Page 100 - Read Online

Page 8 of 19 Ambros et al. Microbiome Res Rep 2023;2:34 https://dx.doi.org/10.20517/mrr.2023.18

Of the 50 intact predicted prophages, 33 prophages harboured no tRNA genes, 13 included one tRNA gene,
and 4 included two tRNA genes. Predicted tRNAs extracted from the NCBI annotation most often coded
for isoleucine (9x) and phenylalanine (7x), but tRNAs for histidine (1x), and methionine (1x), as well as
tRNAs with non-assignable isotype (3x) were also predicted.

The isotype prediction of the formerly undefined tRNA genes failed because of an erroneous secondary
structure prediction (clover leaf form). The gene formerly classified as methionine tRNA was predicted as
an isoleucine tRNA with CAU as anticodon (Ile2) by tRNAscan-SE 2.0. Including the formerly methionine
tRNA classified tRNA, seven of the now ten predicted isoleucine tRNA genes were found to use CAU as an
anticodon, one used AAU, one UAU, and one GAU.

Comparing the codon usage of the eight prophages harbouring isoleucine tRNAs with the codon usage of
their host (including the prophage) revealed that, of the three codons for isoleucine (ATA, ATC, ATT), the
eight phages used ATA (15.68 +/- 2.51)% (mean +/- standard deviation) the least, followed by ATC
(24.58 +/- 2.35)%, and ATT (59.74 +/- 3.19)%. The eight host genomes used ATA (4.31 +/- 0.62)% for
isoleucine coding the least as well, followed by ATC (31.08 +/- 1.47)%, and ATT (64.61 +/- 0.92)%.
Compared to a phage harbouring no tRNA genes, e.g., phage TMW 1.591 P1, a similar distribution was
observable (phage TMW 1.591 P1: ATA 11.54%, ATC 30.34%, ATT 58.12%; host strain L. curvatus TMW
1.591: ATA 3.65%, ATC 31.99%, ATT 64.36%). The analysed phages seem to use the ATA codon more
frequently than their host in relation to the other isoleucine codons.

In contrast to this, the differences between host and phage in the usage of the two phenylalanine codons
TTC and TTT were less pronounced. The phenylalanine tRNA harbouring prophages preferred TTT
(63.65 +/- 3.41)% over TTC (36.35 +/- 3.41)%, as did the hosts at a similar rate [TTC was used by their hosts
(41.41 +/- 0.51)% of the time, TTT (58.59 +/- 0.51)%]. Although TTC was used slightly less frequently by the
phages relative to both phenylalanine codons, phenylalanine tRNA genes harboured by intact prophages
used GAA exclusively as anticodon.

Five putatively intact prophages contained transposases. Four of them contained transposases, which belong
to the IS3 and IS30 families. Both putatively intact prophages in L. curvatus strain ELA214388 shared IS3
family transposases: ELA214388 P1 in the lysogeny gene module and ELA214388 P2 in the replication gene
module. ELA214388 P2 harboured an additional IS30 family transposase in its tail gene module. Phage
TMW 1.1928 P1 contained an IS3-like element (IS1520 family transposase) in the head gene module and
phage KG6 P1 contained an IS30 family transposase in its tail genes module, which may “cut off” the lysis
genes, as a potential attR-site was found between transposase and lysis genes. Phage TMW 1.624 P1
contained four transposases lacking family affiliation according to annotation; three transposases were
located in the replication gene module, and one in the tail gene module. After sequencing the post-induced
strain lysates to verify successful prophage induction, the genome of phage TMW 1.1365 P2 also showed a
gene with a putative transposase activity. This region was located at the ends of a contig within the NGS-
sequenced bacterial genome, and was previously not allocated to the prophage.

Notably, in partially sequenced genomes with an assembly status of “contig” or “scaffold”, some putatively
intact prophages also lacked features located at their boundaries, such as an att-site, or lysogeny or lysis-
related genes, due to incomplete sequencing. For example, phage TMW 1.1447 P1 lacked most of its
lysogeny gene cluster; phages TMW 1.706 P1 and TMW 1.706 P2 both contained non-sequenced parts in
their replication genes; phage NRIC0822 P1 was harboured by two small contigs with most likely missing
sequence information, as well as a non-determinable attR-site; phage TMW 1.1365 P2 had a non-