Streamlining marker-less allelic replacement in Streptococcus pneumoniae through a single transformation step strategy: easyJanus

ABSTRACT

The ability to genetically manipulate bacteria is a staple of modern molecular microbiology. Since the 2000s, marker-less mutants of Streptococcus pneumoniae (Spn) have been made by allelic exchange predominantly using the kan^R-rpsL cassette known as “Janus.” The conventional Janus protocol involves two transformation steps using multiple PCR-assembled products containing the Janus cassette and the target gene’s flanking DNA. We present an innovative strategy to achieve marker-less allelic replacement through a single transformation step. Our strategy involves integrating an additional copy of the target’s downstream region before the Janus cassette, leading to a modified genetic arrangement. This single modification reduced the number of required PCR fragments from five to four, lowered the number of assembly reactions from two to one, and simplified the transformation process to a single step. To validate the efficacy of our approach, we implemented this strategy to delete in Spn serotype 4 strain TIGR4 the virulence gene pspA, the entire capsular polysaccharide synthesis locus cps4, and to introduce a single-nucleotide replacement into the chromosome. Notably, beyond streamlining the procedure, our method markedly reduced false positives typically encountered during negative selection with streptomycin when employing the traditional Janus protocol. Furthermore, and as consequence of reducing the amount of exogenous DNA required for construct synthesis, we show that our new method is amendable to the use of commercially available synthetic DNA for construct creation, further reducing the work needed to obtain a mutant. Our streamlined strategy, termed easyJanus, substantially expedites the genetic manipulation of Spn facilitating future research endeavors.

IMPORTANCE

We introduce a new strategy aimed at streamlining the process for marker-less allelic replacement in Streptococcus pneumoniae, a Gram-positive bacterium and leading cause of pneumonia, meningitis, and ear infections. Our approach involves a modified genetic arrangement of the Janus cassette to facilitate self-excision during the segregation step. Since this new method reduces the amount of exogenous DNA required, it is highly amendable to the use of synthetic DNA for construction of the mutagenic construct. Our streamlined strategy, called easyJanus, offers significant time and cost savings while concurrently enhancing the efficiency of obtaining marker-less allelic replacement in S. pneumoniae.

INTRODUCTION

Streptococcus pneumoniae (Spn), an opportunistic pathogen, is a leading cause of otitis media, community-acquired pneumonia, bacteremia, and meningitis (1, 2). First described by Pasteur and Sternberg in 1881 (3, 4), work with the pneumococcus by Avery et al. in 1944 demonstrated that DNA was the molecule responsible for the transfer of heritable traits and ushered in the modern era of molecular biology (5). In 2001, the first annotated genome of Spn was published (6). Since then, Spn has become one of the most sequenced bacteria in the world, with >40,000 genomes being publicly available (7). This wealth of genomic data has empowered researchers to identify genome regions encoding potential virulence determinants and also to better understand how this pathobiont is able to cause disease (8, 9). Along such lines, the genetic manipulation of Spn, i.e., the creation of genetic mutants, has played a pivotal role in unraveling its physiology and the basis for pathogenesis.

Spn’s ability for natural competence has been and continues to be a boon for its genomic manipulation. Investigators have learned to co-opt this process and replace DNA on the genome using mutagenic constructs that target a gene based on flanking DNA sequences that serve as the sites for homologous recombination. Suicide vector plasmids and PCR-generated mutagenic constructs have in turn been used to delete individual genes, operons, and pathogenicity islands, and even to introduce single-nucleotide replacements (10, 11). Critically, the principal molecular method used today to create marker-less mutants in Spn, as well as many other bacterial species, is the Janus cassette (12). This DNA construct, flanked by the upstream and downstream regions of the target locus, features a kanamycin resistance marker and a counterselectable rpsL gene which confers dominant streptomycin sensitivity in a resistant background. After kanamycin selection, which results in selection of mutants lacking the gene of interest, the second transformation step introduces the allele of choice flanked by the same DNA regions. Streptomycin resistance as a result of rpsL loss is consequently restored, and this is used to select for the loss of Janus and the acquisition of the desired allele, including deletion mutants. It is worth noting that the use of streptomycin for negative selection has posed challenges due to the high frequency of false positives arising from spontaneous mutations in rpsL⁺ in Janus, necessitating a tedious secondary screen. To address this, a modified version of the Janus cassette, termed Sweet Janus, was created by Yuan et al. (13). Sweet Janus incorporates sacB, which blocks growth in sucrose-rich media, alongside rpsL. Other forms of Janus, incorporating a third counterselection marker, have also been developed to further reduce the occurrence of false positives (14). While these approaches have improved the efficiency of negative selection, they still require two transformation steps for the creation of a clean deletion mutant and add to the complexity of the procedure by requiring special media.

In this context, we present a new single-step transformation methodology for allelic replacement that significantly reduces time and costs. What is more, our approach, called easyJanus, can take full advantage of emergent capabilities in synthetic DNA construction, precluding the requirement of PCR to generate the mutagenic construct, further saving time and costs. By simplifying and enhancing the efficiency of genetic manipulation, easyJanus shows great potential for accelerating research and advancing our understanding of this significant bacterial pathogen.

RESULTS

Employing the Janus cassette for allelic replacement involves two steps of transformation: first, with a donor DNA fragment containing both upstream and downstream regions of the target gene flanking the Janus cassette, and subsequently, with the fragment containing the desired allele between the upstream and downstream regions. For the purpose of gene deletion, no DNA is present between the corresponding flanking regions (Fig. 1A). Our hypothesis posited that incorporating an additional downstream sequence of the target gene upstream of Janus would result in the self-excision of the Janus cassette post-integration and first selection, obviating the need for a subsequent transformation step (Fig. 1B). To validate this hypothesis, we aimed to delete the pspA gene using this approach. Initially, our construct featured a configuration of near equal length upstream DNA (up), downstream DNA (down), Janus, and downstream DNA (500 bp-up: 500 bp-down: Janus: 525 bp-down). Note that the downstream fragment includes a 25-bp primer binding site at its 3′ end for the purpose of PCR amplification (see Materials and Methods). Transformation of this construct yielded kanamycin-resistant colonies, and subsequent PCR analysis revealed that 55% of the colonies had appropriate allelic replacement of pspA by the Janus cassette (Table 1), while the remaining 45% resulted from Janus integration downstream of the gene (Fig. S1). Following PCR validation for the correct co-integrant, we allowed self-segregation of Janus by cultivating independent correct co-integrant colonies on Todd-Hewitt broth supplemented with 0.5% yeast extract (THY), followed by selection of segregants through streptomycin selection. PCR analysis demonstrated that 92% of the streptomycin-resistant colonies harbored the desired clean deletion of the pspA gene. The <10% frequency of false positives observed during this step is indicative that the rate of intramolecular homologous recombination between the duplicated downstream region of our construct surpassed the frequency of rpsL revertants. In attempt to reduce the frequency of downstream integration of Janus following the initial transformation step, we adjusted our construct by increasing the proportion of the upstream region of the target gene in our construct relative to the downstream regions, i.e., 1,000 bp-up: 500 bp-down: Janus: 525 bp-down. Following transformation, we observed only a nominal increase in proper chromosomal integration of easyJanus from 55% to 58%. However, the proper segregation rate increased from 92% to 99% (Table 1).

Fig 1.

Schematic comparison between (A) the conventional method for allelic replacement utilizing the Janus cassette and (B) our easyJanus approach. In the traditional method, two donor DNA fragments and two transformation steps are necessary. Conversely, the new strategy involves a single DNA fragment as a donor, with only one transformation step required, facilitated by the self-segregation event through intramolecular homologous recombination.

TABLE 1.

Frequency of successful integration and segregation using easyJanus during construction of ΔpspA, Δcps4, and SP_0837 529 V→E

	Donor DNA	DNA alignmenta	Frequency (%)
			Integration	Segregation
pspA	Assembled PCR	500 bp-up: 500 bp-down: Janus: 525 bp-down	55	91.66
pspA	Assembled PCR	1,000 bp-up: 500 bp-down: Janus: 525 bp-down	58.33	98.88
pspA	Synthetic	500 bp-up: 250 bp: Janus: 275 bp-down	75.79	96
pspA	Synthetic	1,000 bp-up: 500 bp-down: Janus: 525 bp-down	82.10	99
cps	Synthetic	500 bp-up: 250 bp-down: Janus: 275 bp-down	100	90.66
V→E	Synthetic	500 bp-up: 250 bp-down: Janus: 275 bp-down	100	86.66

^a Bold denotes sections with identical DNA sequence, except for the 25 additional bases included for amplification primer design.

The ability to use a single DNA construct to achieve marker-less allelic replacement prompted us to explore the use of synthetic DNA as a donor instead of assembled PCR products. Due to limitations in synthesizing DNA molecules containing duplicated sequences, we divided our construct into two fragments for synthesis that harbored overlapping regions for subsequent HiFi assembly. In this instance, our synthetic DNA donor, comprising upstream DNA, downstream DNA, Janus, downstream DNA (1,000 bp-up: 500 bp-down: Janus: 525 bp-down), generated allelic replacement of pspA with a frequency of 82%, 1.4-fold the rate of the PCR-assembled counterpart, and 99% during segregation (Table 1). What is more, and to reduce the cost associated with generating the synthetic DNA construct, while hopefully maintaining the frequency of allelic replacement, we reduced the synthetic construct to half its size while maintaining the same flanking segment order and relative proportion, i.e., 500 bp-up: 250 bp-down: Janus: 275 bp-down. Employing this construct yielded an equivalent rate of allelic replacement (76%) with 96% segregation (Table 1).

Allelic exchange often entails not only gene deletion, but also the removal of larger DNA segments or single-nucleotide substitutions. We therefore applied our strategy to delete the entire cps4 locus, encompassing 25 genes spanning 21,592 bp. Using synthetic fragments, we successfully deleted cps4, achieving a 100% efficacy in the first step and 91% counterselection (Table 1). Validation of the cps4 deletion was performed through PCR followed by sequencing. We also evaluated the efficiency of inserting point mutations into gene SP_0837, altering the nucleotides TT to AA and thereby changing the encoded amino acid from 529-VRIEL-533 to 529-ERIEL-533. We achieved this mutagenesis using synthetic fragments with a frequency of 100% in the first step and 87% during counterselection (Table 1).

Lastly, we felt that a more automated approach for the creation of mutagenic donor constructs should be available. To this end, we wrote a Perl script that takes a whole-genome sequence (or any sequence spanning regions targeted for modification and their flanks) as input, along with a four-column tab-delimited list of targets in the form of contig identifier, target name, start coordinate, and end coordinate of the target on the input sequence provided. As an example, to generate constructs to delete pspA, the input can be the TIGR4 whole-genome sequence (NCBI NC_003028), then the tab-delimited target list containing one line can be “NC_003028(Tab space)SP_0117(Tab space)118423(Tab space)120657.” The target list can contain multiple targets, one per line. The script then extracts relevant sequences from the genome to generate two constructs of 1,199 bp each: Fragment_1 made of 500 bp-up: 250 bp-down: 449 bp-Janus, and fragment_2 made of 924 bp-Janus: 275 bp-down. The two pieces of Janus overlap by 40 bp for in vitro assembly to create a single fragment prior to transformation. For portability, the script has been embedded in a Docker image (https://hub.docker.com/r/admellodocker/easyjanus_design) that can be installed on a desktop or laptop running Windows, MacOS, or Linux.

DISCUSSION

The successful implementation of allelic replacement strategies, particularly in bacterial genetics, holds significant implications for understanding gene function, microbial pathogenesis, and genetic engineering applications. In this study, we introduce a new strategy aimed to streamline allelic replacement in Spn using the Janus cassette. Our study demonstrates the efficiency of this new methodology, thus paving the way for easier and more rapid genetic manipulations in Spn and other bacteria. Briefly, our streamlined methodology reduces the number of required PCR products (if used) from five to four, consolidates assembly reactions from two to one, and crucially reduces the need for two transformation steps down to one (Table 2). If synthetic DNA is used to create the construct, additional time is saved. Altogether, the easyJanus protocol reduced the time needed to generate mutants from the conventional 6- to 7-day process to as few as 2–3 days.

TABLE 2.

Comparison between the traditional protocol using the Janus cassette and the easyJanus system

Steps	Janus	easyJanus (PCR)	easyJanus (synthetic)
PCR fragments	5	4	0
Assembly	2	1	1
Transformations	2	1	1
Days required	6–7	3–4	2–3

Our initial experiments revealed the presence of undesired integrant stemming from a double crossover event between the two downstream regions of the target gene flanking the Janus cassette and the chromosome’s downstream region (Fig. S1). Consequently, our subsequent effort focused on optimizing the design of the donor DNA fragment to preferentially facilitate efficient allelic replacement. By adjusting the size of the upstream and downstream regions surrounding the target gene, we nominally increased the frequency of obtaining accurate integrants but did lower the rate of false positives. The greatest benefit came due to easyJanus’ applicability to use synthetic DNA for construct creation, which starkly increased the rate of appropriate allelic replacement after transformation. We speculate one reason for this is the absence of competing incomplete PCR amplicons. Our opinion is that the use of synthetic DNA to create the donor construct is also highly advantageous due to its clear advantages in terms of scalability, reproducibility, cost effectiveness, in particular when considering time spent to generate the construct, and ease of sequence customization.

We confirmed the application of our methodology beyond single-gene deletions to encompass larger genetic modifications, such as the removal of entire gene clusters or the introduction of specific point mutations. Our successful deletion of the cps locus, comprising multiple genes spanning a considerable genomic region, underscores the robustness and scalability of our approach for manipulating complex genetic loci. Additionally, our ability to introduce precise point mutations in target genes highlights the versatility and precision afforded by our methodology, thereby enabling fine-tuned genetic modifications with broad implications for functional genomics and molecular biology studies. Based on our findings, we propose a refined protocol: initially selecting five kanamycin-resistant colonies followed by the segregation of each selected colony to identify two streptomycin-resistant colonies, resulting in a total of 10 colonies (Fig. 2). Our analysis indicates that this approach yields the desired mutant with a probability exceeding 98%. Notably, our strategy offers opportunities for enhancement. For instance, our data reveal that increasing the proportion of the upstream region significantly enhances the likelihood of achieving the desired allelic replacement. Furthermore, optimizing the fragment size can potentially reduce the cost of DNA synthesis, particularly when utilizing synthetic DNA. We also have no reason to believe that other selection markers cannot be used as appropriate.

Fig 2.

Schematic representation of the proposed 3-day protocol for achieving allelic replacement in Spn. We recommend selecting five kanamycin-resistant colonies, followed by the segregation of each selected colony to identify two streptomycin-resistant colonies. Our thorough analysis affirms that this method offers a probability of over 98% in obtaining the desired mutant.

Overall, our study demonstrates the efficacy and versatility of the proposed allelic replacement approach in facilitating precise and marker-less genetic manipulations in bacterial systems. For this reason, we have termed this approach easyJanus. By systematically optimizing donor DNA design and leveraging synthetic DNA technologies, we hope to have streamlined the process of genetic engineering for other investigators, opening up new avenues for studying gene function, microbial physiology, and biotechnological applications. It is for this reason that we developed a computer script that automates the generation of two 1,199-bp fragments to be synthesized and directly used following HiFi assembly for mutant generation. As indicated, the script runs in a portable and easy-to-use Docker application that can be deployed on any of the three most common operating systems. We note that sequence coordinates provided to the script can be for any short or long region within the genome and can, for instance, encompass multiple genes such as the cps capsule synthesis locus.

Some important considerations of the described work include that we exclusively used the laboratory strain TIGR4. Importantly, we have found that easyJanus also works in serotype 2 strain D39, another common laboratory strain which relies on competence stimulating peptide (CSP)-1 for competence induction (15, 16). However, we have not systematically quantitated its efficiency in D39 nor in any recent clinical isolate as described herein for TIGR4. Additionally, easyJanus, as used by our group, relied exclusively on rpsL-mediated counterselection. Other counterselection markers, such as SacB, were not tested, but as indicated, we have no reason to expect different results, although this would change the size of the construct. Throughout the text, we mention cost savings. The cost of reagents necessary for creating the DNA constructs and validation of mutants considerably varies between institutions. Differences in the cost of manpower also occur. Based on our in-house estimates, moving from traditional Janus (requiring ~29 h of dedicated time) to easyJanus reduced dedicated technician time needed to generate a mutant by 17 h. Furthermore, shifting from PCR-based to synthetic DNA-based easyJanus constructs saved an additional 7 h of dedicated technician time. We also observed meaningful savings in reagent associated costs due to the reduced number of steps required (Table S1).

In summary, we believe that this innovation, coupled with advances in synthetic DNA technology, now renders it feasible and cost-effective to assemble a comprehensive library of marker-less mutants for Spn. Such a library holds the promise of surpassing existing mutant collections based on insertional disruption of the target gene in random fashion, offering on-demand isogenic in-frame deletion mutants without causing polar effects on neighboring genes.

MATERIALS AND METHODS

Bacterial strain and culture conditions

Spn GPSC27, serotype 4, strain TIGR4 rpsL⁺ was used for the study (6). Bacteria were routinely grown in THY or on tryptic soy agar (5% sheep blood) plates at 37°C with 5% CO₂. Antibiotics were added, when necessary, at the following concentrations: kanamycin at 200 µg/mL and streptomycin at 200 µg/mL.

Construction of DNA fragments

Oligonucleotides used in this study are listed in Table 3. DNA fragments were generated using specific sets of primers designed to amplify the Janus cassette, along with different-sized upstream and downstream DNA regions. Amplification proceeded for 36 cycles as follows: 20 s at 94°C, 30 s at 54°C, and 1.45 min at 72°C, followed by a 5-min extension cycle using Q5 High-Fidelity 2× Master Mix (New England Biolabs). PCR products were then purified using QIAquick Gel Extraction Kit (Qiagen, Valencia, CA). A schematic pertaining to the structure of our easyJanus DNA construct is provided in Fig. 1. Of note, we added an additional 25 bp to the second downstream fragment so that this fragment contains a primer binding site that is absent in the middle downstream fragment of the construct. This creates a single primer binding site for the reverse primer enabling PCR amplification of the entire assembled product, which can then be used as deemed necessary. Custom ordered synthetic DNA fragments were obtained from Twist Biosciences Inc. (South San Francisco, CA). PCR/synthetic DNA fragments were assembled using assembly master mix (NEBuilder HiFi DNA Assembly).

TABLE 3.

Oligonucleotides used in this study

Name	Sequence (5′−3′)
pspA-1	TAG CAA GTT GTT GCA TCG TAG CT
pspA-2	CCA TTG GCA TTG ACT TTA TAA CTT GTT AAA ATC ATT TTT TTC TTA TTC ATC TAA ATT TAC CTC TTT TCT
pspA-3	AAA AAA TGA TTT TAA CAA GTT ATA AAG TCA ATG CCA ATG GTG AAT GGG
pspA-4	TCA AAC GGA TCG ATC CTT AAC TTC TTA TCT TCA TTC AAA TAA CTA TAA TAA ATA AAA CAA CCC ATT TT
pspA-5	ATT TGA ATG AAG ATA AGA AGT TAA GGA TCG ATC CGT TTG ATT TTT AAT GGA TAA TGT GA
pspA-6	CCA TTG GCA TTG ACT TTA TAT TAT GCT TTT GGA CGT TTA GTA CCG TAT TTA GAA C
pspA-7	CTA AAC GTC CAA AAG CAT AAT ATA AAG TCA ATG CCA ATG GTG AAT GGG
pspA-8	AGA TTA TGT GAA TGA TGT GAG TAT ACC TTC TTC TTA TCT
pspA-9	TCT CGT TAA ACC GTT CGG TAA TTC AT
pspA-10	TCA AAC GGA TCG ATC CTT AAT AAG ATA GCG ACG CTG GCT AGA C
pspA-11	TAG CCA GCG TCG CTA TCT TAT TAA GGA TCG ATC CGT TTG ATT TTT AAT GGA TAA TG
pspA-12	CCA TTG GCA TTG ACT TTA TAT TAT GCT TTT GGA CGT TTA GTA CCG TAT TTA GAA C
pspA-13	CTA AAC GTC CAA AAG CAT AAT ATA AAG TCA ATG CCA ATG GTG AAT GGG
pspA-14	GCA AAT CCC TGT CGA GTC TTT CT
pspA-15	CCA TTG GCA TTG ACT TTA TAG ATA GCG ACG CTG GCT AG
pspA-16	GTC TAG CCA GCG TCG CTA TCT ATA AAG TCA ATG CCA ATG GTG AAT GGG
cps-1	AAC AGG TCT TGG CTC CAT GGG
cps-2	CTT CTC ACC TTT AGT GCT TGA ACC TGA
eriel-1	GTG GCT ATT TTG AAA GTC CAA TGA ATG ATA TTC C
eriel-2	GTC TGG TTG CGA CCG CT

CSP-induced transformation of Spn

TIGR4 was grown in THY media until the culture reached an optical density at OD₆₂₁ of ~0.2. Subsequently, 200 µL was transferred to competent media (5-mL THY, 50 µL of 20% glucose, 250 µL of 5% bovine serum albumin, and 10 µL of 10% CaCl₂). Competence was induced with 1 µL of CSP 2 (Eurogentec) and then incubated at 5% CO₂ and 37°C for 10 min. Subsequently, 500 µL of competent bacteria was added to microcentrifuge tubes containing 20-µL assembled mutagenic DNA constructs and incubated for 30 min in 5% CO₂ and 37°C. Following incubation, the bacteria were transferred to 7 mL of THY and grown for 3–4 h. For the selection of transformant colonies, 100 µL of culture was spread on a blood agar media with kanamycin (200 µg/mL). For the segregation step, kanamycin-resistant colonies were grown in 3 mL of THY for roughly 3–4 h. Subsequently, 100 µL of culture was spread on a plate with streptomycin (200 µg/mL). Mutants were confirmed by PCR followed by sequencing.

Calculation of integration and segregation frequency

The integration frequency, i.e., replacement of the desired allele with easyJanus, was calculated by dividing the number of colonies with positive allelic replacements, as confirmed by PCR, by the number of kanamycin-resistant colonies tested. Similarly, the segregation frequency, i.e., self-deletion of easyJanus, was calculated by dividing the number of colonies with confirmed allelic replacement by the total number of streptomycin-resistant colonies tested. The segregation frequency indicates the ratio of allelic replacements to rpsL revertant mutants.

DATA AVAILABILITY

The Perl script written for high-throughput design of easyJanus constructs has been embedded in a Docker image that can be installed on a desktop or laptop running Windows, MacOS, or Linux. This is available at https://hub.docker.com/r/admellodocker/easyjanus_design.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.01010-24.

Supplemental material. aem.01010-24-s0001.pdf.

Table S1; Fig. S1 and S2.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.