4 | Discussion
K. variicola is an emerging human pathogen with increasing antimicrobial resistance and virulence, which poses a threat to public health [1]. CRISPR/Cas system is an acquired immunity system, the most abundant form of innate immunity in prokaryotes. However, the diversity and characterization about CRISPR/Cas system in K. variicola is not well known. Similar to K. pneumoniae [28],K. variicola strains harbored three types of CRISPR/Cas systems, including type I-E, type I-E*, and type Ⅳ-A. Indeed, K. variicola was a member of theK. pneumoniae complex, which phylogenetically related to K. pneumoniae [6, 16, 35]. The consistency of CRISPR types provides new evidence for their evolutionary relationship.
Multiple types of CRISPR/Cas systems coexisted in the same strain inK. variicola . The coexistence of type I-E and IV-A system was found in five strains, and type I-E* and IV-A system coexisted in one strain. An explanation for this phenomenon was that the coexistence of different types of CRISPR/Cas systems contribute to exploiting mutual cas genes to perform corresponding functions [36]. In this study, all type IV-A systems lacked cas1 and cas2 genes. cas1 and cas2genes are indispensable in the acquisition of spacers during adaptation stage [37]. It could be speculated that type IV-A system can utilizecas genes from type I system to perform adaptation roles. Meanwhile, we also noted that the PAM sequences of type I-E and type IV-A systems were identical, which further affirms their functional relationships. Additionally, it has been demonstrated that the synergy between coexisting subtypes was more conducive to avoiding immune evasion of MGEs [38, 39]. One study has found that type III CRISPR/Cas systems could provide redundancy to overcome phage escape from type I systems [38]. Accordingly, the coexistence of type I and type IV-A system in K. variicola reinforces the immune defenses against MGEs. In addition, the absence of cas2 and cas3were found in type I-E CRISPR/Cas system. During CRISPR adaptation, the Cas2 dimer functions as an adaptor protein and forms a binding surface for the protospacer DNA combined with two Cas1 dimers [40]. Thecas3 harbors both nuclease and helicase activities, which is responsible for cutting and degrading invading DNA [41]. The different degree absence of cas2 and cas3 genes in type I-E system may represent a degenerated adaptation immunity role.
MLST has been applied to infer the phylogenetic relationship of strains [42]. Here, we observed the strong association between the distribution of chromosome-encoded type I CRISPR systems and MLST inK. variicola , which was also found in K. pneumoniae[43]. Nevertheless, plasmid-borne type IV-A systems in K. variicola were randomly distributed across MLST. This phenomenon could be attributed to the stable inheritance of chromosome-encoded CRISPR systems to the offspring and the HGT of CRISPR/Cas system mediated by plasmid. Moreover, it has been demonstrated that the scarcity of chromosome-encoded type Ⅰ systems contributes to the global expansion of multidrug-resistant ST11 and ST258 clones in K. pneumoniae [44]. It can be speculated that the uneven distribution of CRISPR systems may be a contributor to the accumulation of ARGs in specific lineages of K. variicola . It has been reported that CRISPR-positive plasmids were rich in ARGs inKlebsiella genus [29]. The adaptive immune function carried by CRISPR might confer drug-resistant plasmids more competitive advantage in survival.
Although different CRISPR loci contain diverse repeat sequences, the base composition and arrangement of repeat sequence are conserved in the same CRISPR/Cas types [45]. We also noticed that the repeat sequences showed conservation within the same subtypes in K. variicola . However, there were a small number of base differences in the repeat sequences of the same CRISPR/Cas system. Although the repeat sequences were diverse during evolution, the function was conserved. The repeat sequences could be transcribed during expression stage, thereby forming a stable RNA secondary structure [45, 46]. In this study, all types of repeat sequence formed relatively stable secondary structure. In secondary structures, conserved motifs interact to generate “stem-loop” that act as processing points for pre-crRNA through the cas6 endonuclease [45-47]. According to MFE value theory, stems with higher GC content and larger match base numbers tend to harbor lower MFE, suggesting more stable secondary structure. Here, the MFE of the type I-E repeat sequence was the smallest, representing the most stable RNA secondary structure.
The adaptation activity of CRISPR/Cas system can be measured by the number of the spacer [22]. The number of the spacers in K. variicola was diverse for different CRISPR/Cas types, ranging from 3 to 41 spacers, which indicated their different abilities to obtain spacers. Type I-E system contained the most spacers, thereby demonstrating its strong adaptation ability. The protospacer sequence was the specific sequence on the foreign gene that was consistent with spacers [48, 49]. About one-third of spacers in K. variicola were homologous to plasmids or phages. The proportion of K. variicola spacers targeting foreign genetic elements was significantly higher than other bacteria, such as Staphylococcus [50] and Clostridium perfringens [51]. It might be owing to that the sequencing ofK. variicola -related MGEs was overrepresented in the current databases. In this study, type I-E spacers showed homology with plasmids more than phages. Previous studies have shown that type I-E CRISPR/Cas systems in K. pneumoniae affected the dissemination of IncF antibiotic-resistant plasmids [44, 52]. Considering the evolution relationship between K. variicola and K. pneumoniae , it is possible that type I-E systems in K. variicola play a similar role in limiting the uptake and survival of antibiotic-resistant plasmids. Different from type I-E system, type I-E* system biasedly targeted phages, suggesting their different responsiveness to diverse MGEs. Uniquely, plasmid-borne type IV-A spacers target plasmids more effectively than phages. Moreover, the target preference for plasmids in type IV-A system was more obvious than chromosome-encoded type I system (type I-E and I-E*). The strong association between type IV-A system and plasmids can be attributed to the fierce battle between similar entities competing for overlapping niches and resources [53-55].
The spacer sequence reflects the history of interaction between bacteria and MGEs [56]. We observed that K. variicola spacers prefer to target MGEs from K. pneumoniae, which suggested that K. varricolla were frequently challenged by K. penumoniae MGEs. Previous study has shown that K. variicola displayed similar biological features to K. pneumoniae [57]. The strong interaction between K. varricola and K. pneumoniae that are implied by spacer target events provides new evidence for their evolutionary affinity. In addition, we analyzed the genes-targeted characteristics. The diversity of coding products indicated the wide range of spacer sources and the diversity of CRISPR functions [58]. Conjugative transfer-related proteins were the most frequently targeted proteins, such as TrbI, TrbD, and TrbH. These Trb proteins were required for pilus extension and conjugal DNA transfer, which were relatively active [59-61]. Conjugative transfer was the main mechanism in the spread of ARGs in which conjugative transfer proteins played an important role in mediating plasmid DNA exchange [62]. In our study, we observed that 15.15% (35/231) spacers targeted conjugative transfer proteins, which further highlight the important role of CRISPR/Cas system in influencing the dissemination of antibiotic resistance. In addition, conserved genes encoding phage tail protein, methyltransferase, kinase, terminase were also frequently targeted by CRISPR spacers. These direct targets to conserved gene regions is beneficial for improving the defense efficiency of CRISPR systems and avoiding immunity escape of MGEs.
In conclusion, our results discovers that K. variicola harbors diverse and complex CRISPR/Cas systems, including three types. The distribution of chromosome-encoded type I CRISPR/Cas system is associated with MLST whereas plasmid-encoded IV-A system is randomly distributed across MLST. Spacer homology analysis reveal that CRISPR/Cas systems targeted diverse MGEs, especially conjugative transfer-related proteins related to antibiotic-resistant plasmids, thereby suggesting the critical role of CRISPR/Cas system in control ARG spread in K. variicola . Overall, this study provides new insights into the evolution of CRISPR/Cas systems in this specie.