DNA repair proteins - m echanism s and functions

H ans E. Krokan, Institute of R esearch and M olecular M edicine, Norw egian U niversity of S cience and T echnology, T rondheim , Norw ay DNA w as assum ed to be inert and not subject to dam age and repair

J . W atson and F. Crick (1953): “It has not escaped our notice that the specific pairing w e have postulated im m ediately suggests a possible copying m echanism for the genetic m aterial.” F. Crick (1974): “W e totally m issed (in 1953) the possible role of enzym es in repair,… … … I later cam e to realize that D N A is so precious that probably m any distinct repair J . W atson F. Crick m echanism s w ould exist.” M utations are required for m alignant transform ation, w hich follow s tw o paths

Gatekeeper pathway Caretaker pathway

Mutation of a 2nd of a 2nd Mutation of a Mutation of a caretaker gene gatekeeper gene caretaker gene → gatekeeper gene allele leads to → → allele leads to allele allele genetic instability tumour initiation

Caretaker genes: Genes required for DNA repair, DNA replication and detoxification of carcinogens Gatekeeper genes: Genes for protooncogenes and tumour suppressor genes

*Adapted from Kinzler and Vogelstein, Nature 386:761-763 (1997) DNA R EPAIR G ENES AND CANCER Rare human cancer forms: - skin cancer; deficiency in one of at least seven different genes for nucleotide excision repair (NER) Fanconi’s anemia - leukemia and solid tumors; four complementation groups, breakage, genes not identified Ataxia teleangiectasia - mostly lymphomas. Very sensitive to ionising radiation. Mutation in ATM- gene. Function not clear Bloom’s syndrome -many cancer forms. Mutation in BLM-gene, a RecQ helicase-gene homologue

More common human cancer forms: Early onset hereditary breast cancer -Mutation in BRCA2-gene, responsible for 50% of these . Brca2 protein may be an essential cofactor for HsRad51, involved in repair of double strand DNA breaks Hereditary nonpolyposis colorectal cancer (HNPCC) - 5% of all colorectal cancers. Deficient mismatch repair. in hMSH2 (45%) or hMLH1 (45%) most common Hereditary colon cancer with polyposis - hMYH-mutations (Y165G and G382D), (not frequent) Sporadic colorectal cancer – defective mismatch repair genes in 12-15% of all cases Sporadic lung cancer: - Low activity of DNA glycosylase hOGG1 and some SNPs enhance risk, Mutations in other BER enzymes genes enhance risk of lung cancer in smokers (APE1 and XRCC1) DNA Dam age and R epair - Overview

From : H oeijm akers 2001, Nature 411:366-74. S PONT ANEOU S DAM AG E OCCU R S FR EQ U ENT LY

II I. Loss of base: Depurination → AP-site ( 10 000/cell/day ) Depyrimidination → AP-site ( 500/cell/day ) I II. Deamination: Adenine → Hypoxanthine ( 10/cell/day ) Cytosine → Uracil ( 500/cell/day ) I III. Oxidative damage: Guanine → 8-oxoGuanine ( 10 000/cell/day ? ) Thymine → Thymine glycol ( 500/cell/day ? ) I IV. Alkylation damage: Guanine→ O6-methylGuanine (spont./induced) Adenine→ 3-methylAdenine ( spont./induced) Consequences: 1. Mispairing in replication (mutation) 2. Block of replication (cytotoxic) 3. Block of transcription (cytotoxic) (BER )

• BER rem oves dam aged or inappropriate bases that do not Short patch Long patch 5‘ 3‘ cause helix distortion • Damaged base excised by DNA glycosylase, leaving AP site 3‘ 5‘ Unmodified Reduced or oxidised • D eficient BER m ay cause cancer AP-sites AP-sites and im m une deficiency AP-sites cleaved by APE1 (stim. by Polβ) APE1 APE1 β β • Im portant for developm ent 5‘-deoxyribose- Strand displacement and phosphate resynthesisby Pol β/δ/ε, - polβ: em bryonic lethal RF-C, PCNA XRCC1 removed and new nucleotide filled in by - X R CC1: em bryonic lethal β Polβ/XRCC1 β RFC Pol

• hM Y H -def. : colorectal cancer Strand ligationby Flap cleavage by FEN-1 XRCC1/LIG3 XRCC1 • hO G G 1-def.: lung cancer LIG3

• hU N G : defect antibody 1 nucleotide patch Ligationby LIG1, aided m aturation (defective CS R and by PCNA/FEN-1 S H M ) LIG1 • m U N G : B-cell lym phom a (late) 2-8 nucleotide patch • m U N G : Increased postischem ic brain injury (Endres et al., 2004) U R ACIL IN DNA • S pontaneous deam ination of cytosine O NH 2 N H N U :G O O NH NH Cytosine U racil

• Enzym atic deam ination of C by AID in B -lym phocytes Enzym atic deam ination of C to U is an early step in affinity m aturation of antibodies - recently discovered

• Incorporation of dU M P during DNA replication (quant.dom .)

dTTP, dCTP, dGTP, dATP d U T P U :A H ow m uch uracil in DNA ?

H ow im portant is nuclear U NG 2 for rem oval of uracil ?

T he structure of U NG catalytic dom ain is 10 years in 2005 Catalytic dom ain of U NG com m on to U NG 1 and U NG 2 Flipping of uracil into tight fitting pocket U racil in DNA analyzed by alkaline elution or com et assay p b 6

0 M EFs - alk. elution 1

0.8 /

s e

t 0.6 i s

-/- -/- . Exp. 1, Ung s 0.4 -/- U ng n Exp. 2, Ung e

s 6 - 0.2 +/+ M EFs: 0.6 per 10 bp , or ~3600

G Ung

N per dipl. genom e; U 0 0 20 40 60 80 100 120 Liver: ~900 per dipl. genom e

p UNG [ng/ml] b

6 Cellspecific 0

1 A lk. elution / +/+ s 0.3 Ung

e -/- Low in sperm atozoa

t Ung i n=6 s

. 0.2 n=5 s n

e +/+ s

- 0.1 n=4 n=6 U ng

G n=3

N n=3 M EFs: Below detection level

U 0 Hepatocytes Splenocytes Spermatozoa (0.02 ± 0.05 per 106 bp); -/-

) A t least 10-fold low er than U N G U 200 Com et assay n=9 A +/+ (

n=6 Ung

l -/-

e 150 Ung v e l 100 l

i n=9

c n=6

a 50 n=8 r n=8 U 0 Kidney Liver Spleen Testis M ore uracil in DNA of proliferating than in nonproliferating cells p

b 0.8 Ung +/+ 6 -/- 0 M EFs Ung 1

/

0.6

s Alk. elution e t i s

e 0.4 v i t i s n

e 0.2 s - G

N 0 U Exponent. 8 days of 24 h 72 h growth confluence after replating ) 3 - Ung +/+ 0 Prim ary splenocytes

1 -/- Ung x 25 300

M Com et assay P ) 250 U C 20 ( A

( .

l 200 p e

r 15 v o e l

c 150 l i n

i 10 c

a 100 r d U h 5

T 50

H

3 0 0 LPS + ConA LPS + ConA Conclusions

U ng2 is very im portant for rem ovalof uracilfrom D N A , but even w hen U ng2 is not present uracilis eventually rem oved

H igher uracilcontent in proliferating cells points to dU M P incorporation as a m ajor quantitative source

U ng2 very im portant for rem ovalincorporated uracil

T his does not exclude a role in repair of deam inated cytosine H ypothesis:

Does B ER take place in organized, preform ed com plexes?

Answ er: Probably yes Assay for short- and long patch BER

1 2 3 Substrate U/G AP/G C/G pGEM-3Zf(+) 3199 bp A HMW

U/G (or U/A, AP/G, AP/A) Short-patch BER

B HMW BamHI XbaI HincII PstI

5’GATCCTCTAGAGTUGACCTGCA3’ 3’CTAGGAGATCTCAGCTGGACGT5’ Long-patch BER

short-patch long-patch (8-mer) (8-mer) C HMW A B

ligation Ligated (22-mer) C Unligated U U D 2 3 3 k N N 5 0 4 D B . . . G G a 5 8 6 2 1 E R (

3 UNG2 P N U 5 C U - n / t 4 e i O q 4 1 r P u

m a UNG1 0 e U M a . 1 1 ) s P u C b Cat. domain L N P ( o 4 U - m E 2 t e 1

m P a r X m a UNG2 o U ) . n

1 - s C ( 2 C c P o UNG1 u a 2 - U m t 7 t A b e a 1

m r a l 0 y m F a o t 1 i

) Cat. domain . c n F I 3 4 5 1 2 N . . . . .

I

( W o T P a I I M s P [ f 3 U i Y n r n s t 2 a

i e o a c 1 U

x P e s p l l I s u ]

h y a N e u a o d b S z t

b r x r G b C a O e e e

t t e T 2

n

i L

r D e b s a i c

P a A c

a y d N e / b s c k n

T s l d p e P t A

l d

a e I T

A a 4 e

, w t c 0 d d

T

x G c

i t i N o s d f t P t l E i

e e m h i e r ,

w c

a s a

f B f A e

i P

v c i f o t Y n s n U T e t h o r o

e

P w

1 r

U w t f s d p s

i r a h N i l N

t u a t a o p n o h f b - h s G m o d r f

t - m

s e t

2 D e r e

a i H x - i e

c r t y p d p c A i s m t e n p a o

e t

c L N a r i t U n r s n o a b o c , i s T N a

n c h p e d

c l I t t b r G

a e N a o n i B i u d o - l a i r o T l O f a n s n t s

n f c i l

P

e D o n - ( e e t s

p c g n I n b r i , a a v

g

z E u u r t i

y t f S r p a a y m f a a l m

y

e c t e ) a t r i c

, l i g

h , c

n

A e B p P a t E - i r R c t )

U NG 2 is the m ajor enzym e for U /A and ”sole” enzym e for U /A

Substrate U/A PU 101 (anti-U N G ) com pletely inhibits U /A UNG2-ARC Extract repair and U /G repair by repairosom e PU101 + + + + (U N G 2-A R C) PSM1 + + + + HMW PS M 1 (anti-S M U G 1) has no effect on U /A

60 min or U /G repair in repairosom e

U/G U/G PU 101 also com pletely inhibits U /A repair

Substrate U/G in extracts, but only partially inhibits 10 min U /G repair

60 min Even in extracts anti-S M U G 1 has a partial y

t effect on U /G repair i BER v BER i

t 10 min 10 min c

a 60 min 60 min

R 1

E 1 B

e

v 0.5 i t 0.5 a l e R C

P 0 P P P P P C P P o S U U U 0 U S S S o n M 1 1 1 1 M n M M t 0 0 0 0 1 t . 1 1 1 . 1 1 1 1

+ + e , s e i s m i G s o s s m s /

s e k t i e o k i y o U l s c o d r l i c d

i i o n o a n r f o a i n

y i s

t b y o i b r p d d a i t e d d t e a n o p n r o n r o n i

p a b e a d a

i b a r a y

i - t

s - 1 t s p 1 b n y e G n e e G t G r a i t G b i a i N r - U s N r - U s a

i U U

G M n c p

G a M i c x i y S i N y s e S

p N e s r t b l U t

b a 2 e U

a n

r p n d b

y G G d b a y a / e a m b A e n a t

b N i n t

f U o / i f c d b c c o U s U d b o

e e r e f e e t i r f m f r e i r t i s m f i i r a o a t o b a o a

a o i b t

t p c i n t t p c s p h s o h e e e o e e e e n e n R r R i n n R i r R r m e l n

i

p U G / U A / + + + + + + m A o

/

c

U U G / A U /

+ +

+ + 1 t

G

o

A i n P - A k c i n - P G / k c / n A U + + t M

u

S

A G P / A A P b / + + t n

a

U G / U A / n + + i b

m U G / U A / o c e 1 1 h 1 1 G 1 h e c G W R 0 e 1 W 0 t t c t U 1 M U t 1 M a a M a M r a S M U r p M U H t S t H

p P S P t s

S P s P

r

t . b . b r o c c u u o e h e S r h S r S S M itochondria are com petent in B ER , but m itochondrial B ER proteins do not form a com plex

BER assays Western blot (mit. extr.) ) ) ) ) ) ) n n n n n n i i i i i i s s s s s s p p p p p p y y y y y y r r r r r r t t t t t t

- + - + + - ( ( ( ( ( (

a a a a a a i i i i i i n r r r r r r n i d d d d d d i s n n n n n n s r r o o o r o o o p p a a a h h h h h h y e e e y c c c c c c r l l l r o o o o o o c c c t t t t t t T i i i i i i T

u u u M M M M M M N N N + -

1 2 3 4 5 6 7 8 9 1 2

UNG1-GFP HMW UNG2 UNG1 APE1 APE2

Ligated PCNA Unligated POLδ

IP (GFP) IP (IgG) Extract Conclusions – BER com plexes

1. N uclear BER proteins can be isolated as com plexes that carry out com plete BER

2. M itochondrialBER does not seem to require stable repair com plexes U NG knockout m ice

• D evelop norm ally and are fertile

• A ccum um ulate uracilin the genom e

• D efective in post-replicative uracilrem oval

• D evelop B-celllym phom as (30-fold↑) late in life (all N ilsen et al., 2000, 2003)

• A bnorm alsom atic hyperm utation (S H M ) and class sw itch recom bination (CS R) in B-cells (R ada et al., 2002)

• Increased postischem ic brain injury (Endres et al., 2004) U ng -/- M ICE W IT H M ACR OS COPIC LY M PH OM A

Diagnosis (n=22) No. Average age (m onths) ± S D Follicular hyperplasia 3 21.0 ± 8.5

Follicular lym phom a 6 19.5 ± 5.6

D iffuse large B-cell lym phom a 13 20.9 ± 3.5 (D LBCL) H IS T OPAT H OLOG Y OF H Y PER PLAS IA AND LY M PH OM A

HES T-cells (CD3) B-cells (CD45) a i 10x 10x 10x s a l p r e p y h r a l u c i l l o F

a 20x 20x 20x m o h p m y l r a l u c i l l o F

40x 40x 40x L C B L D Clonality analysis - exam ples

Fragment size (base pairs) y t i s n M onoclonal e t n i k a e P

Fragment size (base pairs) y t i

s Polyclonal n e t n i k a e P Clonality of hyperplasias and lym phom as

D iagnosis\Clonality M ono Bi Poly n.d.

H yperplasia 1 0 1 1 Follicular lym phom a 5 0 0 1 D iffuse large B-cell 6 2 0 5 lym phom a Im balanced leukocyte populations already in in very youngU ng-deficient m ice (10-12 w eeks)

Ung+/+ Ung-/-

20.22 % 14.93 % CD4 80 CD 19

70

60.91 % 66.44 % 60

50

s CD19 l l e

c 40

f CD 4 o

30 3.02 % 1.18 % % CD 8 20 NK-1.1 NK-1.1 10

0

CD4 CD8 CD19 NK-1.1 Ung+/+ 20.6 9.7 60.8 3.7 Ung-/- 15.6 8.6 65.7 2.2 AB NOR M AL CY T OKINE PR ODU CT ION IN U ng-DEFICIENT M ICE

IFNγ 100 300 3 90 x 0 Ung+/+ Ung+/+ e 1 250 80 Ung-/-

d Ung-/- x

n

l 70 i 200 m 60 n / o g 50 i

t 150 p

a

, 40 l γ γ γ γ 30 u

N 100 m F i

I 20 t

10 S 50 0 0 LPS ConA PHA PMA/ionom. LPS ConA PHA PMA/ionom IL-6 10 IL-2 9 Ung+/+ 700 3

0 8 Ung-/- Ung+/+ 1 600 7 Ung-/- x )

l l 6 500 m m / / 5 400 g U ( p 4

, 2 300 -

6 3 - L

I 200 L 2 I 1 100 0 0 LPS ConA PHA PMA/ionom. PMA/ionom. G eneration of antibody repertoire

Germ line Igh locusV D J

I S C Cδ I S Cγ3 I S Cγ1 µ

VD J Primary I S C Cδ I S Cγ3 I S Cγ1 IgM repertoire µ Class Switch Recombination Somatic Hyper Mutation VD *J** Secondary I S Sγ Cγ1 IgG, IgA, IgE repertoire µ (Slide prepared by Dr. Anne Durandy, Paris) Functions of AID and DNA repair proteins in S H M and CS R U NG 2 has a critical role in S H M and CS R

Apparently AID and UNG are more important than mismatch repair proteins.

UNG2 precedes functions of many other repair factors in SHM and CSR

In humans, CSR is more compromised than SHM when UNG is mutated

UNG2 is important for normal SHM and essential for CSR in humans Patients suffer from HyperIGM syndrome (Im ai et al., 2003, N ature Im m unology 4:1023-1028)

From N euberger et al., 2003 T rends Biochem . S ci. 28:305-312 H IG M patients – their U NG 2 proteins

MIGQKTLYSF FSPSPARKRH APSPEPAVQG TGVAGVPEES GDAAAIPAKK APAGQEEPGT 60 α1 α2 PPSSPLSAEQ LDRIQRNKAA ALLRLAARNV PVGFGESWKK HLSGEFGKPY FIKLMGFVAE 120 α3 β1 ERKHYTVYPP PHQVFTWTQM CDIKDVKVVI LGQDPYHGPN QAHGLCFSVQ RPVPPPPSLE 180 α4 α5 β2 α6 NIYKELSTDI EDFVHPGHGD LSGWAKQGVL LLNAVLTVRA HQANSHKERG WEQFTDAVVS 240 β3 α7 β4 α8 WLNQNSNGLV FLLWGSYAQK KGSAIDRKRH HVLQTAHPSP LSVYRGFFGC RHFSKTNELL 300 s QKSGKKPIDWKEL

frameshift, P1 α2 α7 β3 β1 β4 α4 stop at 141 and 224 β2 α6 s P2 s F251S α3 N α1 α5 P3 frameshift, c stop at 159 α8 UNG∆84 N o U N G 1 or U N G 2 activity is detectable in H IG M -syndrom e patients w ith m utated U N G -gene, but m inor enzym e activity (~ 0.4% ) is detectable in cells carrying the F251S m issense m utation. Problem : W hen expressed in E. coli, U N G F251S i fully active and stable. W hy is it absent in hum an cells? Enzym e activity of F251S 1 2 3 4 + 0-Ab + UNG-Ab + +

SMUG1-Ab + + w estern 600

C1 C2 P1 P2 P3 500

400 UNG2 M P

D 300 UNG1 200 100

0 nuclei m itochondria (cytosol) U NG m utations increase cellular uracil levels Com et assays on B-cells; tw o controls and 3 patients.

Patients with truncated UNG-proteins, as well as Phe251Ser have increased uracil in the

) 200 s t

i All growing cells from patients have n u comet tails, thus uracil content is y r 150 a r increased in all phases t Buffer i b r a UNG∆84 (

e 100 g a m a d 50 A N D

0 C1 C2 P1 P2 P3 C1 P3 141/224 F251S 159 stop stop Localisation of U NG 2 m utants from H IG M patient

Control Phe251S er

UNG2-EYFP UNG2 F251S-EYFP Two S-phase cells with Localisation varies - some cells have UNG2 in nuclear foci mutated protein in both nucleus and cytoplasm, others mainly in cytoplasm

Conclusion: W hen overexpressed as EY FP-tagged protein, Phe251S er is abnorm ally sorted. W hy ? N LS not affected U NG 2 m utants from H IG M patients co-expressed w ith w ild type U NG 1

Yellow and blue tag does not affect the sorting of the proteins; normal sorting of both. UNG2-EYFP UNG1-ECFP MERGED

When co-expressing Phe251Ser with UNG1, all mutant protein is found in the cytoplasm where it co-localises with UNG1. After longer incubations F251S disappears, but wild type is stable UNG2 F251S-EYFP UNG1-EC MERGED

Conclusions: The cellular sorting of the Phe251Ser mutant is clearly abnormal and disappears after longer incubations Hypothesis: UNG2 F251S forms dimers with UNG1 and transported to mitochondria, where it is degraded. The lack of nuclear UNG2 causes the hyperIGM syndrome in the patient. U NG 2 is ~1000-fold m ore efficient than S M U G 1 in rem oval of uracil from single stranded DNA

1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6

UNG2

SMUG1

U:A U:G Uss

A ID only deam inates cytosine in ssD N A , probably explaining the inability of S M U G 1 to com plem ent U N G 2 in U N G -deficient cells. In addition, S M U G 1 is poorly expressed in B-cells M utagenicity of A ID expressed in E. coli is abolished by U N G 2, but not by S M U G 1 w hich is also toxic in an ung - background

ung - ung +

pAID + + + + + + pAID-C87A + + + + pUNG2 + pSMUG1 + + + + pSMUG1-P245A + A ID is m utagenic, but not cytotoxic s

l 500 l e c

in U ng-deficient cells

e 400 l b a i 300 v A ID -C87A , an inactive m utant, has no effect 8 0

1 200

/

s l l 100 U N G 2 abolishes m utagenicity of A ID , e c

R but S M U G 1 does not f 0 i R e r 9 u

t 10 l A ID is not cytotoxic u c

l 8 m 10

/

s l l S urprisingly, S M U G 1 is cytotoxic in e 7 c 10

e l U ng-deficient, but not U ng-proficient cells b a i 6

V 10

αUNG2

αSMUG1 Contributors

T rondheim S outh M im m s, ICR F La J olla, S cripps H ans E. Krokan H ilde N ilsen J ohn A . T ainer G eir S lupphaug Clifford M ol M arit O tterlei Paris S udip Parikh T orkild V isnes A nne D urandy BodilKavli Kohsuke Im ai J avier Peña-D iaz A lain Fischer Em adoldin Feyzi M ansour A kbari M ainz Cathrine B. V ågbø Bernd Epe S onja A ndersen T ina H eine Per A rne A as H ilde N ilsen Lars H agen