Harnessing CRISPR-Cas9 for genome editing in Streptococcus pneumoniae

CRISPR systems provide bacteria and archaea with adaptive immunity against viruses and plasmids by detection and cleavage of invading foreign DNA. Modified versions of this system can be exploited as a biotechnological tool for precise genome editing at a targeted locus. Here, we developed a novel, replicative plasmid that carries the CRISPR-Cas9 system for RNA-programmable, genome editing by counterselection in the opportunistic human pathogen Streptococcus pneumoniae. Specifically, we demonstrate an approach for making targeted, marker-less gene knockouts and large genome deletions. After a precise double-stranded break (DSB) is introduced, the cells’ DNA repair mechanism of homology-directed repair (HDR) pathway is being exploited to select successful transformants. This is achieved through the transformation of a template DNA fragment that will recombine in the genome and eliminate recognition of the target of the Cas9 endonuclease. Next, the newly engineered strain, can be easily cured from the plasmid that is temperature-sensitive for replication, by growing it at the non-permissive temperature. This allows for consecutive rounds of genome editing. Using this system, we engineered a strain with three major virulence factors deleted. The here developed approaches should be readily transportable to other Gram-positive bacteria. Importance Streptococcus pneumoniae (the pneumococcus) is an important opportunistic human pathogen killing over a million people each year. Having the availability of a system capable of easy genome editing would significantly facilitate drug discovery and vaccine candidate efforts. Here, we introduced an easy to use system to perform multiple rounds of genome editing in the pneumococcus by putting the CRISPR-Cas9 system on a temperature-sensitive replicative plasmid. The here used approaches will advance genome editing projects in this important human pathogen.


Introduction 39
Streptococcus pneumoniae (the pneumococcus) is a Gram-positive, human commensal that 40 colonizes asymptomatically the mucosal surfaces of the upper respiratory tract (UTR) 41 (Kadioglu et al. 2008). However, in susceptible groups like children, the elderly and the 42 immunocompromised, it can occasionally become pathogenic causing diseases that range 43 from a mild upper respiratory tract infection, acute otitis media and sinusitis, to a severe and 44 potentially life-threatening condition such as pneumonia, bacteremia and meningitis (Simell et

62
In the case of gene replacement by selection markers, while powerful, this also has 63 drawbacks, preventing further modifications of the genome when there are no further 64 selectable markers available for additional strain development. Also, many important 65 categories of gene mutation, such as missense substitutions and in-frame deletions, usually 66 present no selectable phenotype (Sung et al. 2001). To circumvent these issues, we here 67 established CRISPR genome editing for use as counterselection in the pneumococcus.

68
Clustered regularly interspaced short palindromic repeats (CRISPR) are present in 69 many bacteria and most archaea (Jansen et al. 2002). Naturally, the system provides 70 resistance against foreign genetic elements (e.g. phages or plasmids) via small noncoding 71 RNAs that are derived from CRISPR loci. In class 2 type II CRISPR systems, the mature 72 crRNA that is base-paired to a trans-activating crRNA (tracrRNA) forms a two-RNA structure 73 that directs the CRISPR-associated proteins (e.g. Cas9 from Streptococcus pyogenes) to

77
It has been demonstrated that the endonuclease can be programmed by engineering the 78 mature dual-tracrRNA: crRNA as a single RNA chimera (sgRNA for single guide RNA), to 79 cleave specific DNA sites. Thereby, modified versions of the system can be exploited as a 80 biotechnological tool for precise, RNA-programmable genome targeting and editing (Jinek et

82
After the DSB has been introduced, the cell can utilize two major pathways in order to 83 repair the break and survive: homologous recombination (HR) or non-homologous end-joining 84 (NHEJ). In HR, a second intact copy of the broken chromosome segment, homologous to the 85 DSB site, serves as a template for DNA synthesis across the break. In this mechanism, the 86 crucial process of locating and recombining the homologous sequence is performed by RecA 87 (Shuman and Glickman 2007 (Jiang et al. 2013).

94
In this study, we set out to establish a CRISPR engineering framework for S.

95
pneumoniae. Specifically, we constructed a novel replicative plasmid containing a 96 temperature-sensitive origin of replication (facilitating curing of the plasmid) carrying a genetic 97 system for making targeted, marker-less gene knockouts and large genome deletions, which 98 works with high efficiency in S. pneumoniae. The here developed plasmid system should be 99 readily transportable to other Gram-positive bacteria as the used origin of replication was 100 shown to be functional in L. lactis and B. subtilis (Bijlsma et al. 2007      To assess the efficiency of the system, we first constructed a strain (strain VL3656; Figure 5a) 406 in which we placed the E. coli lacZ gene under a constitutive promoter behind the S.

407
pneumoniae D39V cps locus (encoding the capsule). lacZ encodes for a β-galactosidase that 408 hydrolyzes X-gal to produce a blue product, allowing for blue/white screening on plates.

409
Colonies with blue color would still carry the lacZ gene, while colonies with the standard 410 white/green (on blood agar) color would indicate that the gene has been deleted from the 411 chromosome.

412
Strain VL3656 also carries the pDS07 plasmid, which contains a sgRNA targeting lacZ.

413
Next, we constructed an HR template that consisted of the 1000 bp upstream and 1000 bp 414 downstream region of lacZ (excluding lacZ) (Figure 5a)

427
Additionally, we also used the system to delete an even larger chromosomal fragment.

428
For this, we targeted the operon that encodes the capsule and the lacZ gene that had been 429 inserted downstream of it, which is around 24Kbp long, allowing for blue/white screening.

433
Using the same HR template to delete lacZ, we also performed transformation assays 434 without the counterselection offered, by inducing our CRISPR system (Figure 5f). Thousands

457
To examine whether the system could be used in multiple rounds of genome editing,

458
we attempted to delete the virulence factor pneumolysin. To delete the ply gene, we designed 459 a sgRNA targeting pneumolysin and constructed plasmid pDS12, which we transformed into 460 VL3660. Following the same procedure as used to delete the cps operon, we deleted ply.

461
Again, to confirm the successful deletion, the same principle for the primers set was used. All 462 the colonies from the transformation plate had the expected PCR product demonstrating 463 extremely high selection efficiency using the CRISPR-Cas9 system. Finally, following the 464 same strategy, we also deleted another important virulence factor, lytA resulting in strain 465 VL3665 (Figure 6d; Δcps, Δply, ΔlytA + pDS13).

467
Cas9-dependent genome editing is specific without evidence for off-target cutting in S.  (Figure 6a and b). Additionally, reads from whole-genome 489 sequencing were competitively mapped onto the reference genome, our wild type lab strain 490 D39V (Figure 6c). Direct comparison between the genomes reveals the three chromosomal 491 positions that the deletions have taken place, since in these positions, the chromosomal 492 dosage drops. Therefore, we confirmed that we had successfully performed markerless 493 deletions of these three genes (Figure 6d).  Genetic manipulation of microorganisms has been pivotal for the development of 504 biotechnological tools and the study of microorganisms themselves. In this study, we have 505 developed a novel, replicative plasmid with a temperature-sensitive origin of replication 506 carrying a CRISPR-Cas9 based system for advanced and markerless genome engineering in 507 the bacterium S. pneumoniae. In particular, we demonstrate that we have successfully deleted 508 genes and large chromosomal regions in a precise and sequential way.

509
The here designed plasmid has the temperature sensitive origin of replication pG + host, 510 which is a derivative of pWV01 of L. lactis and can be successfully propagated in 511 pneumococcus at 30°C, while it is not stable at 40°C. Indeed, we show that our pG + host 512 derivative, pDS05, is rapidly lost at 40°C (Fig. 2). We used this feature to eliminate the plasmid 513 from the strains, upon the desired deletion. The fact that the copies of the plasmid vary per 514 cell does not affect our system, since even one copy of cas9 seems to be sufficient to perform

516
Specifically, our approach is to harness this CRISPR and the homologous 517 recombination system, to perform CRISPR/Cas9-Mediated Counterselection. Following the same 518 principle of transformation with antibiotic selection, successful transformants survive the 519 CRISPR/Cas9 induced DSB, like they survive growth in antibiotics, if they uptake the rescue HR 520 template. Applying this, our CRISPR system manages to select for transformants in which 521 single genes or even large chromosomal regions were deleted with very high efficiency.

522
Comparing this to just performing natural transformation without any counterselection, which 523 would be an alternative for clean deletions, we show the advantages of our system (Fig. 4f).

524
Without it, we would need to screen many colonies to find correct transformants, depending 525 on the target. This will have to be done by colony PCR, since in most of the cases, the desired 526 deletion will not give any phenotypic difference in the colonies of the successful transformant, 527 which is a costly and time demanding process. On the other hand, with the CRISPR/Cas9-

528
Mediated counterselection, nearly all the colonies that we obtained were the desired 529 transformant, since very few false positives have been observed.

530
Since we are ultimately interested to remove multiple genes and chromosomal regions 531 from the genome, we also needed to demonstrate that our system is capable of consecutive 532 deletions. The key for this was to easily eliminate the plasmid from the newly constructed 533 strain. By growing the strain still carrying the plasmid at the high, non-permissive temperature,

534
we manage to easily cure it. Next, we can transform a new plasmid and proceed further with 535 our deletions. Specifically, after we deleted the capsule, we next deleted virulence factors ply 536 and lytA, proving that our CRISPR-Cas9 system has flexibility in genetic manipulation of the 537 bacterial genome. Together, the here described plasmid and approach will be useful for the