Science - USA - 03.12.2021

a low concentration (0.005%) of MMS, they
were sensitive to higher concentrations (≥0.01%)
of MMS (fig. S2). We observed thatpol3ITD
cells in a WT RAD27 background were also
sensitive to high concentrations of MMS (fig.
S2). This at least partially explains why the
pol3ITD could not suppress MMS-induced
lethality ofrad27Dcells at high MMS concen-
trations. In addition,pol3ITD did not rescue
the synthetic lethality that occurred in the con-
text ofrad27Dcoupled with deficiency of the
5 ′nucleasesEXO1orDNA2nuclease-helicase
(tables S2 and S3 and supplementary text S2).
Two types of duplications were present in
the revertants:pol3 591 – 598ITD(fig.S3,Aand
B), a previously reported classic duplication re-
sulting from the realignment and ligation of
unprocessed 5′flaps ( 12 ), andpol3 458 – 477 ITD,
which contained a 55–base pair (bp) duplica-
tion with a 5-bp spacer between the duplicated
units (fig. S3, C and D). We named the dup-

lication with an intervening spacer an“alter-

native duplication.”Bothpol3 591 – 598 ITD and

pol3 458 – 477 ITD resembled ITDs detected in

human cancer ( 13 – 15 ). To determine how the

alternative duplicationpol3 458 – 477 ITD was

created, we conducted WGS of WT andrad27D

cells grown at 37° or 30°C for 4 hours. The

mutation frequency of WT cells was the same

at both temperatures (Fig. 2A). By contrast,

the mutation frequency ofrad27Dcells in-

creased by twofold at restrictive temperature;

in particular, the frequency of duplications

and base substitutions was increased (Fig. 2A).

In addition, duplication insertions inrad27D

cells grown at 37°C were considerably longer

than those inrad27Dcells grown at 30°C (Fig.

2B). We found thatrad27Dcells grown at 37°C

exhibited alternative duplications similar

to thepol3 458 – 477 ITD. The alternative dup-

lications were not detected in WT cells (30°

or 37°C) or inrad27Dcells grown at 30°C

(Fig. 2C), suggesting that alternative dupli-

cations were induced by restrictive tempera-

ture stress.

We further noted that the sequences of

these alternative duplications suggested the

formation of three different types of hairpin

structures (Fig. 2D; fig. S4, A to D; and sup-

plementary text S3). This supported a model

of sequential actions, including the conver-

sion of a 5′Okazaki fragment flap to a 3′flap,

annealing of the flap to a complementary se-

quence, extension of the 3′flap, realignment,

and ligation of the extended 3′flap to produce

an alternative duplication, likepol3 458 – 477

ITD. Consistent with this model, our WGS data

indicated that 40% of the alternative duplica-

tions also carried base substitutions at the dup-

lication unit. These substitutions most likely

resulted from failure to remove Pola-generated

errors on the 5′flap. To determine if the re-

strictive temperature induced 3′flap formation

SCIENCEscience.org 3 DECEMBER 2021•VOL 374 ISSUE 6572 1253

Fig. 1.pol3ITD and

missense mutations drive

resistance torad27D-

induced conditional

lethality.(A) Spot assays

of WT,rad27D, orrad27D

revertant (Rev) yeast cells

grown at 30°C (optimal tem-

perature), 25°C (suboptimal

temperature), or 37°C

(restrictive temperature) for

48 hours.rad27D::URA3and

rad27D::LEU2represent the

rad27Dallele with a linked

URA3orLEU2selection

marker gene, respectively.

(B)pol3mutations detected

in independent revertant

strains (n= 31). Circles and

diamonds represent base

substitution and ITD muta-

tions, respectively. The

domain structures were

defined as previously

described ( 23 ). aa, amino

acid; H495Q, His^495 →Gln;

P965T, Pro^965 →Thr. (C) Spot

assays of WT,rad27D, or

rad27Dyeast cells with indi-

catedpol3knock-in mutations

grown at 30°, 25°, or 37°C

for 48 hours.POL3::HIS3rep-

resents thePOL3(WT or

mutant) alleles with a linked

HIS3selection marker gene.

A

WT

rad27Δ::URA3

rad27Δ::LEU2#1

rad27Δ::LEU2#2

Rev

30°C 25°C 37°C

B

(458-477) ITD

(591-598) ITD

R470GR475IA484V S847Y

NT-1 NT-2 Exo I Exo II

Pol IV

Exo III Pol II

Pol VI

Pol III Pol I Pol V

C T- 1

CT-2 CT-3 ZnF1 ZnF2

0

5

20

# of

pol3

mutations

P965T

0 100 200 300 400 500 600 700 800 900 1000 1097 aa

H495Q

C

POL3::HIS3

WT

rad27Δ::URA3

rad27Δ::LEU2

WT

ITD

R475I

S847Y

rad27Δ::LEU2

POL3::HIS3

30°C 25°C 37°C

RESEARCH | REPORTS