Commitment of stem cells to nerve cells and migration of nerve cell precursors in preparatory bud development in Hydra

Transcription

1 /. Embryol. exp. Morph. Vol. 60, pp , Printed in Great Britain Company of Biologists Limited 980 Commitment of stem cells to nerve cells and migration of nerve cell precursors in preparatory bud development in Hydra By SEFAN BERKING From the Zoologisches Institute Universitiit Heidelberg SUMMARY Budding in Hydra starts as an evagination of the double-layered tissue in the parent animal's gastric region. Five hours later the density of nerve cells in the bud's tissue doubles, representing the first detectable difference from the cellular composition of the surrounding tissue. hese new nerve cells derive from multipotent stem cells which are in S-phase one day before evagination starts. Some of the bud's new nerve cells derive from stem cells which have migrated into the future bud's tissue after their commitment, apparently attracted by the bud anlage. he bud anlage recruits precursors of nerve cells even during starvation, during which nerve cell production ceases in other parts of the body. Furthermore, the bud anlage controls the duration of the development from commitment to final differentiation of the resulting nerve cells. Experiments with an inhibitor purified from hydra tissue indicate a tight correlation between stages of preparatory bud development and stages of recruitment of nerve cells for the bud. Whether or not precursors of nerve cells are involved in the control of bud formation in normal hydra, as compared to epithelial hydra which still bud though consisting of epithelial cells only, will be discussed. INRODUCION Budding in Hydra may serve as a model system for processes basic to morphogenesis. A new bud is formed at a certain distance from head and foot in tissue of rather uniform composition (Bode et ah 973). A small area becomes specified to develop into the tip of a new bud. his tip is subsequently able to organize the surrounding tissue (Li & Yao, 945; Berking, 979a). he bud is formed by evagination of the double-layered tissue of the parent animal. his process may be comparable to processes involved in tissue invagination in gastrulation. A useful tool for studying the processes which precede visible bud development was a low-molecular-weight inhibitor purified from hydra tissue (Berking, 977). his substance prevents bud development reversibly in very low con- Author's address: Zoologisches Institut, Universitat Heidelberg, Im Neuenheimer Feld 230, D-6900 Heidelberg, Federal Republic of Germany.

2 374 SEFAN BERKING "Surrounding tissuc v Bud Planes of sectioning Fig.. Excision of the bud and the 'surrounding tissue'. centrations. Upon release from inhibition almost all animals start immediately with the same preparatory developmental step. About half a day later evagination begins. he position of the bud was found to be specified only after the end of the treatment with inhibitor (Berking & Gierer, 977). hus it is possible to investigate preparatory developmental steps involved in budding from the very beginning of this process. he present paper deals with the formation of new nerve cells in the course of budding because this is one of the earliest events in bud development. he first visible change in the cellular composition of the young bud, which starts with the same cellular composition as its surroundings, is an increase in the density of nerve cells (Bode et ah 973). he method was to label animals at various times before the bud evaginates by injection of [ 3 H]thymidine (David & Campbell, 972). he label is incorporated in cells which are in the S-phase at the time of injection. hese include stem cells which can give rise to nerve cells and nematocytes. hus labelling the animals at certain times before or after treatment with inhibitor results in the labelling of the cells of bud anlagen of defined developmental stages. here was a further reason to look at the relation between nerve cell development and budding. In whole animals the inhibitor was found to prevent the determination of stem cells to nerve cells (Berking, 979 ). hus it would be interesting to know whether or not the influence of the inhibitor on preparatory bud development can be correlated with its influence on nerve cell formation during budding. MAERIALS AND MEHODS Hydra attenuata were used for all experiments. All experiments were done at 20 C. For details see Bode et al. (973). o determine the cellular composition of the bud and its surroundings, the tissues were cut off as shown in Fig. and disintegrated in acetic acid, glycerol and water (::3). he cells

3 Nerve cell formation in preparatory bud development in Hydra 375 were classified according to David (973). he category 'nerve cells' includes all types of nerve cells bearing processes which can be detected with the light microscope. 'Interstitial cells' (/-cells) denote single i-cells and nests of two (in able nests of four are included); about half of this population appears to consist of multipotent stem cells (Sproull & David, 979). Hydra cells were labelled with [Me- 3 H]thymidine (20 Ci/mM) by injecting 0-2 /d containing 0002 fic\ of the isotope directly into the gastric cavity (David & Campbell, 972). A single injection result in a pulse of about h duration. Labelled cells were classified in cell macerations by autoradiography using Kodak AR0 stripping film. he experimental error was calculated either by use of the binomial distribution or by ^-analysis. For all experiments purified inhibitor was used which was enriched at least 500-fold from crude extract (Berking, 977). Crude extract was used as standard for comparison with purified extract. An optical density (280 nm) of per ml crude extract was arbitrarily defined as one biological unit (I BU). Retardation of budding was observed to occur by treatment with 0-5- BU (Berking & Gierer, 977). reatment with inhibitor was found to not hinder the incorporation of the label into the stem cells, the precursors of the nerve cells (Berking, 9796). For the transplantation experiments animals were stained by feeding them for some days with Artemia nauplii cultured in a solution of Evans Blue (Merck, Darmstadt, FRG) according to Wilby & Webster (970). Standard procedure. Animals were fed daily with Anemia nauplii. One day after the last feeding animals bearing one bud of age up to 6 h were collected (except first experiment, able ) and were fed. Visible onset of budding. he transition from a smooth surface of the budding region to the protrusion of a tiny tip in this region which is stable upon contraction and elongation of the animal is termed visible onset of budding. his process is completed within h. RESULS In the tissue of the young bud the density of the nerve cells is twice as high as in the surrounding tissue he visible beginning of bud formation is an evagination of the doublelayered tissue of the parent animal's gastric region. he very young bud has the same cellular composition as the tissue surrounding the bud. Later an increase in the density of the nerve cells was observed to occur (Bode et ah 973). About 70 additional nerve cells become detectable between 4 and 8 h after evagination has started. his corresponds to a doubling of the nerve cell number (able ).

4 376 SEFAN BERKING able. In buds 8 h old the frequency of nerve cells has increased (x 2 ', P < 0-00) caused by 70 ±29 additional nerve cells in the bud's tissue (9900 cells per bud, i.e. about 0 % of the animals tissue; mean of 23 buds) Age of bud... 3h45 min ± h 7 h 30 min 8 h 5 min ± h 20 min Bud Surrounding tissue Bud Bud Surrounding tissue No. of tissue pieces No. of counted cells Epithelial cells (%) Big /-cells (%) Little /-cells (%) Nematoblasts (%) Nerve cells (%) Gland cells (%) he new nerve cells appear in 5 h old buds day after their precursors were in S-phase Animals bearing one bud were fed (zero time), labelled at 5 h, treated from 6 to 8 h with inhibitor and then allowed to form buds (Fig. 2). he buds were excised at various times after evagination had started and the frequency of labelled nerve cells per bud was determined. he bud's new nerve cells are formed at a certain developmental stage of the bud (Fig. 2C) and not a fixed period of time after labelling (Fig. 2D). hey appear at 5 h after the bud visibly begins to evaginate. Buds of this age change their morphology (Fig. 2B). he labelling index of 50 % indicates that almost all new nerve cells of the young bud (termed 'primary nerve cells') derive from precursors which are in S-phase at the time of labelling. Effect of feeding and inhibitor treatment on the production of the bud's primary nerve cells Feeding allows stem cells which are just in the middle of their S-phase at the time of feeding to develop into nerve cells (the S-phase has a length of about 2 h (David & Gierer, 974)). Stem cells which are exposed to inhibitor while in the first half of their S-phase (inhibitor sensitive phase, ISP) are prevented from developing into nerve cells after the next mitosis. hose in the second half of their S-phase at the time of treatment are not affected (Berking, 99 b). he development of the primary nerve cells appears to follow the same rules. Labelled primary nerve cells are found if the label is injected less than 7 h after the beginning of the treatment (Fig. 3). hus stem cells which end their S-phase at 7 h or later do not develop into nerve cells. hose which end their S-phase

5 Nerve cell formation in preparatory bud development in Hydra 377 A Feeding Labelling reatment with \ ^ / ^ inhibitor Allowed to form new buds 4 6 ime (h) 8 0 A "* en Bud stage 3 Bud stage " S c «20-20 ^ Age of the bud (h) ^ ime after feeding (h) Fig. 2. New nerve cells appear in buds 5 h old. Starting conditions as usual (see Materials and Methods). (A) Selected animals were fed (zero time) labelled 5-6 h) and treated with inhibitor (6-8 h, hatched area, 2 BU). he next day the animals began to form new buds which were excised at different ages and prepared for autoradiographic analysis (C, D); (O O), percentage labelled nerve cells in the bud; (A A) percentage labelled nerve cells in the respective surrounding tissue. Each point in the graph is obtained from counts of nerve cells of preparations of 5-9 buds. (C) he percentage of labelled nerve cells in buds of different ages and respective surrounding tissues plotted against the age of the buds; (D) the same frequencies plotted against the time interval between feeding (zero time as in most of the following representations) and preparation of the animals. he horizontal bars indicate the range of the age of the used buds. For clarity the standard deviations are omitted. (B) Bud stages were classified according to Otto and Campbell (976). earlier, for instance at 5 h after the beginning of a treatment, are not prevented from developing into nerve cells because they are exposed to inhibitor only during the second half of their S-phase. hus in this experiment (Fig. 3) the primary nerve cells derive mainly from those cells which are just in the middle of their S-phase at the time of feeding. Furthermore, feeding on the day of experiment (before the first treatment) greatly enhances bud formation (Berking & Gierer, 977). hus it is suggested that both effects of feeding, allowance of budding and allowance of nerve cell formation, are coupled to one another. he development into the primary nerve cells of the bud is controlled by the bud anlage he result of the following experiment confirms the notion that the developmental period of the bud's first nerve cells is of variable length. Animals were treated with inhibitor (2 BU) for 2 and 8 h respectively. In both cases the treatment was terminated just before the label was injected at

6 378 SEFAN BERKING A D i B i Feeding " I \ # ime (h) A M " - i i i i i t i ime after feeding (h) Fig. 3. Localization in time of the S-phase of those stem cells which give rise to the primary nerve cells. Standard conditions. (A) Animals were fed (zero time) and treated with inhibitor (2 BU) from 20 min to 8 h (hatched area). he percentage of labelled nerve cells (O, A, #, A) and /-cells ( ) found the next day in buds 5 h old is plotted (A) against the time after feeding at which the label was injected into the animals and (B) against the length of the time interval between feeding and preparation. Each point is the result of countings of at least 00 /-cells and nerve cells respectively. 8 h 20 min after feeding. Labelling following 8 h of treatment causes only a few (9 ± 3 %) primary nerve cells to be labelled, as expected from the above experiment, whereas labelling following 2 h of treatment (after 6 h 20 min inhibitor-free development) results in the labelling of almost all primary nerve cells (40 ± 4-5 %) in buds 5-6 h old. hus commitment of stem cells to nerve cells takes place over a certain period of time following feeding, and the beginning of an inhibitor treatment delimits the period in which commitment can take place. However, the finally formed nerve cells are found to be formed not after a fixed period of time after beginning of the treatment but rather when the bud becomes 5 h old (Fig. 2C, D). In contrast, throughout growth, excluding morphogenetic processes like budding, nerve cell development appears to be accomplished within a rather fixed interval after commitment (Berking, 9796). hus the duration of the development from stem cells to the primary nerve cells appears to be coupled to the process which causes evagination. he precursors of the primary nerve cells can migrate Animals were treated and labelled so as to label the precursors of the primary nerve cells (Fig. 4). hen the animals were sectioned in the budding region and transplanted to unlabelled animals which had been treated and sectioned in the same way. he half-animals were transplanted together so that budless animals of normal morphology were obtained. hus the budding region contains labelled and unlabelled stem cells which have become committed to develop o

7 Nerve cell formation in preparatory bud development in Hydra 379 Bll'F ANIMALS - ll.-ccdinb ""' '" t/ I inhibitor y/\ \ i VMn "* Bucllcss animals obtained ype of transplants from which buds were analysed / No of counted cells \ '.I labelled No. of counted cells Interstitial cells I \ C7 labelled cells No. of counted cells L-lpithdial cells \ ',; labelled Cl C C Fig. 4. he precursors of the primary nerve cells appear to migrate. Animals with one just visible bud, normally red coloured ones and blue, vitally stained ones were fed (zero time) and treated with inhibitor (2 BU). he blue ones were labelled; other conditions as usual. he animals were sectioned and transplanted between 4 h 40 min and 8 h, however, in each case immediately after the end of treatment. he table shows the frequency of labelled cells of various types as found either in excised buds (4-0 per given cell count) of totally labelled animals or of buds formed at transplants. In the buds of transplants formed of unlabelled tissue the frequency of labelled nerve cells is higher than the frequency of labelled /-cells (x"; P < 000) whereas in totally labelled animals the reverse was found (x 2 ; 002 < P < 005). into nerve cells, including those which develop into the primary nerve cells of a bud. he next day buds are formed. hose formed from labelled and unlabelled tissue respectively were excised and analysed separately. Buds formed from unlabelled epithelial cells were found to contain about 5 % labelled primary nerve cells but only very few (about 0-5 %) labelled /-cells (about one half of the /-cells appear to be stem cells (Sproull & David, 979)). he Experiments were designed so that stem cells which are not committed at the time of transplantation to become nerve cells should not give rise to the primary nerve cells of the future bud. But if one argues that the primary nerve cells develop only from stem cells which become committed in the future bud's tissue, one has to explain how a population of stem cells of which 0-4 % are labelled could give rise to nerve cells of which 0 % are labelled (in the bud 5 % of the nerve cells were found to be labelled, half of the bud's nerve cells are new ones). his indicates that mainly commited stem cells immigrate into the future bud's tissue, and not uncommitted stem cells.

8 380 SEFAN BERKING reatment with inhibitor Labelling ype Allowed to form new buds 0 ime, h 60 - B 50 _L 40 ~ i l o o I l ~ r- - C O O i i i i i "~l o o Age of the bud (h) ime after treatment (h) Fig. 5. Stem cells which happen to start their S-phase close to the end of the treatment with inhibitor give rise to the secondary nerve cells. Starting conditions as usual. (A) animals were treated (2 BU), from h to 9 h 50 min and labelled either between 7 h 55 min and 8 h 25 min or between 9 h 50 min and 0 h 25 min. Presentation of the result comparable to Fig Secondary nerve cells derive from stem cells which start their S-phase at the end of the treatment with inhibitor In buds 8-5 h old the nerve cell density doubles a further time; these nerve cells will be termed 'secondary nerve cells'. However, the stepwise increase in nerve cell density is caused by the treatment which starts shortly after feeding (first treatment). Stem cells which start their S-phase close to the end of the treatment with inhibitor give rise to nerve cells (I.e.). Secondary nerve cells were formed in the same way: animals were either labelled immediately before or immediately after the end of a treatment (Fig. 5). In buds of early labelled animals fewer nerve cells were found to be labelled than in buds of late labelled animals. he development of stem cells to secondary nerve cells can be prevented by means of a second treatment with inhibitor Animals were treated twice with inhibitor, as shown in Fig. 6, the first time as usual and the second time at 4 h after the first treatment for a period of 2 h 40 min. he animals were labelled either before or after the second treatment. hen the animals were allowed to form buds which were excised for autoradiographic analysis. he result (Fig. 6B) indicates that the secondary nerve cells derive mainly

9 Nerve cell formation in preparatory bud development in Hydra 38 reatment with inhibitor _.. first second Feeding i, ; Labelling l I I' Excision of the new buds for autoradiographic analysis 0 3 ime, h Development into labelled nerve cells ime necessary to Appearencc Stem cells develop processes of labelled S-Phase G Mitosis, nerve cells a -Phase =0= = =Q l m m m rve cells C o % labellc B i O i i I O i i i O i i o " i - O _ i i I i i i i Age of the bud (h) I 22 Fig. 6. Formation of secondary nerve cells after two treatments with inhibitor. Starting conditions as usual. he further procedure is shown in the upper half of the scheme (A). he animals were treated with inhibitor (20 BU) and labelled either at the first (O) or at the second time (#) shown in the scheme. he lower half of the scheme shows the stem cell populations which give rise to labelled secondary nerve cells in the case the second treatment prevents stem cell determination. he result of the experiment is shown in (B). he 0-20 % labelled secondary nerve cells found in the buds of the early labelled animals (O O) may partly derive from stem cells which have become labelled at the end of their S-phase and have become determined ( t) within their next S-phase. from stem cells which start their S-phase after the end of the second treatment. Most of the stem cells which would give rise to nerve cells after the first treatment (those just starting their S-phase) are prevented from becoming nerve cells by the second treatment. 25 EMB 60 i

10 382 First treatment with inhibitor SEFAN BERKING Labelling I Second treatment -2BU ype A 3-2 BU D Allowed to form 6 ime (h) 0 2 % labelled nerve cells in tissue of ype Bud Bud surrounding 55-6 ± ± ± ± ± ± ±2 9 ±2-0 ± ± ± ime after first treatment (h) Fig. 7. Effect of a second treatment with low concentrations of inhibitor on the formation of nerve cells. Starting conditions as usual. he further procedure is shown in the scheme (A). (B) shows the inhibitory effects of the treatments on bud formation. he table shows the frequency of labelled nerve cells found in the newly formed buds (7-0 per given frequency) 8-3 after evagination has started. Further, the frequency of labelled nerve cells in the tissue surrounding the respective buds is shown. With increasing concentrations of the inhibitor the frequency of labelled nerve cells is only slightly lower in the tissue of the bud (X 2 ; 0-3 < P < 05) whereas in the tissue surrounding the bud less than half of the normal frequency is observed (x 2 ; i* < 000). Low concentrations of inhibitor prevent mainly the formation of new nerve cells in the tissue surrounding the bud he animals were treated twice with inhibitor, the first time as usual and the second time with low concentrations for a period of 2 h. Between both treatment the animals were labelled (Fig. 7). he buds which formed the next day were excised and analysed and compared to the surrounding tissue. he data indicate that the secondary nerve cells were formed at almost the same frequency whether the animals were treated or not. However, in the tissue surrounding the bud fewer nerve cells were formed as a result of the treatment. his indicates that either the sensitivity of stem cells to externally supplied inhibitor is lower in the tissue of the future bud or that committed stem cells within the bud's surrounding tissue do migrate preferentially into the tissue of the future bud. hus, in terms of sensitivity to inhibition of nerve cell recruitment, the bud anlage appears to start acquiring a typical property of a head.

11 Nerve cell formation in preparatory bud development in Hydra 383 able 2. Formation of nerve cells in head, bud and tissue surrounding the bud upon starvation and treatment with inhibitor ime labelled 0 h 2 h 30min Head 3-6±l-4-4±-0 Labelled nerve cells (%) in tissue of Bud 44±6 38 ±5 4 Bud surrounding - ±0 0 <0 Starting conditions as usual. he animals were fed and subsequently treated with inhibitor (8 BU) from 30 min to 22 h. At 7 h 30 min after the end of the treatment the animals were treated a second time and labelled at the times given. he procedure allows stem cells which start their S-phase within a short period after the end of the first treatment to give rise to labelled nerve cells; buds 8-4 h old were analysed. One day after treatment and starvation stem cells become determined to develop into secondary nerve cells in the bud but only rarely into nerve cells in other parts of the body Animals were treated with inhibitor for 22 h after feeding and were then allowed to restart bud development. At 0 or 2 h after the end of the treatment the animals were labelled. he next day the animals were dissected and prepared for autoradiographic analysis. he result (able 2) indicates that only a few nerve cells develop from labelled stem cells in the head and gastric regions, just as in untreated animals starved for the same period of time (I.e.). However, secondary nerve cells were found to be formed in normal frequency. hus it is argued that the bud anlage generates a signal which determines stem cells to nerve cells and/or attracts committed stem cells. DISCUSSION he first nerve cells of the bud (primary and secondary) are recruited by two mechanisms: () by local determination from multipotent stem cells; after bud induction this process proceeds even during starvation, at a time when nerve cell determination ceases in other parts of the animal; (2) by immigration of stem cells committed to become nerve cells from outside the future bud's tissue. Uncommitted stem cells appear to enter the area of the future bud only rarely, as a result of random movement. hese findings raise the question as to what extent the increase of the nerve cell density in head and foot regenerating tissue and the normal replacement of nerve cells during steady state growth of these structures is also caused by immigration of committed stem cells. he duration of the developmental period of the primary nerve cells from end of S-phase onwards varies among animals. hey become detectable, synchronously, at 5 h after the bud's evagination begins. hus the differentiation of the primary nerve cells is under strong control of the bud anlage 25-2

12 384 SEFAN BERKING (whatever that may be). Nerve cells start to protrude processes at about 6 h after mitosis (David & Gierer, 974). hus it is argued that the precursors of the primary nerve cells undergo mitosis synchronously around the beginning of evagination. Budding in normal and epithelial hydra In normal hydra the position of the bud was shown to be specified in two steps. First, a belt-like area is specified at a certain distance from head and foot; second, a small area within the circumference of this belt is specified (Berking & Gierer, 977). he kinetics of bud development were studied by applying inhibitor pulses to the animal at various times. Four qualitatively different developmental phases can be distinguished between a stage where the final position of the future bud is not yet specified and the beginning of evagination (Berking & Gierer, 977). he experiments in this paper indicate a tight correlation between these phases and the development of stem cells to the bud's nerve cells. Both the start of bud development and the commitment of the future bud's new nerve cells are initiated by feeding; both processes are not restricted to the presumptive bud's tissue but extend into a rather large area. In the following time, pulses of inhibitor applied within certain periods are able to cancel both preparatory bud development and the development of stem cells to the nerve cells of the bud. It is certainly possible, although unlikely, that the inhibitor affects both processes independently at the same time. It is more plausible that one process has a strong influence on the other. he occurrence of buds in epithelial hydra (Campbell, 976) which are depleted of interstitial cells and their derivatives, the nerve cells and the nematocytes, has shown that epithelial cells alone can produce all the requirements for bud formation. hus nerve cell commitment may be argued to depend on signals generated by epithelial cells. However, the kinetics of nerve cell commitment as determined by means of inhibitor treatment was found to be the same during preparatory bud development and during growth throughout the whole animal. It thus appears difficult to understand how the inhibitor should affect the bud's pre-pattern which in turn affects the commitment of stem cells to nerve cells when nerve cell commitment alone shows the same kinetics. In addition, the effect of the inhibitor on budding and on nerve cell formation cannot be explained by a single mechanism. On the other hand, a simple possible explanation would be that in normal hydra not only epithelial cells but other cell types also have an influence on budding. he experiments with epithelial hydra do not exclude that in normal hydra cell types which are not present in epithelial hydra effectively control budding. hese cells may just be faster or more efficient than epithelial cells in generating the appropriate signals themselves or in stimulating epithelial cells to generate these signals earlier than epithelial cells would on their own.

13 Nerve cell formation in preparatory bud development in Hydra 385 A possible role of nerve cell precursors in the control of bud formation One possible explanation for the experimental data is that the development of the future nerve cells of the bud is the velocity-determining step in preparatory bud development. According to this hypothesis evagination is triggered if a certain density of nerve cell precursors which have reached G 2 -phase is attained in the presumptive bud's tip. hese precursors will differentiate as primary and secondary nerve cells. Furthermore, within a short period at the very beginning of the inhibitor sensitive phase (ISP) stem cells are most sensitive to inhibitor. Based on these assumptions the influence of the inhibitor on preparatory bud development consists of only one effect, to prevent stem cells from becoming nerve cells: he first phase of bud development starts in all animals immediately after the end of the first treatment. his phase was found to be the most sensitive one. Phase has a variable length between and 0 h. Evagination starts at 2 h after end of phase. he explanation would be that the first phase represents the period in which precursors of nerve cells which later contribute to the threshold density when they reached their G 2 -phase are still in the very early, i.e. the most inhibitor sensitive, part of their ISP. At the end of phase the last of these 'necessary' precursors leaves the most inhibitor sensitive part of its ISP and continues to travel through S-phase, which takes about 2 h. When it enters G 2 -phase the threshold density is reached, therewith triggering evagination. he variable length of phase may reflect differences between the animals in the ability to recruit the appropriate number of stem cells for the development into nerve cells in the future bud's tissue. he second phase represents the period in which necessary precursors are still in their ISP but not at its very beginning. A second treatment within this period was found to prevent stem cells from becoming committed to develop into the secondary nerve cells. After the second treatment a fresh start into secondary nerve cells takes place. his can explain why a second treatment within the first 6 h after the first treatment influences the bud anlagen of all animals. hey are all within their first or second phase of bud development. he treatment cancels the result of the bud's preparatory developmental processes from the end of the first treatment onwards. After the end of the treatment the development starts again with phase. After the 7th hour some bud anlagen have become insensitive to inhibitor (3rd phase). he first buds which will be formed after the second treatment start evagination at the same time as the first of a control group treated only once. However, the number of newly formed buds per time unit is lower in the group treated twice and a plateau, a maximal frequency of formed buds, is reached. he slope of the curve which describes this increase per unit time

14 386 SEFAN BERKING depends on the beginning of the second treatment (the later the steeper), not on its end. Further, the slope of the curve does not depend on the concentration of the inhibitor applied if it is above a certain threshold. he explanation would be that a treatment which starts 7 h after the first treatment is no longer able to prevent the commitment of all stem cells (the ISP has a length of less than 7 h). he result is that the density of committed stem cells in animals treated twice is lower than normal, also in the tissue of the future bud and its surrounding. he later the treatment starts, the more stem cells have been committed and left their ISP. hus the number of stem cells which have been committed is determined by the beginning of the second treatment, not by its duration and not, above a certain threshold, by the applied concentration. Consequently the threshold density in the future bud can be attained only if committed stem cells migrate into the future bud's tissue from an abnormally large area, and this would take more time. hus the treatment causes some bud anlagen to not attain the threshold density at all, and some only after an unusually long period. he fourth phase starts immediately before evagination. he effect of an inhibitor treatment is almost immediately reversible. Its explanation may require a further assumption. It should be emphasized that this discussion of the role of nerve cell precursors in control of budding - as one possible explanation of the data - does not exclude the involvement of epithelial cells. However, it may indicate that if cell types other than epithelial cells are present in animals, some of them can effectively control budding just by being faster or more efficient than epithelial cells to generate the signals which finally push epithelial cells to start evagination. In normal hydra, as in epithelial hydra, epithelial cells may initiate bud development by generating the pre-pattern of the bud which causes, if possible, a local accumulation of nerve cell precursors in the future bud's tissue and later on evagination. In normal hydra, however, it cannot be excluded that even the very beginning of budding is controlled by precursors of nerve cells. It is possible that they control their own accumulation out of the presumptive bud's surroundings, like slime moulds in fruiting body formation control their own accumulation. I would like to thank Dr W. Miiller for support and providing laboratory space; Dr A. Gierer, Dr H. Meinhardt and L. Graf for critical reading of the manuscript, R. Heisner for excellent technical assistance and the Deutsche Forschungsgemeinschaft for support. REFERENCES BERKING, S. (977). Bud formation in Hydra: Inhibition by an endogenous morphogen. Wilhelm Roux' Arch, devl Biol. 8, BERKING, S. (979a). Analysis of head and foot formation in Hydra by means of an endogeneous inhibitor. Wilhelm' Roux Arch, devl Biol. 86,

15 Nerve cell formation in preparatory bud development in Hydra 387 BERKING. S. (9796). Control of nerve cell formation from multipotent stem cells in Hydra. J. Cell Sci. 40, BERKNG, S. & GIERER, A. (977). Analysis of early stages of budding in Hydra by means of an endogeneous inhibitor. Wilhehn Roux Arch, devl Biol. 82, BODE. H. R., BERKING, S., DAVID, C. N., GIERER, A., SCHALLER, H. C. & RENKNER, E. (973). Quantitative analysis of cell types during growth and morphogenesis in Hydra. Wilhehn Roux' Arch. EntwMech. Org. Ill, CAMPBELL, R. D. (976). Elimination of Hydra interstitial and nerve cells by means of colchicine. /. Cell Sci. 2, -3. DAVID, C. N. (973). A quantitative method for maceration of Hydra tissue. Wilhehn Roux' Arch. EntwMech. Org. Ill, DAVID, C. N. & CAMPBELL, R. D. (972). Cell cycle kinetics and development of Hydra attenuata.. Epithelial cells. /. Cell Sci., DAVID, C. N. & GIERER, A. (974). Cell cycle kinetics and development of Hydra attenuata. III. Nerve and nematocyte differentiation. /. Cell. Sci. 6, Li, H. P. & YAO,. (945). Studies of the organizer problem in Peimatohydra oligactis. III. Bud induction by developing hypostome. /. exp. Biol. 2, OO, J. J. & CAMPBELL, R. D. (976). Budding in Hydra attenuata bud stages and fate map. /. exp. Zool. 200, SPROULL, F. & DAVID, C. N. (979). Stem cell growth and differentiation in Hydra attenuata. I. Regulation of self-renewal probability in multiclone aggregates. /. Cell Sci WILBY, O. K. & WEBSER, G. (970). Studies on the transmission of hypostone inhibition in Hydra. J. Embryol. exp. Morph. 24, {Received 27 March 980, revised 4 June 980)

16