Ac i a ion-induced cy idine deaminase ac ion is
s ongly s imula ed by mu a ions o he THO complex
Bele
´nGo
´mez-Gonza
´lez and And e
´s Aguile a*
Depa amen o de Gene´ ica, Facul ad de Biologı´a, and Depa amen o de Biologı´a Molecula , Cen o Andaluz de Biologı´a Molecula y Medicina
Regene a i a, Consejo Supe io de In es igaciones Cien ı´ icas-Uni e sidad de Se illa, A enida Ame´ ico Vespucio s/n, 41092 Se illa, Spain
Communica ed by Thomas D. Pe es, Duke Uni e si y Medical Cen e , Du ham, NC, Ma ch 27, 2007 ( ecei ed o e iew Decembe 9, 2006)
Ac i a ion-induced cy idine deaminase (AID) is a B cell enzyme
essen ial o Ig soma ic hype mu a ion and class swi ch ecombi-
na ion. AID ac s on ssDNA, and swi ch egions o Ig genes, a a ge
o AID, o m R-loops ha con ain ssDNA. Ne e heless, how AID
ac ion is speci ically a ge ed o pa icula DNA sequences is no
clea . Because mu a ions al e ing co ansc ip ional messenge i-
bonucleop o ein (mRNP) o ma ion such as hose in THO/TREX in
yeas p omo e R-loops, we in es iga ed whe he he co ansc ip-
ional assembly o mRNPs could a ec AID a ge ing. He e we
show ha AID ac ion is ansc ip ion-dependen in yeas and ha
s ong and ansc ip ion-dependen hype mu a ion and hype e-
combina ion a e induced by AID i cells a e dep i ed o THO. In
hese s ains AID-induced mu a ions occu ed p e e en ially a
WRC mo i s in he non ansc ibed DNA s and. We p opose ha a
subop imal co ansc ip ional mRNP assembly a pa icula DNA
egions could play an impo an ole in Ig di e si ica ion and
genome dynamics.
hype ecombina ion 兩hype mu a ion 兩messenge ibonucleop o ein
biogenesis 兩R-loops
Ac i a ion-induced cy idine deaminase (AID) is a speci ic B
cell enzyme belie ed o be esponsible o he ini ia ion o
soma ic hype mu a ion (SHM) and class swi ch ecombina ion
(CSR) du ing B cell di e en ia ion (1–3). Many s udies ha e
shown ha AID ac s di ec ly on DNA (4), he na u al a ge o
AID ac ion in he a iable and swi ch (S) egions o SHM and
CSR, espec i ely. The speci ic mechanisms o AID unc ion a e
s ill unclea , bu e idence sugges s ha he p e e en ial a ge o
AID may be ssDNA (5, 6). Thus, in i o expe imen s ha e shown
ha AID deamina es ssDNA and dsDNA ha is ansc ibed
(5–7), and ansc ip ion has been shown o be equi ed o bo h
SHM and CSR in i o (8, 9).
S udies ha e shown a s and p e e ence o he ac ion o AID
du ing in i o ansc ip ion wi h AID-induced mu a ions de-
ec ed p e e en ially in he non ansc ibed (NT) s and (6,
10–12). Analysis o he p oduc s o SHM in B cells, howe e ,
e eals ha bo h DNA s ands a e mu a ed (13). O he in i o
s udies ha e shown ha AID can deamina e bo h s ands wi hin
egions o supe coiled DNA (14) and ha he ssDNA-binding
eplica ion p o ein A is equi ed o deamina ion o SHM a ge s
(15). All o hese indings a e consis en wi h he idea ha
ansien o ma ion o ssDNA du ing ansc ip ion acili a es
AID ac ion.
Along he same line, o ma ion o R-loops du ing ansc ip-
ion o he S egions has been p oposed (16). Such R-loops would
possibly p o ide AID wi h ssDNA subs a es a i s a ge egion
because he e he ansc ibed (T) s and hyb idizes wi h he
nascen RNA and he NT s and is p esen as ssDNA. Because
o i s high G con en , he NT DNA s and a S egions could be
s abilized by he o ma ion o pa allel ou -s anded G qua e s
(17, 18). Indeed, AID appea s o bind speci ically o G-loops
wi hin ansc ibed S egions (19) and can deamina e he dis-
placed s and o a ansc ip ion-induced R-loop in i o (20).
In e es ingly, co ansc ip ional messenge ibonucleop o ein
(mRNP) biogenesis is an essen ial s ep in gene exp ession ha
may in luence gene ic in eg i y (21, 22). In he yeas Saccha o-
myces ce e isiae, hype ecombina ion can be igge ed by an-
sc ip ion in mu an s deple ed o THO/TREX, a conse ed
p o ein complex ha unc ions a he in e ace be ween an-
sc ip ion and mRNA expo (23, 24). THO is a conse ed
complex i s iden i ied in yeas and o med by Tho2, Hp 1,
M 1, and Thp2 (24). I in e ac s wi h he Sub2 RNA-dependen
ATPase and he Y a1 RNA-binding p o ein o o m he TREX
complex (23). THO has been shown o be equi ed o he
ansc ip ion o long genes and genes con aining ei he high GC
con en o mul iple in e nal epea s (25, 26). I is belie ed ha
THO con ols he co ansc ip ional o ma ion o expo -
compe en mRNP du ing ansc ip ion elonga ion by con olling
he assembly o he e ogeneous nuclea ibonucleop o eins on o
he nascen mRNA (21). Hype ecombina ion o THO mu an s
is pa icula ly e iden in long and GC- ich DNA sequences such
as he bac e ial lacZ gene (26). Because hype ecombina ion
depends on he o ma ion o DNA:RNA hyb ids (22), i has been
p oposed ha one unc ion o THO/TREX is o p e en he
nascen mRNA om in e ac ing wi h he empla e DNA o o m
R-loops. The obse a ion ha he ASF/SF2 splicing ac o has a
simila ole in chicken DT40 and HeLa cells (27) sugges s ha
a numbe o RNA p ocessing enzymes ha e an addi ional ole in
p e en ing R-loop o ma ion.
One impo an and un esol ed ques ion is why AID is a ge ed
a speci ic DNA egions du ing B cell di e en ia ion, and we
wonde ed whe he his could be because o subop imal mRNP
o ma ion. To es his possibili y, we con i med ha AID can
induce hype mu a ion and hype ecombina ion a low le els in
yeas cells (28, 29) and demons a ed addi ionally ha his
ins abili y occu s in a ansc ip ion-dependen manne . Using
yeas THO mu an s as a way o gene a e subop imal mRNP
o ma ion, we ound ha bo h hype mu a ion and hype ecom-
bina ion we e s ongly and syne gis ically inc eased in a
ansc ip ion-dependen manne , wi h mu a ions p ima ily oc-
cu ing in he NT s and. We p opose ha a subop imal co-
ansc ip ional mRNP assembly a pa icula DNA egions can
egula e AID ac ion, and we sugges ha he con ol o mRNP
biogenesis may play an impo an ole in Ig di e si ica ion and
genome dynamics.
Resul s
Mu a ions in he THO Complex Con e a Sligh T ansc ip ion-
Dependen Hype mu a ion Pheno ype. Because THO mu an s
sha e pheno ypes o ecombina ion-media ed gene ic ins abili y
Au ho con ibu ions: B.G.-G. and A.A. designed esea ch; B.G.-G. pe o med esea ch;
B.G.-G. and A.A. analyzed da a; and B.G.-G. and A.A. w o e he pape .
The au ho s decla e no con lic o in e es .
Abb e ia ions: AID, ac i a ion-induced cy idine deaminase; CSR, class swi ch e-
combina ion; mRNP, messenge ibonucleop o ein; NT, non ansc ibed; S, swi ch;
T, ansc ibed;SC,syn he iccomple e;SHM,soma ichype mu a ion;TAM, ansc ip ion-
associa ed mu a ion.
*To whom co espondence should be add essed. E-mail: [email p o ec ed].
This a icle con ains suppo ing in o ma ion online a www.pnas.o g/cgi/con en / ull/
0702836104/DC1.
© 2007 by The Na ional Academy o Sciences o he USA
www.pnas.o g兾cgi兾doi兾10.1073兾pnas.0702836104 PNAS
兩
May 15, 2007
兩
ol. 104
兩
no. 20
兩
8409–8414
GENETICS
associa ed wi h ansc ip ion, i was impo an o know whe he
hey showed, in addi ion o hype ecombina ion, a ansc ip ion-
dependen hype mu a ion pheno ype. Two di e en plasmid-
bo ne assays we e de eloped o he analysis o o wa d mu a-
ions. The i s assay is based on a usion o he KAR1 ORF o
he egula ed GAL1 p omo e (GKAR1 sys em). Because KAR1
o e exp ession is le hal in yeas (30), ka 1 mu a ions can be
selec ed on galac ose medium. The second assay is based on a
lacZ::URA3 ansla ional usion unde con ol o he egula ed
Te p omo e (LAUR sys em) (31). In his assay, U a
⫺
mu an s
a e selec ed on syn he ic comple e (SC) medium supplemen ed
wi h 5- luo oo o ic acid, whe eas lacZ
⫺
mu an s and lacZ
⫹
a e
sco ed by colo on SC plus X-Gal. As can be seen in Fig. 1 Aand
B, he mu a ion equencies we e inc eased 13- old when an-
sc ip ion was high in bo h sys ems. T ansc ip ion-associa ed
mu a ion (TAM) was also obse ed in ad52⌬cells al hough o
a mino ex en (Fig. 1B). The esul is in ag eemen wi h p e ious
epo s indica ing ha ansc ip ion inc eases mu a ion in yeas
cells (32) and Esche ichia coli ( e iewed in e . 33).
Nex , we used bo h he LAUR and GKAR1 assays o de e -
mine whe he di e en THO mu a ions s imula ed TAM. As
seen in Fig. 1A,m 1⌬and hp 1⌬, which caused an elimina ion
o he THO complex in yeas cells (34), inc eased he equency
o mu a ion 2- o 3- old in GKAR1 a low ansc ip ion e sus 38-
o 60- old a high ansc ip ion le els. Simila ly, in he LAUR
sys em, he inc ease in mu a ion equency was signi ican ly
s onge o high ansc ip ion (96- o 189- old) han o low
ansc ip ion (2- o 17- old) as assayed in bo h he m 1⌬and
hp 1⌬backg ounds (Fig. 1B). The e o e, THO mu an s sha e a
ansc ip ion-dependen hype mu a ion pheno ype ha is no as
s ong as i s hype ecombina ion pheno ype obse ed in se e al
di ec - epea assays (26). Because all null mu a ions o he genes
encoding he di e en THO subuni s lead o dep i a ion o
THO (34) and he same pheno ypes (24), we decided o con inue
ou analysis wi h m 1⌬, which, in con as o hp 1⌬, has li le
impac on g ow h.
To gain insigh in o he molecula basis o mu a ions caused
by THO mu a ions, we pe o med a gene ic analysis o he
pa e n o mu a ions caused by m 1⌬. As expec ed om he ac
ha ansc ip ion h ough lacZ is he mos sensi i e s ep in THO
mu an s (26), e e y U a
⫺
mu an was also lacZ
⫺
in all condi ions
es ed, implying ha all mu a ions mapped wi hin lacZ (54 in
high ansc ip ion and ou in low ansc ip ion) (Fig. 2),
whe eas, in he wild ype, 34–40% o he U a
⫺
mu a ions we e
lacZ
⫹
(lacZ
⫺
:lacZ
⫹
a ios o 50:32 and 80:41 in low and high
ansc ip ion, espec i ely) (Fig. 2); ha is, hey occu ed wi hin
URA3. The m 1⌬mu a ions mus , he e o e, be ei he ame-
shi s o p ema u e s op codons wi hin lacZ o DNA ea ange-
men s co e ing lacZ and URA3. Gi en he s ong ansc ip ion-
dependen hype ecombina ion pheno ype o THO mu an s, i
was impo an o de e mine whe he lacZ
⫺
U a
⫺
colonies in he
m 1⌬backg ound a ose by ecombina ion-dependen ea -
angemen s, which a e known o be Rad52-dependen . The
spon aneous mu a ion equency unde high ansc ip ion in he
ad52⌬s ain was 2- old lowe han in he wild ype (Fig. 1B),
al hough i has been epo ed ha TAM inc eases in ad52⌬cells
(32, 35), indica ing ha some o he TAM e en s obse ed in he
wild ype may a ise by a Rad52-dependen pa hway in he LAUR
cons uc . The ad52⌬mu a ion causes a decline in he mu a ion
equency o m 1⌬cells o 6.5- old ( om 2.9 ⫻10
⫺5
o 4.5 ⫻
10
⫺6
), indica ing ha mos o TAM e en s in m 1⌬occu by
RAD52-dependen ea angemen s a he han by poin mu a-
ions (Fig. 1B). In con as o wild- ype cells in which none o
eigh analyzed U a
⫺
mu an s con ained de ec able ea ange-
men s, 62% o he mu an s (23 o 37) con ained ea angemen s
in m 1⌬[suppo ing in o ma ion (SI) Fig. 5]. Thus, he absence
o THO s imula es Rad52-independen mu a ions 2- o 3- old
( om 0.6 ⫻10
⫺6
o 1.2 ⫻10
⫺6
and om 1.5 ⫻10
⫺6
o 4.5 ⫻10
⫺6
in low and high ansc ip ion, espec i ely) (Fig. 1B). Indeed,
Fig. 1. E ec o ansc ip ion and AID on he equency o mu a ion in S.
ce e isiae wild- ype and THO mu an s ains. (A) F equency o Ka ⫺mu an s
ob ained in he GKAR1 sys em. (B) F equency o U a⫺mu an s ob ained in he
LAUR sys em and Rad52 dependency. (C) E ec o AID on he equency o
U a⫺mu an s in he LAUR sys em in wild- ype, m 1⌬, ad52⌬, and m 1⌬
ad52⌬s ains. All expe imen s we e pe o med unde high (ON) and low
(OFF) ansc ip ion. Diag ams o sys ems a e shown abo e each g aph. Mean
mu a ion equency and s anda d de ia ion o h ee o ou di e en luc u-
a ion es s a e plo ed.
Fig. 2. Gene ic analysis o spon aneous (⫺AID) and AID-induced (⫹AID)
mu a ions. Dis ibu ion o lacZ⫺:lacZ⫹mu a ions among U a⫺mu an s in
wild- ype and m 1⌬s ains is shown. An as e isk indica es a s a is ically
signi ican di e ence (P⬍0.05) wi h espec o he wild- ype alue unde he
same ansc ip ion condi ions, as de e mined by con ingency
␹
2analysis.
8410
兩
www.pnas.o g兾cgi兾doi兾10.1073兾pnas.0702836104 Go´ mez-Gonza´lez and Aguile a
50% o U a
⫺
mu a ions a e lacZ
⫹
in he m 1⌬ ad52⌬mu an
(7:7 lacZ
⫺
:lacZ
⫹
a io; da a no shown), consis en wi h he
conclusion ha THO mu an s also s imula e poin mu a ions,
al hough sligh ly, in a ansc ip ion-dependen manne .
AID-Induced Mu a ion Is T ansc ip ion-Dependen in Yeas . Taking
ad an age o he LAUR assay, we s udied he abili y o AID o
induce mu a ion and whe he his abili y depended on ansc ip-
ion in wild- ype cells. As can be seen in Fig. 1 Band C,AID
exp ession in wild- ype yeas did no lead o hype mu a ion a
low ansc ip ion, bu caused a 2- old inc ease a high ansc ip-
ion ( om 3.9 ⫻10
⫺6
o 7.6 ⫻10
⫺6
). Consis en wi h a lack o
an AID e ec on mu a ion equency unde low ansc ip ion in
he wild ype, he a io o lacZ
⫺
:lacZ
⫹
among U a
⫺
mu an s was
simila ega dless o whe he cells exp essed AID (11:9 e sus
50:32) (Fig. 2). Howe e , a signi ican change (P⬍0.05) in his
a io was obse ed unde high ansc ip ion i AID was ex-
p essed (66:86 e sus 80:41), indica ing ha he pa e n o
mu a ions was changed by AID in a ansc ip ion-dependen
manne . The e o e, AID induces a weak ansc ip ional-
dependen inc ease in he equency o mu a ion and a change
in he pa e n o mu a ions.
AID-Induced Mu a ion Is S ongly S imula ed by THO Mu a ions. THO
mu a ions may c ea e local egions ha a e suscep ible o
o ming ecombinogenic and mu agenic in e media es in a
ansc ip ion-dependen manne . To de e mine whe he such
egions would be mo e suscep ible o AID ac ion, we es ed he
e ec o AID on he equency and pa e n o mu a ions in
m 1⌬cells. As can be seen in Fig. 1C,inm 1⌬cells he e ec
o AID on he equency o mu a ions was g ea e han in he
wild ype, leading o a 13- old inc ease in mu a ions a low
ansc ip ion ha wen up o 630- old a high ansc ip ion (Fig.
1C). Unde low and high ansc ip ion, AID s imula ed he
mu a ion equency 6.5 imes ( om 0.6 ⫻10
⫺6
o 3.9 ⫻10
⫺6
and
om 2.9 ⫻10
⫺5
o 1.9 ⫻10
⫺4
, espec i ely) in m 1⌬cells.
Impo an ly, his hype mu a o e ec o AID was no dependen
on Rad52. AID exp ession u he inc eased U a
⫺
mu a ions
81- old ( om 4.5 ⫻10
⫺6
o 3.7 ⫻10
⫺4
)inm 1⌬ ad52⌬cells
unde high ansc ip ion (Fig. 1 Band C), indica ing ha hey a e
no due o ea angemen s bu mos likely poin mu a ions.
THO mu a ions enhance AID ac ion p esumably by c ea ing
a local s uc u e in he ansc ibed DNA egion ha wo ks as a
a ge o AID. The e o e, we would expec ha non ameshi
poin mu a ions mapping a lacZ and con e ing a lacZ
⫺
bu no
U a
⫺
pheno ype would occu a a high equency in m 1⌬bu
no in wild- ype cells exp essing AID. As expec ed, we we e no
able o de ec lacZ
⫺
U a
⫹
mu an s om 5,922 colonies in
wild- ype cells exp essing AID ( equency ⬍10
⫺3
), whe eas in
m 1⌬exp essing AID we coun ed 47 lacZ
⫺
U a
⫹
mu an s o
4,692 colonies ( equency o 1 ⫻10
⫺2
)(SI Table 2). This esul
indica es ha he equency o U a
⫺
AID-induced mu a ions in
m 1⌬was unde es ima ed in he LAUR sys em and indica es a
s ong e ec o THO mu a ions on AID ac ion. The e o e,
ansc ip ion and AID esul in a s ong hype mu a o pheno-
ype in m 1⌬cells.
AID Mu a es P e e en ially he NT S and in
m 1
⌬bu Shows No
S and P e e ence in he Wild Type. AID ac s p e e en ially a WRC
(GYW in he opposi e s and) mo i s (11) ha a e pa o he
WRCY mo i s (RGYW in he opposi e s and) ound as ho spo s
o SHM (36) (whe e W is A o T, R is A o G, and Y is C o T).
As R-loops in THO mu an s (19) lea e ssDNA opposi e he
DNA:RNA hyb id, one p edic ion is ha mos o he AID-
induced mu a ions in m 1⌬cells should be a WRC mo i s and
in he NT s and. The e o e, we sequenced spon aneous and
AID-induced mu a ions in wild- ype and m 1⌬cells. As shown
in Fig. 2, all mu an s in m 1⌬cells we e lacZ
⫺
U a
⫺
, which can
be explained i he U a
⫺
pheno ype esul s om ameshi
mu a ions, s op codons, o ea angemen s occu ing a lacZ.
This iew is consis en wi h he ac ha ansc ip ion h ough
lacZ is he mos sensi i e s ep in THO mu an s (26). The e o e,
because poin mu a ions a URA3 a e no de ec ed and lacZ
⫺
U a
⫺
mu an s ep esen a biased class o mu a ions a lacZ,we
decided o sequence he lacZ sequence o he lacZ
⫺
U a
⫹
mu an s ha a ose in m 1⌬cells exp essing AID unde high
ansc ip ion condi ions. The URA3 ORF om 19 and 53
independen lacZ
⫹
U a
⫺
mu an s ob ained om he wild ype
wi h and wi hou AID o e exp ession, espec i ely, was also
sequenced (SI Fig. 6 Aand B). Resul s a e summa ized in Table
1 and SI Table 3. In wild- ype cells no exp essing AID ⬍60%
(11 o 19) we e poin mu a ions and only 9% o hese (1 o 11)
we e wi hin he WRC/GYW a ge mo i o AID. In con as , in
he wild ype and m 1⌬exp essing AID, he pe cen age o poin
mu a ions inc eased o 86% (46 o 53) and 91% (22 o 24),
espec i ely. O hese, 65% (30 o 46) and 54% (12 o 22),
espec i ely, occu ed a he WRC/GYW mo i . Al oge he , he
da a indica e ha AID has a p e e en ial unc ion as a cy osine
deaminase a he WRC/GYW mo i , consis en wi h he p e -
e en ial a ge o AID deamina ion obse ed in i o (11).
As AID deamina es C wi hin ssDNA, mu a ions occu ing a
C should e lec he ac ion o AID di ec ly on he NT s and,
whe eas mu a ions a G e lec he ac ion o AID on C in he T
s and. The e o e, we analyzed he NT:T a io o AID-induced
mu a ions in bo h s ains. Whe eas in he wild ype exp essing
AID he NT:T a io o mu a ions was 0.75:1 (16:21), his a io
inc eased signi ican ly (P⬍0.05) o 3:1 (16:5) in m 1⌬. Impo -
an ly, om hese mu a ions, he WRC AID- a ge mo i was
mu a ed a an NT:T a io o 1:1 in he wild ype (15:15) and 5:1
in m 1⌬(10:2). The lack o s and p e e ence o he mu a ion
pa e n p oduced by AID in he wild ype was no obse ed
p e iously (29) and is consis en wi h he ansien opening o
DNA s ands by ansc ip ion p esumably p omo ed by nega i e
supe coiling. By con as , he pu a i e impai men o mRNP
biogenesis caused by m 1⌬p omo es a mo e s able o ma ion o
ssDNA on he NT s and and makes i an app op ia e a ge o
he ac ion o human AID. The ans e sions: ansi ions a io
also dec eased om 11:0 in he wild ype o 35:11 and 13:9 in
wild- ype and m 1⌬cells exp essing AID, espec i ely (Table 1).
The inc ease in ansi ions is consis en wi h he di ec ac ion o
AID in deamina ing C o U, which leads o a C- o-T ansi ion
(G o A in he opposi e s and). Consis en ly, he changes
p oduced by AID we e mo e e iden o C (o G) wi h a s ong
bias o C in m 1⌬cells (Table 1).
AID-Induced Recombina ion Is T ansc ip ion-Dependen and S ongly
S imula ed by THO Mu a ions in Yeas Cells. Finally, AID has also
been shown o con ibu e o ecombina ion in yeas (28, 29). This
inc ease e lec s he abili y o AID o, in a di ec o an indi ec
Table 1. Spon aneous (ⴚAID) and AID-induced (ⴙAID) base
subs i u ions in he wild ype (lacZ
ⴙ
U a
ⴚ
) and m 1⌬mu an s
(lacZ
ⴚ
U a
ⴙ
) classi ied acco ding o di e en sequence ea u es
Sequence ea u es WT WT ⫹AID m 1⌬⫹AID
Mu a ions a C 0 16 16
Mu a ions a G 7 21 5
Mu a ions a WRC 0 15 10
Mu a ions a GYW 1 15 2
T ans e sions 11 35 13
T ansi ions 0 11 9
Poin mu a ions 11 46 22
To al mu a ions 19 53 24
DNA sequences a e shown in SI Fig. 6 Aand B.
Go´ mez-Gonza´lez and Aguile a PNAS
兩
May 15, 2007
兩
ol. 104
兩
no. 20
兩
8411
GENETICS
way, induce he o ma ion o DNA double-s and b eaks, which
in yeas a e p e e en ially epai ed by ecombina ion, whe eas in
human cells hey a e epai ed by nonhomologous end-joining.
Indeed nonhomologous end-joining is belie ed o be he majo
mechanism ha leads o class swi ching in B cells (see e . 37).
We expec ed ha AID should also inc ease ecombina ion in a
ansc ip ion-dependen manne in wild- ype yeas and much
mo e s ongly in m 1⌬. Fo his pu pose we used a plasmid–
ch omosome ecombina ion assay be ween a CEN-plasmid
leu2- unde con ol o he egula ed Te p omo e and a
ch omosomal leu2-k mu a ion (38) in he wild ype. As seen in
Fig. 3A, AID did no change he equency o gene con e sion
o leu2- a low ansc ip ion. Howe e , i caused a 40- old
inc ease a high ansc ip ion. Thus, like he mu agenic ac i i y,
he hype ecombinogenic ac i i y o AID is ansc ip ion-
dependen in yeas . These indings a e consis en wi h he iew
ha bo h AID-induced mu a ion and ecombina ion a e he
ou pu p oduc s o a common in e media e.
Because di ec - epea ecombinan s a e he ype o e en
p ima ily inc eased by THO mu a ions (24), a di ec - epea
ecombina ion assay was de eloped o he analysis o dele ions
o assess whe he THO mu a ions also s eng hen AID-induced
ecombina ion. The assay is based on wo unca ed GFP
epea s, which a e unde con ol o he GAL1 p omo e and
in e up ed by a lacZ sequence. GFP
⫹
ecombinan s a e sco ed
di ec ly by FACS. As shown in Fig. 3B,m 1⌬and AID inde-
penden ly p oduced a 6- old inc ease in GFP
⫹
dele ions,
whe eas he combina ion o m 1⌬and AID exp ession caused
a s ong syne gis ic 100- old inc ease. The e o e, we conclude
ha impai men o mRNP o ma ion caused by THO mu a ions
enhances he ac ion o AID in a ansc ip ion-dependen man-
ne , esul ing in a s ong induc ion o bo h poin mu a ions and
di ec - epea ecombina ion.
Discussion
In his wo k we show ha AID can unc ion in yeas THO
mu an s simila o ha in human B cells. In wild- ype yeas no
only can AID induce hype mu a ion and hype ecombina ion, as
i was p e iously shown, bu , impo an ly, his induc ion depends
on ansc ip ion. We show ha such ansc ip ion-dependen
hype mu a ion and hype ecombina ion e ec s a e weak in he
wild ype bu s ongly and syne gis ically enhanced in mu an s o
he THO complex. The syne gis ic e ec o AID and THO
mu a ions is also demons a ed a he molecula le el by showing
ha mos AID mu a ions occu ed p e e en ially on he NT
DNA s and. Because THO mu an s a e a ec ed in mRNP
biogenesis (21, 39), ou esul s sugges ha AID ac ion can be
enhanced by subop imal mRNP o ma ion. These esul s open
he possibili y ha a subop imal o ma ion o he mRNP would
enhance he p obabili y o he nascen mRNA o o m a
DNA:RNA hyb id, lea ing he NT s and as single-s anded.
Such a single-s anded egion would be s abilized by he G-loops
in S egions, so ha he inal s uc u e p omo es AID a ge ing
(Fig. 4).
Fo his s udy we i s cha ac e ized new in i o assays o he
analysis o o wa d mu a ions unde he con ol o ansc ip ion-
Fig. 3. E ec o AID on ansc ip ion-associa ed ecombina ion. (A) Spon-
aneous (⫺AID) and AID-induced (⫹AID) equency o gene con e sion unde
high (ON) and low (OFF) ansc ip ion be ween a plasmid leu2- and a ch o-
mosomal leu2-k allele in a wild- ype s ain. (B) E ec o AID exp ession on
di ec - epea ecombina ion in he GLG sys em. The equency o GFP⫹ e-
combinan s (signals inside he box) is indica ed. yaxis, g een luo escence
(FL1H); xaxis, unspeci ic luo escence (FL2H). A diag am o each sys em is
shown on he op o each panel.
Fig. 4. A model o explain s imula ion o genome ins abili y by AID in yeas
THO mu an s compa ed wi h he S egions in B cells. (A) T ansc ip ion and
mRNP o ma ion media ed by THO in wild- ype and THO-deple ed yeas cells.
In wild- ype yeas , AID is able o ac on he DNA when i is ansc ibed.
Nega i e supe coiling behind he elonga ing RNA polyme ase II (RNAPII)
allows a weak AID ac ion. In THO-deple ed yeas , a co ansc ip ional R-loop
can be o med, allowing AID o ac on he non ansc ibed ssDNA, causing
s ong hype mu a ion and hype ecombina ion. (B) A pu a i e example o
ansc ip ion and mRNP o ma ion in he S egion o Ig genes in human B cells.
The G- ich S egion could be e ac o y o a numbe o RNA-binding p o eins,
he e o e o ming a subop imal mRNP a ha pa icula egion. Nascen RNA
S egions could lead o local co ansc ip ional R-loops o G-loops ha would
allow AID ac ion.
8412
兩
www.pnas.o g兾cgi兾doi兾10.1073兾pnas.0702836104 Go´ mez-Gonza´lez and Aguile a
egula ed p omo e s in yeas , wi h which we ha e been able o
show ha ansc ip ion signi ican ly enhances mu a ion, consis-
en wi h a p e ious esul (32), in which his s imula ion was
accompanied by a change in he mu a ion spec um (40). TAM
has also been shown in bac e ia wi h bo h spon aneous and
damage-induced mu a ions ( e iewed in e . 33).
Mu an s o he THO complex ha e a s ong ansc ip ion-
dependen hype ecombina ion pheno ype, as shown in di ec -
epea sys ems (41, 42). Al hough ecombina ion is also en-
hanced by ansc ip ion, he mechanisms unde lying
ansc ip ion-associa ed ecombina ion may be di e en om
hose o TAM. Thus, impai men o eplica ion p og ession by
ansc ip ion may be a c i ical ac o igge ing ansc ip ion-
associa ed ecombina ion (43, 44), bu his is no ob ious o
TAM. Ou s udy e eals ha THO mu a ions lead o a low
hype mu a o pheno ype in he LAUR sys em, which is based on
a highly exp essed lacZ gene (Fig. 1). I is impo an o no e ha
he gene exp ession de ec and hype ecombina ion pheno ypes
o THO mu an s ha e been shown o be pa icula ly s ong a
GC- ich DNA sequences such as lacZ (26). The m 1⌬-induced
mu a ions ha e a clea ly di e en pa e n om hose ob ained
in wild- ype cells. In m 1⌬cells all mu a ions occu ed wi hin
he lacZ gene o he lacZ::URA3 usion, whe eas in he wild ype
40% o he mu a ions occu ed a URA3. In e es ingly, mu a-
ions o igina e a he same DNA egion, lacZ, whe e hype e-
combina ion is s ongly s imula ed (41). Indeed, hype mu a ion
associa ed wi h THO mu a ions is pa ially dependen on Rad52,
and pa o he hype mu a ion e en s obse ed in he LAUR
assay a e caused by Rad52-dependen DNA ea angemen s (see
Fig. 1B), he e o e being a consequence o he s ong hype e-
combina ion pheno ype o hese mu an s. This esul indica es
ha he ecombinogenic s uc u es gene a ed in THO mu an s
may no be mu agenic by hemsel es.
Ou s udy p o ides an app op ia e sys em o he analysis o
hype mu a ion caused by AID. In i o and in i o expe imen s
in E. coli ha e shown he ansc ip ion dependency o AID
deamina ion and hype mu a ion, espec i ely (5–8, 12). A hy-
pe ecombina ion and hype mu a ion e ec o AID was also
obse ed in yeas (28), bu whe he his was ansc ip ion-
dependen was no de e mined. In his wo k we show ha
hype mu a ion and hype ecombina ion caused by AID in yeas
a e also ansc ip ion-dependen , con i ming he hypo hesis ha
he ac ion o AID equi es ansc ip ion o he a ge sequence
(4, 6, 12). This is shown no only by an inc ease in he equency
o mu a ions bu by he ac ha mos o hese mu a ions occu
a he AID-p e e ed DNA mo i WRC (see Table 1). Impo -
an ly, he AID e ec obse ed in yeas is s ill low, which
sugges s ha he e mus be addi ional ac o s in human B cells
esponsible o he speci ic ac ion o AID on i s a ge DNA
sequences. Indeed, ou mu a ion spec um analysis did no
e eal a subs an ial numbe o A-T mu a ions (SI Table 3), which
ypically compose hal o he mu a ional spec um o SHM (13),
consis en wi h he iew ha addi ional ac o s (e.g., misma ch
epai , e o -p one polyme ases, ch oma in s uc u e, e c.) a e
egula ed in B cells o media e SHM (45).
The sligh AID ac ion obse ed in wild- ype yeas may be
explained by he p edic ion ha he nega i ely supe coiled DNA
pu a i ely accumula ed behind an elonga ing RNA polyme ase
can p omo e ansien o ma ion o ssDNA (14, 46), which is he
p e e en ial a ge o AID (5, 6). The esul s indica ing ha
AID-induced mu a ions in wild- ype yeas occu ed a a 1:1 a io
in he NT:T s ands would be consis en wi h his iew, implying
ha bo h s ands ha e equal p obabili y o becoming single-
s anded and, he e o e, o be accessed by AID. P e ious da a
in yeas ha e e ealed a di e en NT:T s and a io o mu a ions
in he CAN1 gene (29). I is likely, he e o e, ha he pa e n o
mu a ions caused by AID in yeas may no be unique bu
dependen on a numbe o pa ame e s ha can include nucle-
o ide sequence, ch oma in s uc u e, le els o exp ession, e c.
Fu he in es iga ions using di e en genes as mu a ion a ge s
may be equi ed o ob ain a mo e comple e iew o AID ac ion
in yeas . Ne e heless, ou esul s clea ly show ha a high impac
o AID ac ion in yeas is obse ed only when mRNP biogenesis
is comp omised by THO mu a ions. We de ec ed a 5:1 bias in
a o o he NT s and, a esul ha can be explained only i a
mo e s able non ansc ibed ssDNA s uc u e is o med and
simul aneously he T s and is p o ec ed om AID ac ion.
Because THO mu a ions ha e p e iously been shown o induce
co ansc ip ional R-loop o ma ion, as de e mined in he same
GC- ich lacZ gene used in his s udy (22), and S egions also o m
R-loops (16), o ma ion o R-loops may be a na u al way o
s abilizing ssDNA o AID ac i i y and o p o ec ing he T s and
in he o m o he DNA:RNA hyb id. Ou da a in yeas show,
he e o e, ha he s ong AID e ec on hype mu a ion in
di e en THO-deple ed yeas cells is consis en wi h he o -
ma ion o R-loops du ing ansc ip ion (22, 34). In e es ingly,
ecen epo s ha e shown ha AID in e ac s in i o wi h he
elonga ion complex (47). Because RNA polyme ase II p og es-
sion is educed in THO mu an s (23), we canno disca d ha , in
addi ion, AID ac ion in THO mu an s could be s eng hened by
a longe ime in which AID could in e ac wi h he elonga ion
complex, he e o e inc easing i s ime o ac on DNA.
AID deamina ion seems o igge he in e media es leading
o class swi ch ecombina ion in B cells, and some e idence
sugges s ha class swi ching occu s by nonhomologous end-
joining (37). Consequen ly, i is belie ed ha AID deamina ion
is he i s e en o a p ocess ending in double-s and b eaks.
Whe he his pu a i e double-s and b eak is pe o med di ec ly
by AID o is o med a e eplica ion is a ques ion ye o be
add essed. Because double-s and b eaks in yeas a e p e e en-
ially epai ed by ecombina ion, i was expec ed ha AID would
a ec ecombina ion simila ly o mu a ion. I was p e iously
epo ed ha AID induces ecombina ion in yeas (28, 29). In
his s udy we show ha AID is able o inc ease he equency o
gene con e sion only unde high ansc ip ion (Fig. 3A), and,
using a di ec - epea ecombina ion assay, which allows us o
de ec hype ecombina ion in THO mu an s, we ha e been able
o see a s ong and syne gis ic e ec o AID and THO mu a ions
on ecombina ion (Fig. 3B). The e o e, in yeas cells THO
mu a ions enhance he ansc ip ion-dependen AID e ec ,
whe he de ec ed as hype mu a ion o hype ecombina ion.
In summa y, ou da a clea ly show ha yeas THO mu an s
can be used as a model in i o sys em o he s udy o AID ac ion.
No only a e hype ecombina ion and hype mu a ion obse ed
as a consequence o AID, bu hey occu in a ansc ip ion-
dependen manne . Mo e impo an ly, he pa e n o mu a ions
i s he p e iously shown pa e n caused by AID in i o (11),
mu a ions occu ing p e e en ially a he WRC mo i . Howe e ,
he pa e n obse ed in i o, in which mu a ions a e p e e en-
ially ound a he NT s and (6, 11, 29), is seen only in THO
mu an s. The e o e, ou esul s open he possibili y ha mRNP
biogenesis con olled AID ac ion. I would be in e es ing o
know whe he he high G con en o he S- egion mRNA in B
cells migh limi i s capaci y o be assembled in an op imal
mRNP, as THO mu a ions in gene al do in yeas , he eby
inc easing he eac i i y o he nascen RNA wi h he DNA
empla e (see Fig. 4). Yeas THO mu an s can, he e o e, be an
excellen ool in deciphe ing he molecula mechanisms by which
AID can induce bo h hype mu a ion and hype ecombina ion.
Ma e ials and Me hods
S ains and Plasmids. We used W303–1A isogenic s ains
WMK-2A (m 1⌬::KAN), U678-1C (hp 1⌬::HIS3), WRS52-4B
( ad52⌬::KAN) [desc ibed p e iously (24)], WMR52-1D and
WMR52-4D ( ad52⌬KAN m 1⌬::KAN) (ob ained in his s udy),
and he BY de i a i e BER08-64A (his3::leu2-k) (38).
Go´ mez-Gonza´lez and Aguile a PNAS
兩
May 15, 2007
兩
ol. 104
兩
no. 20
兩
8413
GENETICS

Cen ome ic plasmids pMR260 (30) ca ying KAR1 unde
con ol o he GAL1 p omo e , pCM184-LAUR (31) con aining
lacZ::URA3 unde he e p omo e , and pCM184-L2HO used
o he plasmid–ch omosome ecombina ion assay (38) we e
desc ibed p e iously. Cen ome ic plasmids p414GALAID and
p413GALAID ca ying he human AID ORF unde he GAL1
p omo e we e ob ained by PCR ampli ica ion o AID om
pGAID (6) using p ime s 5⬘-CTCTGGACGAAATTCCATG-
GACAGCCTCTTC-3⬘and 5⬘-CCTGGAAGCTCGAGT-
CAAAGCTCCAAAGTA-3⬘and cloning in o he EcoRI-XhoI-
diges ed pRS414GAL and pRS413GAL (38), espec i ely.
Cen ome ic plasmid pGLG con aining he GFP di ec - epea
cons uc was ob ained by ampli ying GFP⌬3⬘and GFP⌬5⬘
agmen s om pUG34 and pUG23 plasmids (48) by PCR using
p ime s 5⬘-ACTAGTGCCATGATGTAAACATTG-3⬘,5⬘-
AATACAGGGTCGTCAGAT-3⬘,5⬘-TTAAAGCCTTC-
GAGCGTCGGGCCCCCTTCT-3⬘, and 5⬘-ACTGGTCGA-
CTCCCAATTTTGGTTGAAT-3⬘and subcloning in o he
SpeI-XbaI and ApaI-SalI si es o pRS413GAL, espec i ely. The
lacZ ORF was subcloned a a BamHI be ween he GFP epea s.
Mu a ion and Recombina ion Analysis. Fo he GKAR1 mu a ion
assay, cells we e cul u ed o e nigh in SC medium con aining 2%
glyce ol-lac a e as a ca bon sou ce. A e wa d, he cul u e was
spli in wo, one wi h 2% glucose ( ansc ip ion OFF) and he
o he wi h 2% galac ose ( ansc ip ion ON), and cul u ed o
ano he 9 h be o e mu an selec ion. Fo he LAUR mu a ion
assay and he plasmid–ch omosome ecombina ion assay cells
we e cul u ed in SC pla es wi h ( ansc ip ion OFF) o wi hou
( ansc ip ion ON) 5
␮
g/ml doxycycline, om which indepen-
den colonies we e ob ained o he mu a ion o ecombina ion
analyses. Ka
⫺
and U a
⫺
mu an s we e selec ed on SC con ain-
ing 2% galac ose and SC plus 5- luo oo o ic acid, espec i ely,
and Leu
⫹
ecombinan s on SC-Leu pla es. lacZ
⫹
and lacZ
⫺
we e
dis inguished by colo on SC plus X-gal medium. Median
mu a ion and ecombina ion equencies we e ob ained by
luc ua ion es s as he median alue o six independen colonies
isola ed om SC pla es. The inal equency gi en o each s ain
and condi ion is he mean and s anda d de ia ion o h ee o ou
median alues.
Miscellanea. Yeas me hodology,
␣
-
32
P-labeled DNA p obes, and
Sou he n and No he n blo s we e pe o med by ollowing
s anda d p ocedu es. GFP luo escence was de e mined in a
FACSCalibu (Bec on Dickinson, San Jose, CA) om 10
6
cells
g own in SC o e nigh and esuspended in 1 ml o H
2
O.
We hank A. S. Bhagwa (Wayne Uni e si y, De oi , MI) o kindly
p o iding he hAID clone, J. Escalan e o echnical assis ance, J. F. Ruiz
o help ul discussions, F. Co e´s-Ledesma and H. Gailla d o c i ical
eading o he manusc ip , and D. Haun o s yle supe ision. This wo k
was suppo ed by Spanish Minis y o Science and Educa ion G an s
SAF2003-00204 and BFU2006-05276 and Jun a de Andalucı´a G an s
CVI102 and CVI624. B.G.-G. was ecipien o a p edoc o al aining
g an om he Spanish Minis y o Science and Educa ion.
1. Re y P, Mu o T, Le y Y, Geissmann F, Plebani A, Sanal O, Ca alan N,
Fo eille M, Du ou cq-Labelouse R, Genne y A, e al. (2000) Cell 102:565–575.
2. Okazaki IM, Kinoshi a K, Mu ama su M, Yoshikawa K, Honjo T (2002) Na u e
416:340–345.
3. Mu ama su M, Kinoshi a K, Faga asan S, Yamada S, Shinkai Y, Honjo T
(2000) Cell 102:553–563.
4. Pe e sen-Mah SK, Ha is RS, Neube ge MS (2002) Na u e 418:99–103.
5. Chaudhu i J, Tian M, Khuong C, Chua K, Pinaud E, Al FW (2003) Na u e
422:726–730.
6. Sohail A, Klapacz J, Sama anayake M, Ullah A, Bhagwa AS (2003) Nucleic
Acids Res 31:2990–2994.
7. B ans ei e R, Pham P, Scha MD, Goodman MF (2003) P oc Na l Acad Sci
USA 100:4102–4107.
8. Pe e s A, S o b U (1996) Immuni y 4:57–65.
9. Hein K, Lo enz MG, Siebenko en G, Pe y K, Ch is ine R, Radb uch A (1998)
J Exp Med 188:2369–2374.
10. Ma omo SA, Fu D, Yang WW, Joshi NS, Gea ha PJ (2005) J Immunol
174:7787–7791.
11. Pham P, B ans ei e R, Pe uska J, Goodman MF (2003) Na u e 424:103–107.
12. Rami o AR, S a opoulos P, Janko ic M, Nussenzweig MC (2003) Na
Immunol 4:452–456.
13. Mils ein C, Neube ge MS, S aden R (1998) P oc Na l Acad Sci USA 95:8791–
8794.
14. Shen HM, S o b U (2004) P oc Na l Acad Sci USA 101:12997–13002.
15. Chaudhu i J, Khuong C, Al FW (2004) Na u e 430:992–998.
16. Yu K, Chedin F, Hsieh CL, Wilson TE, Liebe MR (2003) Na Immunol
4:442–451.
17. Dempsey LA, Sun H, Hanakahi LA, Maizels N (1999) J Biol Chem 274:1066–
1071.
18. Duque e ML, Handa P, Vincen JA, Taylo AF, Maizels N (2004) Genes De
18:1618–1629.
19. Duque e ML, Pham P, Goodman MF, Maizels N (2005) Oncogene 24:5791–
5798.
20. Yu K, Roy D, Bay amyan M, Hawo h IS, Liebe MR (2005) Mol Cell Biol
25:1730–1736.
21. Aguile a A (2005) Cu Opin Cell Biol 17:242–250.
22. Hue as P, Aguile a A (2003) Mol Cell 12:711–721.
23. S asse K, Masuda S, Mason P, P anns iel J, Oppizzi M, Rod iguez-Na a o
S, Rondon AG, Aguile a A, S uhl K, Reed R, e al. (2002) Na u e 417:304–308.
24. Cha ez S, Beilha z T, Rondon AG, E djumen -B omage H, Temps P,
S ejs up JQ, Li hgow T, Aguile a A (2000) EMBO J 19:5824–5834.
25. Voyno V, Ve s epen KJ, Jansen A, Runne VM, Bu a owski S, Fink GR
(2006) P oc Na l Acad Sci USA 103:14423–14428.
26. Cha ez S, Ga cia-Rubio M, P ado F, Aguile a A (2001) Mol Cell Biol
21:7054–7064.
27. Li X, Manley JL (2005) Cell 122:365–378.
28. Pol o a sky VP, Wilson SH, Kunkel TA, Pa lo YI (2004) J Immunol 172:4308–
4313.
29. Mayo o VI, Rogozin IB, Adkison LR, F ahm C, Kunkel TA, Pa lo YI (2005)
BMC Immunol 6:10.
30. Rose MD, Fink GR (1987) Cell 48:1047–1060.
31. Jimeno S, Rondon AG, Luna R, Aguile a A (2002) EMBO J 21:3526–3535.
32. Da a A, Jinks-Robe son S (1995) Science 268:1616–1619.
33. Aguile a A (2002) EMBO J 21:195–201.
34. Hue as P, Ga cia-Rubio ML, Wellinge RE, Luna R, Aguile a A (2006) Mol
Cell Biol 26:7451–7465.
35. Mo ey NJ, G eene CN, Jinks-Robe son S (2000) Gene ics 154:109–120.
36. Rogozin IB, Kolchano NA (1992) Biochim Biophys Ac a 1171:11–18.
37. Rooney S, Chaudhu i J, Al FW (2004) Immunol Re 200:115–131.
38. Ga cia-Rubio M, Hue as P, Gonzalez-Ba e a S, Aguile a A (2003) Gene ics
165:457–466.
39. Vincigue a P, S u z F (2004) Cu Opin Cell Biol 16:285–292.
40. Lippe MJ, F eedman JA, Ba be MA, Jinks-Robe son S (2004) Mol Cell Biol
24:4801–4809.
41. Cha ez S, Aguile a A (1997) Genes De 11:3459–3470.
42. P ado F, Pi ua JI, Aguile a A (1997) EMBO J 16:2826–2835.
43. Wellinge RE, P ado F, Aguile a A (2006) Mol Cell Biol 26:3327–3334.
44. P ado F, Aguile a A (2005) EMBO J 24:1267–1276.
45. Woo CJ, Ma in A, Scha MD (2003) Immuni y 19:479–489.
46. D ole M, Bi X, Liu LF (1994) J Biol Chem 269:2068–2074.
47. Besme E, Ma ke E, Papa asiliou FN (2006) Mol Cell Biol 26:4378–4385.
48. Nieden hal RK, Riles L, Johns on M, Hegemann JH (1996) Yeas 12:773–786.
8414
兩
www.pnas.o g兾cgi兾doi兾10.1073兾pnas.0702836104 Go´ mez-Gonza´lez and Aguile a