ArticlePDF Available

Abstract and Figures

Purple photosynthetic bacteria (PPB) are versatile microorganisms capable of producing various value-added chemicals, e.g., biopolymers and biofuels. They employ diverse metabolic pathways, allowing them to adapt to various growth conditions and even extreme environments. Thus, they are ideal organisms for the Next Generation Industrial Biotechnology concept of reducing the risk of contamination by using naturally robust extremophiles. Unfortunately, the potential of PPB for use in biotechnology is hampered by missing knowledge on regulations of their metabolism. Although Rhodospirillum rubrum represents a model purple bacterium studied for polyhydroxyalkanoate and hydrogen production, light/chemical energy conversion, and nitrogen fixation, little is known regarding the regulation of its metabolism at the transcriptomic level. Using RNA sequencing, we compared gene expression during the cultivation utilizing fructose and acetate as substrates in case of the wild-type strain R. rubrum DSM 467T and its knock-out mutant strain that is missing two polyhydroxyalkanoate synthases PhaC1 and PhaC2. During this first genome-wide expression study of R. rubrum, we were able to characterize cultivation-driven transcriptomic changes and to annotate non-coding elements as small RNAs.
Content may be subject to copyright.
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
Available online 20 June 2024
2001-0370/© 2024 The Authors. Published by Elsevier B.V. on behalf of Research Network of Computational and Structural Biotechnology. This is an open access
article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Cultivation driven transcriptomic changes in the wild-type and mutant
strains of Rhodospirillum rubrum
Katerina Jureckova
a
,
1
, Marketa Nykrynova
a
,
1
, Eva Slaninova
b
, Hugo Fleuriot-Blitman
c
,
V´
eronique Amstutz
c
, Kristyna Hermankova
a
, Matej Bezdicek
d
,
e
, Katerina Mrazova
b
,
f
,
Kamila Hrubanova
f
, Manfred Zinn
c
, Stanislav Obruca
b
, Karel Sedlar
a
,
*
a
Department of Biomedical Engineering, Faculty of Electrical Engineering and Communication, Brno University of Technology, Brno, Czech Republic
b
Department of Food Chemistry and Biotechnology, Faculty of Chemistry, Brno University of Technology, Brno, Czech Republic
c
Institute of Life Technologies, University of Applied Sciences and Arts Western Switzerland Valais-Wallis (HES-SO Valais-Wallis), Sion, Switzerland
d
Department of Internal Medicine Haematology and Oncology, University Hospital Brno, Brno, Czech Republic
e
Department of Internal Medicine Haematology and Oncology, Faculty of Medicine, Masaryk University, Brno, Czech Republic
f
Institute of Scientic Instruments of the Czech Academy of Sciences, v.v.i., Brno, Czech Republic
ARTICLE INFO
Keywords:
Rhodospirillum rubrum
RNA-Seq
Transcriptome
Genome
Depolymerase knock-out
Gene ontology
Metabolism
Fructose
Acetate
Polyhydroxyalkanoates
ABSTRACT
Purple photosynthetic bacteria (PPB) are versatile microorganisms capable of producing various value-added
chemicals, e.g., biopolymers and biofuels. They employ diverse metabolic pathways, allowing them to adapt
to various growth conditions and even extreme environments. Thus, they are ideal organisms for the Next
Generation Industrial Biotechnology concept of reducing the risk of contamination by using naturally robust
extremophiles. Unfortunately, the potential of PPB for use in biotechnology is hampered by missing knowledge
on regulations of their metabolism. Although Rhodospirillum rubrum represents a model purple bacterium studied
for polyhydroxyalkanoate and hydrogen production, light/chemical energy conversion, and nitrogen xation,
little is known regarding the regulation of its metabolism at the transcriptomic level. Using RNA sequencing, we
compared gene expression during the cultivation utilizing fructose and acetate as substrates in case of the wild-
type strain R. rubrum DSM 467
T
and its knock-out mutant strain that is missing two polyhydroxyalkanoate
synthases PhaC1 and PhaC2. During this rst genome-wide expression study of R. rubrum, we were able to
characterize cultivation-driven transcriptomic changes and to annotate non-coding elements as small RNAs.
1. Introduction
Bacteria represent a remarkably diverse group of organisms, which is
not surprising, as, according to estimations, there are 10
6
to 10
8
separate
prokaryotic genospecies on Earth [1]. While the majority of these mi-
croorganisms living in a wide range of natural environments seem to be
uncultivable with current techniques, some organisms are very versatile
and can prosper under various cultivation conditions and grown on
various substrates. The ideal example is Rhodospirillum rubrum, a purple,
non-sulfur, Gram-negative facultative anaerobe from the class of
Alphaproteobacteria, which was observed for the rst time by Esmarch in
1887 [2]. It was shown to grow both aerobically and anaerobically. The
absence of oxygen triggering the photosynthesis apparatus for synthesis
of membrane proteins, bacteriochlorophylls, and the carotenoids
turning the culture purple. Its metabolic versatility is further supported
by the fact that it prospers under both heterotrophic and autotrophic
conditions. Besides utilizing various organic substrates as saccharides or
organic ions, R. rubrum can x and metabolize inorganic compounds
such as CO and CO
2
and is therefore considered a prospective strain for
valorization of waste gases such as combustion products or syngas [3,4].
On the other hand, it is capable of producing other gaseous products
utilizable as fuels, particularly hydrogen [5]. Last but not least,
R. rubrum is a potent producer of biopolymers of the class of poly-
hydroxyalkanoates (PHAs) in the form of intracellular granules from
many carbon sources. When supplemented by butyrate, for instance, it
can accumulate up to 50 w/w % of dry weight as PHA [6]. Hence, it
hosts a large variety of metabolic pathways that can be leveraged to
produce sustainable carbon substrates.
* Corresponding author.
E-mail address: sedlar@vut.cz (K. Sedlar).
1
Contributed equally
Contents lists available at ScienceDirect
Computational and Structural Biotechnology Journal
journal homepage: www.elsevier.com/locate/csbj
https://doi.org/10.1016/j.csbj.2024.06.023
Received 15 March 2024; Received in revised form 11 June 2024; Accepted 18 June 2024
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
2682
R. rubrum is considered to be a model strain for studying the con-
version of light energy to chemical energy [7], hydrogen biosynthesis
[8], formation of photosynthetic membranes (PM) [9], and regulatory
pathways of the nitrogen xation system [10,11]. Therefore, it is not
surprising that there are currently 10 genome assemblies of several
R. rubrum in the GenBank database (accessed March 8th, 2024),
including the genome presented in this study. The genes involved in
various metabolic pathways are therefore known. Apart from synthetic
pathways mentioned above, PHA synthesis occurs, and the three genes
coding PHA synthases were identied in R. rubrum genome [12]. While
one of these genes is as part of biosynthetic phaCAB operon, the other
two are located separately in different locations in the genome. The
main challenges related to R. rubrum and its capacity to produce
value-added chemicals are a low specic growth rate, a low biomass and
PHA volumetric productivity, and the need for an organic co-substrate to
increase productivity of the autotrophic pathways in the case of PHAs
synthesis. These hurdles could be overcome by adopting genome engi-
neering strategies, such as the one already applied to R. rubrum, for
example, consisting of overexpressing genes coding PHA synthases [13].
Despite relatively good genome characterization and even avail-
ability of various mutant strains, only a little is known on gene regula-
tions in R. rubrum as studies exploring gene expression on a genome-
wide scale using RNA sequencing (RNA-Seq) are missing and only two
studies based on microarrays are available [14,15]. Thus, in our study,
we compared R. rubrum transcriptomes from cultivation on fructose and
on acetate to describe basic changes in various pathways observed along
the cultivation or between substrates. Moreover, to explore the stability
of expression in engineered strains, we also analyzed transcriptomes of a
mutant strain with two PHA synthases deleted and compared it to the
wild-type strain for both same substrates. Furthermore, as the genome
assembly of the type strain was created relatively long time ago, we
re-sequenced the genome of R. rubrum DSM 467
T
, which was used for
experiments to exert inuence of possible mutations that might be
accumulated over time. We also sequenced a ΔphaC1ΔphaC2 mutant
strain to conrm deletions of PHA synthases and to capture other
genomic changes. Additionally, we improved genome annotation by
small RNA (sRNA) gene inference using RNA-Seq data.
2. Material and methods
2.1. Growth conditions and experiments
A freeze-dried bacterial culture of the type strain Rhodospirillum
rubrum DSM 467
T
(WT strain) was purchased from the Leibnitz Institute
DSMZ-German Collection of Microorganism and Cell Cultures,
Braunschweig, Germany. The double mutant R. rubrum ΔphaC1ΔphaC2
knock-out (KO) strain was obtained from the team of Professor Kevin E.
O΄Connor and Professor Tanja Narancic (University College Dublin,
Ireland). This PHA-negative mutant was designed and constructed as
reported previously in [16].
The rst cultivation step was incubation of both wild-type and
mutant strains on LB broth (tryptone 10.0 g/L, yeast extract 5.0 g/L,
NaCl 5.0 g/L) in Petri dishes at 30 C in the dark for ve days. The
second part of cultivation was inoculated by loop of the bacterial culture
from Petri dishes and the cultivation was performed in 500 mL Erlen-
meyer asks containing 100 mL of LB broth at 30 C in the dark under
shaking at 160 rpm for approximately 72 h till OD
660
=1.5. During the
main part of the cultivation, the cultures were inoculated to OD
660
=0.1
by culture grown in liquid LB broth and cultivated in SYN medium. Its
composition for 1 L of medium was: 250 mg of MgSO
4
7 H
2
O, 132 mg of
CaCl
2
2 H
2
O, 10 g of NH
4
Cl, 21 g of MOPS buffer, 10 mL NiCl
2
(20 µM),
100 mL of a chelated iron-molybdenum solution (0.28 g H
3
BO
3
, 2.1 g of
Na
2
EDTA, 0.7 g of FeSO
4
7 H
2
O and 0.1 g of Na
2
MoO
4
per Liter of
distilled water). In this medium, two carbon sources were used: 2.4 mL
of 1.5 M fructose (18 mM in total) with 4 mL of 50 g/L yeast extract (1 g/
L in total) and 11 mL of 1 M acetate (55 mM in total) with 4 mL of 50 g/L
yeast extract (1 g/L in total). They will be referred to as FY and AY,
respectively. Cultivation was performed in triplicates in 200 mL of
medium in 1 L Erlenmeyer ask at 30 C in the dark under 160 rpm until
the last sample of culture in stationary phase was taken. All the culti-
vations were performed under aerobic conditions.
During the cultivation of wild-type R. rubrum (WT) and R. rubrum
knock-out (KO), samples were taken at three time-points, at different
growth phases. In each sampling time-point, the biological triplicates
were spectrophotometrically screened at λ =660 nm and used for
further DNA and RNA analysis. The culture-specic growth rates and
doubling time were calculated using optical density data. The OD660
measurements for calculating
μ
max
were taken every hour. The
μ
max
values were calculated using a standard method: OD values were
transformed into ln(OD660), and the linear part of the ln(OD660) vs.
time curve was used for regression analysis to determine
μ
max
.
2.2. Transmission electron microscopy
Cultures of wild-type R. rubrum cultivated on both fructose and ac-
etate as carbon sources were xed using the high-pressure freezing
method. Samples were pipetted on the 200 µm side of 3 mm copper-gold
carrier type A, which was closed using the at side of the type B carrier.
Both carriers were pretreated with 1 % solution of lecithin in chloro-
form. Vitrication of the samples was performed using a high-pressure
freezer EM ICE (Leica Microsystems, Austria). Frozen samples were
transferred under liquid nitrogen into a freeze substitution unit (EM
AFS2, Leica Microsystems, Austria). The substitution solution contained
1.5 % OsO
4
in acetone and the protocol was set as previously described
in [17]. Freeze substitution was followed by resin embedding (Epoxy
Embedding Medium kit, Sigma Aldrich, Germany) and curing at 62 C
for 48 h. Cured samples were cut to ultrathin sections (~75 nm) using a
diamond knife (ultra 45, DiATOME, Switzerland) and ultramicrotome
(EM UC7, Leica Microsystems, Vienna, Austria). Since ultrathin sections
were imaged using a low-voltage transmission electron microscope
(LVEM 25 Delong Instruments, Czech Republic) at 25 kV voltage of
electron beam, no post-staining procedure was necessary to achieve
sufcient contrast [18].
2.3. DNA and RNA extraction and sequencing
High molecular weight genomic DNA of the WT strain for long-read
sequencing was extracted using a MagAttract HMW DNA kit (Qiagen,
NL) in accordance with the manufacturers instructions. The DNA purity
was checked using NanoDrop (Thermo Fisher Scientic, USA), while the
concentration was measured using Qubit 4.0 Fluorometer (Thermo
Fisher Scientic, USA), and the length was measured using Agilent 4200
TapeStation (Agilent Technologies, USA). The Ligation Sequencing 1D
Kit (Oxford Nanopore Technologies, UK) was used for library prepara-
tion, and the sequencing was performed using R9.4.1 Flow Cell on the
Oxford Nanopore Technologies (ONT) MinION platform.
Genomic DNA of WT and KO strains for the high-throughput short-
read sequencing was extracted using GenElute Bacterial Genomic DNA
Kit (Sigma-Aldrich, USA) in accordance with the manufacturers in-
structions. Sequencing libraries were prepared using the KAPA Hyper-
Plus kit, and sequencing was carried out using MiSeq Reagent kit v2
(500 cycles) on the MiSeq platform (Illumina, USA).
RNA isolation followed an optimized extraction protocol consisting
of a combination of procedures focused on RNA isolation, where the
crucial point was addition of 1 mL of TRIzol per 40 mg of wet biomass
followed by incubation for 5 min to permit complete dissociation of the
nucleoproteins complex. Then, 0.1 mL of 1-bromo-3-chloropropane per
1 mL of TRIzolReagent used for cell lysis was added, and the tube was
securely capped and incubated for 23 min at room temperature. Af-
terwards, the samples were centrifuged at 11,000 ×g at 4 C for 15 min.
Subsequently, the supernatant containing the RNA was transferred to a
new tube where 70 v/v % EtOH was added at a ratio 1:1. Then, the
K. Jureckova et al.
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
2683
samples were transferred to spin columns and the procedure continued
according to the manual of the NucleoSpin RNA Plus isolation kit
(Macherey-Nagel) with washing and drying steps of silica membrane
and elution of RNA that were stored at 80 C till the sequencing.
Ribodepletion was performed with QIAseq FastSelect 5S/16S/23S Kit
(Qiagen, NL) for WT samples cultivated on fructose (no. 1 9, see
Supplementary Table S1) or with RiboCop rRNA Depletion Kit for Bac-
teria Mixed bacterial samples (Lexogen, AT) for remaining samples (no.
10 33). Strand-specic sequencing libraries were prepared with
NEBNext Ultra II Directional RNA Library Prep Kit (New England Bio-
labs, USA) and sequenced with Illumina NextSeq550 to produce
reversely stranded reads. For samples no. 10 33, Unique Molecular
Identiers (UMIs) were added using xGen Duplex Seq Adapters (IDT,
USA).
2.4. Genome assembly
Nanopore sequencing data of wild-type strain were basecalled using
Guppy (v3.4.4), and the data quality was checked using PycoQC (v2.2.3,
[19]). De novo assembly by Flye (v2.8.1, [20]) was performed, and the
obtained contigs were polished using Minimap2 (v2.17, [21]) combined
with Racon (v1.4.13, [22]) and then nal polishing was performed by
Medaka (v1.1.2).
Both WT and KO strains were sequenced using Illumina MiSeq. Ob-
tained reads were rstly checked for quality using FastQC (v0.11.5)
combined with MultiQC (v1.7, [23]) and secondly, the low-quality ends
of reads together with sequencing adapters were removed by Trimmo-
matic (v0.36, [24]) with following settings: ILLUMINACLIP:TruSeq3-PE.
fa:2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MIN-
LEN:36. Next, the set of reads was checked for contamination by human
DNA, and detected reads were removed using BBMap (v39.01) and
human genome GCF_000001405.40 available in the NCBI RefSeq data-
base. In the case of the WT strain, the reads were mapped to the nano-
pore contigs using BWA (v0.7.17, [25]), and for managing the obtained
assembly Samtools (v1.10, [26]) was employed. Final assembly polish-
ing was made by Pilon (v1.24, [27]). In the last step, the assembled
genome and plasmid were rearranged so that the DnaA gene was the rst
gene in the genome and repB was the rst gene in the plasmid. For the
KO strain assembly, BWA and Samtools were used again, but in this case,
the pre-processed reads were mapped to the previously assembled WT
genome.
The variant calling was conducted to nd differences between both
(WT and KO) analyzed strains. For this purpose, GATK4 (v4.3) was used.
Underrepresented variants, variants with low coverage and false posi-
tive calls were ltered out, and the remaining ones underwent addi-
tional analysis to ascertain their presence within coding regions and, if
applicable, their impact on the phenotype.
2.5. Genome annotation
NCBI Prokaryotic Genome Annotation Pipeline (PGAP) [28] was
used for the wild-type strain chromosome and plasmid annotation.
Protein coding genes functional annotation was performed by classi-
fying them into clusters of orthologous groups (COG) categories from
the eggNOG database via the eggNOGmapper (v2.1.6, [29]). The
DNAplotter [30], which is integrated into Artemis (v18.2.0, [31]), was
used to create the chromosome and plasmid circular maps including GC
content and CG skew plots, which were calculated in a window of size
10,000 bp with 200 bp step. The interspaced short palindromic repeats
(CRISPR) arrays were searched by the CRISPDetect tool (v2.4, [32]), and
Cas genes were manually searched in the annotated genome. Physpy
(v4.2.6, [33]) and online tools Prophage Hunter and Phaster were used
for prophage DNA identication. The restriction-modication systems
were located using REBASE database [34].
The GenBank database was searched for genomes of R. rubrum using
Entrez [35]. Then Roary (v3.13, [36]) was used to identify the core
genome, which consists of genes located in all analyzed strains, the
accessory genome formed by genes presented in at least two strains but
not in all of them, and the number of unique genes. The minimum
percentage identity to assess whether two genes are similar was 95 %,
and all other parameters for Roary were left at their default settings.
2.6. Transcriptomic analysis
The raw RNA-Seq reads for both strains were checked for their
quality using FastQC and MultiQC. Next, the reads trimming was per-
formed to discard low-quality bases and adapters using Trimmomatic
with following parameters for samples no. 10 33: ILLUMINACLIP:
TruSeq3-SE.fa:2:30:10 LEADING:3 TRAILING:3 SLI-
DINGWINDOW:4:15 MINLEN:36 and for samples no. 1 9 HEADCROP
parameter was added with value 5, because rst ve bases of the reads
contained randomized 5 bp long adapters. Remaining contamination by
rRNA was removed using SortMeRNA (v4.3.4, [37]) together with the
default database (smr_v4.3_default_db.fasta) and the SILVA database
[38] with known 16S and 23S rRNA sequences. Processed reads were
mapped to the genome of the wild-type strain using STAR (v2.7.10a,
[39]), and mapped reads were deduplicated using UMI-tools (v1.1.4,
[40]). Finally, the reads were counted using featureCounts function [41]
from Rsubread package (R/Bioconductor). Reads counting considered
two options: Uniquely mapped reads and multimapping reads. For the
multimapping reads the contribution of those reads to the nal count
was always divided by the number of genomic loci to which the read was
mapped, therefore the number of reads remained unchanged.
Created count tables for all samples were further normalized by
calculating RPKM (Reads Per Kilobase per Million mapped reads) and by
using built-in function in R/Bioconductor package DESeq2 [42], which
was also used for differential expression analysis. Normalized count
tables were used for dimension reduction using Barnes-Hut t-SNE
implemented in R package Rtsne [43]. The results were visualized using
ggplot2 [44] R package, which was also employed for creation of Vol-
cano plots. Gene ontology (GO) enrichment analysis was performed by
topGO [45] R/Bioconductor package together with GO map, that was
created with custom Python script based on available GO annotation of
wild-type strain in NCBI RefSeq database (NZ_CP077803.1 for chro-
mosome and NZ_CP077804.1 for plasmid). Finally, the expression pro-
les of selected genes were visualized using heatmaps created using R
packages gplots, RColorBrewer, magick and openxlsx.
To gain insight into the regulatory processes of R. rubrum, mapped
reads were further used for non-coding RNAs (ncRNA) prediction with
baerhunter (v0.9, [46]) The sample depths were normalized using
sizeFactors function from DESeq2 package and visualized using box-
plots. Subsequently, to obtain a count table of a collapsed annotation le
with newly predicted features, barhunters count_features script using
featureCounts function was applied. Furthermore, attention was paid to
putative sRNAs, and their length distribution was visualized by a his-
togram. Also, these predictions were categorized as trans, respectively
cis-acting elements using IRanges R package [47]. Finally, differential
expression analysis was performed on the count table using DESeq2.
Counts of the differentially expressed (up or down) sRNAs between
specic conditions were visualized by a bar chart using R package
ggplot2 as well as for all mentioned graphs.
3. Results
3.1. Genome assembly
The Oxford Nanopore Technologies MinION produced 135,804
reads; from them, 104,955 had Q >7 and were used for further WT
strain assembly. The mean reads length was approximately 3.5 kbp. The
Illumina MiSeq provided 2.5 million 250 bp-long paired reads with an
average Phred score of 35. From them, 296 reads were mapped to the
human genome; thus, they were discarded from further analysis. The
K. Jureckova et al.
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
2684
assembly process of wild-type strain resulted in the nal assembly of one
circular chromosome and one circular plasmid with a coverage of 370×.
The sequences were deposited under accession numbers CP077803.1 for
chromosome and CP077804.1 for plasmid at DDBJ/EMBL/GenBank.
The obtained chromosome sequence length was 4,352,570 bp with a
GC content of about 65.4 %, and the plasmid was 53,835 bp long with a
GC content of around 59.8 %. In total, 3,968 open reading frames (ORFs)
divided into 2,146 operons were identied for both sequences. Most of
the genes were protein coding; however, 49 pseudogenes were also
found. Furthermore, 521 genes overlapped with another neighboring
gene, and 3 overlaps were found between a gene and a pseudogene. The
overlap size was in 390 cases equal to three nucleotides and the longest
overlap was 87 bp between genes KUL73_19715 and KUL73_19720 on
plasmid. See Table 1 for complete statistics for chromosome and
plasmid.
Illumina MiSeq sequencing of KO strain provided about 2.2 million
250-bp long reads with an average Phred score of 33. Of these, 260 reads
that mapped to the human genome were removed. Variant calling of the
KO strain conrmed deletion of PHA polymerases. Moreover, it revealed
three changes in its genome: two single nucleotide mutations and one
insertion. The rst mutation was found in position 823,861 (A G) in
the promoter of cimA gene encoding citramalate synthase, the second
mutation was identied in position 1,305,962 (C T) in the rpoH gene
encoding RHA polymerase sigma factor RpoH and the last third change
was localized in position 3,782,354 (C CGCTTCAGGGGAAA-
CACGTTATGAAG) in the promoter of gene encoding hypothetical
protein.
Ten genomes obtained from GenBank were used for R. rubrum core
genome determination. In total, 1,732 genes were found in all analyzed
strains and thus formed the core genome. Another 2,859 genes were
identied in at least two genomes, and thus, they comprised accessory
genome. Together these 4,591 genes form the R. rubrum pangenome. In
addition, 920 unique genes only identied in one genome were found.
The number of unique genes ranged from 4 to 467, with a median of 44
(see Supplementary Table S2).
3.2. Wild-type strain functional annotation
Protein coding genes and pseudogenes were classied according to
COG into 20 categories. Only 2 CDSs were not assigned to any COG;
however, 1,093 CDSs were classied to class S with unknown function.
The remaining 2,804 CDSs (out of all 3,850 protein coding genes and 49
pseudogenes) were classied to COG class. Individual chromosomal and
plasmidic features are shown in Fig. 1 together with GC content and GC
skew plots for the whole genome. Substantial drop in the average CG
content was observed around position 3,810,000 bp on chromosome,
where are located 5S rRNA (KUL73_17050), 23S rRNA (KUL73_17055)
and 16S rRNA (KUL73_17070) genes. More detailed results can be seen
in Supplementary Table S3.
The genome of R. rubrum DSM 467 was annotated in terms of gene
ontology. Together, 3,047 GO terms were assigned to 1,312 genomic
loci on the chromosome, and 30 GO terms were connected to 14
genomic elements on the plasmid. The most common terms for the three
GO categories were: GO:0006412 translation for biological process
assigned 58 times, for molecular function GO:0005524 ATP binding
with 86 genomic loci and for cellular component GO:0016020 mem-
branewas assigned 107 times.
CRISPR analysis showed 11 arrays with lengths from 337 to 2,467 bp
and a number of spacers from 4 to 40 (see Supplementary Table S4).
CRISPRDetect tool did not identify any Cas genes, but a manual search of
the annotated genes found 34 Cas-like genes, but no Cas9 protein. In the
case of prophage identication, Physpy did not identify any viral DNA,
Prophage Hunter found one ambiguous prophage candidate, and
Phaster localized four incomplete prophages, all of them presenting a
low score (see Supplementary Table S5). The restriction-modication
systems analysis revealed two systems containing restriction endonu-
clease and methylase where one belonged to type II and one to type III.
In addition, another gene coding for methylase type II was also found
(see Supplementary Table S6).
3.3. Cultivation & growth kinetics
To achieve cultivation-driven transcriptome changes in the wild-type
strain of R. rubrum and its knock-out strain, these microorganisms were
cultivated on two substrates, namely acetate and fructose. Growth
curves were determined throughout cultivation, with samples taken
over time and their optical density measured at 660 nm. Except for the
cultivation of the KO strain on acetate, samples were collected in each
cultivation in the mid-exponential phase, at the end of the exponential
phase, and in the stationary phase (for precise sampling times see Sup-
plementary Table S1). For the KO strain on acetate, only two time points
were selected for further characterization (see Fig. 2).
From the obtained growth curves, the growth rate and doubling time
for each cultivation were also calculated (see Table 2 and Supplemen-
tary Fig. S1).
Additionally, pH was of course also analyzed during the cultivations.
The pH of both cultures cultivated on fructose quickly reached values of
7.5 - 7.7 and remained almost constant until the end of cultivation. In
the case of acetate, the pH reached more alkaline values of about 8.6 -
8.8 during the initial stage of cultivation and remained constant until the
end of the experiment.
3.4. Ultrastructural analysis
The presence of cytoplasmic PHA granules in the cultures cultivated
on two different carbon sources as well as the overall ultrastructure of
the cells of R. rubrum wild-type was determined using low voltage
transmission electron microscopy. As seen in Fig. 3, both cultures
contain in their spiral-shaped cells a substantial amount of small
electron-lucent granules chromatophores. However, only cultures
grown on acetate as a carbon source were able to also produce PHA in
their cells, which formed bigger electron-lucent granules.
3.5. RNA-Seq transcriptome
RNA sequencing (Supplementary Table S1) produced 713 millions of
single-end reads averaging on 21.6 million reads per sample. After the
rst part of pre-processing (adapter and quality trimming), the reads
lengths were either 70 bp (samples no. 1 9) or 66 bp (samples no. 10
33) based on the used library construction method, and the average
Phred score was 35 for all samples. The subsequent pre-processing step
removed the remaining contamination of rRNA from the data. The
proportion of reads corresponding to 16S and 23S differed among
samples, e.g., all three samples for WT strain grown on fructose in the
third time-point contained around 46 % of rRNA reads, and on the other
hand 23 samples had less than 5 % of rRNA contamination (see Sup-
plementary Fig. S2).
Finally, reads were mapped to the reference genome of the wild-type
strain. The majority of reads (81 - 99 %) were mapped uniquely, yet
Table 1
Genomic features of Rhodospirillum rubrum DSM 467
T
.
Chromosome Plasmid
Length [bp] 4,352,570 53,835
GC content [%] 65.4 59.8
Number of ORFs 3,919 49
Number of operons 2,116 30
Protein coding genes 3,807 43
Pseudogenes 43 6
rRNA genes (5S, 16S, 23S) 4, 4, 4 -
tRNA 54 -
ncRNA 3 -
K. Jureckova et al.
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
2685
multimapping reads, persisting deduplication, were also considered due
to the existence of overlapping genes in the genome (see Supplementary
Fig. S3). The multimapping reads were in further analysis down
heightened by the number of associated genomic features; therefore, the
original number of reads in each sample remained the same. Overall,
only two pseudogenes located on the plasmid (KUL73_19640 and
KUL73_19670) remained completely silent with RPKM <1 in all tested
conditions.
Reproducibility of our experiments was veried by three biological
replicates for each sampling time-point and all tested condition, and
visualized through dimension reduction of normalized read counts by t-
Distributed Stochastic Neighbor Embedding (t-SNE) method. In most
cases, the data points formed individual clusters representing specic
cultivation condition and sampling time-point (see Fig. 4). Exceptions
can be found in WT AY and KO AY samples, where two bigger clusters
were formed based on the type of cultivated strain and used substrate,
yet within them smaller clusters can be located and they are repre-
senting the time-points.
The main differences between cultivation conditions were identied
through differential expression analysis, which was performed between
cultivations on fructose and on acetate separately for wild-type (WT AY
vs WT FY) and for knock-out strain (KO AY vs KO FY), and between wild-
type and knock-out strain, separately for cultivation on fructose (KO FY
vs WT FY) and on acetate (KO AY vs WT AY). Therefore, four different
combinations of cultivation conditions were tested and for each of them,
comparisons between available time-points were conducted, e.g., for
WT AY vs WT FY differences were found between WT AY T1 vs WT FY
T1, WT AY T2 vs WT FY T2, and WT AY T3 vs WT FY T3, and similarly
for the rest of comparisons (see Supplementary Figs. S4-S13 for corre-
sponding Volcano plots). Results for each combination of cultivation
Fig. 1. Chromosomal maps of Rhodospirillum rubrum DSM 467
T
chromosome and plasmid. The rst, second, and third outermost circles represent CDSs on the
forward and backward strands, and pseudogenes, respectively. Classication of COGs is represented by colors. Next, RNA genes, distinguishing among tRNA, rRNA,
and ncRNA, are represented in the fourth outermost circle. The inner area represents the GC content and GC skew (window size 10,000 bp, step size 200 bp).
Fig. 2. Growth dynamics of R. rubrum WT and KO on acetate and fructose,
obtained in SYN medium at 30
C, 160 rpm in dark.
Table 2
Growth rate and doubling time for R. rubrum WT and KO strains grown on ac-
etate (AY) and fructose (FY).
Sample µ max [h
1
] Td [h]
WT FY 0.057 ±0.001 12.082 ±0.280
WT AY 0.094 ±0.000 7.356 ±0.039
KO FY 0.037 ±0.002 18.992 ±1.299
KO AY 0.017 ±0.001 40.312 ±1.946
K. Jureckova et al.
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
2686
conditions were then separately searched for genes with any statistically
signicant changes in the gene expression levels (either up- or down-
regulated) (adjusted p value <0.05, Benjamini-Hochberg correction)
and they were sorted into two groups based on the number of signicant
regulations. The rst group contained genes with at least one regulation,
i.e., statistically signicant differential expression, among tested pairs of
time-points. The second group consisted of genes that had statistically
signicant change in every performed comparison of available time-
points. The rst group was used as gene universum in GO enrichment
analysis and the second group was used as a set of interesting genes, that
Fig. 3. Morphology of R. rubrum wild-type grown on A) acetate or B) fructose as carbon source, imaged using low voltage transmission electron microscopy. PHA
granules are marked with arrows, smaller electron-lucent granules represent chromophores.
Fig. 4. Comparison of RNA-Seq samples through dimensionality reduction by t-SNE. All samples are represented as points color-coded according to strain (WT or
KO), substrate (FY or AY), time-point (T1, T2 or T3) and text label indicates biological replicate (sfA, sfB or sfC).
K. Jureckova et al.
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
2687
were used to identify GO terms that were signicantly enriched (p value
<0.05, Fishers exact test) between tested conditions.
GO enrichment analysis revealed that for the comparison WT AY vs
WT FY 18 GO terms were enriched in biological process (BP) category
and 11 terms in molecular function (MF) category. BP terms were
related to disaccharide metabolic process, cell cycle, DNA recom-
bination and more, and MF terms corresponded to acetyltransferase
activity, transposase activity or isomerase activity (see Supple-
mentary File 1 and Table 3 for results of differential expression analysis
of selected genes).
Comparison between KO AY and KO FY showed 10 enriched BP GO
terms connected to regulation of biological process, maintenance of
DNA repeat elements and organic acid catabolic processterms and
the only MF term intramolecular transferase activity (see Supple-
mentary File 2).
The difference between WT and KO strains grown on fructose where
characterized with 12 BP terms and 5 MF terms, that where signicantly
enriched. The main BP terms described archaeal or bacterial-type a-
gellum-dependent cell motility or vitamin metabolic process pro-
cesses and MF terms were related to NAD binding, methyltransferase
activityor oxidoreductase activity, acting on CH-OH group of donors
(see Supplementary File 3 and Table 4 for results of differential
expression analysis of selected genes).
Finally, GO enrichment analysis between KO AY and WT AY iden-
tied 21 BP and 6 MF enriched terms. The main BP terms belonged to
aspartate family amino acid metabolic process, regulation of
biological process, regulation of cellular metabolic processterms. MF
terms were pyrophosphatase activityor hydrolase activity, acting on
acid anhydride (see Supplementary File 4 and Table 5 for results of
differential expression analysis of selected genes).
3.6. Small RNA prediction
For the prediction of small RNAs from the strand-specic RNA-Seq
data, a coverage-based detection with baerhunter tool, which requires
setting three input values, was performed. While the low_cut_off
threshold was left on default value of 5, the high_cut_off threshold was set
to 25 as inferred from our dataset (see Supplementary Fig. S14). The last
parameter, min _sRNA_length of predicted elements was set to 40 bp. The
length distribution of putative sRNAs (see Supplementary Fig. S15)
contains a wide variety of lengths ranging from ten to thousands bp.
Table 6 then contains counts of predicted elements in the chromosome
and plasmid sequences of R. rubrum. Small RNAs were further classied
into two groups based on their overlap with other CDSs. Those pre-
dictions that did not overlap with the CDS on either strand, thus were in
intergenic regions, were labeled as trans-encoded sRNAs. On the con-
trary, if there was even a slight overlap with any annotated element on
the opposite strand as an inferred sRNA, the prediction was considered
cis-encoded.
Additionally, differential expression analysis revealed that majority
(>99 %) of predicted sRNA were at least once differentially expressed
among various combinations of available conditions, see Table 6. The
Table 3
Differential expression analysis results of selected genes related to signicantly enriched GO terms for comparison between cultivation on fructose and on acetate for
wild-type strain (WT AY vs WT FY).
WT AY T1 vs WT FY T1 WT AY T2 vs WT FY T2 WT AY T3 vs WT FY T3
Gene
abbr.
Putative physiological function Locus tag log2 fold
change
p-adj log2 fold
change
p-adj log2 fold
change
p-adj
MF GO term: GO:0009055 electron transfer activity
nuoB NADH-quinone oxidoreductase subunit NuoB KUL73_07380 -1.65 4.95E-
02
-1.52 5.11E-
02
-1.04 2.43E-
01
NADH-quinone oxidoreductase subunit C KUL73_08070 -0.33 3.31E-
02
0.54 2.38E-
04
-1.08 2.95E-
13
nuoF NADH-quinone oxidoreductase subunit NuoF KUL73_08085 0.66 6.42E-
04
2.15 5.34E-
31
1.51 3.40E-
15
NADH-quinone oxidoreductase subunit J KUL73_08105 -0.84 7.79E-
03
1.00 1.18E-
03
3.49 4.78E-
25
nuoL NADH-quinone oxidoreductase subunit L KUL73_08115 -1.36 2.39E-
09
0.61 8.42E-
03
1.98 2.17E-
17
NADH-quinone oxidoreductase subunit M KUL73_08120 -2.27 7.70E-
22
-0.35 1.58E-
01
1.23 6.61E-
07
nuoN NADH-quinone oxidoreductase subunit NuoN KUL73_08125 -1.12 1.87E-
06
0.87 1.75E-
04
2.02 6.22E-
17
grxD Grx4 family monothiol glutaredoxin KUL73_03670 -1.67 1.44E-
06
-3.93 1.46E-
31
-2.82 7.84E-
17
c-type cytochrome KUL73_05315 -1.51 1.16E-
12
-1.72 1.44E-
16
-2.66 1.14E-
37
petA ubiquinol-cytochrome c reductase iron-sulfur
subunit
KUL73_06225 0.14 7.21E-
01
-1.77 2.11E-
08
-2.18 9.80E-
12
cytochrome b/b6 domain-containing protein KUL73_06485 -0.55 8.36E-
02
-2.70 5.84E-
20
-2.79 4.51E-
21
ccoP cytochrome-c oxidase, cbb3-type subunit III KUL73_17205 -0.74 7.63E-
03
1.55 6.66E-
09
1.93 3.56E-
12
ccoO cytochrome-c oxidase, cbb3-type subunit II KUL73_17215 -1.91 1.81E-
09
1.25 8.00E-
05
1.03 1.97E-
03
ccoN cytochrome-c oxidase, cbb3-type subunit I KUL73_17220 -0.79 9.68E-
03
1.73 3.21E-
09
1.90 3.12E-
10
pseudoazurin KUL73_05930 -0.51 1.35E-
01
3.25 1.64E-
22
1.42 1.02E-
03
BP GO term: GO:0006950 response to stress
ahpC peroxiredoxin KUL73_07360 -1.56 2.05E-
06
-1.80 2.39E-
08
-1.56 1.57E-
06
msrA peptide-methionine (S)-S-oxide reductase MsrA KUL73_12215 -1.64 1.64E-
04
-5.21 1.09E-
35
-4.50 5.38E-
27
nth endonuclease III KUL73_00795 -1.41 8.15E-
10
-3.08 5.56E-
43
-2.03 2.29E-
16
K. Jureckova et al.
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
2688
further statistics showing counts of differentially expressed sRNAs be-
tween these various combinations also distinguishes whether these el-
ements are up/down regulated (see Supplementary Fig. S16).
4. Discussion
4.1. Genome and transcriptome
R. rubrum DSM 467
T
hybrid assembly combining long Oxford
Table 4
Differential expression analysis results of selected genes related to signicantly enriched GO terms for comparison between wild-type vs knock-out strain for cultivation
on fructose (KO FY vs WT FY).
KO FY T1 vs WT FY T1 KO FY T2 vs WT FY T2 KO FY T3 vs WT FY T3
Gene
abbr.
Putative physiological function Locus tag log2 fold
change
p-adj log2 fold
change
p-adj log2 fold
change
p-adj
BP GO term: GO:0009110 vitamin biosynthetic process
thiE thiamine phosphate synthase KUL73_05625 -1.19 1.32E-
06
-0.44 8.13E-
02
0.88 4.37E-
03
thiD bifunctional hydroxymethylpyrimidine kinase/
phosphomethylpyrimidine kinase
KUL73_05735 0.09 4.91E-
01
0.75 5.81E-
10
-0.74 2.96E-
04
thiL thiamine-phosphate kinase KUL73_09425 1.17 1.44E-
07
3.08 5.67E-
43
1.23 1.96E-
04
thiC phosphomethylpyrimidine synthase ThiC KUL73_10385 1.06 1.49E-
04
0.34 2.45E-
01
-0.65 3.02E-
02
pdxY pyridoxal kinase KUL73_06260 1.05 4.41E-
04
1.91 5.21E-
11
1.74 3.95E-
07
pyridoxine 5
-phosphate synthase KUL73_09595 1.05 4.41E-
04
1.91 5.21E-
11
1.74 3.95E-
07
pdxH pyridoxamine 5
-phosphate oxidase KUL73_13555 0.05 8.79E-
01
0.89 5.05E-
04
-0.71 1.15E-
02
cobS cobaltochelatase subunit CobS KUL73_01095 -0.40 3.57E-
03
-0.79 3.57E-
09
-0.97 2.89E-
11
cobN cobaltochelatase subunit CobN KUL73_17390 0.46 8.99E-
02
2.40 5.11E-
21
1.07 1.19E-
04
cobS adenosylcobinamide-GDP ribazoletransferase KUL73_03465 1.58 1.63E-
08
2.38 5.65E-
18
1.42 1.66E-
04
cobU bifunctional adenosylcobinamide kinase/adenosylcobinamide-
phosphate guanylyltransferase
KUL73_03475 0.42 5.80E-
02
2.44 7.94E-
29
0.52 1.77E-
01
cobW cobalamin biosynthesis protein CobW KUL73_03480 0.33 3.57E-
02
1.68 1.83E-
29
-0.93 1.43E-
08
cobF precorrin-6A synthase (deacetylating) KUL73_15320 -0.13 6.83E-
01
0.73 1.12E-
02
0.22 6.50E-
01
cobalt-precorrin-6A reductase KUL73_15410 0.36 1.54E-
01
0.90 2.06E-
04
-0.79 1.60E-
02
precorrin-2 C(20)-methyltransferase KUL73_15420 1.33 2.04E-
06
2.30 1.57E-
16
2.27 4.18E-
08
cobJ precorrin-3B C(17)-methyltransferase KUL73_15415 1.49 1.35E-
17
2.75 2.30E-
53
2.19 2.69E-
13
BP GO term: GO:0071973 bacterial-type agellum-dependent cell motility
BP GO term: GO:0097588 archaeal or bacterial-type agellum-dependent cell motility
iJ agellar export protein FliJ KUL73_02755 0.84 5.36E-
03
1.88 1.39E-
10
1.28 2.55E-
04
agellin KUL73_13080 -1.35 1.00E-
05
-2.56 5.22E-
18
-2.46 2.30E-
16
agellar basal body-associated FliL family protein KUL73_14640 1.06 2.11E-
02
1.13 1.12E-
02
2.72 1.03E-
09
gG agellar basal-body rod protein FlgG KUL73_14650 -5.77 2.36E-
33
-7.48 2.46E-
57
-1.41 5.68E-
03
agellar basal body L-ring protein FlgH KUL73_14660 -1.21 5.01E-
03
-1.41 8.07E-
04
1.12 1.26E-
02
hypothetical protein KUL73_14700 -0.97 1.51E-
03
-3.52 2.04E-
33
-0.18 5.87E-
01
agellin KUL73_14725 -1.46 7.41E-
39
-2.47 1.91E-
110
-2.37 3.48E-
101
MF GO term: GO:0016614 oxidoreductase activity, acting on CH-OH group of donors
MF GO term: GO:0016616 oxidoreductase activity, acting on the CH-OH group of donors, NAD or NADP as acceptor
guaB IMP dehydrogenase KUL73_01290 -1.21 1.53E-
20
-0.64 6.71E-
07
-0.99 5.12E-
11
NADP-dependent isocitrate dehydrogenase KUL73_01875 -2.10 3.10E-
50
-1.24 1.05E-
18
-1.75 1.80E-
28
mdh malate dehydrogenase KUL73_06315 -1.13 2.58E-
06
-0.72 2.82E-
03
-0.04 8.98E-
01
3-hydroxybutyryl-CoA dehydrogenase KUL73_15880 -0.68 4.10E-
04
-1.41 4.66E-
14
-0.92 1.85E-
06
MF GO term: GO:0051287 NAD binding
NADP-dependent isocitrate dehydrogenase KUL73_01875 -2.10 3.10E-
50
-1.24 1.05E-
18
-1.75 1.80E-
28
nuoF NADH-quinone oxidoreductase subunit NuoF KUL73_08085 0.41 3.86E-
02
2.17 2.66E-
31
0.45 2.96E-
02
K. Jureckova et al.
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
2689
Nanopore and short Illumina reads reveal one circular chromosome and
one plasmid contig as was expected based on the previously published
type strain genome [10]. The chromosome exhibits GC content of
65.4 %, and the plasmid has 59.8 %, which is more than the average for
Gram-negative bacteria [48]; nevertheless, it again corresponds to the
previously published type strain as well as the chromosome length 4.35
Mbp and plasmid length 53.84 kbp. In comparison with type strain, the
overall number of genes is similar; there is only a slight difference be-
tween the number of predicted protein-coding genes (DSM 467
T
:3807 vs
S1
T
:3850), number of RNAs (DSM 467
T
:71 vs S1
T
:83) and predicted
pseudogenes (DSM 467
T
:49 vs S1
T
:9). The differences may be due to the
use of different sequencing platforms for strain sequencing and various
tools for their assemblies and annotations.
RNA-Seq samples for WT cultivation on fructose (samples no. 1 9)
were prepared slightly differently from the remaining ones (samples no.
10 33). Particularly, different rRNA depletion kit was used for the
latter group, which resulted in very low contamination of reads
belonging to 16S and 23S rRNA genes (see Supplementary Fig. S2) and
UMIs were added to allow their deduplication which caused the differ-
ence in read lengths after pre-processing. While the reads without UMIs
collected from the WT cultivation on fructose were 70 bp long, the
remaining ones were only 66 bp long. Yet, these dissimilarities did not
cause any signicant differences in the overall high number of reads per
sample or in the quality of the reads. Furthermore, reads were mapped to
the reference wild-type strain and the results also proved a high quality
of our data as we were able to uniquely map the majority of the reads. In
the worst case, WT_FY_sfB_T3 sample had 81 % of uniquely mapped
reads. This particular sample also had the highest contamination of
rRNA in reads with almost 47 %. On the other hand, many samples had
over 98 % of the reads with unique hit within the genome, and the
multimapping reads were identied only in 12 % of the reads. Multi-
mapping reads could be a consequence of 523 existing overlaps between
neighboring genes within the genome of R. rubrum (see Supplementary
Fig. S3). Nevertheless, our RNA-Seq data proved to be of a very high
quality, thus enabling further genome-wide study of changes in the
expression levels between different cultivation conditions. As main
differences were observed for particular substrates, we rstly used RNA-
Seq data to detect differential expression on a genome-wide scale in
particular time-points for acetate and fructose cultivation for WT
(Supplementary Figs. S4 S6) and KO (Supplementary Figs. S7 and S8)
strains. Additionally, we compared differences between KO and WT
strains in particular time-points on fructose (Supplementary Figs. S9
S11) and acetate (Supplementary Figs. S12 and S13). All volcano plots
showed expected relation among fold changes and statistically signi-
cant differences in expression of protein coding genes where number of
down-regulated and up-regulated genes is roughly the same.
4.2. Comparative analysis of acetate and fructose growth
The GO enrichment analysis comparing acetate and fructose cultures
Table 5
Differential expression analysis results of selected genes related to signicantly enriched GO terms for comparison between wild-type vs knock-out strain for cultivation
on acetate (KO AY vs WT AY).
KO AY T1 vs WT AY T1 KO AY T2 vs WT AY T2
Gene
abbr.
Putative physiological function Locus tag log2 fold
change
p-adj log2 fold
change
p-adj
MF GO term: GO:0016887 ATP hydrolysis activity
bchI magnesium chelatase ATPase subunit I KUL73_02545 -1.00 4.95E-
04
-0.68 1.98E-
02
AFG1 family ATPase KUL73_06305 -0.51 2.44E-
03
-0.28 9.87E-
02
arsA arsenical pump-driving ATPase KUL73_07505 1.89 4.00E-
10
1.50 9.17E-
07
tsaE tRNA (adenosine(37)-N6)-threonylcarbamoyltransferase complex ATPase subunit
type 1 TsaE
KUL73_17750 1.32 1.32E-
05
1.58 2.37E-
07
AAA family ATPase KUL73_19335 -0.63 3.15E-
05
-0.68 7.10E-
06
mfd transcription-repair coupling factor KUL73_08940 -0.96 6.99E-
10
-1.07 7.44E-
12
uvrA excinuclease ABC subunit UvrA KUL73_09080 1.09 1.59E-
04
1.53 1.04E-
07
mutL DNA mismatch repair endonuclease MutL KUL73_15195 -1.16 1.75E-
09
-1.02 1.40E-
07
htpG molecular chaperone HtpG KUL73_00370 1.06 1.54E-
02
1.00 2.13E-
02
groL chaperonin GroEL KUL73_00840 1.71 4.34E-
04
1.88 1.06E-
04
clpB ATP-dependent chaperone ClpB KUL73_03910 1.89 3.75E-
11
2.55 2.95E-
19
dnaK molecular chaperone DnaK KUL73_18350 0.77 1.04E-
02
1.42 1.44E-
06
lon endopeptidase La KUL73_08040 1.04 3.86E-
12
1.14 3.20E-
14
hslU ATP-dependent protease ATPase subunit HslU KUL73_18580 0.76 1.54E-
05
1.06 9.12E-
10
Table 6
Number of predicted small RNAs in Rhodospirillum rubrum DSM 467
T
.
Feature Predicted # in chromosome Differentially expressed Predicted # in plasmid Differentially expressed
Total sRNAs 2,329 2,321 10 10
trans 55 55 0 0
cis 2,274 2,266 10 10
K. Jureckova et al.
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
2690
for the R. rubrum wild-type strain revealed insights into the different
pathways of substrate assimilation. Fructose assimilation undergoes
catabolism mainly via the Embden-Meyerhof-Parnas (EMP) [49]. This
was highlighted by enriched GO terms involved in carbohydrate meta-
bolism, such as carbohydrate biosynthetic process (GO:0016051),
disaccharide metabolic process (GO:0005984) and trehalose metabolic
process (GO:0005991). In contrast, acetate is assimilated via the
ethylmalonyl-CoA (EMC) and methylbutanoyl-CoA (MBC) [5052]
pathways, both of which require an electron-transferring avoprotein
(ETF) to channel electrons to the membrane. The EMP pathway, in the
step of glyceraldehyde-3-phosphate oxidation, produces NADH [53]
which is then oxidized in the membrane through the electron transport
coupled phosphorylation. Therefore, the type of electron carrier used is
a key difference between the assimilation of acetate and fructose. The
term NADH dehydrogenase (ubiquinone) activity (GO:0008137) could
refer to the initial step in the electron transport chain [54,55] and reect
this difference in the assimilation pathway. This change in electron
delivery to the membrane seems to also be found in the term electron
transfer activity (GO:0009055). Investigation of the specic genes
involved in these terms revealed that most of the genes for the subunits
of the NADH-quinone oxidoreductase are more transcribed and regu-
lated under fructose conditions, except for the subunit F (nuoF)
(KUL73_08085) which is upregulated for acetate (see Table 3 and Sup-
plementary File 1). The NADH-quinone oxidoreductase is an important
element of the respiratory chain, therefore a different efciency in the
electron transport chain (ETC) between both carbon sources could be
expected, potentially leading to different energy levels. This hypothesis
is also supported by the different expression and regulation of genes
involved in the electron transport chain, c-type cytochrome
(KUL73_05315), ubiquinol-cytochrome c reductase complex
(KUL73_06225), cytochrome b/b6 domain (KUL73_06485)) found
under GO:0009055 (see Table 3). It is also noteworthy to mention that
the genes for the cbb3-type cytochrome c oxidase complex [56], usually
involved in presence of low oxygen levels, seemed to be transcribed for
both carbon sources. This observation may appear counterintuitive
considering the aerobic culture conditions, but the poor gas exchange
expected in shake ask without bafes could have led to lower dissolved
oxygen levels. Furthermore, the gene coding for pseudoazurin ppaZ
(KUL73_05930), which is encoded by ppaZ [57], is positively regulated
by the Reg/Prr system, related to sensing low oxygen levels, was acti-
vated exclusively under acetate culture.
The GO enrichment analysis also identied genes that could be
relevant in the interpretation of the lower maximum specic growth rate
(µ
max
) measured for fructose (see Table 2). Genes involved in reactive
oxygen species (ROS) scavenging and management of the oxidative
response (peroxiredoxin (KUL73_07360), peptide-methionine (S)-S-
oxide reductase MsrA (KUL73_12215) and endonuclease III
(KUL73_00795)) were more expressed during the growth on fructose
(see GO:0006950 response to stress in Table 3 and Supplementary File
1). While this upregulation of a limited set of three genes suggests a
possible increase in ROS production, this alone does not establish a
direct causal relationship with lower µ
max
. Further investigations are
required to understand the nature of this upregulation. However, it has
been observed that ROS production could be coming from electron
leakage from the ETC [58,59]. The increased presence of ROS may
indicate that electrons are diverted from the ETC, diminishing the ef-
ciency of the proton motive force (PMF), subsequently leading to
reduced energy production efciency (ATP). Moreover, ROS is known to
damage cells and negatively impact their metabolism. A study on Rho-
dobacter sphaeroides, a closely related organism, has shown that, in the
absence of oxygen and aerobic respiratory chain, a higher growth rate
was associated with lower ROS generation in autotrophically growing
cells [60]. Consequently, higher ROS levels could negatively affect cell
metabolism and indicate a less efcient respiratory chain, resulting in
diminished PMF, both contributing to the lower µ
max
.
The enrichment of the high-level BP terms cell cycle (GO:0007049)
and DNA recombination (GO:0006310) could be the result of the
different assimilation pathways between acetate and fructose explaining
different measured specic growth rates (µ
max AY
=0.094 h
1
and µ
max
FY
=0.057 h
1
). This observation suggests a different metabolic reor-
ganization around DNA transcription and RNA translation for acetate
and fructose cultures. Such profound difference seems to also be high-
lighted under the term GO:0006950 response to stress. Various stress
responses mechanisms activations were identied under both conditions
regarding DNA repair functions, transcription repair, recombination
mediator, elements related to DNA methylation and chaperone synthe-
sis. Nevertheless, it is worth noting that the carbon source was most
probably not the only parameter that changed between both cultures.
For example, cultivation using acetate has been shown to result in the
basication of the medium throughout the culture (data not shown). As
mentioned previously different levels of dissolved oxygen and carbon
dioxide should also be expected. These additional changes could nuance
the role of fructose and acetate metabolization in these growth rate and
transcriptomics observations.
Another important difference between acetate and fructose is that
acetate is an excellent substrate for growth associated PHB biosynthesis,
which is not the case for fructose. Although the GO analysis did not
directly reect any changes in PHB cycle activity, PHB granules were
observed only under acetate conditions (see Fig. 3). Therefore, these
observations led us to orient our interpretation of the GO terms also
towards changes induced by the presence of PHB. For instance, the
changes in DNA recombination could have been the result of the pres-
ence of PHB granules associated with growth on acetate. A multifunc-
tional protein (PhaM) has been discovered in C. necator [61,62] whose
role is to link PHB granules to the nucleoid (via PhaC), among other
functions. It was further proposed that PHB granules are associated with
the DNA and are segregated with the nucleoid during cell division [63].
However, no phaM homologs were so far found in R. rubrum.
4.3. Consequences of the deletion of the PHB biosynthesis on fructose
metabolism
The comparison of the cultivation and transcriptomics data analysis
between the polymerases mutant and wild-type strains grown on fruc-
tose led to unexpected results. It started with a signicant decrease in
the maximum specic growth rate for the mutant strain (see Table 2).
This nding is intriguing given the observation that fructose-grown cells
did not present PHB granules. Thus, cellsgrowth was not expected to be
affected by knocking out the PHA synthases. This suggests an unantic-
ipated activity of the PHB metabolism during aerobic growth with
fructose. Indeed, R. rubrum has been described to produce PHB from
fructose mainly in case of nutrient (nitrogen) or oxygen limitation [64,
65], suggesting that activation of the PHB biosynthesis from fructose
necessitates these specic conditions.
The GO enrichment analysis (see Supplementary File 3) indicates
that vitamin biosynthesis (GO:0009110) was inuenced by the presence
of both phaC1 and phaC2. Vitamins are important cofactors in numerous
enzymatic reactions and mutant cells could react to the disrupted PHB
biosynthesis by adjusting its pool of vitamins to favor alternative
metabolic pathways. Transcription of genes associated with vitamins
biosynthesis (see Table 4) suggested that the pool of thiamine, pyri-
doxine, and cobalamin could be modied. Additionally, the terms
GO:0009236 cobalamin biosynthetic process and GO:0033014 tetra-
pyrrole biosynthetic process, which are closely related, have also been
enriched upon deletion of the PHA synthases. Overall, it seems that a
majority of cobalamin biosynthesis genes presented higher expression
for the KO strain, suggesting that the cell could potentially try to favor
cobalamin production (Supplementary File 3). Cobalamin (vitamin B12)
participates in various important reactions (e.g., cofactor of reductase,
acetyltransferase, and isomerase) and can inuence transcriptional
regulation [66]. Interestingly, the cobalamin metabolism was also
modied by the CO adaptation of R. rubrum [67]. Regarding
K. Jureckova et al.
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
2691
tetrapyrrole-related genes, it also seems that a majority of them were
more expressed for the mutant strain compared to the WT strain (Sup-
plementary File 3). Tetrapyrroles [68] are essential cofactors notably for
light absorption, oxidative stress and electron transport. Therefore, the
mutant cells may also redirect their metabolic effort towards tetrapyr-
role production as a consequence of the altered PHB biosynthesis.
Tetrapyrrole is an intermediate of the bacteriochlorophyll biosynthetic
pathway and has been observed to be accumulated and excreted during
high cell density culture of R. rubrum [69], growing on succinate and
fructose. It can be speculated that the accumulation of these pigments
may contribute to photosynthetic membrane repression [70]. Thus,
these observations suggest that shifts in cobalamin and tetrapyrrole
metabolism could hint at changes in the regulation of the photosynthetic
apparatus [71]. Supporting this idea further, the literature also suggests
that pigment synthesis and PHB production may be interconnected via
the cells energy and redox status. A recent study [72] indeed showed
that the pigment synthesis regulatory protein HP1 is able to sense the
intracellular redox state and adjust the pigment synthesis. However, the
interpretation of these results remains challenging as R. rubrum was
cultivated in the presence of oxygen and in the dark. It is known that in
this organism, pigment synthesis is repressed by oxygen to avoid
oxidative stress and therefore occurs at low or zero oxygen levels, which
is not the case in our experiment. We thus hypothesize that an intra-
cellular redox imbalance may have been generated by disrupting the
PHB cycle, which leads to a redox state relatively similar to the one
observed in the absence of oxygen, which consequently impacted
pigment-related regulations.
Changes in oxidoreductase activity were also detected in the oxida-
tion of CH-OH group (GO:0016614) and reduction of NAD(P)
+
(GO:0016616) suggesting again that a perturbation of the PHB biosyn-
thesis could inuence redox-related cellular mechanisms. When looking
closely into the transcription of genes representing these GO terms
(Table 4 and Supplementary File 3), some of them were downregulated
for the KO strain (encoding for IMP dehydrogenase (KUL73_01290),
NADP-dependent isocitrate dehydrogenase (KUL73_01875), malate de-
hydrogenase (KUL73_06315), and 3-hydroxybutyryl-CoA dehydroge-
nase (KUL73_15880)). This interpretation is also conrmed by the
enriched term NAD binding (GO:0051287). In particular, the wild-type
strain presented an upregulation of the NADP-dependent isocitrate de-
hydrogenase gene (KUL73_01875) during the exponential growth phase,
compared to the mutant strain (Table 4). This enzyme is responsible for
the production of NADPH in the tricarboxylic acid (TCA) cycle and is
responsible for the production of NADPH and is also important with
respect to oxidative stress response [73]. Consequently, changes in the
regulation of KUL73_01875 upon deletion of the PHA synthase could
indicate a different intracellular redox state between the WT and KO
strains. This difference could also be one of the reasons why different
maximum specic growth rates were observed, as the lower growth rate
of the mutant strain could be the result of an altered activation of the
TCA cycle. In addition, the KO strain presented a higher transcription of
nuoF (KUL73_08085) (Table 4), responsible for NADH-quinone oxido-
reductase subunit NuoF, which was previously upregulated for acetate
in the comparison between both carbon sources for the WT strain
(Table 3 and Supplementary File 1). The relationship between the PHB
cycle and redox metabolism could originate from a modication of the
activity of acetoacetyl-CoA reductase (PhaB), which uses a reduced
cofactor to produce PHA precursors. The control mechanism of the PhaB
activity would need to be elucidated to establish the link between the
absence of PHA synthases and the overall redox state of the cell.
Moreover, the enrichment analysis also revealed an unexpected
impact of both PHA synthases deletions on the agellum-mediated cell
motility (GO:0071973 and GO:0097588). Further analysis of the genes
involved (see Table 4 and Supplementary File 3) suggested that the as-
sembly or composition of agellar components is modied, impacting
the cell motility. Interestingly, the relationship between the presence of
agellum and PHB accumulation has been studied in the past. For
instance, in the nutrient-limited accumulation of PHB in C. necator, cells
appeared to be agellated during the exponential growth phase and the
agellation became stagnant during the PHB accumulation phase [74].
This is followed by the complete loss of agella during the subsequent
PHB mobilization (after addition of a nitrogen source). The PHB
-
mutant
strain [75] unable to produce PHB also had a complete absence of
agellation under all conditions [74]. Additionally, other studies
showed that the deletion of genes involved in agellum formation
resulted in enhanced PHA production in natural and unnatural PHA
producers [76,77]. Our observation in R. rubrum is therefore coherent
with the literature on other strains and demonstrated that the link be-
tween agellum-mediated cell motility and PHB formation deserves
further exploration.
4.4. Consequences of the deletion of the PHB biosynthesis on acetate
metabolism
The strong growth inhibition observed in KO strain (see Table 2) is
expected to be related to the enrichment of high-level GO terms regu-
lation of biological process (GO:0050789) and regulation of cellular
metabolic process (GO:0031323). Our ndings also suggested that the
growth defect correlates with changes in ATP hydrolysis activity
(GO:0016887). Investigating the genes representing this term (see
Table 5 and Supplementary File 4) suggested variations inexpression of
ATPase elements between strains (magnesium chelatase ATPase subunit
I, AFG1 family ATPase, arsenical pump-driving ATPase, tRNA (adeno-
sine(37)-N6)-threonylcarbamoyltransferase complex ATPase subunit
type 1 TsaE, and AAA family ATPase). Genes involved in the dysfunction
of DNA repair (transcription-repair coupling factor, excinuclease ABC
subunit UvrA, and DNA mismatch repair endonuclease MutL) were also
differently transcribed for the KO strain. In addition, genes involved in
chaperones (molecular chaperone HtpG, chaperonin GroEL, ATP-
dependent chaperone ClpB, and molecular chaperone DnaK) and pro-
teases (endopeptidase La and ATP-dependent protease ATPase subunit
HslU) were upregulated in the KO strain, indicative of stress response
elements. These genes were also present in the GO:0016817 hydrolase
activity, acting on acid anhydride. These observations are expected to be
more related to an impaired substrate assimilation and carbon meta-
bolism in the mutant strain, rather than to the PHB metabolism per se.
Indeed, the similar R. rubrum ΔphaC1ΔphaC2 strain has been reported to
poorly assimilate acetate under aerobic conditions [52], compared with
the wild-type strain. It was also observed elsewhere that the presence of
intracellular PHB was associated with higher growth rates [12,16].
Thus, a perturbation of the EMC pathway and possibly also of the
methylbutanoyl-CoA pathway (MBC) [50] acting under acetate assimi-
lation as an anaplerotic pathway, could impact negatively the TCA cycle
when its biosynthetic precursors are depleted. This diminished TCA
cycle activity would lead to a weaker production of reduced electron
carriers and a lower energy generation despite harbouring a functional
aerobic respiratory chain. This could subsequently lead to the observed
defect in ATP hydrolysis. It is also worth noting that the EMC pathway
shares the 2 rst steps with the PHB biosynthesis.
Based on the absence of expression differences in the redox regula-
tion related terms, the redox metabolism may not be the reason for the
growth defect, as it was previously suggested in Rhizobium etli [78] and
Azotobacter beijerinckii [79] phaC mutants presenting NADH build-up.
This supports the hypothesis that an impaired carbon ow is central to
explain the differences in acetate assimilation between PHB
-
mutants
and wild-type R. rubrum. Indeed, several studies on PHB
-
mutant or-
ganisms [8083] observed a disrupted carbon ow under conditions
conducive to PHB accumulation. This illustrates the role of the PHB
cycle in the central metabolism and provides a framework for under-
standing how the phaC deletion affects acetate assimilation.
K. Jureckova et al.
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
2692
4.5. Small RNAs
Post-transcriptional bacterial regulation by small RNAs has been
described as the fastest response to external stimuli under specic con-
ditions related to the availability of regulatory elements at a given time
[84]. Besides that, Reyer et al. [85] showed by modelling that sRNAs
may be able to also act co-transcriptionally on nascent mRNA molecules
and, thus, making the regulation by sRNAs even more efcient. Their
regulatory role is primarily associated with the bacterial response to
stress conditions [86] and thus making sRNAs the subject of studies in
order to understand the principles of stress responses in organisms.
Direct inference of sRNA from batch RNA-Seq is very sensitive to the
choice of main thresholds used for the peak detection from the coverage
signal. We left low_cut_off threshold on the default value as it seems
reasonable value to both avoid low sequencing noise while not
excluding true positives with lower transcription. High_cut_off threshold
for the baerhunters prediction was inferred from samples depth distri-
butions as this threshold affects the number of predicted features and
thus should be experiment specic. Therefore, it seems to be ideal using
normalized sample depth as in our study. The length of sRNA can vary
but it typically spans within the interval 40500 bp [87]. Our pre-
dictions exceed both boundaries, which is expected. While the shortest
length is given by the threshold min _sRNA_length of predicted elements,
the longest length does not explicitly refer to the length of predicted
element but rather it is situated in that region as the exact prediction
from standard RNA-Seq can be misleading. The high number (>99 %) of
regulated sRNAs (see Table 6) correlates with the regulatory role of
these elements, however from the principle of baerhunter detection and
the fact that differentially expressed elements do not explicitly imply
that there cannot be the noise in the data, further analysis and experi-
ments need to be done to get more precise information about the sRNA
content. On the other hand, a larger number of non-coding regulatory
elements in R. rubrum genome is expected as versatile bacteria were
shown to possess more regulatory elements [88]. As our results
demonstrated, intergenic regions in R. rubrum genome hide probably
regulatory potential that remains to be further studied and understood.
5. Conclusions
In summary, our study focused on the genomic, transcriptomic, and
metabolic aspects of R. rubrum DSM 467
T
and its ΔphaC1ΔphaC2 knock-
out strain unable of PHA accumulation, shedding light on their re-
sponses to different carbon sources, namely acetate and fructose.
Comparative analysis between cultures grown on these substrates un-
veiled distinct pathways for substrate assimilation. The growth on
fructose resulted in an upregulation of carbohydrate metabolism due to
the most likely activity of the Embden-Meyerhof-Parnas pathway.
Indeed, acetate assimilation is believed to utilize the ethylmalonyl-CoA
and methylbutanoyl-CoA pathways. These differences extended to
electron transport chain components, impacting energy production ef-
ciency and growth rates. Generally, the absence of PHA biosynthetic
capability in KO strain affected fructose metabolism unexpectedly,
indicating a broader impact on vitamin biosynthesis and tetrapyrrole
metabolism. Changes in redox-related mechanisms and agellum-
mediated cell motility emphasized the intricate connections within
cellular processes. Similarly, the PHB synthesis deletions consequences
on acetate metabolism elucidated growth inhibition mechanisms. It was
suggested that impaired substrate assimilation resulted in reduced en-
ergy generation and, as a consequence, activated systems for cellular
repair and stress response. Finally, the study delved into metabolism and
stress response regulation by small RNA. A comprehensive analysis of
differentially expressed sRNAs underscored their regulatory role
emphasizing the need for further experiments to rene our
understanding.
CRediT authorship contribution statement
V´
eronique Amstutz: Writing original draft, Investigation. Kris-
tyna Hermankova: Writing original draft, Visualization, Formal
analysis, Data curation. Matej Bezdicek: Investigation. Katerina Mra-
zova: Investigation. Karel Sedlar: Writing original draft, Supervision,
Resources, Data curation, Conceptualization. Katerina Jureckova:
Writing original draft, Visualization, Formal analysis, Data curation.
Marketa Nykrynova: Writing original draft, Visualization, Formal
analysis, Data curation. Eva Slaninova: Writing original draft,
Investigation. Hugo Fleuriot-Blitman: Writing original draft, Inves-
tigation, Formal analysis. Stanislav Obruca: Writing original draft,
Supervision, Resources, Funding acquisition, Conceptualization.
Kamila Hrubanova: Investigation. Manfred Zinn: Writing original
draft, Supervision, Resources, Funding acquisition, Conceptualization.
Declaration of Competing Interest
None.
Data Availability
The genome assembly referred to in this paper is CP077803.1 for
chromosome and CP077804.1 for plasmid. All sequencing data have
been deposited in the NCBI Sequence Read Archive (SRA) under the
project accession number PRJNA742260. Particular SRA accession
numbers for genome sequencing are SRR28268536 (WT strain ONT),
SRR28268535 (WT strain Illumina), and SRR28268534 (KO strain
Illumina). Particular SRA accession numbers for RNA-Seq samples are
listed in Supplementary Table S1.
Acknowledgments
This study was supported by the bilateral grant project GACR
2115958L / SNSF 205321L_197275/1. The authors would like to thank
Professor Kevin E. O΄Connor and Professor Tanja Narancic, University
College Dublin, Ireland for kindly providing the ΔphaC1ΔphaC2 R.
rubrum mutant strain and Jana Hanslikova for help with sequencing this
knock-out strain. The authors thank Delong Instruments a.s., Czech
Republic, namely Jaromir Bacovsky for providing the low voltage
transmission electron microscopy analysis.
Appendix A. Supporting information
Supplementary data associated with this article can be found in the
online version at doi:10.1016/j.csbj.2024.06.023.
References
[1] Simon C, Daniel R. Metagenomic analyses: past and future trends. Appl Environ
Microbiol 2011;77:115361. https://doi.org/10.1128/AEM.02345-10.
[2] Gest H. A serendipic legacy: Erwin Esmarchs isolation of the rst photosynthetic
bacterium in pure culture. Photosynth Res 1995;46:4738. https://doi.org/
10.1007/BF00032302.
[3] Revelles O, Tarazona N, García JL, Prieto MA. Carbon roadmap from syngas to
polyhydroxyalkanoates in Rhodospirillum rubrum. Environ Microbiol 2016;18:
70820. https://doi.org/10.1111/1462-2920.13087.
[4] Karmann S, Panke S, Zinn M. Fed-batch cultivations of Rhodospirillum rubrum under
multiple nutrient-limited growth conditions on syngas as a novel option to produce
poly(3-Hydroxybutyrate) (PHB). Front Bioeng Biotechnol 2019;7. https://doi.org/
10.3389/fbioe.2019.00059.
[5] Najafpour G, Younesi H. Hydrogen production from synthesis gas using the
photosynthetic bacterium Rhodospirillum rubrum. ASEAN J Chem Eng 2005;5:35.
https://doi.org/10.22146/ajche.50162.
[6] Brandl H, Knee EJ, Fuller RC, Gross RA, Lenz RW. Ability of the phototrophic
bacterium Rhodospirillum rubrum to produce various poly (β-hydroxyalkanoates):
Potential sources for biodegradable polyesters. Int J Biol Macromol 1989;11:
4955. https://doi.org/10.1016/0141-8130(89)90040-8.
[7] Kaftan D, Bína D, Koblíˇ
zek M. Temperature dependence of photosynthetic reaction
centre activity in Rhodospirillum rubrum. Photosynth Res 2019;142:18193.
https://doi.org/10.1007/s11120-019-00652-7.
K. Jureckova et al.
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
2693
[8] Najafpour GD, Younesi H. Bioconversion of synthesis gas to hydrogen using a light-
dependent photosynthetic bacterium, Rhodospirillum rubrum. World J Microbiol
Biotechnol 2007;23:27584. https://doi.org/10.1007/s11274-006-9225-2.
[9] Ghosh R, Hardmeyer A, Thoenen I, Bachofen R. Optimization of the Sistrom culture
medium for large-scale batch cultivation of Rhodospirillum rubrum under
semiaerobic conditions with maximal yield of photosynthetic membranes. Appl
Environ Microbiol 1994;60:1698700. https://doi.org/10.1128/aem.60.5.1698-
1700.1994.
[10] Munk AC, Copeland A, Lucas S, Lapidus A, Del Rio TG, Barry K, et al. Complete
genome sequence of Rhodospirillum rubrum type strain (S1
T
). Stand Genom Sci
2011;4:293302. https://doi.org/10.4056/sigs.1804360.
[11] Lehman LJ, Roberts GP. Identication of an alternative nitrogenase system in
Rhodospirillum rubrum. J Bacteriol 1991;173:570511. https://doi.org/10.1128/
jb.173.18.5705-5711.1991.
[12] Jin H, Nikolau BJ. Role of genetic redundancy in polyhydroxyalkanoate (PHA)
polymerases in PHA biosynthesis in Rhodospirillum rubrum. J Bacteriol 2012;194:
55229. https://doi.org/10.1128/JB.01111-12.
[13] Jin H, Nikolau BJ. Evaluating PHA productivity of bioengineered Rhodosprillum
rubrum. PLoS One 2014;9:e96621. https://doi.org/10.1371/journal.
pone.0096621.
[14] Mastroleo F, Van Houdt R, Leroy B, Benotmane MA, Janssen A, Mergeay M, et al.
Experimental design and environmental parameters affect Rhodospirillum rubrum
S1H response to space ight. ISME J 2009;3:140219. https://doi.org/10.1038/
ismej.2009.74.
[15] Mastroleo F, Van Houdt R, Atkinson S, Mergeay M, Hendrickx L, Wattiez R, et al.
Modelled microgravity cultivation modulates N-acylhomoserine lactone
production in Rhodospirillum rubrum S1H independently of cell density. Microbiol
2013;159:245666. https://doi.org/10.1099/mic.0.066415-0.
[16] Klask C. Heterologous expression of various PHA synthase genes in Rhodospirillum
rubrum. Chem Biochem Eng Q 2015;29:7585. https://doi.org/10.15255/
CABEQ.2014.2249.
[17] Kouˇ
rilov´
a X, Schwarzerov´
a J, Pernicov´
a I, Sedl´
aˇ
r K, Mr´
azov´
a K, Krzyˇ
z´
anek V, et al.
The rst insight into polyhydroxyalkanoates accumulation in multi-extremophilic
Rubrobacter xylanophilus and Rubrobacter spartanus. Microorganisms 2021;9:909.
https://doi.org/10.3390/microorganisms9050909.
[18] Mrazova K, Bacovsky J, Sedrlova Z, Slaninova E, Obruca S, Fritz I, et al. Urany-less
low voltage transmission electron microscopy: a powerful tool for ultrastructural
studying of cyanobacterial cells. Microorganisms 2023;11:888. https://doi.org/
10.3390/microorganisms11040888.
[19] Leger A, Leonardi T. pycoQC, interactive quality control for Oxford Nanopore
Sequencing. J Open Source Softw 2019;4:1236. https://doi.org/10.21105/
joss.01236.
[20] Kolmogorov M, Yuan J, Lin Y, Pevzner PA. Assembly of long, error-prone reads
using repeat graphs. Nat Biotechnol 2019;37:5406. https://doi.org/10.1038/
s41587-019-0072-8.
[21] Li H. Minimap2: pairwise alignment for nucleotide sequences. Bioinformatics
2018;34:3094100. https://doi.org/10.1093/bioinformatics/bty191.
[22] Vaser R, Sovi´
c I, Nagarajan N, ˇ
Siki´
c M. Fast and accurate de novo genome assembly
from long uncorrected reads. Genome Res 2017;27:73746. https://doi.org/
10.1101/gr.214270.116.
[23] Ewels P, Magnusson M, Lundin S, K¨
aller M. MultiQC: summarize analysis results
for multiple tools and samples in a single report. Bioinformatics 2016;32:30478.
https://doi.org/10.1093/bioinformatics/btw354.
[24] Bolger AM, Lohse M, Usadel B. Trimmomatic: a exible trimmer for Illumina
sequence data. Bioinformatics 2014;30:211420. https://doi.org/10.1093/
bioinformatics/btu170.
[25] Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler
transform. Bioinformatics 2009;25:175460. https://doi.org/10.1093/
bioinformatics/btp324.
[26] Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The sequence
Alignment/Map format and SAMtools. Bioinformatics 2009;25:20789. https://
doi.org/10.1093/bioinformatics/btp352.
[27] Walker BJ, Abeel T, Shea T, Priest M, Abouelliel A, Sakthikumar S, et al. Pilon: An
integrated tool for comprehensive microbial variant detection and genome
assembly improvement. PLoS One 2014;9. https://doi.org/10.1371/journal.
pone.0112963.
[28] Tatusova T, DiCuccio M, Badretdin A, Chetvernin V, Nawrocki EP, Zaslavsky L,
et al. NCBI prokaryotic genome annotation pipeline. Nucleic Acids Res 2016;44:
661424. https://doi.org/10.1093/nar/gkw569.
[29] Huerta-Cepas J, Szklarczyk D, Heller D, Hern´
andez-Plaza A, Forslund SK, Cook H,
et al. eggNOG 5.0: a hierarchical, functionally and phylogenetically annotated
orthology resource based on 5090 organisms and 2502 viruses. Nucleic Acids Res
2019;47:D30914. https://doi.org/10.1093/nar/gky1085.
[30] Carver T, Thomson N, Bleasby A, Berriman M, Parkhill J. DNAPlotter: circular and
linear interactive genome visualization. Bioinformatics 2009;25:11920. https://
doi.org/10.1093/bioinformatics/btn578.
[31] Rutherford K, Parkhill J, Crook J, Horsnell T, Rice P, Rajandream M-A, et al.
Artemis: sequence visualization and annotation. Bioinformatics 2000;16:9445.
https://doi.org/10.1093/bioinformatics/16.10.944.
[32] Biswas A, Staals RHJ, Morales SE, Fineran PC, Brown CM. CRISPRDetect: a exible
algorithm to dene CRISPR arrays. BMC Genom 2016;17:356. https://doi.org/
10.1186/s12864-016-2627-0.
[33] Akhter S, Aziz RK, Edwards RA. PhiSpy: a novel algorithm for nding prophages in
bacterial genomes that combines similarity- and composition-based strategies.
e126e126 Nucleic Acids Res 2012;40. https://doi.org/10.1093/nar/gks406.
[34] Roberts RJ, Vincze T, Posfai J, Macelis D. REBASE: a database for DNA restriction
and modication: enzymes, genes and genomes. Nucleic Acids Res 2023;51:
D62930. https://doi.org/10.1093/nar/gkac975.
[35] Sayers EW, Cavanaugh M, Clark K, Ostell J, Pruitt KD, Karsch-Mizrachi I. GenBank.
Nucleic Acids Res 2019. https://doi.org/10.1093/nar/gkz956.
[36] Page AJ, Cummins CA, Hunt M, Wong VK, Reuter S, Holden MTG, et al. Roary:
rapid large-scale prokaryote pan genome analysis. Bioinformatics 2015;31:36913.
https://doi.org/10.1093/bioinformatics/btv421.
[37] Kopylova E, No´
e L, Touzet H. SortMeRNA: fast and accurate ltering of ribosomal
RNAs in metatranscriptomic data. Bioinformatics 2012;28:32117. https://doi.
org/10.1093/bioinformatics/bts611.
[38] Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, et al. The SILVA
ribosomal RNA gene database project: improved data processing and web-based
tools. Nucleic Acids Res 2012;41:D5906. https://doi.org/10.1093/nar/gks1219.
[39] Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, et al. STAR: ultrafast
universal RNA-seq aligner. Bioinformatics 2013;29:1521. https://doi.org/
10.1093/bioinformatics/bts635.
[40] Smith T, Heger A, Sudbery I. UMI-tools: modeling sequencing errors in Unique
Molecular Identiers to improve quantication accuracy. Genome Res 2017;27:
4919. https://doi.org/10.1101/gr.209601.116.
[41] Liao Y, Smyth GK, Shi W. featureCounts: an efcient general purpose program for
assigning sequence reads to genomic features. Bioinformatics 2014;30:92330.
https://doi.org/10.1093/bioinformatics/btt656.
[42] Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion
for RNA-seq data with DESeq2. Genome Biol 2014;15:550. https://doi.org/
10.1186/s13059-014-0550-8.
[43] van der Maaten L. Accelerating t-SNE using Tree-Based Algorithms. J Mach Learn
Res 2014;15:322145.
[44] Wickham H. ggplot2. New York, NY: Springer New York; 2009. https://doi.org/
10.1007/978-0-387-98141-3.
[45] Alexa A., Rahnenfurher J. topGO: Enrichment Analysis for Gene Ontology 2023.
doi:10.18129/B9.bioc.topGO.
[46] Ozuna A, Liberto D, Joyce RM, Arnvig KB, Nobeli I. baerhunter: an R package for
the discovery and analysis of expressed non-coding regions in bacterial RNA-seq
data. Bioinformatics 2020;36:9669. https://doi.org/10.1093/bioinformatics/
btz643.
[47] Lawrence M, Huber W, Pag`
es H, Aboyoun P, Carlson M, Gentleman R, et al.
Software for computing and annotating genomic ranges. PLoS Comput Biol 2013;9:
e1003118. https://doi.org/10.1371/journal.pcbi.1003118.
[48] Li X-Q, Du D. Variation, evolution, and correlation analysis of C+G content and
genome or chromosome size in different kingdoms and phyla. PLoS One 2014;9:
e88339. https://doi.org/10.1371/journal.pone.0088339.
[49] Grammel H, Gilles E-D, Ghosh R. Microaerophilic cooperation of reductive and
oxidative pathways allows maximal photosynthetic membrane biosynthesis in
Rhodospirillum rubrum. Appl Environ Microbiol 2003;69:657786. https://doi.org/
10.1128/AEM.69.11.6577-6586.2003.
[50] De Meur Q, Deutschbauer A, Koch M, Bayon-Vicente G, Cabecas Segura P,
Wattiez R, et al. New perspectives on butyrate assimilation in Rhodospirillum
rubrum S1H under photoheterotrophic conditions. BMC Microbiol 2020;20:126.
https://doi.org/10.1186/s12866-020-01814-7.
[51] Leroy B, De Meur Q, Moulin C, Wegria G, Wattiez R. New insight into the
photoheterotrophic growth of the isocytrate lyase-lacking purple bacterium
Rhodospirillum rubrum on acetate. Microbiol 2015;161:106172. https://doi.org/
10.1099/mic.0.000067.
[52] Narancic T, Scollica E, Kenny ST, Gibbons H, Carr E, Brennan L, et al.
Understanding the physiological roles of polyhydroxybutyrate (PHB) in
Rhodospirillum rubrum S1 under aerobic chemoheterotrophic conditions. Appl
Microbiol Biotechnol 2016;100:890112. https://doi.org/10.1007/s00253-016-
7711-5.
[53] Rudolf C, Grammel H. Fructose metabolism of the purple non-sulfur bacterium
Rhodospirillum rubrum: effect of carbon dioxide on growth, and production of
bacteriochlorophyll and organic acids. Enzym Micro Technol 2012;50:23846.
https://doi.org/10.1016/j.enzmictec.2012.01.007.
[54] Erb TJ, Fuchs G, Alber BE. (2S)-Methylsuccinyl-CoA dehydrogenase closes the
ethylmalonyl-CoA pathway for acetyl-CoA assimilation. Mol Microbiol 2009;73:
9921008. https://doi.org/10.1111/j.1365-2958.2009.06837.x.
[55] Erb TJ, Berg IA, Brecht V, Müller M, Fuchs G, Alber BE. Synthesis of C
5
-
dicarboxylic acids from C
2
-units involving crotonyl-CoA carboxylase/reductase:
The ethylmalonyl-CoA pathway. Proc Natl Acad Sci 2007;104:106316. https://
doi.org/10.1073/pnas.0702791104.
[56] Ekici S, Pawlik G, Lohmeyer E, Koch H-G, Daldal F. Biogenesis of cbb
3
-type
cytochrome c oxidase in Rhodobacter capsulatus. Biochim Et Biophys Acta (BBA) -
Bioenerg 2012;1817:898910. https://doi.org/10.1016/j.bbabio.2011.10.011.
[57] Jain R, Shapleigh JP. Characterization of nirV and a gene encoding a novel
pseudoazurin in Rhodobacter sphaeroides 2.4.3. Microbiol 2001;147:250515.
https://doi.org/10.1099/00221287-147-9-2505.
[58] Larosa V, Remacle C. Insights into the respiratory chain and oxidative stress. Biosci
Rep 2018;38. https://doi.org/10.1042/BSR20171492.
[59] Korshunov S, Imlay JA. Two sources of endogenous hydrogen peroxide in
Escherichia coli. Mol Microbiol 2010;75:1389401. https://doi.org/10.1111/
j.1365-2958.2010.07059.x.
[60] Lee YR, Lee W-H, Lee SY, Lee J, Kim M-S, Moon M, et al. Regulation of reactive
oxygen species promotes growth and carotenoid production under autotrophic
conditions in Rhodobacter sphaeroides. Front Microbiol 2022;13. https://doi.org/
10.3389/fmicb.2022.847757.
K. Jureckova et al.
Computational and Structural Biotechnology Journal 23 (2024) 2681–2694
2694
[61] Wahl A, Schuth N, Pfeiffer D, Nussberger S, Jendrossek D. PHB granules are
attached to the nucleoid via PhaM in Ralstonia eutropha. BMC Microbiol 2012;12:
262. https://doi.org/10.1186/1471-2180-12-262.
[62] Pfeiffer D, Wahl A, Jendrossek D. Identication of a multifunctional protein, PhaM,
that determines number, surface to volume ratio, subcellular localization and
distribution to daughter cells of poly(3-hydroxybutyrate), PHB, granules in
Ralstonia eutropha H16. Mol Microbiol 2011;82:93651. https://doi.org/10.1111/
j.1365-2958.2011.07869.x.
[63] Bresan S, Jendrossek D. New insights into PhaM-PhaC-mediated localization of
polyhydroxybutyrate granules in Ralstonia eutropha H16. Appl Environ Microbiol
2017;83. https://doi.org/10.1128/AEM.00505-17.
[64] Liu J, Zhao Y, Diao M, Wang W, Hua W, Wu S, et al. Poly(3-hydroxybutyrate-co-3-
hydroxyvalerate) production by Rhodospirillum rubrum using a two-step culture
strategy. J Chem 2019;2019:18. https://doi.org/10.1155/2019/8369179.
[65] Godoy MS, de Miguel SR, Prieto MA. Aerobic-anaerobic transition boosts poly(3-
hydroxybutyrate-co-3-hydroxyvalerate) synthesis in Rhodospirillum rubrum: the key
role of carbon dioxide. Micro Cell Fact 2023;22:47. https://doi.org/10.1186/
s12934-023-02045-x.
[66] Cheng Z, Li K, Hammad LA, Karty JA, Bauer CE. Vitamin B
12
regulates photosystem
gene expression via the CrtJ antirepressor AerR in Rhodobacter capsulatus. Mol
Microbiol 2014;91:64964. https://doi.org/10.1111/mmi.12491.
[67] Cavazza C, Collin-Faure V, P´
erard J, Diemer H, Cianf´
erani S, Rabilloud T, et al.
Proteomic analysis of Rhodospirillum rubrum after carbon monoxide exposure
reveals an important effect on metallic cofactor biosynthesis. J Proteom 2022;250:
104389. https://doi.org/10.1016/j.jprot.2021.104389.
[68] Busch AWU, Montgomery BL. Interdependence of tetrapyrrole metabolism, the
generation of oxidative stress and the mitigative oxidative stress response. Redox
Biol 2015;4:26071. https://doi.org/10.1016/j.redox.2015.01.010.
[69] Zeiger L, Grammel H. Model-based high cell density cultivation of Rhodospirillum
rubrum under respiratory dark conditions. Biotechnol Bioeng 2010;105:72939.
https://doi.org/10.1002/bit.22589.
[70] Carius L, Carius AB, McIntosh M, Grammel H. Quorum sensing inuences growth
and photosynthetic membrane production in high-cell-density cultivations of
Rhodospirillum rubrum. BMC Microbiol 2013;13:189. https://doi.org/10.1186/
1471-2180-13-189.
[71] Ghosh R, Roth E, Abou-Aisha K, Saegesser R, Autenrieth C. The monofunctional
cobalamin biosynthesis enzyme precorrin-3B synthase (CobZ
RR
) is essential for
anaerobic photosynthesis in Rhodospirillum rubrum but not for aerobic dark
metabolism. Microbiol 2018;164:141631. https://doi.org/10.1099/
mic.0.000718.
[72] Godoy MS, de Miguel SR, Prieto MA. A singular PpaA/AerR-like protein in
Rhodospirillum rubrum rules beyond the boundaries of photosynthesis in response to
the intracellular redox state. MSystems 2023;8. https://doi.org/10.1128/
msystems.00702-23.
[73] Obruca S, Snajdar O, Svoboda Z, Marova I. Application of random mutagenesis to
enhance the production of polyhydroxyalkanoates by Cupriavidus necator H16 on
waste frying oil. World J Microbiol Biotechnol 2013;29:241728. https://doi.org/
10.1007/s11274-013-1410-5.
[74] Raberg M, Reinecke F, Reichelt R, Malkus U, Ko¨nig S, Po¨tter M, et al. Ralstonia
eutropha H16 agellation changes according to nutrient supply and state of poly(3-
Hydroxybutyrate) accumulation. Appl Environ Microbiol 2008;74:447790.
https://doi.org/10.1128/AEM.00440-08.
[75] Schlegel HG, Lafferty R, Krauss I. The isolation of mutants not accumulating poly-
β-hydroxybutyric acid. Arch Für Mikrobiol 1970;71:28394. https://doi.org/
10.1007/BF00410161.
[76] Wang J, Ma W, Wang Y, Lin L, Wang T, Wang Y, et al. Deletion of 76 genes relevant
to agella and pili formation to facilitate polyhydroxyalkanoate production in
Pseudomonas putida. Appl Microbiol Biotechnol 2018;102:1052339. https://doi.
org/10.1007/s00253-018-9439-x.
[77] Wang J, Ma W, Fang Y, Zhang H, Liang H, Liu H, et al. Engineering the outer
membrane could facilitate better bacterial performance and effectively enhance
poly-3-hydroxybutyrate accumulation. Appl Environ Microbiol 2021;87. https://
doi.org/10.1128/AEM.01389-21.
[78] Cevallos MA, Encarnaci´
on S, Leija A, Mora Y, Mora J. Genetic and physiological
characterization of a Rhizobium etli mutant strain unable to synthesize poly-
β-hydroxybutyrate. J Bacteriol 1996;178:164654. https://doi.org/10.1128/
jb.178.6.1646-1654.1996.
[79] Senior PJ, Dawes EA. The regulation of poly-β-hydroxybutyrate metabolism in
Azotobacter beijerinckii. Biochem J 1973;134:22538. https://doi.org/10.1042/
bj1340225.
[80] Raberg M, Voigt B, Hecker M, Steinbüchel A. A closer look on the
polyhydroxybutyrate- (PHB-) negative phenotype of Ralstonia eutropha PHB
-
4.
PLoS One 2014;9:e95907. https://doi.org/10.1371/journal.pone.0095907.
[81] Cook AM, Schlegel HG. Metabolite concentrations in Alcaligenes eutrophus H 16 and
a mutant defective in poly-β-hydroxybutyrate synthesis. Arch Microbiol 1978;119:
2315. https://doi.org/10.1007/BF00405400.
[82] Jung Y-M, Yong-Hyun L. Investigation of regulatory mechanism of ux of Acetyl-
CoA in Alcaligenes eutrophus using PHB- negative Mutant and transformants
harboring cloned phbCAB Genes. J Microbiol Biotechnol 1997;7:21522.
[83] Steinbüchel A, Schlegel HG. Excretion of pyruvate by mutants of Alcaligenes
eutrophus, which are impaired in the accumulation of poly(β-hydroxybutyric acid)
(PHB), under conditions permitting synthesis of PHB. Appl Microbiol Biotechnol
1989;31:16875. https://doi.org/10.1007/BF00262457.
[84] Shimoni Y, Friedlander G, Hetzroni G, Niv G, Altuvia S, Biham O, et al. Regulation
of gene expression by small non-coding RNAs: a quantitative view. Mol Syst Biol
2007;3. https://doi.org/10.1038/msb4100181.
[85] Reyer MA, Chennakesavalu S, Heideman EM, Ma X, Bujnowska M, Hong L, et al.
Kinetic modeling reveals additional regulation at co-transcriptional level by post-
transcriptional sRNA regulators. Cell Rep 2021;36:109764. https://doi.org/
10.1016/j.celrep.2021.109764.
[86] Gottesman S, McCullen CA, Guillier M, Vanderpool CK, Majdalani N,
Benhammou J, et al. Small RNA regulators and the bacterial response to stress.
Cold Spring Harb Symp Quant Biol 2006;71:111. https://doi.org/10.1101/
sqb.2006.71.016.
[87] Huang H-Y, Chang H-Y, Chou C-H, Tseng C-P, Ho S-Y, Yang C-D, et al. sRNAMap:
genomic maps for small non-coding RNAs, their regulators and their targets in
microbial genomes. Nucleic Acids Res 2009;37:D1504. https://doi.org/10.1093/
nar/gkn852.
[88] Bervoets I, Charlier D. Diversity, versatility and complexity of bacterial gene
regulation mechanisms: opportunities and drawbacks for applications in synthetic
biology. FEMS Microbiol Rev 2019;43:30439. https://doi.org/10.1093/femsre/
fuz001.
K. Jureckova et al.
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Photosynthesis (PS) in purple non-Sulfur bacteria (PNSB) is strictly controlled to avoid energy consumption and to prevent oxidative stress when oxygen is present. PpsR, a transcriptional repressor encoded in the photosynthesis gene cluster (PGC), is one of the main players in the regulation of PS-related genes. In some cases, PpsR does not sense the environmental cue by itself but delegates this task to other proteins like PpaA/AerR anti-repressors, which can ultimately affect PpsR affinity to some promoter regions. In this work, the effects of locus Rru_A0625 product (HP1) on PS regulation of Rhodospirillum rubrum were studied by mutation/complementation and transcriptomic strategies. Rru_A0625 is located next to ppsR gene, just like other PpaA/AerR members, and its deletion annuls pigment synthesis in dark micro/anaerobic growth conditions. HP1 shows similarity to PpaA/AerR anti-repressors family, although it does not possess their typical cobalamin binding domain. A transcriptomic analysis of Rru_A0625 deletion mutant showed that HP1 not only has effects on bacteriochlorophyll and carotenoid biosynthesis, but also many other biological processes in the cell. The most notorious is the impact on the transcription of the nitrogenase complex components and accessory proteins. Our results suggest that this new member of the PpaA/AerR family has evolved losing the canonical cobalamin binding domain, but not the redox sensing capability, conserving a not fully understood mechanism of PS regulation. IMPORTANCE Rhodospirillum rubrum vast metabolic versatility places it as a remarkable model bacterium and an excellent biotechnological chassis. The key component of photosynthesis (PS) studied in this work (HP1) stands out among the other members of PpaA/AerR anti-repressor family since it lacks the motif they all share: the cobalamin B-12 binding motif. Despite being reduced and poorly conserved, HP1 stills controls PS as the other members of the family, allowing a fast response to changes in the redox state of the cell. This work also shows that HP1 absence affects genes from relevant biological processes other than PS, including nitrogen fixation and stress response. From a biotechnological perspective, HP1 could be manipulated in approaches where PS is not necessary, such as hydrogen or polyhydroxyalkanoates production, to save energy.
Article
Full-text available
Sample preparation protocols for conventional high voltage transmission electron microscopy (TEM) heavily rely on the usage of staining agents containing various heavy metals, most commonly uranyl acetate and lead citrate. However high toxicity, rising legal regulations, and problematic waste disposal of uranyl acetate have increased calls for the reduction or even complete replacement of this staining agent. One of the strategies for uranyless imaging is the employment of low-voltage transmission electron microscopy. To investigate the influence of different imaging and staining strategies on the final image of cyanobacterial cells, samples stained by uranyl acetate with lead citrate, as well as unstained samples, were observed using TEM and accelerating voltages of 200 kV or 25 kV. Moreover, to examine the possibilities of reducing chromatic aberration, which often causes issues when imaging using electrons of lower energies, samples were also imaged using a scanning transmission electron microscopy at 15 kV accelerating voltages. The results of this study demonstrate that low-voltage electron microscopy offers great potential for uranyless electron microscopy.
Article
Full-text available
Background Microbially produced bioplastics are specially promising materials since they can be naturally synthesized and degraded, making its end-of-life management more amenable to the environment. A prominent example of these new materials are polyhydroxyalkanoates. These polyesters serve manly as carbon and energy storage and increase the resistance to stress. Their synthesis can also work as an electron sink for the regeneration of oxidized cofactors. In terms of biotechnological applications, the co-polymer poly(3-hydroxybutyrate-co-3-hydroxyvalerate), or PHBV, has interesting biotechnological properties due to its lower stiffness and fragility compared to the homopolymer poly(3-hydroxybutyrate) (P3HB). In this work, we explored the potentiality of Rhodospirillum rubrum as a producer of this co-polymer, exploiting its metabolic versatility when grown in different aeration conditions and photoheterotrophically. Results When shaken flasks experiments were carried out with limited aeration using fructose as carbon source, PHBV production was triggered reaching 29 ± 2% CDW of polymer accumulation with a 75 ± 1%mol of 3-hydroxyvalerate (3HV) (condition C2). Propionate and acetate were secreted in this condition. The synthesis of PHBV was exclusively carried out by the PHA synthase PhaC2. Interestingly, transcription of cbbM coding RuBisCO, the key enzyme of the Calvin-Benson-Bassham cycle, was similar in aerobic and microaerobic/anaerobic cultures. The maximal PHBV yield (81% CDW with 86%mol 3HV) was achieved when cells were transferred from aerobic to anaerobic conditions and controlling the CO2 concentration by adding bicarbonate to the culture. In these conditions, the cells behaved like resting cells, since polymer accumulation prevailed over residual biomass formation. In the absence of bicarbonate, cells could not adapt to an anaerobic environment in the studied lapse. Conclusions We found that two-phase growth (aerobic-anaerobic) significantly improved the previous report of PHBV production in purple nonsulfur bacteria, maximizing the polymer accumulation at the expense of other components of the biomass. The presence of CO2 is key in this process demonstrating the involvement of the Calvin-Benson-Bassham in the adaptation to changes in oxygen availability. These results stand R. rubrum as a promising producer of high-3HV-content PHBV co-polymer from fructose, a PHBV unrelated carbon source.
Article
Full-text available
REBASE is a comprehensive and extensively curated database of information about the components of restriction-modification (RM) systems. It is fully referenced and provides information about the recognition and cleavage sites for both restriction enzymes and DNA methyltransferases together with their commercial availability, methylation sensitivity, crystal and sequence data. All completely sequenced genomes and select shotgun sequences are analyzed for RM system components. When PacBio sequence data is available, the recognition sequences of many DNA methyltransferases (MTases) can be determined. This has led to an explosive growth in the number of well-characterized MTases in REBASE. The contents of REBASE may be browsed from the web rebase.neb.com and selected compilations can be downloaded by FTP (ftp.neb.com). Monthly updates are also available via email.
Article
Full-text available
Industrial demand for capture and utilization using microorganisms to reduce CO 2 , a major cause of global warming, is significantly increasing. Rhodobacter sphaeroides is a suitable strain for the process of converting CO 2 into high-value materials because it can accept CO 2 and has various metabolic pathways. However, it has been mainly studied for heterotrophic growth that uses sugars and organic acids as carbon sources, not autotrophic growth. Here, we report that the regulation of reactive oxygen species is critical for growth when using CO 2 as a sole carbon source in R. sphaeroides . In general, the growth rate is much slower under autotrophic conditions compared to heterotrophic conditions. To improve this, we performed random mutagenesis using N -methyl- N ’-nitro- N -nitrosoguanidine (NTG). As a result, we selected the YR-1 strain with a maximum specific growth rate (μ) 1.44 day –1 in the early growth phase, which has a 110% faster growth rate compared to the wild-type. Based on the transcriptome analysis, it was confirmed that the growth was more sensitive to reactive oxygen species under autotrophic conditions. In the YR-1 mutant, the endogenous contents of H 2 O 2 levels and oxidative damage were reduced by 33.3 and 42.7% in the cells, respectively. Furthermore, we measured that concentrations of carotenoids, which are important antioxidants. The total carotenoid is produced 9.63 g/L in the YR-1 mutant, suggesting that the production is 1.7-fold higher than wild-type. Taken together, our observations indicate that controlling ROS promotes cell growth and carotenoid production under autotrophic conditions.
Article
Full-text available
Small RNAs (sRNAs) are important gene regulators in bacteria. Many sRNAs act post-transcriptionally by affecting translation and degradation of the target mRNAs upon base-pairing interactions. Here we present a general approach combining imaging and mathematical modeling to determine kinetic parameters at different levels of sRNA-mediated gene regulation that contribute to overall regulation efficacy. Our data reveal that certain sRNAs previously characterized as post-transcriptional regulators can regulate some targets co-transcriptionally, leading to a revised model that sRNA-mediated regulation can occur early in an mRNA’s lifetime, as soon as the sRNA binding site is transcribed. This co-transcriptional regulation is likely mediated by Rho-dependent termination when transcription-coupled translation is reduced upon sRNA binding. Our data also reveal several important kinetic steps that contribute to the differential regulation of mRNA targets by an sRNA. Particularly, binding of sRNA to the target mRNA may dictate the regulation hierarchy observed within an sRNA regulon.
Article
Full-text available
Poly-3-hydroxybutyrate (PHB) is an environmentally friendly polymer and can be produced in Escherichia coli cells after overexpressing the heterologous gene cluster phaCAB . The biosynthesis of outer membrane (OM) consumes lots of nutrients and influences cell morphology. Here we engineered OM by disrupting all gene clusters relevant to polysaccharide portion of LPS, colanic acid (CA), flagella or/and fimbria in E. coli W3110. All these disruptions benefited PHB production. Especially, disrupting all these OM components improved PHB content to 83.0 wt%, while the wild-type control produced only 1.5 wt% PHB. The improvement was mainly due to the LPS truncation to Kdo 2 -lipid A, which facilitated 82.0 wt% PHB with 25-fold larger cell volume; and disrupting CA facilitated 57.8 wt% PHB. In addition, disrupting LPS facilitated advantageous fermentation features including 69.1% less acetate, 550% higher percentage of autoaggregated cells among the total culture cells, 69.1% less biofilm and higher cellular broken ratio. Further detailed mechanism investigations showed that disrupting LPS caused global regulations on envelope and cellular metabolism: (i) sharply decrease of flagella, fimbria and secretions; (ii) more elastic cell; (iii) much more carbon flux towards acetyl-CoA and cofactors supply including NADP, NAD and ATP; (iv) decrease of byproduct acids but increase of γ-aminobutyric acid by activating σ E factor. Disrupting CA, flagella and fimbria also improved the levels of acetyl-CoA and cofactors. The results indicated that engineering OM is an effective strategy to enhance PHB production, and highlighted the applicability of OM engineering to increase microbial cell factory performance. Importance Understanding the detailed influence of OM on cell envelope and cellular metabolism is important for optimizing E. coli cell factory and many other microorganisms. This study revealed the applicability of remodeling OM to enhance PHB accumulation as representative inclusion bodies. The knowledge generated in this study provided insights concerning the influence and application of OM engineering, and gave essential references for producing other inclusion bodies or chemicals derived from acetyl-CoA or with the need of cofactor NADPH, NADH or ATP supply, and reducing byproduct acids. This study is promising to provide new ideas for the improvement of microbial cell factories.
Article
Full-text available
Actinobacteria belonging to the genus Rubrobacter are known for their multi-extremophilic growth conditions-they are highly radiation-resistant, halotolerant, thermotolerant or even thermophilic. This work demonstrates that the members of the genus are capable of accumulating polyhydroxyalkanoates (PHA) since PHA-related genes are widely distributed among Rubrobacter spp. whose complete genome sequences are available in public databases. Interestingly, all Rubrobacter strains possess both class I and class III synthases (PhaC). We have experimentally investigated the PHA accumulation in two thermophilic species, R. xylanophilus and R. spartanus. The PHA content in both strains reached up to 50% of the cell dry mass, both bacteria were able to accumulate PHA consisting of 3-hydroxybutyrate and 3-hydroxyvalerate monomeric units, none other monomers were incorporated into the polymer chain. The capability of PHA accumulation likely contributes to the multi-extremophilic characteristics since it is known that PHA substantially enhances the stress robustness of bacteria. Hence, PHA can be considered as extremolytes enabling adaptation to extreme conditions. Furthermore, due to the high PHA content in biomass, a wide range of utilizable substrates, Gram-stain positivity, and thermophilic features, the Rubrobacter species, in particular Rubrobacter xylanophilus, could be also interesting candidates for industrial production of PHA within the concept of Next-Generation Industrial Biotechnology.
Article
Full-text available
Background: The great metabolic versatility of the purple non-sulfur bacteria is of particular interest in green technology. Rhodospirillum rubrum S1H is an α-proteobacterium that is capable of photoheterotrophic assimilation of volatile fatty acids (VFAs). Butyrate is one of the most abundant VFAs produced during fermentative biodegradation of crude organic wastes in various applications. While there is a growing understanding of the photoassimilation of acetate, another abundantly produced VFA, the mechanisms involved in the photoheterotrophic metabolism of butyrate remain poorly studied. Results: In this work, we used proteomic and functional genomic analyses to determine potential metabolic pathways involved in the photoassimilation of butyrate. We propose that a fraction of butyrate is converted to acetyl-CoA, a reaction shared with polyhydroxybutyrate metabolism, while the other fraction supplies the ethylmalonyl-CoA (EMC) pathway used as an anaplerotic pathway to replenish the TCA cycle. Surprisingly, we also highlighted a potential assimilation pathway, through isoleucine synthesis and degradation, allowing the conversion of acetyl-CoA to propionyl-CoA. We tentatively named this pathway the methylbutanoyl-CoA pathway (MBC). An increase in isoleucine abundance was observed during the early growth phase under butyrate condition. Nevertheless, while the EMC and MBC pathways appeared to be concomitantly used, a genome-wide mutant fitness assay highlighted the EMC pathway as the only pathway strictly required for the assimilation of butyrate. Conclusion: Photoheterotrophic growth of Rs. rubrum with butyrate as sole carbon source requires a functional EMC pathway. In addition, a new assimilation pathway involving isoleucine synthesis and degradation, named the methylbutanoyl-CoA (MBC) pathway, could also be involved in the assimilation of this volatile fatty acid by Rs. rubrum.
Article
Some carboxydotrophs like Rhodospirillum rubrum are able to grow with CO as their sole source of energy using a Carbone monoxide dehydrogenase (CODH) and an Energy conserving hydrogenase (ECH) to perform anaerobically the so called water-gas shift reaction (WGSR) (CO + H2O- > CO2 + H2). Several studies have focused at the biochemical and biophysical level on this enzymatic system and a few OMICS studies on CO metabolism. Knowing that CO is toxic in particular due to its binding to heme iron atoms, and is even considered as a potential antibacterial agent, we decided to use a proteomic approach in order to analyze R. rubrum adaptation in term of metabolism and management of the toxic effect. In particular, this study allowed highlighting a set of proteins likely implicated in ECH maturation, and important perturbations in term of cofactor biosynthesis, especially metallic cofactors. This shows that even this CO tolerant microorganism cannot avoid completely CO toxic effects associated with its interaction with metallic ions. Significance This proteomic study highlights the fact that even in a microorganism able to handle carbon monoxide and in some way detoxifying it via the intrinsic action of the carbon monoxide dehydrogenase (CODH), CO has important effects on metal homeostasis, metal cofactors and metalloproteins. These effects are direct or indirect via transcription regulation, and amplified by the high interdependency of cofactors biosynthesis.