The potential for transcription to impair DNA replication
Transcription does not stop during DNA replication and
transcription-replication conflicts must arise since both processes
coexist on chromosomes. Furthermore, the transcriptional machinery
includes many DNA-bound protein complexes that might impair the movement
of the replisomes even if active RNA polymerases are removed.
Whereas bacterial genes tend to be ordered co-directionally with DNA
replication (Kunst et al., 1997; McLean et al., 1998; Rocha and Danchin,
2003), in S. cerevisiae almost half of the genes are transcribed
opposite to the normal direction of DNA replication (referred to as
”head-on” genes), so encounters between the replisome and RNA
polymerases must be frequent (Garcia-Muse and Aguilera, 2016; Goehring
et al., 2023). Also unlike bacteria, there is little evidence that
head-on orientated genes in yeast have higher mutation rates so the
eukaryotic replisome must be adept at traversing transcriptional units
(Sankar et al., 2016). In theory, DNA damage systems could alleviate
transcription-replication collisions in head-on genes, but the DNA
damage response is global and delays replication throughout the genome
when forks stall, which must represent an emergency response rather than
the default behaviour (reviewed in (McClure et al., 2022)). Similarly,
while genome-wide profiling of DNA replication has provided evidence
that the replisome pauses at highly expressed loci, the delays required
for the replisome to wait until transcription completes at each head-on
gene would be substantial (often >1 minute given average
gene size and RNA pol II holoenzyme speed (Muniz et al., 2021)), whereas
minimal pausing was detected (Azvolinsky et al., 2009; Claussin et al.,
2022; Kara et al., 2021). Replication profiles show that in S.
cerevisiae many replisomes travel 50 kb through gene rich regions in a
normal S-phase lasting ~ 30 minutes (Kara et al., 2021),
which is at the limit of travel given the measured yeast fork speed of
1.6-1.9 kb/min (Hodgson et al., 2007; Sekedat et al., 2010), so pauses
must be short and infrequent. Therefore, the eukaryotic replisome must
efficiently displace RNA polymerases, just as the bacterial replisome
has been directly observed to do by electron microscopy (French, 1992).
Although we might imagine that head-on collisions between replication
and transcription are inherently recombinogenic, extensive studies of
the budding yeast ribosomal DNA show exactly the opposite. The
unidirectional replication fork barrier (RFB) formed by Fob1 binds the
ribosomal DNA downstream of the massively transcribed 35S region and
prevents the replisome meeting oncoming RNA polymerase I head-on
(Kobayashi and Horiuchi, 1996). However, replication forks stalled at
the RFB are highly recombinogenic (Kobayashi et al., 1998; Stewart and
Roeder, 1989), whereas in RFB-defective mutants replication forks
proceed seemingly unimpeded head-on through the 35S region while
recombination rate is vastly decreased and DNA damage marks are
unchanged (Kara et al., 2021; Keszthelyi et al., 2023). This is also
true for RNA polymerase II as reversion assays using a head-on or
codirectional lys2 frameshift allele failed to demonstrate any
significant difference in the overall rate of reversion between both
orientations (Kim et al., 2007). There are exceptions however, since
tRNA genes transcribed by RNA polymerase III do impede the replisome in
head-on encounters, as detected by both 2D-gels and genome-wide methods
and these encounters can be recombinogenic (Azvolinsky et al., 2009;
Claussin et al., 2022; Deshpande and Newlon, 1996; Kara et al., 2021;
Tran et al., 2017).
Regulatory proteins bound to DNA could also form obstacles for the
replisome. The substantial machinery bound at promoters includes
transcription factors and transcriptional activators (catalogued in
(Rossi et al., 2021)), but as noted above genome-wide studies have
revealed only the mildest impacts of these on replisome progression
(Azvolinsky et al., 2009; Kara et al., 2021).
Helicases including Rrm3 and Pif1
work assist the replisome in removing proteins, easing the path through
difficult features including tRNAs (Azvolinsky et al., 2006; Azvolinsky
et al., 2009; Claussin et al., 2022; Ivessa et al., 2003; Tran et al.,
2017), and it has recently been found that replisome pausing at tRNA is
primarily due to bound TFIIIC complex rather than transcriptionper se (Yeung and Smith, 2020). However, the mild replisome
pausing at highly expressed RNA polymerase II genes is independent of
Rrm3 (Azvolinsky et al., 2009), suggesting that the core replisome is
able to remove protein obstacles associated with RNA polymerase II
activity.
RNA polymerase II and associated factors therefore have the potential to
impede DNA replication, and there is some evidence for this in extremely
highly transcribed genes, but during rapid growth under nutrient rich
conditions the replisome is very proficient in removing such obstacles,
despite transcription being highest under such conditions. RNA
polymerase III genes do have an effect on the replisome, but S.
cerevisiae has dedicated helicases to ensure that replisome pausing at
tRNA is minimal.