Summary Thy-1 is a small glycoprotein of 110 amino acids which, folded in the characteristic structure of an immunoglobulin variable domaid’), are anchored to the lasma membrane via a glycophosphatjdylinosilol (GPI) tailL7’) (Fig. 1). It is a major component of the surface of various cell types, includ- ing neurons, at certain stages of their development(4) . These qualities doubtlessly appeal to certain cognoscenti, but it is not clear why they would raise Thy-1 to the status of a favourite molecule. Indeed, few scientists readily admit to having a favourite. We study individual molecules because science is rooted in specific observations; but we do so in order to discover mechanisms of general importancc. A mol- ecule’s appeal is dependent on its ability to reveal novel aspects of how nature works.

Thy-1 has been unusual in this respect. It was the first lym- phocyte surface antigen shown to be restricted to a functional subset of lymphocytes (T cells in the mouse), a finding cm- cia1 to the development of cellular immunology(5); it was one of the first cell surface molecules to be sequenced and indi- cated the importance of immunoglobulin domains and GPI anchors as structural it has been pivotal in studies dcmonstrating that GPI-anchored molccules are able to sig- nal across the membrane they do not pan(^.^). Thy-1 has revealed this much, however, with the charm of an adroit stripper: it has always promised glimpses of things more exciting than that displayed. In pai-ticular, the function of this molecule has never emerged.

Our current work on Thy-1 in the nervous system is, we believe, finally prizing this last secret into the open. It suggests that Thy-I on the neuronal surfacc interacts with one of the main glial types of adult brain, the astrocyte, lo restrict axonal growth(8). There is no current consensus that astrocytes do regulate such growth, and this is the wider question we are addrcssing from the perspective of Thy- 1. At issue also is how to progress rrom knowing the nature of a molecule to defining its function. This task has already been achieved with molecules which were harder to chardcterise than Thy-1. The problem is a classic Catch 22 situation: Thy- 1 has been readily characterised because it is small and abun- dant (it is ten times more abundant than any other surface glycoprotein on rat thym~cytes(~)); but since there is so much of it, whatever it does i t must do rather badly (or there would be no need to have so much of it!). Defining a function based on low activity or affinity is not trivial.

Here I describe two aspects of work on Thy-1 . One is long completed - the research which led to Thy-1 being isolated and characterised - and included because it demonstrates features which are still applicable to much research using antibodies today. The other is our ongoing work on the ner- vous system, which i \ at that exciting early stage where hypotheses are formed, but the answers are still coming in.

Antisera Too Simple by Half Lymphocytes participate in numerous cellular interactions which must be controlled by their surface molecules. This much was already obvious when I started my P1i.D. work, without it being apparent what these surface molecules would be, or how they would act. What little we knew about the lymphocyte surface came from the analysis of antiscra raised by irnmunising mice of one inbred strain with lympho- cytes of another. These antisera could only detect molecules which showed an immunogenic allelic variation between the two strains, so stringent a restriction that they detected (if anything) a single molecule despite the immunisations being made with whole cells(10). This approach incvitably misscd many molecules, and the antisera produced were of too low an affinity to be used to isolate the antigen concerned. They did demonstrate that lymphocyte surface molecules changed during di ffercntiation. and helped define the different func- tional classes of cells. In molecular terms, however, they amounted to little more than a collection of intriguing anec- dotes about the lymphocyte surface.

What was needed was an approach which would lead to the isolation and characterisation of the full range of surface molecules. I arrived in Oxford in 1972 to work with Alan Williams just as he was formulating these objectives. Today, the lymphocyte is probably the cell whosc surface interac- tions are best understood in molecular terms. This has been due in substantial measure to the application of the approach started in Oxford, and in which Alan’s lab provided many of the important insights. That it got started at all depended on chance, a property of Thy-1 which enabled us to obtain a higher affinity antibody than any other currently available.

Antibodies are Not Created Equal The existing mouse antisera easily recognised their antigen on the cell surface, yet failed to recognise it when solubilised. This was put down to the epitopes concerned being ‘confor- mationally determined’, i.e. the required conformation was destroyed by the detergents used for solubilisation. What was actually happening was an effect first demonstrated in a dif- ferent context: when an antibody binds to multiple antigens on a surface (eg a cell) the overall affinity of binding increascs (by as much as 1 03) with each increase in valency of binding(”). Thus a bivalent IgG antibody can increase its affinity by lo3, an IgM (which for steric reasons can only get 3 of its binding sites to surface antigen) by lo6. Upon solubil- isatinn, however; each antigen binds independently to the antibody; there is no increase in affinity wilh increased valency of binding and the antibody does not recognise its antigen(I2) (sec Fig. 2; strictly speaking, it still binds its anti-

Fig. 1, Our current model of Thy-1, based on the immunoglobulin variable fold and the structure determined for the glycophos- phatidylinositol (GPI) anchor which secures the molecule to the lipid bilayer. The arrows represent the strands of the beta-pleated sheet, which are joined by loop regions. The asparagine residues to which carbohydrate chains are linked (in the mouse and rat) are denoted by the dark bands in the loops after strands R and E, and the white band in strand G. The large sugar chains cover much of the front, left, and rear of the molecule. Residue 89 (indicated with diagonal lines at the top of strand F) is either Arg or Gin in mice, conferring the Thy-1.U1.2 allelic antigenic determinant to the mol- ecule. This model is currently being refined, but is already proving invaluable to identify clusters of residues which are conserved across vertebrate species. Residues in the beta-pleated shect, and occasional residues in the loops, are conserved for structural rea- sons. Most residues in the loops have few structural constraints, and indeed the loops projecting onto the left face of the molccule show considerable variability between the vertebrate species. The loops projecting onto the right, and the upper, face of the molecule have a high proportion of conserved residues, and it is these we are currently focusing on in order to define the astrocyte binding region of the molecule.

gen, but you need times higher concentration 01' anti- body).

These antisera were of low affinity because the task being asked of the immune system - lo detect an allelic variation (normally a single amino acid substitution) in a protein which was a very minor component of a complex immunis- ing dose - was at its very limits of possibility. We wanted to increase the range of surface molecules which could be detected, and therefore immunised across species (rat tissue injected into rabbits). This also increased the affinity of the antibodies produced, due to two effects: the individual clones stimulated by complex immunisation were of higher affinity; and when multiple clones detect different epitopes on a pro- tein, the overall affinity - even for a soluble antigen - is also greatly augmented('2) (Fig. 2). Thus the interspecies anti- bodies were of high enough affinity to identify their solu- bilised antigen during purification.

This property alone would have allowed us to purify the more abundant surface molecules, although the task would not have been easy. Purification of the rarer molecules has

ible because of highly efficient immunoaffinity n (ie the antibodies, imniobilised on a column, pull

out their antigen). The interspecies antibodies were of too high an affinity to be used in this way (their antigen could not be eluted), almost certainly because of the effect of binding to multiple epitopes rather than an exceptional affinity of individual clones (Fig. 2). Monoclonal antibodies, on the other hand, have been highly successful for purifying their antigen because they can be raised (in a heterogeneous immunisation) with a sufficiently high affinity to bind solu- bilised antigen, yet each antigen molecule is usually bound only once, enabling it to be quantitatively recovered.

How Thy-1 Came to Be Purified So what has this lecture on affinity got to do with Thy-l?

Well, in 1972 we had not had this lecture - wisdom came slowly and painfully. However. in addition to looking for new antigens with interspecies antibodies, we were drawn into looking at Thy- 1 as a 'model system', stimulated by the enthusiasm of Ron Acton. newly arrived from Caltech where it had just been shown that one of the allelic forms of Thy-1 defined by alloantisera in the mouse (the Thy-1.1 allele) also occured in the Ron announced that his 'Saturday pro- ject' was to further characterise this rat Thy-1, an aim which progressively spread to the rest of the week and the lab. For one of the first steps done was to produce an anti-Thy-1.1 antibody by irnmunising Thy- 1.2 mice with thymocytes, not from a Thy-1.1 mouse as previously done, but from a rat (which was Thy-1 . I in type). This cross-species immunisa- tion detected rat antigens other than Thy-1, but these were ignored simply by analysing the antiserum on Thy-1 . 1-type mouse thymocytes. Thus a high affinity antibody was pro- duced to a surface all~antigen( '~). This recognised soluble Thy- 1, enabling us to demonstratc that the novel brain -thy- mus antigens detected with rabbit antibodies were simply additional epi topes on the Thy- 1 molecule(15). Alan and col- leagues went on to purify Thy-1 by lectin affinity and gel fil- tration chromatography. On sequencing, it was recognised as being homologous to an immunoglobulin domain(') (Fig. I ) . This property had already been found for another molecule, beta-2 microglobulin, a component of serum which also occured on the cell surface as the smaller chain of the class I major histocompatibility antigen(17). Finding a second Ig- domain structure on a protein confined exclusively to thc cell surface raised the possibility that such domains, known to provide the binding unit for the best characterised 'recogni- tion' molecule, the antibody. might also be used to confer 'recognition' (i.e. selective binding) properties on the cell surface, and lhat indeed the Ig domains might have arisen from a more primitive recognition system used to guide cell movements in development(').

Thy-1 in Brain: It's Not where It Is. . . On leaving Alan's lab in 1975,l wished to change my field of interest and study the cellular interactions controlling neural development. In working with lymphocytes. we had assumed they were homogeneous balls, a simplification which did not impede progress. However, it was obvious that to assume the same of neurons would be to ignore one of the essential prin- ciples underlying neural function. Where an intercellular contact is formed in brain is as crucial a detcrminant of its function as the nature of the molecules, or even the cells, making the contact. To progress in understanding the molec- ular basis for neural development, it would be nccessary to localise molecules, including those of the cell surface, accu- rately within nervous tissue. Conversely, knowing the distri- bution of a surfacc molecule within brain might be expected to shcd light upon its function.

Immunohistochemistry had been successfully applied to the localisation of intracellular molecules in brain. However, surface moleculcs posed a set of new and difficult problems for immunohistochemical l~calisation('~), and an influential body of opinion stated that it would be impossible to obtain

A. Binding to immobilised versus B. Po tyclonal versus monoclonal binding of antigen in

im rn unoaff inity chromatography soluble antigen

Poiyclonal . 4 epitopes bound to column

Monoclonal : 1 epitope bound to column

Fig. 2. Two ways in which multivalent binding affects the outconic of immunological dctection or purification. A. Thc multivalency is that of the antibody, ivtiich in the case shown is divalcnt IgG binding to ccll surfacc (lower) and solubilised (upper) antigen. To dissociate from the cell surface, the antibody nccds to dissociatc sirnultnneousZ>~ from both antigens, an event of great rarity even if the intrinsic affinity for the antigen is low. In solution, howevcr, binding of antigen to one site on the antibody is independent of binding to the others. and there is no increase in affinity with valcncy of binding. B. Polyvalency from the point of view of an antigen exhibiting a numbcr of diffcrent epitopes. When these are recognised by a polyclonal antiserum, the antigcn can only be released (cg elutcd from an immunodfinity column) when all 4 epitopes dissociate simultaneously; the monoclonal antibody recognises only one cpitope, and so dissociation occurs more readily.

meaningful localisation of such molecules. I therefore decided to venture inlo the new field with an old rricnd ~

Thy-1. I did. after all, have good antibodies to it, and it was abundant in brain. The first was a decisive advantage - anti- bodies of a similar affinity and specificity to other neural sur- face molecules only became available in the mid-80's. How- ever. the abundance of Thy- l proved a mixed blessing, for the moleculc was present so ubiquitously in the brain that it was very difficult to identify the elements which expressed the molecule, and those which lacked it.

However. one of the joys the nervous system offers the investigator is that certain regions exhibit unusual character- istics which assist analysis. Thus there are neurons that lack axons, olhers thal lack dendrites. areas where the neurons and their main iniputs are clustered in distinctive layers, others whcre layers are formed by different types of glia. By exploiting these properties, and combining immunohisto- chemistry and radioimmunoassays with grafting and other perturbations of the cellular makeup, we were able to show that in vivo Thy-1 is present on most if not all ncuronal mern- branes, but not detectable on any glial cells of the central or peripheral nervous ~ y s t e m ( ~ 7 ' ~ ) .

their axons which are Thy- l-negative(IX). These are the pii- mary olfactory neurons, located in the olfactory epithelia of the nose; their dendrites are specialised to detect odours, their axons carry the signals so generated into the olfactory bulb at the hont of the brain. They are unique in the mammalian ner- vous system in being able to grow into brain throughout a dult life(19). I was doing this work with Peter Barber, who had spent several years investigating the growth of these axons into adult brain. He suggested this property may be related to their lack of Thy-1. Other observations, on the redistribution of Thy-1 from lesioned axons to glia("), prompted us to form a more specific hypothesis: neuronal Thy-l interacts with some ligand on astrocytes to restrict axonal growth in brain. Nearly 10 years later. we are still try- ing to prove this, although we now have two lines of cvi- dence which I belicvc point firmly in this direction. The first comes from a more detailed understanding of the expression of Thy-1 in developing brain. the second from functional studies on neurite outgrowth by transfected cell lines.

Appearance of Thy-1 on the Neuronal Surface is Related to the Cessation of Axonal Growth In situ hybridisation studies have revealed a very simple pat- tern of expression of Thy-1 mRNA in rats and mice: when neurons have migrated to their final position and started to grow a dcndrite, they express Thy-1 I T I R N A ( ~ ~ ) - ~ ~ ) . Thy-1

. . . But where It Isn't, that Provides the Clue It was difficult to make anything ofthis general neuronal dis- tribution of Thy-1 , until we found one set of neurons and

protein then quickly appears on most neurons; however. somc express comparable levels of message for many days before expressing detectable protein (Fig. 3). Clearly, the appearance of Thy-1 protein is linked to some event addi- tional to expression of its mRNA. That event appears to be the cessation of axonal growth in CNS(*’).

The axon is a long, thin process which connects a neuron to its distant targets (normally other neurons, 01- muscle). The growth of some axons is simple, to a single, preformed target. In many cases in the CNS, however, an axon projeck to mul- tiple targets, each of which develops at different times which may be unrelatcd to when the axon first grows into (or pasl) the target region (see ref. 23). An axon may have finished growing, and formed synapses upon its earlier developing targets, and then (sometimes wceks latcr, a vcry long timc in rodent development) produce a branch to contact a late- developing target. Thus maturation proceeds at a different pace in different regions of the same axon, a property which must be reflected at the molecular level.

The appearance of Thy-1 protein is spatially regulated and linked to development. After expression of its mRNA, Thy- 1 protein can appear on the dendrites of a cell (and even then it can be compartmentalised to parts of the dendrite(?”)) but is excluded from their axons until these have finished growing; where axonal growth occurs in multiple phases, it seems Thy-1 is allowed inlo the regions wherc growth is complete, but remains excluded from growing regions(?’).

But That’s Not Possible This exclusion of Thy-1 from the growing ends of axons poses at least two problems for current models of neuronal membrane biosynthesis. Firstly, it is thought all axonal gly- coproteins are made at the cell body, and transported down the axoplasm to be inserted into the plasma mcrnbrane at the growth cone(”). Newly synlhesiscd surface glycoprotein should appear, within hours of synthesis, at the growing tip, thc very region from which Thy-1 is excluded. An allerna- tive, or additional, mechanism must regulate Thy-I ’s progress down the axon. Secondly. it has been suggested that the GPI anchor on Thy-1 specifies i t to be directed to the axon and not the dendrite, and evidence has been produced from dissociated tissue culture to support this model(2s)). However, there is no doubt in vivo lhal growing dendrites can have extremely high levels of Thy- 1 (and other GPI-linked surface glycoproteins(26)) on them. Thus the growing dendrites of the cerebellar Purkinje cells are intensely Thy- 1 positive (Fig. 3N). During this stage of early postnatal development these cells express high levels of Thy-1 mRNA, predomi- nantly in the lower regions of their dendrites; intense immunolabelling of Thy-1 protcin is similarly found in the cytoplasm of these dendrites, and on their ~urface(~9.’~). No other cell in the cerebellar cortex expresses detectable Thy- 1 mRNA at this stage(22); nor do the developing axons contact- ing the Purkmje dendrites express Thy-1 protein until later, when they have finished growing(4)). Thus the Thy- 1 seen on these dendrites has to be made by the Purkinje cells. A simi- lar case can be made elsewhere in the brain, and the presence

or lhy-1 on growing dendrites is common (although not invariant) in CNS.

The mechanisms by which Thy-1 is incorporated into the different compartments of the neuronal surface in synchrony with different developmental phases of the neuron is as yet unknown. Identifying how the spatial specialisation of the neuronal surface is established is onc of the key challenges facing molecular neurobiology, and here the study of Thy-1 may once again make a seminal contribution.

So What Does Thy-1 Do. . . in Pseudo-neurons? What would happen if Thy-I was present on growing axons? Although this does not occur in normal mammalian develop- mcnt, any adult neuron attempting to regenerate an axon is Thy-1-positive. Such axons do not regenerate within CNS in vivo, but regrow along peripheral nerves if offered the oppor- t ~ i n i t y ( ~ ~ ) . The glial cells of central, but not peripheral, ner- vous system appear to inhibit growth of the axoiis they con- tact. The inhibitory combination is adult neurons contacting adult glia, as transplants have shown that axons can grow quite long distances when one or other element is from embryonic brain(’”.

To approach this problem we have sought to drivc expression of physiological levels of Thy-1 to growing axons. Attempts to cause Thy-1 to be expressed on real axons as they grow in transgenic mice have not been successful, presumably because of regulatory mechanisms indicated in the previous section. However, we have expressed Thy-1 by transfection on a cell line, NGI 15/401L. which is normally Thy- 1 -negative and which extends nenrites when it differen- tiates in culturc. Thcse ncurites are by molecular criteria axonal, rather than dendritic, in character. Three clones expressing Thy-1, and three Thy-1-negative control clones, were assessed for their ability to grow neurites over a number of cellular substrata. When grown on top of the axonal growth-promoting glial cclls of embryonic brain or periph- eral nerve, or on fibroblasts or epithelial cells. or on the plas- tic beside any of these cell types, both Thy-1-positive and ncgativc cells sent out neurites with equal vigour. On astro- cytes, however, the Thy-1 expressing cells attached, but did not send out a neurite, whereas the negative cells did both; on the plastic beside the astrocytes, both cell types grew neu- riles. The inhibition of Thy-1-positive neurite outgrowth could be completely over-ridden by blocking the astrocytic ligand with soluble: pure Thy-I, or by blockin Thy-1 itself

Thus Thy-I on the neural cell line interacts with some molecule on astrocytcs to inhibit neurite outgrowth. It appears the inhibitory signal is transmitted directly across the surface membrane of the neural cell, in a process which in some way involves Thy- 1 cven though it does not span the lipid bilayer. Similar signalling by Thy-1 and other GPT anchored glycoproteins is seen in the lymphoid ~ y s t e m ( ~ . ~ ) , and presumably involves interaction between Thy- 1 and some transmembrane signalling molecule, possibly a tyro- sine kinase(2y) or phosphatase(30).

on the neural cells with a polyclonal antibody (8q

Fig. 3. Expression of Thy-l i n nervous tissue, illustrating the disso- ciation of expression of Thy-1 mRNA and protein (A-L), and the presence of Thy-1 on growing dendrites (M,N). A-D: in situ hybridisation to detect Thy-1 mRNA (the signal is thc black silver grains, cell nuclei have been stained with a blue dyc) in pontine neu- rons at P5 (A) and P12 (B), and in vestibular ganglion at E l 3 (C) and El5 (D). Pontine neurons have Thy-l protcin dctectable inimunohistochemically on their dendrites (arrows) at P5 (E) and P12 (F) (immunopcroxidase labelling, cell nuclei again are visu- alised with a blue stain). Vestibular neurons are barely positive at P8(G ; the two ncuronal cell bodies indicated with asterisks; the three other cells with (blue) nuclei are Thy-1-negative glial cells). nearly 2 weeks aftcr cxprcssion of high levels of mRNA (C.D; G‘ is upper cell of G. present also on an adjacent section, inimunostained for parvalbumin. a marker which identifies the vestibular neurons). H: vestibular neurons at P1.5, now strongly Thy-1 positive. J,K: the growing synaptic terminals of pontine ncuroiis at P14, seen with immunohistochemistrq. for GAP43 (J, curved arrows) but unla- belled with anti-Thy-1 antibodics (K; curved arrows). 1,: Thy-1 itnmunolabelling strongly detects the same synaptic terminals in the adult. M: one of thc first cells in mouse hippocampus CAI area to become Thy-l positivc (at P12) shows intense labelling along its ascending dendritc. Axons coming into the hippocampus (from the enterorhinal cortex) are Thy-1 positive at this age, and (although unlikcly) it could be argued that this dendrite i s not itself positive, but appcars so because of axons surrounding it. N: rat cerebellum at P12, the Purkinje dendrites are intensely Thy-I positive; all axons tcrminaling on them are Thy-1 negative at this age. Scale bars are 10 pin except for N (50 Fm).

. . . and in Real Neurons? Does Thy-1 in vivo interact with astrocytes to restrict axonal growlh? Our work to date suggests such a role, but certainly does not prove it. Neurons interact with astrocytcs, and Thy- 1 will almost certainly be involved in this, but the secondary messenger systems activated may control processes other than, or additional to, axonal growth. Clearly, delineating this may elucidate the role of astrocytes in regulating axonal growth, and provide valuable clues as to the rolc of Thy-1 in other cells (such as lymphocytes) during development. It may also shed light on the function of a wider group of neu- ronal surface molecules. On lymphocytes~ all GPI-anchored surface molecules activate the same signalling pathway, pro- ducing identical metabolic effects(6). If this is so with the neuronal members of this class of glycoprotein, then defining the function of Thy- 1 will identify the intracellular conse- quenccs of activating any of the others with their ligands.

Having outlined our view of the role of Thy-1 in nervous tissue, I should add that others have also suggested roles for the molecule, including roposing that Thy-1 either pro- m o t e ~ ( ~ ’ ) , or axoiial growth. What gives me confidence in our functional analysis using transfected cells is lhat it arises from, and is compatible with, an hypothesis formed by careful study of the in vivo expression pattern of Thy-1 in the same species. We still have much to do, and I hope a few surprises await us. What we do have, in working with Thy-1, are clear, specific reagents and assays, and a good molecular basis from which lo pursue cellular ques- tions. Which, in a nutshell. is the joy of working with this molecule.

Postscript While writing this article, I was conscious of the fact that Alan Williams was dying of cancer. Alan would not have wished me to deviate from a strict scientific assessment of the factors influencing the early work characterising Thy-1, but he has now died, and a more personal assessment of his con- tribution is in order. Prior to Alan’s entering the field, work on surface molecules of lymphocytes (or any other cell less conducive to study than the red blood cell) had been pursued with more vigour and acclaim than rigour. Thy-1, for instance, had been identified both as a glycoprotein of Mr 60,000 and as a glycolipid of Mr 35,000. The other lympho- cyte surface antigens had fared no better. Alan considered he brought little to the field except common sense. Certainly he cut straight lo the essence of a problem, and quickly came up with the practical measures to solve it. His contribution to understanding the structure of thc immunoglobulin super- family molecules has been enormous. Alan sal down at home in the evening and played with sequences, as he dcscribcd it a past-time similar to cross-word puzzles but more interesting. He was morc accurate, and more informative. than powerful computer prograinmes.

As a man, common sense curiously is not the quality which epitomised Alan. He had vast enthusiasm and vitality, was interested in everything, always anxious to know. His endless exuberance would have made many enemies had it not been for the innocence of his intent. The only thing he loved more than an argumcnt was its being settled by deci- sive experimentation. His death at 46 hah deprived all of the insights of a very good scientist. It has deprived many of a good friend.

References 1 WWams, A.F. and Rarclay, A X . (1988). The immunoglobulin superfamily domainb for cell \urlace recognilion. Anna REV. Immunol. 6.381-405. 2 Ilomans. S.W., Ferguson, M.A.J., Dwek. R.A., Rademacher, T.W., Anand, R. and Williams, 4.P. (1 988). Complete structure or [tie glycosyl phospbatidylinositol membrane anchor of rat brain Thy- I glycoprotein. Nnriarr 333,269-272. 3 Ferguson, M.A.J. and Williams, A S . (1 988). Cell surface anchoring of proteins via glgcosyl-phosphatid~linositol structures. Annrr. Rei’. Biorhenr. 57,285-320. 4 Morris, R.J. (I 985). ‘Thy- I iii developing nervous tissue. Dev. iVeurmci. 7. 133-160. 5 Kaff, M.C. (1971). Surface antigenic markers for distinguishing T and B lymphocytcs in mice. Trmy~lnrtt. Rei,. 6. 52-80. 6 Robinson, P.J. (1991 ). Phosphatidylinositol membrane anchors and T cell aclivarion. Immiinol. 7 iduy 12, 35-41, 7 Uarboni, E., Gormley, A.M., Pliego Rivero, F.B., Vidal, M. and hlnrris, R.J. (1991 ) activation of T lymphocytes by cross-linking o f glycophospholipid-;nchured Thy- I inobiliscs separate p o i s of intracellular second messengers to those induced by the antigen-receptorICD3 complex. Immimology 72,457463, 8 Tiveron, M.C., Barboni, E., Pliego Rivero, B., Gormley, A.M., Seeley, P.J., Crosvcld. F. and Morris, K. (1992~. Selective inhibition of nenrite outgrowth on mature astrocytcs by Thy- I glycoprotcin. Nafirre 355.745-748. 9 Williams, A.F. (1982). The predominant surface glycoprotcins of ithyniocytcs and lymphocytes. Kiosci. Rep. 2,277-287. 10 Uoyse, E.A. and Old, LJ. (1969). Some aspects of normal and nbnormal cell surface genetics. A m u . Rev. Generics 3.269-290. I1 Karush. F. (1978). The dfini ty of antibody: range, variability and the role of multivalence. 111 Cwn/xeherisiiv 1n;~nirno~og.v. V d . 5: ~i~irr i rr i io~tohu~ins . (cds G.W. Litman and K.A. Good). pp. 85-1 1 h. N e w Yoi-k: I’lenuni Medical Book Company. 12 Morris, R.J. ( I 992). Antigen-anlibod) irrterachn: how aTTinlty and kineticr affect

design and \election procrdurea. In ! M < J M I c / o ~ ~ Antibodies (rd. M .A. Ritter). Cambridge University Press. Cambridge, in press. 13 Douglas, T.C. (1972). Occnn-cncc of a theta-like antigen in rats. J. Exp. Med. 136, 1054-1062. 14 Acton, R.T., Morris, R.J. and Williams, A.F. [ 1974) Estimation of the amount and tissue distribution of rat Thy- I. I antigcn. Eun .I. hnniunol. 4.59X-602. 15 Morris, R.J., 1,etarte-Muirheatl, M. and Williams, A.V. (1975). Analysis in

deoxycholate of three antigenic spccificitie? associated with the rat Thy1 molecule. Eun .I. Itufniinol. 5, 2x2-285. 16 Peterson, PA. , Cunningham, R.A., Herggard, 1. and Edelman, C.M. (1972). Beta-2 niicmglobolin - a rree Ig domain. Pmc. .Vat. Acnd. Sci. US;Z 69. 1697-1701. 17 Morris, R. and Grosveld, F. (1989j. Expression of Thy-l iii tlic nervous system of the rat and iiiouse. In Gel/ S u r ~ i w Arz/ipm Thy-1. l n t u t i ~ n o l o ~ ~ , iVmrolog? and Thmpeulir Applrcu/ions. (edr. A.E. Rril and h4. Schle\inpcr). pp 121-148 Marcel T)ekherInc. New York 18 Morris, RJ. and Udrher, P.C. (1983). Fixation of Thy-1 in nervous tissue for

quantitative assessment of thc cffcct of diflerenl fixation condition.; upon ietentlon ofantigeiiiciry and lhe crov-linking olThy-I. ./. Hi!Todzriii Cwdzcm. 31. 263 274. 19 Graziadci, P.P.C. and Monti Graziadci, G.A. ( 1978) The olfactory system: a model for the study of neurogenesis and axon regeneration in mammals. In Neurorial Plo.rlici/! (ed C.W. Cotman) p p 131 - 1 51, Ravcn Press, N.Y. 20 Xue, G.P., Calvert. R. and Morris, R.J. (1990). Exprt.s\ion of the neuronal surface glycoprotein Thy- I is tinder post-tranFcriptional control, and is spatially regulated, in the developing olfactory system. Dnelopment 109, 851-864. 21 Xue, G.P.. Pliego Rivero, B. and Morris, R.J. (1991). The surface glycoproteiri Thy-l 1s excluded frorri growing . ixom during development: a d d y LIT the expression ofThy-1 during axogenesis in hippocampus and hindbrain. Dcvelopmentll2, 161-176. 22 Xue, G.P. and Morris. R.J. (1992). Expression of the ncuronal surface glycoprotein Thy-l docs nor follow appearance of its iiiRNA in dweloping inousc Purkinje cell?. J. iVeurocheni. 58.4.30-410. 23 Morris, R.J., Tiveron, M.C. and Xue, G.P. (1992). The relatioii o r the expression and function of the neuronal ylycopmtein Thy- 1 tu sxonsl growth Biocliem Soc. Transoctioris in press. 24 Crafstein. B. and Forman, D.S. (1980). Inrracellular transport in neurons. Phyiol . Reb.. 60, 1167-1283.

I Tel.: 202-319-5274; Fax: 202-319-4467

25 Uotti. C.C., Parlon, R.G. and Simons, K. ( 1991 1. Polarized sorting of glypiated proteius in hippocampal neurons;. Nature 349, 155-161. 26 Faivre-Sarrailh, C., Ccnnarini, G.. Goridis, C. and Rougon, G. (,I99?). FYFI 1 cell surfacc iiiolecule cxprcssioii in the developing inousc ccrcbellum is polansed at synaptic rites and within granule cells. J. Ncrrrorci. 12.2.57-267. 27 Benfey, 11. and Aguayu, A.J. (1982). Extensive elonyation ofawons from rat brain into peripheral nerve grafts. Nrmm 296. 150- I52 28 Zhou, C.-F., Li, Y,Rlorris, R.J. and Raisman, G. i199Oj. Accurate rccoiistruction of three complementary laminar afferents 10 the adult hippocampus by eiiihr)mic

oapp. W. and Slnckinger, I I . (1991). d to protein tyrosine k inaw. Srience

254. 1016-1019. 30 Volarevic, S., Burns C.M., Snssman, J.J. and Ashwell. J.D. (19901. Intiinatc abaoc1;ition ol' Thy-l and the 'I-cell tinlipen receptor w i t h the (:1)45 tyrosiiie phosphatase. Proc. /\'ut/Acad. Sci. US,\ 87.7085-7089. 31 Leiffer, D., Lipton, S.A., Barnstahle, C.J.. Masland, R.H. (19841. h,l(inoclonal

generation of pl-ocesses by rat retinal ganglion cell> in

32 Mahanthappa, N.K. and Patlerson, P.H. ilY97). Thy-l i~iv( i l ie i~ien1 in neurile outgrowth: perturbation by antibodies, phospholipase C and mutation. Drv. U i ~ l . 150, 37-59. 33 Mahanthappa. N.K. and Patterson. P.H. (1992). Thy- I inultimcl-iiarion is correlatedmtth neuritc outgrowth Dm.. Kiol. 150; 60-71.

Roger Morris I S at the Laboratory ot Neurohiology, National Institute Ior Medical Research. Mill Hi11 London N W 7 I AA

Announcement

Courses in Molecular Biology, 1993 Venue: Center for Advanced Training in Cell and Molecular Biology

(CATCMB), Washington, DC

Recombinant DNA Methodology

PCR Techniques

Basic Cell and Tissue Culture

Recombinant DNA Methodology

Receptor Binding Techniques

January 4-8, 1993

January 9-1 1, 1993

January 11-15, 1993

March 8-12, 1993

March 15-19, 1993

For further information, contact:

Office Manager CATCMB1103 McCort- Wart1 The Catholic University of America Washington, DC 20064 United States of America