- Email: [email protected]
During Drosophila embryogenesis the β1 tubulin gene is specifically expressed in the nervous system and the apodemes
Mechanisms of Development, 33 (1991) 107-118 © 1991 Elsevier Scientific Publishers Ireland, Ltd. 0925-4773/91/$03.50
107
M O D 00009
During Drosophila embryogenesis the B1 tubulin gene is specifically expressed in the nervous system and the apodemes Detlev Buttgereit, Dagmar
L e i s s *, F r i t s M i c h i e l s a n d R e n a t e R e n k a w i t z - P o h l
Gentechnologische A rbeitsgruppe am M P I ]'fir Biochemie, Martinsried / Miinchen, F. R. G. (Accepted 3 September 1990)
We determined the in vivo distribution of the fll tubulin from D. melanogaster using isotype specific antibodies. Maternally expressed/~1 tubulin is incorporated into mitotic spindles. Later in development a strong expression in the CNS is observed. Furthermore, all chordotonal organs and the apodemes are marked by fll tubulin. Nuclear run-on assays and stage specific in vitro transcription showed a zygotic expression of the/31 tubulin gene from the extended germ-band stage onwards. Using the P-element system, we identified several elements; upstream between - 2 . 2 kb and the transcription initiation site, elements for low level expression in the CNS are present. In the intron between + 0.44 kb and + 2.5 kb enhancer elements are located that drive the expression in the chordotonal organs and the apodemes. Between the start site and +0.44 kb (273 bp) and +2.5 kb and the second exon (315 bp), maternal and CNS enhancers result in full level expression of a lacZ-fll reporter gene. We show, that the fll tubulin gene is a very early effector gene starting its expression shortly after the commitment of neuroblast cell fate. This gene offers an excellent model system for the identification of neural and apodeme specific transcription factors. /31 Tubulin; Regulatory element; Nervous system; Apodeme; Drosophila
Introduction Microtubules form major structural components in all eukaryotic cells. They participate in structures such as mitotic and meiotic spindles, flagellae, ciliae, the cytoskeleton and neural processes (for review see Roberts and Hyams, 1979). In the nervous system, microtubules are rather abundant. In all multicellular organisms investigated so far, the a and /3 tubulins are encoded by multi-gene families of varying extent (for reviews see Cleveland, 1987; Little and Seehaus, 1988; Sullivan, 1988). In Drosophila, four tubulin genes have been identified with distinct developmental expression (Bialojan et al., 1984; Natzle and McCarthy, 1984). Studies concerning the/32 tubulin during spermatogenesis (Michiels et al., 1987, Kemphues et al., 1982, Raft,
Correspondence to: R. Renkawitz-Pohl, Gentechnologische Arbeitsgruppe am MPI fi.ir Biochemie, A m Klopferspitz 18A, 8033 Martinsried/Miinchen, F.R.G. * Present address: EMBL, 6900 Heidelberg, F.R.G.
1984) and the f13 tubulin gene during embryogenesis (Rudolph et al., 1987; Gasch et al., 1988, Leiss et al., 1988) revealed a tissue and development specific expression of these protein variants. These investigations showed, that during embryogenesis the f13 tubulin is restricted to the mesoderm and its derivatives (Leiss et al., 1988), while the/32 tubulin is solely expressed in the developing testes (Kemphues et al., 1982). Thus, all other essential functions such as spindle formation and microtubules of the nervous system require another 13 tubulin isotype. In Drosophila, a cascade of genes regulates the formation of the central nervous system (CNS) and peripheral nervous system (PNS). Neurogenic loci like Notch (Lehmann et al., 1981; Johansen et al., 1989), Delta (V~issin et al., 1987) or Enhancer of split (Knust et al., 1987) participate in the decision of an ectodermal cell to develop as a neuroblast or epidermoblast. Genes as daughterless (Caudy et al., 1988) and the achaete-scute complex (Garcia-Bellido, 1979) are required for the proper development of CNS (Jimenez and CamposOrtega, 1979) or PNS (Dambly-Chaudiere and Ghysen, 1987) cells. Further downstream, active genes regulate
108 aspects like neuron formation, axone outgrowth, cell identity, e.g., slit (Rothberg et al., 1988), single-minded (Thomas et al., 1988), cut (Bodmer et al., 1987) or fushi tarazu (Doe et al., 1988). During the maturation from the egg to the adult fly, tubulins are needed in large amounts for the structures mentioned above. In this report the distribution of the fll tubulin during embryonic development will be shown. During the first 90 min of embryonic development, all R N A has to be supplied to the egg by maternally synthesized message, because no transcription is observed during this period. For these early stages in situ hybridizations to embryo sections revealed only transcripts of the 131 tubulin gene, which are uniformly distributed throughout the embryo (Gasch et al., 1988). Later, from stage 10 onwards, a concentration of the message is observed in the developing central nervous system and the brain. In order to identify the mechanism of this accumulation, and to decide in which microtubules fll tubulin is incorporated, several independent approaches were chosen. The experiments described in this paper characterize the distribution of the fll tubulin analysed by isotype specific antibodies, and determine the onset of zygotic transcription by run-on assays and in vitro transcription. Furthermore we studied the tissue specificity by P-element mediated transformation with 131-1acZ fusion genes. These analyses revealed that the transcription of the 131 tubulin gene starts shortly after the action of genes which participate in the decision of epidermal versus neural celt fate (for review see Knust and Campos-Ortega, 1989). As we will point out, the expression of the fll tubulin gene is regulated by independent, cell-typespecific enhancers in the intron for the PNS and apodemes, while maternal and CNS expression is directed by cooperation between upstream promoter elements and corresponding enhancers in the intron.
Results
Determination of the wild-type 131 tubulin pattern As a prerequisite for further studies, the distribution of the 131 tubulin in the developing embryo had to be determined. We have shown previously, that the Drosophila 13 tubulin isotypes mainly differ in the last 15 amino acids (Michiels et al., 1987; Rudolph et al., 1987). Therefore, polyclonal antibodies against synthetic peptides corresponding to the C-terminal amino acids of the fil tubulin were raised in rabbits. Their specificity was tested both by competition with specific and unspecific peptides in ELISA assays (data not shown) as well as by analysis of proteins extracted from different embryonic, larval and pupal stages and adults on Western blots. As Fig. 1A shows, a roughly constant amount of fll tubulin is detected in every stage ex-
amined. As a control, the 133 tubulin antibody (Leiss et al., 1988) was used in parallel on a second filter. Only in 8 to 16 h embryos, mid and late pupae and adult testes is 133 tubulin detectable as has been published previously (Leiss et al., 1988). Having confirmed the isotype specificity of the antibody, the distribution of the 131 tubulin was analysed in whole mount stains of embryos from D. melanogaster. Embryos were collected, fixed and stained utilizing taxol to stabilize microtubules as described in Methods. Freshly laid eggs and very early embryos showed an uniform distribution of the 131 tubulin (not shown). As soon as mitotic spindles are visible, the 131 tubulin is concentrated in these structures (Plate IA). In accordance with the known synchronism of the first mitotic divisions (Karr and Alberts, 1986; Foe, 1989), spindles are visible at the same time over the whole embryo. Around stage 8 (Plate 1B) the cytoplasmic staining is still uniform except in those regions that correspond to mitotic domains as determined by Foe (1989). She defines a mitotic domain as a region of the embryo where all cells divide synchronously. For example, Plate 1B shows in the posterior part of the embryo nuclear localization in a region corresponding to mitotic domain 8144. During the next hours of embryonic development, the distribution of the 131 tubulin is uniform and both spindle and cytoplasmic microtubules are decorated by the 131 antibody. At stage 16 when the condensation of the CNS is nearly completed and the PNS is visible, a high concentration of the /31 tubulin is detected in the CNS and the chordotonal organs (Plate 1D). Furthermore, in agreement with the in situ hybridizations published previously (Gasch et al., 1988), a label over the apodemes is found (not shown; due to the strong residual staining of maternal protein, these structures cannot be reproduced photographically).
Zygotic transcription of the fll tubulin gene starts at stage 10 The observed fll tubulin distribution in the embryo led to the question, whether the concentration in the CNS is due to the higher number and density of cells or if it is derived from a tissue-specific zygotic expression of the 131 tubulin gene. A concentration of the /31 message in the CNS is observed round stage 11 (Fig. 2D), thus far earlier than detected with the 131 specific antibody by looking at the protein. Concomitantly, the maternal 131 message in the other parts of the embryo is rapidly degraded, while the protein is still present in large amounts. To determine whether this effect is based on newly synthesized RNA, nuclear run-on assays were performed. Nuclei were isolated from staged embryos and examined for R N A synthesis (see Methods). As Fig. 2B shows, no detectable level of fll message is found in embryos of less than 6 h of development, while a strong signal is seen in the histone H 3 / H 4 control at the same stage. In contrast, 6 to 12 h old embryos yield
109
1
2
3
4
5
6
1
2
3
4
5
6
7
8
9
10
11
12
8
9
10
11
12
13
A
~ii i~iiiiii~ii~!! !I .....
7 ~
13
•
Fig. 1. Distribution of/31 and /33 tubulin lsotypes during Drosophila development. A developmental Western-blot was performed using 60 ~g of total protein per lane. (1) ovaries; (2) embryos 0 h; (3) embryos 8-16 h; (4) embryos > 19 h; (5) first instar larvae; (6) pupae 0 h; (7) pupae 26 h; (8) pupae 49 h; (9) pupae 72 h; (10) pupae 96 h; (11) testes; (12) males; (13) females. (A) Specific reaction with the/31 antibody; this tubulin isotype is detectable in all developmental stages in roughly equal amounts. (B) Specific reaction with the f13 antibody; this tubulin isotype is detectable only during mid-embryogenesis(8-16 h), in pupae (26 h, 49 h, 72 h) and adult testes.
a clearly visible /31 signal. These experiments showed that no zygotic expression of the /31 tubulin gene is observed during early embryogenesis and that detectable levels of transcription are not found before 6 h of embryonic development. To obtain further evidence we confirmed these resuits by in vitro transcription assays of the/31 tubulin gene in nuclear extracts from staged embryos. The template contained /31 sequences from - 2 . 2 kb to + 0.44 kb thus comprising the first exon as well as 250 bases of the intron. Synthesized R N A was analysed by S1 mapping as shown in Fig. 2C. In accordance to the data from the run-on assays, no transcription of the/31 template (Fig. 2A) can be achieved in extracts from embryos staged 0 - 6 h. If the extracts were prepared from 6-12 h embryos, a strong correctly initiated /31 transcript is observed. As a control for transcriptional activity of the extracts, the histone gene promoters for the H3 and H4 genes were chosen (Goldberg, 1979). Fig. 2C shows that a high level of transcription is found in the early extracts in agreement with the in vivo transcription of histone genes (Anderson and Lengyel, 1980). Taken together, these results present clear evi-
dence that the zygotic transcription of the /31 tubulin gene does not start before 6 h of embryonic development, which parallels the observed concentration of R N A and protein in the developing CNS.
- 2 . 2 kb of upstream sequences and the intron are sufficient to achieve the complete expression pattern F r o m the analysis of fll m R N A and /31 tubulin isotype distributions, as well as from the run-on and in vitro transcription assays, we strongly favored the hypothesis that all uniformly distributed /31 tubulin is maternal in origin and that the zygotic expression is restricted to the nervous system. In order to identify cis-acting elements in vivo, we used the P-element transformation system. Eggs were injected with constructs derived from the vector pW8 (Klemenz et al., 1987) which contained the bacterial /3-galactosidase gene as a reporter and varying parts of the/31 gene (see Fig. 3 and Methods). In the fll tubulin gene, a large intron between the codons for amino acids 19 and 20 is present (Michiels et al., 1987). In analogy to the regulatory function of the intron of the mesoderm specifically expressed /33 tubulin gene, we sus-
N
; !
ili~:il~! ;i: :i~
i'L
~i'~i i~ii~i!~i~,i~i~i~i~!
111
A
-2.2 kb
+1
+ 0.44 kb Template
pUC 19
fll
pUC 19 Sl-probe
B
C
NE
n
0-6h +
0-6h
NE
NE
0-6h +
6-12h +
6-9h
D ~-A
J
B1 pUCI9
m
"-81
B3
Act5C H3/H4
iliiiiiiill i¸~¸
i'~i~i!iii
~i!ii!iil;~,,,;~ !i~
"~'i
Fig. 2. Determination of the zygotic transcription of the fll tubulin gene. (A) Template for the in vitro transcription. (B) Transcriptional run-on assay using isolated nuclei. No transcription of the fll tubulin gene is detected in the nuclei from embryos staged 0-6 h, while a clear signal can be seen using embryos staged 6-9 h. (C) In vitro transcription assay. The template (see A) used comprised 2.2 kb of upstream and 0.44 kb of downstream sequences (open box = leader; filled box = coding region; thick bar = intron). Only extracts from embryos staged 6-12 h efficiently transcribe the 131 template. (NE = nuclear extract; a-A = a-amanitine at 0.5/~g/ml). (D) In situ hybridization of embryo sections with a ill-specific probe. (1) Homogeneous distribution of the message in a stage 8 embryo. (2) Concentration of the message in the developing CNS (arrow) in a stage 11 embryo.
pected that within the t u b u l i n gene family, the functional i m p o r t a n c e of the i n t r o n m a y have b e e n conserved. T o look for potential regulatory elements in this region, three different constructs were m a d e which contain either the complete i n t r o n ( = WfllL), 273 b p of 5 ' a n d 315 b p of 3' sequences from the i n t r o n ( = W f l l K ) or are lacking the i n t r o n ( = WfllC). All constructs h a d 2.2 kb of the 131 u p s t r e a m sequences. F o r each construct, expression of the i n d i c a t o r gene was m o n i t o r e d b y i m m u n o s t a i n i n g of e m b r y o s using a m o n o c l o n a l /3-gal antibody. T h e p a t t e r n was c o m p a r e d to the wild-
type d i s t r i b u t i o n of the fll t u b u l i n . The results are s u m m a r i z e d i n Fig. 3. Strains t r a n s f o r m e d with W f l l L show a d i s t r i b u t i o n of the i n d i c a t o r gene p r o d u c t indist i n g u i s h a b l e f r o m the fll t u b u l i n . As was to be expected, n o u p t a k e of the fusion p r o t e i n into spindles was observed since o n l y the first 23 a m i n o acids of the t u b u l i n are c o n t a i n e d in these constructs. I n unfertilized eggs as well as i n t h e early stages of embryogenesis, fl-galactosidase is p r e s e n t i n all cells. I n agreement with the wild type situation, a c o n c e n t r a t i o n of the fll-gal fusion p r o d u c t i n the C N S is observed in later stages of
Plate 1. Determination of the in vivo distribution of the fll tubulin by isotype-specific antibodies. (A) Blastoderm embryo showing the mitotic spindles (arrow). (B) Stage 8 embryo; the distribution of the fll tubulin is nearly uniform except for a nuclear localization in mitotic domain 8144 (arrow). (C) Stage 13 embryo showing an uniform labelling by the antibody. (D) Stage 17 embryo; the ventral cord (vc) and the supra-oesophageal ganglion (spg) show an intense staining. Plate 2. Zygotic expression of the transformed fll-lacZ fusion genes is prominent in the developing CNS. Crosses including the maternal (A, B) or without the maternal expression (C, D) were performed. Without maternal expression, solely the CNS shows a strong label by the lacZ antibody. Comparison of a blastoderm (A, C) and a stage 13 (B, D) embryo are shown.
112
Plate 3.
Plate 4.
113
Expression number of strains -2.2 kb
WB1L
CNS
2
+++
+++
6
++
++
11
+/"
partial
chordotonal organs
apodemes
2.6 kb
I
C ,,~" !
pW8
[31
-2.2 kb
WfllK
maternal expression
iacZ
+++
+++
pW8
0.6 kb
I
[]
pW$
I
131
lacZ
pW$
-2.2 kb
WB1C
,
I pW$
~1
lacZ
.
.
.
.
pW8
Fig. 3. Expression patterns of strains stably transformed with different/31-1acZ fusion constructs. Adjacent to the constructs the expression patterns are outlined. The strains transformed with W/31L show the complete pattern as the/31 tubulin. In W/31K, maternal expression is reduced, and no expression could be detected in chordotonal organs and the apodemes. Maternal expression of the W/31C strains is barely detectable. The expression in the CNS in W/31C strains is not uniform as with the other constructs, but restricted to a number of cells.
embryogenesis. In addition, the chordotonal organs and the apodemes are marked by the expression of the reporter gene as is shown in detail in Plates 2 and 3. A similar pattern emerges from the analysis of the strains transformed with WfllK. As in WfllL strains, the transformed reporter gene is maternally expressed, which is also demonstrated by histochemical staining of ovaries (not shown). Later in embryonic development, the concentration in the CNS is visible as in WfllL. As a most striking difference to WfllL no expression of the reporter gene is seen in the chordotonal organs and the apodemes. Thus, cis-acting elements responsible for the expression in these tissues must reside between + 0.44 and +2.5 kb of the intron; however, both constructs contain all control elements guiding the tissue specific expression in the ovary as well as in the CNS. In
addition, the level of fl-galactosidase is much higher in WfllL than in WfllK strains, indicating a quantitative effect of the deleted parts of the intron on the level of expression. The importance of the intron was further augmented by the analysis of the strains transformed with WfllC. As a major difference, the amount of maternally supplied protein is reduced to a very low level, as freshly laid eggs show only a very faint staining. In addition, the expression in the CNS is detected at the correct time, but at a reduced level. Comparison with WfllL or WfllK shows that the distribution is not so uniform, with a few neuroblasts having a much higher level of the fl-gal than the rest (not shown). Taken together, these data lead to the following conclusions: for a subset of neural cells, all sequences for time and tissue specific expression at low level are present in
Hate 3. Zygotic expression of/31-1acZ fusion genes in W/31L strains. (A) Embryo stage 10; a label is seen in the cephalic region, and in addition single neuroblasts (nb) in the CNS region are stained. (B) Late stage 10 embryo. (C) Embryo stage 14: besides the the ventral cord (vc), precursors of the PNS are labelled (arrows). (D) Stage 17 embryo; the supra-oesophageal ganglion (spg), the ventral cord, the chordotonal organs (co) and the apodemes (ap) are decorated by the antibody. Plate 4. Expression of the reporter gene in the fllWHI strains. (A) Lateral view at a late stage 12 embryo; groups of cells are labelled in the parts of the embryo where the PNS will form. (B) Stage 17 embryo, lateral view; a strong label is seen over the chordotonal organs (co), over axones and the apodemes (ap). In addition, cells are marked in a position (arrow) were the pleural lateral muscles (pl) insert into the epidermis (Campos-Ortega and Hartenstein, 1985). (C) Ventral view at a stage 17 embryo; the visible apodemes iapo, inal and ina2 are marked by arrows. (D) Lateral view on a embryo stained for/33 tubulin. (E, F) Double-staining of embryos with a/33 tubulin specific antibody and the lacZ antibody. /33 stained muscles are light-brown, while the lacZ positive cells are black-brown. Muscles insert at the position of the apodemes iapo (thin arrow), inal and ina2 (thick arrows).
114 the 5' flanking 2.2 kb; for the expression in the complete CNS, elements in the intron included in WfllK (273 bp of 5' and 315 bp of 3' sequences) are necessary. With the additional intron sequences contained in the W/31L constructs, further regulatory units result in a high level of expression, guide the transcription in the chordotonal organs and the apodemes and resemble enhancer elements. The zygotic expression of the fll tubulin gene is restricted to the neroous system and the apodemes The transformed strains offered the possibility to separate the maternal and zygotic expression of the indicator gene under the control of the fll regulatory sequences. Males carrying the W/31L construct were mated to females from the recipient white strain so that the eggs deposited by the females had no maternally supplied/3-gal. As a result, the zygotic expression mediated by the paternally inherited indicator gene can be followed independently as is shown in Plate 2. For comparison embryos from reporter gene positive females are shown. The eggs deposited from the white females show no detectable level of/3-gal. After stage 10, when the development of the CNS proceeds, a strong label over neural tissues and the brain can be seen (Plate 2D). So definitively the neural derivatives are the main tissues of the embryo that zygotically express the reporter gene under the control of the/31 tubulin gene regulatory sequences. This tubulin isotype represents a good marker to follow the development of the CNS and PNS. In addition, as will be shown below, the anchoring points of the muscles to the epidermis, the apodemes, are characterized by 131 tubulin expression. Development of the CNS followed by reporter gene expression In order to determine precisely the developmental stage of first /31-/3gal expression, several crosses using W/31L as described above were performed. Eggs were collected and staged for 0 to 22 h at 25 ° C. At stage 10, when epidermoblast and neuroblast are determined in their cell fate (for review see Knust and Campos-Ortega, 1989), we first observe /3-gal staining. As soon as the first neuroblasts divide, a label over the neuroectoderm can be seen (Plate 3A). The intensity of antibody staining increases very rapidly corresponding to the increasing number of cells in the CNS region (Plate 3B). When germ band retraction has further proceeded, groups of cells are labelled in a region where the PNS will form (Plate 3C). During further development of the embryo, the differentiating neural structures are permanently marked by the expression of the /31-1acZ fusion gene. Structures like axones, ventral cord and chordotonal organs can be visualized in the embryo. Shortly before hatching of the larvae the condensed CNS, the brain hemispheres, the PNS and the apodemes are stained
(Plate 3D). An identical result was obtained by staining of embryos from W/31K strains, except that as was stated above, no expression is detected in the chordotonal organs and the apodemes. For the developing and mature nervous system, microtubules are essential molecules. It represents the tissue with the highest rate of mitotic division from stage 10 onwards and the spindle microtubules contain, in our judgement, exclusively/31 tubulin. The analysis of zygotic/3-gal expression revealed that the/31 gene is a very early effector gene beginning its expression shortly after the commitment of ectodermal cells in epidermoblast and neuroblast. Cell-type specific enhancers are located in the intron In order to address the question whether independently acting elements are situated in the intron of the /31 tubulin gene, a standard enhancer test in the context of a heterologous promoter was performed. We used a truncated hsp70 promoter, cloned a 2.5 kb fragment of the intron (thereby deleting 0.15 kb of the 3' part of the intron) upstream in both orientations /31WHI and fllWHII), and established P-element transformed fly strains. As no difference could be seen in the expression pattern between the two orientations of the intron (5 lines analysed), only the construct (/31WHI is shown. Staining of the embryos revealed that only the apodemes, the chordotonal organs and axones are marked by the expression of the reporter gene (Plate 4B). Development of the chordotonal organs can be followed from stage 12 onwards. In addition, the formation of the apodemes is seen round the same stage, in accordance with the beginning formation of somatic muscles (Leiss et al., 1988). All three types of apodemes (iapo, inal, ina2; Campos-Ortega and Hartenstein, 1985) can be seen (Plate 4C). In order to unambiguously identify the apodemes, double immunostaining using the 133 tubulin antibody to mark the somatic muscles and the /3-gal antibody to show the apodemes in the transformants was performed (see Methods). As Plates 4E and 4F clearly demonstrate, the/33 stained muscles insert at the position were the /3-gal labelled cells are located in the epidermis. So these structures represent the anchoring points of the muscles to the epidermis, the apodemes. As reference, an embryo stained only with the 133 tubulin antibody is shown in Plate 4D. In contrast to the results obtained with the/31 tubulingene promoter, in the/31WHI and/31WHII transformants neither maternal expression nor expression in the CNS is detected. So the elements in the intron that regulate these modes of expression have to cooperate with elements of the/31 promoter, or reside in the 0.15 kb of fll intron sequences not analysed yet. Additional autonomously acting cell-type specific enhancers are found in the intron that are capable of activating a heterologous promoter independent of their orientation,
115 thus resembling classical enhancers. They drive the expression in the PNS and in the apodemes. The next steps in analysing these different elements will be the further subdivision of the intron in order to characterize the enhancers for PNS and apodemes, and the identification of upstream and intron elements for the maternal and CNS expression.
Discussion We have started the analysis of expression of the fll tubulin gene. As a first step, the in vivo distribution using isotye specific antibodies was determined. Staining of whole mount embryos revealed in early stages a uniform distribution of maternally synthesized protein in accordance with the R N A in situ localization published previously (Gasch et al., 1988). When the initial mitotic divisions take place, the fll tubulin is found in the spindles. During the development of the embryo, spindles can also be observed in the mitotic centers as described (Foe, 1989). In later stages, a concentration of the fll tubulin in the CNS and the brain hemispheres is observed. In addition, the anchoring points of the musculature to the epidermis, the apodemes, are marked by fll tubulin expression. Thus, this tubulin isotype represents a ubiquitous, multifunctional tubulin during embryogenesis. In order to evaluate whether all fll tubulin is due to maternal synthesis or if zygotic expression of the fll tubulin takes place, run-on assays were performed. They revealed a transcription of the gene in embryos older than 6 h. Similar results were obtained from in vitro transcription experiments using a fll template containing sequences between - 2 . 2 kb and +0.44 kb and nuclear extracts from staged embryos. So this assay system will help us to identify and purify some of the factors that participate in the regulation of the fll tubulin gene. Heberlein and Tjian (1988) showed a correct transcription initiation for the alcohol dehydrogenase gene in staged embryo extracts, and isolated a factor that binds to regulatory sequences of this gene. Similar experiments showed the presence of several upstream and downstream regulatory sequences in the promoter region of the Ubx gene (Biggin and Tjian, 1988) and the engrailed gene (Soeller et al., 1988). These data showed that the in vitro systems allow at least partially the identification of regulatory sequences, which are also active and necessary in vivo, as well as the identification and the isolation of trans-acting factors. To localize the domains of zygotic fll tubulin expression, P-element transformed fly strains using different constructs were established. We could show that - 2 . 2 kb of upstream sequences and the presence of the 2.6-kb long intron are necessary for maximum level of
maternal and zygotic expression of a fll-lacZ reporter gene. In accordance with the distribution of the fll tubulin, the ventral cord, the chordotonal organs, the brain and the apodemes are the domains of expression. Deletion of the intron nearly completely abolishes the maternal expression, and the expression in the CNS is restricted to a subset of cells. In the PNS, only few cells are labelled. That means that important features of the regulation of the fll gene expression are mediated by elements in the intron. In contrast, in the context of a heterologous promoter, the intron drives only the expression in the PNS and the apodemes. These results clearly show, that at least two different types of enhancer elements are contained within the intron: one type, which can confer cell type specifity autonomously and is able to cooperate with a heterologous promoter, and a second type that requires upstream elements of the fll tubulin gene p r o m o t e r to enhance or activate expression at the correct time and tissue. Whether these modes of regulation are controlled by completely unrelated elements or by specific combinatorial interactions between related elements has to be clarified. Recently, a number of reports for regulation mediated by enhancers in intron regions has been found in Drosophila. Basler et al. (1989) showed that regulatory elements for the transcription of the sevenless gene reside in the intron. Within the fl tubulin gene family, the trait of a large intron is conserved between the fll tubulin gene (2.6 kb intron) and the r 3 tubulin gene (4.5 kb intron). In both genes, this intron is found between codons for amino acids 19 and 20 (Michiels et al., 1987, Rudolph et al., 1987). It was shown that regulatory elements in the intron region of the r 3 tubulin gene are necessary for the expression of the tubulin in the visceral mesoderm (Gasch et al., 1989). The zygotic fll expression, as it is shown in this report, is restricted in the embryo to neural tissues and the apodemes. So the regulatory elements in the introns of these genes are quite different. The fll tubulin gene is certainly not a part of the regulatory cascade driving the formation of the nervous system. But as we could show, it is zygotically expressed in all neural structures. In addition, this expression starts as early as the first neuroblasts in the CNS and the first cells in the PNS region can be identified. That means, that the fll tubulin is one of the first effector genes switched on after the determination of cell fate. So this gene must be under the direct or indirect control of genes that determine cell fate in the ectoderm, and may be common to all neural derivatives. Regulation of transcription of genes expressed in the CNS and PNS is under investigation. Bray et al. (1989) and Dynlacht et al. (1989) have identified cis-acting elements in the dopa decarboxylase promoter and cloned a trans-acting factor that binds to those sequences. But the expression of this factor is restricted to a subset of neural cells and it
116 is not a general neural-specific transcriptional activator. For other genes expressed in the CNS or PNS like ftz, slit, cut, or sim, the modes of expression render it unlikely to easily identify general neural-specific cisand trans-acting elements. Strategies using antihorseradish peroxidase antibodies (Jan and Jan, 1982) or monoclonal antibodies (Fujita et al., 1982) to characterize neural specific genes have led to the identification of new gene products, but their functional analysis has not been completed yet. By application of enhancer trap strategies (Bier et al., 1989, Wilson et al., 1989) expression patterns were identified that show staining in subsets of the neural system. But as the authors stated, differences in the stability between the reporter gene and the gene product of the target site may lead to incorrect estimations in the extent of expression of these genes. In addition, these gene products also are found only in parts of the neural tissues, so that identification of more general neural factors may be difficult. By comparison of in situ hybridization data of the cloned rhomboid gene with lacZ expression patterns of a P-lacW insert near this gene, Bier et al. (1990) showed differences in the extent of the expression. In their P-lacW strain, the blastoderm pattern and the transient expression of the rho gene in two other parts of the embryo is not mimicked by detectable lacZ expression. Thus gene products identified by enhancer traps may underscore the real extent of expression. So most of the genes described above are expressed only in parts of the neural tissues, or are restricted in their expression to certain development stages, or even to single cell types. In contrast, the 131 tubulin gene is expressed continually during the entire period of development of the CNS and PNS. Thus, the/31 tubulin gene offers an excellent system for the identification of cis regulatory sequences, in vivo and in vitro, as well as for the identification and isolation of trans-acting factors that are involved in the regulation of the expression of this tubulin isotype.
Methods
Generation and purification of isotype specific antibodies The 15 carboxy-terminal amino acids were synthesized on an Applied Biosystem peptide synthesizer 430 A according to Meierhofer et al. (1979) and purified by H P L C as reported previously (Leiss et al., 1988). The purified peptides (/31: D A E F E E E Q E A E V D E N ; /33: E F D P E V N Q E E V E G D C I ) were covalently linked to Keyhole Limpet Hemocyanin (KLH, Calbiochem.) according to Tamura et al. (1983). Antibodies were raised in rabbits and purified as outlined by Caroll and Scott (1985). Collected fractions were tested by ELISA and Western blotting.
Western blotting For Western blotting, protein extracts were prepared from the desired tissue or stages. The samples were separated on a 11% SDS polyacrylamide gel and were transferred to nitrocellulose (Towbin et al., 1979). Antibody reactions were essentially done as reported previously (Leiss et al., 1988). Antibody staining of embryos Embryos were dechorionated, permeabilized and fixed essentially as described recently (Leiss et al., 1988). After washing and blocking in BBT (0.15% crystalline BSA, 10 mM Tris-HC1, p H 7.5, 50 mM NaC1, 40 mM MgC1 z, 20 mM glucose, 50 mM sucrose, 0.1% Tween 20) they were incubated with the anti-/31 polyclonal (1 : 50) or the anti fl-galactosidase monoclonal-antibody (purchased from Promega) (1:1000) at 4°C overnight. The bound antibody was detected with a biotinylated secondary antibody and stained with the Vectastain ABC-kit (VectorLabs) using diaminobenzidine following the manufacturers instructions, except that 30 ~tl of 1% NiC12 were added per 600 /~1 staining mix (according to Lawrence et al., 1987). For the double-stains, embryos were incubated with both primary antibodies, than sequentially with the secondary antibody and using Ni-ions only during the first staining procedure. The/33 tubulin antibody was diluted 2500 fold and the /3-gal antibody 10000 fold. Individual embryos were mounted on slides in Epon and viewed under a Leitz microscope. All photographs (Kodak, Ektachrome 50) were taken under Nomarksi optics. Isolation of nuclei Embryos were staged and collected for appropriate times at 25°C and washed with 0.7% NaC1. They were homogenized in HPS (10 mM Hepes-KOH, pH 7.6, 10 mM KC1, 1.5 mM MgC12, 0.25 M sucrose, 0.5 mM DTE, 0.5 mM leupeptine) using a motor-driven glassteflon potter (4-6 strokes at 450 rpm). After removal of the cell debris by low-speed centrifugation (400 × g, 5 min, Heraeus Minifuge) the supernatant was placed on top of a cushion of HPS + 20% ( v / v ) glycerol, and the nuclei were pelleted at 5000 × g for 20 min. They were carefully resuspended in BC100 (20 mM Hepes-KOH, pH 7.6, 5 mM MgCI z, 0.2 mM EDTA, 100 mM KC1, 20% ( v / v ) glycerol, 0.5 mM DTE, 0.5 mM leupeptine, 0.5 mM PMSF), thereby leaving the yellowish yolk proteins behind. The nuclei were either used directly or quick frozen in liquid N 2 and stored at - 8 0 ° C up to several weeks without loss of activity. Transcriptional run-on assay Assay volume was 100 /~1, containing roughly 1 0 6 nuclei. The reaction conditions were 20 mM HepesKOH, pH 7.6, 75 mM KC1, 5 mM MgC12, 0.2 mM
117 EDTA, 0.5 m M DTE, 0.5 m M leupeptine, 0.5 m M P M S F and 5 U of RNasin (Amersham). The assay contained 1 m M of each ATP, CTP and GTP, 10 ~tM U T P and 50 /~Ci [ a - 3 2 p ] U T P (600 C i / m m o l , Amersham). Incubation was performed at 25°C for 2 h. The reaction was terminated by the addition of E D T A at a final concentration of 10 m M and subjected to p h e n o l / c h l o r o f o r m extraction. The synthesized R N A s were precipitated and directly dissolved in the hybridization buffer (50% formamide, 1 x Denhardts, 2 x SSC, 100 t~g/ml herring sperm DNA). D N A s were immobilized to gene-screen using a slot-blot apparatus (Minifold II, Schleicher and Schuell). Hybridization was done in 50% formamide at 42°C overnight. The filter was washed at high stringency (0.1 X SSC, 0.5% SDS, 65°C), dried and exposed overnight using a K o d a k X R film.
Preparation of nuclear extracts Nuclei were isolated as described above. The pellet was resuspended in 1 vol. nuclear extraction buffer (20 m M H e p e s - K O H , p H 7.6, 20% ( v / v ) glycerol, 420 m M NaC1, 1.5 m M MgC12, 0.2 m M EDTA, 0.5 m M PMSF, 0.5 m M DTE, 0.5 m M leupeptine) according to Dignam et al. (1983), and stirred for 30 rain at 0°C. The suspension was centrifuged for 20 min at 25000 x g. The supernatant was dialysed against 50 vol of BC 100 with two changes for 4 - 6 h, and aliquots were quick-frozen in liquid N 2. These extracts could be stored at - 8 0 ° C without loss of activity for several months. In vitro transcription The in vitro transcriptions were performed as described by Heiermann and Pongs 1985. Assay volume was 25/tl. Buffer conditions were 20 m M H e p e s - K O H , p H 7.6, 75 m M KCI, 5 m M MgCI 2, 0.2 m M EDTA, 12% ( v / v ) glycerol, 0.5 m M DTE, 0.5 m M PMSF. Template concentration was 0.5 pmol per assay. Reactions were carried out with 15 ~1 extract (corresponding to 100/~g of protein). Nucleotide concentration was 0.6 m M of each ATP, CTP, GTP, UTP, and the incubation time was 1 h at 25°C. Reactions were terminated by addition of a stop-mix (400 m M sodium acetate, 0.5% sarcosyl, 5 0 / ~ g / m l tRNA); nucleic acids were extracted with phenol/chloroform, precipitated and dissolved in 1 M sodium acetate, 0.1% sarcosyl. For S1 analysis, end-labelled fragments (see legend to Fig. 3) were added, the nucleic acids precipitated with E t O H and dissolved directly in 20 /~1 Sl-hybridization buffer (80% formamide, 40 m M Pipes-KOH, p H 6.8, 400 m M NaC1). They were heated for 30 min to 85°C and hybridized overnight at 49°C. 200 btl ice-cold S1-Mix (30 m M sodium acetate, p H 4.5,280 m M NaC1, 4.5 m M ZnSO 4, 30 /zg/ml herring sperm DNA, 100 U S1/assay) was added and the reaction incubated for 0.5 h at room temperature. They were extracted with p h e n o l / chloroform, precipitated, the pellet dissolved in 3 /~1 8
M urea and electrophorized on 5% PAA urea gels. The gels were fixed, transferred onto W h a t m a n n 3MM paper, dried and autoradiographed.
Construction of lacZ fusion genes Fusion between the lacZ reporter gene and the fll gene was performed as follows. The Sau3A site near the codon for amino acid 25 was used for in frame fusion by means of a BamHI site in p M C 1871 (Gasch et al., 1989). Sequences downstream of a BssHII site in the lacZ gene were replaced by the equivalent sequences of p C H l l 0 , thereby adding the SV40 polyadenylation signal. Genomic sequences from - 2 . 2 kb to + 2.8 kb in the intron were fused via a ScaI site, resulting in the construct WfllL. For WfllK, 2 kb were deleted from the intron by use of internal restriction sites (Avail and PstI), leaving 273 bp of 5' and 315 bp of 3' sequences from the intron. WfllC was constructed using a c D N A clone. Constructs in p W H L (Michiels, F., Hinz, U. and Renkawitz-Pohl, R., unpublished data), namely f l l W H I and f l l W H I I were made by linearizing the vector with StuI and inserting the intron cut with Asp700 at +0.1 kb (amino acid 4) and Bali at +2.7 kb (0.15 kb upstream of the second exon) in both orientations upstream of a truncated hsp70 promoter. P-element transformations P-element transformations were performed as published previously (Michiels et al., 1989).
Acknowledgements We thank Professor Campos-Ortega for helpful discussions, Sabina Kerl for excellent technical assistance, Professor O. Pongs for the gift of the histone gene clones, and Cathy Schindewolf for critical reading of the manuscript. Professor H. Holl~inder made his microscope facilities availaible for us. We acknowledge the financial support of the Deutsche Forschungsgemeinschaft (Re 628/3-1 and SFB 190) and the Bundesministerium fiir Forschung und Technologie.
References Anderson, K.V and Lengyel, I.A. (1980) Cell, 21,717-721. Basler, K., Siegrist, P. and Hafen, E. (1989) EMBO J., 8, 2381-2386. Bialojan, S., Falkenburg, D. and Renkawitz-Pohl, R. (1984) EMBO J., 3, 2543-2548. Bier, E., Vaessin, H., Shepherd, S., Lee, K., McCall, K., Barbel, S., Ackermann, L., Carretto, R., Uemura, T., Grell, E., Jan, L.Y. and Jan, Y.N. (1989) Genes Dev., 3, 1273-1287. Bier E., Jan, Y.J. and Jan, Y.N. (1990) Genes Dev., 4, 190-203. Biggin, M.D. and Tjian, R. (1988) Cell, 53, 699-715. Bodmer, R., Barbel, S., Shepherd, S., Jack, J.W., Jan, L.Y. and Jan, Y.N. (1987) Cell, 51, 293-307.
118 Bray, S.J., Burke, B., Brown, N.H. and Hirsh, J. (1989) Genes Dev., 3, 1130-1145. Campos-Ortega, J.A. and Hartenstein, V. (1985) Springer Verlag, Berlin and Heidelberg. Caroll, S.B. and Scott, M.P. (1985) Cell, 43, 47-57. Candy, M., Grell, E.H., Daubly-Chaudiere, C., Ghysen, A., Jan, L.Y. and Jan, Y.N. (1988) Genes Dev., 2, 843-852. Cleveland, D.W. (1987) J. Cell Biol., 104, 381-383. Dambly-Chaudiere, C. and Ghysen, A. (1987) Genes Dev., 1,297-306. Dignam, J.D., Lebovitz, R.M. and Roeder, R.G. (1983) Nucleic Acids Res., 11, 1475-1489. Doe, C.Q., Hiromi, Y., Gehring, W.J. and Goodman, C. (1988) Science, 239, 170-175. Dynlacht, B.D., Attardi, L.D., Admon, A., Freeman, M. and Tjian, R. (1989) Genes Dev., 3, 1677-1688. Foe, V.E. (1989) Development, 107, 1-22. Fujita, S.C., Zipursky, S.L., Benzer, S., Ferrus, A. and Shotwell, S.L. (1982) Proc. Natl. Acad. Sci. USA, 79, 7929-7933. Garcia-Bellido, A. (1979) Genetics, 91,491-520. Gasch, A., Hinz, U., Leiss, D. and Renkawitz-Pohl, R. (1988) Mol. Gen. Gen., 21, 8-16. Gasch, A., Hinz, U. and Renkawitz-Pohl, R. (1989) Proc. Natl. Acad. Sci. USA, 86, 3215-3218. Goldberg, M.L. (1979) Ph.D. Thesis, Stanford University. Heberlein, U. and Tjian, R. (1988) Nature, 331,410-415. Heiermann, R. and Pongs, O. (1985) Nucleic Acids Res., 13, 27092730. Jan, L.Y. and Jan, Y.N. (1982) Proc. Natl. Acad. Sci. USA, 72, 2700-2704. Jimenez, F. and Campos-Ortega, J.A. (1979) Nature, 282, 310-312. Johansen, K.M., Fehon, R.G. and Artavanis-Tsakonas, S. (1989) J. Cell. Biol., 109, 2427-2440. Karr, T.L. and Alberts, B.M. (1986) J. Cell Biol., 102, 1494-1509. Kemphues, K.J., Kaufmann, T.C., Raft, R.A. and Raft, E.C. (1982) Cell, 31, 655-670. Klemenz, R., Weber, U. and Gehring, W.J. (1987) Nucleic Acids Res., 15, 3947-3959.
Knust, E., Tietze, K. and Campos-Ortega, J. (1987) EMBO J., 6, 4113-4123. Knust, E. and Campos-Ortega, J.A. (1989) BioEssays, 11, 95-99. Lawrence, P.A., Johnston, P., McDonald, P. and Struhl, G. (1987) Nature, 328, 440-442. Lehmann, R., Dietrich, U., Jiminez, F. and Campos-Ortega, J. (1981) Roux's Arch., 190, 226-229. Leiss, D., Hinz, U., Gasch, A., Mertz, R. and Renkawitz-Pohl, R. (1988) Development, 104, 525-531. Little, M. and Seehaus, T. (1988) Comp. Biochem. Physiol., 90B, 655-670. Meierhofer, J., Waki, M., Heimer, E.P., Lambros, I.T., Makoske, R.C. and Chang, C.D. (1979) Int. J. Pept. Protein Res., 13, 35-47. Michiels, F., Falkenburg, D., Miiller, A.M., Hinz, U., Otto, U., Bellmann, R., Gl~itzer, K.H., Brand, R., Bialojan, S. and Renkawitz-Pohl, R. (1987) Chromosoma, 95, 387-395. Michiels, F., Gasch, A., Kaltschmidt, B. and Renkawitz-Pohl, R. (1989) EMBO J., 8, 1559-1565. Natzle, J.E. and McCarthy, B.J. (1984) Dev. Biol., 104, 187-198. Raft, E.C. (1984) J. Cell. Biol., 99, 1-10. Roberts, K. and Hyams, J.S. (1979) Academic Press, London. Rothberg, J.M., Hartley, D.A., Walther, Z. and Artavanis-Tsakonas, S. (1988) Cell, 55, 1047-1059. Rudolph, J.E., Kimble, M., Hoyle, H.D., Subler, M.A. and Raft, E.C. (1987) Mol. Cell. Biol., 7, 2231-2242. Soeller, W.C., Poole, S.J. and Kornberg, T. (1988) Genes Dev., 2, 68-81. Sullivan, K.F. (1988) Annu. Rev. Cell Biol., 4, 687-716. Tamura, T., Bauer, H., Birr, Ch. and Pipkorn, R. (1983) Cell, 34, 587-596. Thomas, J.B., Crews, S.T. and Goodman, C.S. (1988) Cell, 52, 133-141. Towbin, H., Staehelin, T. and Gordon, J. (1979) Proc. Natl. Acad. Sci. USA, 76, 4350-4354. Vassin, H., Bremer, K.A., Knust, E. and Campos-Ortega, J. (1987) EMBO J., 6, 3431-3440. Wilson, C., Pearson, R.K., Bellen, H.J., O'Kane, C.J., Grossniklaus, U. and Gehring, W.J. (1989) Genes Dev., 3, 1301-1313.
107
M O D 00009
During Drosophila embryogenesis the B1 tubulin gene is specifically expressed in the nervous system and the apodemes Detlev Buttgereit, Dagmar
L e i s s *, F r i t s M i c h i e l s a n d R e n a t e R e n k a w i t z - P o h l
Gentechnologische A rbeitsgruppe am M P I ]'fir Biochemie, Martinsried / Miinchen, F. R. G. (Accepted 3 September 1990)
We determined the in vivo distribution of the fll tubulin from D. melanogaster using isotype specific antibodies. Maternally expressed/~1 tubulin is incorporated into mitotic spindles. Later in development a strong expression in the CNS is observed. Furthermore, all chordotonal organs and the apodemes are marked by fll tubulin. Nuclear run-on assays and stage specific in vitro transcription showed a zygotic expression of the/31 tubulin gene from the extended germ-band stage onwards. Using the P-element system, we identified several elements; upstream between - 2 . 2 kb and the transcription initiation site, elements for low level expression in the CNS are present. In the intron between + 0.44 kb and + 2.5 kb enhancer elements are located that drive the expression in the chordotonal organs and the apodemes. Between the start site and +0.44 kb (273 bp) and +2.5 kb and the second exon (315 bp), maternal and CNS enhancers result in full level expression of a lacZ-fll reporter gene. We show, that the fll tubulin gene is a very early effector gene starting its expression shortly after the commitment of neuroblast cell fate. This gene offers an excellent model system for the identification of neural and apodeme specific transcription factors. /31 Tubulin; Regulatory element; Nervous system; Apodeme; Drosophila
Introduction Microtubules form major structural components in all eukaryotic cells. They participate in structures such as mitotic and meiotic spindles, flagellae, ciliae, the cytoskeleton and neural processes (for review see Roberts and Hyams, 1979). In the nervous system, microtubules are rather abundant. In all multicellular organisms investigated so far, the a and /3 tubulins are encoded by multi-gene families of varying extent (for reviews see Cleveland, 1987; Little and Seehaus, 1988; Sullivan, 1988). In Drosophila, four tubulin genes have been identified with distinct developmental expression (Bialojan et al., 1984; Natzle and McCarthy, 1984). Studies concerning the/32 tubulin during spermatogenesis (Michiels et al., 1987, Kemphues et al., 1982, Raft,
Correspondence to: R. Renkawitz-Pohl, Gentechnologische Arbeitsgruppe am MPI fi.ir Biochemie, A m Klopferspitz 18A, 8033 Martinsried/Miinchen, F.R.G. * Present address: EMBL, 6900 Heidelberg, F.R.G.
1984) and the f13 tubulin gene during embryogenesis (Rudolph et al., 1987; Gasch et al., 1988, Leiss et al., 1988) revealed a tissue and development specific expression of these protein variants. These investigations showed, that during embryogenesis the f13 tubulin is restricted to the mesoderm and its derivatives (Leiss et al., 1988), while the/32 tubulin is solely expressed in the developing testes (Kemphues et al., 1982). Thus, all other essential functions such as spindle formation and microtubules of the nervous system require another 13 tubulin isotype. In Drosophila, a cascade of genes regulates the formation of the central nervous system (CNS) and peripheral nervous system (PNS). Neurogenic loci like Notch (Lehmann et al., 1981; Johansen et al., 1989), Delta (V~issin et al., 1987) or Enhancer of split (Knust et al., 1987) participate in the decision of an ectodermal cell to develop as a neuroblast or epidermoblast. Genes as daughterless (Caudy et al., 1988) and the achaete-scute complex (Garcia-Bellido, 1979) are required for the proper development of CNS (Jimenez and CamposOrtega, 1979) or PNS (Dambly-Chaudiere and Ghysen, 1987) cells. Further downstream, active genes regulate
108 aspects like neuron formation, axone outgrowth, cell identity, e.g., slit (Rothberg et al., 1988), single-minded (Thomas et al., 1988), cut (Bodmer et al., 1987) or fushi tarazu (Doe et al., 1988). During the maturation from the egg to the adult fly, tubulins are needed in large amounts for the structures mentioned above. In this report the distribution of the fll tubulin during embryonic development will be shown. During the first 90 min of embryonic development, all R N A has to be supplied to the egg by maternally synthesized message, because no transcription is observed during this period. For these early stages in situ hybridizations to embryo sections revealed only transcripts of the 131 tubulin gene, which are uniformly distributed throughout the embryo (Gasch et al., 1988). Later, from stage 10 onwards, a concentration of the message is observed in the developing central nervous system and the brain. In order to identify the mechanism of this accumulation, and to decide in which microtubules fll tubulin is incorporated, several independent approaches were chosen. The experiments described in this paper characterize the distribution of the fll tubulin analysed by isotype specific antibodies, and determine the onset of zygotic transcription by run-on assays and in vitro transcription. Furthermore we studied the tissue specificity by P-element mediated transformation with 131-1acZ fusion genes. These analyses revealed that the transcription of the 131 tubulin gene starts shortly after the action of genes which participate in the decision of epidermal versus neural celt fate (for review see Knust and Campos-Ortega, 1989). As we will point out, the expression of the fll tubulin gene is regulated by independent, cell-typespecific enhancers in the intron for the PNS and apodemes, while maternal and CNS expression is directed by cooperation between upstream promoter elements and corresponding enhancers in the intron.
Results
Determination of the wild-type 131 tubulin pattern As a prerequisite for further studies, the distribution of the 131 tubulin in the developing embryo had to be determined. We have shown previously, that the Drosophila 13 tubulin isotypes mainly differ in the last 15 amino acids (Michiels et al., 1987; Rudolph et al., 1987). Therefore, polyclonal antibodies against synthetic peptides corresponding to the C-terminal amino acids of the fil tubulin were raised in rabbits. Their specificity was tested both by competition with specific and unspecific peptides in ELISA assays (data not shown) as well as by analysis of proteins extracted from different embryonic, larval and pupal stages and adults on Western blots. As Fig. 1A shows, a roughly constant amount of fll tubulin is detected in every stage ex-
amined. As a control, the 133 tubulin antibody (Leiss et al., 1988) was used in parallel on a second filter. Only in 8 to 16 h embryos, mid and late pupae and adult testes is 133 tubulin detectable as has been published previously (Leiss et al., 1988). Having confirmed the isotype specificity of the antibody, the distribution of the 131 tubulin was analysed in whole mount stains of embryos from D. melanogaster. Embryos were collected, fixed and stained utilizing taxol to stabilize microtubules as described in Methods. Freshly laid eggs and very early embryos showed an uniform distribution of the 131 tubulin (not shown). As soon as mitotic spindles are visible, the 131 tubulin is concentrated in these structures (Plate IA). In accordance with the known synchronism of the first mitotic divisions (Karr and Alberts, 1986; Foe, 1989), spindles are visible at the same time over the whole embryo. Around stage 8 (Plate 1B) the cytoplasmic staining is still uniform except in those regions that correspond to mitotic domains as determined by Foe (1989). She defines a mitotic domain as a region of the embryo where all cells divide synchronously. For example, Plate 1B shows in the posterior part of the embryo nuclear localization in a region corresponding to mitotic domain 8144. During the next hours of embryonic development, the distribution of the 131 tubulin is uniform and both spindle and cytoplasmic microtubules are decorated by the 131 antibody. At stage 16 when the condensation of the CNS is nearly completed and the PNS is visible, a high concentration of the /31 tubulin is detected in the CNS and the chordotonal organs (Plate 1D). Furthermore, in agreement with the in situ hybridizations published previously (Gasch et al., 1988), a label over the apodemes is found (not shown; due to the strong residual staining of maternal protein, these structures cannot be reproduced photographically).
Zygotic transcription of the fll tubulin gene starts at stage 10 The observed fll tubulin distribution in the embryo led to the question, whether the concentration in the CNS is due to the higher number and density of cells or if it is derived from a tissue-specific zygotic expression of the 131 tubulin gene. A concentration of the /31 message in the CNS is observed round stage 11 (Fig. 2D), thus far earlier than detected with the 131 specific antibody by looking at the protein. Concomitantly, the maternal 131 message in the other parts of the embryo is rapidly degraded, while the protein is still present in large amounts. To determine whether this effect is based on newly synthesized RNA, nuclear run-on assays were performed. Nuclei were isolated from staged embryos and examined for R N A synthesis (see Methods). As Fig. 2B shows, no detectable level of fll message is found in embryos of less than 6 h of development, while a strong signal is seen in the histone H 3 / H 4 control at the same stage. In contrast, 6 to 12 h old embryos yield
109
1
2
3
4
5
6
1
2
3
4
5
6
7
8
9
10
11
12
8
9
10
11
12
13
A
~ii i~iiiiii~ii~!! !I .....
7 ~
13
•
Fig. 1. Distribution of/31 and /33 tubulin lsotypes during Drosophila development. A developmental Western-blot was performed using 60 ~g of total protein per lane. (1) ovaries; (2) embryos 0 h; (3) embryos 8-16 h; (4) embryos > 19 h; (5) first instar larvae; (6) pupae 0 h; (7) pupae 26 h; (8) pupae 49 h; (9) pupae 72 h; (10) pupae 96 h; (11) testes; (12) males; (13) females. (A) Specific reaction with the/31 antibody; this tubulin isotype is detectable in all developmental stages in roughly equal amounts. (B) Specific reaction with the f13 antibody; this tubulin isotype is detectable only during mid-embryogenesis(8-16 h), in pupae (26 h, 49 h, 72 h) and adult testes.
a clearly visible /31 signal. These experiments showed that no zygotic expression of the /31 tubulin gene is observed during early embryogenesis and that detectable levels of transcription are not found before 6 h of embryonic development. To obtain further evidence we confirmed these resuits by in vitro transcription assays of the/31 tubulin gene in nuclear extracts from staged embryos. The template contained /31 sequences from - 2 . 2 kb to + 0.44 kb thus comprising the first exon as well as 250 bases of the intron. Synthesized R N A was analysed by S1 mapping as shown in Fig. 2C. In accordance to the data from the run-on assays, no transcription of the/31 template (Fig. 2A) can be achieved in extracts from embryos staged 0 - 6 h. If the extracts were prepared from 6-12 h embryos, a strong correctly initiated /31 transcript is observed. As a control for transcriptional activity of the extracts, the histone gene promoters for the H3 and H4 genes were chosen (Goldberg, 1979). Fig. 2C shows that a high level of transcription is found in the early extracts in agreement with the in vivo transcription of histone genes (Anderson and Lengyel, 1980). Taken together, these results present clear evi-
dence that the zygotic transcription of the /31 tubulin gene does not start before 6 h of embryonic development, which parallels the observed concentration of R N A and protein in the developing CNS.
- 2 . 2 kb of upstream sequences and the intron are sufficient to achieve the complete expression pattern F r o m the analysis of fll m R N A and /31 tubulin isotype distributions, as well as from the run-on and in vitro transcription assays, we strongly favored the hypothesis that all uniformly distributed /31 tubulin is maternal in origin and that the zygotic expression is restricted to the nervous system. In order to identify cis-acting elements in vivo, we used the P-element transformation system. Eggs were injected with constructs derived from the vector pW8 (Klemenz et al., 1987) which contained the bacterial /3-galactosidase gene as a reporter and varying parts of the/31 gene (see Fig. 3 and Methods). In the fll tubulin gene, a large intron between the codons for amino acids 19 and 20 is present (Michiels et al., 1987). In analogy to the regulatory function of the intron of the mesoderm specifically expressed /33 tubulin gene, we sus-
N
; !
ili~:il~! ;i: :i~
i'L
~i'~i i~ii~i!~i~,i~i~i~i~!
111
A
-2.2 kb
+1
+ 0.44 kb Template
pUC 19
fll
pUC 19 Sl-probe
B
C
NE
n
0-6h +
0-6h
NE
NE
0-6h +
6-12h +
6-9h
D ~-A
J
B1 pUCI9
m
"-81
B3
Act5C H3/H4
iliiiiiiill i¸~¸
i'~i~i!iii
~i!ii!iil;~,,,;~ !i~
"~'i
Fig. 2. Determination of the zygotic transcription of the fll tubulin gene. (A) Template for the in vitro transcription. (B) Transcriptional run-on assay using isolated nuclei. No transcription of the fll tubulin gene is detected in the nuclei from embryos staged 0-6 h, while a clear signal can be seen using embryos staged 6-9 h. (C) In vitro transcription assay. The template (see A) used comprised 2.2 kb of upstream and 0.44 kb of downstream sequences (open box = leader; filled box = coding region; thick bar = intron). Only extracts from embryos staged 6-12 h efficiently transcribe the 131 template. (NE = nuclear extract; a-A = a-amanitine at 0.5/~g/ml). (D) In situ hybridization of embryo sections with a ill-specific probe. (1) Homogeneous distribution of the message in a stage 8 embryo. (2) Concentration of the message in the developing CNS (arrow) in a stage 11 embryo.
pected that within the t u b u l i n gene family, the functional i m p o r t a n c e of the i n t r o n m a y have b e e n conserved. T o look for potential regulatory elements in this region, three different constructs were m a d e which contain either the complete i n t r o n ( = WfllL), 273 b p of 5 ' a n d 315 b p of 3' sequences from the i n t r o n ( = W f l l K ) or are lacking the i n t r o n ( = WfllC). All constructs h a d 2.2 kb of the 131 u p s t r e a m sequences. F o r each construct, expression of the i n d i c a t o r gene was m o n i t o r e d b y i m m u n o s t a i n i n g of e m b r y o s using a m o n o c l o n a l /3-gal antibody. T h e p a t t e r n was c o m p a r e d to the wild-
type d i s t r i b u t i o n of the fll t u b u l i n . The results are s u m m a r i z e d i n Fig. 3. Strains t r a n s f o r m e d with W f l l L show a d i s t r i b u t i o n of the i n d i c a t o r gene p r o d u c t indist i n g u i s h a b l e f r o m the fll t u b u l i n . As was to be expected, n o u p t a k e of the fusion p r o t e i n into spindles was observed since o n l y the first 23 a m i n o acids of the t u b u l i n are c o n t a i n e d in these constructs. I n unfertilized eggs as well as i n t h e early stages of embryogenesis, fl-galactosidase is p r e s e n t i n all cells. I n agreement with the wild type situation, a c o n c e n t r a t i o n of the fll-gal fusion p r o d u c t i n the C N S is observed in later stages of
Plate 1. Determination of the in vivo distribution of the fll tubulin by isotype-specific antibodies. (A) Blastoderm embryo showing the mitotic spindles (arrow). (B) Stage 8 embryo; the distribution of the fll tubulin is nearly uniform except for a nuclear localization in mitotic domain 8144 (arrow). (C) Stage 13 embryo showing an uniform labelling by the antibody. (D) Stage 17 embryo; the ventral cord (vc) and the supra-oesophageal ganglion (spg) show an intense staining. Plate 2. Zygotic expression of the transformed fll-lacZ fusion genes is prominent in the developing CNS. Crosses including the maternal (A, B) or without the maternal expression (C, D) were performed. Without maternal expression, solely the CNS shows a strong label by the lacZ antibody. Comparison of a blastoderm (A, C) and a stage 13 (B, D) embryo are shown.
112
Plate 3.
Plate 4.
113
Expression number of strains -2.2 kb
WB1L
CNS
2
+++
+++
6
++
++
11
+/"
partial
chordotonal organs
apodemes
2.6 kb
I
C ,,~" !
pW8
[31
-2.2 kb
WfllK
maternal expression
iacZ
+++
+++
pW8
0.6 kb
I
[]
pW$
I
131
lacZ
pW$
-2.2 kb
WB1C
,
I pW$
~1
lacZ
.
.
.
.
pW8
Fig. 3. Expression patterns of strains stably transformed with different/31-1acZ fusion constructs. Adjacent to the constructs the expression patterns are outlined. The strains transformed with W/31L show the complete pattern as the/31 tubulin. In W/31K, maternal expression is reduced, and no expression could be detected in chordotonal organs and the apodemes. Maternal expression of the W/31C strains is barely detectable. The expression in the CNS in W/31C strains is not uniform as with the other constructs, but restricted to a number of cells.
embryogenesis. In addition, the chordotonal organs and the apodemes are marked by the expression of the reporter gene as is shown in detail in Plates 2 and 3. A similar pattern emerges from the analysis of the strains transformed with WfllK. As in WfllL strains, the transformed reporter gene is maternally expressed, which is also demonstrated by histochemical staining of ovaries (not shown). Later in embryonic development, the concentration in the CNS is visible as in WfllL. As a most striking difference to WfllL no expression of the reporter gene is seen in the chordotonal organs and the apodemes. Thus, cis-acting elements responsible for the expression in these tissues must reside between + 0.44 and +2.5 kb of the intron; however, both constructs contain all control elements guiding the tissue specific expression in the ovary as well as in the CNS. In
addition, the level of fl-galactosidase is much higher in WfllL than in WfllK strains, indicating a quantitative effect of the deleted parts of the intron on the level of expression. The importance of the intron was further augmented by the analysis of the strains transformed with WfllC. As a major difference, the amount of maternally supplied protein is reduced to a very low level, as freshly laid eggs show only a very faint staining. In addition, the expression in the CNS is detected at the correct time, but at a reduced level. Comparison with WfllL or WfllK shows that the distribution is not so uniform, with a few neuroblasts having a much higher level of the fl-gal than the rest (not shown). Taken together, these data lead to the following conclusions: for a subset of neural cells, all sequences for time and tissue specific expression at low level are present in
Hate 3. Zygotic expression of/31-1acZ fusion genes in W/31L strains. (A) Embryo stage 10; a label is seen in the cephalic region, and in addition single neuroblasts (nb) in the CNS region are stained. (B) Late stage 10 embryo. (C) Embryo stage 14: besides the the ventral cord (vc), precursors of the PNS are labelled (arrows). (D) Stage 17 embryo; the supra-oesophageal ganglion (spg), the ventral cord, the chordotonal organs (co) and the apodemes (ap) are decorated by the antibody. Plate 4. Expression of the reporter gene in the fllWHI strains. (A) Lateral view at a late stage 12 embryo; groups of cells are labelled in the parts of the embryo where the PNS will form. (B) Stage 17 embryo, lateral view; a strong label is seen over the chordotonal organs (co), over axones and the apodemes (ap). In addition, cells are marked in a position (arrow) were the pleural lateral muscles (pl) insert into the epidermis (Campos-Ortega and Hartenstein, 1985). (C) Ventral view at a stage 17 embryo; the visible apodemes iapo, inal and ina2 are marked by arrows. (D) Lateral view on a embryo stained for/33 tubulin. (E, F) Double-staining of embryos with a/33 tubulin specific antibody and the lacZ antibody. /33 stained muscles are light-brown, while the lacZ positive cells are black-brown. Muscles insert at the position of the apodemes iapo (thin arrow), inal and ina2 (thick arrows).
114 the 5' flanking 2.2 kb; for the expression in the complete CNS, elements in the intron included in WfllK (273 bp of 5' and 315 bp of 3' sequences) are necessary. With the additional intron sequences contained in the W/31L constructs, further regulatory units result in a high level of expression, guide the transcription in the chordotonal organs and the apodemes and resemble enhancer elements. The zygotic expression of the fll tubulin gene is restricted to the neroous system and the apodemes The transformed strains offered the possibility to separate the maternal and zygotic expression of the indicator gene under the control of the fll regulatory sequences. Males carrying the W/31L construct were mated to females from the recipient white strain so that the eggs deposited by the females had no maternally supplied/3-gal. As a result, the zygotic expression mediated by the paternally inherited indicator gene can be followed independently as is shown in Plate 2. For comparison embryos from reporter gene positive females are shown. The eggs deposited from the white females show no detectable level of/3-gal. After stage 10, when the development of the CNS proceeds, a strong label over neural tissues and the brain can be seen (Plate 2D). So definitively the neural derivatives are the main tissues of the embryo that zygotically express the reporter gene under the control of the/31 tubulin gene regulatory sequences. This tubulin isotype represents a good marker to follow the development of the CNS and PNS. In addition, as will be shown below, the anchoring points of the muscles to the epidermis, the apodemes, are characterized by 131 tubulin expression. Development of the CNS followed by reporter gene expression In order to determine precisely the developmental stage of first /31-/3gal expression, several crosses using W/31L as described above were performed. Eggs were collected and staged for 0 to 22 h at 25 ° C. At stage 10, when epidermoblast and neuroblast are determined in their cell fate (for review see Knust and Campos-Ortega, 1989), we first observe /3-gal staining. As soon as the first neuroblasts divide, a label over the neuroectoderm can be seen (Plate 3A). The intensity of antibody staining increases very rapidly corresponding to the increasing number of cells in the CNS region (Plate 3B). When germ band retraction has further proceeded, groups of cells are labelled in a region where the PNS will form (Plate 3C). During further development of the embryo, the differentiating neural structures are permanently marked by the expression of the /31-1acZ fusion gene. Structures like axones, ventral cord and chordotonal organs can be visualized in the embryo. Shortly before hatching of the larvae the condensed CNS, the brain hemispheres, the PNS and the apodemes are stained
(Plate 3D). An identical result was obtained by staining of embryos from W/31K strains, except that as was stated above, no expression is detected in the chordotonal organs and the apodemes. For the developing and mature nervous system, microtubules are essential molecules. It represents the tissue with the highest rate of mitotic division from stage 10 onwards and the spindle microtubules contain, in our judgement, exclusively/31 tubulin. The analysis of zygotic/3-gal expression revealed that the/31 gene is a very early effector gene beginning its expression shortly after the commitment of ectodermal cells in epidermoblast and neuroblast. Cell-type specific enhancers are located in the intron In order to address the question whether independently acting elements are situated in the intron of the /31 tubulin gene, a standard enhancer test in the context of a heterologous promoter was performed. We used a truncated hsp70 promoter, cloned a 2.5 kb fragment of the intron (thereby deleting 0.15 kb of the 3' part of the intron) upstream in both orientations /31WHI and fllWHII), and established P-element transformed fly strains. As no difference could be seen in the expression pattern between the two orientations of the intron (5 lines analysed), only the construct (/31WHI is shown. Staining of the embryos revealed that only the apodemes, the chordotonal organs and axones are marked by the expression of the reporter gene (Plate 4B). Development of the chordotonal organs can be followed from stage 12 onwards. In addition, the formation of the apodemes is seen round the same stage, in accordance with the beginning formation of somatic muscles (Leiss et al., 1988). All three types of apodemes (iapo, inal, ina2; Campos-Ortega and Hartenstein, 1985) can be seen (Plate 4C). In order to unambiguously identify the apodemes, double immunostaining using the 133 tubulin antibody to mark the somatic muscles and the /3-gal antibody to show the apodemes in the transformants was performed (see Methods). As Plates 4E and 4F clearly demonstrate, the/33 stained muscles insert at the position were the /3-gal labelled cells are located in the epidermis. So these structures represent the anchoring points of the muscles to the epidermis, the apodemes. As reference, an embryo stained only with the 133 tubulin antibody is shown in Plate 4D. In contrast to the results obtained with the/31 tubulingene promoter, in the/31WHI and/31WHII transformants neither maternal expression nor expression in the CNS is detected. So the elements in the intron that regulate these modes of expression have to cooperate with elements of the/31 promoter, or reside in the 0.15 kb of fll intron sequences not analysed yet. Additional autonomously acting cell-type specific enhancers are found in the intron that are capable of activating a heterologous promoter independent of their orientation,
115 thus resembling classical enhancers. They drive the expression in the PNS and in the apodemes. The next steps in analysing these different elements will be the further subdivision of the intron in order to characterize the enhancers for PNS and apodemes, and the identification of upstream and intron elements for the maternal and CNS expression.
Discussion We have started the analysis of expression of the fll tubulin gene. As a first step, the in vivo distribution using isotye specific antibodies was determined. Staining of whole mount embryos revealed in early stages a uniform distribution of maternally synthesized protein in accordance with the R N A in situ localization published previously (Gasch et al., 1988). When the initial mitotic divisions take place, the fll tubulin is found in the spindles. During the development of the embryo, spindles can also be observed in the mitotic centers as described (Foe, 1989). In later stages, a concentration of the fll tubulin in the CNS and the brain hemispheres is observed. In addition, the anchoring points of the musculature to the epidermis, the apodemes, are marked by fll tubulin expression. Thus, this tubulin isotype represents a ubiquitous, multifunctional tubulin during embryogenesis. In order to evaluate whether all fll tubulin is due to maternal synthesis or if zygotic expression of the fll tubulin takes place, run-on assays were performed. They revealed a transcription of the gene in embryos older than 6 h. Similar results were obtained from in vitro transcription experiments using a fll template containing sequences between - 2 . 2 kb and +0.44 kb and nuclear extracts from staged embryos. So this assay system will help us to identify and purify some of the factors that participate in the regulation of the fll tubulin gene. Heberlein and Tjian (1988) showed a correct transcription initiation for the alcohol dehydrogenase gene in staged embryo extracts, and isolated a factor that binds to regulatory sequences of this gene. Similar experiments showed the presence of several upstream and downstream regulatory sequences in the promoter region of the Ubx gene (Biggin and Tjian, 1988) and the engrailed gene (Soeller et al., 1988). These data showed that the in vitro systems allow at least partially the identification of regulatory sequences, which are also active and necessary in vivo, as well as the identification and the isolation of trans-acting factors. To localize the domains of zygotic fll tubulin expression, P-element transformed fly strains using different constructs were established. We could show that - 2 . 2 kb of upstream sequences and the presence of the 2.6-kb long intron are necessary for maximum level of
maternal and zygotic expression of a fll-lacZ reporter gene. In accordance with the distribution of the fll tubulin, the ventral cord, the chordotonal organs, the brain and the apodemes are the domains of expression. Deletion of the intron nearly completely abolishes the maternal expression, and the expression in the CNS is restricted to a subset of cells. In the PNS, only few cells are labelled. That means that important features of the regulation of the fll gene expression are mediated by elements in the intron. In contrast, in the context of a heterologous promoter, the intron drives only the expression in the PNS and the apodemes. These results clearly show, that at least two different types of enhancer elements are contained within the intron: one type, which can confer cell type specifity autonomously and is able to cooperate with a heterologous promoter, and a second type that requires upstream elements of the fll tubulin gene p r o m o t e r to enhance or activate expression at the correct time and tissue. Whether these modes of regulation are controlled by completely unrelated elements or by specific combinatorial interactions between related elements has to be clarified. Recently, a number of reports for regulation mediated by enhancers in intron regions has been found in Drosophila. Basler et al. (1989) showed that regulatory elements for the transcription of the sevenless gene reside in the intron. Within the fl tubulin gene family, the trait of a large intron is conserved between the fll tubulin gene (2.6 kb intron) and the r 3 tubulin gene (4.5 kb intron). In both genes, this intron is found between codons for amino acids 19 and 20 (Michiels et al., 1987, Rudolph et al., 1987). It was shown that regulatory elements in the intron region of the r 3 tubulin gene are necessary for the expression of the tubulin in the visceral mesoderm (Gasch et al., 1989). The zygotic fll expression, as it is shown in this report, is restricted in the embryo to neural tissues and the apodemes. So the regulatory elements in the introns of these genes are quite different. The fll tubulin gene is certainly not a part of the regulatory cascade driving the formation of the nervous system. But as we could show, it is zygotically expressed in all neural structures. In addition, this expression starts as early as the first neuroblasts in the CNS and the first cells in the PNS region can be identified. That means, that the fll tubulin is one of the first effector genes switched on after the determination of cell fate. So this gene must be under the direct or indirect control of genes that determine cell fate in the ectoderm, and may be common to all neural derivatives. Regulation of transcription of genes expressed in the CNS and PNS is under investigation. Bray et al. (1989) and Dynlacht et al. (1989) have identified cis-acting elements in the dopa decarboxylase promoter and cloned a trans-acting factor that binds to those sequences. But the expression of this factor is restricted to a subset of neural cells and it
116 is not a general neural-specific transcriptional activator. For other genes expressed in the CNS or PNS like ftz, slit, cut, or sim, the modes of expression render it unlikely to easily identify general neural-specific cisand trans-acting elements. Strategies using antihorseradish peroxidase antibodies (Jan and Jan, 1982) or monoclonal antibodies (Fujita et al., 1982) to characterize neural specific genes have led to the identification of new gene products, but their functional analysis has not been completed yet. By application of enhancer trap strategies (Bier et al., 1989, Wilson et al., 1989) expression patterns were identified that show staining in subsets of the neural system. But as the authors stated, differences in the stability between the reporter gene and the gene product of the target site may lead to incorrect estimations in the extent of expression of these genes. In addition, these gene products also are found only in parts of the neural tissues, so that identification of more general neural factors may be difficult. By comparison of in situ hybridization data of the cloned rhomboid gene with lacZ expression patterns of a P-lacW insert near this gene, Bier et al. (1990) showed differences in the extent of the expression. In their P-lacW strain, the blastoderm pattern and the transient expression of the rho gene in two other parts of the embryo is not mimicked by detectable lacZ expression. Thus gene products identified by enhancer traps may underscore the real extent of expression. So most of the genes described above are expressed only in parts of the neural tissues, or are restricted in their expression to certain development stages, or even to single cell types. In contrast, the 131 tubulin gene is expressed continually during the entire period of development of the CNS and PNS. Thus, the/31 tubulin gene offers an excellent system for the identification of cis regulatory sequences, in vivo and in vitro, as well as for the identification and isolation of trans-acting factors that are involved in the regulation of the expression of this tubulin isotype.
Methods
Generation and purification of isotype specific antibodies The 15 carboxy-terminal amino acids were synthesized on an Applied Biosystem peptide synthesizer 430 A according to Meierhofer et al. (1979) and purified by H P L C as reported previously (Leiss et al., 1988). The purified peptides (/31: D A E F E E E Q E A E V D E N ; /33: E F D P E V N Q E E V E G D C I ) were covalently linked to Keyhole Limpet Hemocyanin (KLH, Calbiochem.) according to Tamura et al. (1983). Antibodies were raised in rabbits and purified as outlined by Caroll and Scott (1985). Collected fractions were tested by ELISA and Western blotting.
Western blotting For Western blotting, protein extracts were prepared from the desired tissue or stages. The samples were separated on a 11% SDS polyacrylamide gel and were transferred to nitrocellulose (Towbin et al., 1979). Antibody reactions were essentially done as reported previously (Leiss et al., 1988). Antibody staining of embryos Embryos were dechorionated, permeabilized and fixed essentially as described recently (Leiss et al., 1988). After washing and blocking in BBT (0.15% crystalline BSA, 10 mM Tris-HC1, p H 7.5, 50 mM NaC1, 40 mM MgC1 z, 20 mM glucose, 50 mM sucrose, 0.1% Tween 20) they were incubated with the anti-/31 polyclonal (1 : 50) or the anti fl-galactosidase monoclonal-antibody (purchased from Promega) (1:1000) at 4°C overnight. The bound antibody was detected with a biotinylated secondary antibody and stained with the Vectastain ABC-kit (VectorLabs) using diaminobenzidine following the manufacturers instructions, except that 30 ~tl of 1% NiC12 were added per 600 /~1 staining mix (according to Lawrence et al., 1987). For the double-stains, embryos were incubated with both primary antibodies, than sequentially with the secondary antibody and using Ni-ions only during the first staining procedure. The/33 tubulin antibody was diluted 2500 fold and the /3-gal antibody 10000 fold. Individual embryos were mounted on slides in Epon and viewed under a Leitz microscope. All photographs (Kodak, Ektachrome 50) were taken under Nomarksi optics. Isolation of nuclei Embryos were staged and collected for appropriate times at 25°C and washed with 0.7% NaC1. They were homogenized in HPS (10 mM Hepes-KOH, pH 7.6, 10 mM KC1, 1.5 mM MgC12, 0.25 M sucrose, 0.5 mM DTE, 0.5 mM leupeptine) using a motor-driven glassteflon potter (4-6 strokes at 450 rpm). After removal of the cell debris by low-speed centrifugation (400 × g, 5 min, Heraeus Minifuge) the supernatant was placed on top of a cushion of HPS + 20% ( v / v ) glycerol, and the nuclei were pelleted at 5000 × g for 20 min. They were carefully resuspended in BC100 (20 mM Hepes-KOH, pH 7.6, 5 mM MgCI z, 0.2 mM EDTA, 100 mM KC1, 20% ( v / v ) glycerol, 0.5 mM DTE, 0.5 mM leupeptine, 0.5 mM PMSF), thereby leaving the yellowish yolk proteins behind. The nuclei were either used directly or quick frozen in liquid N 2 and stored at - 8 0 ° C up to several weeks without loss of activity. Transcriptional run-on assay Assay volume was 100 /~1, containing roughly 1 0 6 nuclei. The reaction conditions were 20 mM HepesKOH, pH 7.6, 75 mM KC1, 5 mM MgC12, 0.2 mM
117 EDTA, 0.5 m M DTE, 0.5 m M leupeptine, 0.5 m M P M S F and 5 U of RNasin (Amersham). The assay contained 1 m M of each ATP, CTP and GTP, 10 ~tM U T P and 50 /~Ci [ a - 3 2 p ] U T P (600 C i / m m o l , Amersham). Incubation was performed at 25°C for 2 h. The reaction was terminated by the addition of E D T A at a final concentration of 10 m M and subjected to p h e n o l / c h l o r o f o r m extraction. The synthesized R N A s were precipitated and directly dissolved in the hybridization buffer (50% formamide, 1 x Denhardts, 2 x SSC, 100 t~g/ml herring sperm DNA). D N A s were immobilized to gene-screen using a slot-blot apparatus (Minifold II, Schleicher and Schuell). Hybridization was done in 50% formamide at 42°C overnight. The filter was washed at high stringency (0.1 X SSC, 0.5% SDS, 65°C), dried and exposed overnight using a K o d a k X R film.
Preparation of nuclear extracts Nuclei were isolated as described above. The pellet was resuspended in 1 vol. nuclear extraction buffer (20 m M H e p e s - K O H , p H 7.6, 20% ( v / v ) glycerol, 420 m M NaC1, 1.5 m M MgC12, 0.2 m M EDTA, 0.5 m M PMSF, 0.5 m M DTE, 0.5 m M leupeptine) according to Dignam et al. (1983), and stirred for 30 rain at 0°C. The suspension was centrifuged for 20 min at 25000 x g. The supernatant was dialysed against 50 vol of BC 100 with two changes for 4 - 6 h, and aliquots were quick-frozen in liquid N 2. These extracts could be stored at - 8 0 ° C without loss of activity for several months. In vitro transcription The in vitro transcriptions were performed as described by Heiermann and Pongs 1985. Assay volume was 25/tl. Buffer conditions were 20 m M H e p e s - K O H , p H 7.6, 75 m M KCI, 5 m M MgCI 2, 0.2 m M EDTA, 12% ( v / v ) glycerol, 0.5 m M DTE, 0.5 m M PMSF. Template concentration was 0.5 pmol per assay. Reactions were carried out with 15 ~1 extract (corresponding to 100/~g of protein). Nucleotide concentration was 0.6 m M of each ATP, CTP, GTP, UTP, and the incubation time was 1 h at 25°C. Reactions were terminated by addition of a stop-mix (400 m M sodium acetate, 0.5% sarcosyl, 5 0 / ~ g / m l tRNA); nucleic acids were extracted with phenol/chloroform, precipitated and dissolved in 1 M sodium acetate, 0.1% sarcosyl. For S1 analysis, end-labelled fragments (see legend to Fig. 3) were added, the nucleic acids precipitated with E t O H and dissolved directly in 20 /~1 Sl-hybridization buffer (80% formamide, 40 m M Pipes-KOH, p H 6.8, 400 m M NaC1). They were heated for 30 min to 85°C and hybridized overnight at 49°C. 200 btl ice-cold S1-Mix (30 m M sodium acetate, p H 4.5,280 m M NaC1, 4.5 m M ZnSO 4, 30 /zg/ml herring sperm DNA, 100 U S1/assay) was added and the reaction incubated for 0.5 h at room temperature. They were extracted with p h e n o l / chloroform, precipitated, the pellet dissolved in 3 /~1 8
M urea and electrophorized on 5% PAA urea gels. The gels were fixed, transferred onto W h a t m a n n 3MM paper, dried and autoradiographed.
Construction of lacZ fusion genes Fusion between the lacZ reporter gene and the fll gene was performed as follows. The Sau3A site near the codon for amino acid 25 was used for in frame fusion by means of a BamHI site in p M C 1871 (Gasch et al., 1989). Sequences downstream of a BssHII site in the lacZ gene were replaced by the equivalent sequences of p C H l l 0 , thereby adding the SV40 polyadenylation signal. Genomic sequences from - 2 . 2 kb to + 2.8 kb in the intron were fused via a ScaI site, resulting in the construct WfllL. For WfllK, 2 kb were deleted from the intron by use of internal restriction sites (Avail and PstI), leaving 273 bp of 5' and 315 bp of 3' sequences from the intron. WfllC was constructed using a c D N A clone. Constructs in p W H L (Michiels, F., Hinz, U. and Renkawitz-Pohl, R., unpublished data), namely f l l W H I and f l l W H I I were made by linearizing the vector with StuI and inserting the intron cut with Asp700 at +0.1 kb (amino acid 4) and Bali at +2.7 kb (0.15 kb upstream of the second exon) in both orientations upstream of a truncated hsp70 promoter. P-element transformations P-element transformations were performed as published previously (Michiels et al., 1989).
Acknowledgements We thank Professor Campos-Ortega for helpful discussions, Sabina Kerl for excellent technical assistance, Professor O. Pongs for the gift of the histone gene clones, and Cathy Schindewolf for critical reading of the manuscript. Professor H. Holl~inder made his microscope facilities availaible for us. We acknowledge the financial support of the Deutsche Forschungsgemeinschaft (Re 628/3-1 and SFB 190) and the Bundesministerium fiir Forschung und Technologie.
References Anderson, K.V and Lengyel, I.A. (1980) Cell, 21,717-721. Basler, K., Siegrist, P. and Hafen, E. (1989) EMBO J., 8, 2381-2386. Bialojan, S., Falkenburg, D. and Renkawitz-Pohl, R. (1984) EMBO J., 3, 2543-2548. Bier, E., Vaessin, H., Shepherd, S., Lee, K., McCall, K., Barbel, S., Ackermann, L., Carretto, R., Uemura, T., Grell, E., Jan, L.Y. and Jan, Y.N. (1989) Genes Dev., 3, 1273-1287. Bier E., Jan, Y.J. and Jan, Y.N. (1990) Genes Dev., 4, 190-203. Biggin, M.D. and Tjian, R. (1988) Cell, 53, 699-715. Bodmer, R., Barbel, S., Shepherd, S., Jack, J.W., Jan, L.Y. and Jan, Y.N. (1987) Cell, 51, 293-307.
118 Bray, S.J., Burke, B., Brown, N.H. and Hirsh, J. (1989) Genes Dev., 3, 1130-1145. Campos-Ortega, J.A. and Hartenstein, V. (1985) Springer Verlag, Berlin and Heidelberg. Caroll, S.B. and Scott, M.P. (1985) Cell, 43, 47-57. Candy, M., Grell, E.H., Daubly-Chaudiere, C., Ghysen, A., Jan, L.Y. and Jan, Y.N. (1988) Genes Dev., 2, 843-852. Cleveland, D.W. (1987) J. Cell Biol., 104, 381-383. Dambly-Chaudiere, C. and Ghysen, A. (1987) Genes Dev., 1,297-306. Dignam, J.D., Lebovitz, R.M. and Roeder, R.G. (1983) Nucleic Acids Res., 11, 1475-1489. Doe, C.Q., Hiromi, Y., Gehring, W.J. and Goodman, C. (1988) Science, 239, 170-175. Dynlacht, B.D., Attardi, L.D., Admon, A., Freeman, M. and Tjian, R. (1989) Genes Dev., 3, 1677-1688. Foe, V.E. (1989) Development, 107, 1-22. Fujita, S.C., Zipursky, S.L., Benzer, S., Ferrus, A. and Shotwell, S.L. (1982) Proc. Natl. Acad. Sci. USA, 79, 7929-7933. Garcia-Bellido, A. (1979) Genetics, 91,491-520. Gasch, A., Hinz, U., Leiss, D. and Renkawitz-Pohl, R. (1988) Mol. Gen. Gen., 21, 8-16. Gasch, A., Hinz, U. and Renkawitz-Pohl, R. (1989) Proc. Natl. Acad. Sci. USA, 86, 3215-3218. Goldberg, M.L. (1979) Ph.D. Thesis, Stanford University. Heberlein, U. and Tjian, R. (1988) Nature, 331,410-415. Heiermann, R. and Pongs, O. (1985) Nucleic Acids Res., 13, 27092730. Jan, L.Y. and Jan, Y.N. (1982) Proc. Natl. Acad. Sci. USA, 72, 2700-2704. Jimenez, F. and Campos-Ortega, J.A. (1979) Nature, 282, 310-312. Johansen, K.M., Fehon, R.G. and Artavanis-Tsakonas, S. (1989) J. Cell. Biol., 109, 2427-2440. Karr, T.L. and Alberts, B.M. (1986) J. Cell Biol., 102, 1494-1509. Kemphues, K.J., Kaufmann, T.C., Raft, R.A. and Raft, E.C. (1982) Cell, 31, 655-670. Klemenz, R., Weber, U. and Gehring, W.J. (1987) Nucleic Acids Res., 15, 3947-3959.
Knust, E., Tietze, K. and Campos-Ortega, J. (1987) EMBO J., 6, 4113-4123. Knust, E. and Campos-Ortega, J.A. (1989) BioEssays, 11, 95-99. Lawrence, P.A., Johnston, P., McDonald, P. and Struhl, G. (1987) Nature, 328, 440-442. Lehmann, R., Dietrich, U., Jiminez, F. and Campos-Ortega, J. (1981) Roux's Arch., 190, 226-229. Leiss, D., Hinz, U., Gasch, A., Mertz, R. and Renkawitz-Pohl, R. (1988) Development, 104, 525-531. Little, M. and Seehaus, T. (1988) Comp. Biochem. Physiol., 90B, 655-670. Meierhofer, J., Waki, M., Heimer, E.P., Lambros, I.T., Makoske, R.C. and Chang, C.D. (1979) Int. J. Pept. Protein Res., 13, 35-47. Michiels, F., Falkenburg, D., Miiller, A.M., Hinz, U., Otto, U., Bellmann, R., Gl~itzer, K.H., Brand, R., Bialojan, S. and Renkawitz-Pohl, R. (1987) Chromosoma, 95, 387-395. Michiels, F., Gasch, A., Kaltschmidt, B. and Renkawitz-Pohl, R. (1989) EMBO J., 8, 1559-1565. Natzle, J.E. and McCarthy, B.J. (1984) Dev. Biol., 104, 187-198. Raft, E.C. (1984) J. Cell. Biol., 99, 1-10. Roberts, K. and Hyams, J.S. (1979) Academic Press, London. Rothberg, J.M., Hartley, D.A., Walther, Z. and Artavanis-Tsakonas, S. (1988) Cell, 55, 1047-1059. Rudolph, J.E., Kimble, M., Hoyle, H.D., Subler, M.A. and Raft, E.C. (1987) Mol. Cell. Biol., 7, 2231-2242. Soeller, W.C., Poole, S.J. and Kornberg, T. (1988) Genes Dev., 2, 68-81. Sullivan, K.F. (1988) Annu. Rev. Cell Biol., 4, 687-716. Tamura, T., Bauer, H., Birr, Ch. and Pipkorn, R. (1983) Cell, 34, 587-596. Thomas, J.B., Crews, S.T. and Goodman, C.S. (1988) Cell, 52, 133-141. Towbin, H., Staehelin, T. and Gordon, J. (1979) Proc. Natl. Acad. Sci. USA, 76, 4350-4354. Vassin, H., Bremer, K.A., Knust, E. and Campos-Ortega, J. (1987) EMBO J., 6, 3431-3440. Wilson, C., Pearson, R.K., Bellen, H.J., O'Kane, C.J., Grossniklaus, U. and Gehring, W.J. (1989) Genes Dev., 3, 1301-1313.