Correia Repeat Enclosed Elements and Non-Coding RNAs in the Neisseria Species

Neisseria gonorrhoeae is capable of causing gonorrhoea and more complex diseases in the human host. Neisseria meningitidis is a closely related pathogen that shares many of the same genomic features and virulence factors, but causes the life threatening diseases meningococcal meningitis and septicaemia. The importance of non-coding RNAs in gene regulation has become increasingly evident having been demonstrated to be involved in regulons responsible for iron acquisition, antigenic variation, and virulence. Neisseria spp. contain an IS-like element, the Correia Repeat Enclosed Element, which has been predicted to be mobile within the genomes or to have been in the past. This repeat, present in over 100 copies in the genome, has the ability to alter gene expression and regulation in several ways. We reveal here that Correia Repeat Enclosed Elements tend to be near non-coding RNAs in the Neisseria spp., especially N. gonorrhoeae. These results suggest that Correia Repeat Enclosed Elements may have disrupted ancestral regulatory networks not just through their influence on regulatory proteins but also for non-coding RNAs.


Introduction
Gonorrhoea poses a global threat that may become virtually untreatable due to antibiotic resistance, including increasing prevalence of azithromycin-resistant isolates [1][2][3][4] resulting in unsuccessful first-line dual treatment with ceftriaxone and azithromycin. Treatment for meningococcal meningitis and septicaemia must be rapid and effective or the infection will be fatal [5][6][7][8]. Concerns over the growing problem of antibiotic resistance in several bacterial species have driven investigations into RNA-based therapeutics to manipulate bacterial gene expression [9,10].
Small ncRNA aniS was annotated as a coding sequence (NMB1205 in the meningococcus and NGO0796 in the gonococcus), leading to its inclusion in the design of microarrays that have shown its differential expression [13,[21][22][23][24][25][26]. The ncRNA nrrF has been shown to be part of the fur regulon [13,16]. RNA-seq, using next-generation sequencing, has identified several potential ncRNAs due to the depth and specificity of transcript data that can be identified [17,20,27]. As a result, a ncRNA adjacent to pilin gene pilE was identified and shown to be required for pilin antigenic variation [14,19]. Deletion of
The output of SIPHT was combined with the CREE location information for each genome and those within 1000 bp were identified (Table S17 to Table S24). Starting from the distance of 1000 bp, the distances between the CREE and ncRNAs were assessed. This revealed that most of the CREE were within 300 bp of a predicted ncRNA (Table 3). For example, in strain NCCP11945 86 of the 131 CREE are within 300 bp of a predicted ncRNA (66%) ( Table 3; Table S17). There are fewer CREE within 300 bp of a predicted ncRNA in the meningococci (average 56%) than in the gonococci (average 71%) and N. lactamica (64%) ( Table 3; Table S17 to Table S24).
Many of the CREE overlap the sequences of predicted ncRNAs (Table 3; Table S17 to Table S24). In gonococcal strain NCCP11945, for example, 57% of all of the CREE (75 out of 131) in the genome overlap with predicted ncRNAs (Table 3; Table S17). There are fewer overlaps seen in N. meningitidis and N. lactamica, where on average 37% and 39%, respectively, of all CREE overlap with predicted ncRNAs, compared to a 58% average in N. gonorrhoeae (Table 3).  (Table S9 to Table S16); 3 CREE that have one or more SIPHT predicted ncRNAs within 1 kb (Table S17 to Table S24); 4 Percentage of CREE associated with predicted ncRNAs within 1 kb; 5 CREE that have one or more SIPHT predicted ncRNAs within 300 bp (Table S17 to Table S24); 6 Percentage of CREE associated with predicted ncRNA within 300 bp; 7 CREE that overlap one or more SIPHT predicted ncRNAs (Table S17 to Table S24); 8 Percentage of CREE that overlap with predicted ncRNAs.
Using all of the CREE start or end locations for strain NCCP11945 matched to the nearest corresponding ncRNA start or end locations (Table S25), the Kolmogorov-Smirnov test indicated that the CREE locations were not normally distributed (p < 0.001). Using the same test, CREE distribution around the genome was shown to be neither random (Z = 5.292; p < 0.001) nor uniform (Z = 1.557; p = 0.016). Spearman's Rho correlation demonstrated that CREE locations were very strongly correlated with ncRNA locations (ρ = 1; p < 0.001). The ncRNA locations were all either located within or very close to CREE regions ( Figure 1).
that the CREE locations were not normally distributed (p < 0.001). Using the same test, CREE distribution around the genome was shown to be neither random (Z = 5.292; p < 0.001) nor uniform (Z = 1.557; p = 0.016). Spearman's Rho correlation demonstrated that CREE locations were very strongly correlated with ncRNA locations ( = 1; p < 0.001). The ncRNA locations were all either located within or very close to CREE regions (Figure 1).

Discussion
SIPHT predicts between 760 and 996 ncRNAs in the Neisseria spp. genomes investigated present within the 1782 to 2255 intergenic regions (Table 1; Table S1 to Table S8). Amongst these predictions are several previously reported and verified ncRNAs [13,14,16,19,24,25,39], which supports the accuracy of the SIPHT predictions. The predictions reported here include some that span the same region of sequence, where only one is likely to be a ncRNA, and some that may be over-predictions. However, these predictions provide a starting point for investigations into ncRNAs in these species.
Correia Repeat Enclosed Elements are unique to the Neisseria spp. [29,30]. They have been demonstrated to insertionally inactivate genes [31] and to disrupt ancestral regulatory systems through CREE-associated promoters [40]. There are, on average 127 CREE in a N. gonorrhoeae genome, 249 in a N. meningitidis genome, and there are 92 CREE in the N. lactamica strain 050-20 (Table 2; Table S9 to  Table S16). CREE tend to be located within 300 bp of predicted ncRNAs, with many overlapping (Table 3; Table S17 to Table S25). The frequency with which CREE overlap predicted ncRNAs is higher in N. gonorrhoeae (average 58%) than in N. meningitidis (average 37%) and N. lactamica (39%). This is particularly of note because there are only half as many CREE in N. gonorrhoeae, yet those that are present are far more likely to be located close to ncRNAs and to overlap the ncRNA sequences. In N. lactamica, the number of CREE in the genome is closer to that seen in the gonococcus and yet the overlaps between CREE and ncRNAs are on par with N. meningitidis (Table 3). Given the known roles of CREE in gene regulation [40][41][42], it is possible that the presence of CREE and its associated promoters may influence the expression of small ncRNAs through insertional inactivation and/or disruption of ancestral regulatory networks through introduction of CREE-associated promoters.
RNA-seq data supports the SIPHT predictions (Table S26), demonstrating transcription of 95% of the ncRNAs that have adjacent or overlapping CREE in N. gonorrhoeae strain NCCP11945. Those with the highest detected transcription are upstream of virulence determinants (Table S27), further supporting the important role of ncRNAs in these pathogens.

Conclusions
In the Neisseria spp., CREE are frequently found near or overlapping ncRNAs in the genome. CREE may influence the expression of ncRNAs through the presence of their promoters and/or insertional activation, similar to the role of CREE in gene regulation. It may be possible to exploit differences between the species with regards to ncRNAs and their interaction with CREE in the design of RNA-based therapies to restrict the meningococcus to the mucosal surface, where it is as harmless as N. lactamica, and to prevent gonococcal antigenic variation, enabling rapid clearance and immunity.
Supplementary Materials: The following are available online at www.mdpi.com/2076-2607/4/3/31/s1, Table S1: SIPHT predicted small non-coding RNAs in Neisseria gonorrhoeae strain NCCP11945, Table S2: SIPHT predicted small non-coding RNAs in Neisseria gonorrhoeae strain FA1090, Table S3: SIPHT predicted small non-coding RNAs in Neisseria meningitidis strain MC58, Table S4: SIPHT predicted small non-coding RNAs in Neisseria meningitidis strain Z2491, Table S5: SIPHT predicted small non-coding RNAs in Neisseria meningitidis strain FAM18, Table S6: SIPHT predicted small non-coding RNAs in Neisseria meningitidis strain alpha14, Table S7: SIPHT predicted small non-coding RNAs in Neisseria meningitidis strain 53442, Table S8: SIPHT predicted small non-coding RNAs in Neisseria lactamica strain 020-06,  Table S18: Association of CREE and ncRNAs in Neisseria gonorrhoeae strain FA1090, Table S19: Association of CREE and ncRNAs in Neisseria meningitidis strain MC58, Table S20: Association of CREE and ncRNAs in Neisseria meningitidis strain Z2491, Table S21: Association of CREE and ncRNAs in Neisseria meningitidis strain FAM18, Table S22: Association of CREE and ncRNAs in Neisseria meningitidis strain alpha14, Table S23: Association of CREE and ncRNAs in Neisseria meningitidis strain 53442, Table S24: Association of CREE and ncRNAs in Neisseria lactamica strain 020-06, Table S25: Locations of the start or end points of the CREE and the closest predicted ncRNA in Neisseria gonorrhoeae strain NCCP11945, Table S26: RNA-seq data supporting candidacy of ncRNAs with nearby CREE in Neisseria gonorrhoeae strain NCCP11945, Table S27: Genes nearby to CREE that are nearby to ncRNAs in Neisseria gonorrhoeae strain NCCP11945.