The Drosophila modigliani (moi) gene encodes a HOAP-interacting protein required for telomere protection.

Abstract

Several proteins have been identified that protect Drosophila telomeres from fusion events. They include UbcD1, HP1, HOAP, the components of the Mre11-Rad50-Nbs (MRN) complex, the ATM kinase, and the putative transcription factor Woc. Of these proteins, only HOAP has been shown to localize specifically at telomeres. Here we show that the modigliani gene encodes a protein (Moi) that is enriched only at telomeres, colocalizes and physically interacts with HOAP, and is required to prevent telomeric fusions. Moi is encoded by the bicistronic CG31241 locus. This locus produces a single transcript that contains 2 ORFs that specify different essential functions. One of these ORFs encodes the 20-kDa Moi protein. The other encodes a 60-kDa protein homologous to RNA methyltransferases that is not required for telomere protection (Drosophila Tat-like). Moi and HOAP share several properties with the components of shelterin, the protein complex that protects human telomeres. HOAP and Moi are not evolutionarily conserved unlike the other proteins implicated in Drosophila telomere protection. Similarly, none of the shelterin subunits is conserved in Drosophila, while most human nonshelterin proteins have Drosophila homologues. This suggests that the HOAP-Moi complex, we name "terminin," plays a specific role in the DNA sequence-independent assembly of Drosophila telomeres. We speculate that this complex is functionally analogous to shelterin, which binds chromosome ends in a sequence-dependent manner.

Full text

Page 1

The Drosophila modigliani (moi) gene encodes
a HOAP-interacting protein required for
telomere protection
Grazia D. Raffaa, Giorgia Siriacoa,1, Simona Cugusia, Laura Ciapponia, Giovanni Cencib, Edward Wojcikc,
and Maurizio Gattia,2
aIstituto di Biologia e Patologia Molecolari del CNR and Dipartimento di Genetica e Biologia Molecolare, Sapienza, Universita ` di Roma, 00185 Rome, Italy;
bDipartimento di Biologia di Base ed Applicata, Universita ` dell’Aquila, Coppito, L’Aquila, 67010, Italy; andcLouisiana State University Health Sciences
Center, Department of Biochemistry and Molecular Biology, Louisiana State University, New Orleans, LA 70803
Communicated by Dan L. Lindsley, University of California at San Diego, La Jolla, CA, December 16, 2008 (received for review August 29, 2008)
Several proteins have been identified that protect Drosophila
telomeres from fusion events. They include UbcD1, HP1, HOAP, the
components of the Mre11-Rad50-Nbs (MRN) complex, the ATM
kinase, and the putative transcription factor Woc. Of these pro-
teins, only HOAP has been shown to localize specifically at telo-
meres. Here we show that the modigliani gene encodes a protein
(Moi) that is enriched only at telomeres, colocalizes and physically
interacts with HOAP, and is required to prevent telomeric fusions.
Moi is encoded by the bicistronic CG31241 locus. This locus pro-
duces a single transcript that contains 2 ORFs that specify different
essential functions. One of these ORFs encodes the 20-kDa Moi
protein. The other encodes a 60-kDa protein homologous to RNA
methyltransferases that is not required for telomere protection
(Drosophila Tat-like). Moi and HOAP share several properties with
the components of shelterin, the protein complex that protects
human telomeres. HOAP and Moi are not evolutionarily conserved
unlike the other proteins implicated in Drosophila telomere pro-
tection. Similarly, none of the shelterin subunits is conserved in
Drosophila, while most human nonshelterin proteins have Dro-
sophila homologues. This suggests that the HOAP-Moi complex,
we name ‘‘terminin,’’ plays a specific role in the DNA sequence-
independent assembly of Drosophila telomeres. We speculate that
this complex is functionally analogous to shelterin, which binds
chromosome ends in a sequence-dependent manner.
HP1 ? MRN complex ? telomeric fusions ? telomeric proteins
T
distinguishnaturalchromosometerminifrombrokenDNAends(1,
2).Inmostorganisms,telomerescontainarraysofGC-richrepeats,
which are added to chromosome ends by telomerase (2). These
repeatsbindadiscretenumberofspecializedproteins,whichrecruit
additional polypeptides to form a complex protein network that
caps chromosome ends (1). Drosophila telomeres are elongated by
transpositionof3specializedretroelements,ratherthantelomerase
activity and do not terminate with GC-rich repeats (3–4). In
addition, several studies indicate that Drosophila telomeres are
epigenetically determined structures than can be assembled inde-
pendently of the sequence of the DNA termini (5–7).
Genetic and molecular analyses have thus far identified several
loci that are required for Drosophila telomere maintenance. Fre-
quent telomeric fusions have been observed in mutants in 8 loci:
effete (eff; also called UbcD1), Su(var)205, caravaggio (cav), rad50,
mre11, nbs, telomere fusion (tefu; also called atm), and without
children (woc) (6–19). eff/UbcD1 encodes a highly conserved E2
ubiquitin conjugating enzyme. However, the role of UbcD1 at
Drosophila telomeres is still unclear, and the putative telomere-
associated UbcD1 target(s) remains to be identified (8, 15). Su-
(var)205 encodes heterochromatin protein 1 (HP1), a well-known
component of centric heterochromatin that is also enriched at
telomeres and many euchromatic sites (6, 20). HP1 interacts with
elomeres are nucleoprotein complexes that counterbalance
incomplete replication of terminal DNA and allow cells to
HOAP (HP1/ORC-associated protein), a cav-encoded DNA-
binding polypeptide that shares some similarity with the HMG
proteins and is specifically associated with all Drosophila telomeres
(7, 21). Recent work has shown that HOAP is not only required to
protect telomeres from fusion events but also to prevent activation
of both the DNA damage and the spindle assembly checkpoint
(SAC)(22).TheWocprotein,whichcontains8zincfingerand2AT
hook motifs, is enriched at all telomeres and most polytene chro-
mosome interbands; in interbands, Woc precisely colocalizes with
theinitiatingformofRNApolymeraseII.Thus,Wocappearstobe
a transcription factor with telomere-capping properties, just as
Rap1 in yeast (16).
Accumulation of HOAP-HP1 at telomeres requires the wild-
type function of the Mre11-Rad50-Nbs (MRN) DNA repair com-
plex (9, 10, 14, 17, 18). Mutations in the tefu (ATM) and mei-41
(ATR)genesdonotsubstantiallyaffectHOAP-HP1localizationat
telomeres. However, tefu mei-41 double mutants display greatly
reduced amounts of HOAP-HP1 at chromosome ends, suggesting
that the ATM and ATR kinases play a partially redundant function
required for telomeric localization of HOAP-HP1 (14). The mech-
anism by which these kinases and the MRN complex mediate HP1
and HOAP recruitment at telomeres is unclear. It has been
suggested that interactions between these DNA repair factors and
the DNA termini result in conformational changes of telomeric
chromatin that facilitate association of HOAP-HP1 with chromo-
some ends (10, 14, 15, 19).
Here we describe another gene, modigliani (moi), required for
Drosophila telomere protection. The Moi protein interacts with
both HOAP and HP1, and localizes specifically to all Drosophila
telomeres. Collectively, our results suggest that the HOAP-Moi
complex, we name terminin, is specifically required for the DNA
sequence-independent assembly of Drosophila telomeres. We pro-
pose that Drosophila terminin is the functional analogue of human
shelterin.
Results
Isolation and Characterization of Modigliani. The modigliani1(moi1)
mutation was isolated by a cytological screen of 1680 late lethals
mappingtothethirdchromosome(seesupportinginformation(SI)
Text for details). Examination of DAPI-stained brain preparations
from moi1homozygous larvae revealed that mitotic cells display
Author contributions: G.D.R. and M.G. designed research; G.D.R., G.S., S.C., L.C., and G.C.
performed research; E.W. contributed new reagents/analytic tools; and M.G. wrote the
paper.
The authors declare no conflict of interest.
1Presentaddress:MolecularCellandDevelopmentalBiology,UniversityofCalifornia,Santa
Cruz CA.
2To whom correspondence should be addressed. E-mail: maurizio.gatti@uniroma1.it.
This article contains supporting information online at www.pnas.org/cgi/content/full/
0812702106/DCSupplemental.
© 2009 by The National Academy of Sciences of the USA
www.pnas.org?cgi?doi?10.1073?pnas.0812702106PNAS ?
February 17, 2009 ?
vol. 106 ?
no. 7 ?
2271–2276
GENETICS

Page 2

frequent telomeric fusions (Fig. 1A). Multiple telomeric associa-
tions (TAs) in the same metaphase spread often resulted in
multicentric linear chromosomes that resemble little ‘‘trains’’ of
chromosomes.Themodiglianigenewasnamedafterthisphenotype
just as caravaggio (7), as Caravaggio and Modigliani are the names
of 2 Italian trains.
An analysis of colchicine-treated brains from moi1mutants
revealed that mutant metaphases exhibit an average of 5.6 TAs per
cell, which involve all telomeres (Figs. 1 A and B; and data not
shown). Double telomere associations (DTAs), in which 2 sister
telomeres fuse with another pair of sister telomeres, were approx-
imately 4-fold more frequent than single telomere associations
(STAs) involving a single telomere that associates with either its
sister or a nonsister telomere (Fig. 1C). Comparable patterns of
TAs were previously observed in eff/UbcD1, Su(var)205, cav, woc,
mre11, rad50, nbs, and tefu/atm mutants (6–8, 10, 16, 17).
Recombination analysis with visible markers and deficiency
mapping placed moi in the 90C2–91B1 polytene chromosome
interval uncovered by Df(3R)P14. Complementation tests with
P-element-induced insertional mutations mapping to the same
intervalshowedthatmoi1failstocomplementbothl(3)S096713and
l(3)CB02140 (designated as CG31241CB02140in FlyBase). moi1/
l(3)S096713andmoi1/l(3)CB02140fliesdiedatlarval/pupalbound-
ary and showed telomeric fusions in their larval brains. Thus,
l(3)S096713 and l(3)CB02140 will be henceforth designated as
moiS09andmoiCB0,respectively.ThemoiS09allelecarriesaP{lacW}
construct inserted upstream of the first ORF (ORF1) of the
CG31241/dtl locus; this ORF1 has a coding capacity for a 178-aa
polypeptide (Fig. 2). moiCB0is a protein trap line with a P{PTT-
GB} construct inserted within the same ORF1 (Fig. 2A) (23). We
sequenced the ORF1 of the moi1allele and found a G3A
transition that converts a glycine codon at position 45 into a
glutamicacidcodon(Fig.2A).Wealsogeneratedanadditionalmoi
allele (moiM12) by imprecise excision of the P element inserted into
moiCB0. Genomic DNA sequencing showed that moiM12differs
from moiCB0only for the size of the P insertion. While moiCB0
contains a large P construct (presumably a complete P{PTT-
GB}construct), moiM12retains only 261 bp of the P construct and
a 9-bp duplication of the insertion site (Fig. 2A). The DNA
fragment inserted into the moiM12mutant allele contains a stop
codon leading to a truncation of the ORF1-encoded protein.
The frequency of TAs observed in moi1homozygotes was
comparable to that seen in moi1/Df(3R)P14 hemizygotes, but
substantially higher than those caused by the P element-induced
mutations and their derivatives (moiM12). This suggests that moi1is
a genetically null allele and that moiCB0, moiM12, and moiS09are
hypomorphic alleles; the moi mutations can be ordered in an allelic
series with moi1? moiCB0? moiM12? moiS09(Fig. 1B and data not
shown).
Organization and Function of the moi/dtl Locus.Previousstudieshave
shown that the CG31241/dtl locus produces a single transcript of
2600 nucleotides that can be detected throughout Drosophila de-
velopment (24). This transcript, in addition to ORF1, contains a
second ORF (ORF2) that encodes a 491-aa protein. ORF1 and
ORF2 are in different reading frames and have a 79-bp overlap
(Fig.2A)(24).ExpressionofacDNAincludingbothORFsineither
bacteria or mammalian cells gave rise to 2 different polypeptides of
20 and 60 kDa, corresponding to ORF1 and ORF2, respectively
(24). The polypeptide encoded by ORF1 is conserved in fly species
buthasnohomologywithknownproteins.TheORF2productisan
RNA-binding protein that shares homology with yeast Tgs1 and
mammalian PIMT (24).
Thesefindingspromptedustoinvestigatetheorganizationofthe
CG31241/dtl locus and the roles of the ORF1 and ORF2 products
in telomere protection. We first examined whether the peculiar
organization of this locus was conserved among the Drosophila
species whose genomes were recently sequenced (25). An analysis
of the published DNA sequences revealed that in some species the
2ORFsoverlaplikeinDrosophilamelanogaster,whileinothersthey
are separated by variable numbers of nucleotides (1–25; see Fig.
S1). We next checked whether a D. melanogaster cDNA including
both ORF1 and ORF2 produces 2 proteins in Drosophila cells, as
it does in bacterial and mammalian cells (24). We transfected
Drosophila S2 cells with constructs comprising both ORFs, con-
taining a HA-coding sequence (3HA) attached either upstream of
ORF1ordownstreamofORF2(Fig.2A).Westernblotsofextracts
from transfected cells revealed the presence of ?20 and ?60 kDa
HA-labeled proteins but not of a larger protein resulting from
ORF1 and ORF2 cotranslation (Fig. 2B). These results indicate
that ORF1 and ORF2, although transcribed by a single mRNA,
produce different proteins.
ToanalyzetherolesoftheORF1andORF2proteinproductswe
made 4 constructs that were inserted in flies by germline transfor-
mation and used in complementation tests with moi mutants (Fig.
2 A and C). Two of these constructs included both ORF1 and
ORF2. In 1 of them (R1), both ORFs were wild type; in the other
construct (R2), ORF1 was normal but ORF2 was carrying a
mutation in the putative RNA methyltransferase site (see SI for
details). The other 2 constructs carried wild-type copies of either
ORF1(R3)orORF2(R4).Complementationtestsshowedthatthe
R1,R2,andR3constructsrescuethetelomerefusionphenotypeof
allmoiallelesandmutantcombinations(Fig.2C).moi1mutantflies
bearing 1 of these constructs showed frequencies of TAs ranging
from 0.02 to 0.04 per cell. The R1 construct also rescued the lethal
phenotypeofallmoimutantcombinationsbutdidnotcomplement
the lethality of moi1homozygotes. However, the same construct
rescuedthelethalityofmoi1/Df(3R)P14mutants.Thissuggeststhat,
in addition to the moi1mutation in the ORF1, the moi1-bearing
Fig. 1.
associations in moi mutant neuroblasts. (a) Control (Oregon R) male meta-
phase. (b) Metaphase showing a 3-XL STA (arrowhead), a 2–4 DTA (asterisk),
and a multicentric chromosome (arrow) generated by YL-4–3 DTAs and a 3–2
STA. (c) Metaphase showing a 4–4 DTA (asterisk), a 2–2 dicentric ring chro-
mosome (arrowhead), and a 3–3 DTA (diamond); the XL and 3R telomeres
exhibit sister union STAs (arrows). (d) Metaphase with a multicentric chromo-
some (arrow) containing 3–3 and XR-XR DTAs and 3 STAs involving both sister
telomeres of an XL arm and individual telomeres of chromosomes 2 and 3. (e)
Metaphase containing a chromosome 3 ring generated by a DTA (arrow) and
a multicentric chromosome (arrowhead) generated by a 2-XLXR-4–4-XRXL-
2–3 DTAs. (f) Metaphase with a 4-XR DTA (arrowhead), a chromosome 3 ring
(arrow), and a complex multicentric (asterisk) chromosome involving the rest
of the complement. (Scale bar, 5 ?m.) (B) Frequencies of TAs observed in
homozygous moi mutants and in heteroallelic mutant combinations. Df is
Df(3R)P14.Foreachmutantaminimumof100cellsfromatleast3brainswere
analyzed.
Mutations in moi cause telomeric fusions. (A) Examples of telomeric
2272 ?
www.pnas.org?cgi?doi?10.1073?pnas.0812702106 Raffa et al.

Page 3

chromosome carries a second site lethal independent of moi1. The
R2 and R3 constructs rescued the lethality of moi1/moiCB0, moi1/
moiS09, moi1/Df(3R)P14, and moiM12/moiM12mutants but did not
complement the lethal phenotype of moiS09/moiS09and moiCB0/
moiCB0homozygotes. Finally, the R4 transgene did not rescue the
telomerefusionandthelethalphenotypeinanyofthemoimutants
(Fig. 2C). Thus, point mutations in ORF1 such as moi1, or muta-
tionsleadingtotruncationoftheORF1productsuchasmoiM12,are
fully complemented by a transgene that contains only ORF1. In
contrast,transgenesbearingeitherORF1alone(R3)orORF1plus
a mutated form of ORF2 (R2) did not rescue the lethal phenotype
associated with P element-insertions upstream of or within ORF1
(moiS09and moiCB0). However, moiS09and moiCB0homozygotes
were fully viable in the presence of a transgene that contains the
wild-typeformsofbothORF1andORF2(Fig.2C).Together,these
resultsindicatethatORF1andORF2specify2differentgenesboth
required for fly viability, and suggest that P-element insertions
within the ORF1 inactivate both genes, consistent with their
transcription in a single bicistronic mRNA. In addition, our results
clearly show that ORF1 is required to prevent telomeric fusions.
ORF2 does not appear to be involved in telomere protection
because moiS09and moiCB0homozygotes bearing an R2 construct
are lethal but do not exhibit TAs. Thus, we propose to name
modigliani the gene identified by ORF1, and to retain the Drosoph-
ila Tat like (dtl) designation (24) for the gene identified by ORF2.
Moi Specifically Localizes at Telomeres and Interacts with HOAP. To
investigate the subcellular localization of the Moi protein we
transformedflieswithaconstructcontainingORF1fusedinframe
with the GFP coding sequence (R5, Fig. 2A). Complementation
experiments showed that this construct rescues the telomere fusion
phenotype of moi mutants. Analysis of mitotic chromosomes from
bothlivingandfixedlarvalbrainsdidnotrevealanyclearGFP-Moi
signal at chromosome ends (data not shown). However unfixed
polytene chromosome nuclei displayed 6 discrete GFP-Moi signals
(Fig. 3A). The same type and number of fluorescent signals were
observed in unfixed polytene chromosome nuclei from salivary
glandsoffliesthatexpressHOAP-GFP(Fig.3A).Thesesignalsare
likely to correspond to telomeres, as previous studies have shown
that HOAP is specifically enriched at all polytene chromosome
ends (9–11, 16). We thus fixed polytene chromosomes with meth-
anol-free formaldehyde to preserve GFP fluorescence and immu-
nostained them with both anti-GFP and anti-HOAP antibodies.
GFP immunostaining was aimed at increasing the natural GFP
fluorescent signal. In these preparations, the GFP-Moi and the
HOAP signals fully overlapped, suggesting that HOAP and Moi
colocalizeatthetelomeres(Fig.3B).Becausethefixationtechnique
used in the latter experiments does not preserve polytene chromo-
some morphology (Fig. 3B), we analyzed HOAP and Moi local-
ization in formaldehyde/acetic acid fixed preparations (see Mate-
rials and Methods) that allows clear visualization of individual
polytene chromosome telomeres (20). Immunostaining of these
preparations with both anti-GFP and anti-HOAP antibodies re-
vealed that the signals elicited by these antibodies are exclusively
telomericandfullycoincident(Fig.3C).Stainingwiththeanti-GFP
antibody produced some weak signals along the polytene chromo-
somearms.Howeverthesamesignalswerealsoobservedincontrol
polytene chromosomes from salivary glands that do not express
GFP-Moi, indicating that they are background signals that do not
reflect the GFP-Moi localization. Collectively, these experiments
Fig. 2.
intron (thin line) while ORF2 is intronless. MT, methyltransferase domain. The triangles above and below the diagram indicate the sites of P-element insertions
in moiS09, moiCB0, and moiM12; the vertical line indicates the point mutation in the moi1allele. The T1 and T2 constructs that contain 3HA-coding sequences at
theirendshavebeenusedfortransfectionexperimentsofS2cells(seeTextandpartBofthefigure).TheR1-R5constructswereusedforgermlinetransformation
and subsequent complementation analysis (part C of the figure), and for GFP-Moi localization (R5). The cross marks the S398R W401G mutations that are likely
toinactivatetheMTdomain.(B)WesternblotsofextractsfromS2cellstransfectedwitheithertheT1orT2constructshowingthatthe2ORFsofCG31241encode
distinct polypeptides. ORF1 and ORF2 indicate polypeptides of sizes corresponding to HA-ORF1 and ORF2-HA detected with an anti-HA antibody. The asterisk
marks an additional band in the T2 lane generated either by degradation of the ORF2-HA product or by transcription from an internal ATG. Note the absence
ofaproteinresultingfromcotranslationofORF1andORF2(arrow)L.C.,loadingcontrol.(C)AbilityoftheR1-R4constructstorescuethemoimutantphenotypes
in lines carrying the indicated allelic combinations. Df, Df(3R)P14;TF, telomeric fusions; VB, viability; ?, rescues mutant phenotype; ?, does not rescue mutant
phenotype
Organization of the CG31241 locus. (A) Schematic representation of the 2 overlapping ORFs that comprise the CG31241 locus: the ORF1 contains an
Raffa et al.
PNAS ?
February 17, 2009 ?
vol. 106 ?
no. 7 ?
2273
GENETICS

Page 4

indicate that HOAP and Moi are enriched only at telomeres where
they precisely colocalize.
TheseresultspromptedustoaskwhetherMoiphysicallyinteracts
with HOAP and its binding partner HP1. We thus performed a
GST-pulldown assay by incubating GST-Moi with extracts from S2
cells expressing HOAP-FLAG. As shown in Fig. 4A, HOAP was
capturedbyGST-MoibutnotbyGSTalone.WethenmixedS2cell
extracts expressing HOAP-FLAG, GFP-FLAG, and HA-Moi in
variouscombinationswitheitherGST-HP1orGSTalone(Fig.4B).
GST-HP1efficientlyprecipitatedbothHOAP-FLAGandHA-Moi
butnotGFP-FLAG(usedasacontrolforbindingspecificity),while
GSTalonedidnotinteractwitheitherofthesetaggedproteins(Fig.
4B). We finally performed a pulldown assay using purified bacte-
riallyexpressedGST-HOAP,GST-HP1,andHis-Moi.His-Moiwas
capturedbybothGST-HOAPandGST-HP1butnotbyGSTalone
(Fig. 4C), suggesting that Moi interacts directly with both HOAP
and HP1. These results strongly suggest that HOAP, Moi, and HP1
form a complex that specifically localizes at chromosome ends,
protecting them from fusion events.
Recruitment Dependencies of the Moi Protein. To define the role of
moi in telomere protection, we asked whether it is required for
proper localization of HP1, HOAP, and Woc. In polytene chro-
mosomes, HOAP is primarily bound to chromosome ends (9–11,
14,16–18).Inmitoticchromosomes,HOAPisspecificallyenriched
at telomeres, while HP1 and Woc associate with multiple hetero-
chromatic and euchromatic sites (6, 9–11, 14, 16–18). Thus, while
accumulations of HOAP at mitotic telomeres can be clearly
visualized, unambiguous detection of HP1 or Woc signals is
extremely difficult (6, 16). Owing to these technical problems,
we analyzed HP1 and Woc localization only in polytene
chromosomes; HOAP localization was examined in both mi-
totic and polytene chromosomes.
The analysis of mitotic chromosomes revealed that mutations in
moi do not substantially affect HOAP localization at telomeres
(Fig. 5A). In Oregon R controls, 91% of telomeres (n ? 280)
displayed a clear HOAP signal. Similarly, in moi1homozygous
mutants, 86% of the telomeres not involved in fusion events (n ?
110)and60%oftheattachedtelomeres(n?150)showedaHOAP
signal. Consistent with these results, the polytene chromosome
telomeres of moi1mutants showed normal concentrations of
HOAP (Fig. 5B). Our immunostaining experiments also showed
that moi1homozygous mutants display normal amounts of both
Fig. 3.
in salivary glands nuclei from third instar larvae expressing GFP-Moi or GFP-HOAP. Note the 6 fluorescent signals in the nucleus that are likely to correspond to
the euchromatic Drosophila telomeres (XL, 2L, 2R, 2L, 3R, and 4R). (B) Salivary glands nuclei from a GFP-Moi expressing strain fixed with methanol-free
formaldehyde and coimmunostained with anti-GFP and anti-HOAP antibodies. Note that GFP-Moi colocalizes with HOAP in 6 discrete polytene chromosome
regions. (C) Examples of polytene chromosome ends fixed according to ref. 20 and immunostained with anti-GFP and anti-HOAP antibodies. Note the precise
colocalization of GFP-Moi and HOAP at the telomeres.
Moi specifically localizes at the telomeres of polytene chromosomes. (A) In vivo localization of GFP-Moi (top panels) and GFP-HOAP (bottom panels)
2274 ?
www.pnas.org?cgi?doi?10.1073?pnas.0812702106Raffa et al.

Page 5

HP1andWocattheirpolytenechromosomesends(Fig.5BandC).
Collectively,theseresultsindicatethatthewild-typefunctionofmoi
is not required for telomeric localization of HOAP, HP1, and Woc,
and that the strong telomere fusion phenotype observed in moi
mutants is not because of the absence of any of these proteins. We
also found that the impairment of the Moi function does not result
inarobustspindleassemblycheckpoint(SAC)activation(TablesS1
and S2), indicating that the SAC is specifically triggered by HOAP-
depleted telomeres (22).
We also analyzed Moi localization in polytene chromosomes
from cav and mre11 mutants that express GFP-Moi. GFP-Moi
failed to accumulate at the telomeres of both mutants (data not
shown), indicating that Moi recruitment requires both HOAP and
Mre11. However, because the MRN complex mediates HOAP
localization at telomeres (9, 10, 14, 17, 18), Mre11 is likely to play
an indirect role in Moi recruitment.
Discussion
We have shown that the Moi protein is enriched exclusively at
telomeres, where it colocalizes and physically interacts with both
HOAP and HP1. We have also shown that Moi is not required for
HOAP accumulation at telomeres, whereas Moi localization re-
quires the wild-type functions of cav and mre11. These results
suggestamechanismforMoilocalizationattelomeres.Wepropose
thattheDrosophilachromosomeends,whichcontainvariableDNA
sequences, are processed and shaped by the MRN complex so as to
allow binding of HOAP, which would in turn recruit Moi.
TheMoi-HOAPcomplexsharesseveralanalogieswithshelterin,
a 6-protein complex that protects human chromosome ends, allow-
ing cells to distinguish telomeres from sites of DNA damage (1).
Shelterin is comprised of 3 polypeptides that directly bind the
TTAGGG telomeric repeats (TRF1, TRF2, and POT1) intercon-
nected by 3 additional proteins (Tin2, TPP1, and Rap1). The
shelterin subunits share 3 properties that distinguish them from the
nonshelterin telomere-associated proteins. They are specifically
enrichedattelomeres;theyarepresentattelomeresthroughoutthe
cell cycle; and their functions are limited to telomere maintenance
(1).WiththeexceptionofTin2andTPP1,shelterin-relatedproteins
have been found in most eukaryotes. However, none of the
shelterinsubunitsareconservedinDrosophila.Thisisnotsurprising
asDrosophilatelomeresareDNAsequence-independentstructures
(4, 5, 15, 19), while the core subunits of shelterin are sequence-
specific DNA binding proteins (1).
The Moi and HOAP proteins have the same properties of the
shelterinsubunits:theyaccumulateonlyattelomeres;theyarelikely
to be associated with telomeres throughout the cell cycle, as they
colocalizeindiscreteaggregatespresentinallinterphasenucleiand
are enriched at polytene chromosome telomeres; and they appear
to function only at telomeres. HP1 interacts with both Moi and
HOAP but does not share their properties; it localizes to multiple
chromosomal sites and its function is not limited to telomere
maintenance (6, 20). Notably, Moi and HOAP are not conserved
in either yeasts or mammals, consistent with the fact that both
proteins associate with telomeres in a sequence-independent fash-
ion. Thus, we propose that Moi and HOAP are the founding
components of a Drosophila telomere complex, we name ‘‘termi-
nin,’’ which acts like human shelterin. We suggest that terminin
accumulation at chromosome ends prevents both checkpoint acti-
vation and telomere fusion and helps in recruiting nonterminin
componentsofDrosophilatelomeres.Thishypothesispositsthatthe
nonterminin proteins of Drosophila telomeres should be conserved
in humans and play roles in telomere maintenance. Similarly,
nonshelterin components of human telomeres should have con-
served Drosophila homologues. Indeed, all of the nonterminin
proteins specified by the Drosophila telomere-fusion mutants so far
identified have human counterparts. UbcD1 and Woc have highly
conserved human homologues but it is currently unknown whether
any of them is involved in telomere maintenance. HP1 too is
conserved in humans, and HP1 homologues have been found at
mouse telomeres where they appear to control telomere length
(26). The ATM kinase and the components of the MRN complex
(Mre11, Rad50, and Nbs) have highly conserved human ortho-
logues, which bind shelterin and help to regulate human telomere
organization (27, 28).
Fig.4.
HOAP-FLAG from S2 cell extracts. (B) GST-HP1 precipitates both HOAP- FLAG
and HA-Moi from S2 cell extracts. (C) GST-HOAP and GST-HP1 precipitate
bacterially purified His-Moi. Input is 10% in all assays. WB, Western blotting;
P, Ponceau staining.
MoiphysicallyinteractswithHOAPandHP1.(A)GST-Moiprecipitates
Fig. 5.
WOC. (A) moi1mitotic chromosomes immunostained for HOAP (red) and
counterstained with DAPI. Note that HOAP localizes at both free and fused
telomeres. (B) Wild-type (WT) and moi1polytene chromosomes immuno-
stained for both HOAP and HP1 and counterstained with DAPI. HP1 staining
isshowninblackandwhite;HOAPstaining(red)ismergedwithDAPIstaining
(black and white). Note that the telomeres of mutant chromosomes accumu-
late normal amounts of HOAP and HP1. (C) Wild-type (WT) and moi1polytene
chromosomes (distal regions of XL) immunostained for Woc. The telomeric
Woc signal in moi1is comparable to the wild-type signal (arrows).
Moi is not required for telomeric localization of HOAP, HP1, and
Raffa et al.
PNAS ?
February 17, 2009 ?
vol. 106 ?
no. 7 ?
2275
GENETICS

Page 6

In addition to ATM, Mre11, Rad50, and Nbs1, the human
nonshelterin factors include Ku70 and Ku80 and their associated
DNA-dependentproteinkinasecatalyticsubunit(DNA-PKcs),the
ATR kinase, PARP1 and PARP2, Rad51, the ERCC1/XPF endo-
nuclease,theApollonuclease,andtheRecQfamilymembersWRN
and BLM, which are mutated in the Werner and Bloom syndrome,
respectively (1). With the exception of DNA-PKcs, all these non-
shelterinproteinshaveDrosophilahomologues(4,15,19,29).There
is also evidence that Drosophila ATM and ATR cooperate to
prevent telomere fusion (14, 17, 18), and that Ku70 and Ku80 act
as negative regulators of Drosophila telomere elongation by trans-
position (30). However, it is not currently known whether the fly
homologuesofRad51,ERCC1,Apollo,WRN,andBLMplayroles
at Drosophila telomeres (4, 15, 19, 29).
In summary, it is clear that the terminin and shelterin compo-
nentsarenotevolutionarilyconserved.Incontrast,thenonterminin
and nonshelterin proteins are largely conserved from flies to
mammals, and many of them play telomere-related functions in
bothDrosophilaandhumans.Thissuggeststhatthemaindifference
between Drosophila and human telomeres is in the protective
complexes that specifically associate with the DNA termini. Thus,
apart from the different mechanisms of elongation, Drosophila and
humantelomeresmightnotbeasdifferentasitisgenerallythought.
Materials and Methods
Drosophila Strains. The moi1allele was isolated from a collection of 1680 EMS-
induced third chromosome late lethals generated in Charles Zuker’s laboratory.
Thel(3)S0967–13,CG31241CB02140,andDf(3R)P14aredescribedinFlyBase(http://
flybase.bio. indiana.edu/). moiM12was generated by incomplete excision of the
P{PTT-GB} element inserted into the CG31241 locus. The moi1and moiM12alleles
were characterized by DNA sequence analysis using standard methods. The cav1
mutation has been described previously (7).
Rescue Constructs and Complementation Analysis. The R1-R5 DNA fragments
wereclonedintothepCASPER4plasmidunderthecontrolofatubulinpromoter
(31). Molecular details on these DNA fragments are reported in SI. The R2
fragment was mutagenized in vitro to introduce the S398R and W401G amino
acid substitutions within the methyltransferase domain of Dtl. Germline trans-
formation with the R constructs was performed by standard methods. Comple-
mentation analysis was carried out using 2 different transgenic lines for each
construct.
Cell Transfection and Western Blotting Analysis. The T1 and T2 transfection
constructs were generated like the R constructs described above. The genomic
fragment of T1 was cloned in frame to the 3? end of a 3HA epitope tag; in T2, a
genomic fragment lacking the stop codon was cloned in frame at the 5? end of
3HA. S2 cells were transfected using Cellfectin (Invitrogen) and harvested 72 h
after transfection; extracts were prepared by standard methods and immuno-
blotted as described in ref. 32. The HA epitope was detected with the anti-HA
12CA5 antibody (Roche).
Chromosome Cytology and Immunostaining. Preparation and immunostaining
of mitotic and polytene chromosomes were described previously (7, 16). For in
vivo detection of GFP-Moi and GFP-HOAP, salivary glands were dissected in
Voltalefoilandimmediatelyanalyzedunderafluorescencemicroscope(Fig.3A).
For immunodetection of GFP-Moi, salivary glands were either fixed with meth-
anol-freeformaldehyde(Fig.3B)orpreparedasdescribedinref.20(Fig.3C);they
were then costained with mouse anti-HOAP and rabbit anti-GFP polyclonal
antibodies, both generated in our laboratory. See SI for details on fixation and
staining techniques.
GST-Pulldown Assays. To obtain GST fusion proteins, full-length moi, cav, and
Su(var)205 cDNAs were cloned in pGEX-3X in frame with the GST sequence.
HA-MoiexpressingcellswereobtainedbytransfectionwiththeT1construct(Fig.
2A). HOAP-FLAG and GFP-FLAG expressing cells were obtained by transfection
with a pAWF vector (Drosophila Genomics Resource Center, Indiana University).
GST puldown was carried out by standard methods. FLAG-HOAP, HA-Moi, and
6His-Moiweredetectedwithanti-FLAGHRP-conjugated(1:500,Roche),anti-HA
12CA5 (1:1000, Roche) and anti-His HRP-conjugated (1:500, Roche) antibodies,
respectively. See SI for details on GST pulldown assays.
ACKNOWLEDGMENTS. We thank Rebecca Kellum (University of Kentucky, Lex-
ington, KY) and Sally Elgin (Washington University, St. Louis) for providing the
anti-HOAPandanti-HP1antibodies,respectively;andGianlucaCestra(University
of Rome La Sapienza, Rome) for the anti-GFP antibody. We also thank Maria
Grazia Giansanti for her advice on fixation techniques for GFP visualization.
G.D.R.wassupportedbyafellowshipfromtheCenciBolognettiFoundation.This
work was supported in part by grants from the Italian Association for Cancer
Research and the Italian Telethon (to M.G.).
1. Palm W, de Lange T (2008) How shelterin protects mammalian telomeres. Annu Rev
Genet 42:301–334.
2. McEachern MJ, Krauskopf A, Blackburn EH (2000) Telomeres and their control. Annu
Rev Genet 34:331–358.
3. Pardue ML, DeBaryshe PG (2003) Retrotransposons provide an evolutionarily robust
non-telomerase mechanism to maintain telomeres. Annu Rev Genet 37:485–511.
4. Mason JM, Frydrychova RC, Biessmann H (2008) Drosophila telomeres: an exception
providing new insights. Bioessays 30:25–37.
5. Mason JM, Biessmann H (1995) The unusual telomeres of Drosophila. Trends Genet
11:58–62.
6. Fanti L, Giovinazzo G, Berloco M, Pimpinelli S (1998) The heterochromatin protein 1
prevents telomere fusions in Drosophila. Mol Cell 2:527–538.
7. Cenci G, Siriaco G, Raffa GD, Kellum R, Gatti M (2003) The Drosophila HOAP protein is
required for telomere capping. Nat Cell Biol 5:82–84.
8. Cenci G, et al. (1997) UbcD1, a Drosophila ubiquitin-conjugating enzyme required for
proper telomere behavior. Genes Dev 11:863–875.
9. Bi X, Wei SD, Rong YS (2004) Telomere protection without a telomerase: the role of
ATM and Mre11 in Drosophila telomere maintenance. Curr Biol 14:1348–1353.
10. Ciapponi L, et al. (2004) The Drosophila Mre11/Rad50 Complex is required to prevent
both telomeric fusion and chromosome breakage. Curr Biol 14:1360–1366.
11. Oikemus SR, et al. (2004) Drosophila atm/telomere fusion is required for telomeric
localization of HP1 and telomere position effect. Genes Dev 18:1850–1861.
12. Silva E, et al. (2004) ATM is required for telomere maintenance and chromosome
stability during Drosophila development. Curr Biol 14:1341–1347.
13. Song YH, Mirey G, Betson M, Haber DA, Settleman J (2004) The Drosophila ATM
ortholog,dATM,mediatestheresponsetoionizingradiationandtospontaneousDNA
damage during development. Curr Biol 14:1354–1359.
14. Bi X, et al. (2005) Drosophila ATM and ATR checkpoint kinases control partially
redundant pathways for telomere maintenance. Proc Natl Acad Sci USA 102:15167–
15172.
15. Cenci G, Ciapponi L, Gatti M (2005) The mechanism of telomere protection: a compar-
ison between Drosophila and humans. Chromosoma 114:135–145.
16. Raffa GD, Cenci G, Siriaco G, Goldberg ML, Gatti M (2005) The putative Drosophila
transcription factor woc is required to prevent telomeric fusions. Mol Cell 20:821–831.
17. Ciapponi L, Cenci G, Gatti M (2006) The Drosophila Nbs protein functions in multiple
pathways for the maintenance of genome stability. Genetics 173:1447–1454.
18. Oikemus SR, et al. (2006) Epigenetic telomere protection by Drosophila DNA damage
response pathways. PLoS Genet 2:e71.
19. Rong YS (2008) Telomere capping in Drosophila: dealing with chromosome ends that
most resemble DNA breaks. Chromosoma 117:235–242.
20. James TC, et al. (1989) Distribution patterns of HP1, a heterochromatin-associated
nonhistone chromosomal protein of Drosophila. Eur J Cell Biol 50:170–180.
21. Shareef MM, et al. (2001) Drosophila heterochromatin protein 1 (HP1)/origin recog-
nition complex (ORC) protein is associated with HP1 and ORC and functions in hetero-
chromatin-induced silencing. Mol Biol Cell 12:1671–1685.
22. Musaro M, Ciapponi L, Fasulo B, Gatti M, Cenci G (2008) Unprotected Drosophila
melanogaster telomeres activate the spindle assembly checkpoint. Nat Genet 40:362–
366.
23. Buszczak M, et al. (2007) The carnegie protein trap library: a versatile tool for
Drosophila developmental studies. Genetics 175:1505–1531.
24. Komonyi O, et al. (2005) DTL, the Drosophila homolog of PIMT/Tgs1 nuclear receptor
coactivator-interacting protein/RNA methyltransferase, has an essential role in devel-
opment. J Biol Chem 280:12397–12404.
25. Clark AG, et al. (2007) Evolution of genes and genomes on the Drosophila phylogeny.
Nature 450:203–218.
26. Garcia-Cao M, O’Sullivan R, Peters AH, Jenuwein T, Blasco MA (2004) Epigenetic
regulationoftelomerelengthinmammaliancellsbytheSuv39h1andSuv39h2histone
methyltransferases. Nat Genet 36:94–99.
27. Zhu XD, Kuster B, Mann M, Petrini JH, de Lange T (2000) Cell-cycle-regulated associa-
tion of RAD50/MRE11/NBS1 with TRF2 and human telomeres. Nat Genet 25:347–352.
28. Karlseder J, et al. (2004) The telomeric protein TRF2 binds the ATM kinase and can
inhibit the ATM-dependent DNA damage response. PLoS Biol 2:E240.
29. CiapponiL,CenciG(2008)Telomerecappingandcellularcheckpoints:Cluesfromfruit
flies. Cytogenet Genome Res, 10.1159/000167824.
30. Melnikova L, Biessmann H, Georgiev P (2005) The Ku protein complex is involved in
length regulation of Drosophila telomeres. Genetics 170:221–235.
31. Bruckner K, et al. (2004) The PDGF/VEGF receptor controls blood cell survival in
Drosophila. Dev Cell 7:73–84.
32. Somma MP, Fasulo B, Cenci G, Cundari E, Gatti M (2002) Molecular dissection of
cytokinesis by RNA interference in Drosophila cultured cells. Mol Biol Cell 13:2448–
2460.
2276 ?
www.pnas.org?cgi?doi?10.1073?pnas.0812702106Raffa et al.