Reduced aerobic capacity and locomotory endurance in thyroid-deficient lizards

Thyroid hormones are present in all classes of vertebrates and are associated with a diverse array of metabolic processes (see Shambaugh, 1978). One of the most characteristic effects of thyroid hormones is a stimulatory effect on standard metabolic rate (SMR), and this calorigenic effect has been interpreted as evidence for a thermoregulatory role of the thyroid gland in endothermic vertebrates. A more recent view is that thyroid hormones may not play a major role in cold acclimation or other physiological adjustments to environmental temperature (Galton, 1978), but rather that primary effects of thyroid hormones may be alterations in the activities of enzymes in energetic pathways (Gorbman, 1978). Functional benefits accrued by an animal through these enzymatic effects are still unclear. Interpreting the stimulatory effect of thyroid hormones on SMR in ectothermic vertebrates is even more problematical, because metabolic heat production makes a negligible contribution to the control of body temperature in these animals.

Lizards present a good model for studying non-thermoregulatory actions of thyroid hormones because SMR is influenced by thyroid status in these animals (Maher & Levedahl, 1959; Maher, 1965; Wilhoft, 1966) but there is no obligatory change in body temperature or thermal conductance. Recently, John-Alder (1983) reported a stimulatory effect of thyroxine (T4) not only on SMR but also on maximal rates of O₂ consumption (⁠max) and muscle oxidative capacity in the desert iguana, Dipsosaurus dorsalis. An enhancement of max suggests an improvement in submaximal endurance (John-Alder & Bennett, 1981), and improved endurance could translate into improved performance during natural activities. If this were true, then selection for the stimulatory effects of T4 on metabolic capacities could have operated indirectly through selection for improved endurance.

The present experiments were designed to extend our understanding of metabolic effects of thyroid hormones and the resultant effects on activity capacity in lizards. Previous experimental T4-injection protocols have resulted in non-physiological plasma T4 concentrations (John-Alder, 1983). Such data are difficult to interpret since physiological responses to progressively higher levels of thyroid hormone levels are generally biphasic, being anabolic within some lower range of hormone concentrations and catabolic at higher concentrations (Shambaugh, 1978). It is difficult to interpret the resuts of earlier studies within the natural biological context of lizards. I have found that surgical thyroidectomy can be performed easily to induce hypothyroidism in Dipsosaurus, and the range of experimental plasma T4 concentrations induced by thyroidectomy is very similar to that of Dipsosaurus retrieved from winter hibernation (H. B. John-Alder, in preparation). Thus, similar physiological responses may occur in experimentally thyroidectomized lizards and in field-active lizards during the seasons marked by low thyroid activity.

The study was made upon Dipsosaurus dorsalis, for which there is considerable information on ecology (Norris, 1953; Muth, 1980), thermal biology (DeWitt, 1967), and physiology (Moberly, 1963; Bennett & Dawson, 1972; Gleeson, Putnam & Bennett, 1980; John-Alder & Bennett, 1981). This species is the subject of a continuing series of experimental and field studies designed to identify the physiological importance of thyroid hormones in lizards (John-Alder, 1983, and in preparation).

Adult male desert iguanas (mean body mass ±S.E. = 65·2 ± l·9g) were captured in May 1982 in Riverside County, CA (CA Scientific Collecting Permit No. 2005) and were transferred to the laboratory at the University of California, Irvine. Animals were maintained on a 13:11 h L: D photoperiod, the onset of light being at 06.00 PDT. The daily temperature cycle was as follows: 24 °C from 00.00 to 07.30; increasing temperature from 07.30 to 09.00; 40°C from 09.00 to 17.00; decreasing temperature from 17.00 to 18.30; 24°C from 18.30 to 24.00. These temperatures were selected because the average activity temperature of Dipsosaurus is 42·1 °C (Norris, 1953) and the substrate temperature at a depth of 20 cm was 24 °C at the time of capture. Animals were fed a mixed diet of lettuce and blossoms of common butterweed (Senecio sp.) thoughout the experimental period, with the exception of 36-h fasts prior to measurements of SMR. Water was available ad libitum.

Initial measurements of SMR, max during locomotory activity, and submaximal endurance were made within 4 days of capture before animals had been segregated into experimental groups. All measurements were made at 40°C. At the completion of these initial measurements of metabolic and performance capacities, animals were arbitrarily assigned to one of two groups for experimental treatment. Blood samples were collected via the infraorbital sinus for determinations of initial T4 concentrations. Subsequently, one group underwent surgical thyroidectomy, the other a sham operation. Lizards were maintained as described for 35 days until the final measurements were made. Metabolic and performance capacities were measured again during the sixth week after surgery. After the final metabolic and performance measurements were completed, lizards were killed by decapitation and were exsanguinated via the neck wound. Blood was collected in heparinized tubes, and plasma was immediately separated from formed elements. Plasma was frozen at −70 °C and stored for determination of T4 concentrations by radioimmunoassay. The heart, liver and hindlimb muscles were quickly dissected free of the carcass and were frozen and stored at −70 °C for subsequent determinations of enzyme activities.

Lizards were initially anaesthetized by having them rebreathe air inside a chamber saturated with vaporous Halothane. Subsequent administrations of Halothane were done as required via the barrel of a 20-ml syringe placed over the lizards head. All surgical equipment and the lizard’s neck were bathed with 95 % ethanol prior to surgery. A ventral midline incision was made through the skin anteriorly from the caudal plane of the pectoral girdle to a point above the posterior third of the hyoglossum. Muscles of the throat and the underlying fat body were separated by blunt dissection. The thyroid gland in Dipsosaurus is a bi-lobed, encapsulated structure lying on either side of the trachea and connected by a narrow isthmus on the ventral surface of the trachea. The entire thyroid was removed. An electrically-heated loop of steel wire was used to cauterize superficial vessels and the thyroid arteries and veins as they were encountered. Blood loss was usually negligible by this procedure. The throat cavity was then bathed with sterile saline, and the incision was closed with three or four disconnected ties of 3-0 silk. Sham operations involved anaesthesia and exposure of the thyroid gland. Lizards were given several hours in isolation for recovery and then were returned to the common housing facilities. Thyroidectomized and sham-operated animals were not segregated for housing.

Standard metabolic rates were measured on lizards placed individually into opencircuit metabolic chambers constructed from 22-cm sections of transparent Plexiglas cylinders (i.d. 8 cm). Five metabolic chambers were arranged inside an environmental chamber regulated at 40 ± 0·1 °C. Air flowed sequentially through a column containing Drierite and Ascarite for the removal of H₂O vapour and CO₂, respectively, through an upstream rotameter for measuring incurrent flow rate, through the metabolism chamber, through a downstream column of Drierite and Ascarite, and finally into a sampling manifold connected to a switching device. Excurrent air was vented either directly into the room or into a sampling port, as determined by the switching device. The switching device sampled the excurrent air from each chamber once each hour for 5–8 min. Three samples of room air were interspersed with the chamber air samples. Flow rates through the metabolism chambers were 120–150 ml min⁻¹. Samples of excurrent air were pumped through an Applied Electrochemistry model S-3A O₂ Analyzer at a rate of 20 ml min⁻¹. Animals were fasted for 36 h prior to entry into the chambers. Body masses were recorded before and after metabolic measurements, and the average of the two masses was used in subsequent calculations. Animals were placed inside the chambers at about 17.00 h. The lowest single record of O₂ consumption was used to calculate SMR.

Maximal values were measured on animals forced to run through a series of graded speeds on a treadmill as described previously (John-Alder, 1983). The initial speed was 0·9 km h⁻¹, and increments of about 0·2 km h⁻¹ were adjusted at approximately 2-min intervals. Air was drawn through the mask covering the lizard’s head at a rate of 470mlmin⁻¹ (STPD; about 500mlmin^-1 ambient). Continuous records of O₂ and CO₂ concentrations in the excurrent air were made on a recorder, and the record of the highest O₂ consumption was used to calculate max. Recordings not less than 1 min in duration were used for calculations. Rates of CO₂ production were calculated only insofar as they are required for calculating max when CO₂ is not removed from the air. Equation 3b of Withers (1977) was used to calculate ⁠, and equation 2 of Gleeson (1979) was used to calculate ⁠.

Endurance was measured as walking time to exhaustion at 1·1 km h⁻¹ on a motordriven treadmill. This speed was selected because it was the lowest speed that was still adequate to discriminate among animals with high and low endurance capacities. Endurance trials were terminated by exhaustion of the animal, as indicated by the second failure to sustain the tread speed, or at 30 min by the investigator.

Citrate synthase and pyruvate kinase activities were assayed on the same tissue homogenates. Tissues were homogenized in 100 mm potassium phosphate buffer containing 5 mm-EDTA (pH 7·4 at 0·4°C). All assays were done at 40°C. Dilution factors for the final homogenates of each tissue were as follows: liver, 50×; heart 500×; red iliofibularis muscle, 100×; white iliofibularis muscle, 25×; gastrocnemium muscle, 50 ×. Citrate synthase was assayed by the method of Srere (1969) using the reduction of 5,5′-dithiobis-(2-nitrobenzoic acid). Assays were performed in a Beckman model 25 thermostatted spectrophotometer in 1 ml reaction volumes, and the change in absorbance at 412 nm was recorded. Pyruvate kinase activity was assayed in all tissues as described by Somero & Childress (1980). Assays were performed in 2 ml reaction volumes in thermostatted cuvettes, and the change in absorbance at 366 nm was recorded. All enzyme activities for these two assays are expressed as Ug⁻¹ at 40 °C. Protein concentrations in homogenates were measured using Biuret reagent and BSA standards.

Myofibrillar ATPase activity was assayed using the white iliofibularis from the remaining hindlimb. Myofibrils were prepared at 0–4°C as described by Marsh & Wickler (1982). The ATPase assays were conducted in stirred 1-dram vials at 40 °C with a reaction volume of 15 ml containing 100 mm-KCl, 20mm-Tris, 2mm-MgCl₂, 0·25 mm-CaCl₂, 2mm-MgATP at pH 7·4 (adjusted at 40 °C). The assay was initiated by the addition of ATP and was terminated by the addition of 0·1 ml of 50% trichloroacetic acid (TCA) after 30 s. Zero time controls were done by adding 50% TCA first. Vials were placed on ice and subsequently analysed for free phosphate by the method of LeBel, Poirier & Beaudoin (1978). Protein was measured by the Biuret method with BSA standards. Activities were expressed as U mg⁻¹ myofibrillar protein.

Plasma T4 concentrations were determined by a radioimmunoassay (RIA) modified from MacKenzie, Licht & Papkoff (1979). Antiserum T4-15 was obtained from Endocrine Sciences (Tarzana, CA) and ¹²⁵I-labelled T4 (NEX-11IX) from New England Nuclear. Standards were prepared from crystalline L-thyroxine, Na-salt (Sigma T-2501) in 0·11 M-Na-barbital buffer (Barbital Sodium C-IV, Fisher B-22), pH 8·-6, containing 1 gl⁻¹ bovine gamma-globulins (Sigma G-3500) and 0·01 g l⁻¹ Thimerosal (Sigma T-5125). Assays were conducted in polystyrene gamma-counter vials. To begin the assay, 25 μl of sample or standard were pipetted into the vial, and 150 μl of ¹²⁵I-T4, diluted 2300× in barbital buffer containing 2·0 mg ml⁻¹ 8-anilino-1-naphthalene sulphonic acid (sodium salt; Pfaltz & Bauer A32940) in addition to the gamma globulins and Thimerosal, were added. Vials were vortexed and preincubated at room temperature for 1 h. Antiserum was diluted 1300X in barbital buffer, and 150 μl of diluted antiserum was pipetted into each vial. Vials were covered and incubated in the dark for 12– 15 h. Vials were immersed in an ice-water slurry for 20min, and 500 μl of polyethylene glycol (25%w/v; Carbowax PEG 8000; Fisher P-156) was pipetted into each vial. Each vial was vortexed for 5 s and returned to the ice-water slurry for 20 min. Vials were centrifuged at maximum speed in a Beckman model TJ-6 refrigerated bench top centrifuge for 20 min. Supernatants were decanted and discarded. Vials were inverted over absorbent paper for 2– 3 h prior to counting. Vials were counted for 1 min in a Beckman model 4000 gamma counter. Results were expressed as the quoteint of counts in the sample or standard divided by counts in the vial containing no unlabelled T4 (B_o). Quotients for standards were plotted on loglogit plots, and sample T4 levels were determined directly from the resultant linear plot. The limit of detection of the assay was 0·55 ng ml⁻¹. Sample T4 concentrations were linear through a series of four serial dilutions. Preparation of standards in Dipsosaurus plasma stripped of thyroid hormones had no effect on the standard curve.

All measured variables were analysed by analysis of covariance (Kleinbaum & Kupper, 1978) to adjust for differences in body mass among individuals. Initial measurements of SMR and max were pooled for analysis. In cases where body mass was not a significant covariate, subsequent comparisons were done by t-tests. Endurance measurements were analysed by the Wilcoxon signed ranks test. All statistical analyses were done on an Apple II Plus computer uing STATPRO (Blue Lake Computing, Madison, WI).

Surgical thyroidectomy resulted in a significant decrement in plasma T4 concentrations pre-and post-surgically (mean±s.E. 13· 0 ± 5· 1 vs 0· 83 ± 0· 28 ng ml⁻¹ ; P = 0·044). There was also a significant decrease in plasma T4 of sham-controls (9·5 ±2·2 vs 5·8 ± 1·6 ngml⁻¹; P = 0·015). Nevertheless, the final mean T4 concentration of thyroidectomized lizards was significantly less than that of shamcontrols (0·83 ± 0·28 vs 5·8 ± 1·6 ngml^{− 1}; P = 0·007) (Table 1). By the direct criterion of plasma T4 concentration, hypothyroidism was successfully induced in all thyroidectomized lizards relative to sham-controls. Individual final T4 concentrations in thyroidectomized animals ranged from below the limit of detection of the T4 assay to 1·8 ng ml⁻¹, a level below the range of sham-controls. The most likely explanation for the highest T4 levels in thyroidectomized animals is that some extra-thyroidal follicles had partially restored normal thyroid secretory activity by the end of the experiment.

The average body mass of sham-operated controls showed a small decline during the experiment (60·9 ± 2·8g vs 56·6 ± 3·3 g; P < 0·02), whereas there was no significant change in body mass of thyroidectomized animals (66·1 ± 3·0 g vs 63·5 ±3·0 g; 0·4 < P< 0·5). The differences between final mean body masses of sham-operated controls and thyroidectomized animals were not significant.

By analysis of covariance, SMR was lower in thyroidectomized animals than in sham-controls (P< 0·001) and lower in sham-controls than in the initial group (P< 0·005). These differences are illustrated in Fig. 1, in which log SMR is plotted as a function of log mass. This presentation allows comparisons of SMRs at all body masses, and differences among groups are seen as adjustments in intercepts. The 95 % confidence interval around the slope of log SMR on log mass is large (±0·48), and the slope should be interpreted with appropriate caution.

The final Vozmax of thyroidectomized animals was significantly less than that of both sham-controls and the initial measurements (P < 0·005 ; Fig. 2). Sham-operated controls showed no significant change in max during the experiment and have *een pooled with the initial measurements for this analysis.

Endurance is presented as walking time to exhaustion in Fig. 3. Some walking trials were terminated at 30 min by the investigator, and this protocol necessitated nonparametric data analysis. Endurance of thyroidectomized lizards declined significantly between the initial and final measurements (one-tailed P = 0·0078; Wilcoxon signed ranks test). There was no significant change in endurance of controls (P >0·10).

Liver, heart and three skeletal muscles were analysed for tissue and cellular responses to thyroid deficiency. The relative heart mass of thyroidectomized lizards, i.e. the ratio of heart mass to body mass, (gg⁻¹), was significantly lower than in shamcontrols (mean ± S.E. 0·0019 ± 0·0008 vs 0·00214 ± 0·0006; P = 0·015: Student’s t test on arcsin-transformed values) (Fig. 4). The relative masses of liver and skeletal muscles were not different between the two experimental groups.

Protein contents of heart, red iliofibularis muscle and gastrocnemius muscle were significantly lower in thyroidectomized lizards than in sham-controls (Table 2).

There were no significant differences between the two groups in white, non-oxidative iliofibularis muscle nor, surprisingly, in liver.

The activities of citrate synthase (Table 3) and pyruvate kinase (Table 4) were calculated both in mass-specific and in protein-specific units. On a mass-specific basis, citrate synthase activity was significantly lower in all tissues of thyroidectomized lizards compared to controls. These results indicate that the oxidative capacity per gram of all tissues was lowered by thyroid deficiency. On a protein-specific basis, however, there were significant decreases in citrate synthase activity only in liver and white iliofibularis muscle. The percentage decreases in citrate synthase activity of heart, red iliofibularis and gastrocnemius were, however, greater than the percentage decreases in protein content of these tissues.

Pyruvate kinase activity was not significantly changed by thyroid deficiency (Table 4). This result is unchanged whether pyruvate kinase activity is expressed on a massspecific or on a protein-specific basis. It should be noted, however, that all tissues of thyroidectomized animals had somewhat higher protein-specific pyruvate kinase activity than did tissues of controls, whereas the same tissues had lower proteinspecific citrate synthase activity than did controls.

Myofibrillar ATPase activity of white iliofibularis muscle is reported per mg of myofibrillar protein (Table 5) due to the nature of the assay. There was no significant difference between thyroidectomized and sham-control animals (mean ± S.E. 0.957 ± 0.063 vs 0.809 ± 0.070; P = 0.137).

Capacities for aerobic energy metabolism in Dipsosaurus are dependent on thyroid status. Both standard (−40%) and maximal (−16%) rates of O₂ consumption of thyroid-deficient lizards were below those of controls. In an earlier study, John-Alder (1983) reported a 60 % increase in SMR and a 15 % increase in max of T4-injected lizards. However, since T4-injection protocols used previously in lizards (Maher, 1965; Wilhoft, 1966; John-Alder, 1983) resulted in plasma T4 concentrations that were transiently two orders of magnitude above normal, the conclusions of such studies can be criticized for the non-physiological experimental T4 levels. In the present study, plasma T4 concentrations of thyroidectomizedDipsosaurus were in the normal range observed in these lizards retrieved from hibernation (mean ± S.E. 0·862 ±0·183 ng ml⁻¹ ; H. B. John-Alder, in preparation). Physiological responses to thyroid deficiency can thus be interpreted in the context of a naturally realistic thyroid status.

Standard metabolic rates of sham-operated controls decreased significantly during the experiment (Fig. 1 ). It is intriguing that there is a similar decrease in SMR of fieldactive Dipsosaurus at a comparable time of year (H.B. John-Alder, in preparation). One interpretation of this observation is that natural seasonal variation in physiological functions was not interrupted by the experimental animal-housing conditions. Two other results of the present experiments support this interpretation. First, there was a significant decrease in plasma T4 concentrations of sham-controls (Table 1), similar to the decrease in plasma T4 of field-active Dipsosaurus at comparable dates (H. B. John-Alder, in preparation). Secondly, citrate synthase activities reported here for sham-controls (Table 3) are comparable to those of field-active Dipsosaurus in mid-June and higher than those of field-active animals at any other time of year (H. B. John-Alder, in preparation).

The functional significance of reduced max in thyroid-deficient lizards can be seen in the concomitant reduction in locomotory endurance. On the basis of demonstrated correlations between max and endurance in lizards (John-Alder & Bennett, 1981; John-Alder, Lowe & Bennett, 1983) and mammals (Davies, Packer & Brooks, 1981), it was argued that T4-supplemented Dipsosaurus would probably have had improved locomotory endurance. Lowered endurance has been reported in thyroid-deficient rats (Baldwin, Hooker, Herrick & Schrader, 1980), and improved endurance has been reported in T4-supplemented rats (Naito & Griffith, 1977). Thus, it would seem clear that thyroid hormonal effects on aerobic capacity are directly reflected in submaximal locomotory performance. However, reduced endurance has been reported not only in thyroidectomized dogs (Kaciuba-Uscilko, Brzezinska & Kobryn, 1979) but also in T4-and T3-supplemented dogs (Brzezinska & Kaciuba-Uscilko, 1979a,EXBIO_109_1_175C4b). In the latter two reports, reduced endurance was associated with lowered pre-exercise levels of liver and muscle glycogen. Reduced liver and muscle glycogen have also been reported in T4-supplemented frogs (McNabb, 1969; Packard & Randall, 1975). Given the potential for a reduction in energy stores in hyperthyroid animals, it is important that any prediction of an improvement in endurance of T4-supplemented animals be verified experimentally.

Thyroid hormones affect both the potential for cardiovascular O₂ delivery capacity and the oxidative capacity of skeletal muscle. The reduced relative heart mass of thyroid-deficient lizards suggests a reduced cardiovascular capacity for muscle perfusion during activity and consequently a reduced maximum work capacity (see Davies et al. 1981). Mass-specific citrate synthase activity, a frequently-used index of mitochondrial oxidative capacity, was lower in all tissues of thyroidectomized Dipsosaurus than in controls. The reduction in muscle oxidative capacity is directly associated with impaired submaximal locomotory endurance. High oxidative capacity favours oxidation of fatty acids and sparing of glycogen. Glycogen sparing is important in the context of endurance because glycogen depletion is likely to be the ultimate factor responsible for exhaustion during submaximal activity (Holloszy et al. 1977). Baldwin et al. (1980) and Kaciuba-Uscilko et al. (1979) presented evidence of impaired lipid mobilization in thyroid-deficient rats and dogs, respectively, and Paul (1971) reported that glucose provides a greater contribution to substrate utilization during exercise in thyroidectomized than in control dogs. Thus, thyroidectomized dipsosaurus would be expected to have reduced submaximal endurance even if preactivity glycogen levels were unaffected by thyroid deficiency. Protein contents of liver and white iliofibularis muscle were not affected by thyroidectomy (Table 2). The decrements in citrate synthase activity of these tissues (Table 4) thus appear to be specific effects of thyroid deficiency. In heart, red iliofibularis and gastrocnemius muscles, however, decreases in protein contents can account for the decreases in citrate synthase activity.

It is of interest that the mass-specific citrate synthase activity of all skeletal muscles was lower in thyroidectomized lizards compared to controls, whereas T4-supplementation of intact animals failed to induce an increase in citrate synthase activity of white, non-oxidative iliofibularis muscle (John-Alder, 1983). Similar findings have been reported in mammals. Baldwin et al. (1980) and Janssen, van Harde-veld & Kassenar (1978) have reported decrements in the oxidative capacities of red, oxidative and white, non-oxidative mammalian skeletal muscles, whereas Janssen et al. (1978) and Winder & Holloszy (1977) have reported white mammalian skeletal muscle to be much less responsive to thyroid hormones than red muscle. These differences among muscles may be explained on the basis of different concentrations of thyroid hormone receptors in different muscles.

The effects of thyroid hormones on energetic pathways are apparently restricted to enzymes of aerobic metabolism. Pyruvate kinase activity, an index of glycolytic capacity (Zammit, Beis & Newsholme, 1978), was unchanged in all tissues that were examined. Furthermore, myofibrillar ATPase, the dominant ATPase of skeletal muscle was not affected by thyroidectomy. It appears, then, that there is a dichotomy between responses in those energetic pathways used in sustained activities and responses in those pathways of rapid ATP synthesis and hydrolysis used during intense activity. It would be expected that the capacity of Dipsosaurus to engage in brief periods of strenuous activity requiring high power output would be unimpaired by thyroid deficiency.

This new experimental information on physiological responses to thyroid manipulation in lizards (see also John-Alder, 1983) establishes a new selective context for understanding the evolution of thyroid function. Previous field studies on seasonality of thyroid glandular activity in field-active lizards were not successful in identifying physiological correlates of thyroid glandular activity, and laboratory studies have failed to establish the functional significance of changes in SMR in thyroid-manipulated lizards (see Lynn, 1970). However, an intriguing, recurrent observation has been that peak thyroid activity in field-active lizards occurs at the time of highest physical activity (Lynn, 1970). Recently, John-Alder (1982) reported preliminary findings showing that seasonal changes in thyroid function parallel seasonal changes in aerobic capacity and endurance in Dipsosaurus (H. B. John-Alder, in preparation). Selection for metabolic responses to thyroid hormones in lizards may have operated on the advantages associated with improved endurance in addition to some other aspects of increased metabolic rate per se. Increases in SMR associated with thyroid activity would thus be seen as metabolic costs associated with the maintenance of high energetic capacities, and these costs would be outweighed by the advantages of im-proved endurance.

Dr Kenneth W. Baldwin, Dr Albert, F. Bennett, Dr Paul Licht and Dr Richard L. Marsh made contributions to the development of this manuscript. Ms Kathleen L John-Alder skillfully prepared the figures and helped in the preparation of the manuscript. Mr Theodore Garland provided advice on statistical analyses. Mr Brett A. Adams read a preliminary draft of the manuscript. This study is part of my doctoral dissertation. I am grateful to Dr Albert P. Bennett for chairing my dissertation committee and for opening his laboratory to me. Supported by NSF Grant DEB 8119797 to HBJ-A and AFB and by an award from the Patent Fund of the Regents of the University of California to HBJ-A.