Isolation of a chicken HMG2 cDNA clone and evidence for an HMG2-specific 3′-untranslated region

Gene. 113 (1992) 251-256 © 1992 Elsevier Science Publishers B.V. All rights reserved. 0378-1119/92/$05.00

251

GENE 06375

Isolation of a chicken region

HMG2 c D N A clone and evidence for an HMG2-specific 3'-untranslated

(High-mobility group proteins; liver ontogeny; nonhistone chromosomal proteins; recombinant DNA)

Dorene L. Davis and John B.E. Burch blstitute for Cancer Research, Fox Chase Cancer Center, Philadelphia, PA 19111 (U.S.A.) Received by A.-M. Skalka: 9 August 1991 Revised/Accepted: 2 December/5 December 1991 Received at publishers: 16 January 1992

SUMMARY

HMGI and HMG2 (high-mobility group proteins) are two of the most abundant nonhistone chromosomal proteins in higher eukaryotes. Mammalian HMGI eDNA sequences have the unusual feature of being conserved not only over their coding regions, but also over large segments of their 3'-untranslated regions (3' UTRs) as well. In contrast, the only reported mammalian HMG2 eDNA clone has a distinct 3' UTR. We now report the isolation of a chicken HMG2 eDNA clone and show that it is markedly similar to the mammalian HMG2 eDNA clone over both its coding regions and 3' UTRs. We therefore infer that the 3' UTRs of the HMGI and HMG2 genes are subject to distinct evolutionary pressures. Our data, along with published data, also serve to highlight 26 amino acid positions where HMG1 and HMG2 are distinctly conserved, and we note that trout HMG-T conforms to the HMGI paradigm at most of these diagnostic positions.

INTRODUCTION

The high-mobility group (HMG) proteins, which were historically defined as a group due to a number of shared biochemical and physical properties, include three distinct families (HMGI/2, HMG14/17 and HMGI/Y) of nuclear proteins that are each prevalent and evolutionarily well conserved in higher eukaryotes (for an informative review,

Correspmldence to: Dr. J.B.E. Burch, Institute for Cancer Research, 7701 Burholme Ave., Philadelphia, PA 19111 (U.S.A.) Tel. (215)728-3696; Fax (215)728-3574. Abbreviations: a~, amino acid(s); bp, base pair(s); HMG, high-mobility group; HMG, gene (DNA) encoding HMG protein; kb, kilobase(s) or 1000 bp; Myr, 106 years; nt, nucleotide(s); ORF, open reading frame; SDS, sodium dodecyl sulfate; SSC, 0.15 M NaCl/0.015 M Na3.citrate pH 7.6; UTR, untranslated region; u, unit(s); UWGCG, University of Wisconsin Genetics Computer Group; VBP, vitellogenin gene binding protein.

see Bustin et al., 1990). HMGI/2, which are implicated in such diverse events as DNA replication, nucleosome assembly and transcription, have tripartite structures comprised of two related DNA-binding domains (here denoted HMG1/2-boxes to underscore the fact that this motif is confined to the HMG1/2 family of HMG proteins) and a C-terminal domain of 20 or more consecutive acidic residues. Interest in HMG1/2-boxes has broadened recently with the finding of related motifs in functionally diverse DNA-binding proteins such as nucleolar (Jantzen et al., 1990) and mitochondrial transcription factors (Parisi and Clayton, 1991), a yeast mating-type protein (Kelly et al., 1988) and the putative testis-determining factor (Sinclair et al., 1990). One unexpected finding to emerge from the analysis of mammalian HMGI eDNA clones is that the 3' UTRs are remarkably well conserved (Bustin et al., 1990). Based on a sequence comparison between a chicken HMG2 eDNA clone (that we now report) and a previously reported pig

252 from this comparison that these two 3' UTR are remarkably similar. Indeed, when the chicken HMG2 3' UTR sequence was used to search the database, the pig HMG2 3' UTR was identified as being the most similar. Since mammals and birds have evolved independently for about 300 Myr (McLaughlin and Dayhoff, 1972) we infer that there must be appreciable selective pressure to conserve both coding and noncoding sequences in the HMG2 transcription unit. Obviously, this information could be relevant to controls that operate at the DNA or RNA levels or both. To evaluate further the significance of the apparent conservation of 3' UTR for the chicken and pig HMG2 eDNA clones, a computer program was used to search for regions of similarity between a representative sample of both HMG2 and HMGI eDNA sequences. As with the analysis presented in Fig. 2, O R F and 3' UTR sequences were analyzed separately. As indicated by the data shown in the lower left portion of Fig. 3, the ORFs for all of the HMGI and HMG2 eDNA clones examined were found to have extended sequence similarity, reflecting the fact that these proteins are well conserved. The 3' UTR comparisons summarized in the upper right part of this figure quantify both the high degree of extended sequence similarity seen above in Fig. 2 for the pig and chicken HMG2 eDNA clones (75?/o over 374 bp), as well as the even higher degree of extended sequence similarity previously noted among the mammalian HMGI cDNA clones (94% or higher over 648 bp), In contrast, only small regions (less than 80 bp) of similarity were identified between the 3' UTRs of HMG2 and HMGI cDNAs. Thus it would appear that there has

HMG2 cDNA clone (Shirakawa et al., 1990) we conclude that HMG2 3' UTRs are also well conserved. A comparison between the predicted chicken H M G 2 aa sequence and published H M G 1/2 sequences also allows us to identify aa positions where HMG1 and HMG2 are distinctly conserved.

EXPERIMENTAL AND DISCUSSION (a) Isolation and sequence of a cDNA clone for chicken HMG2 A phage clone that we isolated from a 2gtl 1 chicken embryo eDNA library on the basis of its harboring a eDNA insert for the VBP leucine-zipper transcription factor (lyer et al., 1991) was found to contain in addition an unrelated insert, the sequence of which is shown in Fig. 1. That this eDNA clone happens to encode chicken HMG2 was indicated by a search which revealed that the most similar sequence in the database is a pig HMG2 eDNA clone. A comparison of the coding regions for the chicken eDNA clone and the pig HMG2 eDNA clone (Shirakawa et al., 1990) is shown in Fig. 2A. Aside from the 3' end of the ORF, where the chicken clone has a small in-frame deletion relative to the pig HMG2 eDNA clone, these two sequences are obviously strikingly similar.

(b) The 3' UTR of the chicken HMG2 eDNA clone is similar to the 3' UTR of a pig HMG2 eDNA clone but distinct from the 3' UTRs found in HMG! eDNA clones An optimal alignment of the 3' UTRs of the chicken and pig HMG2 eDNA clones is shown in Fig. 2B. It ts clear

M ~ K

2

GGGC~ACCCCAACAAGCCGCG~GGCAA~ATGTCCTCGTA~:~C~T~CTTCGTGCA~ACGTGCCGc~G~AGCAC~AAGAAGCACCCGGA G D P N K P R G K M ~ S Y A Y F V Q T C R E E H K K K H P D

1B0 3~

CTCG~?~AGTCAACTTCGCCGAGTTCTCGCGG~GTG~TCGGAGCGGTGG~GACGATGT~CAGC~GGAGAAAGGGAAGTTCGAG~AGAT S S V N F A E F S R K C ~ E R W K T M S S K E K G K F E E M

270 62

GGCC~GG~AGACA~GCTCGCTACGACAGGGAGATG~G~CTATGTTC~TcCCA~GGAGAG~GAAGGGA~GAAGAAGGACC~AA

360 92

A

K

G

D

K

A

R

Y

D

R

E

M

K

N

Y

V

P

P

K

G

E

K

K

G

K

K

K

D

P

N

CG~TCCT~GACCACCATCTGCATTCTT~CTTTT~TGTTC~G~CAc~GT~C~ATC~AAATGAT~AT~cCGGCTTGT~TATCGG A

P

K

R

P

P

S

A

F

F

L

F

C

S

E

H

R

P

K

I

K

N

D

H

P

G

L

S

I

G

AGACA~GGC~AGAAGTTGGGTGAAATGTGGTCTG~CAGT~GGCCA~AC~GC~CCTAT~AACAG~GGCTGC~AG~T~%GGA D

T

A

K

K

L

G

E

M

W

S

E

Q

L

A

K

D

K

Q

P

Y

E

Q

K

A

A

K

L

K

E

GA~TATGAAAAGGATATTGCAGCATATCGTGCCAAGAG~GAGTGATGCAGG~GGGCC~AGGTAG~CTGCAGGAT~T~GAA K

Y

E

K

D

!

A

A

Y

R

A

K

S

K

S

D

A

G

R

K

G

P

G

R

P

A

G

S

K

K

GA~GCAGAACCAG~GAGGAGGAGGAGGAGGAGG~GATGAGG~GAGGAGGAAG~G~GAGGATGAGGAGTA~TGAATGCCCTG~T K A E P E E E E ~ E E E D E E E E E E E E D E E •

450 122 540 152 630 182 720

GT~AGATTTATGCTCG~ATCC~TATTTTGCTAA~TGTGAACTCAAGTGCAGCT~ATTGTTAGCTTCAGTATAAAAAT~TGTACAG

810

ATTGTGTATAGGTAATGTGGTT•TTTGTAGGGAGACCTCTATTTCTAAGTAAGTAGTAG•TGAGTG•TGCTACAGTTCAGTTTTAAAAGA

900

AG~G~GTGG~TGTTTCTGCATATTT~TGGTTTTGTTGA~TTCAGTG~TGATAGCCAGGTGTTTTTTTTTTTAGCTAAGT~T

990

~AGGAGGTAGCTT~GCTGATCTTATATTTTGTAGTATATTCATTGTATTACTTTTTTTT~C~TTTTGTAATAAAATGTTGGGAA

1080

TTC

1083

Fig. 1. A ;.gtI ! chicken embryoeDNA clone has the completeORF for chicken HMG2. The eDNA fragmentwas subcioned into a plasmid vector and dideoxy sequenced on both DNA strands. The chicken ttMG20RF encodes a 207-aa protein beginningat a Kozak (1984) sequence but the aa sequence is numbered beginning at the next codon since the N-terminal Met is likelycleaved off (Bustin et al.. 1990). The aa are aligned with the first nt of each codon. Asterisk represents a stop codon. Note that a polyadenylation signal (5'-AATAAA)is present near the 3' end (nt 1065-1070) of the eDNA. This sequence has been deposited in the EMBL, GenBank and DDBJ Nucleotide Sequence Databases under accession No. M83235.

253

A 03 ^¢C,C.GC~C,C~CCCC,~,CA~CCC.CG~C,C~A~÷CCrCGrACC~ 13~ Illll

illlll

Illlllilllllllll

Illlll

IIllllllll

"0

144 ATGGGGA~.GGGCC~CCCCAACA~GCCGCGGGGCAAGATGTCCTCCTACC, C 193

-~ Q-

-~ 0_

om

I

133 GTACY'I~GTGCAG&CGYGCCGCGAGG&GCAC.~Gi~GAAGC&CCCGGACT 182 194 CTTCTYCGTGCAGACCTGCCGGGAGGAGCACAAG/OkG.q~O.C&CCCCGATT243 183 CGTC>CN~CTTCGCCGAGTTCTCGCGG~AGTGCTCGG~GCGGTGGA#,G232 ttt t~t~ ttt,l~**~, t,t i~tt~ Jolt t t t t t

244 CCTCGGTC~qCTTCGCCGAGTTCTCCAAGAAGTGCTCCGAGCGATGGkAG 293

CHKHMG2 83 - 706; 707 - 1076

PIG I:-IMG2 144 - 776; 777 - 1153

233 ACG&TGTCC~C~AAGGAGM~GGG/I~GTTCGAGGAGATG~¢~GGGA~d~ 202

~94 ~C¢~TGTCTGCC~GG~TCC~TrT~T~T~C,C~TC~ 343 203 C&~GCT~GCT&CG&C&GGGAG~TGAAG~ACT&TGTTCCTCCCAJ~GGAG 332

Illll Jilt I Illltl l i t tt Ili I J | l t l II 344 CAAAGCTCC~TATGACCGGGAG&TGAAAAATTACGTCCeT~¢¢AAGGGTG 393 333 AGAAGJtOkGGGNkAG.qAGNkC,GACCCCNkCGCTCCTARAAC.O,CC&CCATCi 382 I Ill II I I I II i l IIII I II I I II IIII 394 ACN~GAAGGGCAAGAAAkAAGATCCCAATGCTCCCA/U~GGCCTCCATCT 443

303 GCATTCTTCCTTTTCTGTTCCGAACACCGiCC~AAAATC~d~ATGATC~432 It

II

Ill

I

II

It

II

II

II

Ill

lilt

It

PIGHMG1 9 - 656; 6 5 7 - 1250

B O V H MG 1 119 - 766; 7 6 7 - 1 2 3 6 HUMHMG1 53 - 700; 701 - 1089

l

444 GCCTTCTTCCTGTTTTGCTCTG~CATCGCCC/~G&rC~GTG~CA 493 433 TCCCGGCTrGTCTATCGG~GACACGGCAA~GAAGTTGGGiG~JO, ATGTGGi 402

o 4 CCCT~TT~TCC~T~T~CTGC~',~TTr.,G~T~TGTG~T

~4~

483 CTG~CAGTTGGCC~GAC~GC~C¢CIATG&ACAGA~GC~TGC~G 532 544 CTGAGC>CAGCCAA~G~TAAACAACCGTATG&&CAGA~.GCAGCTA~G 593 533 CTA/O,C~AG~.TATG~AGG&TATTGCA~CATATCGTGCCAAGAGCM~582 594 CT/Id~AGGkAAAAT&TGAAAAGGATATTGCTGCATACCGTGCCAAGGGT,~ 643 583 GRGTGATGCAGG/O,N,O.AGGGCCCAGGTAGGCCTGCAGGATCT/OkGNkG~632 l till llilt I lilllll II II i i i i l i l l l l t i 644 GGGTG/~GCAGGJld~GA~GGGCCCTGGCAGGCCAACAGGCTCTAAG,qAG~ 693

694 AGN%TG&ACCAGAAG&TGAGGAGGAAGAGGAAGAGG&GGNm, GAAGATGAA 743 674 G~AG&GG&GGN~GAAGa~GAGGATGAGGAGTN~706 il

lli*

liliillil

II

llltliil

ill

744 G~CG&GG&GGN~GAAGaTGF.~GATGAGGAAT~776

B

ORF x ORF

Fig. 3. A matrix comparison of sequence similaritiesbetween coding and 3' UTR for HMGI and HMG2 cDNAs. The BESTFIT program (Devereux et al., 1984) was used to evaluate the degree of similarity between pairwise combinations of HMG! and HMG2 eDNA clones. The chicken HMG2 eDNA sequence is fi'om Fig. 1, whereas the other eDNA sequences are taken from published reports: pig HMG2 (Shirakawa et al., 1990), pig HMGI (Tsuda et al., 1988), cow HMGI (Kaplan and Duncan, 1988) and human HMG1 (Wen et al., 1989). The abbreviations used to denote the chicken, bovine and human sequences are CHK, BOV and HUM, respectively. Data from the ORF comparisons are shown below the diagonal (lower left); data from the 3' UTR comparisons are shown above the diagonal (upper right). Percent similarityis noted for each pair and the size (in bp) of the region identified by the computer is indicated in parentheses. The data in the upper portion of the figure indicate that the 3' UTRs for the HMG2 eDNA clones are well conserved, a~ are the 3' UTRs for the HMG! cDNA clones. Moreover, the data indicate that these two classes of UTR sequences are distinct.

744 A.IkT&TTTTGA~TAAGFdS, TGTG&ACTCAACTGC&GCTC&TTGTT&GCTTCAG793 i

IIIliil

li*lllllilil

II

IIIIIIltiillll*!iiillJ

026 aTTaTTTTGCTAaG~TCTC~TTC^aGTGCAGCTCkqTaTraGCTTC^Ge75 ~94 TaT~aT~TCTa¢aGa.÷TCTCTaTaCGT~TGTCGTTeTTT~TAC~ e42 lillllll

IIIIIIII

II

Itl

III

IIII

! IIIllitlll

I

076 TATAA/~k&.CTGTAC&G&TTTTTGTAT&~CT/~TAAG&TTCTTTGTAGAG924 843 AGACCTCTATTTCTAA...~GTAAGTAGTAGCTG>GC~'GC ....... I I

II

II

!It

i I I

I lilt

,I

~ 080

I

925 AA/~TACTTTTTTTAAA>GCAGGTTGCAGCTTTTTGAG~C~GCTACTAC 974 001 .TACAGTTCAGTTTTN~AAGAAGAAGTTGIGGf~TGTTTCTGCAT&TTTA 929 lilllll I Ill lilt lit lilllllill lilllll 9?5 &TACA~TT.AG~TTTTN~AC.,CTTCTGATGTTG~kTGTTTCT~T&TTTA1023 930 AT~TTTTGIT~AAATTC>G,SaqTG&TAC,~CAGGTGTTiTTTTTTTTAG979 lllllil I I I Ill II tl I II 1024 ATGGTTTCTTTAATTTCTTGTGTATGGTAAC ................. AG 1056 980 CTA/~T~J~TAAAGGAGGiA~CTTAATACJ~TGATCTTAiATTTTGTAGI1029 1057 C/~L~,CTCGTAGAAATAAG...T&TCA&TAC, CAAATTTTGGGTTTTCTAGA1103 1030 ATATTCATTGTATTACTTTiTTTTAACAAiTTTGTAATA~TGTTG1076 It I t I II tillll i ttttttittttiltlt it

1104 &TGTTGCAT...TTTGTTTTTTAAAAAAATTTTGTAATAAAATT&TG 1147

Fig. 2. The BESTFIT program from the UWGCG software package (Devereux et al., 1984) was used to align nt sequences of chicken HMG2 cDNA with the pig HMG2 cDNA (Shirakawa et al., 1990). Comparisons of the ORFs (panel A) and 3' UTR (panel B) are shown with the chicken sequence on the upper line in each case. The numbering of the chicken sequence is from Fig. 1, whereas the numbering of the pig sequence is taken from the database (GenBank accession No. J02895).

been evolutionary pressure to maintain distinct 3' U T R s for H M G I and H M G 2 m R N A s .

(c) A comparison of tbe deduced aa sequences of HMG! and HMG2 highlights residues that serve to distinguish these highly conserved proteins As noted above, H M G 1 and H M G 2 are very similar proteins. However, since all but one of the higher eukaryotic H M G 1 / 2 sequences have been obtained from mammais, c o m p a r i s o n s between these sequences have been limited in the range o f information they offer. Moreover, since only one H M G 2 aa sequence has been reported, the significance o f differences in the deduced H M G 1 and H M G 2 aa sequences was difficult to gauge. The H M G 2 sequence shown in Fig. 1 is thus useful in that it is only the second H M G 2 sequence available, as well as being only the seco n d H M G 1 / 2 sequence determined for a n o n m a m m a l i a n higher eukaryote. A comparison o f the deduced chicken H M G 2 aa se-

254 (3) proximal to the acidic aa tail. And fourth, it is clear that trout HMG-T (Pentecost et al., 1985) conforms to the HMG1 paradigm at most of the diagnostic positions.

quence with pig HMG2 and mammalian HMG1 aa sequences highlights a number of aa positions where the HMG1 and HMG2 sequences are distinctly conserved as well as positions where differences do not strictly correlate with either protein (Fig. 4). Regarding the former set, these can be further classified with respect to whether the changes involve nonconservative aa (marked by circled asterisks, of which there are 15) or conservative aa (marked by asterisks, of which there are eleven). Several points are noteworthy with respect to the distribution of putative diagnostic residues for HMGI/2. First, each of the three structural domains of HMG1/2 harbor diagnostic residues. Second, the diagnostic aa in the two HMG 1/2-boxes do not map to analogous positions, except in one case. Third, the diagnostic positions involving nonconservative aa changes tend to cluster in three regions: (/)the link region between the two HMG1/2-boxes, (2) a region (aa 133-138) within HMG1/2 BOX B and

(d) H M G 2 is encoded by a single-copy gene in chickens and ItMG2 mRNA levels are downregulated during liver ontogeny

Evidence that HMG2 is encoded by a single-copy chicken gene is provided by the Southern-blot analysis shown in Fig. 5B. These data appear to be in conflict with the results of a biochemical analysis which indicated that two HMG2-1ike proteins are produced in chicken thymus and erythrocytes (Mathew et al., 1979). However, this discrepancy may simply reflect the fact that very stringent hybridization conditions were used in our Southern-blot analysis, as evidenced by the lack of cross-reactivity with HMGI sequences. The Northern blot shown in Fig. 5A indicates that a

H M G 1/2 B O X A i

Z.o

'"

.

LINK

5v

,.

.

'

e ez

Cll~2 ,_.

PIGHM~I BOY TRTHMGT

•

t/i

~

-R

'

~

~

z,c

a

DL ,-

~

~Z~l

~,,

H M G 1/2 B O X B .

sZ.o

.

,,;o (zz

¢RF~tG2

e.

,,po

.

.... ~'~ ~ .......

PIG nOV [.trim

'

TRTHMGT

"

s

z

~

QV ~ T

~

K~W

v" ~iZ sz~

"T

DOMAIN C r

170

190

c,~2

]KSKSI~G~r~r.,~a~;S~~¢-~.=.._-

~ ~nv.,z

.or

I a P ~

~

~

zva~nooz

Fig. 4, A comparison of deduced HMGI and HMG2 aa sequences reveals a number of residues where different aa are conserved for HMG I and HMG2. The data in this figure are taken from the references noted in the legend to Fig. 3 and the same abbreviations are used. In addition, the trout (TRT) HMG-T sequence (Pentecost et al,, 1985) is shown, Note that the entire aa sequence of chicken HMG2 is shown, whereas only distinct residues are shown for the other HMGI/2 sequences, Positions where a single aa is used for HMG:~' and a distinct single aa is used for H M G I are indicated by asterisks: the subsets that involve nonconservative aa differences are circled, Note that troat HMG-T has HMG l-like residues at almost all of the highlighted positions. The tripartite structure of HMG 1/2 is also indicated. Those aa that are conserved between the two DNA-binding domains (HMG l/2-boxcs A and B), as well as the two yeast HMG1/2-1ike proteins, NHP6A and NHP6B (Kolodrubetz and Burhum, 1990) are marked beneath the sequences with blackened (identical aa) and open (conservative aa changes) triangles. The region between the two DNA-binding domains and the acidic C-terminal tail region are denoted LINK and DOMAIN (3, respectively.

255

A

ACKNOWLEDGEMENTS

B

kb

~ORIGIN

j23.1 B

9.4 6.6

-"%4.4 ~

I i

2.3 2.G

--18S HM(;;2 ~

~!

We th,ank Leonard Cohen, John Pehrson and Alfred Zweidler for constructive comments on the manuscript and Judy Herman and Kathy Truesdale of the Secretarial Services at the Fox Chase Cancer Center for help in preparing the manuscript. This work was supported by NIH grants 35535 and CA-06927, a grant from the Pew Charitable Trust and an appropriation from the Commonwealth of Pennsylvania.

i{~i!~!: '

REFERENCES

GAD--

Q,Q,W,~

, ~,,~

1 2 3 4 5 6

123

Fig. 5. Northern- and Southern-blot analysis of chicken HMG2. The Northern blot shown in the upper portion of panel A was hybridized to a random-labeled chicken H M G 2 0 R F probe to reveal the complexity and relative abundance of HMG2 mRNA in 5-#g aliquots of liver poly(A) + RNA isolated at day 15 of embryonic development (lane I), day I post hatching (lane 2) or one, two, four or six weeks post hatching (lanes 3-6). The top of the gel and the positions of the resolved 18S and 28S ribosomal RNA species are indicated. The Northern-blot strip at the bottom of panel A was hybridized with a random-labeled GAD probe (for glyceraldehyde-3-phosphate dehydrogenase) (Dugaiczyk et al., 1983) to confirm that equal amounts of RNA (in this case I pg) were loaded in each lane. Both Northern blots were washed under conditions of high stringency (two 30-min washes at 22~C in 2 x SSC/I o~ SDS followed by two 30-min washes at 65 ° C in 0.1 x SSC/1 ~/0 SDS). The autoradiograms were exposed to film between two intensifying screens at -70 ~C for either three days (the HMG2 blot) or 16 h (the control blot). The Southern blot in panel B was hybridized to the same random-labeled ORF probe from the HMG2 eDNA ,:lone. The restriction enzymes used to digest the 10-/~g aliquots of chicken genomic DNA were BamHl (lane I), EcoRI (lane 2) and Hindlll (lane 3). The blot was washed using the same conditions as noted above for the Northern blots and was exposed at -70°C between intensifying screens for 16 h.

single mRNA species for HMG2 is expressed in the chicken liver and that this mRNA decreases in its abundance during liver ontogeny. In the future it should be interesting to determine the relative levels of HMG2 and HMGI mRNAs in a variety of tissues as a function of ontogeny. Given that the 3' UTR sequences of HMGI and HMG2 appear to be subject to distinct evolutionary pressures, one might expect that this feature will prove to be relevant to the discordant synthesis of H M G 1 and HMG2 protein that is seen for a number of terminal differentiating tissues (Seyedin and Kistler, 1979).

Bustin, M., Lehn, D.A. and Landsman, D.: Structural features of the HMG chromosomal proteins and their genes. Biochim. Biophys. Acta 1049 (1990) 231-243. Devereux, J., Hacberli, P. and Smithies, O.: A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12 (1984) 387-395. Dugaiczyk, A., Haron, J.A., Stone, E.M., Dennison, O.E., Rothblum, K.N. and Schwartz, R.J.: Cloning and sequencing of a deoxyribonucleic acid copy of glyceraldehyde-3-phosphate dehydrogenase messenger ribonucleic acid isolated from chicken muscle. Biochcmi.qry 22 (1983) 1605-1613. lyer, S.V., Davis, D.L., Seal, S.N. and Butch, J.B.E.: Chicken VBP, a leucine zipper transcription factor that binds to an important control element in the chicken vitellogenin II promoter, is related to rat DBP. Mol. Cell. Biol. 11 (1991)4863-4875. Jantzen, H.M., Admon, A., Bell, S.P. and Tjian, R.: Nucleolar transcription factor hUBF contains a DNA-binding motif with homology to HMG proteins. Nature 344 (1990) 830-836. Kaplan, D.J. and Duncan, C.H.: Full length cDNA sequence for bovine high mobility group 1 (HMGI) protein. Nucleic Acids Res. 16 (1988) 10375. Kelly, M., Burke, J., Smith, M., Klar, A. and Beach, D.: Four matingtype genes control sexual differentiation in the fission yeast. EMBO J. 7 (1988) 1537-1547. Kolodrubetz, D. and Burhum, A.: Duplicated NHP6 genes of Saccharomyces cerevisiae encode proteins homologous to bovine high mobility group protein 1. J. Biol. Chem. 265 (1990) 3234-3239. Kozak, M.: Compilation and analysis of sequences upstream from the translational start site in eukaryotic mRNAs. Nucleic Acids Res. 12 (1984) 857-872. Mathew, C.G.P., Goodwin, G.H., Gooderham, K., Walker, J.M. and Johns, E.W.: A comparison of the high mobility group non-histone chromatin protein HMG2 in chicken thymus and erythrocytes. Biochem. Biophys. Res. Commun. 87 (1979) 1243-1251. McLaughlin, P.J. and Dayhoff, M.O.: Evolution of species and proteins: a time scale. In: Atlas of Protein Sequence and Structure, Vol. 5, The National Biomedical Research Foundation, Silver Spring, MD, 1972, pp. 47-52. Parisi, M.A. and Clayton, D.A.: Similarity of human mitochondrial transcription factor 1 to high mobility group proteins. Science 252 (1991) 965-969. Pentecost, B.T., Wright, ,i.M. and Dixon, G.H.: Isolation and sequence of cDNA clones coding for a member of the family of high mobilib group proteins (HMG-T) in trout and analysis of HMG-T mRNA's in trout tissues. Nucleic Acids Res. 13 (1985) 4871-4888. Seyedin, S.M. and Kistler, W.S.: Levels of chromosomal protein high mobility group 2 parallel the proliferative activity of testis, skeletal muscle, and other oigans. J. Biol. Chem. 254 (1979) ll_"t,4-11271.

256 Shirakawa, H., Tsuda, K. and Yoshida, M.: Primary structure of nonhistoric chromosomal protein HMG2 revealed by the nucleotide sequence. Biochemistry 29 (1990) 4419-4423. Sinclair, A.H., Berta, P., Palmer, M.S., Hawkins, J.R., Griffiths, B.L., Smith, M.J., Foster, J.W., Frischauf, A.M., Lovell-Badge, R. and Goodfellow, P.N.: A gene from the human sex-determining region encodes a protein with homologyto a conserved DNA-binding motif. Nature 346 (1990) 240-244.

Tsuda, K., Kikuchi, M., Mori, K., Waga, S. and Yoshida, M.: Primary structure of non-histone protein HMG1 revealed by the nucleotide sequence. Biochemistry 27 (1988) 6159-6163. Wen, L., Huang, J.K., Johnson, B.H. and Reeck, G.R.: A full length human placental eDNA clone for nonhistone chromosomal protein HMG-1. Nucleic Acids Res. 17 (1989) 1197-1214.