No el ea u es o elome e biology e ealed
by he absence o elome ic DNA me hyla ion
Alejand o Vega-Vaque o,
1,4
Gianca lo Bono a,
2,4
Ma co Mo selli,
2,4
Ma ía I. Vaque o-Sedas,
3
Liudmilla Rubbi,
2
Ma eo Pelleg ini,
2
and Miguel A. Vega-Palas
3
1
Technical Supe io School o In o ma ics Enginee ing, Uni e si y o Se ille, 41080 Se ille, Spain;
2
Depa men o Molecula , Cell and
De elopmen al Biology, Uni e si y o Cali o nia, Los Angeles, Cali o nia 90095, USA;
3
Ins i u e o Vege al Biochemis y and
Pho osyn hesis, CSIC-Uni e si y o Se ille, IBVF (CSIC-US), 41092 Se ille, Spain
Cy osine me hyla ion egula es he leng h and s abili y o elome es, which can a ec a wide a ie y o biological ea u es,
including cell di e en ia ion, de elopmen , o illness. Al hough i is well es ablished ha sub elome ic egions a e me hyl-
a ed, he p esence o me hyla ed cy osines a elome es has emained con o e sial. He e, we ha e analyzed mul iple bisul-
i e sequencing s udies o add ess he me hyla ion s a us o A abidopsis haliana elome es. We ound ha he le els o
es ima ed elome ic DNA me hyla ion a ied among s udies. In e es ingly, we es ima ed highe le els o elome ic DNA
me hyla ion in s udies ha p oduced C- ich elome ic s ands wi h lowe e iciency. Howe e , hese high me hyla ion es-
ima es a ose due o expe imen al limi a ions o he bisul i e echnique. We ound a simila phenomenon o mi ochond ial
DNA: The le els o mi ochond ial DNA me hyla ion de ec ed we e highe in expe imen s wi h lowe mi ochond ial ead
p oduc ion e iciencies. Based on expe imen s wi h high elome ic C- ich s and p oduc ion e iciencies, we concluded ha
A abidopsis elome es a e no me hyla ed, which was con i med by me hyla ion-dependen es ic ion enzyme analyses.
Thus, ou s udies indica e ha elome es a e e ac o y o de no o DNA me hyla ion by he RNA-di ec ed DNA me hyl-
a ion machine y. This esul , oge he wi h p e iously epo ed da a, e eals ha sub elome ic DNA me hyla ion con ols
he homeos asis o elome e leng h.
[Supplemen al ma e ial is a ailable o his a icle.]
Telome es gua an ee he comple e eplica ion o ch omosomal e -
mini, p e en genome ins abili y, and in luence ele an sys emic
p ocesses like aging, cance , o illness (Blackbu n 2010). The
leng h o elome es and he ch oma in o ganiza ion o elome ic
egions in luence elome e unc ions. Hence, he epigene ic ma ks
ha label elome ic egions, which include elome es and sub elo-
me es, play impo an oles in elome e biology (Blasco 2007;
Gala i e al. 2013; Gi aud-Panis e al. 2013).
One o he majo epigene ic signa u es ound in euka yo es is
cy osine me hyla ion. This DNA modi ica ion egula es mul iple
p ocesses in plan s and animals, including he homeos asis o elo-
me e leng h (Blasco 2007; Suzuki and Bi d 2008; Ooi e al. 2009;
Law and Jacobsen 2010; Cas el and Ma ienssen 2013; Og ocká
e al. 2014; Vaque o-Sedas and Vega-Palas 2014). Mammalian
DNA me hyla ion is p ima ily ound in he CG con ex
(Ramsahoye e al. 2000; Lis e e al. 2009). In con as , plan s
ha e signi ican le els o DNA me hyla ion in all sequence con-
ex s (CG, CHG, and CHH, whe e H can be A, C, o T) (Law and
Jacobsen 2010).
Al hough sub elome ic DNA me hyla ion has been epo ed
in animals and plan s, he p esence o DNA me hyla ion a elo-
me es emains an open ques ion in bo h kingdoms (Blasco 2007;
V bsky e al. 2010; Vaque o-Sedas e al. 2011; Og ocká e al.
2014). The me hyla ion s a us o mammalian elome es has no
been in es iga ed because, as men ioned abo e, mammals ha e
low le els o non-CG me hyla ion, which is he ype o DNA me h-
yla ion ha should be associa ed wi h elome ic sequences
(CCCTAA in mammals and CCCTAAA in plan s). In u n, al-
hough he me hyla ion le els o plan elome es ha e been s ud-
ied by di e en g oups, hey emain con o e sial (V bsky e al.
2010; Maje o á e al. 2011a,b; Vaque o-Sedas and Vega-Palas
2011a,b; Vaque o-Sedas e al. 2011, 2012; Og ocká e al. 2014).
The e o e, i is impo an o se le he me hyla ion s a us o
elome es.
The expe imen al analysis o he epigene ic ma ks ha label
elome es is complica ed by he in luence o sub elome es and/
o he In e s i ial Telome ic Sequences (ITSs), which a e usually
p esen a pe icen ome ic egions and sub elome es (Vaque o-
Sedas and Vega-Palas 2011b). On he one hand, elome es and sub-
elome es canno be di e en ia ed by mic oscopy echniques. On
he o he hand, ITSs can in e e e wi h he analyses o elome ic
ch oma in s uc u e by ch oma in immunop ecipi a ion ollowed
by hyb idiza ion wi h a elome ic p obe. Mo eo e , ITSs migh be
iden i ied as elome es in massi ely pa allel DNA sequencing s ud-
ies (Vaque o-Sedas e al. 2012; Vega-Palas and Vaque o-Sedas
2013). Hence, he analysis o he epigene ic modi ica ions p esen
a elome es should be ca e ully designed.
4
These au ho s should be conside ed equal i s au ho s.
Co esponding au ho s: [email p o ec ed], ma eop@mcdb.
ucla.edu
A icle published online be o e p in . A icle, supplemen al ma e ial, and publi-
ca ion da e a e a h p://www.genome.o g/cgi/doi/10.1101/g .202465.115.
© 2016 Vega-Vaque o e al. This a icle is dis ibu ed exclusi ely by Cold
Sp ing Ha bo Labo a o y P ess o he i s six mon hs a e he ull-issue publi-
ca ion da e (see h p://genome.cshlp.o g/si e/misc/ e ms.xh ml). A e six
mon hs, i is a ailable unde a C ea i e Commons License (A ibu ion-
NonComme cial 4.0 In e na ional), as desc ibed a h p://c ea i ecommons.
o g/licenses/by-nc/4.0/.
Resea ch
26:1047–1056 Published by Cold Sp ing Ha bo Labo a o y P ess; ISSN 1088-9051/16; www.genome.o g Genome Resea ch 1047
www.genome.o g
Cold Sp ing Ha bo Labo a o y P ess on Sep embe 11, 2016 - Published by genome.cshlp.o gDownloaded om
The s udy o elome es independen ly o ITSs may be acili a -
ed by he ac ha hey usually ha e di e en sequence o ganiza-
ions. Al hough elome es a e essen ially composed o andem
a ays o pe ec elome ic epea s, ITSs usually con ain pe ec
elome ic epea s in e spe sed wi h degene a e epea s. In ac , i
is uncommon o ITSs o con ain long s e ches o pe ec andem
elome ic epea s (Lin and Yan 2008; Gámez-A jona e al. 2010).
He e, we ha e add essed he me hyla ion s a us o A abidopsis
haliana elome es by analyzing da a p oduced by genome-wide
bisul i e sequencing s udies and by pe o ming me hyla ion-
dependen es ic ion analyses. These s udies e ealed ha
A abidopsis elome es a e no me hyla ed.
Resul s
In silico analysis o elome ic DNA me hyla ion
To gain insigh in o he me hyla ion s a us o A abidopsis elo-
me es, we es ima ed hei me hyla ion le els om di e en ge-
nome-wide bisul i e sequencing s udies (Supplemen al Table S1).
These s udies had been pe o med in di e en labo a o ies and in-
ol ed he ea men o DNA wi h sodium bisul i e, he PCR am-
pli ica ion o he esul ing DNA samples, and he sequencing o
he bisul i e modi ied DNA s and. Since bisul i e deamina es
unme hyla ed cy osines gene a ing u acil, unme hyla ed cy o-
sines a e de ec ed as hymines a e PCR ampli ica ion. In con as ,
me hyla ed cy osines a e no modi ied by bisul i e and emain as
cy osines a e ampli ica ion (F omme e al. 1992; Cla k e al.
1994).
The eads ep esen ing elome es in he bisul i e sequencing
s udies should ollow a pe ec andem elome ic epea pa e n,
ep esen ed as (YYYTAAA)
n
, in which Y is C o T depending on
whe he he elome ic cy osines a e con e ed o no . We es ima -
ed ha eads con aining abou se en pe ec andem elome ic e-
pea s should essen ially ep esen elome es in he bisul i e
expe imen s (Me hods). Only 20 genomic (YYYTAAA)
7
sequences
can be ound in he A abidopsis genome, wi h one o hem being
300 bp in leng h (Fig. 1).
Since he leng h o he eads in many o he s udies selec ed
was 50 bp, we conside ed elome ic eads as hose in which he i s
50 bp ollow he (YYYTAAA)
n
pa e n, e en i hey we e longe . We
deno ed hese eads o agmen o eads as 50-bp Telome ic Reads
(50TRs).
Va iable le els o elome ic DNA me hyla ion a e es ima ed om
di e en bisul i e sequencing s udies
Be o e s udying he me hyla ion s a us o A abidopsis elome es,
we decided o e alua e he e iciency o elome ic C- ich s and
gene a ion a e bisul i e ea men . We ound ha he p oduc ion
e iciency o he C- ich s and a ied among he di e en bisul i e
expe imen s. These e iciencies we e calcula ed as he numbe o
50TRs di ided by he o al numbe o mapped eads and we e ex-
p essed as pe cen ages (Supplemen al Table S1).
We nex decided o de e mine wo pa ame e s o es ima e he
deg ee o cy osine me hyla ion a elome es: i s , he pe cen age
o uncon e ed cy osines wi hin 50TRs, and second, he pe cen -
age o 50TRs ha con ained a leas one uncon e ed cy osine.
Ou analyses e ealed ha he pe cen ages o uncon e ed cy o-
sines wi hin 50TRs de i ed om wild- ype A abidopsis plan s a -
ied depending on he s udy analyzed (Fig. 2). These pe cen ages
anged om 1.5% o 40%. In e es ingly, he compa ison o he di -
e en s udies indica ed ha he pe cen age o uncon e ed cy o-
sines a ied wi h he elome ic C- ich s and p oduc ion
e iciency. High pe cen ages o uncon e ed cy osines, om 5%
o 40%, we e de ec ed when he elome ic C- ich s and p oduc-
ion was < 8% × 10
−4
. In u n, he pe cen ages o uncon e ed cy-
osines we e be ween 1.5% and 5% when he elome ic C- ich
s and p oduc ion was > 8% × 10
−4
(Fig. 2). In ag eemen wi h
his obse a ion, we no iced ha he pe cen age o 50TRs con ain-
ing cy osines also dec eased when he p oduc ion e iciency o he
C- ich s and inc eased (Fig. 2). Hence, he le els o cy osine me h-
yla ion es ima ed a elome es dec eased wi h he p oduc ion e i-
ciency o he elome ic C- ich s and.
Expe imen al limi a ions o he bisul i e echnique lead
o o e es ima ion o elome ic DNA me hyla ion
To de e mine whe he he an i-co ela ion de ec ed be ween p o-
duc ion e iciency and cy osine uncon e sion was also obse able
a o he genomic loci, we assessed bo h a iables in mi ochond ial
Figu e 1. Majo A abidopsis ITSs con aining pe ec elome ic epea s. Genomic DNA sequences con aining (YYYTAAA)
7
s ings and se e al nucleo ides
ups eam and downs eam we e aligned using he compu e p og am, CLUSTALW. These sequences we e iden i ied using he Pa Ma ch ool and he
TAIR10 Whole Genome da abase (BAC clones) in The A abidopsis In o ma ion Resou ce (TAIR; h p://www.A abidopsis.o g). The name o he clones co -
esponding o each sequence is indica ed as well as he hi co esponding o each clone. The igh bo de o he 300-bp ITS is included in he alignmen . Red
le e s label DNA sequences ha con inuously ollow he (YYYTAAA)
n
pa e n. As e isks ma k conse ed esidues.
Vega-Vaque o e al.
1048 Genome Resea ch
www.genome.o g
Cold Sp ing Ha bo Labo a o y P ess on Sep embe 11, 2016 - Published by genome.cshlp.o gDownloaded om
DNA. We de ec ed highe le els o uncon e ed mi ochond ial cy-
osines in expe imen s wi h lowe mi ochond ial ead p oduc ion
e iciencies (Fig. 3A). In addi ion, we also obse ed highe le els o
uncon e sion in cy osines wi h lowe ead co e age in all he ex-
pe imen s analyzed (e.g., Fig. 3B). These esul s lead us o conclude
ha he me hyla ion le els o elome es should be be e es ima ed
om he expe imen s wi h he high elome ic C- ich s and p o-
duc ion e iciencies.
Be o e u he analyzing he me hyla ion s a us o A abidopsis
elome es, we decided o in es iga e elome ic p oduc ion e icien-
cy and cy osine uncon e sion in DNA me hyla ion mu an s. We
analyzed se e al mu an s a ec ed in CG, CHG, and CHH me hyl-
a ion. These mu an s we e al e ed in DNA me hyl ans e ases o
ch oma in emodeling p o eins. Among he DNA me hyl ans e -
ase mu an s we e me 1-3, cm 2, a double me 1-3,cm 3-11 mu an ,
and a iple d m1-2, d m2-2, cm 3-11 mu an . Among he ch oma-
in emodeling mu an s we e ddm1-2, d d1-7, and a double ddm1-
2,d d1-7 mu an . P e ious esul s ha e shown ha DNA me hyla-
ion is a ec ed a a genome-wide le el in all hese mu an s.
Whe eas CG me hyla ion is ca alyzed by MET1, non-CG me hyla-
ion is ca alyzed by CMT3 (CHG) and DRM1, DRM2 o CMT2
(CHH and also CHG). These DNA me hyl ans e ases access DNA
wi h he help o ch oma in emodeling p o eins like DDM1 and
DRD1 (Vongs e al. 1993; Finnegan e al. 1996; Ronemus e al.
1996; Jeddeloh e al. 1999; Ba ee e al. 2001; Lind o h e al.
2001; Cao and Jacobsen 2002a,b; Law and Jacobsen 2010; S oud
e al. 2013; Zemach e al. 2013).
No su p isingly, we ound ha he pe cen age o uncon e -
ed cy osines in 50TRs and he pe cen age o 50TRs con aining cy-
osines de ec ed in all he me hyla ion mu an s a ied wi h he
elome ic C- ich s and p oduc ion e iciency (Fig. 2). Hence, hese
esul s con i med ha he me hyla ion s a us o elome es is con-
ounded by expe imen al limi a ions o he bisul i e echnique.
The double me 1-3,cm 3-11, and d m1-2,d d1-7 mu an s
ha e been shown o lack signi ican le els o CHH me hyla ion.
Howe e , we de ec ed elome ic DNA me hyla ion in hese mu-
an s, which e ealed ha hei elome ic DNA was no ully con-
e ed by bisul i e. To u he explo e con e sion e iciencies, we
pe o med bisul i e sequencing expe imen s using a syn he ic
nonme hyla ed elome ic DNA empla e. We used wo di e en
empe a u es (54°C and 40°C) du ing he bisul i e ea men
s ep. The pe cen age o uncon e ed cy osines de ec ed in hese
expe imen s was ∼1% a 54°C and 2.3% a 40°C, e ealing ha
bisul i e con e sion was mo e e ec i e a highe empe a u e
(Fig. 4A). These da a demons a e ha he expe imen al condi-
ions in luence he deg ee o elome ic cy osine con e sion.
We nex wan ed o add ess why high le els o cy osine me h-
yla ion we e es ima ed om some o he A abidopsis bisul i e se-
quencing s udies. Two di e en kinds o 50TRs could be de ec ed
in all he A abidopsis s udies acco ding o hei cy osine con en :
hose wi h a low con en o cy osines, and hose wi h a high
A
B
Figu e 3. Mi ochond ial DNA me hyla ion es ima es a y wi h he mi o-
chond ial eads’p oduc ion e iciency. Da a om di e en genome-wide
bisul i e sequencing expe imen s we e analyzed. (A) Sca e plo ep esen-
a ion o he pe cen age o uncon e ed cy osines (Cs) wi hin mi ochon-
d ial eads e sus he mi ochond ial ead p oduc ion e iciency in wild-
ype plan s. The p oduc ion e iciencies we e calcula ed as he numbe
o mi ochond ial eads di ided by he o al numbe o mapped eads.
They a e exp essed as pe cen ages. (B) Densi y plo ep esen a ion o
uncon e sion e sus co e age o all he mi ochond ial cy osines analyzed
in expe imen SRR578938 (Zemach e al. 2013). The dis ibu ion o
uncon e sion and co e age is ep esen ed in bo h axes.
Figu e 2. Telome ic DNA me hyla ion es ima es a y wi h he elome ic
eads’p oduc ion e iciency. Da a om di e en genome-wide bisul i e se-
quencing expe imen s we e analyzed. The uppe panel shows he sca e
plo ep esen a ion o he pe cen age o uncon e ed cy osines (Cs) de-
ec ed wi hin 50TRs (es ima ed me hyla ion le els) e sus he elome ic
C- ich s and p oduc ion e iciency. P oduc ion e iciencies we e calcula ed
as he numbe o 50TRs di ided by he o al numbe o mapped eads and
exp essed as pe cen ages. The lowe panel shows he sca e plo ep esen-
a ion o he pe cen age o 50TRs con aining cy osines e sus he elo-
me ic C- ich s and p oduc ion e iciency. Da a ob ained om wild- ype
expe imen s a e ep esen ed as ed ci cles and da a ob ained om DNA
me hyla ion mu an s a e ep esen ed as blue ci cles. These mu an s in-
clude single (me 1-3, cm 2, ddm1-2, d d1-7), double (me 1-3 and cm 3-
11, ddm1-2 and d d1-7), and iple (d m1-2, d m2-2 and cm 3-11)
mu an s.
A abidopsis elome es a e unme hyla ed
Genome Resea ch 1049
www.genome.o g
Cold Sp ing Ha bo Labo a o y P ess on Sep embe 11, 2016 - Published by genome.cshlp.o gDownloaded om
con en (Fig. 5A). This kind o dis ibu ion has been p e iously ob-
se ed a e bisul i e ea men o nonme hyla ed phage DNA
(Guo e al. 2013). Al hough phage eads wi h ew cy osines we e
p oposed o o igina e om andom bisul i e con e sion ailu e
o sequencing e o s, phage eads wi h high con en o cy osines
we e hough o a ise om seconda y s uc u e o ma ion and
o e all ailu e o bisul i e con e sion (Guo e al. 2013).
We obse ed ha he low elome ic C- ich s and p oduc ion
expe imen s gene a ed highe pe cen ages o 50TRs wi h a high
con en o cy osines han he high p oduc ion expe imen s (Fig.
5A). These esul s suppo he hypo hesis ha he low p oduc ion
e iciency expe imen s migh be ela ed, a leas in pa , o second-
a y s uc u e o ma ion and o e all ailu e o bisul i e con e sion
o he A abidopsis elome ic DNA. Indeed, we ound a s ong posi-
i e co ela ion be ween he pe cen age o uncon e ed cy osines
de ec ed wi hin 50TRs o mi ochond ial eads and he pe cen age
o uncon e ed eads using BS-Seeke 2 (Fig. 5B; Guo e al. 2013).
We nex decided o analyze he p oduc ion e iciency o he
elome ic empla e DNA s ands. We ound ha he elome ic C-
ich s and was p oduced wi h lowe e iciency han he G- ich
s and (Fig. 4A). In addi ion, we also ound his elome ic s and
imbalance in he expe imen s pe o med wi h A abidopsis geno-
mic DNA, al hough i was mo e p onounced in his la e case.
Al hough he le els o uncon e ed cy osines a A abidopsis elo-
me es we e lowe in expe imen s wi h lowe s and imbalance,
he le els o uncon e ed cy osines de ec ed o he syn he ic em-
pla e did no dec ease wi h he a io o he elome ic s ands (Fig.
4A). These esul s could be explained i we assume ha he elo-
me ic empla e had a lowe endency o emain double s anded
o o o m seconda y s uc u e han he A abidopsis elome ic ag-
men s du ing bisul i e expe imen s. This assump ion is suppo ed
by he ac ha uncon e ed cy osines do no end o accumula e
wi hin he same elome ic empla e eads and by he i ual ab-
sence o elome ic empla e eads wi h a high densi y o cy osines
(Fig. 4A,B). The na u e o he syn he ic elome ic empla e could
ha e in luenced i s capabili y o emain double s anded o o
o m seconda y s uc u es as i was syn he ized by annealing wo
elome ic oligos and, mos impo an ly, was sho e (105 bp)
han he A abidopsis DNA agmen s analyzed in he whole-ge-
nome bisul i e sequencing s udies ( ypically se e al hund ed bp).
We ound ha cy osines wi hin mi ochond ial DNA egions
wi h high CG densi y had a highe endency o be uncon e ed
and, o a lesse ex en , unco e ed han cy osines wi hin low CG
densi y egions (e.g., Fig. 6A). This endency was pa icula ly
A
B
Figu e 4. A syn he ic nonme hyla ed elome ic DNA empla e is no ul-
ly con e ed du ing bisul i e ea men . (A) Sca e plo ep esen a ions o
he pe cen ages o uncon e ed cy osines (Cs) wi hin 50TRs and o 50TRs
con aining cy osines e sus he elome ic s ands a io. Ra io alues we e
calcula ed as he numbe o 50TRs di ided by he numbe o eads ha ol-
low he TTTAGGG pa e n along he i s 50 bp. They we e exp essed as
pe cen ages. Whole-genome bisul i e sequencing s udies pe o med
wi h he A abidopsis wild- ype s ain ( ed do s) and wi h DNA me hyla ion
mu an s (blue do s) we e analyzed. Simila ly, bisul i e sequencing expe i-
men s pe o med wi h a 105-bp-long elome ic syn he ic agmen a
54°C (ligh g een do s) o a 40°C (da k g een do s) we e also analyzed.
Da a co esponding o he me 1-3, cm 3-11 and ddm1-2, d d1-7 double
mu an s a e indica ed. (B) Ba plo ep esen a ion showing he dis ibu ion
o cy osines wi hin 50TRs de i ed om he syn he ic elome ic agmen
ea ed wi h bisul i e a 54°C (ligh g een ba s) o a 40°C (da k g een
ba s). The pe cen ages o 50TRs con aining di e en numbe s o cy osines
a e ep esen ed. A e age alues co esponding o ou independen expe -
imen s a e shown o each empe a u e.
A
B
Figu e 5. Low le els o con e ed eads unde lie he o e es ima ion o
cy osine me hyla ion. (A) Ba plo ep esen a ion showing he dis ibu ion
o cy osines wi hin 50TRs. The pe cen ages o wild- ype 50TRs con aining
di e en numbe s o cy osines a e ep esen ed. Low ( < 8% × 10
−4
) and
high ( > 8% × 10
−4
) elome ic C- ich s and p oduc ion e iciency expe i-
men s we e g ouped and hei a e age pe cen ages we e independen ly
ep esen ed. (B) Sca e plo ep esen a ion o he pe cen age o uncon-
e ed cy osines de ec ed wi hin 50TRs o mi ochond ial eads e sus
he pe cen age o mapped eads lagged as po en ially uncon e ed.
Po en ial uncon e ed eads we e lagged by BS-Seeke 2 as ollows:
#(mCH si es) > 5 and [#(mCH si es)/#(all CH si es)] > 0.5. P- alues o
50TRs and mi ochond ial eads a e 0.0008 and 0.0241, espec i ely.
Vega-Vaque o e al.
1050 Genome Resea ch
www.genome.o g
Cold Sp ing Ha bo Labo a o y P ess on Sep embe 11, 2016 - Published by genome.cshlp.o gDownloaded om
e iden in expe imen s wi h high le els o o al mi ochond ial un-
con e ed cy osines (Fig. 6B). I will be in e es ing o asce ain
whe he high CG densi y egions ha e a high p opensi y o emain
as double-s anded DNA o o o m noncanonical seconda y s uc-
u es du ing bisul i e expe imen s.
High-con idence da a e eal he absence o DNA me hyla ion
in A abidopsis elome es
As men ioned abo e, he expe imen s wi h high elome ic C- ich
s and p oduc ion e iciencies should be e ep esen he me hyl-
a ion s a us o A abidopsis elome es. Indeed, hese expe imen s e-
ealed ha A abidopsis elome es a e no me hyla ed o unde go
e y low le els o DNA me hyla ion (Fig. 2). We ocused on he
only high C- ich s and p oduc ion e iciency s udy analyzed
he e ha gene a ed eads longe han 50 bp (Supplemen al Table
S1). The Sequence Read A chi e accession numbe o his expe i-
men is SRR578938 (Zemach e al. 2013). The pe cen ages o un-
con e ed cy osines wi hin 50TRs and o 50TRs con aining
cy osines de ec ed in SRR578938 we e 1.5% and 10%, espec i ely
(Fig. 2). Hence, he le els o elome ic cy osine me hyla ion es i-
ma ed om SRR578938 a e wi hin he echnical backg ound le els
achie ed o he whole genome in bisul i e expe imen s (Lis e
e al. 2008).
Since he eads om SRR578938 we e 100 bp in size, we decid-
ed o use he comple e sequence o all he 50TRs om his expe -
imen o map hei genomic o igin. We ound ha all he 50TRs
mapped o elome es, o ITSs, o o he le elome e-sub elome e
junc ion o Ch omosome I (Fig. 7; Supplemen al Fig. S1). Mos
o he elome ic eads de ec ed lacked
cy osines along hei en i e leng h (Fig.
7; Supplemen al Fig. S1). In con as ,
he eads co esponding o ITSs o o
he elome e-sub elome e junc ion o
Ch omosome I exhibi ed high le els o
cy osines, which is in ag eemen wi h
he high le els o DNA me hyla ion p e-
iously epo ed o hese loci (Fig. 7;
Cokus e al. 2008; V bsky e al. 2010;
Vaque o-Sedas e al. 2011, 2012;
Og ocká e al. 2014). Thus, he analysis
o he comple e 100-bp eads om
SRR578938 con i ms he i ual absence
o DNA me hyla ion in A abidopsis
elome es.
Res ic ion enzyme sensi i i y analyses
con i m he absence o DNA
me hyla ion in A abidopsis elome es
To e i y ou in silico analyses, we decid-
ed o s udy DNA me hyla ion a
A abidopsis elome es using an al e na-
i e echnique. We pe o med Sou he n
blo hyb idiza ion expe imen s using
h ee es ic ion enzymes wi h di e en
DNA me hyla ion dependence/sensi i -
i y. One o he enzymes (HpaII) ecogniz-
es and cu s he sequence CCGG only
when bo h cy osines a e unme hyla ed.
The o he wo enzymes (Mc BC and
FspEI) ecognize me hyla ed cy osines
and cu he su ounding nucleo ides. Mc BC ecognizes he se-
quence R
m
C(
N40-3000
)R
m
C and should a ge he i s cy osine o
he elome ic epea uni (CCCTAAA) i i we e me hyla ed.
FspEI ecognizes he sequence C
m
C and should a ge he second
and hi d cy osines o he elome ic epea uni i hey we e me h-
yla ed (Fig. 8A).
We diges ed equal amoun s o A abidopsis genomic DNA wi h
he es ic ion enzyme T u9I o wi h T u9I plus one o he enzymes
men ioned abo e. We hen an he esul ing DNA samples on an
aga ose gel and hyb idized hem wi h a elome ic p obe o wi h
a cen ome ic p obe. T u9I ecognizes and cu s he sequence
TTAA, which localizes in sub elome ic egions a ∼200 bp om
he pe ec elome ic epea s, on a e age. The e o e, his enzyme
cu s he A abidopsis genome equen ly, bu lea es elome es es-
sen ially uncu . We obse ed a 2- o 5-kbp uzzy elome ic band a -
e diges ing A abidopsis genomic DNA wi h T u9I, as p e iously
epo ed (Richa ds and Ausubel 1988; Fi zge ald e al. 1999;
Vaque o-Sedas e al. 2011). This elome ic band emained una -
ec ed a e diges ion wi h HpaII, FspEI, o Mc BC (Fig. 8B,C). In
con as , h ee cen ome ic DNA bands gene a ed a e T u9I
diges ion we e eadily diges ed wi h FspEI and Mc BC, as would
be expec ed due o he high le els o DNA me hyla ion p e iously
obse ed o he cen ome ic epea s. As expec ed, hese bands e-
mained una ec ed in he p esence o HpaII (Vongs e al. 1993;
Lind o h e al. 2001). Al hough hese esul s canno comple ely
ule ou he exis ence o a e y mino po ion o me hyla ed elo-
me es, we should conside ha a single cu wi hin a elome ic ag-
men should sho en his agmen . In addi ion, he e y low
p opo ion o 50TRs con aining one o wo uncon e ed cy osines
A
B
Figu e 6. High CG densi y challenges cy osine con e sion and co e age o mi ochond ial DNA. Da a
om di e en genome-wide bisul i e sequencing expe imen s we e analyzed. (A) Densi y plo ep esen-
a ions o cy osine uncon e sion o co e age e sus he pe cen age o CG esidues wi hin 100-bp iles
(CG densi y). All he mi ochond ial cy osines om expe imen SRR492932 (Dowen e al. 2012) we e
g ouped in 100-bp iles and analyzed. The dis ibu ions o uncon e sion, co e age, and CG densi ies
a e ep esen ed in bo h axes. Pea son co ela ion coe icien s (R) a e also indica ed. (B) Sca e plo ep-
esen a ions o Pea son co ela ion coe icien s be ween cy osine uncon e sion o co e age and CG den-
si y pe 100-bp iles (PCC alues) e sus he mean mi ochond ial me hyla ion es ima ed in e e y
expe imen (pe cen age o uncon e ed Cs). P- alues o cy osine uncon e sion and co e age a e
0.0064 and 0.0802, espec i ely.
A abidopsis elome es a e unme hyla ed
Genome Resea ch 1051
www.genome.o g
Cold Sp ing Ha bo Labo a o y P ess on Sep embe 11, 2016 - Published by genome.cshlp.o gDownloaded om

ha we ha e obse ed in he wild- ype high p oduc ion e iciency
expe imen s should be be e explained by expe imen al limi a-
ions o he bisul i e echnique, as should be he 50TRs con aining
uncon e ed cy osines obse ed in he expe imen s pe o med
wi h he DNA me hyla ion mu an s lacking CHH me hyla ion.
Thus, aken oge he ou esul s suppo ha DNA me hyla ion
is essen ially absen om A abidopsis elome es.
Discussion
Limi a ions o he bisul i e sequencing echnique
Al hough bisul i e sequencing is p obably one o he mos accu a e
echniques o analyze cy osine me hyla ion, some p oblems can
a ise du ing he bisul i e ea men . We belie e ha hese p ob-
lems may explain he e y high le els o cy osine de ec ion ha
we ound a A abidopsis elome es in expe imen s wi h low elo-
me ic C- ich s and p oduc ion e iciencies. I some unme hyla ed
cy osines do no eac wi h bisul i e, hey will be de ec ed as me h-
yla ed cy osines leading o o e es ima ion o he me hyla ion
e en s. Thus, o e es ima ion o elome ic me hyla ion should be
in luenced by he deg ee o elome ic DNA dena u a ion o he o -
ma ion o elome ic C- ich s and seconda y s uc u es like he i-
mo i (Le oy e al. 1994; Fe nández e al. 2011; Guo e al. 2013).
In addi ion, since bisul i e ea men modi ies DNA and e en de-
g ades i (F omme e al. 1992; G unau e al. 2001), some un-
me hyla ed cy osines migh pa ially o imp ope ly eac wi h
bisul i e eagen s leading o he gene a ion o abe an cy osines.
Figu e 7. A high-con idence expe imen e eals he absence o elome ic DNA me hyla ion. The comple e 100-bp sequences o all he 50TRs om
SRR578938 (Zemach e al. 2013) con aining one o mo e cy osines a e displayed, oge he wi h he comple e sequences o some o he 50TRs lacking
cy osines. The numbe o cy osines wi hin 50TRs and he genomic o igin es ima ed o all he eads a e indica ed in blue. Among he 222 50TRs iden i ied
in SRR578938, 199 did no con ain cy osines (see Supplemen al Fig. S1), 19 con ained one o wo cy osines, and ou con ained i e cy osines o mo e. The
pa s o he eads ha ollow a con inuous pe ec elome ic (YYYTAAA)
n
pa e n a e shown in black wi h he excep ion o cy osines ha a e labeled ed. The
emaining bases a e shown in g een and ha e been used o mapping pu poses. All he eads could be mapped o he TAIR10 da abase excep one, which
mapped o he le elome e-sub elome e junc ion o Ch omosome I (Kuo e al. 2006). No e ha all he 50TRs lacking cy osines (Supplemen al Fig. S1) and
hose con aining one o wo cy osines ollow a quasi-pe ec elome ic (YYYTAAA)
n
pa e n along hei en i e 100-bp leng h. Mos o hese eads (218 o
222) should o igina e om elome es. The 50TR con aining 19 cy osines also ollows he pe ec elome ic pa e n along i s en i e 100-bp leng h and
should be elome ic and o igina e om expe imen al limi a ions o he bisul i e echnique o o igina e om he 300-bp ITS. The 50TRs ha con ain
mo e han i e cy osines o igina e om ITSs o om he le elome e–sub elome e junc ion o Ch omosome I (IL). Since he equency o hese 50TRs
(3 o 222; 1.4%) is expec ed acco ding o he dis ibu ion o ITSs and elome e–sub elome e junc ions in he A abidopsis genome, hei cy osines migh
e lec eal me hyla ion e en s.
Vega-Vaque o e al.
1052 Genome Resea ch
www.genome.o g
Cold Sp ing Ha bo Labo a o y P ess on Sep embe 11, 2016 - Published by genome.cshlp.o gDownloaded om
These abe an cy osines migh s ill be ecognized as cy osines bu
unde go ampli ica ion wi h less e iciency han u acil du ing he
i s cycle o PCR. Al e na i ely, hese abe an cy osines migh
be ampli ied e icien ly du ing he i s cycle o PCR bu inco ec ly
copied. Such phenomena migh explain he an i-co ela ion ob-
se ed o he elome ic C- ich s and p oduc ion e iciency and
he uncon e sion o cy osines a A abidopsis elome es.
The hi d cy osine o he A abidopsis elome ic epea uni has
been ound o be uncon e ed mo e equen ly han he i s and
second cy osines a e bisul i e ea men (Cokus e al. 2008). This
cy osine is uncon e ed mo e equen ly han he i s and he sec-
ond in he 50TRs o igina ing om he double me 1-3,cm 3-11 and
ddm1-2, d d1-7 mu an s, which lack signi ican le els o asymme -
ic DNA me hyla ion (Supplemen al Fig. S2; S oud e al. 2013;
Zemach e al. 2013). Hence, he lowe con e sion a e o his cy o-
sine should esul om expe imen al limi a ions like hose men-
ioned abo e. The hi d cy osine o he A abidopsis elome ic
epea uni is also uncon e ed mo e equen ly han he i s
and second cy osines in sub elome ic egions and ITSs o wild-
ype plan s (Fig. 7; Cokus e al. 2008; V bsky e al. 2010; Og ocká
e al. 2014). Howe e , hese cy osines e lec eal me hyla ion
e en s. Thus, aken oge he , ou esul s sugges a special DNA con-
o ma ion o he elome ic epea a ays.
We hypo hesize ha incomple e elome e dena u a ion and/
o elome ic seconda y s uc u e could p e en bisul i e con e -
sion, enhance bisul i e d i en DNA b eakage, and/o al e he e i-
ciency/accu acy o he i s cycle o PCR ampli ica ion a e
bisul i e ea men . On he o he hand, he e icien dena u a ion
o elome es and he absence o elome ic seconda y s uc u es
could a o cy osine con e sion, al hough he con e sion e icien-
cies achie ed depend on di e en expe imen al pa ame e s, such
as he empe a u e o incuba ion wi h bisul i e.
Telome es a e e ac o y o de no o DNA me hyla ion
Ou esul s indica e ha DNA me hyla ion is absen om
A abidopsis elome es. Howe e , sub elome ic egions display a
high le el o cy osine me hyla ion, which is in ag eemen
wi h p e iously published esul s (Fig. 9A; V bsky e al. 2010;
Vaque o-Sedas e al. 2011; Og ocká e al. 2014).
Small in e e ing RNAs (siRNAs) can a ge DNA me hyla ion
in all sequence con ex s h ough he RNA-di ec ed DNA
Me hyla ion pa hway (RdDM). This pa hway in ol es mul iple
p o eins including DNA and his one me hyl ans e ases, ch oma-
in emodeling p o eins, RNases, and DNA o RNA polyme ases
(Law and Jacobsen 2010). Plan s ha e wo speci ic DNA poly-
me ases in ol ed in RdDM: pol IV and pol V. Pol IV oge he
wi h he RNA dependen RNA polyme ase 2 (RDR2) and he
Dice -like RNase 3 (DCL3) pa icipa es in he gene a ion o 24-bp
siRNAs. In u n, pol V, he siRNA single s and binding p o ein
A gonau e 4 (AGO4) and he non-CG Domains Rea anged DNA
me hyl ans e ase 2 (DRM2) a e equi ed o siRNA-dependen
DNA me hyla ion. The RdDM pa hway is in ol ed in de no o
DNA me hyla ion. Th ough his p ocess, siRNAs con aining a spe-
ci ic nucleo ide sequence a e belie ed o d i e DRM2-dependen
DNA me hyla ion o he same DNA sequence a di e en genomic
loci. Fo example, 24-bp siRNAs con aining pe ec elome ic
Figu e 8. Me hyla ion-dependen es ic ion enzyme analyses con i m
heabsenceo elome icDNAme hyla ion.(A)Ca oon ep esen ing hese-
quence speci ici y o he es ic ion enzymes used o diges A abidopsis ge-
nomic DNA. (B) Sou he n blo hyb idiza ions o A abidopsis genomic DNA
diges ed wi h he es ic ion enzymes indica ed in A. The uppe panel shows
equal amoun s o undiges ed (Con ol) o diges ed (HpaII, FspEI, o Mc BC)
DNA samples hyb idized wi h a elome ic p obe. The ou DNA samples
we e alsodiges ed wi h T u9I p io o hyb idiza ion. The middlepanel shows
he hyb idiza ion o he same samples wi h a 180-bp cen ome ic epea
p obe. The e hidium b omide s aining o he samples is shown in he lowe
panel. The con ol DNA sample was un in he same gel as he es o he
samples, so ha hei co esponding hyb idiza ion signals we e p ocessed
equally. (C) Ba plo ep esen a ion o he elome ic and cen ome ic bo -
om band hyb idiza ion signals exp essed as pe cen ages o he con ol.
C, H, F, and M ep esen Con ol, HpaII, FspEI, and Mc BC, espec i ely.
Figu e 9. Model ep esen ing he in luence o DNA me hyla ion on he
homeos asis o elome e leng h. (A) The DNA me hyla ion machine y o
wild- ype plan s can a ge sub elome ic egions (g ay ba ) and elo-
me e-sub elome junc ions (da k g ay ba ) bu no elome es (black ba ).
In his scena io, sub elome ic DNA me hyla ion con ibu es o posi i ely
egula e elome e leng h. (B) The homeos asis o elome e leng h is los
when sub elome ic DNA me hyla ion canno be achie ed due o he im-
pai men o he DNA me hyla ion machine y.
A abidopsis elome es a e unme hyla ed
Genome Resea ch 1053
www.genome.o g
Cold Sp ing Ha bo Labo a o y P ess on Sep embe 11, 2016 - Published by genome.cshlp.o gDownloaded om
epea s ha e been desc ibed in A abidopsis (Gus a son e al. 2005;
V bsky e al. 2010). These siRNAs d i e RDR2, DCL3, AGO4, and
DRM2-dependen DNA me hyla ion o elome ic epea s a ITSs
and sub elome es (Cokus e al. 2008; V bsky e al. 2010; S oud
e al. 2013). Howe e , we ha e shown he e ha he RdDM pa h-
way canno a ge DNA me hyla ion a elome es, which p obably
e lec s hei special a chi ec u e (Fig. 9A).
Sub elome ic DNA me hyla ion con ols elome e
leng h homeos asis
We ha e p e iously epo ed ha A abidopsis elome es ha e lowe
le els o DNA me hyla ion han ITSs o sub elome es (Vaque o-
Sedas e al. 2011). He e, we ha e desc ibed ha DNA me hyla ion
is indeed absen om A abidopsis elome es. No ably, A abidopsis
mu an s a ec ing CG and non-CG DNA me hyla ion ha e sho e
elome es han wild- ype plan s (Og ocká e al. 2014; Vaque o-
Sedas and Vega-Palas 2014). The e o e, DNA me hyla ion mus
be in ol ed in he con ol o elome e leng h homeos asis in
A abidopsis. Since he exp ession o p o eins ha egula e elome e
leng h ha e no been ound o be al e ed in he a o emen ioned
mu an s (Zhang e al. 2006; Hudson e al. 2011), we conclude
ha sub elome ic DNA me hyla ion posi i ely con ols he ho-
meos asis o elome e leng h in A abidopsis (Fig. 9). In con as ,
DNA me hyla ion con ols elome e leng h nega i ely in mam-
mals (Gonzalo e al. 2006). Thus, long elome es a e obse ed in
mammalian DNA me hyla ion mu an s. Howe e , hese elome es
migh a ise a e elome ic sho ening, as p e iously epo ed
o he Saccha omyces ce e isiae es 1 (e e sho e elome es 1)
su i o s (Lundblad and Blackbu n 1993). In addi ion, i is un-
known whe he he con ol o elome e leng h exe ed by DNA
me hyla ion in mammals occu s a elome es o sub elome es
(Blasco 2007).
Cy osine me hyla ion has been ela ed o a wide a ie y o bi-
ological ea u es om he main enance o genome s abili y o cell
di e en ia ion, de elopmen , o illness. Many o hese ea u es a e
also a ec ed by elome e leng h (Osbo ne and Boub iak 2002; Ooi
e al. 2009; Law and Jacobsen 2010; Guy e al. 2011; Bodelon e al.
2014; Hea d and Ma ienssen 2014; Plasschae and Ba olomei
2014). Hence, i will be in e es ing o asce ain he ele ance o
sub elome ic DNA me hyla ion in all hese p ocesses in mammals
and plan s.
Me hods
In silico analysis o elome ic DNA me hyla ion
Da a om di e en genome-wide bisul i e sequencing s udies we e
analyzed (Schmi z e al. 2011; Ausin e al. 2012; Dowen e al. 2012;
Moissia d e al. 2012; S oud e al. 2012, 2013, 2014; Zhong e al.
2012; G eenbe g e al. 2013; Law e al. 2013; Zemach e al.
2013). Since he o iginal Wa son and C ick s ands, bu no hei
e e se complemen s, we e sequenced in hese s udies, only he
elome ic C- ich s and was moni o ed o es ima e me hyla ion
le els. The e o e, he elome ic G- ich s and did no in luence
he me hyla ion analyses. The Gene Exp ession Omnibus and
he Sequence Read A chi e accession numbe s o he s udies a e in-
dica ed in Supplemen al Table S1, oge he wi h he codes o he
speci ic uns analyzed.
Conside ing ha he leng h o he eads in many o he s ud-
ies selec ed was 50 bp, we decided o s a by s udying only he i s
50 bp o he eads in all he expe imen s e en i hey we e longe .
In p inciple, we conside ed elome ic eads as hose ha ollowed
he (YYYTAAA)
n
pa e n, whe e Y is C o T, along he 50 bp. These
sequences could s a a any nucleo ide (C, T, o A) and co e-
sponded o abou se en me hyla ed/unme hyla ed andem elo-
me ic epea s. We deno ed hese eads/ agmen o eads as 50-
bp Telome ic Reads (50TRs).
To e i y ha he 50TRs we e gene a ed essen ially om elo-
me es, we de e mined he numbe o (YYYTAAA)
7
nono e lapping
sequences in he A abidopsis haliana Col-0 genome, including ITSs
and elome es. Fi s , we de e mined he numbe o hese sequenc-
es a ITSs. We used he Pa Ma ch ool in The A abidopsis
In o ma ion Resou ce (TAIR; h p://www.A abidopsis.o g) and
sea ched o he sequence (YYYTAAA)
7
in he TAIR10 Whole-
Genome (BAC clones) da abase, which con ained 135,670,227
by es. This sea ch was pe o med using he de aul op ions ha in-
cluded sea ching in bo h s ands wi h no misma ches allowed. The
sea ch esul ed in 41 hi s, om which only 31 we e no epea ed
in di e en BAC clones. Six o he unique hi s co esponded o se-
quences ha we e o ganized in andem a he pe icen ome ic e-
gion o Ch omosome III gene a ing a 300-bp ITS composed o
pe ec andem elome ic epea s. In addi ion, six o he hi s
we e elome ic ( i e om pA T51 and one om TEL3N). An align-
men o he emaining 19 and he igh bo de o he 300-bp ITS is
shown in Figu e 1. This alignmen e eals ha none o he 19
(YYYTAAA)
7
a o emen ioned sequences a e included in a DNA
s ing con aining mo e han 11 epea s.
Once we de e mined he numbe o ITSs con aining he
(YYYTAAA)
7
sequence, we es ima ed he numbe o imes ha
his sequence is ound a elome es. We es ima ed ha elome es
con ained abou 765 nono e lapping (YYYTAAA)
7
sequences
(3750 bp o a e age elome e leng h di ided by 7 bp o he elo-
me ic epea uni and mul iplied by 10 elome es). The e o e,
only ∼3% o he (YYYTAAA)
7
sequences a e p esen a ITSs (25 di-
ided by 765 and mul iplied by 100). Howe e , he pe cen age o
50TRs co esponding o ITSs in he bisul i e s udies should be low-
e han 3% as he size o he DNA inse s used o cons uc he DNA
lib a ies a e ypically 150–300 bp in size. Simila ly, a low pe cen -
age o he 50TRs should co espond o he elome e–sub elome e
junc ions, which usually con ain pe ec elome ic epea s in e -
spe sed wi h degene a e epea s. In summa y, i he bisul i e expe -
imen s wo k p ope ly, we expec ha a leas 97% o he 50TRs
co espond o elome es.
To pe o m he analyses o cy osine me hyla ion a elome es
we used an in-house compu e p og am ha allowed he iden i i-
ca ion o 50TRs, he de e mina ion o he numbe o uncon e ed
cy osines wi hin 50TRs, and he numbe o eads ha ollowed he
(TTTAGGG)
n
pa e n along he i s 50 bp (Supplemen al Table S1,
Supplemen al Telome ic analyses sou ce code).
In silico analysis o mi ochond ial DNA me hyla ion
Reads o he a o emen ioned A abidopsis da a se s we e download-
ed om he GEO eposi o y and we e mapped o he en i e TAIR10
e e ence genome wi hou allowing misma ches using BS-Seeke
2.0.32 (Guo e al. 2013), which calls Bow ie 0.12.9 (Langmead
e al. 2009) in u n. Reads we e no il e ed o poo con e sion p i-
o o base calling. Coun s o C and T base calls pe mi ochond ial
cy osine o bo h s ands we e ex ac ed om he esul ing CG
map iles. Subsequen s a is ical analysis and plo s we e gene a ed
using a cus om R en i onmen (R Co e Team 2008).
Telome ic empla e syn hesis, bisul i e ea men , lib a y
p epa a ion, and nex gene a ion sequencing
Equal amoun s o oligos con aining he sequences (TTTAGGG)
15
and (CCCTAAA)
15
we e mixed, hea ed o 10 min o 98°C and
Vega-Vaque o e al.
1054 Genome Resea ch
www.genome.o g
Cold Sp ing Ha bo Labo a o y P ess on Sep embe 11, 2016 - Published by genome.cshlp.o gDownloaded om
allowed o cool down o 15 min a oom empe a u e. Then, he
annealed DNA agmen s we e gel pu i ied and hei quali y as-
sessed using he D1000 assay on he TapeS a ion 2200 (Agilen
Technologies). End epai , A- ailing, and adap e liga ion was pe -
o med on he annealed oligos acco ding o he Illumina T uSeq
Nano ki p o ocol (ca alog #15041759, Illumina). Adap e -liga ed
oligos we e dena u ed in he p esence o bisul i e (Zymo
Ligh ning MagP ep, ca alog #D5046, Zymo Resea ch) o 8 min
a 98°C, and wo empe a u es we e used o he con e sion
s ep: 54°C (manu ac u e ’s sugges ed empe a u e) and 40°C.
Desul ona ion was pe o med acco ding o he manu ac u e ’s in-
s uc ions. Con e ed DNA was ampli ied using MyTaq Mix
(Bioline, Taun on) and Illumina T uSeq PCR P ime Cock ail ac-
co ding o he ollowing p o ocol: ini ial dena u a ion o 30 sec
a 98°C; 12 cycles o 15 sec a 98°C, o 30 sec a 60°C, o 30 sec
a 72°C; inal ex ension o 5 min a 72°C. AMPu e XP beads
we e used o clean up he inal PCR eac ions be o e sequencing.
Lib a ies we e sequenced wi h an Illumina HiSeq 2500 sys em us-
ing 100-bp single-end eads.
Me hyla ion-dependen es ic ion enzyme analyses o elome ic
DNA me hyla ion
Equal amoun s o genomic DNA (500 ng) we e undiges ed o di-
ges ed wi h he es ic ion enzymes HpaII, FspEI, o Mc BC ollow-
ing he supplie ’s ins uc ions (New England Biolabs). The
esul ing DNA samples we e pu i ied, u he diges ed wi h
T u9I, esol ed on an aga ose gel, ans e ed o a Hybond-N
+
memb ane, and hyb idized wi h a elome ic p obe, as p e iously
desc ibed (Gámez-A jona e al. 2010). Then, he same memb ane
was s ipped and hyb idized again wi h a 180-bp cen ome ic e-
pea p obe. The 180-bp cen ome ic empla e was syn he ized as
p e iously epo ed (Nagaki e al. 2003).
Da a access
Illumina bisul i e sequencing da a om his s udy ha e been sub-
mi ed o he NCBI Gene Exp ession Omnibus (GEO; h p://www.
ncbi.nlm.nih.go /geo/) unde accession numbe GSE83510.
Acknowledgmen s
We hank Suhua Feng o c i ical eading o he manusc ip and
NCBI and TAIR o sha ing hei esou ces.
Re e ences
Ausin I, G eenbe g MV, Simanshu DK, Hale CJ, Vashish AA, Simon SA, Lee
TF, Feng S, Espanola SD, Meye s BC, e al. 2012. INVOLVED IN DE
NOVO 2-con aining complex in ol ed in RNA-di ec ed DNA me hyla-
ion in A abidopsis.P oc Na l Acad Sci 109: 8374–8381.
Ba ee L, Malagnac F, Bende J. 2001. A abidopsis cm 3 ch omome hylase
mu a ions block non-CG me hyla ion and silencing o an endogenous
gene. Genes De 15: 1753–1758.
Blackbu n EH. 2010. Telome es and elome ase: he means o he end
(Nobel lec u e). Angew Chem In Ed Engl 49: 7405–7421.
Blasco MA. 2007. The epigene ic egula ion o mammalian elome es. Na
Re Gene 8: 299–309.
Bodelon C, Sa age SA, Gadalla SM. 2014. Telome es in molecula epidemi-
ology s udies. P og Mol Biol T ansl Sci 125: 113–131.
Cao X, Jacobsen SE. 2002a. Locus-speci ic con ol o asymme ic and
CpNpG me hyla ion by he DRM and CMT3 me hyl ans e ase genes.
P oc Na l Acad Sci 99: 16491–16498.
Cao X, Jacobsen SE. 2002b. Role o he A abidopsis DRM me hyl ans e ases
in de no o DNA me hyla ion and gene silencing. Cu Biol 12:
1138–1144.
Cas el SE, Ma ienssen RA. 2013. RNA in e e ence in he nucleus: oles o
small RNAs in ansc ip ion, epigene ics and beyond. Na Re Gene 14:
100–112.
Cla k SJ, Ha ison J, Paul CL, F omme M. 1994. High sensi i i y mapping o
me hyla ed cy osines. Nucleic Acids Res 22: 2990–2997.
Cokus SJ, Feng S, Zhang X, Chen Z, Me iman B, Haudenschild CD, P adhan
S, Nelson SF, Pelleg ini M, Jacobsen SE. 2008. Sho gun bisulphi e se-
quencing o he A abidopsis genome e eals DNA me hyla ion pa e n-
ing. Na u e 452: 215–219.
Dowen RH, Pelizzola M, Schmi z RJ, Lis e R, Dowen JM, Ne y JR, Dixon JE,
Ecke JR. 2012. Widesp ead dynamic DNA me hyla ion in esponse o
bio ic s ess. P oc Na l Acad Sci 109: E2183–E2191.
Fe nández S, E i ja R, A iñó A, Jaumo J, Ga gallo R. 2011. In luence o pH,
empe a u e and he ca ionic po phy in TMPyP4 on he s abili y o he
i-mo i o med by he 5′-(C
3
TA
2
)
4
-3′sequence o he human elome e.
In J Biol Mac omol 49: 729–736.
Finnegan EJ, Peacock WJ, Dennis ES. 1996. Reduced DNA me hyla ion in
A abidopsis haliana esul s in abno mal plan de elopmen . P oc Na l
Acad Sci 93: 8449–8454.
Fi zge ald MS, Riha K, Gao F, Ren S, McKnigh TD, Shippen DE. 1999.
Dis up ion o he elome ase ca aly ic subuni gene om A abidopsis in-
ac i a es elome ase and leads o a slow loss o elome ic DNA. P oc Na l
Acad Sci 96: 14813–14818.
F omme M, McDonald LE, Milla DS, Collis CM, Wa F, G igg GW, Molloy
PL, Paul CL. 1992. A genomic sequencing p o ocol ha yields a posi i e
display o 5-me hylcy osine esidues in indi idual DNA s ands. P oc
Na l Acad Sci 89: 1827–1831.
Gala i A, Micheli E, Cacchione S. 2013. Ch oma in s uc u e in elome e dy-
namics. F on Oncol 3: 46.
Gámez-A jona FM, López-López C, Vaque o-Sedas MI, Vega-Palas MA.
2010. On he o ganiza ion o he nucleosomes associa ed wi h elo-
me ic sequences. Biochim Biophys Ac a 1803: 1058–1061.
Gi aud-Panis MJ, Pisano S, Bena och-Popi ke D, Pei B, Le Du MH, Gilson
E. 2013. One iden i y o mo e o elome es? F on Oncol 3: 48.
Gonzalo S, Jaco I, F aga MF, Chen T, Li E, Es elle M, Blasco MA. 2006. DNA
me hyl ans e ases con ol elome e leng h and elome e ecombina-
ion in mammalian cells. Na Cell Biol 8: 416–424.
G eenbe g MV, Dele is A, Hale CJ, Liu A, Feng S, Jacobsen SE. 2013.
In e play be ween ac i e ch oma in ma ks and RNA-di ec ed DNA
me hyla ion in A abidopsis haliana.PLoS Gene 9: e1003946.
G unau C, Cla k SJ, Rosen hal A. 2001. Bisul i e genomic sequencing: sys-
ema ic in es iga ion o c i ical expe imen al pa ame e s. Nucleic Acids
Res 29: e65.
Guo W, Fizie P, Yan W, Cokus S, Sun X, Zhang MQ, Chen PY, Pelleg ini M.
2013. BS-Seeke 2: a e sa ile aligning pipeline o bisul i e sequencing
da a. BMC Genomics 14: 774.
Gus a son AM, Allen E, Gi an S, Smi h D, Ca ing on JC, Kasschau KD.
2005. ASRP: he A abidopsis small RNA p ojec da abase. Nucleic Acids
Res 33: D637–D640.
Guy J, Che al H, Sel idge J, Bi d A. 2011. The ole o MeCP2 in he b ain.
Annu Re Cell De Biol 27: 631–652.
Hea d E, Ma ienssen RA. 2014. T ansgene a ional epigene ic inhe i ance:
my hs and mechanisms. Cell 157: 95–109.
Hudson K, Luo S, Hagemann N, P euss D. 2011. Changes in global gene ex-
p ession in esponse o chemical and gene ic pe u ba ion o ch oma in
s uc u e. PLoS One 6: e20587.
Jeddeloh JA, S okes TL, Richa ds EJ. 1999. Main enance o genomic me hyl-
a ion equi es a SWI2/SNF2-like p o ein. Na Gene 22: 94–97.
Kuo HF, Olsen KM, Richa ds EJ. 2006. Na u al a ia ion in a sub elome ic
egion o A abidopsis: implica ions o he genomic dynamics o a ch o-
mosome end. Gene ics 173: 401–417.
Langmead B, T apnell C, Pop M, Salzbe g SL. 2009. Ul a as and memo y-
e icien alignmen o sho DNA sequences o he human genome.
Genome Biol 10: R25.
Law JA, Jacobsen SE. 2010. Es ablishing, main aining and modi ying DNA
me hyla ion pa e ns in plan s and animals. Na Re Gene 11: 204–220.
Law JA, Du J, Hale CJ, Feng S, K ajewski K, Palanca AM, S ahl BD, Pa el DJ,
Jacobsen SE. 2013. Polyme ase IV occupancy a RNA-di ec ed DNA
me hyla ion si es equi es SHH1. Na u e 498: 385–389.
Le oy JL, Gué on M, Me gny JL, Hélène C. 1994. In amolecula olding o a
agmen o he cy osine- ich s and o elome ic DNA in o an i-mo i .
Nucleic Acids Res 22: 1600–1606.
Lin KW, Yan J. 2008. Endings in he middle: cu en knowledge o in e s i-
ial elome ic sequences. Mu a Res 658: 95–110.
Lind o h AM, Cao X, Jackson JP, Zilbe man D, McCallum CM, Heniko S,
Jacobsen SE. 2001. Requi emen o CHROMOMETHYLASE3 o main e-
nance o CpXpG me hyla ion. Science 292: 2077–2080.
Lis e R, O’Malley RC, Ton i-Filippini J, G ego y BD, Be y CC, Milla AH,
Ecke JR. 2008. Highly in eg a ed single-base esolu ion maps o he epi-
genome in A abidopsis.Cell 133: 523–536.
Lis e R, Pelizzola M, Dowen RH, Hawkins RD, Hon G, Ton i-Filippini J, Ne y
JR, Lee L, Ye Z, Ngo QM, e al. 2009. Human DNA me hylomes a base
esolu ion show widesp ead epigenomic di e ences. Na u e 462:
315–322.
A abidopsis elome es a e unme hyla ed
Genome Resea ch 1055
www.genome.o g
Cold Sp ing Ha bo Labo a o y P ess on Sep embe 11, 2016 - Published by genome.cshlp.o gDownloaded om