AN ELECTROPHORETIC STUDY OF AVIAN

EYE-LENS PROTEINS

CHARLES G. StBLEY ANt) ALAN H. BRUSH

THE search for evidence of genetic relatedness among various groups of animals which can be used as the basis for taxonomic conclusions has

resulted in comparative studies of many different characters. Within the past few years it has become clear that the structure of protein molecules is a potentially productive source of systematic data, because of the rela- tionship which exists between the genetic material (DNA) and the linear sequence of the amino acids composing a protein molecule. The rationale behind the use of protein structure in taxonomic studies has been dis- cussed in previous papers (Sibley, 1960, 1962, 1964, 1965).

Several protein systems in vertebrate animals have been explored as sources of data for classification. Among the more extensive electrophoretic comparisons are those of Dessauer and Fox (1956) on the serum pro- teins of amphibians and reptiles, Johnson and Wicks (1959) on mam- malian serum proteins, and Sibley (1960) on avian egg-white proteins. The symposium volume edited by Leone (1964) contains several papers relating to proteins as sources of taxonomic data.

The lens of the vertebrate eye possesses several characteristics which suggest that its proteins should be an excellent source of taxonomic in- formation. It is a discrete, easily isolated structure which is extremely simple to collect even under field conditions and it contains the highest concentration of protein of any structure in the vertebrate body, 35 per cent by weight. The lens becomes cytologically isolated early in embryonic development, contains only epithelial cells, and grows slowly throughout the life of the animal, thus becoming more dense with age. There is no direct blood supply to the lens and its metabolism is sluggish, with metabolic processes limited almost exclusively to glycolysis in which about one-fifth of the glucose pathway is through the pentose-phosphate shunt (Ely, 1949). Aerobic metabolism may be confined to the epithelium and cortex (van Heyningen and Pirie, 1953). The only other significant metabolic pathway is proteolysis, a process mediated by peptidase enzymes with esterase activity. Recently Swanson (1966) demonstrated the ex- istence of lysosomes in the lens epithelium. The functions of other sub- stances, for example ophthalmic acid, ascorbic acid, and glutathione are obscure. Van Heyningen (1962) suggests that they may function as co- enzymes.

The first comparative study of vertebrate lens proteins that included considerable avian material was the extensive work of Rabaey (1959)

203 The Auk, [14: 203-219. April, 1967

204 SmL•¾ ̂ •D BRusu, Avian Eye-lens Proteins [ Auk [ Vol. 84

in which the lens proteins of 29 species of birds, 37 species of mammals, and various numbers of species from other groups were compared using agar-gel electrophoresis. Rabaey's paper stimulated the senior author to begin the collection of avian eye lenses, because an examination of the electrophoretic patterns suggested th'at they contained "similarities and differences correlated with systematic relationships" (Sibley, 1962: 113). Rabaey has published several additional papers (1962, 1964) and his student, Gysels, has published a series of taxonomic papers on eye-lens proteins (1962, 1963, 1964a, b, c, d, 1965). Gysels and Rabaey (1964) used eye-lens proteins in a study of the relationships of certain alcids. Gysels' monograph (1964d) presents agar-gel electrophoretic and im- munoelectrophoretic data on the eye-lens proteins of 233 avian species. This paper was translated for us from the original Flemish by Dr. Ari van Tienhoven of Cornell University to whom we are deeply indebted. Dr. van Tienhoven's translation has been checked by Dr. Gysels.

METHODS AND RESULTS

In the present study the electrophoretic properties of the eye-lens proteins were compared using the vertical starch-gel method of Smithies (1955, 1959). A discontinuous buffer system was employed with a Tris- citric acid buffer, pH 7.95, in the gel, and a boric acid-lithium hydroxide combination at pH 7.98 as the "bridge" buffer (Ashton and Braden, 1961; Ferguson and Wallace, 1961). Bromphenol blue dye was used as an inert marker to follow the buffer front. All separations were carried out at 4øC. The lens proteins of more than 1,400 specimens representing over 400 species, and 22 of the 27 orders recognized by Wetmore (1960), were examined. Lenses that were collected in the field were placed in a 1: 1000 Merthiolate solution and kept as cold as possible. Such speci- mens were shipped by air from many parts of the world, without refrigera- tion, but were stored at -10øC when received. Lenses collected from fresh specimens in the laboratory were placed in a small volume of Tris- citric acid buffer, mashed, diluted to a concentration of 1-2 grams/100 ml with the buffer, and applied directly to the starch gel. There were no differences between the electrophoretic patterns obtained from fresh material that was centrifuged after the above preparation and from that which was not. Freshly collected lenses went completely into solution but a small residue of insoluble material was usually apparent when lenses were prepared following either preservation in Merthiolate or after storage.

All lenses which were stored for any period of time exhibited changes in the electrophoretic behavior of the proteins. Frozen material changed more slowly than unfrozen specimens but the net ch'anges were similar (Figure 1). Comparisons between stored specimens and fresh material

April ] SIBLEY AND BRUSH, Avlan Eye-lens Proteins 205 1967 J

I

8

8'

Cyano½itta ½ristata

0 RT 1.0

Melospiza rnelod/a

0 Rf 1.0

Figure 1. Effects of storage on the lens patterns in two species. The horizontal axis is indicative of electrophoretic mobility. The numbers on the vertical axis in- dicate the number of days the specimens were stored at -10øC. In the upper diagram 8 and 8' are the patterns after eight days of storage of two individuals which originally had identical patterns.

showed that most of the denaturation occurs between the time of collec-

tion and arrival in the laboratory. Under all storage conditions there is some denaturation, presumably due either to the autolytic activity of proteases in the lens or to the effects of freezing. Wood and Burgess (1961), who used agar-gel electrophoresis, also reported changes in lens proteins related to freezing. The patterns of denatured samples were easily recognized but the denatured patterns were inconsistent--samples from the same species could give quite different patterns after exposure to a similar amount of denaturation. Controlled denaturation experiments gave some insight into the nature of the patterns produced, but the great amount of uncertainty involved makes interpretation of the patterns of samples with unknown histories both risky and difficult.

Specific stains were used to locate and to identify various proteins in the lenses. The over-all pattern of proteins was developed with amido- black 10 B (0.7 gm of stain in methanol: water: acetic acid, 50: 50: 10).

206 S•BLEY A•D BR•JSX•, Avian Eye-lens Proteins [ Auk L Vol. 84

Esterases were located with the method described by Manwell and Baker (1963: 41).

Glucose-6-phosphate dehydrogenase (G-6-PD) was located by a method modified slightly from Shows et al. (1964). Sodium nitroprusside was used as a general reagent to identify compounds rich in sulfhydryl (SH) groups (Grunert and Phillips, 1951). In the lens one of the most com- mon molecules which contains SH groups is glutathione. In addition both the alpha- (Bj/Srk, 1961) and beta- (van Heyningen, 1962) crystallins are reported to have high sulphur and sulfhydryl contents. Attempts to locate glycogen by the method of Rabaey (1963) were unsuccessful.

Identification of lens proteins.--Mammalian and avian eye lenses con- tain three principal fractions in paper electrophoresis (Fran$ois et al., 1954; Maisel and Langman, 1961b) but, under certain conditions (low ionic strength, increased voltage), more may appear. Agar microelectro- phoresis at high voltage (40Vfcm) has revealed as many as 17 lenticular protein fractions (Fran$ois and Rabaey, 1959).

Three major protein fractions from the chick lens have been isolated by paper-strip electrophoresis and analyzed in the ultracentrifuge (Maisel and Langman, 1961b). Alpha-crystallin is the largest molecule (molecular weight [MW]: 1,000,000; Sedimentation Coefficient: 17.0-19.0) and apparently consists of several subunits. It also has the greatest mobility of the soluble proteins in paper, starch-block, agar, and free electrophoresis at pH 8 (Bloemendal and Bont, 1962; Bloemendal et al., 1963). Alpha- crystallin beh'aves as a single component in vertical starch-gel electro- phoresis, but has the lowest mobility (see below). Chick beta-crystallin (MW = 200,000; Sed. Coeff. -- 9.3) and gamma-crystallin (MW: 60,000; Sed. Coeff.=4.3) are smaller and more heterogeneous both electrophoretically and immunologically. In paper-strip electrophoresis and in continuous-flow electrophoresis the mobility of the lens proteins is related directly to their molecular weight (Maisel and Langman, 1961a) and presumably to their total charge.

The lens of the adult domestic fowl (Gallus gallus) has been sh•own to contain at least seven substances which act as antigens (Langman, 1959). Other immunological studies have indicated the presence of from 5 to 10 antigens (Maisel and Langman, 1961b). Maisel and Langman grouped these antigens into three main fractions. Fraction I consists of one antigen and corresponds with alpha-crystallin. Fraction II consists of at least four components and is identical with beta-crystallin. Frac- tion III contains two antigenic components and is identical with gamma- crystallin. The sequence of formation during organogenesis is alpha, beta, and gamma (Langman and Maisel, 1962). Nuclear beta-crystallins ac-

April ] SIBLEY AND BRUSH, Avlan •Eye-lens Proteins 207 1967 I

I II III

0 Rf

MW. I0 6 2.10 5

IV

6.10 4

Figure 2. General pattern of avian lenticular proteins in starch-gel electrophoresis. Molecular weights (MW) were determined from Sedimentation Coefficients of the lens proteins of Gallus gallus. Roman numerals indicate fractions described in the text. The application point is indicated by the 0 at the left side of the diagram.

cumulate prior to hatching and their appearance coincides with • the ap- pearance of the primary lens fibers (Maisel and Goodman, 1965).

Agar diffusion and immunoelectrophoretic techniques have demonstrated lens protein antigens in the cornea, iris, and retina of the chick eye. Lens antigens have also been detected in the aqueous humor, brain, and skin of the eyelid. Such extra-lenticular antigens are identical electrophoreti- cally with those of the lens proteins. Antigens from the other ocular tissues are immunologically identical also; however, corresponding electrophoretic components from brain and skin show only partial immunological identity (Maisel, 1962a, b).

We found six to eight lens proteins by starch-gel electrophoresis, and we have located and identified a number of enzymes in the electrophoretic pattern of avian eye-lens proteins. The variation in the location of the enzymes is negligible, indicating that they are relatively consistent in structure. Multiple molecular forms of some of these enzymes were present.

Figure 2 represents the important protein fractions of the avian eye lens and their relative mobilities. Under the conditions of this study most fractions migrate towards the anode. The anodal direction is to the right and the cathodal to the left. The fractions, starting with that nearest the application point, are:

I. Pre-alpha-crystallins.--These are always narrow, lightly staining bands that are irregular both in their appearance and location.

II. Alpha-crystallin (Rf= .250-.375).--This is the broadest, darkest staining, and most consistent of the protein fractions. It ap- pears as a rather broad, uniform band. This band and the area im- mediately cathodal to it are major sites of sulfhydryl activity. Towards the anode a series of smaller fractions occasionally appears. These are thought to be denaturation products and are much less consistent in appearance than the main alpha-crystallin fraction.

208 SIBLE¾ AIgD BRXJS•, Avian •Eye-lens Proteins [ Auk L Vol. 84

C,vana½itta •ristata

I

2

3

0 Rf 1.0

M•lospixa m•lodia

0 Rf 1.0

Figure 3. Representative patterns of the lens proteins of the Blue Jay (Cyanocitta cristata) and of the Song Sparrow (Melospiza melodia). These patterns demonstrate the low degree of intraspecific and interspecific variation in fresh material. Numbers on vertical axes indicate different specimens; m indicates an average pattern based upon three or more fresh specimens.

III. Beta-crystallin (R l: .425-.525).--This is the site of beta- crystallin and this region also contains G-6-PD activity. In the House Sparrow, Passer domesticus, the G-6-PD was present as a two- band isozyme. G-6-PD is not always associated with a protein band as detected by amido-black. This is due to the different sensitivities of the staining techniques used to locate specific fractions.

Ri = .600- .800.--This region in many groups, appears as a space and therefore acts as a marker which separates the major proteins. Th'ere are several orders (Strigiformes, Caprimulgiformes, and Cu- culiformes) that do not show this break but which may have an addi- tional protein band in this location. The variability of this spacing is high.

IV. Gamma-crystallin ( R• : .800-.950).--The last important protein complex is gamma-crystallin. This fraction consists of three or fewer bands of variable widths. Gamma-crystallin has been re- ported to occur only infrequently or in low concentrations in birds (see Papaconstantinou, 1964, and pers. comm.). Although it appeared

April '[ SIBLEY AND BRVS•, Avian Eye-lens Proteins 209 1967 J

to be the most variable of the fractions, in both number of compo- nents and mobility, it was always present in fresh samples.

Esterases usually appear close to the cathodal boundary of the gamma-crystallin bands. In one specimen of the Black-collared Barbet (Lybius torquatus), two bands appeared which suggests the existence of multiple molecular forms of this enzyme. However, be- cause of the diversity in the nature of esterases these are not neces- sarily isozymes (Hunter and Strachan, 1961). The enzyme responsi- ble for the esterase activity is thought to be an aminopeptidase (Spec- tor, 1962).

In general, variation in the electrophoretic patterns among the indi- viduals of a species was slight (Figure 3) but variation increased as the material aged during storage. The major bands (Figure 2) were less transient than the minor ones and the hierarchy of persistence of the major bands was alpha > beta > gamma. Many of the older samples gave only a single band, located in the Rf = .300--.350 range which is assumed to be alph'a-crystallin. The insoluble material in old or Merthiolate-treated samples was presumably the "albuminoid" of early investigators. This material is apparently closely related to alpha-crystallin and may be a denaturation product. This material showed no electro- phoretic mobility under the conditions of our study.

In various gel media such as starch (Smithies, 1962) and acrylamide gel (Ornstein, 1964) the pore size of the matrix is of molecular dimen- sions. In these systems the size of the molecules being separated is as important a physical factor as is their net charge. Depending upon the pore size of the gel, protein molecules will be retarded to a degree pro- portional to their size. However, the effects of the introduction of a starch-gel matrix into the electrophoretic field are actually more com- plex than merely an over-all increase in the viscosity of the medium with the resultant proportional slowing of all migrating molecules. The effect of the gel matrix can be non-linear, although this depends mainly upon the concentration of the gel. Within certain limits increased gel concen- tration provides better resolution of protein fractions without adversely affecting other parameters (Krotoski et al., 1966). Although the rela- tionships among gel concentration, molecular size, and mobility are complex, we have, for present purposes, assumed that in starch gel the pro- teins are separated by size as well as by charge. If this is true then the rate of migration of a protein in starch gel will be, in part, an inverse function of the density of the starch matrix (Smithies, 1962). That is, the migration of molecules is proportional to the reciprocal of the starch concentration and the molecules with the highest molecular weight migrate

210 S•B•rY AND BRUSa, Avian Eye-lens Proteins [ Auk / Vol. 84

I/S

Figure 4. The effect of starch concentration upon the electrophoretic mobility of

avian lens proteins. M--• relative mobility, 1/S--• the reciprocal of the starch con- centration. The slope of the lines is a function of molecular size. Molecules with a slope approaching zero are too small to be retarded by the starch-gel medium.

the shortest distance. Lenticular proteins, when separated in gels of vary- ing concentration, conform to this hypothesis (Figure 4). The slope of the curve that relates relative migration to the reciprocal of the starch concentration (l/S) can be defined as the retardation coefficient, r. The slope of this curve is proportional to molecular size, thus the relative retardation of eye-lens proteins under these conditions appears to be a function of molecular size. From this evidence we conclude that both

molecular size and net charge contribute to the relative electrophoretic mobilities of lenticular proteins in starch gel. Thus, the starch-gel tech'- nique provides a comparative index to more molecular parameters than any other electrophoretic medium except other molecular sieve systems such as "disc" or acrylamide-gel electrophoresis.

Variation in patterns.---There is little intraspecific variation in the pat- terns produced by freshly collected lenses (Figure 3). Differences among the species of related groups are also small (Figure 5). The patterns produced by the lens proteins of some 60 species of the order Passeriformes, representing many families, show remarkably little variation. There are no consistent differences between oscines and sub-oscines. Because varia-

April ] 8IBLEY AND BRUSH, Avlan Eye-lens Proteins 211 1967 .l

tion in a character is paramount to its usefulness in taxonomic comparisons a statistical analysis of variation in the lens-protein patterns was under- taken.

Measurements of the mobilities of th'e proteins in an electrophoretic pattern are indexes to some aspects of molecular structure but it must not be forgotten that two proteins with different amino acid sequences can have identical mobilities and, conversely, two proteins differing by only a single amino acid can have different mobilities. One safeguard against either of these possibilities is to compare only homologous systems composed of several proteins, such as the lens proteins. The complex patterns so produced are likely to be completely identical only if they are derived from genetically very similar organisms. Thus small differ- ences in the mobility of homologous proteins from related species usually mean very little taxonomically.

In our comparisons of the mobilities of the passerine lens proteins we found that all species were similar in over-all pattern but that certain proteins sh'owed consistently greater variability than others. In general the cathodal boundary was more variable than the anodal. This is to be expected since "tailing" will produce this effect. For example, in a sample of lenses from seven individuals of the Song Sparrow (Melospiza melodia) the beta-crystallin component had a standard deviation of 0.022 for the anodal boundary and 0.021 for the cathodal boundary. The co- efficients of variability for these were 2.54 (anodal) and 3.14 (cathodal). Other fractions may be more variable. For example, the pre-alpha-crystal- lins from a sample of 10 White-throated Sparrows (Zonotrichia albicollis) had a standard deviation of 0.034 (anodal) and 0.092 (cathodal) with coefficients of variability of 5.18 and 23.52, respectively. The low vari- ability among related species was demonstrated by a sample of 13 in- dividuals of nine species representing six families of passefines. In these patterns the anodal boundary of the beta-crystallin had a standard deviation of 0.0039 and for the cathodal boundary 0.048. The mean of this sample compared with that of the Song Sparrows, mentioned above, showed no statistically significant difference at the 95 per cent level. It is somewhat surprising to find th'at the variability in the sample of these nine species is actually lower than in the sample of 10 White-throated Sparrows but this demonstrates further the homogeneity of passefine patterns.

Statistically significant differences between higher categories are equally difficult to demonstrate because of the difficulties inherent in quantify- ing patterns and the lack of obvious and consistent differences among patterns. It can be shown that there are statistically significant differences in the mobilities of some of th'e larger, more consistent fractions but there

212 SIBLEY ^•D BRUS•, Arian Eye-lens Proteins [ Auk [ Vol. 84

N. nycticorax

Pelecanus occidentalis

Anas platyrhynchos

Butso jamaicansis

Phasianus colchicus

P. pardix

.: Charadrius wilsonia

.. ....... _ $treptopelia decaocto

Bubo virginianus

Megaceryle alcyon

Cyanocitta cristata

Passer domesticus

t•ichmondena cardinali3

Zonotrichia albicollis

dunco hyemalis

Melospiza melodia

t•. rattus

Hana piplens

April ] SIBLEY ̂NI• BRUS•, Avian Eye-lens Proteins 213 1967 .I

Fish

Fro g

Turtle

Bird

Rabbit

Rat

Rf

Figure 6. Starch-gel electrophoretic patterns of the lens proteins from various classes of vertebrates.

is no regularity to these differences. This is complicated by the apparent splitting of certain fractions into subunits. The degree to which the lens proteins appear to break up into subunits may be due to differences in intramolecular binding and handling. No reliable index was devised to estimate the total differences between patterns, although homologous frac- tions can be compared. However, there is no fully satisfactory method to assess the taxonomic value of the presence or absence of fractions or to sum the statistical differences between individual fractions. These dif-

ficulties apply to taxonomic assessments of th'e electrophoretic patterns of any protein system but in such systems as avian egg white or hemo- globins there are consistent patterns within related groups and con- sistent, often large differences between distantly related groups. These make it possible to assess the taxonomic value of similarities and differ- ences in the total pattern. In the eye-lens patterns, the great similarity in the patterns of fresh material and the increased but random variability which follows storage make it extremely difficult, if not impossible, to demonstrate taxonomically valid similarities and differences. To some degree the differences are apparently the result of denaturation and there- fore are not valid bases for taxonomic conclusions.

To determine the extent of variation in lens patterns among the vertebrates, fresh material from several other classes was obtained. A

Figure 5. Starch-gel electrophoretic patterns of the lens proteins of various birds, a mammal, and a frog. Apparent mobility differences are not necessarily real unless the patterns are side by side in the same gel. In this figure, Phasianus and Perdix, Zonotrichia and Junco, and Rattus and Rana are the pairs which were side by side in the same gel.

214 S•B•E¾ ^•D BRrrs•r, Avian Eye-lens Proteins [ Auk Vol. 84

Pelecaniformes

Procellariiformes

Ciconiiforrnes

Anseriformes

Falconiforrnes

Galliforrnes

Gruiforrnes

Charadriiforrnes

Colurnbiforrnes

Psittaciforrnes

Strigifor rnes

Caprirnulgifor rnes

Cuculiforrnes

Coliifor rnes

Coraciiforrnes

Piciforrnes

Passerlforrnes

8 M•lospiza

I0 Cyanocitta

8 fornilies

0 0.5 1.0

Rf

Figure 7. Composite starch-gel electrophoretic patterns of the lens proteins of various orders of birds. Although the gamma-crystallins show considerable variation, the patterns throughout are essentially similar. Taxonomically correlated variation may be present but is obscured by the effects of rapid denaturation.

great deal of variability was present although certain similarities were also apparent (Figure 6). Comparisons among the patterns from the various orders of birds often failed to demonstrate either consistent sim- ilarities within an order or consistent differences between orders (Figure

7). In general the differences between orders were no greater than the variation within an order. However, the different proteins show differ-

April 1967 ] SIBLEY ANI) BRUSH, Arian Eye-lens Proteins 215

ing degrees of variation and therefore it is necessary to consider them individually.

The pre-alpha-crystallins are the most variable in mobility and also in presence and absence. Alpha-crystallin is the most consistent while the beta-crystallin complex is only slightly less so. Both of the latter are quite regular in occurrence in all birds.

The gamma-crystallin complex is consistently less concentrated than the alpha- or beta-crystallins. In most orders the gamma-crystallin con- sists of three fractions but both the mobilltles and numbers of fractions

are variable with the most cathodal one displaying the greatest varia- tion.

The Strigiformes, Caprlmulglformes, and Cucullformes have similar patterns which are somewhat different from other orders. A fraction at Rf = 0.600 - 0.700 iS the distinguishing feature of the pattern. The origin and nature of this fraction is unknown. The Coliiformes and the Coraci-

iformes also have a band in this region but with a somewhat higher mo- bility (Figure 7).

Under the conditions of this study, cathodally migrating lens proteins have been found only in the Accipitridae, in the domestic pigeon, Columba livia, and in the Hawk Owl, Surnia ulula.

DISCUSSION

At the time this study was begun it seemed certain that the lens pro- teins would prove to be taxonomically useful and we fully expected to find consistently similar patterns within closely related groups and con- sistent differences between less closely related groups. This had been the case in previous studies of egg-white proteins, blood-serum proteins and hemoglobin and there was no reason to expect the lens proteins to be different in this respect. We were therefore surprised at the unusual amount of random heterogeneity which soon became apparent at all levels. This eventually proved to be due, at least in part, to the effects of dena- turation. A series of experiments using only fresh lenses demonstrated that the patterns of absolutely fresh lens proteins were remarkably alike in most birds, and that changes due to denaturation occur very soon after death, even in frozen material. We have therefore been forced to conclude that the results of our present study cannot safely be used as the basis for taxonomlc decisions. To the extent that electrophoretic behavior is an index to protein structure the eye-lens proteins of birds are more uni- form throughout the class than, for example, the egg-white proteins or the hemoglobins. No doubt there are differences in the amino acid se- quences of homologous proteins in different groups but these differences must be relatively small. Further studies using only absolutely fresh

216 SIBLEY AND BRUS•, Arian Eye-lens Proteins [ Auk 1. Vol. 84

lenses may uncover consistent and informative patterns of variation which could be taxonomically useful.

Because Rabaey and Gysels have used a different technique, namely, agar-gel electrophoresis which does not have the molecular sieving prop- erties of starch gel, we cannot claim to have disproved their results. As did we, they have often used lens specimens that were not absolutely fresh, but (pers. comm.) they have not found evidence of such' large or rapid changes as we have encountered. Until the two techniques can be compared using the same lens specimens, we must conclude that at least some of our differences in results are due to differences in experi- mental techniques. On the other hand we feel compelled to express our conviction that the use of small mobility differences as the basis for major taxonomic changes is unwarranted. In our opinion the extremely interest- ing proposals which have been made by Gysels and Rabaey (e.g., 1964) for modifications in avian classification should be examined with great caution. Some of these proposals may well prove valid and all should be given consideration but, as of now, we are unable to accept them with- out additional proof.

ACKNOWLEDGMENTS

We are grateful to many persons for assistance in the collection of lens specimens. During field work in Australia in 1963, the senior author was aided by D. F. Dorward, A. J. Marshall, H. Middleton, and other staff members and students at Monash University. In Africa (1964) the kind assistance of R. Liversidge, N. van der Merwe, R. Boulton, A. Forbes-Watson, D. Owen, and L. Grimes was greatly appreciated. Many specimens from Spain were contributed by P. Garayalde and from Argentina by F. Contino. R. A. Norris and H. L. Stoddard sent us a large number of species from Florida. We are also pleased to thank J. E. Ahlquist, R. C. Banks, J. Brown, C. T. Collins, K. W. Corbin, J. M. Forshaw, D. J. Futuyma, V. Grimes, W. Grow, A. Harkabus, H. T. Hendrickson, H. P. Hoffman, C. Lacey, L. J. Loomis, H. G. Lumsden, A.M. Morgan-Davies, E. S. Morton, B. G. Murray, L. Nichols, S. M. Patten, L. Pearsall, D. A. Rose, J. S. Rowley, W. C. Russell, B. K. Seavey, L. L. Short, Jr., F. C. Sibley, H. Sick, H. K. Springer, the late D. Stamm, C. Sutherland, G. A. Swanson, R. Weisbrod, C. White, R. Wood, and J. B. Woodford.

Most of the work reported in this paper was carried out at Cornell University with the support of grants from the National Science Foundation (G-18958; GB-2229) and the National Institutes of Health (GM-6889). The junior author was a post-

doctoral trainee (1964-65) under the Cornell N.I.H. Genetics Training Grant (T1- GM-1035) directed by Prof. A.M. Srb. For assistance with the laboratory work we are indebted to Susan Rauchway, Louise Barr, and Martha Maxwell.

SUMMARY

The starch-gel electrophoretic patterns of the eye-lens proteins of over 1,400 specimens, representing more than 400 species of birds, were compared in a search for evidence of taxonomic relationships. Rapid

April] SlBLEY ^ND BRUS•t, Avlan Eye-lens Proteins 217 1967

denaturation of the lens proteins made it difficult to obtain trustworthy comparative data from specimens collected in the field. Fresh material, examined immediately after death, revealed that the eye-lens proteins of birds are probably more uniform throughout the class than, for example, the egg-white proteins. Although no taxonomic conclusions were reached on the basis of this study it should not be assumed that the lens proteins are lacking in taxonomically significant variation. However, in future comparative studies of the lens proteins it would be prudent to use only absolutely fresh material.

LITE1L•.TURE CITED

AS•tTON, G. C., ANn A. W. H. BWDEN. 1961. Serum-globulin polymorphism in mice. Australian J. Biol. Sci., 14•: 248-253.

Bj6R•c, L. 1961. Studies on alpha-crystallin from calf lens. Exp. Eye Res., 1: 145- 154.

BLOEMENDAL, H., AND W. S. BONT. 1962. 'Purity' of alpha-crystallin. Nature, 193: 437-439.

BLOEMF. NnAL, H., W. S. BONT, J. F. JUNGKn'•D, AND J. H. WISSE. 1963. Splitting and recombination of alpha-crystallin. Exp. Eye Res., 1: 300-305.

DESSAUE•, H. C., AND W. FOX. 1956. Characteristic electrophoretic patterns of plasma proteins of orders of amphibia and reptiles. Science, 1'/4•: 225-226.

ELY, L. O. 1949. Metabolism of the crystalline lens. II. Respiration of the intact lens and its separated parts. Amer. J. Ophthal., 32: 22(3-224.

FE•OUSON, K. A., AND A. L. C. WALLACE. 1961. Starch-gel electrophoresis of an- terior pituitary hormones. Nature, 190: 629-630.

FraNCois, J., R. WIEME, M. RaBAE¾, AND A. NEETENS. 1954. L'electrophorese sur papier des proteines hydrosolubles du cristallin. Experientia, 10: 79-80.

FraNCois, J., AND M. RABA•¾. 1959. Agar electrophoresis at high tension of soluble lens proteins. Arch. Ophthal, 61: 351-360.

G•tYN•T, R. R., ANO P. H. PI•LL•PS. 1951. A modified method of the nitroprusside method of analysis for glutathione. Arch. Biochem., 30: 217-225.

G¾SELS, H. 1962. Contributions to the biochemical taxonomics of birds. Giervalk, 52: 576-585.

GYSELS, H. 1963. New biochemical techniques applied to avian systematics. Ex- perientia, 19: 107-109.

G¾s•Ls, H. 1964a. Immunoelectrophoresis of avian eye lens proteins. Experientia, '•0: 145-148.

G¾s•Ls, H. 1964b. Avian relationships studied by immunoelectrophoresis of lens proteins. Giervalk, 54: 16-28.

G¾SELS, H. 1964c. A biochemical evidence for the heterogeneity of the family Psittacidae. Bull. Royal Soc. Zool. d'Anvers, 33: 29-41.

G¾SELS, H. 1964d. Bijdrage tot de Systematiek van de Vogels, aan de hand van de elektroforese in agar van de oplosbare lens- en spierproteYnen. Natuurwet.

Tijdschr., 46: 43-178. G¾SELS, H. 1965. An electrophoretic component from the lens proteins of the

Passeriformes as an important taxonomic characteristic. J. f. Orn., 106: 208-217. G¾SELS, H., ann M. R^BaE¾. 1964. Taxonomic relationships of Alca torda, Fratercula

218 SIBLEY AND B•JS•, Arian Eye-lens Proteins [ Auk [ Vol. 84

arctica and Uria aalge as revealed by biochemical methods. Ibis, 106-' 536-539. HUNTER, R. L., ANn D. S. ST•C•A•. 1961. The esterases from mouse blood. Ann.

New York Acad. Sci., 94: 861-867. JoaNsoN, M. L., ANn M. J. W•c•:s. 1959. Serum protein electrophoresis in mam-

mals--taxonomic implications. Syst. Zool., 8: 88-95. KROTOS•:•, W. A., D.C. BENJAmiN, AND H. E. WErrd•R. 1966. Effects of starch

concentration on the solution of serum proteins by gel electrophoresis. Canadian J. Blochem., 44: 545-555.

LANCMA>•, J. 1959. Appearance of antigens during development of the lens. J. Embryol. Exp. Morph., 7: 264-274.

LA>•½MA>•, J., AND H. MA•SE[. 1962. Formation and distribution of chick lens pro- tein. Invest. Ophth., 1: 86-94.

LEON•, C. A. (ed.). 1964. Taxonomic biochemistry and serology. New York, Ronald Press.

MA•SEL, H. 1962a. The specificity of lens antigens. Fed. Proc., 21 (abstract). MA•SE[, H. 1962b. Lens antigens in ocular tissue. Arch. Ophth., 68: 254-260. MA•SEL, H., ANn M. Goon•^•. 1965. Analysis of cortical and nuclear lens proteins

by a combination of paper and starch gel electrophoresis. Anat. Rec., 161: 209- 216.

M^•SE[, H., ANn J. LAN½•^N. 1961a. An immuno-embryological study on the chick lens. J. Embryol. Exp. Morph., 9: 191-201.

MA•SE[, H., AND J. L^NCMAN. 1961b. Lens proteins in various tissues of chick eye and lens of animals throughout the vertebrate series. Anat. Rec., 140: 185-198.

MANWE[L, C., AND C. M. A. B^XER. 1963. A sibling species of sea cucumber dis- covered by starch gel electrophoresis. Comp. Biochem. Physiol., 10: 39-53.

ORNST•N, L. 1964. Disc electrophoresis. 1. Background and theory. Ann. New York Acad. Sci., 121: 321-349.

PAPACONST^NT•NOU, J. 1964. The formation of gamma-crystallins during lens fiber differentiation. Amer. Zool., 4: 279.

RABAE•r, M. 1959. Over de eiwitsamenstelling der ooglens. Ghent, Klinick voor oogziekten der Rijksuniversiteit.

R•B^E¾, M. 1962. Electrophoretic and immunoelectrophoretic studies on the soluble proteins in the developing lens of birds. Exp. Eye Res., 1: 310-316.

RAB^E¾, M. 1963. Glycogen in the lens of birds' eyes. Nature, 198: 203. R^B^E•r, M. 1964. Comparative study of tissue proteins (lens and muscle) in fish.

Pp. 273-277 in Protides of the biological fluids, vol. 12 (H. Peeters, ed.). New York, Elsevier Publ. Co.

S•{ows, T. B., R. E. TASm^N, G. S. BREWE•, A>•D R. J. DEAN. 1964. Erythrocyte glucose-6-phosphate dehydrogenase in Caucasians: new inherited variant. Sci- ence, 14S: 1056-1057.

S•B[E•r, C. G. 1960. The electrophoretic patterns of avian egg-white proteins as taxonomic characters. Ibis, 102: 215-284.

S•B[E¾, C. G. 1962. The comparative morphology of protein molecules as data for classification. Syst. Zool., 11: 108-118.

S•B[E¾, C. G. 1964. The characteristics of specific peptides from single proteins as data for classification. Pp. 435-450 in Taxonomic biochemistry and serology (C. A. Leone, ed.). New York, Ronald Press.

Sm•E¾, C.G. 1965. Molecular systematics: new techniques applied to old problems. L'Oiseau et la Rev. Frangaise d'Ornith., •: 112-124.

April] SIBLEY AND ]•RUStt, Avian Eye-lens Proteins 219 1967

SM•T•XES, O. 1955. Zone electrophoresis in starch gel: group variations in the serum proteins of normal human adults. Biochem. J., 61: 629-641.

SMiTh,ES, O. 1959. An improved procedure for starch gel electrophoresis: further variation in the serum proteins of normal individuals. Biochem. J., 71: 585- 587.

SMXT•q•ES, O. 1962. Molecular size and starch-gel electrophoresis. Arch. Blochem. Biophys., Suppl. 1: 125-131.

SrECTOR, A. 1962. A study of peptidase and esterase activity in calf lens. Exp. Eye Res., 1: 330-335.

SWANSON, A. A. 1966. The identification of lysosomal enzymes in the bovine lens epithelium. Exp. Eye Res., 5: 145-149.

van HEY•X•CrN, R. 1962. The lens. Pp. 213-287 in The eye (H. Davson, ed.). New York, Academic Press.

van HEYN•NCEN, R., ^Nl) A. Pxmr. 1953. Reduction of glutathione coupled with oxidative decarboxylation of malate in cattle lens. Biochem. J., 53: 436-444.

WrT•dO•E, A. 1960. A classification for the birds of the world. Smiths. Misc. Coll., 139(11): 1-37.

Wool), D.C., ^Nl) L. BustESS. 1961. An electrophoretic study of soluble lens pro- teins from different species. Amer. J. Ophth., 51: 305-314.

Department o] Biology, Yale University, New Haven, Connecticut, and Department oJ Zoology and Entomology, University oJ Connecticut, Storrs, Connecticut.