Introduction

The genus Streptococcus, comprising a group of commensal and pathogenic low GC Gram-positive bacteria, typically has spherical or ovoid cells with homofermentative, anaerobic/aerotolerant, non-motile, catalase negative, and non-spore forming characteristics [1]. More than 55 species of Streptococcus are found in a wide variety of habitats including the mouth, respiratory tract, and skin surfaces of animals and humans, as well as in various environments such as soil and plants [2]. According to its hemolytic properties, Streptococcus is divided into three categories: (I) α-hemolytic Streptococcus e.g., S. mutans; (II) β-hemolytic Streptococcus, mainly including S. pyogenes (also called group A Streptococcus or GAS) [3] and S. agalactiae; and (III) γ-hemolytic Streptococcus, which are also called non-hemolytic Streptococcus and are generally not pathogenic [1]. Some species of Streptococcus significantly affect both humans and animals. For example, S. pyogenes, S. pneumoniae, and S. agalactiae are human pathogens causing serious acute infections [1,2,3], whereas S. thermophilus is an important probiotic lactic acid bacterium and the only “generally recognized as safe” species of Streptococcus [4].

Small non-coding RNAs (sRNAs) are a novel class of regulatory RNAs that fine-tune biological function networks, e.g., responses to environmental changes [5, 6]. Many sRNAs have been discovered [7, 8], and their detailed regulatory functions have been studied in bacteria [7]. However, compared with model bacteria such as Escherichia coli and Bacillus subtilis, current knowledge of sRNAs and sRNA-mediated regulation in Streptococcus is limited. In this review, we provide a comprehensive overview of the functions and regulatory mechanisms of known sRNAs in Streptococcus.

Small Non-coding RNAs

As an important class of regulators, sRNAs 20–500 nucleotides (nt) in length exist in bacteria but are generally not translated into proteins [7, 9]. sRNAs are transcribed from the intergenic regions (IGRs) located between open reading frames, or in the 5′ or 3′ non-coding regions (UTRs) of coding genes [10]. sRNAs play regulatory roles at the post-transcriptional level in the bacterial response to environmental stress, through complementary base-pairing with mRNA molecules or interaction with the corresponding protein molecules [11, 12]. sRNAs are important nodes in many signaling pathways and physiological metabolic pathways [5, 13]. Most known functional sRNAs, in the form of antisense RNAs, specifically bind target mRNAs and regulate target gene expression. Most known bacterial sRNAs act as regulatory elements that respond to extracellular environment stresses and play key roles in intracellular physiological processes [14].

The process of sRNA recognition is not easy to identify through experimental methods. First, sRNAs themselves do not encode proteins, their lengths are short, and they are not easily inhibited by single nucleotide mutations. Second, sRNAs are not translated, and therefore their sequences cannot be obtained by simply recognizing open reading frames. Hence, many sRNAs are discovered and identified through a combination of bioinformatics and experimental methods [15, 16]. For example, Rath et al. [15] have created an RNase III null mutant and used RNA sequencing (RNA-Seq) to analyze the differential expression between this mutant and the wild-type of S. pyogenes. Twelve significantly differential transcripts have been obtained using the analysis of Cufflinks and Stringtie softwares, and six putative sRNAs have been visually verified by Artemis and Bamview genome viewers.

The targets of bacterial sRNAs is commonly predicted by bioinformatics methods, mainly on the basis of sequence length, base matching, RNA secondary structure, and sequence conservation and location [17]. Currently used software programs for predicting sRNA target mRNAs include RNAPredator [18], sRNATarBase [19], CopraRNA [20], RNAup [21], RNAhybrid [22], TarPicker [23], IntaRNA [24], and Target RNA2 [25]. The main difference of these in silico sRNA target discovery programs is due to the use of different prediction models, including models for general RNA–RNA interaction and models specifically designed for sRNA–target mRNA interactions in bacteria [17]. Many experimental methods have been used to verify sRNA target genes, among which genetic methods, affinity technology, immune co-precipitation methods, and microarray technology are the most mature and widely used [16, 26]. Ribosome profiling (Ribo-seq) technology is a recently developed method that simultaneously determines transcriptional and translational levels in vivo [27]. Wang et al. [28] have used Ribo-seq to fully validate the known targets of sRNA RyhB from E. coli and have identified many novel sRNA targets. Guo et al. [29] have also employed Ribo-seq to discover the target of a new σE-dependent sRNA MicL. The experimental data have indicated that Ribo-seq is an effective method for identifying sRNA targets and detecting the extent of regulation by sRNA.

With the progress of bacterial sRNAs, the regulatory mechanisms of sRNAs are gradually being elucidated (Fig. 1A and B). Bacterial sRNAs regulate the virulence of bacteria in host cells and initiate internal mechanisms in bacterial cells, thus allowing bacteria to respond to changes in growth conditions, adapt to the environmental changes in infected cells [30], and control bacterial density by regulating the expression of quorum sensing system-regulated proteins [31]. In the absence of nutrients in the growth environment, bacterial sRNAs such as CsrB and CsrC [32] play important regulatory roles in anti-nutritional stress. Trace elements are indispensable in cell metabolism, and sRNAs regulate the relative balance of trace elements in bacteria. For example, sRNA RyhB regulates the gene expression of Fur (an iron-uptake regulator) in E. coli and YdeP (which is responsible for regulation of acidity) in Shigella flexneri [33]. In addition to regulating acid tolerance, sRNAs regulate other environmental stress responses, such as MicA, RybB, and MicL to Membrance stress [34], SgrS to phosphosugar stress [34], TisB to SOS stress [34], DsrA, ArcZ, RprA, and OxyS to oxygen stress [35]. sRNAs may have different regulatory mechanisms in different bacteria, thus leading to different functional effects [36].

Fig. 1
figure 1

Regulatory mechanisms of sRNAs. A Mechanisms of inhibition of translation. a sRNAs prevent the binding of 30S ribosomes and the initiation of translation, through incomplete complementary base-pairing with the ribosome binding site (RBS) located at the 5′-UTR of the target mRNA. b sRNAs bind the mRNA being translated with the aid of Hfq, thus resulting in structural recombination of mRNA. The RBS region undergoes complementary base-pairing to form a secondary structure that prevents ribosome binding, thus leading to translational inhibition. c After the sRNA base-pairs with the mRNA, the mRNA is degraded by RNase E. In target mRNA degradation, RNase E specifically acts on AU-rich single-stranded RNA, and can simultaneously degrade target mRNA and sRNA. B Regulatory mechanisms of activation of translation. a Some sRNAs activate translation by binding to the 5′-untranslated region (5′-UTR) of the target mRNA, making the RBS becomes available, allowing initiation of translation. b Some sRNAs directly and stringently base-pair with target mRNA with the help of Hfq, thus forming a double-stranded structure, thereby increasing the stability of mRNA and protecting it from degradation by RNase E

Full size image

The Biological Functions of sRNAs in Streptococcus

sRNAs from S. pyogenes

Streptococcus pyogenes is a bacterial pathogen that causes diseases ranging from superficial infections (e.g., impetigo and pharyngitis) to systemic diseases (e.g., necrotizing fasciitis and toxic shock) [3]. Several sRNAs have been identified in S. pyogenes (Table 1), most of which are involved in virulence regulation [37].

Table 1 Overview of sRNAs in Streptococcus
Full size table

sRNA FasX

FasX is an sRNA (~ 200 nt) that affects the expression of proteins (e.g., streptokinase) involved in the conversion of plasminogen to plasmin [37]. The mechanism of FasX regulation has been well described elsewhere [38,39,40,41]. FasX is part of an operon including genes encoding two histidine kinases (FasBC) and one response regulator (FasA) [37]. FasX influences the expression of virulence factors through three steps [38,39,40,41]: (I) FasX enhances the expression of streptokinase (a virulence factor) by stabilizing the mRNA of ska, encoding streptokinase; (II) FasX regulates pilus biosynthesis and adherence through destabilizing the mRNA of the pilus biosynthesis operon through base-pairing; and (III) FasX inhibits the translation of cpa, which encodes a pilin protein (Fig. 2a). FasX regulation is serotype specific and depends on the presence of target genes [38,39,40,41]. For example, FasX can posttranscriptionally regulate the expression of the adhesion- and internalization-promoting, fibronectin-binding proteins PrtF1 and PrtF2. The mechanism is through base-pairing to the prtF1 and prtF2 mRNAs within their 5′ untranslated regions, where blocks ribosome access and leads to an inhibition of mRNA translation [41]. As a molecular switch, FasX governs the transition of S. pyogenes between the colonization and dissemination stages of infection [41].

Fig. 2
figure 2

Regulation of sRNAs in Streptococcus. a FasX regulates the expression of virulence in S. pyogenes. FasX is a main effector of the FasBCA three-component system. FasB and FasC are histidine kinases that phosphorylate the response regulator FasA after signal sensing and subsequently upregulate expression of the fas operon. FasX inhibits the expression of sag (encoding streptolysin S, SLS), fbp54 (encoding a fibronectin-binding protein), and mrp (encoding M-related protein, a fibrinogen-binding protein). FasX increases the stability of ska (encoding streptokinase) mRNA through direct targeting, thereby increasing streptokinase synthesis. b Pel controls the expression of virulence factors in S. pyogenes. The direction of the arrow of ralp3 represents its direction of transcription. c RivX and rivR activate the expression of an operon encoding Mga, M protein (encoded by emm), and Sic (streptococcal inhibitor of complement) in S. pyogenes. d MarS activates the expression of mga (encoding multiple gene activator) mRNA

Full size image

sRNAs Pel/sagA and RivX

Pleiotropic effect locus/streptolysin-associated gene A (Pel/sagA) regulates M and M-related proteins (virulence factor streptolysin S, SLS) in S. pyogenes [8]. Pel activates virulence gene expression with strain specificity. RofA-like protein IV regulator X (RivX) is an sRNA that is derived from a longer mRNA encoding the transcription regulator RivR (RALP4, a RofA-like protein family of transcriptional regulators) and corresponds to a processed form of the rivRX transcript [8]. Both RivR and RivX positively regulate the mga regulon (encoding multiple gene activator), which activates the expression of virulence-associated regulators and factors (Fig. 2b and c).

mga-Activating Regulatory sRNA (MarS)

Recently, Pappesch et al. [42] have reported the novel sRNA MarS in S. pyogenes M49591, which modulates the expression of Mga-dependent virulence factor and affects capsule biosynthesis and the fate of S. pyogenes. Deletion of MarS results in downregulation of the expression of mga and Mga-activated genes. The mode of action of MarS is shown in Fig. 2d.

sRNAs from S. pneumoniae

Streptococcus pneumoniae is an opportunistic pathogen associated with a wide range of human diseases [43]. It can cause serious infections such as sepsis, endocarditis, and meningitis. Furthermore, it is a leading cause of pneumonia, which kills more young children than any other infectious diseases [44].

sRNAs F20 and F32

The sRNAs F20 (also called as srn157) and F32 (previously identified as a tmRNA) are associated primarily with the virulence regulation of S. pneumoniae. F20 and F32 deletion mutants show decreased adhesion and invasion of nasopharyngeal or endothelial cells, thus resulting in a lack of fitness and competitive index in the environment of the nasopharynx and lungs [44]. Through proteomic analysis, Acebo et al. [45] have found that purine metabolism is significantly downregulated, whereas the pathways of DNA synthesis and repair are strongly upregulated, in an F20 deletion mutant. Kumar et al. [46] have reported that metabolic pathways encompassing the lactose transport system and multiple phosphoglucose transferase systems are depressed in an F32 mutant. F20 deletion thus has pleiotropic effects on virulence attenuation. The F32 mutant also shows a strong effect on bacterial pathogenesis. F32 is associated with deficiencies in the stress response and pathogenicity and has a core role in the trans-translation mechanism, which resolves stalled ribosomes on non-stop mRNAs [47]. However, the specific regulatory mechanisms of these two sRNAs and their target genes remains unclear.

cia-Dependent Small sRNAs (csRNAs)

csRNAs are a family of Cia, competence induction and altered cefotaxime susceptibility (CiaR)-regulated sRNAs in Streptococcus. CiaR-targeted promoters have been identified through mapping the genome of S. pneumoniae [48]. CiaR is a response regulator in the CiaRH two-component system (TCS), which is responsible for β-lactam susceptibility, competence, bacteriocin biosynthesis, autolysis and virulence in bacteria. Sixty-one csRNAs with lengths of 51–202 nt have been identified and classified into 40 different types through a search for CiaR-activated promoters in the genomes of 14 Streptococcus species [49]. Among these csRNAs, five pneumococcal csRNAs (csRNA1 to 5) are transcribed by cia-controlled non-coding RNA (ccnA-E) in the CiaR regulon. Halfmann et al. [48] have reported that csRNA4 and csRNA5 are involved in stationary-phase autolysis in S. pneumoniae R6. Microarray analysis of csRNA1 inactivation has uncovered a cis-acting effect on the mRNAs of adjacent genes, thus suggesting that csRNA1 is co-transcribed in strain D39 [50]. However, csRNA1 inactivation or activation does not affect cell growth, stress responses, global transcription or virulence.

sRNAs from S. mutans

Streptococcus mutans is a major clinical pathogen in dental caries that possesses various virulence factors and consequently can accumulate and colonize the surfaces of teeth [51]. It can utilize diverse carbohydrate sources, and produce and tolerate various acids; therefore, it can survive at low pH. It adheres to the surfaces of teeth, thereby resulting in demineralization of tooth surfaces [52]. Currently, sRNAs are known to regulate adherence and polysaccharide biosynthesis in S. mutans. Moreover, sRNAs in S. mutans are induced by different concentrations (1–5%) of sucrose during planktonic growth [53]. An array of sRNAs has been found to be induced under acid stress, in comparison with the sRNAs reported by Lee et al.[54]. The study of sRNAs that regulate bacterial virulence in S. mutans may be applied to solving the problem of dental caries (Table 1).

sRNA L10-Leader

Xia et al. [55] have predicted 334 sRNAs and verified the existence of L10-Leader in S. mutans UA159. The upregulation of L10-Leader in an acidic environment may indicate that L10-Leader binding to target mRNA improves the stability and translation of target mRNA. Consequently, cells can quickly adapt to acidic stress, and correct DNA matching is ensured under various environment stresses. However, the mechanism of L10-Leader in acidic environments remains unclear in S. mutans.

sRNA133474

Streptococcus mutans harbors diverse stress response pathways including TCSs to mitigate acid stress in the oral cavity [56]. Nine acid tolerance-associated sRNAs have been validated in S. mutans UA159 and other clinical strains cultured at pH 5.5 and 7.5. Compared with S. mutans culture at pH 7.5, in culture at pH 5.5, sRNA133474 is the most significantly downregulated among these sRNAs [56]. The mechanism of action of sRNA133474 involves regulation of the expression of LiaSR, CiaRH, and CovRS—TCSs responsible for acid tolerance.

sRNA srn225147

The bacteriocin mutacin IV from S. mutans antagonizes numerous non-mutans streptococcal species (e.g., S. gordonii). A target of sRNA srn225147 has been found to be the mutacin IV formation-associated gene comD, on the basis of RNAhybrid and RNAPredator prediction [57]. As compared with a negative control condition, comD expression has been found to significantly increase with 1400-fold increased srn225147 expression but to decrease with 400-fold increased srn225147 expression, thus suggesting that srn225147 has a two-way regulatory effect on comD expression. However, srn225147′s regulatory effect on mutacin IV biosynthesis is weak [57].

sRNA0426

The virulence of S. mutans is dependent on the formation of biofilms, which provide better conditions for bacteria to adapt to the changing environment in the oral cavity [58]. According to bioinformatics analyses, qRT-PCR, and crystal violet staining assays, Yin et al. [58] have found that sRNA0426 has a positive relationship with dynamic biofilm formation and exopolysaccharide production. In the synthesis of exopolysaccharide, sRNA0426 expression is positively correlated with the expression of three target mRNAs (GtfB, GtfC, and CcpA). Therefore, sRNA0426 has been found to play an important positive role in biofilm formation, thus providing novel insights into the biofilm regulatory network in S. mutans.

sRNAs from S. suis

Although S. suis serotype 2 (S. suis 2) is the main cause of porcine diseases, it can directly infect humans and cause hearing loss, meningitis, and septic shock [59, 60]. Zhang et al. [59] have performed RNA-seq in three strains of S. suis 2: P1/7 (reference strain) as well as 05ZYH33 and 98HAH33 (highly virulent isolates). Fifty-six sRNAs were predicted in three isolates, of which 12, particularly rliD and RatA, in all three isolates as well as cspA and rli38 in 05ZYH33 and 98HAH33 were associated with bacterial virulence. For the remaining sRNAs, over 27% (12/44) were found to be involved in riboswitches. Recently, sRNA rss04 has been confirmed to repress capsular polysaccharide biosynthesis in P1/7 [60], thus increasing adhesion and invasion of mouse brain microvascular endothelial cells. Moreover, rss04 regulates catabolite control protein A (CcpA) and the virulence factor LuxS, thereby affecting biofilm formation, capsular polysaccharide biosynthesis, and meningitis progression [61, 62]. In addition, Gong et al. [63] have verified that sRNA34 regulates the virulence of S. suis 05ZYH33. Transcriptomic analysis has indicated that deletion of sRNA34 represses the expression of HPA2 (encoding histone acetyltransferase) and 05SSU0013 (a cell cycle control gene) as well as 05SSU0003 and 05SSU1265, which are involved in the synthesis of lipoteichoic acids and contribute to cellular chain formation and elongation. Thus, the inactivation of sRNA34 results in longer cellular chains and attenuated virulence [63]. However, the mechanism of action of sRNAs in S. suis remains to be elucidated.

sRNAs from S. sanguinis

Streptococcus sanguinis, such as S. mutans, has been recognized as a pathogen involved in dental caries and periodontal diseases. Choi et al. [9] have predicted 219 sRNAs and characterized five putative sRNAs by RNA-seq, northern blotting, and qRT-PCR in S. sanguinis SK36. In the genome of SK36, among these five characterized sRNAs, the sRNA S.S-1964 is located to the right of the gene SSA_0513, encoding a putative ATP:cob (I) alamin denosyltransferase, and it may be involved in the conversion of vitamin B12 to coenzyme B12 through regulation of SSA_0513 expression [9]. Additionally, Ota et al. [64] have used computational target prediction and luciferase reporter assays to identify csRNA1-1 (a csRNA) as a negative regulator of the expression of pilT (a constituent of the type IV pilus operon) in S. anguinis. csRNA1-1 directly binds pilT mRNA, on the basis of RNA–RNA electrophoretic mobility shift assays. Moreover, csRNA1-1 and csRNA1-2 have been found to negatively regulate biofilm formation, thus suggesting that csRNAs are involved in the colonization process in S. sanguinis [64].

Perspectives

Although several sRNAs have been predicted and identified in Streptococcus, most current studies on sRNAs have focused on those involved in the regulation of virulence, such as FasX, Pel, RivX, and MarS in S. pyogenes, as well as F20 and F32 in S. pneumoniae. Understanding the roles of sRNAs in other cellular functions, particularly in the control of toxin–antitoxin systems and the responses to environmental changes, such as those in pH, oxygen, and nutrients, is very limited in Streptococcus but abundant in other bacteria [65]. Moreover, the current understanding of sRNA-mediated regulation, including modes of action and target identification, is largely insufficient in Streptococcus. sRNAs and their regulatory mechanisms reported in other bacteria have not yet been studied in Streptococcus. Thus, several questions regarding sRNAs in Streptococcus remain to be answered.

RNA chaperone protein Hfq (also known as HF-1) is required for the interaction between sRNA and target mRNAs in most sRNAs in Gram-negative bacteria [66]. For example, at least one-third of the experimentally verified E. coli sRNAs are Hfq-binding sRNAs [26]. Hfq is also at the core of sRNA-mediated post-transcriptional regulation, although the need for Hfq in the regulation of sRNA depends on multiple factors. The presence of Hfq promotes the pairing of most bacterial sRNAs with target mRNAs [26]. Hfq is involved in regulating the stability, polyadenylation, and translation of many RNAs, as well as in RNA processing. Deletion of the Hfq gene results in growth defects, decreased resistance to environmental stress, and changes in toxicity [67]. Hfq homologues also exist in some Gram-positive bacteria. However, the only known Hfq-dependent sRNA in Gram-positive bacteria is LhrA from L. monocytogenes [68]. On the basis of our comparative genome analysis, Hfq and its homologues are not found in Streptococcus. Currently known Streptococcus sRNAs do not require Hfq to function [66]. It suggests that Hfq and its homologues are not necessary for sRNA regulation in Streptococcus. How are sRNAs stabilized and annealed to the targets in Streptococcus? Generally, sRNAs are more likely to be associated with RNA-binding proteins than naked in the cellular niche [16]. Thus, whether other RNA chaperone proteins are necessary or helpful for sRNA regulation in Streptococcus needs further study.

sRNA-mediated gene silencing with an sRNA-assisted framework can provide a basis to construct artificial synthetic sRNA. The rational design of synthetic sRNAs based on natural sRNAs, can achieve high-throughput knockdown without genetic alterations; this method has been widely applied in metabolic engineering of model bacteria such as E. coli and B. subtilis [69]. Na et al. [70] have used synthetic sRNAs for combinatorial repression of four genes in 14 strains of E. coli. Repression of the expression of tyrR and csrA caused the mutant strain to produce 2 g/L of tyrosine. Moreover, repression of murE led to a 55% increase in the cadaverine titer. Meng et al. [71] have used synthetic sRNA to increase 6-deoxyerythronolide B production in E. coli, thus achieving the highest reported titer for heterologous polyketide biosynthesis. Liu et al. [72] have used an sRNA scaffold from E. coli to design synthetic sRNA to increase the yield of N-acetylglucosamine in B. subtilis. Thus, synthetic sRNA might potentially be used for gene regulation in Streptococcus. With deeper understanding of sRNAs and their regulation in Streptococcus, we believe that sRNAs could be designed and engineered as a genetic toolkit for metabolic engineering and synthetic biology in Streptococcus and other prokaryotes.