J. Mol. Biol. (1991) 218. 55-67

Structure of the Core Promoter of Human and Mouse Ribosomal RNA Gene Asymmetry of Species-specific Transcription Yoshiaki Ishikawa, Geza Safranyt, Koji Hisatake, Nobuyuki Tanaka Yasushi Maeda, Hiroyuki Kato, Ryo Kominamif and Masami Muramatsu$ Department of Biochemistry The University of Tokyo Faculty of Medicine Hongo 7-3- 1, Bunkyo-ku, Totkyo 113. Japan (Received 4 July 1990; accepted 8 November 1990) In vitro transcription of the ribosomal RNA gene (rDNA) shows a remarkable species specificity such t.hat human and mouse rDNA cannot use heterologous extracts of each other. The region that is responsible for this specificity has been studied using human-mouse chimeric genes and characteristic structures of both core promoters are presented. When the mouse sequence is substituted by the corresponding human sequence from upstream. bhe promoter activity in the mouse extract begins to decline at’ nucleotide -32 or -30. decreasing gradually and is lost completely at - 19. A similar gradual decrease was noted for t,he 3’ side substitution, which started at nucleotide - 14 and was completed when up to the nucleotide -22 mouse position was replaced by the corresponding sequence from human. Thus, in the mouse rDNA core promoter, the sequence that is involved in species specificity resides only in a stretch encompassing the non-conserved region between the distal conserved sequence (DCS) and the proximal conserved sequence (PCS), plus two altered nucleotides in the PCS. When human rDNA is transcribed with human cell extract, the mouse sequence cannot substitute for the human sequence within the region from nucleotide -43 to + 17 without affecting promoter activity significantly. This asymmetry of species specificity is due to the presence of nucleotides -43, + 1 and + 17, which are sensitive to change in only the human core promoter. The difference in the 5’ border is ascribed to the species specificity of a transcription factor TFID, which recognizes this region. But the large difference of the 3’ border is apparently due to another factor, possibly RNA polymerase I itself, because this region is not recognized by TFID in either human or mouse. Mammalian rDNA core promoter appears to consist of a tandem mosaic in which three evolutionarily conserved sequences alternate with non-conserved sequences having certain functionally important nucleotides. Not only non-conserved sequences and non-conserved nucleotides in conserved sequences, but also the spacings between I he three conserved regions, play ii crucial role in species specificity.

tpresent address: I)epartment of Biochemistry. Frederic ,Joliot-(‘uric National Research Instit’ute Radiohggiene

and Radiobiology,

Budapest,

IPresent address: 1st Department Niigata

University

School

of Medicine.

for Hungary.

1. Introduction The fundamental mechanism by which transcription of the eukaryotic ribosomal RNA gene (rDSAI1) is initiated is yet) to be elucidated (for reviews. see Muramatsu, 1985; Sollner-Webb & Tower, 1986). In mammalian and frog cells, the promoter may be separated into two parts, the core and upstream sequences (Haltiner et al., 1986; Windle & Sollner-Webb, 1986; Haltiner-Jones et al.. 1988). The core promoter is present, near and surrounding the transcription start, sit#e of rDNA and is required for basic. arcurat,e and minimal but,

of Biochemistry. Niigata.

NA could not replace the human counterpart in transcription hy human core promotter with human extract (Hell et nl.. 1988: Jantzen rt nl.. 1990). Both the human and mouse core promot,ers show eficient transcription initiat’ion. albeit, lower. in the absence of CCE. The core promoter also requires homologous extracts in both the human and mouse transcription system. Therefore, to elucidate the interact,ion between the core promot,er and TFII) (or a,n equivalent factor). and to analyze t’hr strucatural features of t,he rDh’A promot’er. it’ is important, to determine the sequences that are responsible for t’he species specificity. Tn t,his work, WC have constructed a variety of human-mouse chimerica genes and determined those regions that are required for t,his species-specific transcription initiation using both human and mouse transcription systems. The results revealed an interesting feature of the transcription machinery, namely asymmet,ry. and the characteristic structure of t’he core promoter

of mammalian

rl)NA.

2. Materials

and Methods

Their construction has been described (Safrany r/ CL/.. 1989). M-wild is thta same as pMrSI (Kishimoto ~1 nl.. 1985). (ii) MHM-type

chimeric

ZIX.4

To prepare the MHM-type chimeric DNA. the mouse sequencefrom nucleotidr - 166 to + 133 was inserted into a vector derived from pHR322. Any portion nucleotide -40 and f26 was exchangeable

between by t,hr

cassette method (Grundstrom e/ nl., 1985). (iii) f~Tpstream-type chinreric DNA To prepare these types of chimeric DNA, 5 fragments were first prepared. They were a human rDN4 upstream fragment’ (@), a part of HM-type chimeric DNA (0). a part of the pBR32P vector (a). a mouse rDNA upstream

fragment (@), and a part of t’he MHM-type rhimeric DNA (a). By an appropriate ligation of 3 fragments, rachDNA(H1: 0. @ and @;H2 0, Q and 0: Ml: 0. @ and 0: M2: 0.

0 and 0. was constructed.

To prepare the H-type chimerio IINS. the human sequence from nucleotide - 595 to + 698 was inserted into a vector derived from pCC 9 (H-wild had the same sequence in pBR322). Any portion between nucleotide -72 and f77 was exchangeable by the cassette met,hod (Grundstrom et a,Z., 1985). The det,ailed procedure for the const~ruction of chimeri\ desired

point so as notj to overlap any nu&otidr hv synt htlsizing t>he c*himeric DNA (SW Matjr&ls nrrd Methods). The chimeric genes are named after the horder between the human and mouse sequences by mouse co-ordinate: e.g. the HM -47 chimeric gene consists of a human sequence upstream from human nucleotide -49 (up to - 155) and mouse sequence down stream from nucleot’ide -47 (to + 133). There is a nucleotide length difference at this point because of t’hr deletions in the mouse sequence for alignment to maximal homology with human (Fig. l(a)). The in oitro template activity of these chimeric genes was tested by a mouse FM3A cell extract (Mishima et al., 1981, 1982). To standardize the conditions for transcription. a mouse wild-type template (M-wild) was included in every sample as an internal control; the results are summarized in Figure l(b) with their constructs. Control transcription became weaker when put together with t)ranscriptionally active chimeras. presumably due to the competition of factors. The chimeric gene (HM -38), carrying the mouse sequence downstream from nucleotide - 32 together wit)h a conserved region between -38 and -33 (mouse co-ordinat)es), designated DCS (distal clonsrrved sequence), could be transcribed at about 95°{j of HM-47, which had a nearly wild-type level of activity under these conditions. Stepwise replacement of the mouse sequence with the human c&ounterpart caused a gradual decrease in promoter activity elicited by the mouse extract. as shown in Figure l(b). The activity became about 50%, when t’he upstream sequence was replaced with a human sequence up to mouse nucleotide -25 (hereinafter, we refer to the nucleotide position by the coordinates of the original species from which the was derived (here mouse), unless stat’ed promoter otherwise). A close examination of the sequence difference in this region reveals that only four base changes are present, including one deletion. Further replacement of the mouse sequence between -25 and -20 with the corresponding region of human rDNA caused a rather sharp decrease in promoter activity from about 50% to lo:& (compare Fig. l(b), lanes 5 and 7). Here the effect of chimerizat’ion is twofold. Not only are the bases changed. but, the distance or spacing between the two conserved regions, DCS and another conserved sequence designated PCS (proxirnal conserved sequence). is altered by the three ext,ra bases in the human promot at - I4 decreased the activit.y to only about 500,, (confirmed by a longer exposure, data not shown). This means that, although this region (PCS) is important for t,he core promot’er activity. t.hr - 19 base plays a greater role than t’he - 14 hasc in determining species specificity. Further replacements from - 11 up t,o + 18 produced completely inactive promoters for the mouse extract (Fig. 1(b). lanes 11 to 13). The 5’ border of the mouse speciesspecific promoter element has thus been rst’ablished at nucleotide -32 (note t,hat nucleotides --38 through -33 are conserved between these species) for total and nucleotide -20 for minimal in &ro transcription activity. iirld

iLW

(b) Z’hr 3’ horder of the mouse species-spa:@ promoter

rlrme~nt

To determine the 3’ border of the mouse speciesspecific promoter elements, a chimera HM - 47. was chosen as a starting construct because it had almost the same promoter activity as the wild-type and was easy to manipulate further. Preliminary experiments had shown that the replacement of the region downstream from + 17 with the corresponding human sequence did not, change the promoter activity in the mouse extract. Therefore, we started to change the upstream sequence of + 1 to the human counterpart’, leaving the downstream of +2 in HM-47 as it was. (Note that the sequence from + 2 through + 16 is conserved between mouse and human.) These chimeric genes were named after the mobile horder by the 3’ end of the mouse sequence upstream from t’he human insert (Fig. 2). The replacement from + I to - 11 did not change the template act’ivit’y at all (Fig. 2. lanes 2, 3 and 5). The nucleotide order of this 12 nucleotide region thus appears not t’o be involved in species-specific recognition by the mouse transcription machinery, although we must note that three nucleotides (- 1. -5 and - 7) are conserved bet)ween human and mouse. Indeed, we have reported that C at’ -7 is essential (Kishimoto et u.Z.. 1985). Clos et CLI. (19866) also reported the importance of T at - 1. The in vitro transcription activity was reduced to a great extent when - 14 (MH - 15, data not shown) or bot*h mouse nucleotides - 14 and - 19 (MH -20) were replaced by human nucleotides (Fig. 2). The transcription activity was lost completely when the human insert reached the -24 human position (MH-22). The 3’ border of the mouse npeciesspecific element has thus been identified at’ nucleotide - 14. from the maximal transcription activity of MH - 12. because - 1% and - 13 are conserved nucleotides. The 3’ border of the minimal activity is -20, since MH -22 lost the activity completely. We not,e that the difference between MH -20 and MH -22 is only the addit,ion of three nucleotides

Asymmetry of Species-specilc Transcription

1

of

59

rDNA

234578

“_ *

I’(i‘ : i (.

Transcription

(%I

-166 1

+291

M-wild

m -155

2

MH-4

01

3

MH-7

01

-47

-4

cm

-7

+2

100

+I33

100 -m

100

-m

50

Drn

100

-m

30

-10

4 5 6 7

MH-lO(-12del) MH-12

01 -12

or -15

ql

MH-15 MH-20

-20

01

n

10

-22

8

MH-22

01

mm

0

Figure 2. Transcription activity of the 3’ human replacement chimeras of mouse rl)X;A in mouse extract. Autoradiographs of gel electrophoresis of transcription products from the various constructs shown underneath. Numbers on the top of thr gel correspond to the numbers of the con&u&s shown in the left column. Arrowhead a, product of wild-type promoter M-wild digested with AccI (470 nucleotides); arrowhead b. products from various chimeras digested with HincII (592 nucleotides). The arrow shows the bands of control plasmid Nwild digested with /‘vu11 (291 nucleotides). Open and filled bars represent human and mouse sequences. respectively. Xumbers on the bars denok the 3’ and 5’ ends of mouse and human sequences. The rights column shows promoter activity (expressed as a percentagr of M-wild) of chimeric constructs.

(2 G and 1 A) in the non-conserved region between DCS and PCS. and therefore infer that the size increase would have been fatal to the promoter for the recognition by the mouse transcription machinery. When - I BC was deleted accidentally during the preparation of chimeras, the resultant’ chimeric gene (MH - 1O( - 12del)) showed significantI) decreased promoter activity (--5Oo/,; Fig. 2, lane 4). In a previous study, we found that the substitution of - 12C to T did not affect the promoter act’ivity at all in mouse rDNA (Kishimoto et al., 1985). The above results therefore suggest that the spacing between the start site conserved sequence (SCS) and PCS or DCS is also important.

(c) The 5’ border of the human species-specijk promotw

element

Xext we proceeded to determine the speciesspecific region of the human rDNA promoter. For this purpose, a series of reverse-type human-mouse chimeric genes were constructed and tested for the promoter activit,y with a human cell (HeLa) extract. First,. a series of 5’ deletion mutants (upstream replaced by a pRR sequence) obtained from R. Tjian and S. P. Bell (University of California.

Berkeley CA, V.S.A.) were tested for the promoter activity in our system. The data agreed well with their previous results (Learned et al.. 1983). Then we examined the effect of substitut’ion of the upstream sequences with the mouse counterpart. Figure 3(a) shows that) replacement of the sequence upstream from -91 with the mouse counterpart diminished the promoter activity by a factor of -2.5 (lanes 3 and 4). This is as expected by the species specificit,y of UCI? reported (Learned et al., 1986). Incidentally, the reverse replacement of the corresponding region of mouse rDNA did not change the promoter activity appreciably (Fig. l(b): \‘amamoto et al., 19X4). The reason for this is not clear at present. Unexpectedly, Ml -41 had a very low promoter activity, less than 005 of M2-91 (Fig. 3(a), lanes 4 and 5). This indicates that some crucial sequence t,hat is species-specific resides in between -41 and - 91. An internal chimera in which only this portion is replaced with the corresponding mouse sequence (HB-proto) showed a very low act’ivity. crorroborating the above conclusion (Fig. 3(a,). lane 1). Further replacement up to -33 (MHM -33) or tnore downstream eliminated the adivity almost completely (Fig. 3(a), lanes 7 to 9). In order to locate the essential stjruc:ture for the above specificity we narrowed down t)hr effective

1

(a)

23456789

I

b a

Transcription

(%) -505

1

H2-proto

CZll

2

Hl -proto

IZlrl

-92

-41

+l +17

+133

mm

1-2

mm

80

-155

3

HM-proto

100

Irn -330

4

-91

M2-91

Drn

40

ma

1-2

-41

5

Ml -41 -166

6

MHM -proto

7

MHM-33

-41

am

l-2

-33

Cl

-29

8

MHM-29

Cl

mm -20

9

MHM-20

t

Cl

+18

Fig. 3. various mouse-in-human region I by constructing chimeras and testing the promoter activit’y each time. Figure 3(b) clearly indicates that the crucial region is confined to a short sequence between -45 and -42. Exchange of this sequence to t’he mouse count’erpart alone decreased the promoter act’ivit’y to (r25 of the wild-type (Fig. 3(b). lane 5). We further dissected this region by dividing it’ into two using best-matched alignment, and exchanging each part’ separately. However, both changes caused the same drastic decrease on the promoter activity as that caused by total exchange. The mutant H-1 GTA + CTT] may be regarded as being created by two point mutations at - 43 (A -+ T) and -45 ((S --) C), and H-(T -+ GTG] is due to two insertions of (: on both sides of -42T. Since bot’h of these mutants showed decreased activities. we next, exameach. ined this region by changing one nucleotide The results shown in Figure 3(c) indicate that reduced the promoter only the change at. -43A

activitv drastically. The hypothesis t)hat nucleotide -43A *is responsible was confirmed by synthesizing insert,ion mutants having C and A at -43, respecively (H-Iins-GC] and H-[ins-AT]: lanes 6 and 7). Clearly, when an A residue came to position -43 (lane 7). t’he promoter activity was retained well. whereas when that, position was occupied by (: was remarkably decreased. (lane 6) t’he activity Double point mutants H-[GTA -+ C’TT] and H-IT -+ GTG] again confirmed this rule. Altogether. the 5’ border of the human species-specific core which is promoter element extends to -43, upstream from IXS, by the presence of t’his specific nucleotide. (d) T/k

3’ hordw of thr human prom&r ekmrnt

species-speciJLic

The 3’ side of’ the human and mouw core promot,er shares a relat,ively long homologous region

dsymmetry qf Species-spec$c

1

(b)

2

3

4

Transcription

5

6.

7

of rDNA

8

61

@)I @

b

a

Transcription

(%) Hl -proto H2-proto

-505 I

II

I

II

+1

2

+133

80

am

t +18

H-wild H-C-71

5

w I

I

1

+17

I'

I

=I

I

+698 ICI

100

ICI

20

I

~-421 I

20

4

H-C-50--421 H-C-50--461

5

H-C-45--421

25

6

H-CGTA-

7

H-C

8

H-C-71

3

90

CTTI

30

T -GTGl

25

u-461

175

Human Mouse Human coordinate

; ~GGGAG~GTA~ T IATcTTT ; ~TTATG$TT/GTG:ATcTTT ‘t ’ ‘t -41 -45 ; 120 Fig. 3.

- 1 to + 18 (S(S), in which only nucleospanning tides + 1 and + 17 are different between these species (Financsek et al.. 1982b). In the experiments shown in Figure 4, effects of these mutations were tested using pseudo-wild-type promoters having different lengths of human upstream sequence. Figure 4 indicates that one point mutation at + 17 from human to mouse nucleotide (C + T) decreased the promoter activity in human cell extract to a small extent (to about 75 to -5O%), irrespective of the presence of the upstream sequence. Additional change at + 1 (C; -+ A) further decreased it t’o a much larger extent (to 5 to ~10~/~ of the intact

human promoter). The effects of + I and + 17 positions are also demonstrated with the chimeric promoters (MHM series) having a mouse sequence in the region upstream from -41. Although the overall activity was much lower than intact, human rDNA (Fig. 4, lanes 7, 8 and 9). the relative effects of substitution of nucleotides + 17 and + 1 were almost the same as those found with the wild-type promoter. When only nucleotide + 1 was altered the activity decreased to about @I: indicating clearly that nucleotide + 1 affected promoter function more strongly than nucleotide + 17 (data not shown). The above-mentioned results indicate that the effects of

1

23456789

-45

1

H-wild

2

H-C-45G-C H-C-44T-,A

3 4 5 6 7 8 9

GAGGTA

-42

-41

T

Transcript ion (%)

ATC

100

1 1

CTA

T

90

GAA

T

80

H-C-43A-,T 1 H-C-42 T-,A 1 H-C ins-GC 1 H-C ins-AT 1 H-CGTA-CTTI H-C T -GTGl

GTI

T

20

GTA

11 TGC

80

Ti;-T

50

GTA GTA CT1

15 20

G TAG_TG_

20

Figure 3. Transcription activity of (a) the 5’ mouse replacement chimeras of human rI)NA and (b) and (c) mouse-inhuman chimeras in human extract. Autoradiographs of gel electrophoresis of transcription products from various constructs are shown underneath (a), (b) and (c). Numbers on the top of the gels correspond to the numbers of the constructs shown in the left column. Bands a. phosphorylated 18 S RNA; bands b, products of end-to-end transcription. Arrowhead. products from chimeras digested with PslI ((a), 885 nucleotides). and those wild-type (lanes 1 in (b) and (c)) and chimeras (lanes 2 to 8 in (b), lanes 2 to 9 in (c)) digested with ScaT (698 nucleotides). and those from chimeras (lanes 0. @ in (b)) digested with ScaI (650 nucleotides). Arrow. products of H I-proto (lane 2 in (a). lane @ in (b)) as a control digested with ScaI ((a), 650 nucleotides) or PstI ((b) and (c), 885 nucleotides). Open and filled bars in (a) and (b) represent the human and mouse sequence, respectively. Numbers on the bars in (a) and (b) denote the 5’ and 3’ ends of the human and mouse sequence. In (b), nucleotide numbers are expressed by human co-ordinates except for + 133 on HI -proto (lane 0). In (c), underlines indicate the nucleotide substitution. The right column shows promoter activity (expressed as a percentage of HM-proto (a) or H-wild ((b) and (c))) of chimeric constructs.

the 3’ side of the core promoter are independent of those of the upstream sequences and suggest the involvement of different transacting factors (possibly RNA polymerase I itself) in these regions. In any event, we conclude that the 3’ border of the human element species-specific promoter extends to + 17 for partial effect and t,o + 1 for a

more drastic effect. nucleotides in SCS.

Hoth

are

the

non-conserved

4. Discussion In this work mouse chimeric

we have used a variety of human-genes and determined the 5’ and 3’

Asymmetry of Species-speci$ic Transcription

1

of rDNA

63

23456789

b i

a

Transcription

(%) 1

Hl +l,t17

2

-505

+l +17

+133

I

II

5

( IO)

HI +17

1

II

45

( 75)

3

H 1-proto

I

II

60

(100)

4

HM +1,+17

5

HM -f-l7

50

6

HM -proto

100

7

MHM +I, +17

8

MHM +17

1-2

9

MHM - proto

1-2

-155 r

5

0

Figure 4. Transcription activity of the 3’ mouse replacement chimeras of human rDNA in human extract. Aut,oradiographs of gel elect’rophoresis of transcription products from various const’ructs are shown underneath. Numbers on t’hr top of the gels correspond to the numbers of the constructs shown in the left column. Bands a, phosphorylated 18 S RNA; bands b, products of end-to-end transcripbion. Arrowhead, products from chimeras digested with I’stI (885 nucleotides): arrow, products of Hl-proto as a control digested with ScnI (650 nucleotides). Open and filled bars represent the human and mouse sequence, respectively. Numbers on the bars denote thp 5’ and J’ ends of the human and mouse seyuencrs. Dots represent the mouse nucleotide. and the numbers are shown above. The right column shows the promoter activit,y (expressed as a percentage of HM-proto) of chimeric constructs. The activity relative to HI-proto is presented in parentheses. Hl + 1, + 17. H 1 $ I7 and Hl -proto are t~he same vhimeric HM- 1. KM+2 and HM+ 18 (in Fig. l(b)). respectively.

borders of mouse and human rDNA core promoter regions that show species specificity for each other’s transcription machinery. In mouse rDNA, the region that) is related to the species-specific transcription machinery was identified as a stretch -32 and - 14, including the nonbetween conserved region between DCS and PCS and the two non-conserved nucleotides in PCS (Fig. 5). This region, together with DC& forms the upper half of the core promoter and the binding site of the TFTD.

expressed penes as

a species-specitic transcription factor previously described (Mishima et al., 1982; Kato et al.. 1986; Safrany et al., 1989; Tanaka et al., 1990). In human rDNA, however, the region involved in species specificity was much larger; it encompassed a region from -43 to + 17 which covered the whole core promoter (Fig. 5). What is the nature of this asymmetry’! What causes it? 1Mammalian rDNA core promoters are made up of several discernible elements arranged in a tandem

Replacement Human 1 Mouse

:

-I””

I

7”

7”

+1 I

+20 I

I

ACGCGACCTCCCGGCCCCGGGGGAGGTA T ATCTTT CGCTC~GAGTCGGCATTTTGGGCCG~CGGGTTATTGCTGACACGCTGTCCTCTGG II IIIIII I I I II IIiIIII/ III II IIlIIlII II I IlIIIIIlIIIIlII I ACTTT CCTCCCTGTCTCTTTTATGCTTGTG-TCTATCTGT TC C TATTGGACCTGGAGATAGGTACTGACACGCTGTCCTTTCC I I I 1 DCS ' I' scs' -40 +I +20 -60 -20

Figure

5. Comparison of the regions that relate to species specificity between thr human and mouse transcription The percentage of intact core promoter activit,y shown by the chimeras derived from the replacements of nucleotides to the mouse or human seyuence is plotted by open (human s-+em) and closrd (mouse system) circBles. against thr replaced point. Below arp shown approximate nucleotidt, posItIons of both the rDNA (aorc promottsrs. Abbreviations are as for Fig. I (a)).

s.ystrm.

mosaic manner (Fig. l(a)). Three conserved regions identified as DCS, PCS and SCS exist in at least all the mammalian rDNA promoters examined including human. rat, mouse, hamster and rabbit (Financsek et al.. 1982a); Dumenco bt Wejksnora. 1986; De Winter, 1987) and seem to be functionally important because changes in these regions affect’ the promoter act,ivity drastically, as demonstrated by deletion as well as linker-scanning mutants ((irummt. 1982; Yamamoto et al., 1984; Skinner et nl., 1984; Miller et al., 1985; Haltiner it al.. 1986). These sequences must be important for the intera&on with some transcription factor(s) that recognizes the characteristic structure of the sequence. Interestingly, the relatively infrequent base substitutions in these regions have much to do with the species specificity of transcription. For instance, the A to T substitution at - 19 in mouse PCS completely eliminates the remaining t’ranscription activity of a human-mouse chimeric gene by mouse extract (Fig. 1 (b), lane 9). This is presumably due to the lowered affinity of mouse species-specific factor TFI D to the core promoter, since the - 19 position is in the region where TFID binds (Safrany ef al.. 1989; Tanaka et al., 1990). Mutation G to A at -t 1 in human SCS causes a considerable decrease (to - 10%) in the promoter activity by the human extract (data not shown). By contrast. a similar decrease is not encountered when the + 1 nucleotidr is changed from A to G in mouse SCR and transcribed by mouse extract). Substitution of nucleotidr + 17 had a similar, though not as great, effect on human rDNA. This is consistent with the results of a deletion experiment of Learned et al. (1983). This

apparent asymmet,ry of the transcriptional species specificity is notewort’hy but the underlying mrchanism is yet, t,o be clarlhed. One possibility may be that, human poI1 rat)her rigorously requires (iTI’ as the initiating nucleotide, while mouse poll can utilize both ATP and GTP. In one experiment, we added a mouse polI fraction to the reaction where template Hl + 1, + 17 (Fig. 4. lane 1) was being transcribed by HeLa extract. No recovery of transcrip tion was seen, however (data not shown). Probably, mouse ~011 was not able to function with human TFTI) under t’hese conditions. The requirement,, though to a lesser extent, of a specific nucleot’ide (C’) at’ position + 17 for human initiation may also argue for the more stringent’ sequence requirements of human poll than of mouse polT in the initiation process. Non-conserved regions alternating with IXS. PCS and SCS are evolutionarily diverged. but also are an important integral part of the core promoter. These regions contain certain essential nucleotides for transcription initiation such as -7G (in both human and mouse) and -43A (in human alone). Mutation of the former nucleotide to A cornpletelq destroyed the promoter activity of mouse rDNA et al., 1985). The same change in human (Kishimoto rDNA decreases the promoter activity to lO(% but does not eliminate it completely (N. Tanaka. unpublished results). Incidentally, the mutation from G to A at - 16 in PCS of mouse rI)NA drastically impairs the promoter activity (to -10%) (Skinner et al., 1984; Kishimoto et al., 1985), whereas in human rI)NA it decreases only to a half (-50%) (TV. Tanaka, unpublished results). Thus.

Asymmetry of Species-speci$c Transcription the effect of mutations on these conserved nucleotides is not always the same among different species. The -43A nucleotide in the human rUNA core region is particularly interesting. This nucleotide is located outside the three conserved complexes but nevertheless is essent,ial for the human core promoter (Fig. 3(c)). This is in contrast to the mouse COW promoter, the border of which is mapped at the 5’ end of the IKR. A linker-scanning experiment by Haltiner et al. (1986), showing that replacement of a region from -52 to - 47 did not change promoter activity. is compatible with our resu1t.s. Sot onl) the substitution of the upstream region of DCS wit,h the human counterpart but also its substitution with indifferent sequencxes from a bacteria1 plasmiti do not affect appreciably the core promoter act,ivit.! of mouse rl)NA ((jrummt,, 1982; Learned et al., 1983; Yamamoto et al., 1984). Thus. the human core promoter apparently has a larger size than that of the mouse. Tndeed, the former has a three-nucleotide larger spacing between JXS and PC’S, Alternatively. it is also possible that although the mouse core Jjromoter extends upstream from 1X%. the r+I+c~t.of this region just, cannot, be detected by this transcription svWm in vitro btbc,ause of its intrinsic: lower srnsitlvit y. The length of’ the non-conserved region, or rat,hcr the sJ)acing between the conserved regions, is very important for the promoter activity and is spe(*icsspecbific. This conc*lusion is corroborated by the fin& ings shown in Figures I(b) and 2. The stringent requirements for spacing constraints as well as the core promoter has been sequence uit.hin experiment suggest,ed by a linker-scanning (Haltiner ut al.. 1986). The spacing between UCE and thr, core promoter is also rather critical (hliller uf a,l.. 1985: Haltiner rt al., 1986), suggesting that the stereospecific interactions are crucial in each hierarchy of transcription machinery. I’(1E was also reported to have sJ)ecirs specificity, because it could not be exchanged between human and mouse (Haltiner et al.. 1986): it, is also caonfirmed in this work (see Fig. 3(a). lanes 3 a,nd 4). Presumably. sJ)ecies sJ)ecificity is Jjresent in both struc*tures: one core promoter governs t.hc basic minimal J)romotion and the ot’her U(IE augment’s the former t.o it full aactivity. It could. howrvcr. sometimes b(> overcome, as seen in a mouse system which could utilize the Xunopus rT)SA J)romoter (Culotta rt nl.. 1987). It was also reported t’hat the XUUY~~US rI)Nd4 J)romotrr could be converted into a mouse-type one by cha,nging the spacing between I’CE and the core promoter by half a helical turn of a J>NA double strand (T’ape et nl., 1990). It will be the stereospecific structural fitness between the promoter and t,hr transcription factors which drtermines the eficient transcription initiation. From these results, in combination with the previous results of deletion mutants. a picture emerges of‘ the general st.rurture of mammalian rI)NA core promoter. M’e postulate t,hat the mammalian rl)SA core promoter csonsists of at least three evolutionarily conserved and functionally

of rl>NA

65

important element’s that are rigorously spaced characteristically of the species. In addition. nonconserved elements also cont,ain a number of functionally import,ant nucleotides. some of which are common among different’ species. Hell rt nl. (1989) reJ)orted that the protein-protein interaction between transcription factors is c>rucial for the determination of species specificity in rl>?;A transcription. However. the result,s presented here. together wit’h the previous data of footprinting (Safrany Pf nl.. 1!)89: Tanaka rf al.. 19!)0), indichate that the yrot.eirr-I)NA interaction is also very imJ)ortant’ in determining the species spec*ificity. This st’udy has also provided somp csvitience for t.he (*oncurrent molecular c>volution of I)NA sequenc~e and the cognate t.ransac+ing proteins. Accvording to the analyses by gradual replac*ement’ with ceorresponding human sequencesfrom uJ)stream. the mouse core promoter loses its act,ivit,y gradually from - 32 to - 19. Tn other words. thttrtl is no key nuc+leotide in the non-conserved region whose change alone abolishes the promoter activity comJ)lct,ely in mouse extract. This is J)art,icularly int,eresting from the viewpoint, of evolution. since it indiratrs that, t,hr loss of mouse r1)N.A promoter ilctiVil> in heterologous transcription machinery took place not by one step but hy mult,ist,eJ) changes in t hc promoter sequences among dif%rent spwies. This is compatible wit.h Dover’s h!,J)ot’hrsis (Dover & Flavell, 1984) on the coevolution of t-DNA and its t,ransc*ription machinery. Accumulation of mutations, each ant’ of which affec%s t,hr func%ion only slightly. must have changed these promoters so as noi t,cl be recognized by the fact,or(s) from the other species. In human rl)NA. the effec*t,of’ mutation at -43X was so large t,hat, it was not J)ossiblr to evalu:tt,e the effect of further internal changes. However. the results of chimeras of t)he 3’ side region suggest similar gradual c.hallyes for t,he human

rl)NA

ww

promoter.

12’~ t,hank I)r Robert Tjian and St,ephrrl Bell fen providing us with the human deletion mut,antri used in this study. \Yr also thank Drs Kosuke Tashiro and Ttrshimitsu Kishimot.0 for their help in some aspects of t.hr experiments. This research is support,rd in part by grants from thts Ministry of Education. Science and t’ulturr of .lapan, and from the Found&ion for Promotion of’ (‘a.nccLrResrarch backed by t.he ~Iapan Shipbuilding

Industry liout~dation.

References Hell, S. I’., Learned, R. M., *Jantie. H. M. rYrTjian. R. (1988). Functional co-operatirity between transcription factors URFI and SLl mediates humxn ribosomal RIVX synt,hesis. Science. 241. 11X-l 197. Bell, P. I’., I’ikaard. C. S., Reedrr. R. H. & Tjian. R. (1989). Molecular mechanisms #overninp speciesspecific transcription of’ ribosomai R?;A. (‘~11, 59, 489-497.

~ios. ,I.. Buttgereit. D. 8: Urummt. I. (198ticl). A puritied tr;mscription f’actor (TTF-TR) hinds to essential sequencesof’ the mouse rDRiA promoter. I’TOC.Sut. Acad. fki., I ‘.S.A

83. 604-608.

(‘los, .J.. Xormann. A.. ijhrlein, A. & Grummt~. I. (1986h). The (bore promoter of mouse rDNA consists of two functionally dist)inct domains. S&. .1&s Res. 14. 7r,81- 7595. (‘ulot,ta. \‘. C’.. Wilkinson. ?J. K. bi Sollnrr-Webb. t3. (1987). Mouse and frog violate the paradigm of species-specific transcription of ribosomal RNA genes. f’roc. Na,t. Acad. Sci.. 7’.S.A. 84, 7498.750Z. I)e Winter. R. F. ,I. (1987). The funrtion of the ribosomal DIVA spacer in t,he promotion of rRNA transcaript,ion in Xenopua In&s. Ph.D. thesis. Vrijr ITrriversiteit. Amsterdam. 1)ovrr. (:. A. 8: Flavell. R. 1s. (1984). Molecular coevolution: DNA divergence and t*he maintennncse of fun?tion. C’rll. 38, 622-623. Ihunenc~o. \:. M. & Wrjksnora, P. ,J. (1986). (:haracterizabion of the region around t,he start point, of transcription of ribosomat RNA in t)he Chinese hamster. Gene. 46. ?%7-235. Financssek. 1.. Mizumoto. K. & Muramatsu, M. (1982a). Nuclrotide sequence of the transcription init,iation region of a rat ribosomal RNA gene. G’rnw. 18. 1l&122. Financsek. 1.. Mizumot’o, K., Mishima. Y. & Muramatsu. M. (198%). Human ribosomat RNA gene: Xucleotide sequence of the transcription initiation region and comparison of three mammalian genes. Proc. ‘Vat. Arad. A’ci., i:.S.A. 79. 3092-3096. Grunimt~. I. (1982). Nucleotide sequence requirements for specific initiation of t,ranscription by RNA polymerase I. Pror. Mat. Acrid. Sci.. I’.S.A 79. 6908-69 11. Grummt. I., Roth. E. & Paule. M. R. (1982). Ribosomal RNA transcription ill vitro is spec%-specific. Nature (London), 296. 17% 174. (irundstrom. ‘I‘.. Zenke, M. W.. Wintzerith, M., Yatthes, A. & Chambon. P. (198.5). H. W. I).. Staub. Oligonucteotide-directed mutagenesis b? microscale “shot,-gun” gene synthesis. Nucl. rlclds Rus. 13. 3305-33 16. Haltiner. MM.M.. Smalr. S. 1’. & Tjian. R. (1986). Two distinct promoter elements in the human rRNA gene ident,ified by linker scanning mutagenesis. ;Mol. (‘~11. lhl. 6. 227 23.7. Halt&r-*Jones. M.. Learned. R. M. 8r Tjian, R,. (1988). Analysis of clustered point mutations in t,he human ribosomal RNA gene promoter by transient expression in m’vo. Proc. A’&. Acad. Aci.. I:.S..4. 85. 669-673.

Horikoshi. M.. Wang. c*. K.. Fujii. H., Clromlish. ,J. A.. \Yeil. P. A. & Roeder, R. G. (1989). Purification of a yeast TATA box-binding protein that exhibits human transcription factor IT 1) activity. Proc. Sat. .-lead. Sri., i:.S.A 86. 4843-4847. Iida. (‘. T.. Kownin. P. 8: Paule, M. R. (1985). Ribosomal KXA transcription: Proteins and DNA sequences involved in preinitiation complex formation. I’roc. Sat. Amd. Zci., I’.S.A. 82, 1668-1672. .Jantzen. H.. Admon. A.. Bell, R. P. & Tjian. R. (1990). Nucleolar transcription factor hUBF contains a DNA-binding motif with homology to HMG proteins. Natwre

(London),

344. X30-836.

Kakidani, H. & Pt,ashne, M. (1988). GAL4 activates gene expression in mammalian cells. Cell. 52. 161-167. Kate. H.. Nagamine. M., Kominami, R. $ Muramatsu. M. (1986). Formation of the transcription initiation complex on mammalian rDNA. Mol. Cell. &o/. 6. 34 l&3427. Kishimot,o. T.. Nagamine. M.. Sasaki. T.. Takakusa. N..

Miwa, ‘I’.. Kominami. 1-t. B Muramat’su. It (I!)X6). Presence of a limited number of essential nu&otides in the promoter region of mouse rihosomal RNA gene. Sucl. Acids EPR. 13, 351~~3:i3%. Learned, R. M., Smate, S. T.. Haltiner. 81. M. & Tjian. R. (1983). Regulation of human rit~osomal RN,\ trawlscription. f’tw. Snt. :Icrrd. )Sci.. 1YS.. I. 80. 3558 -3562. It. (l!+X.?). Learned. R. hl. ~‘ordrs. S. & Tjian. Purific*ation and characterization of a transc~ril)tion factor that) caonfers promoter specificity to human 5. 135X t X6!). JZXA polymerast: I. ~lilol. C’f4. Hid Learned. Ii. M,. Learned, T. K.. Halt,illrr. M. M. & 1’jian. R. ‘I’. (1986). Human rRNA transcription is nlottrllatrd by thtl cbo-ordinate binding of two factors to ~II upstream control element. (‘ell, 45. 847 X.iS. Maniatis. T.. Fritsch. E. F. & Sambrook. .I. (t!JX:!). :llolrct~lar f ‘loniny: .3 Lnhorntwy Manual. ( ‘old Spring Harbor Labora,tor,v t’ress. (‘old Sljring Harbor. ?r’\.. Miller. K. (i.. Tower. ,J. HL Soltner-Webb. K. (19%). A ~orrrples cnontrol region trf the mouse rRNA gene directs accurate initiation by RNr\ poly~~w~m~ I. Mol. (‘~11. Niol. 5. 3% 562. 0.. Komirrami. I{. & Mishima. le.. Yamamot,o. Muramatsu. M. (1981). Ire vitro t~ranscription of it ~lonrd mouse ribosornal RX;;\ gene. .V/LC/. .-lcids Ifrs. 9. 6773-67XT,. .Mishima, I-.. Financsek. I.. Kominami. R. & .Mur~amatsu. Yl. (1982). l+actionation and rrc,oflstitrlt,iori of fftc*t,ors reyuirrtl for arcuratr t,ranscription of mammalian ribosomal RNA genes: ident,ification of a sprc>ies.Acids Krs. 10. dependent initiation factor. Sufl. 665!)&6670. Muramatsu. M. (19%). Transcriptional regulaticnl of mammalian ribosomal RNA genes. 12ioE.ssfr!ys. 3. 263&26T,. Paptl. I,. K.. ~$~indle. ,I. .J. & Solhier-\Vc:bb. I%. (l!)!)(1)‘. Half helicaal turn spacing changes convcart a frog illt,o a mouse rJ)N’A promoter: a distant upstream domain determines t.hr helix face of the init,iatiou site. (/p/Lp.s IhIvp. 4, 52 62. Safrany. C:.. Tanaka. X., Kishimoto. ‘I’.. lshikawa, Y.. Kate. 1-I.. Kominami, R. & Muramat’su. 11. (19X!)). Struc’tural determinant, of the sprcirs-spwific~ transc*ription of’ the n1ouse rRSA gene promoter,. Xo!. (‘rll. /Go/. 9. 349- 353. Schnapp. i\.. (‘10s. .I.. Hgdelt. IV.. Schrrck. K.. (‘vrkl. A. I!! (:rummt. 1. (1990). Isolation and functional cnharof TIF-113. a factor that c,onfers artrrizatiori promoter specificaity to mouse RNA polyrrrwitw 1. A’trd. ;tcid.s r%.s. 18, 13X5- 1393. Skinner. .I. .A.. Ohrlein. A. & (irummt, I. (19X4). /r/ ~it,o mutagenesis and transcriptional analysis of a mouse ribosomal promotIer element. Pror. .c~trf. drnd. ,Sci.. ly.S..t. 81, 2137~~2111. Sollnrr-L\‘ebb. t%. & Tower. .I. (1986). Transcription of’ doned rukaryotic ribosomal RNA genes. Av)//I. 11!r~. HiochPnl.

55.

WI

X30.

Tanaka, N., Kate. H lshikawa. Y.. Hisa,take. K.. Tashiro, K.. Kominami. Ii. 8.z Muranratsu. 11. (19!tO). Seyuencafa-specific binding of a transcriptiotl fac*tor TFID to the promoter region of mouse ribosomal RX:\ gene. .1. Hiol. C’hm. 265. 13836 13842. Tower. ,I.. (‘ulot,ta. V. (‘. B Sollner-Webb. B. (1986). Factors and nucleotide sequences t,hat direcat ribosomal DNA transcription and their relationship to the st,ablr t,ranscript,ion complex. ;l/lo/. (‘rll. fGo/. 6. 3451~3462.

Asymmetry of Species-spai& Tsujimoto, Y.. Hirose, S.. Tsuda, M. & Suzuki. Y. (1981). Promoter sequence offibroin gene assigned by in vitro transcription system. Proc. Nat. ,4cnd. Sri., lT.S..4, 78. 48383842. Webster. N.. Jun. ,J. K.. Green. S.. Hollis. hl. 8z (‘hambon, T’. (1988). The yeast UAS, is a t.ranst%ptional enhancer in human HeLa cells in the presence of the GAS4 trans-activator. Cell. 52. 169~-I58. Witrdlt~, tJ. ,J. & Sollner-Webb. K. (I 986). Tuo distant and

Transcriptotioa of rDNA

67

precisely positioned domains promote transcription of Xenopus Zaevis rRNA genes: analysis with linkerscanning mutants. Mol. Cell. Biol. 6. 4.585 4593. Yamamoto, 0.. Takakusa. N.. Mishima, Y.. Kominami. R. & Muramatsu. M. (1984). 1)etermination of the promoter region of mouse riboxoma,l RN4 gene t>y an irt rifro transcription system. I’roc. S/r/.