MOLECULAR AND CELLULAR BIOLOGY,
0270-7306/01/$04.00⫹0 DOI: 10.1128/MCB.21.20.7054–7064.2001
Oc . 2001, p. 7054–7064 Vol. 21, No. 20
Copy igh © 2001, Ame ican Socie y o Mic obiology. All Righ s Rese ed.
Hp 1 Is P e e en ially Requi ed o T ansc ip ion o Ei he Long o
G⫹C-Rich DNA Sequences in Saccha omyces ce e isiae
SEBASTIA
´N CHA
´VEZ, MARI
´A GARCI
´A-RUBIO, FE
´LIX PRADO, AND ANDRE
´S AGUILERA*
Depa amen o de Gene´ ica, Facul ad de Biologı´a, Uni e sidad de Se illa, 41012 Se ille, Spain
Recei ed 21 May 2001/Re u ned o modi ica ion 26 June 2001/Accep ed 24 July 2001
Hp 1 o ms, oge he wi h Tho2, M 1, and Thp2, he THO complex, which con ols ansc ip ion elonga ion
and genome s abili y in Saccha omyces ce e isiae. Mu a ions in genes encoding he THO complex con e s ong
ansc ip ion-impai men and hype ecombina ion pheno ypes in he bac e ial lacZ gene. In his wo k we
demons a e ha Hp 1 is a ac o equi ed o ansc ip ion o long as well as GⴙC- ich DNA sequences. Using
di e en lacZ segmen s used o he GAL1 p omo e , we show ha he nega i e e ec o lacZ sequences on
ansc ip ion depends on hei dis ance om he p omo e . In pa allel, we show ha ansc ip ion o ei he a
long LYS2 agmen o he S. ce e isiae YAT1 GⴙC- ich open eading ame used o he GAL1 p omo e is
se e ely impai ed in hp 1 mu an s, whe eas ansc ip ion o LAC4, he Kluy e omyces lac is o holog o lacZ bu
wi h a lowe GⴙC con en , is only sligh ly a ec ed. The hype ecombina ion beha io o he DNA sequences
s udied is consis en wi h he ansc ip ional de ec s obse ed in hp 1 cells. These esul s indica e ha bo h
leng h and GⴙC con en a e impo an elemen s in luencing ansc ip ion in i o. We discuss hei ele ance
o he unde s anding o he unc ional ole o Hp 1 and, by ex ension, he THO complex.
The con ol o genome s abili y is essen ial o ensu e main-
enance o gene ic in o ma ion in all cells o a li ing o ganism.
Dys unc ion o his con ol causes mu a ions and ch omosomal
abe a ions ha can gi e ise o loss o gene unc ion, cell
dea h, o i e e sible changes in he cell p og am.
Gene ic ecombina ion is equi ed o mi o ic DNA epai
and o p ope meio ic ch omosome seg ega ion. In addi ion,
i may also be esponsible o p ocesses o gene ic ins abili y. A
numbe o animal diseases, including cance , o igina e by
e en s o mi o ic ecombina ion be ween epea s ha lead o
ch omosomal abe a ions (34). Se e al elemen s ha e been
desc ibed o enhance mi o ic ecombina ion, including DNA
damage, eplica ion de ec s, al e a ion o ch oma in s uc u e,
and ansc ip ional ac i i y ( e iewed in e e ence 3). Ikeda
and Ma sumo o (26) i s desc ibed he in luence o ansc ip-
ion on ecombina ion showing ha ecombina ion o phage ␭
was s imula ed by ansc ip ion. In yeas , he i s example o
ansc ip ion-associa ed ecombina ion was he inding ha a
ho spo o ibosomal DNA ( DNA) ecombina ion, HOT1, was
dependen on RNA polyme ase I-d i en ansc ip ion (55, 60).
Thomas and Ro hs ein (56) ex ended ansc ip ion-induced
ecombina ion o sequences ansc ibed by RNA polyme ase II
(RNAPII). Addi ional examples o RNAPII-dependen e-
combina ion ha e been subsequen ly desc ibed in yeas (21, 36,
50) and mammalian cells (37, 57). Special men ion mus be
made o he modula ion o ecombina ion a he immunoglob-
ulin loci, as bo h V(D)J ecombina ion (7, 31, 38) and class
swi ching (15) a e posi i ely con olled by ansc ip ion.
A gene linking ansc ip ion and genome ins abili y in Sac-
cha omyces ce e isae is HPR1,ashp 1 mu an s show bo h in-
c eased le els o ecombina ion be ween di ec epea s and
ch omosome loss (2, 49) as well as s ong ansc ip ional de-
ec s (11, 44, 67). De ailed cha ac e iza ion o hese de ec s has
shown ha he absence o Hp 1 causes impai men o an-
sc ip ion elonga ion. The in ensi y o such a ansc ip ional
impai men depends on he ansc ibed DNA sequence (11).
The e is a close co ela ion be ween he eluc ance o a DNA
sequence o be ansc ibed in hp 1 cells and he abili y o such
a sequence o p omo e ecombina ion when inse ed be ween
di ec epea s (11, 44).
Biochemical and gene ic analyses ha e con ibu ed o iden-
i ying se e al ac o s ha pa icipa e in RNAPII-media ed
ansc ip ion elonga ion ( e iewed in e e ence 14). Acco ding
o hei unc ion in ansc ip ional elonga ion, hese ac o s can
be classi ied in di e en g oups. TFIIS p e en s RNAPII a es
and induces nascen ansc ip clea age ( e iewed in e e ence
64). Some o he ac o s, like TFIIF, CSB, ELL, and elongin,
in luence elonga ion by supp essing he pausing o RNAPII (5,
46, 52, 53). P-TEFb s imula es ansc ip ion elonga ion in e-
sponse o ansac i a o s ( e iewed in e e ence 45) by an ag-
onizing nega i e ac o s like DSIF and NELF (22, 61, 65).
Finally, some ansc ip ion elonga ion ac o s like FACT and
Elonga o play a ole in acili a ing RNAPII-d i en ansc ip-
ion on ch oma in empla es (39, 40).
Al hough hp 1⌬cells a e a ec ed in ansc ip ion elonga ion
in i o, Hp 1 does no seem o be physically associa ed wi h
any o he known elonga ion ac o s. I has been demons a ed
ha Hp 1 is physically p esen in a new o m o RNAPII ho-
loenzyme ha has been p oposed o espond o p o ein kinase
C-media ed signal ansduc ion (10). Hp 1 o ms he THO
complex in i o oge he wi h he p oduc s o he THO2,
MFT1, and THP2 genes (12). The absence o any o he ou
p o eins con e s simila pheno ypes o ansc ip ional elonga-
ion impai men and hype ecombina ion, indica ing ha he
THO complex is a unc ional uni in gene exp ession and
genome s abili y (12). Howe e , he way THO con ols hese
p ocesses emains obscu e.
* Co esponding au ho . Mailing add ess: Depa amen o de Ge-
ne´ ica, Facul ad de Biologı´a, Uni e sidad de Se illa, A d. Reina Me -
cedes 6, 41012 Se ille, Spain. Phone: 34 954 557 107. Fax: 34 954 557
104. E-mail: [email p o ec ed].
7054
on July 4, 2017 by USE/BCTA.GEN UNIVERSITARIAh p://mcb.asm.o g/Downloaded om
As ele ance o he THO complex in ansc ip ion depends
on he ansc ibed DNA sequence, in es iga ion o he ea u es
ha make ansc ip ion o a pa icula DNA sequence depen-
den on Hp 1 can p o ide some clues o unde s anding i s
p ecise unc ion. We ha e ound ha ansc ip ional impai -
men in hp 1 occu s p ima ily in long ansc ip ion uni s as well
as in DNA sequences wi h a high G⫹C con en used o he
GAL1 p omo e . The ele ance o hese esul s o unde -
s anding he unc ional ole o he THO complex is discussed.
MATERIALS AND METHODS
Yeas s ains and plasmids. The wo isogenic yeas s ains used in his s udy
we e W303-1A (MATaade2-1 can1-100 his3-11,15 leu2-3,112 p1-1 u a3-1) and
U768-4C (MATaade2-1 can1-100 his3-11,15 leu2-3,112 p1-1 u a3-1 hp 1⌬::
HIS3). All plasmids used a e monocopy CEN-based plasmids and a e lis ed in
Table 1.
Analysis o gene exp ession and ecombina ion. Fo he analysis o GAL1-
d i en exp ession, mid-log phase cells we e inocula ed wi h 3% glyce ol–2%
lac a e syn he ic medium a an op ical densi y a 600 nm (OD
600
) o 0.1. A e
16 h o incuba ion a 30°C, 2% galac ose was added and incuba ion was con in-
ued o ano he 8ha 30°C. Acid phospha ase ac i i y and mRNA le els we e
de e mined as desc ibed p e iously (11). I is impo an ha all ansc ip ion
analyses shown in his s udy, wi h ew excep ions, we e made in monocopy
CEN-based plasmids, which a e los a highe equencies in hp 1 e sus wild-
ype s ains (11). Howe e , since in all expe imen s cells we e g own unde he
selec ion condi ions o he plasmid, mo e han 90% o he hp 1 cells s ill
con ained plasmids. The e o e, all obse ed ansc ip ional e ec s a e no caused
by plasmid loss.
Fo exp ession analysis o EGT2,CDC48,KAR2,OLE1, and GOG5 cells we e
g own in yeas ex ac -pep one-dex ose (YEPD)- ich medium o an OD
600
o
1.0 and subsequen ly sampled. DNA p obes o No he n expe imen s we e
ob ained by PCR ampli ica ion using he ollowing pai s o p ime s: TCATTTC
GATACTCGGCCTAG and GCAGCATCAGAGCTAGTTGTG o EGT2;
AAACCACTTTTGGACGCCTC and TCTTGTCTCTCTTTGGAGCT o
CDC48; TTCAACAGACTAAGCGCTGG and CAATTTCAATACGGGTGG
ACA o KAR2; ATGCCAACTTCTGGAACTAC and CCGAAAGTAACAAT
GGCAGT o OLE1; and TTGAAAACAGGTCATGCAGG and TGGGCTT
GTTGCTTCTTTTG o GOG5.
Recombina ion equencies we e calcula ed as he median o six independen
cul u es as p e iously published (43).
Mapping o MNase clea age si es. Yeas sphe oplas s and mic ococcal nucle-
ase (MNase) diges ions we e pe o med acco ding o Fedo and Ko nbe g (19)
wi h he modi ica ions o Cha´ ez e al. (13). Sphe oplas s p epa ed om mid-log
phase cul u es ans o med wi h p416GAL1lacZ and g own in he app op ia e
selec i e medium con aining 2% glucose o 2% galac ose we e lysed and imme-
dia ely diges ed wi h 6.25 o 800 mU o MNase. Fo naked DNA con ols,
genomic DNA was ex ac ed as p e iously desc ibed (28) and diges ed wi h 0.003
o 1.6 mU o MNase unde he same condi ions.
MNase-clea ed genomic DNA was diges ed wi h ei he EcoRI ( o he en-
dogenous GAL1 gene) o ClaI ( o he GAL::lacZ usion) and esol ed in 1.5%
aga ose. As in e nal size ma ke s, we used genomic DNA diges ed wi h SacIo
XbaI ( o he GAL1 p omo e used o lacZ). Fo he analysis o he endogenous
GAL1 gene, he p obe used was he 196-bp GAL1 agmen loca ed immedia ely
downs eam o he EcoRI si e and ob ained by PCR wi h he oligonucleo ides
ATTCGACAGGTTATCAGCAAC and TTAAACTTCTTTGCGTCCATC.
Fo he analysis o GAL1::lacZ he p obe used was he 202-bp lacZ agmen
immedia ely ups eam o he ClaI si e and ob ained by PCR wi h he oligonu-
cleo ides TCGTTGCTGCATAAACCG and TCGATAATTTCACCGCCG.
Miscellaneous. Se ial dele ions o he PHO5::lacZ usion cons uc s we e
cons uc ed using a double-s anded nes ed dele ion ki om Ame sham Pha -
macia. Published me hods we e used o RNA and DNA hyb idiza ions (13, 44).
RESULTS
T ansc ip ion impai men h ough lacZ caused by hp 1⌬is
no dependen on pa icula lacZ sequences bu on hei dis-
ance om he p omo e . T ansc ip ion o he Esche ichia coli
lacZ gene in S. ce e isiae is se e ely impai ed in hp 1 mu an s
TABLE 1. Plasmids
Plasmid Desc ip ion Sou ce o e e ence
pRS416 YCp ec o based on he URA3 gene 54
p416GALI-lacZ pRS416 con aining he lacZ egion used o he GAL1 p omo e 35
pSCh202 pRS416 con aining he PHO5 egion used o he GAL1 p omo e 11
pSCh212 pSCh202 wi h lacZ ansc ip ionally used o he 3⬘end UTR
a
o PHO5 11
pRS314-L YCp ec o pRS314, based on he TRP1 gene and con aining wo 598-bp LEU2 in e nal agmen s
epea ed in di ec o ien a ion, sepa a ed by a polylinke
43
pSCh211 pSCh212 wi h lacZ in opposi e o ien a ion This s udy
pSCh211⌬17-1 pSCh202 wi h a ⬃0.4-kb agmen o he 3⬘end o lacZ used o he 3⬘end UTR o PHO5 This s udy
pSCh229 pSCh202 wi h he i s 439 bp o lacZ ORF used o he 3⬘end UTR o PHO5 This s udy
pSCh226 pSCh202 wi h he HpaI-BssHII 447-bp agmen o lacZ inse ed in he 3⬘end UTR o PHO5 This s udy
pSCh251 pSCh202 wi h a 393-bp agmen o GAL1 ( om posi ion 843 o 1236 o he ORF) inse ed in he
3⬘end UTR o PHO5
This s udy
pSCh215 pRS416 con aining a 439-bp agmen o he 5⬘end o lacZ used o he GAL1 p omo e This s udy
pSCh213 pRS416 con aining he 447-bp HpaI-BssHII agmen o lacZ used o he GAL1 p omo e This s udy
pSCh216 pRS416 con aining he 352-bp P uII-EcoRI agmen o lacZ used o he GAL1 p omo e This s udy
pSCh218 pSCh202 con aining he i s 439-bp agmen o he 5⬘end o lacZ inse ed be ween he GAL1
p omo e and he PHO5 ORF
This s udy
pSCh219 pSCh202 con aining he 447-bp HpaI-BssHII agmen o lacZ inse ed be ween he GAL1
p omo e and he PHO5 ORF
This s udy
pSCh220 pSCh202 con aining he 352-bp P uII-EcoRI agmen o lacZ inse ed be ween he GAL1
p omo e and he PHO5 ORF
This s udy
pSCh221 pRS314-L con aining he i s 439 bp o lacZ ORF inse ed be ween he epea s This s udy
pSCh222 pRS314-L con aining he 447-bp HpaI-BssHII agmen o lacZ inse ed be ween he epea s This s udy
pSCh223 pRS314-L con aining he 352-bp P uII-EcoRI agmen o lacZ inse ed be ween he epea s This s udy
pSCh205 pRS314-L con aining he en i e lacZ gene inse ed be ween he epea s 11
pSCh227 pRS416 con aining he 3.7-kbp EcoRV agmen o LYS2 used o he GAL1 p omo e This s udy
pSCh230 pRS314-L con aining he 3.7-kbp EcoRV agmen o LYS2 inse ed be ween he epea s This s udy
pSCh255 pRS416 con aining he en i e LAC4 coding egion used o he GAL1 p omo e This s udy
pSCh254 pRS314-L con aining he en i e LAC4 coding egion inse ed be ween he epea s This s udy
pSCh247 pRS416 con aining he en i e YAT1 coding egion used o he GAL1 p omo e This s udy
pSCh248 pRS314-L con aining he en i e YAT1 coding egion inse ed be ween he epea s This s udy
a
UTR, un ansla ed egion.
VOL. 21, 2001 Hp 1-DEPENDENT TRANSCRIPTION ELONGATION 7055
on July 4, 2017 by USE/BCTA.GEN UNIVERSITARIAh p://mcb.asm.o g/Downloaded om
a he elonga ion le el. T ansc ip ion h ough PHO5 is no
app eciably a ec ed in hp 1 cells, bu i becomes sensi i e o
hp 1 when lacZ is used o PHO5 in a single ansc ip ion uni
(11). In o de o ind ou which s uc u al elemen s o se-
quence mo i s p esen in lacZ a e esponsible o his an-
sc ip ional elonga ion impai men , we cons uc ed se ial dele-
ions o he lacZ gene in he PHO5::lacZ ansc ip ional usion
unde con ol o he GAL1- egula ed p omo e . The esul ing
dele ions we e in oduced in o bo h wild- ype and hp 1 cells.
Yeas ans o man s g own in galac ose-con aining medium
we e hen assayed o acid phospha ase ac i i y. Exp ession
was clea ly lowe in he ans o man s ha bo ing PHO5-lacZ
usions han in hose con aining only PHO5, in bo h wild- ype
and hp 1 cells (Fig. 1A). The p esence o a lacZ agmen as
sho as 1 kb downs eam o PHO5 educed PHO5 exp ession
o 40% in wild- ype cells. Howe e , he educ ion was consid-
e ably s onge in hp 1, eaching ansc ip ion alues below
10% o hose o PHO5 alone (Fig. 1A). E en he sho es
usion, encompassing ⬇0.4 kb o he 5⬘end o lacZ, showed
educed le els o phospha ase ac i i y in he wild- ype (65%)
and, o a g ea e deg ee, hp 1 (35%) cells (Fig. 1A).
Se ial dele ions we e also made in a PHO5::lacZ usion
ca ying lacZ in an opposi e o ien a ion. A simila p o ile o
phospha ase ac i i ies was ob ained (Fig. 1B). Al hough all
usions showed lowe exp ession le els han PHO5 alone, he
nega i e ansc ip ional e ec was clea ly s onge in hp 1 han
in wild- ype cells. A ⬇0.4-kb lacZ agmen , o example, was
enough o educe he phospha ase ac i i y o unde 15% o he
le el shown by PHO5 alone in hp 1 s ains (Fig. 1B). Thus, he
wo end agmen s o lacZ we e able o impai ansc ip ion in
hp 1 cells.
The p e ious esul s sugges ha he e is no a pa icula
lacZ sequence esponsible o he ansc ip ional elonga ion
impai men caused by hp 1⌬. On he con a y, ansc ip ional
impai men could occu h ough any lacZ egion. To con i m
his and o show ha he nega i e e ec o hp 1⌬on acid
phospha ase exp ession eally akes place a he ansc ip ional
a he han pos ansc ip ional le el, we pe o med No he n
analyses o selec ed PHO5::lacZ- agmen cons uc s. We in-
se ed h ee di e en ⬇0.4-kb agmen s o lacZ co esponding
o he wo ends and he cen e o he gene (plasmids pSCh229,
pSCh226, and pSCh251; Table 1) immedia ely downs eam o
aPHO5 gene unde GAL1 con ol. Galac ose-induced an-
sc ip ion o he esul ing usion cons uc s was analyzed in
wild- ype and hp 1 cells. The esul s shown in Fig. 2 indica e a
subs an ial dec ease in he accumula ion o ull-leng h mRNA
o he h ee usion cons uc s in hp 1 cells (13 o 25% o he
wild- ype le els), whe eas only a mino e ec was obse ed
wi h PHO5 alone. In addi ion o he ull-leng h PHO5-lacZ
mRNA, a sho e ansc ip exhibi ing he same size as PHO5
was de ec ed in hp 1. The p esence o his sho e ansc ip
sugges s ha hp 1 cells ansc ibe poo ly h ough lacZ se-
quences,downs eam o he PHO5 open eading ame (ORF).
The same esul s we e ob ained when he lacZ agmen s we e
loca ed in he opposi e o ien a ion (da a no shown). To con-
i m ha his phenomenon was due o lacZ i sel and no o he
3⬘end o PHO5, we eplaced he lacZ segmen wi h a 416-bp
agmen o he 5⬘end o GAL1, a gene whose exp ession is
no a ec ed in hp 1 cells (67) (see Fig. 9A). No he n analysis
o he esul ing PHO5::GAL1⌬ usion was ca ied ou in wild-
ype and hp 1 cells (Fig. 2). A weake educ ion in accumula-
ion o ull-leng h mRNA was measu ed in hp 1 (60% o wild-
ype le els), and no sho ansc ip was de ec ed. This
con i ms ha he ansc ip ional de ec s o he PHO5-lacZ
usion cons uc s we e mainly due o he p esence o lacZ
agmen s in he ansc ip ion uni s.
Al oge he , hese esul s sugges ha he longe dis ance
be ween he p omo e and lacZ sequences he g ea e he
ansc ip ional elonga ion impai men . To es his possibili y,
we inse ed ups eam o PHO5 he same h ee lacZ agmen s
FIG. 1. Exp ession pa e ns o se ial dele ions o GAL1p ::PHO5-lacZ usion cons uc s. Acid phospha ase ac i i ies unde induced condi ions
o wild- ype (W303-1A) and hp 1 (U768-4C) s ains ans o med wi h lacZ-dele ed a ian s o plasmids pSCh212 (A) o pSCh211 (B) ha con ain
he en i e lacZ coding sequence used o PHO5 in he same and opposi e o ien a ions, espec i ely, unde he GAL1 p omo e . The a e age alue
and s anda d de ia ion o ou di e en ans o man s is shown o each s ain. Ve ical lines ac oss he lacZ sequences indica e he end poin s
o he dele ion cons uc s analyzed.
7056 CHA
´VEZ ET AL. MOL.CELL.BIOL.
on July 4, 2017 by USE/BCTA.GEN UNIVERSITARIAh p://mcb.asm.o g/Downloaded om
used p e iously. The esul ing ansc ip ional usions exhibi ed
simila ansc ip ion le els and pa e ns in wild- ype and hp 1
cells (Fig. 3). The e o e, he dis ance be ween lacZ and he
p omo e can modula e he nega i e e ec o he lacZ ag-
men s on ansc ip ion in hp 1 cells.
T ansc ip ion h ough long DNA sequences is nega i ely
a ec ed by hp 1⌬.The da a shown in Fig. 1 indica e ha he
longe he cons uc s con aining lacZ agmen s a e, he lowe
he ansc ip ional yield exhibi ed in bo h wild- ype and hp 1
cells. To e alua e he in luence o ansc ip leng h on he
ansc ip ional e ec o hp 1, we pu he same h ee abo e-
men ioned lacZ agmen s immedia ely downs eam o he
GAL1 p omo e . The kine ics o accumula ion o hese h ee
sho lacZ agmen s was iden ical in wild- ype and hp 1 cells
(Fig. 4A). Full-leng h mRNA accumula ed sho ly a e induc-
ion in bo h s ains and a simila le els in all h ee cons uc s,
in con as wi h he clea di e ence be ween he wild ype and
hp 1 shown by he en i e lacZ (Fig. 4A). The same esul s we e
ob ained when he lacZ agmen s we e cloned in he opposi e
o ien a ion (da a no shown). These esul s s ongly sugges
ha sho ansc ip ion uni s a e no impai ed by hp 1, e en i
hey con ain DNA agmen s ha hinde ansc ip ion in a
di e en con ex .
We ha e p e iously shown ha in hp 1 and o he mu an s
a ec ed in he THO complex, he abili y o a gi en DNA
segmen , like lacZ, o impai ansc ip ional elonga ion co e-
la es wi h hype ecombina ion o a di ec - epea sys em con-
aining ha segmen . This hype ecombina ion is ansc ip ion
dependen (11, 42, 44). We es ed, he e o e, he ecombina-
ion equency o di ec - epea sys ems con aining ei he one
o he h ee lacZ agmen s used in he p e ious expe imen s
lanked by wo leu2 epea s. In ag eemen wi h he absence o
e ec o he hp 1 mu a ion on ansc ip ion o such lacZ ag-
men s, we did no de ec a signi ican s imula ion o ecombi-
na ion when he agmen s we e loca ed be ween he leu2
epea s (Fig. 4B). Again, in his case he e was a clea di e -
ence be ween he sho lacZ agmen s and he en i e lacZ,
which s imula es ecombina ion be ween di ec epea s up o
200- old in hp 1 cells (11) (Fig. 4B).
The p e ious esul s sugges ha ansc ip leng h is an
impo an ea u e in de e mining he equi emen o Hp 1
in ansc ip ion. To con i m his, we cons uc ed compa able
ansc ip ion and ecombina ion sys ems con aining only yeas
long DNA sequences. We andomly chose a agmen om a
long yeas ORF, a 3.7-kb agmen o he S. ce e isiae LYS2
gene. We ei he used i o he GAL1 p omo e o inse ed i
FIG. 2. T ansc ip ion analysis o GAL1p ::PHO5, h ee di e en GAL1p ::PHO5-lacZ usion cons uc s, and a GAL1p ::PHO5-GAL1 usion in
wild- ype (W303-1A) and hp 1 (U768-4C) cells. (A) No he n blo analyses o PHO5-con aining mRNAs d i en om he GAL1 p omo e .
Plasmids used we e pSCh229, pSCh226, and pSCh211⌬17-1 (ca ying ⬇0.4 kb o he 5⬘end, middle pa , and 3⬘end o lacZ used o PHO5,
espec i ely), pSCh202 (ca ying he PHO5 gene), o pSCh251 (ca ying ⬇0.4 kb o he 5⬘end o GAL1 used o PHO5). Mid-log phase cells we e
cul u ed in 3% glyce ol–2% lac a e syn he ic comple e (SC)-U a medium and dilu ed in o iden ical esh media o an OD
600
o 0.3 and incuba ed
o 16 h. Galac ose (Gal) was hen added and samples we e aken o No he n analysis a di e en imes, as speci ied. A 0.9-kb EcoRV PHO5
in e nal agmen and a 589-bp 28S DNA in e nal agmen ob ained by PCR ( RNA) we e used as DNA p obes. (B) Kine ics o induc ion o
mRNAs as de e mined by quan i ica ion o No he n blo s in a Fuji FLA3000. The mRNA alues a e gi en in a bi a y uni s (A.U.) wi h espec s
o RNA le els. Fo any gi en cons uc , RNA le els a e ela ed o he wild- ype (w ) le els a 90 min, which was se a 100 o each panel.
VOL. 21, 2001 Hp 1-DEPENDENT TRANSCRIPTION ELONGATION 7057
on July 4, 2017 by USE/BCTA.GEN UNIVERSITARIAh p://mcb.asm.o g/Downloaded om
be ween he leu2 epea s o ansc ip ional and ecombina-
ional analyses, espec i ely. The new cons uc s we e in o-
duced in wild- ype and hp 1 cells (Fig. 5). Full-leng h mRNA
om he GAL1p ::LYS2 ansc ip ional usion accumula ed
sho ly a e ans e ing wild- ype cells o galac ose. Howe e ,
only a smea o incomple e ansc ip was de ec ed in simila
No he n expe imen s pe o med wi h hp 1 samples (Fig. 5A),
a e y simila pa e n o ha ob ained o he en i e 3-kb-long
lacZ in hp 1 (11) (Fig. 4A). As expec ed o a DNA sequence
ha canno be p ope ly ansc ibed in hp 1 cells, LYS2 p o-
mo ed a s ong hype ecombina ion in hp 1 when loca ed
be ween leu2 epea s (L-LYS2 sys em). The ecombina ion
equency eached in hp 1 (5%) is 60 imes highe han he
wild- ype le els, bu s ill 3- o 10- old lowe han ha o anal-
ogous sys ems con aining lacZ (11) (Fig. 4B). This esul sup-
po s ou hypo hesis o he in luence o ansc ip leng h on
hp 1 sensi i i y o ansc ip ion.
I he con ibu ion o Hp 1 o he accumula ion o long
ansc ip s ini ia ing a he GAL1 p omo e is no es ic ed o
he a i icial cons uc ha we ha e analyzed, we should expec
he genome-wide e ec o hp 1 o be mo e d ama ic on long
genes han on sho ones in highly exp essed genes. To es his
idea we analyzed he e ec o hp 1⌬on ansc ip ion o i e
endogenous ch omosomal genes wi h sizes anging om 0.5 o
3.1 kb. They we e selec ed because hey ha e high and com-
pa able exp ession le els in YEPD- ich medium in wild- ype
cells (be ween 10 and 14 mRNAs pe cell, acco ding o Hol-
s ege e al. [24]). The longes genes, EGT2,CDC48, and KAR2,
showed signi ican ly lowe exp ession le els in hp 1 han in he
wild ype. The sho es ones, OLE1 and GOG5, exhibi ed e en
highe exp ession le els in hp 1 han in he wild ype (Fig. 6).
As we a e no con olling ansc ip ion, such as wi h he eg-
ula able GAL1 p omo e in hese expe imen s, we canno ex-
clude he possibili y ha such highe exp ession le els a e an
indi ec e ec o hp 1. The co ela ion be ween ansc ip size
and he hp 1:wild- ype ansc ip a io is no pe ec . This may
e lec he ac s ha (i) each ORF is unde he con ol o a
di e en p omo e , (ii) he ORFs a e in di e en ch omosomal
loca ions, and (iii) he di e en DNA sequence con ex o each
gene may a ec i s ansc ip ion pa e n. These esul s a e
consis en wi h ansc ip leng h being a leas one ea u e
pa ially esponsible o impai ing ansc ip ion d i en om
s ong p omo e s in hp 1 cells.
T ansc ip ion o GⴙC- ich DNA sequences is se e ely im-
pai ed by hp 1⌬.Al hough he leng h o a gene is an impo an
ea u e in luencing he sensi i i y o i s ansc ip ion o he
hp 1 mu a ion, he e a e se e al pieces o e idence indica ing
ha i canno be he only ea u e. Fi s , eplacemen o he
lacZ agmen s by GAL1 in he PHO5 ansc ip ional usion
cons uc s la gely supp essed he hp 1 e ec (Fig. 2). In addi-
ion, he wo se s o PHO5 usions ha we cons uc ed, in
which lacZ agmen s we e loca ed ei he a he 3⬘o he 5⬘
end, sha e he same leng h bu beha e di e en ly in hp 1 cells.
Finally, al hough bo h lacZ and LYS2 a e hype ecombino-
genic when lanked by di ec epea s, lacZ is signi ican ly mo e
ecombinogenic han LYS2 (Fig. 3B and 4B). The e o e, we
decided o explo e o he ea u es o lacZ, he mos hp 1-sen-
si i e sequence de ec ed so a , in o de o iden i y addi ional
elemen s in luencing ansc ip ional impai men by hp 1.
The mos e iden di e ence be ween lacZ and he bulk o
S. ce e isiae genes is he G⫹C con en . The majo i y o yeas
genes show a G⫹C con en o a ound 40%, whe eas ha o
lacZ is 56.2%. In o de o in es iga e whe he he G⫹C con-
en in luences he ansc ip ional impai men o lacZ in hp 1
cells, we used he Kluy e omyces lac is LAC4 gene, a yeas ho-
mologue o lacZ wi h 40% G⫹C (1). We placed LAC4 unde
GAL1 con ol, c ea ing a GAL1p ::LAC4 usion simila o
hose p e iously used in his wo k. T ans o man s o wild- ype
and hp 1 isogenic s ains we e used o de e mine he kine ics o
accumula ion o mRNA. The esul s p esen ed in Fig. 7A show
ha accumula ion o LAC4 ull-leng h mRNA was only mod-
e a ely diminished in hp 1 cells (50% o he wild- ype le el
a e 90 min o induc ion). None heless, in he same back-
g ound and a e an iden ical induc ion ime, lacZ ull-leng h
mRNA was almos absen (Fig. 4A) (11). The o e all compa -
ison o LAC4 and lacZ ansc ip ional beha io s showed ha
ansc ip ion h ough LAC4 was a leas i e old mo e e icien
han ha h ough lacZ in hp 1 cells. Thus, wo ansc ip ion
uni s, iden ical in leng h and di e ing in G⫹C con en , we e
di e en ially a ec ed by hp 1. This indica es ha , in addi ion o
ansc ip leng h, he G⫹C con en o a DNA sequence may be
an impo an ea u e in luencing ansc ip ional impai men by
hp 1.
To de e mine he e ec o LAC4 in ecombina ion, we
placed he en i e LAC4 ORF be ween he leu2 di ec epea s.
FIG. 3. T ansc ip ion analysis o h ee di e en GAL1p ::lacZ⌬-
PHO5 usion cons uc s in wild- ype (W303-1A) and hp 1 (U768-4C)
cells. (A) No he n blo analyses o PHO5-con aining mRNAs d i en
om he GAL1 p omo e . Plasmids used we e pSCh218, pSCh220, and
pSCh219 (ca ying ⬇0.4 kb o he 5⬘end, middle pa , and 3⬘end o
lacZ used o PHO5, espec i ely). As a con ol we used pSCh202
(ca ying he PHO5 gene; da a no shown), which ga e iden ical esul s
as hose shown in Fig. 2. The ansc ip le els o each cons uc wi h
espec o PHO5 we e simila o hose o he wild ype shown in Fig. 2.
O he de ails we e as desc ibed o Fig. 2. (B) Quan i ica ion o No h-
e n analyses.
7058 CHA
´VEZ ET AL. MOL.CELL.BIOL.
on July 4, 2017 by USE/BCTA.GEN UNIVERSITARIAh p://mcb.asm.o g/Downloaded om

The esul ing L-LAC4 sys em exhibi ed a high equency o
ecombina ion in hp 1 cells (90- old abo e wild- ype le els
[Fig. 7B]). This equency was eigh old lowe han ha shown
by L-lacZ bu was compa able o he equency eached by
L-LYS2 (Fig. 3B and 4B). An inc ease in ansc ip ion e i-
ciency is accompanied he e o e by a lowe ecombina ion
equency. I is impo an ha he L-LAC4 sys em shows a
hype ecombina ion pheno ype, because he size o he ull
ansc ip in his sys em is app oxima ely 5 kb. Indeed, and
consis en wi h ou hypo hesis, we ha e shown by No he n
analysis (Fig. 7C) ha ansc ip ion o he L-LAC4 sys em is
impai ed in hp 1 cells.
To u he in es iga e he in luence o G⫹C con en on
ansc ip ion o a DNA sequence, ega dless o whe he com-
ing om bac e ia o yeas , we decided o analyze a yeas gene
wi h a G⫹C con en compa able o ha o lacZ. YAT1, a 2-kb
long ORF, is he gene in he S. ce e isiae genome wi h he
highes G⫹C con en (58%). We placed YAT1 unde GAL1
con ol in a ansc ip ional sys em simila o hose used in he
p e ious expe imen s. No he n analysis showed ha high le -
els o YAT1 ull-leng h mRNA we e eached a e galac ose
induc ion in he wild ype, bu only a minimal accumula ion
was de ec ed in hp 1 cells (Fig. 8A). In ag eemen wi h his
ansc ip ional impai men , a ecombina ion sys em bea ing
YAT1 as in e ening sequence (L-YAT1) displayed an ex-
emely high equency o ecombina ion in hp 1 (12%) ha
was 173 imes he equency eached in he wild ype (Fig. 8B).
FIG. 4. T ansc ip ion and ecombina ion analyses o se e al GAL1p ::lacZ usion cons uc s in wild- ype (W303-1A) and hp 1 (U768-4C) cells.
(A) No he n analyses o lacZ-con aining mRNAs ansc ibed om he GAL1 p omo e . Plasmids used we e pSCh215, pSCh213, and pSCh216
(ca ying ⬇0.4 kb o he 5⬘end, middle pa , and 3⬘end o lacZ, espec i ely) o p416GAL1-lacZ (ca ying he en i e lacZ ORF). O he de ails
we e as desc ibed o Fig. 2. (B) Recombina ion equencies o leu2-based di ec - epea sys ems con aining he same sho lacZ agmen s used
in he p e ious No he n expe imen s. Plasmids used we e pSCh221, pSCh222, and pSCh223 (ca ying ⬇0.4 kb o he 5⬘end, middle pa , and 3⬘
end o lacZ, espec i ely) o pSCh205 (ca ying he en i e lacZ ORF). A schema ic diag am o he ecombina ion p oduc s ob ained wi h he
di ec - epea LEU2 ecombina ion sys ems used is shown a he op o panel B. The LEU2 p omo e (P m) and ansc ip ional e mina o (Te )
as well as he RNA (a ow) p oduced by he sys em a e indica ed. The median ecombina ion equency o six independen alues is gi en in each
case. All median equencies we e calcula ed in duplica e wi h wo independen ans o man s. Recombinan s we e selec ed in SC-Leu-T p. Da a
om he L-lacZ sys em con aining he en i e lacZ gene (bo om) a e aken om Cha´ ez and Aguile a (11).
FIG. 5. T ansc ip ion and ecombina ion analyses o LYS2 se-
quences in wild- ype and hp 1 cells. (A) No he n blo analyses o
LYS2 mRNAs in s ains ans o med wi h plasmid pSCh227 con aining
a 3.7-kb agmen o he LYS2 coding sequence unde he con ol o
he GAL1 p omo e . (B) Recombina ion equencies o s ains ans-
o med wi h plasmid pSCh230 ha bo ing a leu2-based di ec epea
sys em con aining as in e ening sequence he same 3.7 kb agmen o
LYS2 used o he ansc ip ion assays. O he de ails a e as desc ibed
o Fig. 4.
VOL. 21, 2001 Hp 1-DEPENDENT TRANSCRIPTION ELONGATION 7059
on July 4, 2017 by USE/BCTA.GEN UNIVERSITARIAh p://mcb.asm.o g/Downloaded om
This le el o hype ecombina ion is compa able o ha o
L-lacZ (Fig. 4B) and highe han he le els o L-LYS2 (Fig. 5B)
and L-LAC4 (Fig. 7B). Thus, ansc ip ion h ough a medium-
size G⫹C- ich gene is clea ly hp 1 sensi i e, indica ing ha
G⫹C con en can modula e he Hp 1 dependency o gene
ansc ip ion.
Nucleosome posi ioning is lacking in lacZ sequences. The
o ganiza ion o DNA in a p ope nucleosome-posi ioned ch o-
ma in s uc u e has been shown o be a o ed by A⫹T- ich
mo i s (27) and p e en ed by G⫹C- ich sequences (62). In
o de o es whe he he e is a ela ionship be ween ch oma in
s uc u e and ansc ip ional e iciency in hp 1 cells, we de e -
mined whe he he ch oma in s uc u e o G⫹C- ich se-
quences such as lacZ was di e en om ha o low-G⫹C-
con en sequences. We pe o med MNase sensi i i y assays o
he GAL1p ::lacZ usion cons uc and he GAL1 endogenous
genes, in which ansc ip ion was s ongly and poo ly impai ed
in hp 1 cells, espec i ely (11, 17, 67) (Fig. 9A). As p e iously
shown (19), clea and speci ic nucleosome posi ioning along
he endogenous GAL1 gene was obse ed (Fig. 9B). Such a
pa e n o MNase sensi i i y was iden ical o bo h wild- ype
and hp 1⌬cells. The mo e di use pa e n o MNase diges ion
unde induced condi ions in bo h wild- ype and hp 1 cells e-
lec s he des abiliza ion o ch oma in s uc u e caused by an-
sc ip ion (9). In e es ingly, he pa e n o MNase sensi i i y o
lacZ shows no nucleosome o ganiza ion in ei he wild- ype o
hp 1⌬cells unde bo h induced and ep essed condi ions o
ansc ip ion. Nucleosome posi ioning is only limi ed o he
GAL1 p omo e (Fig. 9C). Indeed, a lack o nucleosome po-
si ioning is also obse ed o e he bac e ial sequences up-
s eam o he GAL1 p omo e , h ough which ansc ip ion has
also been shown o be impai ed in hp 1 mu an s (44). Ou
esul s, he e o e, show ha lacZ adop s a andom nucleoso-
mal o ganiza ion in yeas and ha hp 1 has no e ec on nu-
cleosome posi ioning.
DISCUSSION
In his wo k we ha e in es iga ed why ansc ip ion o DNA
sequences like E. coli lacZ is especially sensi i e o hp 1.We
ha e shown ha 0.4-kb lacZ agmen s used o PHO5 unde
he GAL1 p omo e a e su icien o inc ease he Hp 1 depen-
dency o ansc ip ion, bu no when hey a e ansc ibed
alone. Such an e ec is posi ion dependen : he longe he
dis ance be ween he lacZ sequence and he GAL1 p omo e ,
he s onge he impai men o ansc ip ion caused by hp 1. In
addi ion, we see ansc ip ion o long yeas DNA sequences
like LYS2 used o he GAL1 p omo e is nega i ely a ec ed by
hp 1. We ha e also shown ha ansc ip ion o K. lac is LAC4,
a euka yo ic homologue o lacZ ha is equal in leng h bu wi h
a much lowe G⫹C con en , exhibi s a milde Hp 1 depen-
dency in S. ce e isiae, whe eas YAT1,anS. ce e isiae G⫹C- ich
gene sho e han lacZ, is d ama ically a ec ed by hp 1. Taken
FIG. 6. T ansc ip ion analyses o i e yeas endogenous genes,
EGT2,CDC48,KAR2,OLE1, and GOG5, ha ing high le els o ex-
p ession and di e en ansc ip sizes, in wild- ype and hp 1 cells. To al
RNA was isola ed om mid-log phase cells, g own in YEPD b o h,
and used o No he n analyses. In e nal agmen s o each gene and
o he 23S DNA, ob ained by PCR, we e used as DNA p obes. The
hp 1:wild- ype ansc ip a io was ob ained om he mRNA le els
ha we e quan i ied in a Fuji FLA3000 and no malized wi h espec o
he RNA le els.
FIG. 7. T ansc ip ion and ecombina ion analyses o LAC4 in wild-
ype and hp 1 cells. (A) No he n analyses o LAC4 mRNAs in s ains
ans o med wi h he plasmid pSCh255, which con ains he en i e
LAC4 coding sequence unde he con ol o he GAL1 p omo e . (B)
Recombina ion equencies o cells ans o med wi h plasmid pSCh254,
which ha bo s he leu2-based di ec - epea L-LAC4 cons uc con ain-
ing LAC4 as he in e ening egion. (C) No he n analyses o he
L-LAC4 epea cons uc in wild- ype and hp 1 cells. O he de ails a e
as desc ibed o Fig. 4.
7060 CHA
´VEZ ET AL. MOL.CELL.BIOL.
on July 4, 2017 by USE/BCTA.GEN UNIVERSITARIAh p://mcb.asm.o g/Downloaded om
oge he , hese esul s indica e ha bo h leng h and G⫹C
con en a e impo an elemen s in luencing gene ansc ip ion
in i o and ha Hp 1 is an impo an ac o con olling an-
sc ip ion o ei he long o G⫹C- ich DNA sequences used o
a s ong p omo e such as GAL1p .
Hp 1 is equi ed o p ope ansc ip ion o long DNA
sequences. We ha e shown ha ansc ip ion o long DNA
sequences is comp omised in hp 1 cells, whe eas sho e se-
quences a e ei he una ec ed o mildly in luenced. This con-
clusion is suppo ed by he inapp eciable e ec o hp 1 on
ansc ip ion o sho lacZ agmen s di ec ly used o he
GAL1 p omo e , whe he o no ups eam o PHO5, whe eas
he same lacZ agmen s do con e hp 1 dependency when
used downs eam o PHO5 (Fig. 2 and 3). This conclusion is
also suppo ed by he ma ked nega i e e ec o hp 1 on an-
sc ip ion o he 3.7-kb-long LYS2 agmen unde con ol o
he GAL1 p omo e (Fig. 5). The ansc ip ional analysis o
i e highly ansc ibed ch omosomal genes showed a nega i e
e ec o hp 1 on ansc ip ion o he h ee longes ones, EGT2
(3.3 kb), CDC48 (2.7 kb), and KAR2 (2.1 kb), bu no nega i e
e ec on he o he wo, OLE1 (1.6 kb) and GOG5 (1.1 kb)
(Fig. 6). Fu he expe imen s would be equi ed o know
whe he hese esul s a e also alid o poo ly exp essed genes.
P ocessi i y de ec s o an RNA polyme ase can be mo e
easily de ec ed wi h ansc ip ion o long a he han sho
DNA empla es. Compa ison be ween long and sho an-
sc ip s is in ac a common me hod o quan i y he e ec o
ansc ip ion ac o s on RNAPII-media ed elonga ion (66).
Consequen ly, a leng h-dependency e ec o hp 1 on ansc ip-
ion is expec ed i Hp 1 con ols elonga ion. The longe he
ansc ip ion uni , he highe he p obabili y o RNAPII each-
ing a DNA egion equi ing he unc ion o Hp 1. E en in he
wild- ype s ain, long ansc ip ion uni s a e less e icien ly
exp essed han sho ones, as we ha e obse ed by compa ing
FIG. 8. T ansc ip ion and ecombina ion analyses o YAT1 in wild-
ype and hp 1 cells. (A) No he n blo analyses o YAT1 mRNAs in
cells ans o med wi h plasmid pSCh247 con aining he en i e YAT1
coding sequence unde he con ol o he GAL1 p omo e . (B) Re-
combina ion analyses o cells ans o med wi h plasmid pSCh248,
which ha bo s he leu2-based di ec - epea sys em con aining he en-
i e YAT1 gene as in e ening sequence. O he de ails a e as desc ibed
o Fig. 4.
FIG. 9. MNase diges ion pa e n o he GAL1 gene and he
GAL1p ::lacZ usion in wild- ype and hp 1 s ains unde ep ession
and ac i a ion condi ions. (A) No he n blo analysis o GAL1 mRNAs.
(B) Nucleosome posi ioning o e he GAL1 gene. (C) MNase diges ion
pa e n o he GAL1p ::lacZ usion. A scheme o he analyzed egions o
GAL1 and lacZ indica ing he posi ion o nucleosomes and he mos
ele an egula o y elemen s is shown. As e isks indica e he MNase hy-
pe sensi i e si es associa ed wi h he ac i a ion o ansc ip ion o GAL1.
VOL. 21, 2001 Hp 1-DEPENDENT TRANSCRIPTION ELONGATION 7061
on July 4, 2017 by USE/BCTA.GEN UNIVERSITARIAh p://mcb.asm.o g/Downloaded om
he exp ession le els o se e al PHO5::lacZ usion cons uc s
ha di e in he size o he lacZ agmen (Fig. 1). Thus, i is
possible ha he absence o Hp 1 enhances elonga ion de ec s
al eady p esen in he wild ype, he occu ence o such de ec s
being mo e likely as he DNA sequence o be ansc ibed
becomes longe .
T ansc ip ion o GⴙC- ich DNA sequences is Hp 1 depen-
den . Ou esul s suppo a co ela ion be ween G⫹C con en
and HPR1 unc ion. This conclusion is based on h ee se s o
da a. Fi s , al hough mos es ed genes a e ei he sligh ly o no
a ec ed by hp 1 (Fig. 7) (11, 67), ansc ip ion o lacZ (56%
G⫹C) in Saccha omyces is s ongly a ec ed and di ec - epea
sys ems con aining lacZ sequences ha a e ansc ibed exhibi
hype ecombina ion in hp 1 cells (11). Second, ansc ip ion o
LAC4, a 40% G⫹C- ich lacZ o hologue o K. lac is, is a leas
i e imes mo e e icien han lacZ in hp 1 cells (Fig. 7A), and
di ec epea s lanking LAC4 ecombine a equencies eigh -
old lowe han hose con aining lacZ (Fig. 7B). Finally, YAT1,
a 2-kb-long gene om Saccha omyces wi h a 58% G⫹C con-
en , is s ongly a ec ed by hp 1 bo h in ansc ip ion and in
ecombina ion (Fig. 8).
As a as we know, nei he a di ec in luence o he G⫹C
con en o he empla e on ansc ip ion elonga ion e iciency
no he equi emen o auxilia y ac o s o ansc ip ion o
G⫹C- ich genes has been desc ibed. The only sequence ea-
u es ha ha e been shown o a ec elonga ion a e hose o
speci ic ansc ip ional pausing si es (59). Some pause si es
iden i ied in bac e ia a e G⫹C ich, like he ops signals in
E. coli (4), bu o he s a e no . I is, he e o e, unlikely ha
G⫹C- ich genes exhibi , in gene al, a highe p obabili y o
con aining a pause signal. None heless, a high G⫹C con en
migh a ec ansc ip ional elonga ion by s abilizing seconda y
s uc u es in he nascen RNA ha can unc ion as pausing
signals. A leas in he case o T7 phage, a lowe numbe o
hyd ogen bonds in he RNA hai pins elimina es some pauses
(32), sugges ing ha a high G⫹C con en migh con ibu e o
s onge RNA-media ed elonga ion impai men s.
AG⫹C- ich nascen RNA migh also o m mo e s able
RNA-DNA hyb ids wi hin he empla e. Twel e-nucleo ide-
long RNA-DNA hyb ids nega i ely a ec RNAPII p ocessi i y
in i o (29). Ano he class o RNA-DNA hyb ids a e R-loops,
p oduced by he associa ion o nascen mRNA wi h ups eam
empla e DNA. I has been p oposed ha hese R-loops a e
o med du ing ansc ip ion o he G⫹C- ich human immuno-
globulin swi ch egion in i o (15), and hey ha e been de-
ec ed a e in i o ansc ip ion (58). The clea age o R-loops
by speci ic nucleases migh ini ia e class-swi ch ecombina ion
(58), p o iding a mechanism o explain ansc ip ion-associ-
a ed ecombina ion. No hing is known abou he in luence o
RNA-DNA s uc u es on RNAPII-dependen ansc ip ion
(20). Howe e , i has been p oposed ha R-loops o med du -
ing RNA ansc ip ion elonga ion in E. coli cons i u e oad-
blocks o he nex ansc ibing RNA polyme ase (25).
Molecula ea u es making ansc ip ion elonga ion Hp 1
dependen . Unless Hp 1 plays mo e han one unc ion, we
should expec a common mechanis ic equi emen du ing an-
sc ip ion o long e sus G⫹C- ich genes equi ing i s ac ion. I
is no e iden which kind o ansc ip ion-impai ing signal
migh link long and G⫹C- ich DNA sequences. One possibili y
would be he exis ence o sho e G⫹C- ich egions hinde ing
ansc ip ion wi hin long genes. Howe e , we can exclude his
possibili y wi h he esul s o his wo k. Conside ing a 300-bp-
long window ( he minimal lacZ agmen con e ing an e ec
o hp 1 in ansc ip ion), he maximum G⫹C con en is 46%
o LYS2, 45.5% o LAC4, and 59% o lacZ. Sho e o
longe windows show simila esul s. In addi ion, no di e ence
in he G o C con en is ound on he cDNA s ands. As LYS2
is no mo e G⫹C- ich han LAC4, leng h is he mos likely
eason he wo genes beha e di e en ly in hp 1 mu an s. Con-
sis en ly, when LAC4 is loca ed be ween he leu2 epea s in he
L-LAC4 cons uc , he ansc ip ion uni con aining LAC4 be-
comes longe and ansc ip ion becomes signi ican ly a ec ed
in hp 1 cells (Fig. 7).
Al e na i ely, he link be ween long and G⫹C- ich se-
quences migh be un ela ed o he DNA sequence i sel . T an-
sc ip ion o long and G⫹C- ich sequences may p oduce some
kind o ansc ip ional e en ha would be o e come by he
ac ion o Hp 1. Gene ic analyses ha e p o ided some hin s as
o he na u e o his kind o e en . Se e al mu an s ha e been
desc ibed ha display syn he ic pheno ypes wi h hp 1 (2, 67).
The ac ha opoisome ase mu an s ( op1, op2, and op3)
become sick in an hp 1 backg ound may es ablish a link wi h
DNA opology (2, 48). Fo example, he accumula ion o neg-
a i e supe coiling impai s ansc ip ional elonga ion o bac e-
ial genes in i o (30), and posi i e supe coiling diminishes
RNAPII-dependen ansc ip ion in yeas cells. R-loops a e
o med in he absence o DNA opoisome ase I in E. coli (16),
and hey ha e been p oposed o impai ansc ip ion elonga-
ion (25). Elonga ion by RNAPII al e s empla e opology (8),
p oducing an accumula ion o posi i e and nega i e supe coil-
ing ahead o and behind he RNAPII, espec i ely (33). I is,
he e o e, expec ed ha posi i e supe coiling will be s onge
a he 3⬘ egion o long ansc ip ion uni s. Examples o DNA
sequences ha impai ansc ip ion elonga ion mo e e icien ly
in dis al loca ions ha e been epo ed (63). The s ong e ec o
hp 1 on ansc ip ion o long DNA sequences ag ees wi h his
iew.
Ch oma in s uc u e is ano he sou ce o s ess a ec ing
RNAPII-media ed ansc ip ion elonga ion ha equi es he
ac ion o speci ic auxilia y ac o s ( e iewed in e e ence 39).
Some mu a ions esul ing in a poo g ow h pheno ype wi h
hp 1 a e in ac ela ed o ch oma in s uc u e, like SIN1-2
o hose causing his one imbalance (67). The o ganiza ion
o DNA in a p ope , nucleosome-posi ioned ch oma in s uc-
u e is a o ed by some A⫹T- ich mo i s (27) and is p e en ed
by some G⫹C- ich sequences (62). A G⫹C- ich sequence
migh , he e o e, be biased agains a p ope ch oma in s uc-
u e. The lacZ gene is in ac unable o suppo s ably
o ganized ch oma in in S. ce e isiae (Fig. 9). T ansc ip ion
o G⫹C- ich genes migh hen be impai ed in hp 1 due o an
abe an ch oma in s uc u e.
I hese hypo heses a e ue, he loca ion o a G⫹C- ich
egion in he 3⬘ egion o a gene should p oduce a s onge
e ec , since i would combine wo sou ces o ansc ip ional
s ess in he same place: supe helici y and abe an ch oma in
s uc u e. Indeed, nega i e supe coiling can induce changes o
DNA s uc u e in CG sequences, wi h his al e ed s uc u e
being able o p oduce ansc ip ional elonga ion blocks (41).
Ou esul s wi h he PHO5::lacZ usion cons uc s suppo his
iew, since he loca ion o he G⫹C- ich lacZ agmen s a he
7062 CHA
´VEZ ET AL. MOL.CELL.BIOL.
on July 4, 2017 by USE/BCTA.GEN UNIVERSITARIAh p://mcb.asm.o g/Downloaded om