MHC Class III RNA Binding Proteins and Immunity

ABSTRACT

Here we review data suggestive of a role for RNA-binding proteins in vertebrate immunity. We focus on the products of genes found in the class III region of the Major Histocompatibility Complex. Six of these genes, DDX39B (aka BAT1), DXO, LSM2, NELFE, PRRC2A (aka BAT2), and SKIV2L, encode RNA-binding proteins with clear roles in post-transcriptional gene regulation and RNA surveillance. These genes are likely to have important functions in immunity and are associated with autoimmune diseases.

KEYWORDS:

• RNA-binding proteins

Introduction

In the last few decades we have learned remarkable new facts about RNAs – they are not just messengers of genetic information or important viral genomes, but can be enzymes (e.g., ribosomes and telomerases [1]), critical structural components of macromolecular machines (e.g., phage motors [2]) and important trans-acting regulators of gene expression (e.g., microRNAs (miRNAs) [3]). Most importantly, we have come to recognize that RNA was the original molecule of life and this realization has important implications. A salient and fascinating consequence of this ancestry is the role that RNA plays in immunity. It follows then that the first attack of one genome (pathogen) on another (host) represented the first infections. We argue these invasions are now recapitulated daily (especially in these difficult days of COVID-19) when an RNA virus infects a human, a mosquito or a plant. The argument can be extended by concluding that the defences of the invaded RNA genome represent the earliest forms of immunity. Thus, we postulate that RNA genomes were the first to define ‘self’ and ‘non-self’: the basis of immunity. We posit that today we can discern this RNA ancestry in many extant immune systems and thus propose, as have others [4–6], exceptional roles for RNAs and RNA-binding proteins (RBPs) in immune systems.

Immunity has been arbitrarily divided into, innate immunity and acquired immunity, which are defined by the fact that the latter involves responses driven by memory of a specific pathogen. Although this classification is being re-examined [7,8], it serves as a useful scaffold to discuss the role of RBPs. Innate immune mechanisms have been associated with critical roles for RNA and RBPs [9]. RNA interference (RNAi), which is mediated by short double-stranded RNAs (dsRNAs) called small interfering RNAs (siRNAs), is a potent innate immune mechanism in many animals, particularly invertebrates, and in plants [10]. In vertebrates, the interferon system is a major innate immune pathway that is highly regulated by dsRNAs and involves a large number of proteins, many of which are RBPs [11]. Amongst them, RIG-I and MDA5 are pattern recognition receptors that mediate the IFN response by recognizing pathogen-associated molecular patterns (PAMPs) on foreign nucleic acids [12]. 2ʹ-5ʹ oligoadenylate synthetases (OAS) are effectors of this pathway and they bind to and are activated by these PAMPS [13]. OAS synthesize atypical RNA oligoadenylates, with 2ʹ-5ʹ linkages, that activate ribonuclease L (RNAse L), which degrades viral RNAs [14]. Even the innate sensor of foreign DNA, the cyclic GMP-AMP synthase (cGAS), is homologous to OAS and thus likely evolved from a family of RBPs [15].

A special role for RBPs in acquired immunity is evident in the CRISPR system in prokaryotes where guide RNAs derived from sequences of previously encountered pathogens direct endonucleases to cleave the genomes of the same or related pathogens upon reinfection [16,17]. Elegant studies of immunity in mammals have demonstrated that there are robust regulatory programmes of post-transcriptional control in the development, maturation, and activation of immune cells [6,18]. For instance, it is clear that alternative splicing factors (e.g., CELF2) control important immune gene expression networks required for the action of T cells [19,20]. Equally, one can point to interesting examples of RBPs regulating mRNA turnover and/or translation of short-lived cytokine mRNAs [21]. Indeed, these roles are expected for a system as complex as the adaptive immune system. In this review, we focus on a group of RBPs that provide further evidence for a special role of these proteins in both innate and adaptive immunity.

The intriguing case of RNA-binding protein genes in the major histocompatibility complex

A clue to a special role for RBPs in immunity came from studies of the DExD-Box Helicase 39B (DDX39B), also known to immunologists as HLA-B associated transcript-1 protein (BAT1) and to RNA biologists as U2AF associated protein 56kD (UAP56). Early work from immunologists indicated that DDX39B promoted gene expression of anti-inflammatory pathways [22]. RNA biologists showed that DDX39B plays roles in constitutive pre-mRNA splicing, and subsequently in alternative splicing and transport of both mRNAs and circular RNAs (see references below). The gene encoding DDX39B resides in the class III region of the major histocompatibility complex (MHC) locus in humans and this location is conserved in all vertebrates (Fig. 1).

Figure 1. MHC class III genes encoding RNA-binding proteins. A UCSC genome viewer schematic of the chromosome 6 region spanning the MHC class III region. Genes encoding proteins implicated in RNA metabolism are boxed in red. Genes encoding proteins with a potential role in RNA metabolism and mentioned in the text are boxed in green. Genes indicated in black correspond to proteins with an entry in the Protein Data Bank (PDB); genes indicated in dark blue represent transcripts that have been reviewed or validated by either the Reference Sequence collection (RefSeq), SWISS-PROT or the Consensus Coding Sequence Project (CCDS); genes indicated in light blue correspond to non-RefSeq transcripts

The MHC is a large genetic locus first discovered because of its role in determining tissue and organ compatibility in transplantation experiments in mice and now known to coordinate a large number of immune processes [23]. The locus, which in humans is located in the short arm of chromosome 6, has been traditionally divided into three regions based on the classes of genes that predominate in these. MHC class I genes encode ubiquitously expressed proteins involved in antigen presentation, and MHC class II genes encode proteins engaged in antigen presentation by so-called professional antigen-presenting cells (APCs) (e.g., macrophages). The class III region, which has very high gene density [24], lies between regions I and II, and its genes are not easy to classify based on their known or presumed functions. Among class III genes some are clearly immune modulators such as components of the complement system (e.g., C4), cytokines (e.g., TNF-a) or regulators of gene expression implicated in immune function (e.g., NFKBL1) [23] (Fig. 1). Other MHC class III genes have unknown function or functions that have not been directly associated with immunity. The conservation of synteny in the MHC class III region in all jawed vertebrates suggests that many of these genes of unknown function will turn out to play important roles in immunity. As mentioned above and described in more detail below this is the case for DDX39B, an RNA helicase that promotes immune tolerance.

DDX39B could have been considered an exceptional case of an RBP encoded in the class III region, but an interesting report from Sanderson and colleagues proposed that a large number of MHC class III genes encode proteins involved in RNA transactions [25]. A similar suggestion was later made by Yu and colleagues noting that four MHC class III genes, NELFE, SKIV2L, DXO, and STK19, are involved in RNA surveillance, which could reflect a role in recognition of foreign RNA [26]. A review of published observations suggests that six class III gene products, DDX39B (BAT1), DXO, LSM2, NELFE, PRRC2A (BAT2), and SKIV2L are RBPs and play important roles in RNA metabolism, and four, ABHD16A (BAT5), CLIC1, PBX2, and STK19 have connections to RNA biology. This means that out of the 60 expressed protein-coding genes in the human MHC class III region [24] between 10% and 16% are associated with RNA metabolism. This is higher than the calculated frequency of proteins engaged in RNA metabolism, which has been calculated to be between 3% and 11% of the proteome in a broad survey of organisms, with complex organisms like humans on the lower end of the range [27]. Here we review what is known about these MHC class III gene products in the context of RNA biology and immunology.

DExD-Box Helicase 39B (BAT1)

In 1989, DExD-Box Helicase 39B (DDX39B) was identified as the protein product of the MHC class III HLA-B-associated transcript −1 [28], and in 1995 it was predicted to be a member of the ancestral DEAD-box family of RNA helicases [29]. Early studies suggested a genetic association between DDX39B and several autoimmune disorders [28,30], and functional studies where DDX39B expression was reduced using antisense oligonucleotides suggested that it was a negative regulator of inflammation [22]. In parallel studies published in 1997, DDX39B was identified as a constitutive splicing factor that interacted with U2AF2 and named 56 kDa U2AF2 associated protein (UAP56) [31]. Several studies extended the roles of DDX39B in RNA metabolism to include nucleocytoplasmic transport of mRNAs [32–34] and circular RNAs (circRNAs) [35], alternative splicing [36,37], and resolution of R-loops [38]. It should be noted that DDX39A is a paralog of DDX39B and the two have overlapping, but not completely redundant, functions [35,37,39].

Our work provides a connection between the RNA biology and immunology roles of DDX39B. DDX39B activates splicing of interleukin-7 receptor-a (IL7R) exon 6 and reduces levels of the soluble form of IL7R (sIL7R) [36]. High levels of sIL7R are associated with increased risk of multiple sclerosis (MS) [40,41] and other autoimmune diseases [42,43], and sIL7R has been shown to enhance the severity of experimental autoimmune encephalomyelitis, an animal model of MS [41]. Furthermore, a strong genetic association between alleles of DDX39B that lead to high expression of the protein and protection from MS-risk and there is strong genetic and functional epistasis between DDX39B and IL7R. Additionally, DDX39B is required for the expression of FOXP3, which is the master transcriptional controller of T regulatory cells (Tregs) (Hirano, Galarza-Muñoz et al., manuscript in preparation). FOXP3 introns have cytidine-rich polypyrimidine tracts and these tracts make their splicing exquisitely sensitive to DDX39B (Ibid). These reports support a critical role for DDX39B as a guardian of immune tolerance (Fig. 2).

Figure 2. DDX39B, a guardian of tolerance. (Left) DDX39B is a potent activator of IL7R exon 6, which leads to lower levels of sIL7R. Since sIL7R promotes autoimmunity, the effect of DDX39B is to decrease the tendency towards autoimmunity. (Centre) DDX39B mediates nucleocytoplasmic transport of circular RNAs, which repress innate immune and autoimmune reactions. (Right) DDX39B promotes expression of Forkhead Box P3 (FOXP3) by enhancing the splicing of weak introns. FOXP3 is a master regulator of the development and function of CD4⁺/CD25⁺ T regulatory (T reg) cells, which downregulate autoimmunity. Arrows indicate activation and bar-headed lines indicate repression, and green arrows or bar-headed lines indicate processes that promote tolerance and red arrows indicate processes that promote autoimmunity

Other functions of DDX39B are also connected to immune function. In human cells DDX39B has been shown to control the nucleocytoplasmic transport of circRNAs [35]. A role for circRNAs in innate immunity is now recognized [44,45]. CircRNAs induced innate immune gene expression in a RIG-I dependent fashion and this induction depended on the origin of the circRNAs [46]. CircRNAs bind and inhibit the dsRNA dependent kinase, PKR (EIF2AK2), and patients with systemic lupus erythematosus (SLE) show lower levels of circRNAs in their PBMCs and increased PKR activity [47]. Circular RNAs have been shown to modulate the function and activation of innate immune cells. For example, binding of circANRIL (Circular antisense non-coding RNA in the INK4 locus) to pescadillo ribosomal biogenesis factor 1 (PES1) prevents the uncontrolled proliferation of vascular tissue macrophages and may protect against atherosclerosis [48]. Given that DDX39B mediates nucleocytoplasmic transport of circRNAs and that circRNAs can downregulate immune responses, we speculate that DDX39B can promote tolerance by maintaining robust levels of circRNAs in the cytoplasm.

Additional roles for DDX39B in immunity are suggested by the fact that this helicase has been shown to resolve R loops [38], which are believed to be a source of cytoplasmic DNA that is sensed by cGAS and induces innate immune responses [49]. Another intriguing connection is the requirement for the DDX39B homologue in Oryza sativa (Asian rice) in the stress response [50]. All of these studies together suggest an ancient and important role of DDX39B in genome protection and a modern role in immunity.

Proline-rich coiled-coil protein 2A (BAT2)

The proline-rich coiled-coil protein 2A (PRRC2A, aka BAT2) [51] was shown to interact with hnRNP A1 and the eukaryotic translation initiation factor-3 suggesting it could have roles in RNA metabolism [25]. More recently, PRRC2A has been shown to be a reader for the common epitranscriptome mark N⁶-methyladenosine (m⁶A) in mRNAs and to regulate the stability of the Olig2 RNA [52]. PRRC2A has been associated with lupus nephritis [53], coeliac disease [54], and type 1 diabetes [55]; however, these genetic linkages cannot be formally sorted away from those of other MHC class III genes.

Decapping exoribonuclease

The decapping exoribonuclease (DXO, aka DOM3Z) is involved in removing the 5ʹcap of RNAs synthesized by RNA polymerase II (RNAPII) (decapping) and particularly removal of the alternative nicotinamide adenine dinucleotide (NAD+) cap (deNADding) [56] and other non-canonical cap structures [57]. Remarkably DXO also has 5ʹ-to-3ʹ exoribonuclease activity and therefore can both decap and degrade RNAs [58]. DXO plays an antiviral role in Hepatitis C virus (HCV) infected cells, cleaving HCV genomes not protected at the 5ʹend by miR-122 [59]. This suggests that DXO may be able to sense viral RNAs as foreign.

Like Sm protein-2

The Like Sm protein-2 (LSM2) assembles into ring-like heteroheptamers with other Sm-like proteins [60]. The cytoplasmic LSM1-7 complex mediates RNA decay and the nuclear LSM2-8 interacts with U6 snRNA and is involved in pre-mRNA splicing [61]. As part of the LSM complexes, LSM2 plays roles in RNA surveillance and regulates decay of AU-rich elements (AREs) containing mRNAs, which encode many genes involved in immunity and inflammation [62]. As described above for DDX39B, LSM rings also play important roles in regulating the stress response in plants [63].

SKI2-like RNA helicase

The SKI2-like RNA helicase (SKIV2L, aka as SKI2W) is named because of its homology to a Saccharomyces cerevisiae nucleolar factor that controls the replication of RNA viruses of yeast [64]. A similar antiviral role for SKIV2L was observed in human cells infected with hepatitis B virus (HBV); SKIV2L binds the HBV X-mRNA and mediates it degradation [65]. SKIV2Lis a subunit of the human RNA exosome, a complex that degrades unstable RNAs, including ARE-containing RNAs [66], is involved in translational surveillance [67], can have antiviral activity [68], and is a negative regulator of the RLR-mediated antiviral response (RIG-I like receptor) [69]. Mutations in SKIV2L have been associated with Trichohepatoenteric syndrome-2 (THES2), which is characterized by intrauterine growth retardation, facial dysmorphism, hair abnormalities, intractable diarrhoea, and combined immunodeficiency [70,71]. SKIV2L has been connected to innate immunity since depletion of SKIV2L in cells in culture enhances the interferon (IFN) response to ligands of RIG-I and patients with THES2 have an overactive blood IFN gene expression signature [69]. Further, SKIV2L has been associated with early onset of presumed monogenic inflammatory bowel disease (IBD) as well as SLE [72–74].

Negative elongation factor, member E

Negative elongation factor, member E (NELFE) was also named RD RNA-binding protein because it contains an arginine and aspartic acid dipeptide repeat also found in the U1 snRNP 70-kDa protein and an RNP consensus motif found in many RBPs [75]. Indeed, subsequent evidence supports that NELFE binds RNA [76–78]. NELFE is the smallest subunit (member) of the NELF complex, which is involved in the regulation of transcription elongation by RNAPII [79]. NELF interacts with the nuclear cap-binding complex (CBC), which binds the 5ʹ cap and is important for many RNA processing events, and this interaction is mediated by binding of NELFE to the 80 kDa CBC subunit [80]. NELFE has been proposed as a sarcoidosis susceptibility gene [81] and may be critical for antiviral immunity in insect cells [82].

Other MHC class III proteins with potential connections to RNA metabolism

In addition to the MHC class III RBPs described above, there are several proteins encoded by genes in the class III region that have been connected, more or less definitively, to RNA metabolism. Abhydrolase domain-containing protein 16A (ABHD16A, aka BAT5) was shown to interact with hnRNP A1 [25], but there is not further evidence to suggest that it plays a role in RNA metabolism. The chloride intracellular channel 1 (CLIC1) binds the Sm-like protein LSM1 [25], which is found in Processing bodies (P-bodies) [83], but as with ABHD16A, CLIC1 has not been connected further to RNA metabolism. The serine/threonine kinase-19 (STK19) interacts with SF3B subunit 4 [25], a splicing factor, and although it has been proposed to be involved in DNA repair, this function is vigorously debated [84,85]. A role for STK19 in the synthesis or processing of RNAs has been suggested [26], but is not well documented. The class III region including NELFE, SKIV2L, and STK19 was shown to genetically associate with metabolic syndrome linked inflammation [86]. Finally, the pre-B cell leukaemia transcription factor-2 (PBX2), which contains a homeobox domain, was shown to interact with the nucleolar MTR4 exosome RNA helicase (MTREX) [25], which is a homologue of SKIV2L [87]. Although the aforementioned connection and the fact that PBX2 is expressed at high levels in murine bone marrow and thymus suggest a potential immune function, PBX2^-/- mice have normal hematopoeisis and immune function [88]. Therefore, for ABHD16A, CLIC1, STK19, and PBX2, the connections to RNA metabolism and/or immune function remain tantalizing but unproven.

Conclusions

In this brief review we highlight the intriguing connections between RNA-binding proteins encoded by genes in the major histocompatibility complex and immunity. We conjecture that immune functions of these RNA-binding proteins are fossils of ancestral immune mechanisms in the RNA world.

Acknowledgments

We thank many colleagues at UTMB for discussions about the ideas presented here, in particular Eric J. Wagner. We acknowledge support from NIH R01 CA204806, Mildred Hajek Vacek and John Roman Vacek Distinguished Chair, and UTMB startup package (MGB).

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the National Cancer Institute [R01 CA204806].