Summary:
Legionella pneumophila (L.p.) is a gram-negative, intracellular pathogen and a common cause
of severe community-acquired pneumonia. In humans, L.p. replicates primarily within alveolar
macrophages. It manipulates vital host cell functions such as vesicle trafficking and gene
expression by the secretion of over 300 effector proteins into the host cell cytosol. Thus, L.p.
modifies its host cell to promote its own replication. An unbiased and global analysis of the
molecular changes and biological processes that are associated with bacterial infections of
human cells can provide new insights into host-pathogen interactions. Therefore, one goal of
this study was to characterize expression changes of different RNA species in response to
infection with L.p. in human primary blood-derived macrophages (BDMs) or differentiated
THP-1 cells. This work is structured into two parts: (1) a functional study on how miRNA
manipulations can alter L.p. replication in macrophages and (2) an in depth analysis of
transcriptomic events in host and pathogen during infection.
(1) In the last few decades, miRNAs have been established as critical modulators of immune
function. Therefore, one aim of this study was to identify the miRNA profile of L.p.-infected
macrophages and to determine the functional impact of a miRNA manipulation on L.p.
replication. BDMs of healthy donors were infected with L.p. strain Corby. Small RNA
sequencing revealed the miRNA profile in BDMs following L.p. infection. An upregulation of
miR-146a and miR-155, as well as downregulation of miR-221 and miR-125b was validated by
qPCR in macrophages. miRNA regulation in response to infection seems to be due to
transcriptional regulation of miRNA promoters, since the histone acetylation levels at the
promoter and the pri-miR expression correlated with the miRNA expression upon L.p.-
infection. Overexpression and knock down experiments of miR-125b, miR-221 and miR-579 in
combination were performed for functional characterization and showed an influence of all
three miRNAs on bacterial replication. A SILAC approach revealed the protein MX1 as
downregulated following simultaneous overexpression of all three miRNAs. MX1 is an
interferon-induced GTP-binding protein important for antiviral defence. As shown by
validation experiments, MX1 knockdown in macrophages led to an increased replication of
L.p., as seen following overexpression of the miRNAs. Since in silico analysis predicted no
binding sites for either miRNA in the 3’UTR of MX1, Ingenuity pathway analysis was performed
to find the linking molecules. DDX58 (RIG-I), a sensor for cytosolic RNA, was validated as a
target for miR-221, while the tumour suppressor TP53 was shown to be targeted by miR-125b
via luciferase reporter assays. An siRNA-mediated knockdown of both, TP53 and DDX58,
respectively, led to an enhanced replication of L.p. in macrophages. Thus, DDX58 and TP53
were validated as linking molecules between the three miRNAs and MX1. Additionally, the aforementioned SILAC approach revealed a downregulation of LGALS8 which was later
validated as a target of miR-579. LGALS8 is a cytosolic lectin which binds carbohydrates and
localizes to damaged vesicles. Knockdown of LGALS8 enhanced intracellular replication in
macrophages. Thus, MX1 and LGALS8 were identified as targets of the three miRNAs
(miR-125b, miR-221, miR-579) and to be responsible for the restriction of L.p. replication
within human macrophages.
(2) The transcriptional profile of L.p. during the course of infection in human macrophages was
next to be established. Dual RNA-Sequencing was performed to determine the regulation of
coding and non-coding RNA species during the course of infection of both, host and pathogen,
simultaneously. After adaptation and optimization of existing protocols, macrophages were
infected using a GFP-expressing L.p. strain Corby. To separate infected cells (gfp+) from the
non-invaded bystander cells (gfp-), flow cytometry sorting was performed. Furthermore,
Pam3CSK4 was used to generate TLR2-activated cells. RNA from all different samples, and also
RNA from cultivated Legionella, was sequenced. Differential gene expression analysis was
performed using DESeq2 resulting in 4,144 differentially expressed human genes (across
multiple conditions) and 2,707 differentially expressed Legionella genes (across two time
points). The DESeq analysis of the separated RNA fractions from host cells revealed
differentially expressed mRNAs (3,504), lncRNAs (495), and miRNAs (145). 1,128 differentially
expressed genes were exclusively significantly regulated in invaded cells (gfp+ at 8 and 16 h).
Some of these were validated via qPCR including BCL10, SOD1, IRS1, CYR61, ATG5, RND3 and
JUN. In addition, the simultaneous upregulation of the genes ZFAND2A and HSPA1 in the
bystander and in Legionella-invaded cells was validated. The analysis of the bacterial mRNAs
revealed a switch of gene usage, i.e. inverse regulation at 8 and 16 h post infection. This switch
included genes which are involved in iron metabolism, stress response, glycolysis and lipid
biosynthesis. Hence, differentially expressed genes within different growth phases of the
infection cycle were identified. This dataset is the first of its kind to cover a respiratory
pathogen. The dual RNA-Sequencing performed in this study provides data to encapsulate the
RNA landscape of coding and non-coding RNAs in pathogen and host.
In summary, the results have deepened our insight into the infection process and the
molecular interaction of L.p. and its host cells and will help to understand the complex
interplay between host and pathogen by allowing for the in silico re-construction of an RNA
interaction network. Furthermore, the present study will help to establish potential new
candidates for diagnosis and therapy.