Developmental regulation of globin gene expression

15

Journal of Cell Science, S upplem ent 16, 15-20 (1992) P rinted in G reat B ritain © The Com pany o f B iologists Lim ited 1992

Developmental regulation of globin gene expression MARK MINIE1, DAVID CLARK1, CECELIA TRAINOR1, TODD EVANS1*, MARC REITMAN1, ROBERT HANNON2, HANNAH GOULD2 and GARY FELSENFELD1 1Laboratory o f M olecular Biology, NIDDK, National Institutes o f Health, Bethesda, M D 20892, USA and 2Biom olecular Sciences Division, King ’s College, London WC2B 5RL, England *Present address: D epartm ent o f Biological Sciences, University o f Pittsburgh, Pittsburgh, PA 15260, USA

Summary

We have used the globin family of genes in chicken to study developmental regulation of gene expression, both at the level of individual interaction of trans-acting fac­ tors with local promoters and enhancers, and at the level of chromatin structure. Regulation of all members of the a- and (3-globin clusters is affected by the erythroid regulatory factor GATA-1. Separate mechanisms exist for regulation of individual members of the family. As an example, we describe the control mechanisms that play a role in the expression of the p-globin gene, which

is expressed only in primitive lineage erythroid cells. In addressing the involvement of chromatin structure in gene activation, we have examined the role of locus con­ trol elements, and also considered the way in which RNA polymerase molecules might accommodate to the presence of nucleosomes on transcribed genes.

Introduction

Our earliest studies of globin gene expression were con­ centrated on the PA-globin gene, and were greatly aided by the development of a method for transfection of plasmids into primary erythroid cells (obtained from the embryonic circulation) with little or no damage to the cells (Hesse et al., 1986; Lieber et al., 1987). The method permits testing of the activity of globin promoters and enhancers in cells of both the primitive and definitive lineages. This led to the identification of a strong, erythroid-specific enhancer ele­ ment just 3' of the pA-globin gene (Hesse et al., 1986; Choi and Engel, 1986), which preferentially activates that gene in definitive lineage cells. As was subsequently shown (Nickol and Felsenfeld, 1988; Choi and Engel, 1988), the enhancer also functions in primitive lineage cells to acti­ vate the embryonic e gene, located downstream of the enhancer.

The globin gene family has served as a model for the study of developmentally regulated gene expression (Evans et al., 1990), allowing investigators to address the questions of how the entire family is activated in an erythroid-specific manner, and how individual family members are turned on and off at successive stages of erythroid development. We have focussed our attention on erythroid development in the chicken, in part because embryonic erythroid develop­ ment in chickens is well described, and because cells at each developmental stage can be obtained in abundance. Furthermore, since avian erythrocytes retain their nuclei, many regulatory factors can be isolated readily from cells found in the adult circulation. Two distinct erythroid lineages are produced in the devel­ oping chick embryo (Dieterlen-Lievre, 1988; Nikinmaa, 1990). The primitive lineage predominates in the circula­ tion until about day 5 after fertilization; at that point the definitive lineage appears. Primitive lineage cells express the embryonic P-globin genes p and e, and the a-globin genes, a n, a A and a D. Definitive cells on the other hand express (3H and pA, as well as a A and a D (Fig. 1). The steps leading to the activation of these genes require the partic­ ipation of typical trans-acting factors that bind to promot­ ers and enhancers, as well as the modification of chromatin structure over the a and p domains. Understanding of the globin regulatory pathways thus requires an analysis of both local and long-range interactions.

Key words: erythroid development, GATA-1, chromatin, transcription.

The general erythroid factor GATA-1 DNase I footprinting in vitro and site mutagenesis allowed us to identify essential elements of the enhancer (Evans et al., 1988; Reitman and Felsenfeld, 1988). Of particular interest was a pair of DNA sequence motifs of the form (A/T)GATA(A/G), which we had already identified in the promoter regions of the p and a D genes (Kemper et al., 1987). Related binding sites for this factor (now called GATA-1, originally named Eryfl in chicken, and NF-E1 or GF-1 in human and mouse) were identified in the promot-

M. M inie and others

16

P

Cluster

a Cluster Fig. 1. Organization o f the a-globin and (3-globin clusters in chicken. The letters e and i mark the p- and a-cluster enhancers respectively.

ers and enhancers o f all the globin genes (Evans et al., 1988). The cD N A from m ouse (Tsai et al., 1989) and both the cD N A and gene from chicken, cGATA-1 (Evans and Felsenfeld, 1989; H annon et al., 1991), have been cloned. The hum an GATA-1 cD N A and gene (Trainor et al., 1990; Zon et al., 1990; Zon and O rkin, 1992) have also been cloned. cGATA-1 is a 304 am ino acid protein w ith two metal ion ‘finger’ m otifs, w hich binds typically as a m onom er to an asym m etric site (Evans and Felsenfeld, 1989). 7>a«.s-activation studies in chick em bryo fibroblasts (Evans and Felsenfeld, 1991) show that som e prom oters carrying a single copy o f a G ATA-1 binding site are strongly stim ulated by expression o f G A TA -1. W hen plas­ mids carrying such a GATA-1 expression vector are co ­ transfected with an appropriate C A T reporter plasm id, a 360-fold stim ulation o f C A T activity is observed relative to antisense controls (Fig. 2). Rem arkably, the presence of additional GATA-1 sites in the prom oter does not increase this response in fibroblasts. A quite different effect is seen when the sam e group o f reporters is introduced into p ri­ mary em bryonic erythroid cells: the reporter carrying m ul­ tiple GATA-1 sites is m uch m ore active (Evans and F elsen­ feld, 1991). This result has led us to suggest that cGATA-1 is interacting either w ith an activator present in fibroblasts, or an erythroid cell-specific repressor. It seem s clear that cGATA-1 m ust play an im portant role in the activation o f the erythroid program . W e do not know how cGATA-1 itself is first activated in the early stages of developm ent. Exam ination o f the cGATA-1 prom oter

REPORTER

ACTIVITY IN CEF + PRSV20-1

+pRSV20-2

p a D /J

12 (5.5; 4)

980 (180:4)

paD 1

15 (7.0;4)

960 (200; 4)

paD 3

1.0 (N/A; 10)

360 (100; 10)

paD 4

0.8 (0.4; 6)

4.6 (1.8; 6)

paD 5

0.2 (0.2; 2)

270 (100; 4)

paD 6 ■

0.5 (0.3; 6)

130 (51; 6)

paD 7

0.1 (0.06; 2)

1.1 (0.4; 4)

0.9 (0.1; 2)

1.5 (0.02; 2)

p a D 8 r ' r r n r l i - . — a t-

a

(H annon et al., 1991; Fig. 3) reveals the presence of a clus­ ter o f three strong G A TA -1-binding sites upstream o f the site o f transcription initiation. 7Vans-activation studies in prim ary chick em bryo fibroblasts confirm that these sites contribute to the expression o f G A TA -1. A lthough this autoregulatory function may w ell serve to m aintain expression o f the gene, it clearly cannot initiate it. The first steps in activation o f G ATA-1 expression m ay involve either the binding to these sites o f other m em bers o f the G A TA fam ily, or interaction o f other factors w ith putative sites that w e have identified in or near the GATA-1 pro­ moter. N otable am ong these is a c-m yb site and a site for an unidentified m em ber o f the steroid horm one receptor superfam ily in the first intron (Fig. 3). W e are presently investigating the possible role o f these sites in control of G ATA-1 expression in early erythroid progenitors.

Control of individual globin genes GATA-1 clearly can serve as part o f a general erythroid ‘sw itch’ m echanism , but how can it contribute to lineagespecific expression o f individual globin fam ily m em bers? W e choose as an exam ple the p-globin gene, a m em ber o f the P-globin fam ily that is expressed only in cells o f the prim itive lineage. W hen plasm id constructions containing the p-globin prom oter are transiently transfected into pri­ m ary erythroid cells, the sam e lineage specificity is observed (Fig. 4). T he construction pC A T is active in prim ­ itive lineage cells, but not in the definitive lineage. As show n in Fig. 4, constructions carrying the (3A prom oter (p AC A T E ) or the a A prom oter ( a AC A T I) also are expressed in a way consistent with their behavior in vivo. T hese results indicate that the inform ation for lineagespecific ( ‘stag e’-specific) control o f p-globin gene expression is contained in the 456 bp-prom oter segm ent o f pCA T. To identify the controlling elem ents, w e have m ade a series o f deletion and site m utations in the pC A T p ro ­ m oter (M inie et al., 1992). T he effect o f the site m utations is sum m arized in Fig. 5. T w o elem ents critical for expression from the p prom oter in prim itive cells are a GATA-1 m o tif at -2 2 1 bp from the cap site, and an S p lbinding site 50 bp 5 ' o f the transcriptional start site. Since G A TA -1 and S p l are present in both prim itive and definitive lineages, it is perhaps difficult to see how they

TRANS - ACTIVATION

Fig. 2. Effect o f expressing cGATA-1 in primary chicken embryo fibroblasts on transcription from various promoters containing GATA-1-binding sites. GATA-1 expression vectors were transiently co­ transfected with CAT reporter plasmids carrying various promoters. Results were normalized for transfection efficiency by co­ transfection of a (3-galactosidase expression vector. The solid black boxes represent GATA-1-binding sites. (From Evans and Felsenfeld, 1991).

R egulation o f globin gene expression

17

C A T G G C T G T C G C T G A T A A G C A G GA G T T G C i A G G A A G G G C A C A G T G C C T T C C T T T A A C C C G G G GAGGGGGGGGTGGGGGGGGGAT G A C A C C A T G A C C C A C C C

-554

n n n A r j n T C n r T T T A A n r T C n n C A n n A n C A C T C O n n C O T C O C C A C C C C A A A A G C G G A G G T C T C A G T G T G C C CC C C C C C A A T AÀlSAACCCACTATGGGGTGG

- 454

{10}

GGGGGAGGTAAAGCAAGGGGGGGGGCACCCCCTGATGCGTGTGTGTGTCCCCCCCCCCCGTTCCCAACTGCCTCCCGCTC TTJ Tal [ c - my b 1

TI

[gata - r

T G C C C T A C A G C C C T G G G G C C G T T A C C C A A A A G G G C A G G G G G G C A C C C T A A A G T G C C C G C C C C C C C C C A C A C C A A A A A A A A A C T T C T A C A AGGAGGGG^

a m

- 254

G G G G G G G G G G T G C A C C G T C T G £ G |G C G G G G C C G T G T C C G T C C G T C T G T C C G T C G G T G C G A C C G C G G T G A C A C A C G G C T G C A G G T C A G T G A C C C C C C C ACG I n t r on 1 TTT TeT

- 154

C C G G G G G G G G G G G G G G G C G A C T G T A T T A T T T T GGGG GGGGGGG AGGGGG AGG TC A T CGG T GGG T T T A A T G G A G T T A A T GC AGGG AGGGGGC AC T GCJLGAJ.

- 54

m

{2}

in

JG AGTTCGTGGCGCTG od i ng

AG AT TCACACACACACACACACCCTCCTTCCTTTTTCTTTCAGGCCCCCCCCCj ♦-■i n t r o n 1

Fig. 3. Protein-binding sites on the promoter o f the cGATA-1 gene. Lines represent sequences protected in DNase I footprinting experiments with extracts from erythroid cells. The steroid hormone receptor superfamily binding site is the palindromic sequence GGTCA.. at the 5 ' end of the first intron. (From Hannon et al., 1991).

A

Primitive lineage specific DNase I hypersensitive site

B Definitives

Primitives if Sma\

€

NeoI

pCAT a ACATI

Smd :

:

\

:p -g lo b in gene

\ 456 b p /

^CA TE 100 75 50 25

0 25 50 75 too

0

CAT Activity P

C

A

^ I C A T/SV40

Fig. 4. (A) Plasmid construction used in transient expression studies o f the p-globin promoter. (B) Lineage specific expression of cat under the control of the p-, a A- and PA-globin promoters. (From Minie et al., 1992).

Definitives

Primitives Spl/feGPl

C = 5 H :^ Spl

pCAT

pC AT3.1

p C A T l.l

pC A Tl .5

ß CAT ßACATE J ___ L

300

200

100

i

0

CAT Activity

i

100

could be responsible for lineage-specific expression. H ow ­ ever, m easurem ent o f the nuclear concentrations o f both factors in prim itive and definitive lineages reveals that there are m ajor differences betw een the two lineages. As show n in Fig. 6, the concentrations o f GATA-1 and S p l (as m easured by their ability to bind to their specific D N A sites in a gel shift assay) decrease by alm ost an order o f m ag­ nitude as the prim itive cells are replaced by the definitive

i

i

200

_u

300

Fig. 5. Effect of mutation in GATA-1-, S p l- and TFIID-binding sites o f the p-globin promoter on expressed CAT activity. (Minie et al., 1992).

lineage. W e have also m easured GATA-1 m R N A levels: these decrease by 3- to 4-fold. Since we know the approxim ate nuclear concentrations o f the factors, and their specific binding constants, w e can estim ate the effect o f the decrease in nuclear abundance of factors on the occupancy o f sites on the p prom oter w ithin the nucleus. W e calculate that the observed concentration decrease m ight result in a reduction in the num ber o f doubly

18

M. M inie and others

A Primitives

CD Cl

>i—

•

o Definitives a