More precise characterization of pos- 3, r;. Matsumura and M. Ilayashi, Science 153, animals. After whole chain analysis, 757 (1966); F. Matruniura and K. C. Patil siblc DDT molecular association phe- 121 (1969); F. T, A: both variant a chains wcre distinctly nomcna could be obtained from studies Br;ltkowski, K. C. Patil, Ball. h'nviron. unil~ual in thc proportions of particu- Contam. Toxicol. 4, 262 (1969); 11. B. Koch, of solution behavior in which nuclear J . Neurocke,n. l(i, 269 (1969). lar amino acids among the total of magnetic resonance spectrometry (13, 15) is used and from correlations of colligativc properties with charge- transfer charactcristic~ of appropriate molecular complexes of pesticides. W. E. WILSON L. FISHBEIN S. T. CLBMENTS N~tionul 1n)titrlte of Environmental Hpalth Sciences, Research Triungle Park, Nor tk Carolina 27709 References and Notes 1. Abhrcviations arc: DDT, l,1,1-tricllloro-2,2- !?ir(r~-chloroplicnyI)etIia~~e: DDD. 1,l-dichloro- 2.2-bis(p-c11loroplienyI)ct11ane; YDMS, I-chlo- ro-2.2-his@-chlorophcny1)cthanc; DDA, 1.1- bis(p-clilorop11cn~~l)ncctic acid. 2. T. Na~.aliaslii and 11. G. Ilass, J . Grn. Phy- .%id. 51, 177 (1968); B. Hille, ibirl., p. 199. 4. R. M. Welch, W. Levln, A. H. Conney, Toricol. Apyl. PJiairnacol. 14, 358 (1969). 5 4 H. C onncv. K. M Welch. K Kunt/man. ., J . . J . Burns, Clin. ~1zarmar;l. Thrr. 8, 2 (1967). 6. G. Flolan, Nature 221, 1025 (1969); L. J. Mullins, Cli~nr. Rev. 1954, 280 (1954). 7. R . D. O'Bricn and F. Matsumnra, Science 146, 657 (1964). 8. F;. Matsumura and R. D. O'Brien, J . Agr. Food Chem. 14, 39 (1966). 9. P. 0. P. Ts'o, Ann. N.Y. Acarl. Sci. 153, 785 (1969). 10. .I. C. Skou. Biockiin. Biophys. Acta 58. 314 (1962). 11 D. P Stcven\on and G . M. Coppingel, J . Amer. CIiem. Soc. 84, 149 (1962). 12. C. I. Andrcws and R. M. ~ c e ~ c r , Moleclzlar Complrxes in Organic Chemistry (Holdcn- Day, San Francisco, 1964). 13. R. T. Ross and P. J. Biros, Biochem. Bio- phys. Krs. Comnrrm. 39, 723 (1970). 14. I. Walkley, D. N. Glcn, J . H. Hildchrand, J . Chem. J'hys. 33, 621 (1960). 15. R. Il;~que, W R. Coshow, L. F. Johnson, J . Anler. Chem. Soc. 91, 3822 (1969). Silent Hemoglobin Alpha Genes in Apes: Potential Source of Thalassemia Abstract. Small quantities of unusz~al hemoglobins were found in I o f 37 chimpanzees and 2 of 6 gorillas. In each genus these hernoglohins contain unique a chains that difler from the ordinary hy eight to nine sctrttered amino acid chaiiges. Tlie ur1usual chains arise from a hitherto undetected hemoglobin S a locus. No " t u products are fo~lnd iiz most apes; accordingly, "a is considered syiitheticc~lly inactive in ~ l l hut a few reversion mrttcznts. Indirect evidence that the inactive % locus is juxtaposed to an active a locus together with the supposi- tion that .'a exists in nzan provides a setting wherein thalassenzia rnight he pro- d ~ i c ~ d Oy rzoiil~o~~rologous rccombinutioii between two loci. S~lent genes, that is, genetic loci with- O L I ~ demonstrablc prod~rcts in most indi- viduals of a specie\, havc not been heretoforc idcntrfied in higher orga- ni\ms. In this report we provide rcawns for believing that a silcnt locus, termed her,roglohin :Gx, existr in grcat apcs and probably also in man. In most ind~vid- uals ;!a seems to be inactive and pro- duces no evident product; howcvcr, in a few mutants the locus is active and produces an unusual a chain. Dur~ng an clcctrophoretic survey of adult liernoglobins from great apes, thrcc exceptional animals wcre en- countered. One of 37 ( I ) chin~panzees (Par7 troglocly~es) and 1 of 5 ( 1 ) un- related lowland gorillas (Gorilla gorilla gorilla) cxhibitcd not only hemoglobins A and A, but also small quant~ties (2.4 to 3.4 percent, Fig. 1 legend) of an unusual form of hemoglobin A and still smaller quantities (0.04 to 0.1 percent) o l an unusual A,. Both components differed from the wual by a net g,iin of about four elcctro- static changes per hemoglobin mole- cule. Idcntical amounts of these unusual cornponcnts wcrc alro found in the son of the variant gorilla (2) . Elcctro- pllorc\is of isolated ( 3 ) co~lccntratcs of the principal unrls~~al componentu, dcsignatcd hemoglobin Hyzoo in the chimpanzee and hemoglobin Wazoo (4) in thc gorilla, are shown in Fig. 1 ( 5 ) . Parallel variation of both hemo- globin A (ru,P,) and A , (ru,S,) in all affccted individualr ruggerted that an unusual (X chain was prcrcnt in thcrc animals. This was corroboratcd by column chromatographic separation of conrtitutive hemoglobin chains ( 6 ) . Chromatographic behavior and amino acid composition of /3 chains from both Hyzoo and Wazoo were identical to A-P. In contrast, thc a cha~ns of Hyzoo and Wazoo cach showed net gains of about two electrostat~c changes when compared w ~ t h A-a from variant 14 1 residues present ( 7 ) . The differences bctwecn variant a and A-a sequences were further clis- sectcd through amino acid analysis ot purified tryptic peptides (7). The net number of various residues realized from the sum of tryptic peptidcs exact- ly matched those obtained by whole chain analysis, thereby suggesting that character~zation of variant chains is rea5onably complete. A synopsis of dif- fcrcnccs is shown in Fig. 2. A remark- able feature-pivotal to our later in- terpretation-is the similarity between chimpanzee (Pan) Hyzoo-a and Goril- la Wazoo-a. These chain, share a pre- st~mcd constellation of eight scattered amino acid diffcrcnces, outlined in Fig. 2, wlth respect to the A-a sequences charactcristic of each genus. The extent and diffuse distribution of d~ffcrences shown In Fig. 2 makc ~t most unlikely that cithcr Hyzoo-a or Wazoo-a, let alone both, have arisen s11nply as allelic mutat~ons at tho locur for A-(2. Detectable hemoglobin mu- tants differ from wild-type alleles, either through changes in one nucleotide or, In a few instances, through dclctlon of \hart runs of nuclcotldes in multiple\ of three ( 8 ) . Aside from a few in- stances of within-locus recombillants between two separate nucleotide ch,~nge\, multiple scattercd changes are not found 'Inlong uncommori v'lrrants. M~lltiple scattered dlffel-enccs may, however, dcvclop between common al- leles (9) whcn thcx havc been main- tained by natural rclection tor millions of gcncrations. In this connection both FIy7oo and Wnzoo arc distinctly un- common; nothing l ~ k c them was dc- tccted in other rulveys involving samples from substantial numberr of great apcr (10). As persistently rarc allcler at thc locus for A-ru thcsc vari- ants would, by Fisher's prediction ( I ] ) , be lost long beforc thcy could accumu- late stcp by stcp the pattern of change \how11 in Fig. 2. Accordingly, Hy7oo-a and Wa7oo-a can only be regardcd au the products of an ru locus that is sep- arate from the locus for A-ru. It is lrkely that this additional a locus has a com- rnon ancertry in the two spccics, that is, rt arorc from a single gene duplication in some common ancestor of apes. Although six of the cight positions wherein Hyzoo-a and Wazoo-a are seemingly alike and dlffercnt from A-(X 182 SCIPNCE, VOI.. 171 havc not becn definitively placed by exact sequcnce analysis, the pattern of differences shown in Fig. 2 is nonethe- less sufficient to support bclief in a com- mon origin. In the context of later in- terpretation a commonality of anccstry is all that matters; whether prescnt day Hyzoo-a and Wazoo-a differ as shown (Fig. 2) by two changes or in fact by, say four changes, is immaterial. In man there appear to be two gea- erally indistinguishable a loci in some (12), but not necessarily all (13), popu- lations. Whcre there are two activc a loci, cach locus is thought to produce about one-half and cach allele about one-fourth of all a chains. An analogous statc of affairs may cxist in chim- panzees where a hcalthy individual was found to havc about 25 pcrccnt of an clcctrophoretically fast n~oving n chain variant (10, 14 ) . Wc designate the syn- thetically vigorous a loci as la and ha, but in so doing recognize that only onc of these loci may bc activc in some mcn and pcrhaps in all gorillas. The synthetically impoverished (5) locus rc- sponsible for Hyzoo-a and Wazoo-a is designated %. Ternlinology follows that adopted for the several Izernoglohirz y loci of man (15) and allows nlolecular formulas to be written -in an unambigu- ous fashion, for example, hcmoglobitz A: mixture of l ~ , A * p , A and 2tu,A.p,A; hcmoglobin Hyzoo: b2TTrzoo'P,*. The hcmoglobin "a locus, like I n and %, presumably arose through genc du- plication via the successive processes of nonhomologous meiotic pairing and rc- combination. The antiquity of the cvents producing "a may be judged first by the appearance of % chains in cach of two genera and second by the minimum of nine to ten nuclcotidc changes calcu- lablc from genetic code for the amino acid differences (Fig. 2) between A-a and the a chains of Pan Hyzoo and Gorrilla Wazoo. Thc first observation indicates that the age of % antcdatcs the separation of evolutionary lines leading to chimpanzee and gorilla, whcrcas the second finding suggests 'a has had a much longer history. The nine to ten n~icleotide changcs cxceed the minimum of four nuclcotidc diffcr- cnccs cxisting bctwecn thc a gcncs of man and Rhesus monkey (8) and, in addition, approximate the minimum of 9 to 14 ilucleotide difTercnces existing bctwecn hcmoglobin ,I3 and 6 gcnes in several primates (14). The latter two loci probably arose before the ancestors of apcs and New World monkeys had diverged from one another (14, 16). Fig. 1. Amido schwarz stain of pH 8.6 EB'T [EDTA, boric acid, tris ( 9 ) ] starch- gel electrophoresis of chromatographically isolated (3) concentrates of great ape hemoglobins: 1, chimpanzee hemoglobin A:; 2, chimpanzee hemoglobin Hyzoo; 3, chimpanzee hemoglobin A; 4, gorilla AL; 5, gorilla Wazoo; 6, gorilla A; 7, whole hemolyzate from gorilla. All chimpanzee samples are from animal No. 9, in whom Ilyzoo forms 3.4 k 0.4 percent (iV = 3 ) of total hemoglobin. All gorilla samples come from Tomoka in whom hemoglobin Wn7oo forms 2.5 percent of the total (com- pare 2.4 pelcent in his father, Nikumba). The PI opo~tion of Ag is 1.8 percent in No. 9 and 2.0 percent in Tomoka. V refcrs to variant hemoglobin. In light of such comparisons the 'a locus seems quite old and one that man is entitled to by right of descent. If thc 'a locus is indccd ancient and widespread how has it remained hidden from view, and how has it become vis- ible in uncornmon individuals from two different genera? There arc threc alter- native explanations. First, the % locus may only occur in a few individuals. Sccond, ::a may be present and activc in all, but its product in most may lie clcc- trophorctically and chron~atographically buricd within the cnvclopc of hcmo- globin A. Under thcsc conditions sparse quantities of %Y chain might easily have cscapcd detection during separation and analysis oC A-a peptides. Third, %a may he prcscnt but synthetically inactive in most individuals because long ago, in some com~non ancestor of apcs, it underwent, for cxamplc, a nonscnse or a profound missensc mutation that has only bccn overcome in a few back mutants. As we shall indicatc, the first explanation sccms decidedly improb- able, the second explanation can be eliminated, and thus the third explana- tion is favorcd. The first cxplanation, limitation of "a locus to a few individuals, is sta- tistically unattractive. I t requires the ,preservation for millions of generations !of what would now bc a rare locus in chimpanzces and, at best, an uncommon locus in gorillas. Just as noted for thc case of rare alleles, the chance for loss of a rare locus is enormous in each species (11) . In the spccific case of a rare locus the opportunity for loss is compoundcd by risk of extinction dur- ing obligatc mispairing that must occur in cach carricr during each meiotic di- vision. I t is anticipated from thc second ex- planation that uncommon variant ani- mals are probably hctcrozygotes, that is, :iaA/3alIyzoo or 3,A/:la\\ :woo. Tf this is thcn it is supposed that Hyzoo and Wazoo havc bccome elcctrophoretically visible through single but scparate 'a mutations producing nct gains of about four electrostatic chargcs per hcmo- globin n~olcculc (Fig. 1). Thc only re- motely appropriate candidates (Fig. 2) for such charge changes are aspartic acid to lysinc (Asp; Lys) mutations at R,.~D~ (17). Under the tcrms of this cx- planation hypothetical wild-type "a" allcles produce and their prod- ucts arc buricd in the mass of hcmo- globin A; but the variant :3a1'rz00 and :ia""zoo alleles arc concordallt mutants that produce : ' a G i Ly911d are thereby elcctrophorctically visiblc. The undoing of this hypothesis comes from starch- gcl clcctrophoretic analysis at p H 7.1 (la), whcrc thc imidazole group of histidinc is more positively charged than at pH 8.5. Differcnccs in histidine con- tent (Fig. 2) between A-a and thc a of Hyzoo and Wazoo that are nearly silent at p H 8.6 arc cxpresscd at pH 7.1. The same should be true for pH 7.1 electrophoretic comparisons involv- ing products of hypothetical %A. In an cxplanation that accounts for Hyzoo and Wazoo visibility through the occur- rcncc of single but separate muta- tions, it is expected that othcr rcsiducs, including those involving diffcrenccs in histidine, will remain unchangcd. As alrcady noted, most known diffcrellccs betwecn normal and mutant alleles in- volve single, not n~ultiple, residues (8). Despite such expectations, bascd on differences bctwcen the histidinc con- tent of A-a and hypothetical %*, no new products appear after p H 7.1 electrophoresis of hcmolyzates and hemoglobins from a nun~ber of difercnt chimpanzces and gorillas. Failure to uncovcr buricd products occurs despite the unexpcctcd (19) finding that Hyzoo and Wazoo are slightly lcss n~obile than A at pH 7.1. Thus a buricd %aA prod- uct-if it existed and differed from both Hyzoo and Wazoo only at a"-should be distinctly lcss mobilc than A at p H 7.1 and easily discernible. For such 15 JANUARY 1971 Pig. 2. Minimum amino acid differences bctweeii hemoglobin A-a chains of man-Pan-Gorilla ( 8 ) and the variant a chains \ i n 23 64 6ib 67 70' 102 1 0 5 ~ 1 0 6 ~ 110. lid peptlde~ (7). Positions 65-90 derive f ~ o m trypt~c pepl~des a-9b; postttons 105-1 27 a] e f~ om peptide a- 12b. Distinction between actds and amides depcncls on peptide eIectrophoresis. Posit~on aswnments ale infe~ied f ~ o m homology w~th human A-a se- of hcmoglobin Hyzoo from chimpanzee Pan, animal No. 9, Chain \ and hemoglobin Wazoo from Gorilla, animal Nikumba. Posi- tions where all chains are similar are omitted. Results in Hyzoo- A - a a and Wazao-n are based on amino acid compositions of tryptic quence. The superscript letlcrs indicate-the following: (a) Cilu in man md ~ u f i A-a, ASP in Coii//a ~ - a ( 8 ) ; (b) position? 6.5, GM ~ 0 z o 0 - a 60, 71, 79, 82, or 88; (c) position 70 or 73; (d) diffcrences in- volve any two of pocitions 105, 106, 109, 113, and 125; (c) pocttions 110, 11 1, 115, 120, or 123; (f) positions 112 or 122. Boxcs dcnote the recidues where Hyzoo-a and Wazoo-a arc alike and different from A-a. Abbreviation\: Ala, alaninc; Asn, asparagine; Asp. aspartic acid; Glu, glutamic acid; HIS, histid~ne; Ly\, lycinc; Phc, phenylalanine; Ser, scrine; 'Thr, thrconine; Val, valine; x, not asce~taincd whether ac~d or amide. GIU" Asp Asp Ala Thr Val Ser Leu Leu Ala HIS rea\ons we discount the possibility of an active *'n locus with a hidden product and favor the third explanation, namely, a :IN locus that is us~ially synthetically inactive. Consequently, hemoglobins Hyzoo and Wazoo are each regardcd as the product of a mutation whose effect is to rcvcrse a regulatory muta- tion, for cxample, a nonsense or a pro- found missense n l~~ta t ion for which 3a is usually hon~ozygo~rs. Although our clctection of such concordance of re- versal in scparate gencra scems remark- able we do not know whether the samc kind of back mutation has been opera- tive in each genus. For example, a non- scn\e mutation might be reversed either by back mutation at :b o r by a sup- pressor mutation in a transfer RNA that translates the nonsensical codon of "a. Thus concorctance of revcrsal may be less cxact and a little less remarkable than it first seems. In terms of thalassemia the signifi- cance of an inactive "a locus depends on the assumption that an array of two or more nearly identical and genetically juxtaposed loci--for example, la, %a -favors recurrent meiotic mispairing (20). In fact, two findings suggest that meiotic mispairing betwecn adjacent a loci nlay have occurred in Gorilla dur- ing comparatively reccnt times. First, Gorilla is unique among primates, and in particular anlong apes, in possession of A-(yZ3 AFT' ( 8 ) ~ This individuality pre- sunkably reflects an a2" C'U+ A" muta- tion (Glu, glutamic acid) in the com- paratively short evolutionary period since taxono~nic separation of great apes. Second, ruxLi5p occurs in both the A-a and Wa7oo-a of Gorilla (Fig. 2). Although such concordance at aX%ay reflect isologous mutations, there has hcen relatively little evolutionary time for a pair of these to develop. It seems inore likely that a single Glu + Asp inutation occurred at a':: in one CY locus followed by its insertion into another and ncighhoring a locus t h r o ~ ~ g h non- homologous crossing over. I t Is notable that "a is left silent in the process. Ac- cordingly-if our rccon~tructions are correct--the ~ynthetic ~ilcnce of does not depend on information contained in the ei~rly portion of its message. By extenhion we presume that 3 a silence dcpcncls on information following spcc- ification of a". It may thus bc expected that a portion of new intergenic rccom- binants beginning as 'a or 2a and cnd- ing as % will be silent. Where only 'a or % are present, hetcrozygotcs for a nonhomologou~ recombirlant fuqion of two loci, for cxample, la-", will have only one synthetically active a locus. Homozygotcs will have nonc. Such hy- pothetical la-%/Ia-"a homozygotes for a fusion gcne might be a source of a lethal form of a thalassemia associated with hydrops fetalis. Infants with this conctition have no demonstrable a chain synthesis (2I) . This postulated origin of a thalassemia is in some ways anal- ogous to the P-6 fusion recombinant producing hemoglobin Lepore (22) and as\ociated in hotnozygotes with /3 thal- assemia. Finally, a thalassemia might also develop even if the discounted notion of an active hut synthetically impoverished %I locus, discussed carlier, is correct. In this situation we suspect that only small quantities of henlo- globin, possibly similar to Nyzoo and Wazoo and perhaps indistinguishable from hemoglobin A,, would be pro- duced by a lor-3a fusion locus. What- ever the case, we reiterate that the sig- nificance of 3a as a source of thalas- scmia lies within its supposed tight juxtaposition to an active a locus. In this setting the frequency with which mciotic mispairing lcads to loss of a synthesis is likely to be common (20) and may greatly exceed the frequency with which point mutations lead to the same condition. The virtue of thc ad- mittedly elaborate :h: nod el thus lies with its potential rate of appearance: a feature that may help to account lor the stcadily increasing evidcnce of het- erogcncity among thalasscn~ics. In all such interpretations, the crucial 1111- known is the question of b cxisterlce in man. Whilc we clo not know whether or not :'a cndurcs in ourselvcs, it noac- thelcss seems likcly that a locus that bas apparently perqisted for pcrhapc 40 million ycars or more in our remote anccqtors will have survived the last scveral million since our taxonomic divergence from the ape stem line. SAMUEI. H. BOYBK ANDREA N. NOYES GEORGE R. VRABLITC LOIS J. DONALDSON EDWARD W. SCHAEFTIK, JR. Dil~ision o f Medical G~iretics and Clayton La/~ortrtories, Department of Mpdicine, .lohirs Hopkins Univecsity School of Medicine, Btrltinzore, M~irylancl2 1205 CI,INTON W. GRAY National Zoologicnl Park, Smithsonian Institzition, Wasiziizgtor7, D.C. 20009 TIIEODOKE F. THUI~MON Department o f Prdiutrics, Loriisiana State University School of Medicine, New Orleans, Loui.viuna 701 I2 References and Notes 1. Chimpanzee samples derive from the :Zener- nsity of J. Moore, Baltirliore Zoo: f i~ur . animals; Dr. A. Eldadah, Johns Hopkins University Scbool of Hygiene: nine animals including the Hyzoo individual (No. 9) since transferred and bled for 11s by Dr. C. Oibbs, Patuxent, Md.; Delta Rcgional Primate Ccr~. tcr. Covington, 1.a.: 24 animals. Gorillas cx- anlined include Jacky, Hercules, and Sylvia from the Baltimore Zoo; and F:cmclle, Nikumba, and Tomoka from Smithsonistl's Washington Zoo. 2. Ilematocrits, llcmoglohin concentrations, erytlt- rncyle concentrations, ergtllrocyte morphol- ogy, and A, percentages in tlte variant apcs we]-e unremarkable. Accordingly, there is no rcason lo suspcct that the variant hemoglobins in themsclvcs prodnce disease. 3. A. M. Dozy, 13. F. Kleihaner, T. H. J. fIuis- man. J. Chron~atogr. 32, 723 (1968). 4. Specific names arc given to uncommon hemo- globin variants Irom great apes. These names 184 SCIENCE, VOL. 171 are hyphenates with the suffix zoo being used to denote the usual source and to emphasize that the variant occurs in a nonhuman subject. The prefix denotes the placc of first sampling. Thus Wazoo is found in gorillas at the Washington Zoo and Hyzoo was found in a chimpanzee in the Johns Hopkins University School oE Hygicne zoological collcction. 5. The observed quantities of hcmoglobin EIyzoo and Wazoo present in wholc hemolyzates are little more than the 1.6 to 2.0 percent of A, prcsent in chitnpanzccs and gorillas, and, motaovcr, considerably less than the amounts of variant hemoglobin found in almost any of a variety of human hcmoglobin heterozy- gote.: (14). The source of such scant quantities of Hyzoo and Waaoo seems to lie with im- poverished synthesis rather than with prema- iurc d e s t ~ ~ ~ c t i o n of abundantly produced hcmoglobin. Estimates of the specitic activity of purified a and B chains from thc gorilla Tomoka, aftcr incubation in vitro of bone marrow aspirates (obtained by Dr. S. Char- ache) with 1 mc of L2Hlleucine at 3S?C for 100 minutes. indicatc that both chains of hc~noglobin Wazoo-like those of hemoglobin A from the same animal-are synthesized by marrow approximately in proportion to their final concentration in peripheral blood. 6. J. B. Clegg, M. A. Naughton, D. J. Wcather- 311, J. Mol. Biol. 19, 91 (1966). 7. S. H. Boycr, A. N. Noyes, G. R. Vrablik, 1,. J. Donaldson, E. W. Schacfer, Jr., in prcparation. 8. M. 0. t>ayhoff, Atlas of P~o te in Seqrrence arid Sti.uctz(i~ I969 (National Bioineclical Re- search Fonndation, Silver Spring, Md., 1969). 9. S. H. Boyer, P. Hathaway, F. Pascasio, C . Orton, J. Bordley, M. A. Naughton, Scie~zce 153, 1539 (1966). 10. H. A. Iloffman, A. J. Gottlcib, W. G. Wise- cup, ibid. 156, 944 (1967); P. T. Wade and N. A. Barnicot, liiochem. Genet., in prcss. 11. R. A. Fisher, The Geizetical Theor], of Na- triral Selection (Oxford University Press, Ox- ford, 1930), chap. 4 . 12. TI. Lchman ant1 R. W. Carrell, Rrft. Mcd. J . 4, 748 (1968); R. Brimhall, S. I-Iollan, R. T. Jones, R. D. Koler, J. G. S~elenyi, Clin. Rer. 18, 184 (1970). 13. R. K. Abratnson, D. L. Rucknagcl, D. C. Shremcr, Sclerrcc 169, 194 (1970). 14. S. II. Boycr, E. F. Crosby, A. N. Noyes, G. F. Fuller, S. E. Leslie, L. J. Donald\on, Ci . R. Vrablik. E. W. Shaefcr. Jr.. Broclzent. , Genet., in press. 15. T. H. J. I-Iuisman, W. A. Schroeder, G . Stamatovannoooulos. N. Bouver. J . R. Shel- ton, G. ~ p e l l , ' ~ . ~ l h . Iizvest. 49; 1635 (1970). 16. S. H. Boyer, E. F. Crosby, T. F. Thurmon, A. N. Noyes, G. F. Fuller, S. E. Leslie, M. K. Shepard, C. N. Hcrndon, Science 166, 1428 (1969). 17. An Asp+ Lys changc is a two-nuclcotidc change. The presumption that the hypotheti- cal "a" bcars a""""P is a convenient fiction; it might just as well, for example, be a"""!" in one or both genera. I t must, however, be cither Asp or Glu if the hypothesis of a buried ::a" product is to remain tenable. 18. P. S. Gcrald and M. L. Eiron, Proc. Nut. -4cad. Sci. U.S. 47, 1758 (1961). 19. The relative order of cathodic mobilities at pH 7.1 was expected, on the basis of amino acid composition (7), to be A,> Hyzoo -- Wazoo > A. The observed order, A, > A > Ilyzoo -= Wazoo, is attributed to conforma- tional clepcnclcnt d~ffcrences between A and thc variant hemoglobins in ionization of histi- dine residues. 20. W. E. Nance, Science 141, 123 (1963); W. E. Nance and O. Smithies, Nafiwe 198, 869 (1963). 21. D. J. Wealherall, J. B. Clcgg, Wong Hock Boon, Rrit. J. Haemafol. 18, 357 (1970). 22. C. Raglioni, Proc. Nat. Acad. Sci. U.S. 48, 1880 (1962). 23. Supported in part by PI-IS grant HD-02508-04 and I'HS career development award K3-C;M- 6308-03 to S.H.B. 4 September 1970; rcvised 14 October 1970 Functional Sequences Modulated by Morphological Transitions in Human Lymphoid Cells Grown in vitro Abstract. Immunoglobulirz-prod~~cirzg cells undergo n series of morp/zologicnl iransitions; each corzfigulntion displ~~ylys pecific l~mctional attributes. TIze life c y d e o f imrnunocytes may he ~.~i,runlizcd as a series of functional compart~neafs expre.\sed by morplzological scqucnces. A long-tern~ c ~ r l t ~ ~ r c of human lym- dominant. Binuclcate cells and transi- phoid cells derived from a patient with tional forms are a freq~rent finding (Fig. lymphoma was established in 1966 (1). 1). Indirect imn~unofluoresce~it s udies These cells now havc been maintained havc demonstrated that these cclls syn- in monolayer cultures for 5 years. Retic- thesize gamma globulin. Cells growing uloid fusifor~n cclls and lymphocytoid on Leighton cover slips were rinsed in and plasinacytoid round cclls arc prc- saline and fixed in acctonc for 10 min- utes. The cclls were incubated with un- Fig 1. TI cells in culture exhibiting the en- tire gamut of nlorphological configura- tions (Wright's stain; x 480). labeled goat antiserum to human gam- ma globulin for 30 millutes at room temperature. The cells were then washed twice with a buffer solution (pH 7.2) and incubated for 30 rninubs with fluorescein tagged rabbit antise- rum to goat gamma globulin. The cells, without counterstaining, were examined under an ultraviolet n~icroscope. Posi- tive apple-green fluore\ccnee was easily distinguishable from autofluorescencc. Time-lapse photographic studies wcre performed with a cdture system (2). Bccnusc of the prolonged doubling time of these cells (52 rirr 2 hours), one pic- ture was taken every 30 minutes for 165 hours. Cell pedigrees were gener- ated from enlargements of the nega- tives for morphologic analysis. Gener- ation times were measured from one cell division to the next daughter-cell division. The median generation time was 36 hours. Rare cells, still metabolically active as evidenced by mobility and changes in shape, failed to divide during the ex- periment. Most of the cells undergo changes in shape, from round to fusi- form and often hack to round. Each of these changes lasts for several hours which allows for morphologic defini- tion. When spindle-shaped cells he- come round prior to mitosis, they do so very rapidly within a single hall- hour interval. Upon division, a fusi- form ccll can givc ri\e to onc elongated and one round daughter cell (Fig. 2) or more commonly, to two fusiform cells. Round cells may also givc rise to a mor- phologically mixed population; some- times thcy produce only smaller round cells. Apparently these smaller round cells are terminal because, aftcr a brief period of rapid inoven~cnt, thcy beconle immobile and never divide. Occasion- ally a ccll that remained round lor rzu~~~crous hours will adopt a fu\iform shape for a few hours and then start to divide vigorously. An unusual finding is that two daughter cells may come into close contact and fuse, and a single binu- cleated cell will en~crgc (Fig. 3). After several hours this cell may either ells- sociatc into a rapidly mobile round cell and a static spindle-shaped form, or it will divide giving four round daughter cclls. This fusion of cells is different lrorn the mechanisms of emperipolesis (3), pcripolesis (4), and uropodapsis (5) be- cause the fusion is long-lived, distin- guishable cell boundaries disappear, and the process may sometimes con- Tablc 1. All of the cell< in t hc examined fields wcre arh~trarily a\signed to a morplrological category aceortlnig t o the prevalent f e a t u ~ e and cla\sifeil a< Iloore\cent o r nonfluorescent. -- -- - - - - Mot phological distribution - - - - - --- -- - -- Cell Fluores- Nonfluo- 'IUo- types cent rc\cent re'cent cell\ cells ( 76 ) Jntcrmcdiate forms 114 93 65.5 Fusiform 32 265 10.7 Total 329 403 44.6 15 JANUARY 1971