin propels blood through the circulatory network during diastole. Elastin is therefore crucial to proper cardiovascular function. Gosline and French (1979) used cow ligament elastin to assess the efficiency of elastic energy storage. Using dynamic testing, they found elastin was capable of storing energy elastically over a wide temperature range (0 to 70OQ and frequency range (0 to 200 Hz). Further, because elastin increased its water content dramatically (60%) over this temperature range, it was proposed that the swelling behaviour had a plasticizing 43 effect, offsetting any decrease in elastic energy storage efficiency which would occur with decreasing temperature. In a previous paper, Gosline (1978b) showed that the swelling behaviour could be directly attributed to the changing solubility of non-polar groups over the temperature range. Gosline and French (1979) proposed a swelling-temperature compensation hypothesis, in which it was suggested that elastin evolved as a highly hydrophobic protein in order to provide ectothermic vertebrates with a protein capable of storing elastic energy over a wide temperature range. Gosline and French (1979) and Gotte et al. (1968) have used cow ligament elastin for mechanical tests. Cow is an endothermic homeotherm and is not given to a wide range of body temperatures. However, vertebrate body temperatures range from -2°C in antarctic fish (Wilson, 1979) to 42°C in birds (Sturkie, 1976). In order to assess the swelling-temperature compensation hypothesis, the elastins from these vertebrates must be examined. In addition to physiological differences amongst the vertebrates with which elastin must contend, biochemical differences exist across the vertebrate sub-phylum. Sage and Gray (1979, 1980, 1981) used a hydrophobic index to quantify these biochemical differences. They showed that the hydrophobicity of elastin varied with evolutionary advancement, the criterion for advancement being the magnitude of the pressure within the circulatory system. Vertebrates with higher hydrophobic indices had higher circulatory system pressures than lower vertebrates with lower hydrophobic indices. Again, if the hydrophobic hypothesis is to be evaluated, the effect of biochemical diversity on the mechanical properties must be examined. It is the purpose of this study to investigate the ability of elastin from various vertebrates to store energy elastically over a wide frequency and temperature range. From this study, a good evaluation of Gosline and French's hypothesis can be made. In chapter II, it was shown that the physical properties amongst the vertebrate elastins are quite different in accordance with their biochemical differences. Thus, the extreme cases 44 with respect to hydrophobic index, salmon and turkey elastin, were tested mechanically to assess their ability to store elastic energy over a wide temperature range. Pig elastin was also used as a control measurement, since it has been examined in detail elsewhere (Lillie, unpublished results). 45 B. MATERIALS AND METHODS 1. Sample Preparation. Purified elastin samples were cut into small pieces approximately 3-4 mm wide, 1 mm thick and 4-5 mm long. They were then held flat and dehydrated in 100% ethanol for several hours. Note, there was no significant difference in mechanical behaviour between ethanol-treated tissue and untreated tissue. The samples were then carefully trimmed with a razor blade and potted with epoxy into grooves cut into aluminum mounts. The elastin was finally rehydrated by autoclaving for 20 minutes. Some tissue was stored in distilled water for several days and autoclaved again prior to use, and no changes in mechanical properties were observed. For open system dynamic tests, the mounted samples of turkey, salmon and pig were used straight from distilled water. For closed system dynamic tests, mounted samples of turkey elastin were placed in sterilized glass canning jars over a 0.83M NaCl solution. This solution gives a 97% relative humidity at 37oC. The canning jars were sealed and placed in a constant temperature chamber (370C + 0.5OQ for 10 days. Along with the mounted specimen, three pieces of similar tissue were placed in the canning jar in order to know the approximate water content (g water/g elastin) of the dehydrated sample. These pieces of tissue were treated as described in chapter 2 and found to have water contents between 0.30 and 0.35 g water/g elastin, similar to absorption isotherm data. Samples were tested while immersed in mineral oil so that no water could be absorbed into the elastin network as the temperature was decreased. Closed system then refers to a elastin sample which is tested at constant water content. 2. Dynamic Tests. The theory and operation of the apparatus used in this study has been well documented (Gosline and French, 1979; Denny and Gosline, 1980; Shadwick and Gosline, 1985; DeMont and Gosline, 1988). The length, width and thickness of each 46 sample was measured with vernier calipers before every experiment. The sample was pre-strained by 20% and suspended between a strain gauge force transducer and an electromagnetic vibrator with attached displacement transducer. Preliminary experiments showed that the mechanical properties did not change between pre-strain values of 20 and 80%. The mounting of tissue is difficult and by using the pre-strain value of 20%, the probability of breaking the sample is lower. The vibrator was driven by a noise function generator (Wavetek Cross Channel Spectrum Analyser, model 5820A, Rockleigh, New Jersey) which deformed the sample in tension with a nominal strain of 4% peak-to-peak over a frequency span of 0 to 200 Hz. The noise generator showed a constant power spectrum over this frequency range. Resonances of unknown origin present in the testing apparatus limited the experimental frequency span to about three decades. At each frequency of forced vibration, both the ratio of Fourier coefficients of the force and displacement signals and the phase angle (d) between them were calculated by the spectrum analyser. Stress (force/cross-sectional area) and strain (change in length/original length) were calculated by a Digital PDP 11-23 computer from measured sample dimensions. The ratio of stress to strain gives the complex dynamic modulus, E * . This complex modulus can be broken down into two components: the storage modulus, E ' , and the loss modulus, E " . The storage modulus is a measure of how much energy is stored per cycle, whereas the loss modulus measures how much energy is lost per cycle. Both moduli were calculated as follows: E ' = E*(cosd) (1) E " = E*(sind) (2) The tangent of the phase angle (tan d) is a useful measure of how efficiently 47 energy is stored, since it provides an indication of the ratio of energy lost through viscous processes to energy stored through elastic processes. Tan d is often referred to as damping and can be calculated as follows: tand = E " / E ' (3) The resilience or elastic efficiency for the three elastin types can be calculated using the following equation (Wainwright etal, 1976): Re/100 = exp(-nxtand) (4) Resilience (Re) is defined as the percentage of the total energy applied to a system which is recovered and subsequently available to power the elastic recoil. 3. Time-Temperature Superposition Principle. The data were collected over a temperature range of 0°C to about 65°C in approximately 5°C increments. The sample was held at each temperature for about 20 minutes and assumed to be in swelling equilibrium. Swelling equilibration experiments show that elastin attains swelling equilibrium in 20 minutes or better following a step increase or decrease in temperature from 1°C to 65°C. Further, the data were combined according to the time-temperature superposition principle. The equivalence of time and temperature has been well documented (Ferry, 1970; Aklonis et al., 1972; Wainwright et al., 1976; Gosline and French, 1979; Dorrington, 1980). Briefly, molecular chains require both energy and time to move or respond to external forces because they must overcome frictional forces. For peptide chains with non-polar amino acids, frictional forces consist of hydrogen bonds between backbone peptide groups and polar side chains, and to a lesser extent, hydrophobic interactions between non-polar side chains. These frictional forces represent energy barriers to 4 8 molecular chain movement. At a given temperature, and thus, a given amount of energy, molecular chains require a certain time to overcome energy barriers to movement. Time is required because the chains overcome these barriers in a statistical manner. With shorter time periods, the chains have less chance of overcoming the barrier. In fact, if the times are too short, the chains will not respond at all, and a rubber-like material will behave like a rigid glass. With longer times, random thermal motion allows the chains to respond, and the material to deform elastically. At a given time period or rate of deformation, changes in temperature alter the amount of energy the molecular chains possess. Thus, at low temperatures the chains will behave as though the time period was short and at higher temperatures the chains will respond as if the time period was long. This equivalence of time and temperature is known as the time-temperature superposition principle. In this study, the time-temperature superposition principle was used indirectly to extend the experimental time span over a much wider range of frequencies than could be measured directly. For example, the properties at low frequency or long time periods can be observed by increasing the temperature and vice versa. The experimental apparatus had resonances of unknown origin above 200 Hz. Therefore, the time-temperature superposition principle was used to expand the experimental frequency range by several orders of magnitude. In temperature shifting of dynamic tests, modulus versus frequency curves obtained at different temperatures are shifted horizontally until they overlap, and the horizontal or time shift factor (log aT) is determined. Log aT represents the shift in time by the molecular chains caused by a change in temperature in order to maintain equivalent mechanical behaviour. 49 C. RESULTS 1. Closed System. Figure 3.1 shows a typical plot of dynamic data for closed system turkey elastin hydrated at 97% relative humidity at 37°C. Recall that closed system means that the elastin is held at a fixed water content, thus, it is not able to absorb water and swell as the temperature is lowered. The dynamic data consist of log moduli and tan d values as a function of log frequency. The spectrum analyser samples the 0 to 200 Hz frequency range at 200 points, and in this figure only half of these points are plotted. Therefore, these points can be considered to be lines without resorting to complex statistics. Figure 3.1 shows typical behaviour for a polymer approaching its glass transition (Tg). With increasing frequency, the storage modulus (E') and the loss modulus (E") both increase and tend to converge. The point of convergence, when the storage and loss modulus are equal, is the mid-glass transition. The bottom panel in figure 3.1 shows the elastic efficiency (tan d) decreasing with increasing frequency. Figure 3.2 illustrates the use of the time-temperature superposition principle. The lines in the left panel show storage modulus (E') versus frequency data as shown in figure 3.1, at four of a possible 8 to 10 different temperatures. Notice that the modulus increases with decreasing temperature. In the right panel, the lines are shifted horizontally until they overlap by at least one decade of frequency, creating a master curve with a reference temperature of 37°C. This curve predicts the mechanical properties of turkey elastin over the frequency range of 102 to 106 Hz. Figure 3.3 shows the mechanical shift (log aT) data for turkey elastin plotted as a function of temperature. This curve was constructed by shifting the moduli curves at each temperature horizontally according to the time-temperature superposition principle. The total mechanical shift is about 6 decades of frequency between 0 to 40°C which is similar to the results of Gosline and French (1979) for cow 50 TEMP =37 .0 LOG FREQUENCY Fig. 3.1. Typical dynamic data for turkey elastin used to generate the shifted data curve in figure 3.2. 50a Fig. 3.2. The storage modulus data for turkey elastin plotted as a function of frequency used to illustrate the time-temperature superposition principle. The modulus data from 4 of a possible 8 to 10 temperatures are plotted and shifted from a reference temperature of 37°C. Data i i i i - 1 0 1 2 Shifted Data 52 CO o 20 40 Temperature (°C) 6 0 S;t?n3/m-MeC?^Cal S h l ? d a ti Plotted as a function of temperature for turkey system) a t 3 W 3 t e r 0 0 1 1 1 6 1 1 1 ° f 0 3 0 g w a t e r / g e l a s t i n ( c l o s e d 53 ligament elastin and Lillie (unpublished results) for pig elastin. In other words, a 40°C temperature change corresponds to a one million-fold change in time scale. 2. Open System. Figure 3.4 shows typical open system modulus data for salmon, turkey and pig elastin given at 5°C and 40°C. These data can be shifted horizontally using the time-temperature superposition principle just like closed system turkey elastin to produce log aT curves as a function of temperature. Figure 3.5 shows the log aT curves plotted as a function of temperature for salmon, pig and turkey elastin tested as open systems; that is, in swelling equilibrium. All three elastin types had mechanical shifts that were essentially linear functions of temperature and showed a total temperature dependence of one decade of frequency over the 0 to 40°C temperature range. Thus, the temperature dependence of the closed system turkey elastin is about 100,000 times greater than the turkey elastin in swelling equilibrium. These results contradict the swelling-temperature compensation hypothesis. Salmon elastin does not need to swell like turkey or pig elastin in order to be mechanically temperature-independent. 3. Mechanical Efficiency of Salmon and Turkey Elastin. Figure 3.6 shows the tan d plotted as a function of temperature at low (1 Hz) and high (100 Hz) frequency to demonstrate the effect of temperature on the elastic efficiency of the elastins. At low frequency, turkey and salmon elastins both show low damping indicated by the very low tan d values (0.07 or less) which increase slightly as temperature decreases. The damping indicates that the resilience at 1 Hz for both elastins is greater than 80% at all temperatures. At 100 Hz, the data for turkey and salmon elastin were fitted to second order polynomial equations. The tan d lines for turkey elastin reaches a maximum value of 0.1 at 20°C whereas salmon increases continuously with temperature 54 TURKEY - i . o. i. i L O G F R E Q U E N C Y _L _ J -1. 0. I . 2. L O G F R E Q U E N C Y 4a] SALMON L O G F R E Q U E N C Y L O G F R E Q U E N C Y Fig. 3.4. Typical data for turkey and salmon elastin at 5°C and 40°C in equilibrium with water. 54a Fig. 3.5. Mechanical shift data for salmon (•, N=5, n = 47), turkey (•, N=4, n=36) and pig elastin (o, N=5, n=49) tested m swelling equilibrium with water (open system) and fitted to a linear regression (r^  = 0.83, 0.95, 0.90, respectively). 1 0 n. o • H 2H 0 1-£ OH -2-2^ Temperature (°C) • S °1 -H Salmon i i i i i • 10 20 30 40 50 60 Temperature (°C) ~70 Turkey I I I | | | T — 10 20 30 40 50 60 70 Pig -2-* — — r — i — - i 1 i i i 0 10 20 30 40 50 60 70 Temperature («C) 55a Fig. 3.6. 100 H z (dark symbols) and 1 H z (open symbols) mechanical efficiency (tan d) data for turkey (squares; N = 4, n = 39), and salmon (triangles; N = 4, n=36). The 100 H z data were fitted to a second order polnomial equation while the 1 H z data were fitted to a linear equation. Temperature (•€) 57 to a maximum of 0.13 at IOC. Again, even at 100 Hz, both elastins are capable of storing energy efficiently, as the calculated resiliences are about 65 to 70%. Interestingly, the peak for turkey elastin coincides with the temperature where dramatic swelling occurs, suggesting that swelling may be over-compensating for increases in damping. 58 D. DISCUSSION The swelling-temperature compensation hypothesis of Gosline and French (1979) states that in order to exhibit temperature-independent mechanical properties, elastin must be a very hydrophobic protein. By being hydrophobic, elastin will demonstrate a dramatic temperature-dependent swelling (Chapter 2; Gosline, 1977; Gosline, 1978b), in which an increase in water content is brought about by the solubility changes of non-polar amino acid side chains causing the weakening of hydrophobic interactions as the temperature is lowered. This increase in water content with decreasing temperature allows water molecules to reduce frictional forces and offset any reduction in mechanical efficiency which would otherwise occur with temperature. Thus, the hypothesis suggests that the hydrophobic nature of elastin evolved in order to provide vertebrates with protein capable of storing elastic energy efficiently over a wide temperature range. In chapter 2, swelling tests show that lower vertebrate elastins are not necessarily all highly hydrophobic systems. Salmon shows a very reduced temperature-dependent swelling, but it maintains a high water content at all temperatures. The amino acid compositions, as described by hydrophobic index, of various vertebrate elastins from Sage and Gray (1979, 1980, 1981), as well as those done in chapter 2, correlated well with the swelling results. The strong correlation between swelling and hydrophobic index forces the rejection of Gosline and French's swelling-temperature compensation hypothesis because ectothermic lower vertebrates which would need a temperature independent elastin achieve this property without being hydrophobic. This leaves us with the question: how do these biochemical differences affect the mechanical properties of elastin? Fish elastins are not hydrophobic, but the results of this chapter indicate that all elastins, regardless of hydrophobic index, are capable of displaying mechanical efficiency over a wide temperature range. Elastin, like all amorphous polymers, exhibits three regions of molecular 59 behaviour: glass-like, rubber-like and a transition state (Hoeve and Flory, 1974; Kakivaya and Hoeve, 1975; Wainwright et al., 1976; Gosline, 1980). Figure 3.7 shows a diagram of the three mechanical regions typical of amorphous polymers indicated by dynamic modulus at constant frequency as a function of temperature. At the far left of figure 3.7, the polymer is in its rigid, glass-like state. Thermal energy is low and insufficient to surmount the frictional forces to movement, and so, there is no chain mobility. As indicated by tan d values, or damping, in figure 3.7, the efficiency of a polymer in its glass-like state is reasonably high since frictional forces are strong enough so that no energy is lost in overcoming them. As the temperature increases, chain mobility begins due to increasing thermal energy. However, large frictional forces still exist which require large amounts of energy to overcome. Thus, the mechanical efficiency is low as indicated by the increase in the damping on figure 3.7. The polymer is in the transition state or glass transition, referring to the point when a mechanical response shows a dramatic increase in frictional forces. From figure 3.7, the glass transition is characterized by large increases in damping and about a thousand fold decrease in stiffness. Thus, in the glass transition, mechanical properties are very temperature-dependent. Further increases in temperature supply the chains with adequate energy to overcome frictional forces. The chains are now fully mobile, and the polymer is located somewhere in the rubber-like plateau region as illustrated in figure 3.7. The mechanical efficiency is high, as reflected in low tan d values, and the mechanical properties are essentially temperature-independent, as reflected by a temperature-independent modulus. The mechanical region of a polymer is dependent on polymer chemistry, solvent content and temperature. The glass transition of elastin is known to be water content dependent (Kakivaya and Hoeve, 1975). For example at 0.30 g water/g elastin, the glass transition temperature (Tg) is approximately 10°C whereas at 0.20 g water/g elastin, Tg is 35°C. When hydrated at 36°C and 55oC under 99.8% relative 60 Temperature (°C) Fig. 3.7. Typical mechanical behaviour for an amourphous polymer is shown with the storage modulus and tan d values plotted as function of temperature (solid line). The temperature axis can be replaced with frequency according to the time-temperature superposition principle. The three characteristic regions (glass-like state, transition state, and rubber-like state) are shown. The black circles show 100 Hz turkey elastin data at a fixed water content (0.30 g water/g elastin). 61 humidity conditions, Gosline and French (1979) calculated a Tg of -45°C and -25°C for water contents of 0.46 and 0.41 g water/g elastin respectively. Clearly, the dramatic swelling of elastin which occurs with decreasing temperature increases the water content sufficiently to delay the onset of the glass transition and keep the polymer in its rubbery region. When tested as closed systems, at fixed water contents, cow ligament (Gosline and French, 1979), pig (Lillie, unpublished results) and turkey elastin (fig. 3.1) all entered the glass transition. Since water contents were fixed and no water could enter the network, the mechanical properties were highly dependent on temperature, as indicated by mechanical shifts of about 6 decades of frequency (see fig. 3.3) and a one hundred fold increase in modulus (see fig. 3.7, black circles). Further, they all demonstrated low resiliences at low temperatures with damping values approaching 1.0 in the mid-glass transition region (see fig. 3.7, black circles). These results show that the hydrophobic elastins from higher vertebrates are operating close to their glass transitions and would have a very low mechanical efficiency if the swelling-temperature compensation mechanism were not available as the temperature decreased. When tested as open systems, in swelling equilibrium with water, turkey and pig elastin showed limited mechanical shift and turkey elastin was shown to have a high elastic efficiency over the entire biological temperature range, 0 to 40°C. These hydrophobic systems are operating close to their glass transitions (fig. 3.2, 3.7) and only an increase in temperature or water content will delay the onset of the glass transition. As the temperature decreases, the temperature-dependent swelling increases the water content and actually takes the elastins away from their glass transitions. The increase in water content effectively creates a new material, with a curve similar to the solid line in figure 3.7, but shifted to the left away from its glass transition. For direct comparison, figure 3.8 shows the 1 Hz and 100 Hz damping of open and closed system turkey 62 Temperature (°C) Fig. 3.8. Comparison of the effects of temperature on the mechanical efficiency (tan d) of open ( J ) and closed ( g ) system turkey elastin at 100 Hz (squares) and 1 Hz (circles). 63 elastin as a function of temperature. The elastic efficiency is high for turkey elastin when tested as an open system but very low when tested as a closed system. Figure 3.8 illustrates the essence of the swelling-temperature compensation hypothesis. A hydrophobic system like turkey elastin operates close to its Tg and must swell dramatically with decreasing temperature in order to maintain a high elastic efficiency at low temperatures. While this may not be important for a turkey or pig since they maintain a constant body temperature, hibernating mammals or birds which enter periods of torpor may require swelling compensation. Salmon elastin, from chapter 2, shows a very reduced temperature-dependent swelling, although, its total water content is high over the entire temperature range. Thus, salmon elastin displays mechanical shifts similar to those of pig, turkey and cow elastins, which increased their water contents by 35, 40 and 60% respectively over this temperature range. Salmon elastin is an amorphous system which is not nearly as hydrophobic. Yet, when the temperature decreases, it demonstrates comparable mechanical properties. These results suggest that the salmon elastin, by virtue of its high water content, is in the rubbery plateau region and is farther away from its glass transition than the other elastins (refer to fig. 3.7). Thus, salmon elastin does not require a swelling compensation mechanism. Since both lower and higher vertebrate elastins are capable of achieving mechanical efficiency over a wide temperature range, the results suggest two possible molecular mechanisms for ensuring this elastic efficiency. Higher vertebrates, with a high hydrophobic index, have a low water content, and as a consequence, function close to their Tg. Increasing their water contents through temperature-dependent swelling, which offset increasing frictional interactions with decreasing temperature, is thus an essential mechanism for maintaining mechanical efficiency at low temperatures. From the results of chapter 2, not all elastins must necessarily possess a hydrophobic nature. Salmon elastin showed mechanical efficiency, yet, it was not 64 a hydrophobic system like turkey or pig elastin. Salmon elastin demonstrated a second molecular mechanism for strain energy storage efficiency with decreasing temperature. Salmon elastin, having a higher percentage of polar residues, maintained a high water content at all temperatures. This shifts the material well away from its Tg and therefore, it did not require the dramatic increase in water content through swelling. So, if the hydrophobic nature of elastin did not evolve in order to ensure temperature-independent mechanical efficiency, what is the driving force for elastin hydrophobic diversity? Sage (1982) observed a good correlation between circulatory system pressure and elastin hydrophobic character. Sage pooled hydrophobic indices for all representatives within each vertebrate class and calculated an average hydrophobic index for the class. Also, I have selected mean systolic pressure as an indicator of circulatory system pressure. The numbers given for mean systolic pressure are recognizably variable. Figure 3.9 shows mean systolic pressure as a function of hydrophobic index. Higher vertebrates tend to have higher hydrophobic indices than lower vertebrates, which was established in chapter 2. Gosline (1978a) showed that the elastic recoil mechanism of elastin had a substantial hydrophobic component. Briefly, when the elastin chains are pulled apart, hydrophobic groups are forced to increase their average hydration. This unfavourable situation has a positive free energy change, thus, energy can be stored and used to power the elastic recoil. From elastin sequence data, a large proportion of the network chains are made up of hydrophobic amino acid residues (Gosline and Rosenbloom, 1984). It seems plausible that a less hydrophobic elastin would have a smaller hydrophobic contribution to the elastic recoil than a more hydrophobic elastin and vice versa, assuming that the flexible chains are composed of high proportions of hydrophobic amino acids. 65 0 10 20 30 40 Hydrophobic Index Fig. 3.9. The relationship between systolic circulatory pressures and elastin hydrophobic index. The hydrophobic indices, taken from Sage and Gray (1981) and this study, represent means of different vertebrate classes (birds,*, n=4; mammals,«, n = 9; reptiles,A,n = 2; amphibians,o,n = 2; teleosts.a, n = l l ; non-teleosts,A, n = 4). Systolic pressure values were taken from Prosser (1973) and are recognizably variable. 66 According to the law of LaPlace, tension is equal to the pressure within the vessel multiplied by the radius of the vessel. Therefore, tension within the arterial wall may be directly affected by both circulatory pressure and vessel size. Assuming changes in wall tension require modification of the arterial wall, three options exist which could counteract changes in arterial wall tension. The first option is to regulate the amount of material within the arterial wall. The second option is to regulate the type of material within the arterial wall. A third option of course is a combination of the first and second. Within a given vertebrate class, evidence suggests that the differences in wall tension with animal size are overcome by changing the number of lamellar units in the arterial wall. A lamellar unit consists of a concentric layer of smooth muscle and collagen separated by an elastin lamina (Stehbens, 1979). In a comparative study of mammals, Wolinsky and Glagov (1967) found that the tension per lamellar unit remains constant for 10 different species of mammals. Thus, if the arterial wall has greater tension, the number of lamellar units increases in order to compensate. Moreover, they found that the number of lamellae tapered off at about 70 at the maximum distending pressure of 130 mm Hg. The area of the vessel scales with mammalian size (Schmidt-Nielsen, 1985), therefore, variation in circulatory pressure is also compensated for by increasing the amount of material, ie. the number of lamellar units. In fact, table 3.1 shows that the mean hydrophobic index for 9 different mammalian species does not vary by more than two HI units suggesting that it is the amount of elastin within the wall and not the elastin biochemistry that is variable. Between vertebrate classes, however, it may be that the differences in relative blood pressures between vertebrates of comparable size may have been too great to cope with by simply increasing the number of lamellar units. Further, the hydrophobic indices for representative vertebrate classes, given in table 3.1, shows that while the HI within a class does not change very much, the HI between classes is considerably variable. I propose that there has been evolutionary selection towards changing the molecular 67 Table 3.1. Hydrophobic indices (HI) and Mean systolic pressures (MSP) of representative vertebrate classes. Vertebrate Class HI+/-S.E. MSP Mammalia (9) 23+A2 120 Aves (4) 33+/-2 150 Reptilia (2) 20+/-3 80 Amphibia (2) 16+/-1 30 Osteichthyes (11) 11+/-3 40 Chondricthyes (4) 9+/-2 30 HI values from Sage and Gray (1981). MSP values from Prosser (1973). The number in brackets represents the number of individuals used for each class. 68 structure of the elastin, as reflected by HI, rather than the amount of material. Thus, in order to cope with changing circulatory system pressure, the molecular structure was changed so that a given amount of elastin in a bird would have a greater elastic recoil than a mammal of comparable size. It is however, conceivable that the architecture of the lamellar units in other vertebrate classes is different. Sage and Gray (1980), in a histological study, found that of all vertebrates containing elastin, only teleosts did not have the concentric lamellae. Instead, teleost elastin is arranged in loops and whorls which may or may not have the same basic structural organization. While these statements are speculative, they indicate that further research is required for a full understanding of elastin evolution in the lower vertebrates. The above statements lead to the formulation of a testable hypothesis. Since fish elastin has a lower hydrophobic index than higher vertebrate elastin, then, they should show a greater proportion of elastin per unit wall tension than an equivalent bird or mammal. This study has shown that elastins with very different hydrophobic characters are capable very high elastic efficiencies when tested mechanically. Calculated resiliences indicate that at a strain rate of 1 Hz, 80% of the input energy can be stored to power the elastic recoil. Indeed, only at high frequencies (100 Hz) at very low temperature were there any significant differences between the elastins. The highly hydrophobic turkey and pig elastins are capable of higher resiliences at low temperatures than salmon elastin since the former adopt the swelling-temperature compensation mechanism. 100 Hz frequencies represent the very extreme range and so the loss of resilience is probably insignificant for most vertebrates as far as physiological function is concerned. The highest measurements of heart rate are about 22 Hz in birds (hummingbirds, Lasiewski, 1964) and small mammals (shrews, Morrison et al., 1959). The fact that elastin can still 69 function efficiently at high frequencies says much for the elegant design of this biological material. Only vertebrates with higher blood pressures, like birds and mammals, may have high frequency vibrations set up within the arterial wall as a result of turbulent blood flow (Roach, 1979). Thus, elastin may have to contend with vibrations which are orders of magnitude greater than the heart rate. In conclusion, elastin is capable of storing elastic energy efficiently over a wide temperature range, regardless of its hydrophobic index or swelling profile. 70 CHAPTER IV. GENERAL CONCLUSIONS The aim of this study was to examine the validity of the swelling-temperature compensation hypothesis as proposed by Gosline and French (1979). This study supplied three lines of data which forces the rejection of the hypothesis. First, amino acid composition data, as measured by hydrophobic index, from this study and the literature, shows that the hydrophobic nature of elastin varies amongst the vertebrates. Higher vertebrates have a larger hydrophobic index than lower vertebrates. By themselves, these results do not contradict the hypothesis because the relationship between elastin biochemistry and physical chemistry had not been established. Thus, all elastins may have sufficient hydrophobic character to swell in a temperature-dependent manner, consistent with the hypothesis. Second, the relationship between elastin amino acid composition, as measured by hydrophobic index, and physical chemistry, swelling and water content, was established. In general, both the swelling behaviour and water content showed a strong inverse correlation with the hydrophobic index. In contrast to the other elastins, shark elastin has a low hydrophobic index, similar to salmon, but shows the lowest water content of all elastins tested and a dramatic temperature-dependent swelling profile. In fact, the measured hydrophobic index was about 10 but the predicted hydrophobic index would be at least 25. Based on these results, a network structure for shark elastin consisting of regular, random-coil network interspersed with stable, globular structure is proposed. Finally, the elastins with the most different physical chemistries, namely turkey and salmon elastin, were tested dynamically in order to estimate their mechanical efficiency over a wide range of temperatures, 0 to 60°C. Pig elastin was also tested as a control since its dynamic mechanical behaviour is known. At biological frequencies of 1 Hz, turkey, salmon and pig elastin show equivalent mechanical behaviour, with efficiencies of 80% or better. At high frequency, 100 Hz, these three elastins show 71 efficiencies of about 65 to 70% with pig and turkey showing marginally higher efficiency than salmon. Considering these three bodies of data, the swelling-temperature compensation hypothesis must be rejected. The evolution of elastin towards increasing hydrophobic character or higher hydrophobic index was not driven by the requirement for efficiency since a lower vertebrate elastin was equally capable of efficient spring-like behaviour without having high hydrophobic character. Further, an alternate hypothesis is suggested for the evolution of elastin based on the correlation between vertebrate circulatory pressures and hydrophobic indices. The requirement for a protein which can exert a greater elastic force per unit material may have driven the evolution of elastin toward a higher hydrophobic index to meet the demands for increasing circulatory pressures. 72 REFERENCES Aaron, B.B. and Gosline, J.M. (1980). Optical properties of single elastin fibres indicate random protein conformation. Nature, 287 (5785), 865-867. Aaron, B.B. and Gosline, J.M. (1981). Elastin as a random-network elastomer: a mechanical and optical analysis of single elastin fibres. Biopolymers, 20, 1247-1260. Aklonis, J.J., MacKnight, W.J. and Shen, M. (1972). Introduction to Polymer Viscoelasticity. Toronto: John Wiley and Sons, Inc. Ayers, J.P. (1964). Elastic tissue. Int. Rev. Connect. Tissue Res., 2, 33-100. Berne, R.M. and Levy, M.N. (1987). Physiology. Second edition. St. Louis: The C.V. Mosby Company. Biedka, R.F., Gosline, J.M. and DeWreede, R.E. (1987). Biomechanical analysis of wave-induced mortality in the marine alga Pterogophora californica. Mar. Ecol. Prog. Ser., 36,163-170. Bigelow, CC. (1967). On the average hydrophobicity of proteins and the relation between it and protein structure. J. Theoret. Biol., 16,187-211. Cantor, CR. and Schimmel, P.R. (1980). Biophysical Chemistry. Part I. The conformation of biological macromolecules. San Francisco: W.H. Freeman and Company. Ceccorulli, G.M., Scandola, M. and Pezzin, G. (1977). Calorimetric investigation of some elastin-solvent systems. Biopolymers, 16,1505-1512. DeMont, M.E. and Gosline, J.M. (1988). Mechanics of jet propulsion in the hydromedusan jellyfish, Polyorchis penicillatus. J. Exp. Biol., 134, 313-332. Denny, M.W. and Gosline, J.M. (1980). The physical properties of the pedal mucus of the terrestial slug, Ariolimax columbianus. J. Exp. Biol., 88, 375-393. 73 Dorrington, K.L. (1980). The theory of viscoelasticity in biomaterials. In: The Mechanical Properties of Biological Materials, J.F.V. Vincent and J.D. Curry (eds), 289-314. Cambridge: Cambridge University Press. Edelhoch, H. and Osborne, J.C. (1976). The thermodynamic basis of the stability of proteins, nucleic acids, and membranes. Adv. Prot Chem., 30,183-250. Eyre, D.R., Paz, M.A. and Gallop, P.M. (1984). Crosslinking in collagen and elastin. Ann. Rev. Biochem., 53,717-748. Ferry, J.D. (1970). Viscoelastic Properties of Polymers. New York: John Wiley and Sons, Inc. Fleming, S.S., Sullivan, C.E. and Torchia, D.A. (1980). Characterization of molecular motions in Cl3-labelled elastin by 13C-1H magnetic double resonance. Biopolymers, 19, 597-618. Flory, P.J. (1953). Principles of Polymer Chemistry. London: Cornell University Press. Gosline, J.M., (1976). The physical properties of elastic tissue. In: Int. Rev. Conn. Tissue Res., 7,211-249. New York: Academic Press. Gosline, J.M. (1977). The role of hydrophobic interaction in the swelling of elastin. In: Elastin and Elastic Tissue, L.B. Sandberg, W.R. Gray and C. Franzblau (eds), 621-631. New York: Plenum Press. Gosline, J.M., (1978a). Hydrophobic interaction and a model for the elasticity of elastin. Biopolymers, 17, 677-695. Gosline, J.M., (1978b). The temperature-dependent swelling of elastin. Biopolymers, 17, 697-707. Gosline, J.M. and French, C.J. (1979). Dynamic mechanical properties of elastin. Biopolymers, 18, 2091-2103. Gosline, J.M. (1980). Elastic properties of rubber-like protein and highly extensible tissues. In: Mechanical Properties of Biological Materials, J.F.V. Vincent and J.D. Currey (eds), 331-357. Cambidge: Cambridge University Press. 74 Gosline, J.M. and Rosenbloom, J. (1984). Elastin. In: Extracellular Biochemistry, K.A. Piez and A.H. Reddi (eds), 191-227. New York: Elsevier. Gosline, J.M. (1987). Structure and mechanical properties of rubber-like proteins in animals. Rubber Chem. and Technol., 60(3), 417-438. Gotte, L., Mammi, M. and Pezzin, G. (1968). Some structural aspects of elastin displayed by x-ray diffraction and other physical methods. In: Symposium on Fibrous Proteins, Crewther, W.G. (ed), 236-245. Butterworth: Australia. Green, R.W. (1948). The adsorption of water vapour on collagen and elastin. Trans. Royal Soc. New Zealand, 77 (part 1), 24-46. Grut, W. and McCrum, N.G. (1974). Liquid drop model of elastin. Nature, 251,165. Guantieri, V., Jaques, A.M., Serafini-Fracassini, A. and Tamburro, A.M. (1987). Conformational transitions of solubilized trout elastins. Biopolymers, 26, 1957-1963. Hoeve, C.A.J, and Flory, P.J. (1974). The elastic properties of elastin. Biopolymers, 13, 677-686. Kakivaya, S.R. and Hoeve, C.A.J. (1975). The glass point of elastin. Proc. Nat. Acad. Sci., 72(9), 3505-3507. Katz, S., Roberson, L.C. and Crissman, J.K. (1974). Structure-volume relationships for proteins. Dilatometric study of the reaction of bovine hemoglobin with sodium dodecylsulfate, hydrogen and hydroxylions. Biochim. Biophys. Acta, 342,4-14. Lansing, A.L, Rosenthal, T.B., Alex, M. and Dempsey, E.W. (1952). The structure and chemical characterization of elastic fibers as revealed by elastase and electron microscopy. Anat. Records, 114, 555-575. Lasiewski, R.C. (1964). Body temperatures, heart and breathing rate, and evaporative water loss in hummingbirds. Physiol, zool., 37,212-223. 75 Lyerla, J.R. and Torchia, D.A. (1975). Molecular mobility and structure of elastin deduced from the solvent and temperature dependence of 13C magnetic resonance relaxation data. Biochemistry, 14, 5175-5183. Marshall, A .G. (1978). Biophysical Chemistry. Principles, Techniques and Applications. New York: John Wiley and Sons. Martin, P.J. and Strauss, U.P. (1980). Thermodynamics of hydrophobic polyacids. Biophys. Chem., 11, 397-402. McDonald, D.A. (1974). Blood Flow in Arteries. London: Edward Arnold. Minns, R.J. and Steven, F.S. (1978). The effect of calcium on the mechanical behaviour of aorta media elastin and collagen. Br. J. Exp. Path., 58, 572-579. Mistrali, F., Volpin, D., Garibaldo, G.B. and Ciferri, A. (1971). Thermodynamics of elasticity open systems. Elastin. J. Physical Chem., 75(1), 142-149. Morrison, R.H., Ryser, A.B. and Dave, J.K. (1959). Studies on the physiology of the masked shrew Sorex cinereus. Physiol, zool., 32, 256-271. Nemethy, G. and Scheraga, H.A. (1962). The structure of water and hydrophobic bonding in proteins. III. The thermodynamic properties of hydrophobic bonds in proteins. J. Prot. Chem., 66, 1773-1789. Prosser, C L . (1962). Comparative Animal Physiology, vol. 1. Philadelphia: W.B. Saunders Co. Queslel, J.P. and Mark, J.E. (1986). Elasticity. In: Encyclopedia of Polymer Science and Engineering, 5, 365-408. Roach, M . (1979). Poststenotic dilatation. In: Hemodynamics and the Blood Vessel Wall, W.E. Stehbens (ed), 439-465. Springfield: Charles C. Thomas Publishers. Ross, R. and Bomstein, P. (1969). The elastic fiber. I. The separation and partial characterization of its molecular components. J. Cell Biol., 40, 366-381. Sage, H. and Gray, W.R. (1979). Studies on the evolution of elastin-I. Phylogenetic distribution. Comp. Biochem. Physiol., 64B, 313-327. 76 Sage, H. and Gray, W.R. (1980). Studies on the evolution of elastin-II. Histology. Comp. Biochem. Physiol., 66b, 13-22. Sage, H. and Gray, W.R. (1981). Studies on the evolution of elastin-DI. The ancestral protein. Comp. Biochem. Physiol., 68b, 473-480. Sage, H. (1982). Structure function relationships in the evolution of elastin. J. Invest. Dermatology, 79,146s-153s. Sandberg, L.B. (1976). Elastin structure in health and disease. In: Int. Rev. Conn. Tissue Res., 7,159-210. Scandola, M. and Pezzin, G. (1980). Density of elastin-water system. In: Water in Polymers, S.P. Rowland (ed), 225-234. Americal Chemical Society symposia series. Scatchard, G., Hamer, WJ. and Wood, S.E. (1938). Ionic solutions I. J. Amer. Chem. Soc, 60, 3061-3070. Serafini-Fracassini, A., Field, J.M., Spina, M., Garbisa, S. and Stuart, R.J. (1978). The morphological organization and ultrastructure of elastin in the arterial wall of trout (Salmo gairdneri) and salmon (Salmo salar). J. Ultrastructural. Res., 65, 1-12. Serafini-Fracassini, A., Field, J.M. and Campbell, A.M. (1985). Quaternary structure of elastin: characterization of multichain peptide fragments obtained by elastase digestion. Biopolymers, 24,1515-1526. Shadwick, R.E. and Gosline, J.M. (1985). Mechanical properties of the octopus aorta. J. Exp. Biol., 114, 259-284. Schmidt-Nielson, K. (1985). Scaling - Why is Animal Size So Important? Cambridge: Cambridge University Press. Spina, M., Garbisa, S., Field, J.M. and Serafini-Fracassini, A. (1979). The salmonid elastin fibril. An investigation of some chemical and physical parameters. Arch. Biochem. Biophys., 192(1), 430-437. 77 Starcher, B.C., Saccomani, G. and Urry, D.W. (1973). Coacervation and binding sites on aortic elastin. Biochim. Biophys. Acta, 310,481-486. Starcher, B.C. and Galione, M.J. (1976). Purification and comparison of elastins from different animal species. Analytical Biochem., 74,441-447. Stehbens, W.E. (1979). Anatomy and structure of blood vessels. In: Hemodynamic and the Blood Vessel Wall, W.E. Stehbens (ed), 3-74. Springfield: Charles C. Thomas. Steven, F.S., Minns, R.J. and Thomas, H. (1974). The isolation of chemically pure elastins in a form suitable for mechanical testing. Conn. Tissue Res., 2, 85-90. Sturkie, P.D. (1976). Avian Physiology, 3rd edition. Berlin: Springer Verlag. Sugai, S. and Ohno, N. (1980). Conformational transitions of the hydrophobic polyacids. Biophys. Chem., 11, 387-395. Tanford, C , (1962). Contribution of hydrophobic interactions to the stability of the globular conformations of proteins. J. American Chem. Soc, 84,4240-4247. Tanford, C. (1970). Protein denaturation. Part C. Theoretical models for the mechanism of denaturation. Adv. Prot. Chem., 24, 1-97. Tanford, C , (1980). The Hydrophobic Effect, 2nd edition. New York: John Wiley and Sons, Inc. Wainwright, S.A., Biggs, W.D, Currey, J.D. and Gosline, J.M. (1976). Mechanical Design in Organisms. London: Edward Arnold. Welscher, H.D. (1969). Correlations between amino acid sequence and conformation of immunoglobulin light chains. Int. J. Prot. Res., 1,253-265. Wilson, J.A. (1979). Principles of Animal Physiology, 2nd edition. New York: MacMillan Publishing. Wolinsky, H. and Glagov, S. (1967). A lamellar unit of aortic medial structure and function in mammals. Circ. Res., 20,99-111. 78 Wright , G . M . (1984). Structure o f the conus arteriosus and ventral aorta i n the sea lamprey, Petromyzon marinus, and the Adan t i c hagfish, Myxine glutinosa: microf ib r i l s , a major component. Can. J . Zoo l . , 62,2445-2456. APPENDIX I. Temperature CC) Volume Fraction (vj) S£. PIG 1 0.54 0.01 10 0.49 0.01 20 0.44 0.01 30 0.40 0.01 40 0.37 0.01 50 0.36 0.01 60 0.35 0.01 TURKEY 1 0.54 0.008 10 0.4g 0.008 20 0.42 0.008 30 0.37 0.008 40 0.34 0.008 50 0.32 0.008 60 0.31 0.008 FROG 1 0.51 0.03 10 0.46 0.03 20 0.42 0.03 30 0.39 0.03 40 0.38 0.03 50 0.36 0.04 60 0.35 0.04 TURTLE 1 0.49 0.04 10 0.46 0.04 20 0.42 0.04 30 0.45 0.04 40 0.38 0.05 50 0.37 0.05 60 0.36 0.05 SALMON 1 0.47 0.015 10 0.46 0.015 20 0.45 0.015 30 0.44 0.01 40 0.44 0.015 50 0.44 0.015 60 0.44 0.02 SHARK 1 0.48 0.02 10 0.44 0.02 20 0.39 0.02 30 0.35 0.015 40 0.31 0.02 50 0.29 0.02 60 0.28 0.02 COW LIGAMENT 1 0.59 0.005 10 0.52 0.005 20 0.44 0.005 30 038 0.005 40 035 0.005 50 0.34 0.01 60 0.33 0.01 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items