Refine
Has Fulltext
- yes (26)
Is part of the Bibliography
- yes (26)
Year of publication
Document Type
- Doctoral Thesis (26)
Keywords
- SMN (4)
- miRNS (4)
- Genexpression (3)
- Messenger-RNS (3)
- Small nuclear RNP (3)
- Messenger-RNP (2)
- Multiproteinkomplex (2)
- RNS (2)
- SMN complex (2)
- Spleißosom (2)
Institute
- Graduate School of Life Sciences (16)
- Lehrstuhl für Biochemie (8)
- Fakultät für Biologie (3)
- Fakultät für Chemie und Pharmazie (2)
- Institut für Organische Chemie (2)
- Institut für Klinische Neurobiologie (1)
- Institut für Molekulare Infektionsbiologie (1)
- Medizinische Fakultät (1)
- Neurologische Klinik und Poliklinik (1)
- Theodor-Boveri-Institut für Biowissenschaften (1)
Sonstige beteiligte Institutionen
Complex formation between macromolecules constitutes the foundation of most cellular processes. Most known complexes are made up of two or more proteins interacting in order to build a functional entity and therefore enabling activities which
the single proteins could otherwise not fulfill. With the increasing knowledge about
noncoding RNAs (ncRNAs) it has become evident that, similar to proteins, many of
them also need to form a complex to be functional. This functionalization is usually executed by specific or global RNA-binding proteins (RBPs) that are specialized
binders of a certain class of ncRNAs. For instance, the enterobacterial global RBPs
Hfq and ProQ together bind >80 % of the known small regulatory RNAs (sRNAs),
a class of ncRNAs involved in post-transcriptional regulation of gene expression.
However, identification of RNA-protein interactions so far was performed individually by employing low-throughput biochemical methods and thereby hindered the discovery of such interactions, especially in less studied organisms such
as Gram-positive bacteria. Using gradient profiling by sequencing (Grad-seq), the
present thesis aimed to establish high-throughput, global RNA/protein complexome resources for Escherichia coli and Streptococcus pneumoniae in order to provide a
new way to investigate RNA-protein as well as protein-protein interactions in these
two important model organisms.
In E. coli, Grad-seq revealed the sedimentation profiles of 4,095 (∼85 % of
total) transcripts and 2,145 (∼49 % of total) proteins and with that reproduced
its major ribonucleoprotein particles. Detailed analysis of the in-gradient distribution of the RNA and protein content uncovered two functionally unknown
molecules—the ncRNA RyeG and the small protein YggL—to be ribosomeassociated. Characterization of RyeG revealed it to encode for a 48 aa long, toxic protein that drastically increases lag times when overexpressed. YggL was shown to
be bound by the 50S subunit of the 70S ribosome, possibly indicating involvement
of YggL in ribosome biogenesis or translation of specific mRNAs.
S. pneumoniae Grad-seq detected 2,240 (∼88 % of total) transcripts and 1,301
(∼62 % of total) proteins, whose gradient migration patterns were successfully reconstructed, and thereby represents the first RNA/protein complexome resource
of a Gram-positive organism. The dataset readily verified many conserved major
complexes for the first time in S. pneumoniae and led to the discovery of a specific
interaction between the 3’!5’ exonuclease Cbf1 and the competence-regulating ciadependent sRNAs (csRNAs). Unexpectedly, trimming of the csRNAs by Cbf1 stabilized the former, thereby promoting their inhibitory function. cbf1 was further shown
to be part of the late competence genes and as such to act as a negative regulator of
competence.
In mammals, a major fraction of the genome is transcribed as non-coding RNAs. An increasing amount of evidence has accumulated showing that non-coding RNAs play important roles both for normal cell function and in disease processes such as cancer or neurodegeneration. Interpreting the functions of non-coding RNAs and the molecular mechanisms through which they act is one of the most important challenges facing RNA biology today.
In my Ph.D. thesis, I have been investigating the role of 7SK, one of the most abundant non-coding RNAs, in the development and function of motoneurons. 7SK is a highly structured 331 nt RNA transcribed by RNA polymerase III. It forms four stem-loop (SL) structures that serve as binding sites for different proteins. Larp7 binds to SL4 and protects the 3' end from exonucleolytic degradation. SL1 serves as a binding site for HEXIM1, which recruits the pTEFb complex composed of CDK9 and cyclin T1. pTEFb has a stimulatory role for transcription and is regulated through sequestration by 7SK. More recently, a number of heterogeneous nuclear ribonucleoproteins (hnRNPs) have been identified as 7SK interactors. One of these is hnRNP R, which has been shown to have a role in motoneuron development by regulating axon growth. Taken together, 7SK’s function involves interactions with RNA binding proteins, and different RNA binding proteins interact with different regions of 7SK, such that 7SK can be considered as a hub for recruitment and release of different proteins. The questions I have addressed during my Ph.D. are as follows: 1) which region of 7SK interacts with hnRNP R, a main interactor of 7SK? 2) What effects occur in motoneurons after the protein binding sites of 7SK are abolished? 3) Are there additional 7SK binding proteins that regulate the functions of the 7SK RNP?
Using in vitro and in vivo experiments, I found that hnRNP R binds both the SL1 and SL3 region of 7SK, and also that pTEFb cannot be recruited after deleting the SL1 region but is able to bind to a 7SK mutant with deletion of SL3. In order to answer the question of how the 7SK mutations affect axon outgrowth and elongation in mouse primary motoneurons, we proceeded to conduct rescue experiments in motoneurons by using lentiviral vectors. The constructs were designed to express 7SK deletion mutants under the mouse U6 promoter and at the same time to drive expression of a 7SK shRNA from an H1 promoter for the depletion of endogenous 7SK. Using this system we found that 7SK mutants harboring deletions of either SL1 or SL3 could not rescue the axon growth defect of 7SK-depleted motoneurons suggesting that 7SK/hnRNP R complexes are integral for this process.
In order to identify novel 7SK binding proteins and investigate their functions, I proceeded to conduct pull-down experiments by using a biotinylated RNA antisense oligonucleotide that targets the U17-C33 region of 7SK thereby purifying endogenous 7SK complexes. Following mass spectrometry of purified 7SK complexes, we identified a number of novel 7SK interactors. Among these is the Smn complex. Deficiency of the Smn complex causes the motoneuron disease spinal muscular atrophy (SMA) characterized by loss of lower motoneurons in the spinal cord. Smn has previously been shown to interact with hnRNP R. Accordingly, we found Smn as part of 7SK/hnRNP R complexes. These proteomics data suggest that 7SK potentially plays important roles in different signaling pathways in addition to transcription.
Das Spleißen von prä-mRNAs stellt in der Expression eukaryontischer Gene einen essentiellen Reifungsschritt dar. Erst durch das exakte Entfernen von nicht-kodierenden Introns und Verbinden der kodierenden Exons kann die genetische Information am Ribosom in funktionelle Proteine umgesetzt werden. Spleißen wird durch das Spleißosom katalysiert, welches sich aus den small nuclear ribonucleoproteins (snRNPs) U1, U2, U4, U5 und U6 und einer großen Anzahl weiterer Proteinfaktoren zusammensetzt. Die snRNPs bestehen aus einer Uridin-reichen snRNA, spezifischen und generellen (Sm-)Proteinen. Die Sm-Proteine B/B`, D1, D2, D3, E, F, und G bilden einen heptameren Ring um die sog. Sm-Bindungsstelle der snRNAs. Während die Zusammenlagerung von Sm-Proteinen mit der RNA in vitro spontan ablaufen kann, wird dieser Prozess in vivo von zwei makromolekularen Proteinkomplexen assistiert, die als PRMT5- bzw. SMN-Komplex bezeichnet werden. Der PRMT5-Komplex (bestehend aus PRMT5, WD45 und pICln) agiert in der frühen Phase der Zusammenlagerung. Seine Hauptfunktion ist die symmetrische Dimethylierung der Sm-Proteine und die Stabilisierung von Sm-Proteinkomplexen durch das Chaperon pICln in zwei Intermediaten. Einhergehend mit dieser Aktivität werden auch Aggregation bzw. unspezifische Wechselwirkungen der Sm-Proteine mit RNAs verhindert. In der späten Phase der Zusammenlagerung löst der SMN-Komplex (bestehend aus SMN, Gemin2-8 und unrip) pICln-Intermediate auf, wobei dieser die Sm-Proteine en bloc übernimmt und sie auf die snRNA überträgt. Während dieser Reaktion wird pICln aus den Komplexen verdrängt. Ein Fehlen des SMN-Proteins, einer Schlüsselkomponente des SMN-Komplexes, führt zur autosomal rezessiven Erbkrankheit `Spinale Muskelatrophie` (SMA) wobei der Schweregrad der Krankheit invers mit der Menge an funktionellem SMN-Protein korreliert. Es wird vermutet, dass eine gestörte snRNP-Biogenese die Ursache der SMA ist.
In der vorliegenden Arbeit sollte die U snRNP-Zusammenlagerungsmaschinerie aus rekombinanten Bausteinen rekonstituiert werden und so funktionellen und strukturellen Studien zugänglich gemacht werden. Folgende Resultate wurden in dieser Arbeit erhalten:
1) Im ersten Teil der Arbeit wurde eine experimentelle Strategie etabliert, welche die Rekonstitution des humanen SMN-Komplexes aus rekombinanten Untereinheiten erlaubte. Entscheidend hierfür war die Definition von Subkomplexen aufgrund einer Protein-Interaktionskarte. Die Subkomplexe konnten separat hergestellt und anschließend zum Gesamtkomplex vereinigt werden.
2) Die erfolgreiche Etablierung eines rekonstitutiven Systems erlaubte eine detaillierte biochemische Charakterisierung des SMN-Komplexes. Es konnte gezeigt werden, dass der rekombinante Komplex alle für die Biogenese von U snRNPs nötigen Schritte bewerkstelligen konnte. Dies schließt sowohl die Übernahme der Sm-Proteine aus den pICln-Intermediaten als auch das Verdrängen des Chaperons pICln und die Übertragung der Sm-Proteine auf die snRNAs ein.
3) Durch die Reduzierung des SMN-Gesamtkomplexes um Gemin3-5 auf einen SMN-Pentamer konnte dieser als ein funktioneller Kernbereich identifiziert werden, der die einzelnen Schritte der U snRNP-Biogenese vergleichbar mit dem gesamten Komplex bewerkstelligen konnte. Zudem agierte dieser reduzierte Komplex als notwendiger und ausreichender Spezifitätsfaktor der RNP-Zusammenlagerung.
4) Das rekombinante System ermöglichte erstmals SMN-Komplexe mit SMA-pathogenen Mutationen herzustellen und einer eingehenden funktionellen und strukturellen Untersuchung zu unterziehen. Die detaillierte Analyse der SMA-verursachenden Punktmutation SMN(E134K) offenbarte spezifische Defekte im Zusammenlagerungsprozess und damit Einblicke in die Pathophysiologie der Krankheit.
Mit der im Rahmen dieser Arbeit etablierten Rekonstitution des rekombinanten SMN-Komplexes wurde die Grundlage für die detaillierte biochemische und strukturbiologische Untersuchung der Zusammenlagerungsmaschinerie spleißosomaler U snRNPs gelegt. Dieses experimentelle System wird auch bei der Aufdeckung der biochemischen Defekte hilfreich sein, die zur neuromuskulären Krankheit SMA führen.
The transcription factor MYC is a onco-protein, found to be deregulated in many human cancers. High MYC levels correlate with an aggressive tumor outcome and poor survival rates. Despite MYC being discovered as an oncogene already in the 1970s, how MYC regulates transcription of its target genes, which are involved in cellular growth and proliferation, is not fully understood yet.
In this study, the question how MYC influences factors interacting with the RNA polymerase II ensuring productive transcription of its target genes was addressed using quantitative mass spectrometry. By comparing the interactome of RNA polymerase II under varying MYC levels, several potential factors involved in transcriptional elongation were identified. Furthermore, the question which of those factors interact with MYC was answered by employing quantitative mass spectrometry of MYC itself. Thereby, the direct interaction of MYC with the transcription elongation factor SPT5, a subunit of the DRB-sensitivity inducing factor, was discovered and analyzed in greater detail. SPT5 was shown to be recruited to chromatin by MYC. In addition, the interaction site of MYC on SPT5 was narrowed down to its evolutionary conserved NGN-domain, which is the known binding site for SPT4, the earlier characterized second subunit of the DRB-sensitivity inducing factor. This finding suggests a model in which MYC and SPT4 compete for binding the NGN-domain of SPT5.
Investigations of the SPT5-interacting region on MYC showed binding of SPT5 to MYC’s N-terminus including MYC-boxes 0, I and II.
In order to analyze proteins interacting specifically with the N-terminal region of MYC, a truncated MYC-mutant was used for quantitative mass spectrometric analysis uncovering reduced binding for several proteins including the well-known interactor TRRAP and TRRAP-associated complexes.
Summarized, ...
Cellular proteome profiling revealed that most biomolecules do not exist in isolation, but rather are incorporated into modular complexes. These assembled complexes are usually very large, consisting of 10 subunits on an average and include either proteins alone, or proteins and nucleic acids. Consequently, such macromolecular assemblies rather than individual biopolymers perform the vast majority of cellular activities. The faithful assembly of such molecular assemblies is often aided by trans-acting factors in vivo, to preclude aggregation of complex components and/or non-cognate interactions. A paradigm for an assisted assembly of a macromolecular machine is the formation of the common Sm/LSm core of spliceosomal and histone-mRNA processing U snRNPs. The key assembly factors united in the Protein Arginine Methyltransferase 5 (PRMT5) and the Survival Motor Neuron (SMN) complexes orchestrate the assembly of the Sm/LSm core on the U snRNAs. Assembly is initiated by the PRMT5-complex subunit pICln, which pre-arranges the Sm/LSm proteins into spatial positions occupied in the mature U snRNPs. The SMN complex subsequently binds these Sm/LSm units, displaces pICln and catalyses the Sm ring closure on the Sm-site of the U snRNA.
The SMN complex consists of the eponoymous SMN protein linked in a modular network of interactions with eight other proteins, termed Gemins 2-8 and Unrip. Despite functional and structural characterisation of individual protein components and/or sub-complexes of this assembly machinery, coherent understanding of the structural framework of the core SMN complex remained elusive. The current work, employing a combined approach of biochemical and structural studies, aimed to contribute to the understanding of how distinct modules within the SMN complex coalecse to form the macromolecular SMN complex.
A novel atomic resolution (1.5 Å) structure of the human Gemin8:7:6 sub-complex, illustrates how the peripheral Gemin7:6 module is tethered to the SMN complex via Gemin8’s C-terminus. In this model, Gemin7 engages with both Gemin6 and Gemin8 via the N- and C-termini of its Sm-fold like domain. This highly conserved interaction mode is reflected in the pronounced sequence conservation and identical biochemical behaviour of similar sub-complexes from divergent species, namely S. pombe and C. elegans.
Despite lacking significant sequence similarity to the Sm proteins, the dimeric Gemin7:6 complex share structural resemblance to the Sm heteromers. The hypothesis that the dimeric Gemin7:6 functions as a Sm-surrogate during Sm core assembly could not be confirmed in this work. The functional relevance of the structural mimicry of the dimeric Gemin7:6 sub-complex with the Sm heterodimers therefore still remains unclear.
Reduced levels of functional SMN protein is the cause of the devastating neurodegenerative disease, Spinal Muscular Atrophy (SMA). The C-terminal YG-zipper motif of SMN is a major hot-spot for most SMA patient mutations. In this work, adding to the existing inventory of the human and fission yeast YG-box models, a novel 2.2 Å crystal structure of the nematode SMN’s YG-box domain adopting the glycine zipper motif has been reported. Furthermore, it could be assessed that SMA patient mutations mapping to this YG-box domain greatly influences SMN’s self-association competency, a property reflected in both the human and nematode YG-box biochemical handles. The shared molecular architecture and biochemical behaviour of the nematode SMN YG-box domain with its human and fission yeast counterparts, reiterates the pronounced conservation of this oligomerisation motif across divergent organisms.
Apart from serving as a multimerization domain, SMN’s YG-box also acts as interaction platform for Gemin8. A systematic investigation of SMA causing missense mutations uncovered that Gemin8’s incorporation into the SMN complex is influenced by the presence of certain SMA patient mutations, albeit independent of SMN’s oligomerisation status. Consequently, loss of Gemin8 association in the presence of SMA patient mutations would also affect the incorporation of Gemin7:6 sub-complex. Gemin8, therefore sculpts the heteromeric SMN complex by bridging the Gemin7:6 and SMN:Gemin2 sub-units, a modular feature shared in both the human and nematode SMN complexes.
These findings provide an important foundation and a prospective structural framework for elucidating the core architecture of the SMN complex in the ongoing Cryo-EM studies.
The expression of genetic information into proteins is a key aspect of life. The efficient and exact regulation of this process is essential for the cell to produce the correct amounts of these effector molecules to a given situation. For this purpose, eukaryotic cells have developed many different levels of transcriptional and posttranscriptional gene regulation. These mechanisms themselves heavily rely on interactions of proteins with associated nucleic acids. In the case of posttranscriptional gene regulation an orchestrated interplay between RNA-binding proteins, messenger RNAs (mRNA), and non-coding RNAs is compulsory to achieve this important function.
A pivotal factor hereby are RNA secondary structures. One of the most stable and diverse representatives is the G-quadruplex structure (G4) implicated in many cellular mechanisms, such as mRNA processing and translation. In protein biosynthesis, G4s often act as obstacles but can also assist in this process. However, their presence has to be tightly regulated, a task which is often fulfilled by helicases.
One of the best characterized G4-resolving factors is the DEAH-box protein DHX36. The in vitro function of this helicase is extensively described and individual reports aimed to address diverse cellular functions as well. Nevertheless, a comprehensive and systems-wide study on the function of this specific helicase was missing, so far.
The here-presented doctoral thesis provides a detailed view on the global cellular function of DHX36. The binding sites of this helicase were defined in a transcriptome-wide manner, a consensus binding motif was deviated, and RNA targets as well as the effect this helicase exerts on them were examined. In human embryonic kidney cells, DHX36 is a mainly cytoplasmic protein preferentially binding to G-rich and G4-forming sequence motifs on more than 4,500 mRNAs. Loss of DHX36 leads to increased target mRNA levels whereas ribosome occupancy on and protein output of these transcripts are reduced. Furthermore, DHX36 knockout leads to higher RNA G4 levels and concomitant stress reactions in the cell. I hypothesize that, upon loss of this helicase, translationally-incompetent structured DHX36 target mRNAs, prone to localize in stress granules, accumulate in the cell. The cell reacts with basal stress to avoid cytotoxic effects produced by these mis-regulated and structured transcripts.
MicroRNAs are endogenous ≈22 nt long non coding RNA molecules that modulate gene expression
at the post transcriptional level by targeting mRNAs for cleavage or translational repression.
MicroRNA-mRNA interaction involves a contiguous and perfect pairing within complementary
sites usually in the 3’ UTR of the target mRNA. Heart failure is associated with myocyte
hypertrophy and death, due to compensatory pathological remodeling and minimal functional repair
along with microRNA deregulation.
In this study, we identified candidate microRNAs based on their expression strength in
cardiomyocytes and by their ability to regulate hypertrophy. Expression profiling from early and
late stages of heart failure showed several deregulated microRNAs. Of these microRNAs, miR-378
emerged as a potentially interesting microRNA that was highly expressed in the mouse heart and
downregulated in the failing heart. Antihypertrophic activity of miR-378 was first observed by
screening a synthetic miR library for morphologic effects on cardiomyocytes, and validated in vitro proving the tight control of hypertrophy by this miR. We combined bioinformatic target prediction analysis and microarray analysis to identify the targets of miR-378. These analyses suggested that factors of the MAP kinase pathway were enriched among miR-378 targets, namely MAPK1 itself (also termed ERK2), the insulin-like growth factor receptor 1 (IGF1R), growth factor receptor bound protein 2 (GRB2) and kinase suppressor of ras 1 (KSR1). Regulation of these targets by miR-378 was then confirmed by mRNA and protein expression analysis. The use of luciferase reporter constructs with natural or mutated miR-378 binding sites further validated these four proteins as direct targets of miR-378. RNA interference with MAPK1 and the other three targets prevented the prohypertrophic effect of antimiR-378, suggesting that they functionally relate to miR-378. In vivo restoration of disease induced loss of miR-378 in a pressure overload mouse model of hypertrophy using adeno associated virus resulted in partial attenuation cardiac hypertrophy and significant improvement in cardiac function along with reduced expression of the four targets in heart.
We conclude from these findings that miR-378 is an antihypertrophic microRNA in cardiomyocytes, and the main mechanism underlying this effect is the suppression of the MAP kinase-signaling pathway on four distinct levels. Restoration of disease-associated loss of miR-378 through cardiomyocyte-targeted AAV-miR-378 may prove as an effective therapeutic strategy in myocardial disease.
Galectin-1 (hGal-1) is overexpressed by numerous cancer types and previously conducted studies confirmed that the β-galactoside-binding protein mediates various molecular interactions associated with tumor growth, spread and survival. Upon interaction with carbohydrate-based binding epitopes of glycan structures on human cell surfaces galectin-1 induces proliferative, angiogenetic and migratory signals and modulates negative T cell regulation which essentially helps the tumor to evade the immune response. These findings attributed galectin-1 a pivotal role in tumor physiology and strongly suggest the protein as target for diagnostic and therapeutic applications.
Within the scope of this work a strategy was elaborated for designing tailor-made galectin-1 ligands by functionalizing selected hydroxyl groups of the natural binding partner N-acetyllactosamine (LacNAc) that are not involved in the sophisticated interplay between the disaccharide and the protein. Synthetic modifications intended to introduce chemical groups i) to address a potential binding site adjacent to the carbohydrate recognition domain (CRD) with extended hGal-1-ligand interactions, ii) to implement a tracer isotope for diagnostic detection and iii) to install a linker unit for immobilization on microarrays.
Resulting structures were investigated regarding their targeting ability towards galectin-1 by cocrystallization experiments, SPR and ITC studies. Potent binders were further probed for their diagnostic potential to trace elevated galectin-1 levels in microarray experiments and for an application in positron emission tomography (PET).
Eukaryotische messenger-RNAs (mRNAs) müssen diverse Prozessierungsreaktionen durchlaufen, bevor sie der Translationsmaschinerie als Template für die Proteinbiosynthese dienen können. Diese Reaktionen beginnen bereits kotranskriptionell und schließen das Capping, das Spleißen und die Polyadenylierung ein. Erst nach dem die Prozessierung abschlossen ist, kann die reife mRNA ins Zytoplasma transportiert und translatiert werden. mRNAs interagieren in jeder Phase ihres Metabolismus mit verschiedenen trans-agierenden Faktoren und bilden mRNA-Ribonukleoproteinkomplexe (mRNPs) aus. Dieser „mRNP-Code“ bestimmt das Schicksal jeder mRNA und reguliert dadurch die Genexpression auf posttranskriptioneller Ebene.
Für das La-verwandte Protein LARP4B (La-related protein 4B) wurde kürzlich eine direkte Interaktion mit den Translationsfaktoren PABPC1 (poly(A) binding protein, cytoplasmic 1) und RACK1 (receptor for activated C kinase) gefunden. Diese Befunde sowie die Assoziation mit aktiv translatierenden Ribosomen lässt vermuten, dass LARP4B zum mRNP-Code beiträgt. Die Domänenstruktur des Proteins legt darüber hinaus nahe, dass LARP4B direkt mRNAs bindet.
Um einen Einblick in die Funktion von LARP4B und seiner in vivo RNA-Bindungspartner zu erhalten, wurde die mRNA-Assoziation transkriptomweit mit Hilfe von PAR-CLIP (Photoactivatable-Ribonucleoside-Enhanced Crosslinking and Immunoprecipitation)-Experimenten bestimmt. Diese Daten zeigten, dass LARP4B ein spezifisches Set an zellulären mRNAs über Sequenzbereiche in deren 3’-untranslatierten Regionen bindet. Die bioinformatische Auswertung der PAR-CLIP-Daten identifizierte ein LARP4B-Bindemotiv, welches durch in vitro Bindungsstudien validiert werden konnte. Darüber hinaus belegten pSILAC (pulsed stable isotope labeling with amino acids in cell culture)-Experimente und eine transkriptomweite Analyse der mRNA-Level, dass LARP4B die Expression der Ziel-mRNAs beeinflusst, indem es die Stabilität der gebundenen Transkripte erhöht. LARP4B konnte somit als positiver Faktor der eukaryotischen Genexpression identifiziert werden.
While TGF-β is able to regulate miRNA expression in numerous cell types, TGF-β-dependent changes in the miRNA profile of CD8+ T cells had not been studied before. Considering that TGF-β suppresses CD8+ T cell effector functions in numerous ways, we wondered whether induction of immune-regulatory miRNAs could add to the known transcriptional effects of TGF-β on immune effector molecules. In this study, we used miRNA arrays, deep sequencing and qRT-PCR to identify miRNAs that are modulated by TGF-β in human CD8+ T cells. Having found that the TGF-β-dependent downregulation of NKG2D surface expression in NK cells and CD8+ T cells does not go along with a corresponding reduction in mRNA levels, this pathway appeared to be a possible target of TGF-β-inducible miRNAs. However, this hypothesis could not be confirmed by miRNA reporter assays. Instead, we observed that DAP10 transcription is suppressed by TGF-β which in turn negatively affects NKG2D surface expression. In spite of promising preliminary experiments, technical difficulties associated with the transfection of primary NK cells and NK cell lines unfortunately precluded the final proof of this hypothesis.
Instead, we focused on the TGF-β-induced changes in the miRNome of CD8+ T cells and confirmed the induction of the miR-23a cluster members, namely miR-23a, miR-27a and miR-24 by three different techniques. Searching for potential targets of these miRNAs which could contribute to the immunosuppressive action of TGF-β in T cells, we identified and confirmed a previously unknown regulation of IFN-γ mRNA by miR-27a and miR-24. Newly generated miRNA reporter constructs further revealed that LAMP1 mRNA is a target of miR-23a. Upon modulation of the miR-23a cluster in CD8+ T cells by the respective miRNA antagomirs and mimics, significant changes in IFN-γ expression confirmed the functional relevance of our findings. Effects on CD107a/LAMP1 expression were, in contrast, rather minimal. Still, overexpression of the miR-23a cluster attenuated the cytotoxic activity of antigen-specific CD8+ T cells. Taken together, these functional data reveal that the miR-23a cluster not only is induced by TGF-β, but also exerts a suppressive effect on CD8+ T-cell effector functions, even in the absence of TGF-β signaling.