120 G.B. KITTO ET AL.THE JOURNAL OF EXPERIMENTAL ZOOLOGY 282:120–126 (1998)

© 1998 WILEY-LISS, INC.

Evolution of Cooperativity in Hemoglobins: WhatCan Invertebrate Hemoglobins Tell Us?

G. BARRIE KITTO,* P.W. THOMAS, AND M.L. HACKERTDepartment of Chemistry and Biochemistry, University of Texas, Austin,Texas 78712

ABSTRACT While vertebrate hemoglobins typically are tetrameric and show highly regulatedand cooperative ligand binding, little is known of the evolution of these properties. We are studyingthe structural and functional properties of the hemoglobins from Caudina arenicola, an echino-derm. The echinoderms are in the lineage most closely related to the vertebrates to express hemo-globin. C. arenicola has three sets of red cells, in the water vascular system, the coelomic cavity,and in an intestinal vein. Each of these expresses a distinct array of globins. The hemoglobins arecooperative and exhibit unusual ligand-linked associative properties, being dimeric when oxygen-ated and forming tetramers and higher aggregates on deoxygenation. The major coelomic hemoglo-bins have been subjected to a detailed examination by a combination of ligand binding analysesand protein and DNA sequencing, as well as X-ray crystallography. Two typical globin introns wereidentified, along with a unique intron that bisects an N-terminal extension of the globin from theremainder of the gene. X-ray crystallographic analysis shows that the subunit interfaces of C.arenicola hemoglobins differ radically from those of vertebrate hemoglobins and indeed from someother invertebrate hemoglobins, but closely resemble the packing arrangements found in a clamhemoglobin (Scapharca). However, the residues implicated in cooperativity in these two types ofhemoglobins differ substantially. J. Exp. Zool. 282:120�126, 1998. © 1998 Wiley-Liss, Inc.

Vertebrate hemoglobins and myoglobins are ar-guably the most extensively studied family of pro-teins in biochemistry. From them we have learnedmuch about the nature of protein folding, sub-unit-subunit interactions, the effects of point mu-tations on functionality, and about the modulationof activity by ligand binding. By contrast, muchless is known about invertebrate hemoglobins, butthey too have much to tell us, particularly withrespect to the evolution of both structural andfunctional aspects of the globin family of respira-tory proteins.

Among nonvertebrates, myoglobins and hemo-globins are found in a wide variety of organisms,from bacteria to plants and an array of inverte-brates, but the distribution is extremely scattered.Many phyla appear to lack any expressed globingenes, while in others such as the Insecta, ex-pression is very spotty with just a few speciessynthesizing these proteins. All the globin-basedrespiratory pigments share a common myoglobinfold, but the sizes of the hemoglobins range fromsimple monomers of approximately 100–150 resi-dues, such as those found in the bloodwormGlycera and Chironomid insect species, to the ex-tremely large multimeric extracellular hemoglo-bins present in annelids with molecular weights

ranging into the millions of daltons; hemoglobinswith multiple covalently linked globin domainssuch as those found in the shrimp Artemia, andyet others with both heme and flavin prostheticgroups (Candida). Comprehensive reviews of theamino acid sequences and globin gene structuresof nonvertebrates have been provided by Vino-gradov and his colleagues (’92, ’93).

Our own work has centered on the structuraland functional properties of echinoderm hemoglo-bins, in particular the intracellular hemoglobinsof Caudina (Molpadia) arenicola, a sea cucum-ber. We chose to work with this organism becausethe echinoderms are the most closely relatedgroup to the vertebrates to express hemoglobins.C. arenicola is a burrowing sea cucumber thatlives in relatively sandy or muddy sediment withonly its mouthparts and anus exposed. Among theechinoderms hemoglobins are found primarily insome burrowing sea cucumbers and a few brittle-

Grant sponsors: Foundation for Research and the Welch Founda-tion; Grant sponsor: National Institutes of Health; Grant number:GM 30105.

*Correspondence to: G.B. Kitto, Department of Chemistry & Bio-chemistry, University of Texas, Austin, Texas 78712.

Received 24 April 1998; Accepted 24 April 1998

INVERTEBRATE HEMOGLOBINS 121

stars. C. arenicola obtains its supply of oxygenby taking in sea water through the cloaca, whichworks in conjunction with an anal sphincter. Thewater is pumped into a respiratory tree throughwhich there is gaseous exchange with the coelo-mic fluid.

Interestingly, there are three separate hemoglo-bin-containing systems in C. arenicola, each withnucleated red cells of distinct morphology andcomplement of globins. The coelomic cavity con-tains by far the majority of red cells, but othersare found in the water vascular system surround-ing the oral cavity and in an apparently closedend vein wound around the intestinal tract. Ofclose to a dozen expressed globins, some areshared by all red cell types while others are re-stricted to just one of the hemal systems. Eachred cell type has distinctive oxygen binding char-acteristics (Kitto et al., ’77; Quittner, ’86).

FUNCTIONAL STUDIESWe have concentrated our efforts on the coelo-

mic red cells where four major globin types, whichwe have termed A, B, C, and D, are expressed. Inthe oxygenated form, the coelomic hemoglobinsare dimeric, and chromatographic and electro-phoretic analysis of coelomic cell lysates indicatesthat both homo- and heterodimers are present(Whitfill, ’73). While the oxy-hemoglobins aredimeric at physiological concentrations, there isa marked tendency to dissociation to monomersat concentrations below approximately 6 mg/ml.Upon deoxygenation, these hemoglobins undergoa ligand-linked association with the formation oftetramers and higher aggregates (Whitfill, ’73).While whole coelomic cells and cell lysates showmoderate cooperativity (n = 1.4) in oxygen bind-ing, and a p50 of approximately 9 mm, the singleglobin types, i.e., A, B, C, or D chains, exhibitconsiderably greater oxygen affinity (p50 ≅ 2–3mm) but still exhibit measurable cooperativity.Ligand binding kinetics, examined by rapid mix-ing, flash photolysis, temperature jump methods,carbon monoxide binding equilibria, and kineticsstudied by steady-state photochemical methods,showed a protein concentration dependence thatindicated the aggregated hemoglobin has a lowerligand affinity than do the dissociated subunits(Bonaventura et al., ’76). No effect on ligand bind-ing by organic phosphates was observed. Even atthis early point in our investigations, it was pos-tulated that the cooperativity in ligand bindingseen with the Caudina hemoglobins might havea different molecular basis than found in typical

vertebrate hemoglobins. As will be discussed be-low, this proved prescient with respect to ourrecent X-ray crystallographic studies on theCaudina hemoglobins, which are described later.When chain mixing studies were carried out withpairs of the coelomic globins, it was found thatthere was a marked increase in cooperativitywhen the D globin was added to any of the otherthree chains (Fig. 1), suggesting that the D chainlikely differed considerably in structure from theA, B, and C chains (J. Bonaventura, C. Bonaven-tura, and G.B. Kitto, unpublished observations).Immunological studies and peptide mapping(Dodds, ’77; Kitto et al., ’77) confirmed a cluster-ing of the A, B, and C chains relative to the Dchain. The data suggested a separation of thesetwo groups approximately 130 mya.

PROTEIN, CDNA AND GENOMICSEQUENCE ANALYSES

Structural studies carried out both by directamino acid sequencing (Omnaas, ’77; Mauri, ’85;McDonald, ’90; Mauri et al., ’91; McDonald et al.,’92) and DNA analyses (Thomas, ’94) confirm therelationship of the C. arenicola globins indicatedby immunological studies. Each of the C. arenicolaglobins is N-acetylated (Kitto et al., ’76), whichhampered amino acid sequencing, but helps to ex-plain the lack of effect of organic phosphates onligand binding. Glycoslylation of these hemoglo-bins also was detected (McDonald and Kitto, ’90).Each of the C. arenicola globins possesses an elon-

Fig. 1. Oxygen binding curves for C. arenicola chain Dand an equimolar mixture of chains C and D. Comparablebinding curves were obtained with mixtures of chains A andB with chain D.

122 G.B. KITTO ET AL.

gated NA region, a feature seen in a number ofother invertebrate globins, including those of an-other sea cucumber, and several vertebrateglobins (Li and Riggs, ’70; Nash et al., ’76; Comoand Thompson, ’80; Suzuki, ’89a,b). Sequencingof cDNAs derived from the mRNAs of the C andD globins (Thomas, ’94) revealed precoding re-gions of 77 and 73 nucleotides, respectively. Theseuntranslated regions are relatively long comparedwith those of other globins, where the regionsrange from 34 to 72 nucleotides. Remarkably long3´ noncoding regions were found for both globinsC and D. The 3´ untranslated segment of globinC is 837 nucleotides, while that of globin D is747 nucleotides (not including the stop codon orpoly A tail). Two other minor coelomic globins(termed E and F) were found to have even longer3´ untranslated regions (900 bp and 1100 bp,repectively). The 3´ untranslated segments of theC. arenicola globins are the largest of any globinsstudied to date. Most other globins have 3´untranslated regions of from 80 to 300 bp, al-though the Lumbricus globin C has a 3´ untrans-lated region of 559 bp (Jhiang and Riggs, ’89).Some myoglobins, such as those from humansand seals, have extended 3´ untranslated regions(Blanchetot et al., ’83; Weller et al., ’84), but noneis as long as those for the C. arenicola globins.An upstream AAUAAA sequence element asso-ciated with pre-mRNA processing was identifiedin the C. arenicola C globin mRNA, with the Upreceding the polyadenylation site by 24 nucle-otides, but the D globin lacks this sequence. Thislack of an AAUAAA sequence is rare amongglobins, but not unique. Many plant mRNAs andsome mammalian and yeast mRNAs lack thiselement (Henikoff et al., ’83; Joshi, ’87). The cod-ing regions of the C. arenicola coelomic C and Dglobins share an identity of approximately 62%(Mauri, ’85; McDonald, ’90; Thomas, ’94), whilethe 5´ and 3´ untranslated regions are consider-ably less homologous (44 and 43% identity, re-spectively). If one assumes a relatively constantrate of globin evolution, then the relatively highdegree of amino acid sequence identity of the Cand D C. arenicola globins (≥60%), taken to-gether with the fact that vertebrate α and βglobins share an identity of only 43% (Dayhoff,’72), suggests that the divergence of the C.arenicola C and D globins was a more recentevent than the vertebrate α, β divergence.

One of the constant features of vertebrate α andβ globin genes is the presence of a three-exon, two-intron structure. The sites of these introns are

very precisely constrained and the introns are ofsimilar sizes in all species examined (approxi-mately 100–140 bp for Introns I and II of the α-chains and Intron I of the β-chain; 820–900 bpfor Intron II of the β-chain). Vertebrate myoglo-bins also have two introns at positions homolo-gous to those in the vertebrate hemoglobins, buthere the introns are much longer (3,400–5,800 bp)(Hardison, ’91). By contrast with the vertebrateglobins, the intron structure of invertebrateglobins, including those from bacteria and plants,is much more varied.

A number of invertebrate globins share thethree-exon, two-intron structure found in verte-brates, with in some instances, such as plants,the addition of a third central intron. In othercases, as in several Chironomid flies, hemoglobingenes lack introns entirely, even though relatedChironomids (which are not the earliest in theevolutionary chain) have intron-containing globins(see Vinogradov et al., ’92 for a comprehensivereview of invertebrate globin genes). Because ofthe proximity of the echinoderms to the vertebratelineage, we were interested in exploring the types,locations, and lengths of any introns that mightbe present in the C. arenicola globins. For thispurpose we employed exploratory probes basedon the cDNA globin sequences we had determined.Pairs of primers were constructed for the regionon each side of the two common vertebrate in-trons and for the central intron found in plantsand some lower invertebrates. These were usedto PCR amplify these regions from C. arenicolacoelomocyte DNA. An overview of our results ispresented schematically in Fig. 2. For hemoglo-bins C and D, Intron I sequences of 932 and 241bp, respectively, were identified. The splice siteswere within the codon for amino acid 41, exactlyas expected by analogy with vertebrate globingenes. Similarly, an Intron II was identified inboth C and D C. arenicola globins, with lengths

Fig. 2. A schematic of the intron structure of the C.arenicola hemoglobin C gene.

INVERTEBRATE HEMOGLOBINS 123

of 671 and 676 bp, respectively, located betweenamino acids 115 and 116, as in vertebrate globins.No evidence for a central intron, of the plant type,was found in the C. arenicola globins. Riggs andhis colleagues recently identified a novel intron ina globin gene from the clam Barbatia reeveana(Naito et al., ’91), located just two bases upstreamfrom the start codon. To determine whether sucha precoding intron might exist in the C. arenicolaglobins, we designed appropriate primers to spanthis region. Unexpectedly, we discovered a long(≅2,200 bp) intron, not in the precoding region, butwith a splice site within the codon for amino acid10 of the coding region. This intron neatly dividesthe N-terminal extension found in some inverte-brate and lower vertebrate globin genes from theremainder of the coding region (Thomas, ’94). Theevolutionary origins of this unique intron are un-clear, but it appears possible that it could havearisen by an alternative splicing of a 5´ intron suchas that found in B. reeveana. These precoding in-trons themselves may have arisen by a conserva-tive transposition into a 5´ untranslated region.

X-RAY CRYSTALLOGRAPHIC STUDIESThe functional and sequencing studies of the

C. arenicola hemoglobins also have been comple-mented by a series of X-ray crystallographicanalyses. Initial work concentrated on the struc-ture of the monomeric hemichrome form of he-

Fig. 3. C. arenicola hemoglobin C, monomeric hemi-chrome form.

Fig. 4. Hemoglobin homodimers from (top) Caudinaarenicola (sea cucumber), (center) Scapharca inaequivalvis(clam), and (bottom) Urechis caupo (innkeeper worm).

124 G.B. KITTO ET AL.

moglobin C (Carson et al., ’79; Mitchell et al.,’95a,b). The dimeric oxy-forms of the coelomic C.arenicola hemoglobins (high-spin Fe2+) dissociateto monomers when oxidized to the met-forms, andthese in turn rapidly and spontaneously convertto hemichromes (low-spin Fe3+) in which both thefifth and sixth coordination sites of the Fe arebound to the imidazole groups of histidines. Thehemichrome hemoglobin C was solved using mul-tiple isomorphous replacement and refined to 2.5Å resolution. It shows the classic myoglobin foldstructure with eight α-helices, and a pocket forthe heme group between the E and F helices. Un-usual features of this molecule are the relativelylong N and C terminal tails, which are relativelyunstructured (Fig. 3). The structure of a dimericform of the C. arenicola hemoglobins also hasbeen determined, using in this case cyanomet-liganded hemoglobin D. The structure was solvedusing molecular replacement techniques and re-fined to 2.9 Å resolution (Mitchell et al., ’95b,c).The monomeric unit in this dimer was similar inoverall structure to the hemoglobin C monomer,except that the D hemoglobin lacks a D helix, be-ing in this respect similar to the α-chain of ver-tebrate hemoglobins. As was anticipated, theliganded iron moves more into the heme planethan in the unliganded state. By far the most in-teresting aspect of the dimer structure, however,is the nature of the intersubunit contacts of thiscooperative hemoglobin (Fig. 4). In marked con-trast to the dimer contacts found in vertebratehemoglobins, the C. arenicola D dimer has sub-unit-subunit contacts involving the E and F heli-ces, which brings the heme groups in proximity(≅ 19 Å) and which places the G and H helices onthe surface. Thus, one must expect a totally dif-ferent mechanism for the cooperativity of thisinvertebrate hemoglobin than found in its verte-brate counterparts. In the homotetrameric hemo-globin from the innkeeper worm Urechis caupo(Kolatkar et al., ’92, ’94), the F and E helices areinvolved in subunit contacts like the C. arenicoladimeric hemoglobin. In Urechis, the helices arealmost at right angles to each other, comparedwith an orientation of approximately 50 degreesin the C. arenicola protein. The C. arenicola Ddimer bears a much stronger resemblance to thedimeric hemoglobin from Scapharca inaequivalvis(Royer et al., ’89; Royer, ’94) (Fig. 4). This simi-larity extends to the general nature of the dimerinterfaces (Fig. 5), which both exhibit close heme-heme contacts. Royer (’94) proposed a detailedmechanism for the cooperativity of the S. inaequi-

Fig. 5. Comparison of the dimer interfaces of Caudinaarenicola hemoglobin D (top) and Scapharca inaequivalvishemoglobin (bottom).

INVERTEBRATE HEMOGLOBINS 125

valvis hemoglobin, which differs greatly from thatfor vertebrate hemoglobins. However, a close ex-amination of the residues lining the intersubunitcontact regions of the C. arenicola dimer revealsthat this mechanism is inapplicable here sincethe key residues implicated in the Scapharcamechanism are not present in the sea cucumberhemoglobin.

CONCLUSIONSClearly much more remains to be examined with

respect to these unusual sea cucumber hemoglo-bins. It will be of interest to determine the exactnature of the specific subunit interactions inmixed chain oxy-dimers, both with and withoutthe D chain, which we can only infer from ourcurrent knowledge of the homodimeric cyanomet-Hb-D and monomeric hemichrome-Hb-C struc-tures. Likewise, we hope to crystallize and solvethe structure of the tetrameric deoxy forms of thecoelomic C. arenicola hemoglobins. Of more gen-eral interest, the cloning of the C. arenicola globingenes provides us with some important tools tofurther explore the nature of hemoglobin evolu-tion. As noted earlier, hemoglobin is expressedonly sporadically among the invertebrates. In anumber of lineages, periods of hundreds of mil-lions of years appear to have passed with hemo-globin genes being silent, only for them to bereactivated when needed. Are these genes reallypresent in the nonhemoglobin expressing species,and if so, how is functionality maintained for ex-tremely long periods in the absence of any selec-tive pressure? Perhaps the globin genes areexpressed at minimal levels and these proteinsare used for purposes other than oxygen trans-port and storage. Or is it possible that those spe-cies expressing hemoglobins gain the necessarygenes by horizontal transfer? The C. arenicolagene sequences we have determined should allowus to construct appropriate probes to investigatethese questions, initially through examination ofsea cucumbers closely related to C. arenicola, butwhich do not express hemoglobins, and then withmore distantly related organisms.

ACKNOWLEDGMENTSWe acknowledge the contributions of the many

undergraduates, graduate students, post-doctoralfellows, and faculty colleagues who have contrib-uted over a number of years to the studies de-scribed here. Their numbers are too great for usto provide our individual thanks, but we expressour particular appreciation for the encourage-

ment, collaboration, and critical evaluations ofDrs. Joseph and Celia Bonaventura, Duke Uni-versity, and Dr. Austen Riggs, The University ofTexas at Austin.

LITERATURE CITEDBlanchetot, A., V. Wilson, D. Wood, and A.J. Jeffreys (1983)

The seal myoglobin gene: An unusually long globin gene.Nature, 301:732–734.

Bonaventura, C., J. Bonaventura, G.B. Kitto, M. Brunori, andE. Antonini (1976) Functional consequences of ligand-linkeddissociation in hemoglobin from the sea cucumber Molpadiaarenicola. Biochim. Biophys. Acta, 428:779–786.

Carson, W.M., T.R. Bowers, G.B. Kitto, and M.L. Hackert(1979) Preliminary crystallographic data on monomeric anddimeric hemoglobins from the sea cucumber Molpadiaarenicola. J. Biol. Chem., 254:7400–7402.

Como, P.F., and E.O.P. Thompson (1980) Amino acid sequenceof the α-chain of the tetrameric haemoglobin of the bivalvemollusc Anadara trapezia. Austra. J. Biol. Sci., 33:653–664.

Dayhoff, M.O. (1972) Atlas of Protein Sequence and Struc-ture (Data Section), Vol. 5, National Biomedical ResearchFoundation, Washington, D.C., p. 64.

Dodds, J.M. (1977) Immunological comparison of invertebratehemoglobins. Ph.D. dissertation, The University of Texasat Austin.

Hardison, R.C. (1991) Evolution of globin gene families. In:Evolution at the Molecular Level. R.K. Selander, A.G.Clark, and T.S. Whitman, eds. Sinauer, Sunderland, MA,pp. 272–290.

Henikoff, S., J.D. Kelly, and E.H. Cohen (1983) Transcriptionterminates in yeast distal to a control sequence. Cell,33:607–614.

Jhiang, S.M., and A.F. Riggs (1989) The structure of the geneencoding chain c of the hemoglobin of the earthworm, Lum-bricus terrestris. J. Biol. Chem., 264:19003–19008.

Joshi, C.P. (1987) Putative polyadenylation signals in nucleargenes of higher plants: A compilation and analysis. NucleicAcids Res., 15:9627–9640.

Kitto, G.B., D. Erwin, R. West, and J. Omnaas (1976) N-ter-minal substitution of some sea cucumber hemoglobins.Comp. Biochem. Physiol., 55B:105–107.

Kitto, G.B., J. Dodds, J. Omnaas, W.M. Carson, and M.L.Hackert (1977) Hemoglobins of the sea cucumber Molpadiaarenicola. Fed. Proc., 36:758.

Kolatkar, P.R., S.R. Ernst, M.L. Hackert, C.M. Ogata, W.A.Hendrickson, E.A. Merritt, and R.P. Phizackerley (1992)Structure determination and refinement of homotetramerichemoglobin from Urechis caupo at 2.5 Å resolution. ActaCrystallogr. B, 48:191–199.

Kolatkar, P.R., M.L. Hackert, and A.F. Riggs (1994) Struc-tural analysis of Urechis caupo hemoglobin. J. Mol. Biol.,237:87–97.

Li, S.L., and A. Riggs (1970) The amino acid sequence of he-moglobin V from the lamprey, Petromyzon marinus. J. Biol.Chem., 245:6149–6169.

Mauri, F.C. (1985) The structure and functional studies of anechinoderm hemoglobins. Ph.D. dissertation, The Univer-sity of Texas at Austin.

Mauri, F., J. Omnaas, L. Davidson, C. Whitfill, and G.B. Kitto(1991) Amino acid sequence of a globin from the sea cu-cumber Caudina (Molpadia) arenicola. Biochim. Biophys.Acta, 1078:63–67.

McDonald, G.D. (1990) Sequence and evolutionary studies on

126 G.B. KITTO ET AL.

hemoglobin from the sea cucumber Molpadia arenicola.Ph.D. dissertation, The University of Texas at Austin.

McDonald, G.D., and G.B. Kitto (1990) Glycosylation of inver-tebrate hemoglobins. Comp. Biochem. Physiol., 96B:235–237.

McDonald, G.D., L. Davidson, and G.B. Kitto (1992) Amino acidsequence of the coelomic C globin from the sea cucumberCaudina (Molpadia) arenicola. J. Protein Chem., 11:29–37.

Mitchell, D.T., S.R. Ernst, W.-X. Wu, and M.L. Hackert (1995a)Three-dimensional structure of a hemichrome hemoglobinfrom Caudina arenicola. Acta Crystallogr., D51:647–653.

Mitchell, D.T., G.B. Kitto, and M.L. Hackert (1995b) Structuralanalysis of monomeric hemichrome and dimeric cyanomet he-moglobins from Caudina arenicola. J. Mol. Biol., 251:421–431.

Mitchell, D.T., S.R. Ernst, and M.L. Hackert (1995c) X-raystructure determination of a dimeric hemoglobin fromCaudina arenicola. Acta Crystallogr., D51:760–766.

Naito, Y., C.K. Riggs, T.L. Vanderoon, and A.F. Riggs (1991)Origin of a “bridge” intron in the gene for a two-domainglobin. Proc. Natl. Acad. Sci. USA, 88:6672–6676.

Nash, A., W. Fisher, and E.O.P. Thompson (1976) Haemo-globins of the shark, Heterodontus portusjacksoni. II. Aminoacid sequence of the α-chain. Aust. J. Biol. Sci., 29:73–97.

Omnaas, J.R. (1977) Sequence studies on a hemoglobin fromMolpadia arenicola. Ph.D. dissertation, The University ofTexas at Austin.

Quittner, A.L. (1986) Comparison of the hemoglobins fromthe three vascular systems of the sea cucumber M. arenicola.M.A. thesis, The University of Texas at Austin.

Royer, W.E. Jr., W.A. Hendrickson, and E. Chiancone (1989)The 2.4 Å crystal structure of Scapharca dimeric hemoglo-bin. J. Biol. Chem., 264:21052–21061.

Royer, W.E. Jr. (1994) High resolution crystallographic analy-sis of a co-operative dimeric hemoglobin. J. Mol. Biol.,235:657–681.

Suzuki, T. (1989a) Amino acid sequences of a major globinfrom the sea cucumber Paracaudina chilensis. Biochim.Biophys. Acta, 998:292–296.

Suzuki, T. (1989b) Amino acid sequence of the monomer sub-unit of the giant multisubunit hemoglobin from the earth-worm Pheretima sieboldi. Eur. J. Biochem., 185:127–134.

Thomas, P.W. (1994) Genomic analysis and evolutionary stud-ies on an echinoderm hemoglobin family. Ph.D. disserta-tion, The University of Texas at Austin.

Vinogradov, S.N., D.A. Walz, and B. Pohajdak (1992) Organi-zation of non-vertebrate globin genes. Comp. Biochem.Physiol. B, 103:759–773.

Vinogradov, S.N., D.A. Walz, B. Pohajdak, L. Moens, O.H.Kapp, T. Suzuki, and C.N.A. Trotman (1993) Adventitiousvariability? The amino acid sequences of nonvertebrateglobins. Comp. Biochem. Physiol. B, 106:1–26.

Weller, P., A.J. Jeffreys, V. Wilson, and A.J. Blanchetot (1984)Organization of the human myoglobin gene. EMBO J.,3:439–446.

Whitfill, C.E. (1973) The isolation and properties of hemoglo-bins from the sea cucumber Molpadia arenicola. Ph.D. dis-sertation, The University of Texas at Austin.