CRISPR loci-PCR as Tool for Tracking Azospirillum sp. Strain B510

Azospirillum-based plant and soil inoculants are widely used in agriculture. The inoculated Azospirillum strains are commonly tracked by both culture-dependent and culture-independent methods, which are time-consuming or expensive. In this context, clustered regularly interspaced short palindromic repeats (CRISPR) loci structure is unique in the bacterial genome, including some Azospirillum species. Here, we investigated the use of CRISPR loci to track specific Azospirillum strains in soils systems by PCR. Primer sets for Azospirillum sp. strain B510 were designed and evaluated by colony and endpoint PCR. The CRISPRloci-PCR approach was standardized for Azospirillum sp. strain B510, and its specificity was observed by testing against 9 different Azospirillum strains, and 38 strains of diverse bacterial genera isolated from wheat plants. The CRISPRloci-PCR approach was validated in assays with substrate and wheat seedlings. Azospirillum sp. strain B510 was detected after of two weeks of inoculation in both sterile and nonsterile substrates as well as rhizosphere grown in sterile substrate. The CRISPRloci-PCR approach was found to be a useful molecular tool for specific tracking of Azospirillum at the strain level. This technique can be easily adapted to other microbial inoculants carrying CRISPR loci and can be used to complement other microbiological techniques.


Introduction
Biofertilizer products based on microbial inoculants have been marketed since the late 1800's and are currently commercialized in many countries. It has been estimated that by 2025, worldwide revenues obtained from biofertilizers could reach 4.9 billion USD, with an annual growth rate of 11.46% [1]. Several types of microbial inoculants containing nitrogen-fixing microorganisms are now marketed and include symbiotic legume root and stem nodule bacteria (rhizobia) as well as free-living diazotrophic plant growthpromoting rhizobacteria (PGPR), such as Azospirillum. In this context, members of the genus Azospirillum are among the most studied and used soil microbial inoculants for non-legume crops (e.g., cereals). These inoculants are extensively used because of their growth-promoting functions, mainly due to their ability to fix atmospheric nitrogen (N) and to produce phytohormones (e.g., auxins, gibberellins and cytokinins).
An efficient colonization of crop and pasture plants by inoculated Azospirillum spp. is essential to increase yields in agroecosystems. It has been reported that intrinsic factors, such as exo-and lipo-polysaccharide production and motility can be determinants of competence and colonization [2,3]. Other extrinsic factors can also be relevant determinants

Primer Design
The CRISPR locus NC_013854_8 in Azospirillum sp. strain B510 was selected for primer design. The repeat sequence (5 -GCT TCA ATG AGG CCC AAG CAT TTC TGC CTG  GGA AGA C-3 ) was used as template for the design of three forward and three reverse  primers (Table 2). 'Fixed' (FP) and 'repeat' (RP) primers were designed, by using Primer3 software [12]. The RP was complementary to a conserved sequence. Four combinations of the six designed primers were established considering the repeat sequence as an annealing zone for the forward primers F1-R1 and F2-R1 ( Figure 1) and R1F-R1R and R1F-R2R as the reverse primers ( Figure 1). Prior to other standardization and validation experiments, a first PCR assay was conducted for all primer sets using GoTaq TM Flexi DNA polymerase (Promega Inc., Madison, WI, USA), using the manufacturer suggested concentrations and temperatures.

Primer Design
The CRISPR locus NC_013854_8 in Azospirillum sp. strain B510 was selected for primer design. The repeat sequence (5′-GCT TCA ATG AGG CCC AAG CAT TTC TGC CTG GGA AGA C-3′) was used as template for the design of three forward and three reverse primers ( Table 2). 'Fixed' (FP) and 'repeat' (RP) primers were designed, by using Primer3 software [12]. The RP was complementary to a conserved sequence. Four combinations of the six designed primers were established considering the repeat sequence as an annealing zone for the forward primers F1-R1 and F2-R1 ( Figure 1) and R1F-R1R and R1F-R2R as the reverse primers ( Figure 1). Prior to other standardization and validation experiments, a first PCR assay was conducted for all primer sets using GoTaq TM Flexi DNA polymerase (Promega Inc., Madison, WI, USA), using the manufacturer suggested concentrations and temperatures.

DNA Extraction and PCR Reaction
The purity of strains was checked prior to DNA extraction by streaking on R2A agar plates [13] and incubating at 30 • C for 3 days. One mL of overnight liquid cultures of each strain was used for chromosomal DNA extraction by using the cetyltrimethylammonium bromide (CTAB)-Proteinase K method [14]. The PCR reactions (50 µL) contained 15 ng DNA, 1× PCR Buffer, 1.5 mM MgCl 2 , 1 mM dNTPs, 0.5 mM of repeat targeted primer, 0.5 mM of fixed primer and 0.5 U µL −1 Promega GoTaq TM DNA polymerase (Promega Inc.). The PCR conditions were: 95 • C for 7 min, followed by 35 cycles of denaturation at 94 • C for 30 min, annealing at 60 • C for 30 min, extension at 72 • C for 30 min and a final extension at 72 • C for 7 min. PCR products were separated on 1% agarose gels made in TBE buffer and stained with GelRed ® 1× Nucleic Acid Gel Stain (Biotium Inc., Fremont, CA, USA).

PCR Reaction Standardization
Improvement of DNA fingerprint patterns was accomplished by varying concentrations of DNA polymerase, primers and dNTPs, changing annealing temperatures and by use of DMSO to enhance denaturation during PCR. All reactions used 15 ng of purified DNA from Azospirillum sp. strain B510 as template. Three commercial brands of DNA polymerases were tested: GoTaq TM Flexi (Promega Inc., USA), Invitrogen TM Taq DNA Polymerase (ThermoFisher Scientific Inc., Wartham, MA, USA) and KAPA Taq PCR Kit (KAPA Biosystems, Switzerland). The concentration of primers (IDT Inc., Newark, NJ, USA) and dNTPs (Promega Inc., Madison, WI, USA) tested ranged from 0.1 to 1 µM and from 0.1 to 1 mM, respectively. Repeat-targeted primer (RP) pairs were set at 2× the concentration of the fixed primers (FP). The PCR reactions were run at 95 • C for 10 min, followed by 35 cycles of denaturation at 94 • C for 1 min, annealing from 60 to 65 • C for 1 min, extension at 72 • C for 1 min and final extension of 72 • C for 15 min. Dimethyl sulfoxide (DMSO) was added from 1 to 10% for denaturation enhancement. In addition, the PCR amplicons were visualized using 1.5% of agarose gels prepared with 1% TAE (Tris-acetate-EDTA), TBE (Tris-Borate-EDTA) and/or SB (Sodium borate) buffers, and stained with GelRed ® Nucleic Acid Gel Stain 1× (Biotium Inc., San Francisco, CA, USA).

Specificity of CRISPR loci -PCR
The specificity of CRISPR loci -PCR reactions was evaluated by colony-and endpoint-PCR targeting all strains listed in Table 1. Fifteen ng DNA were used as template for endpoint PCR reactions, whereas for colony PCR reactions, the templates were obtained by scraping colonies from nutrient broth (NB) agar plates cultured at 28 • C for 24 h. Each reaction tube contained 1× PCR buffer, 1.5 mM MgCl 2 , 1 mM dNTPs, 0.1 mM of repeat targeted primer and 0.05 mM of fixed primer, 0.25 U µL −1 DNA polymerase and 10% DMSO (only for endpoint PCR reactions). The PCR conditions were: 95 • C for 10 min, followed by 35 cycles of denaturation at 94 • C for 1 min, annealing at 60 • C for 1 min, extension at 72 • C for 1 min and a final extension at 72 • C for 15 min. PCR products were visualized on 1% agarose gels in TBE buffer and stained with GelRed ® 1× Nucleic Acid Gel Stain.

Detection Limit of CRISPR loci -PCR
The limit of detection of the CRISPR loci -PCR assay was ascertained by using Azospirillum sp. strain B510 cells that were cultured overnight in nutrient broth at 28 • C, with shaking at 100 rpm. Cells were washed three times and resuspended in sterile saline solution (0.8% NaCl) prior to use. Washed cell suspensions were homogenized and cell numbers determined, in triplicate, by flow cytometry using a Facs Canto II instrument (BD Life Sciences, NJ, USA). In parallel, 1 mL aliquots of cell suspensions were used for triplicate DNA extractions with the DNeasy UltraClean ® Microbial Kit (QIAGEN N.V., Düsseldorf, Germany). The concentration of DNA was quantified by using samples that were diluted to 10, 5, 2.5, 1.25 and 0.625 ng DNA in TE buffer. Endpoint PCR was performed using the designed the F1-R1, F2-R1, R1F-R1R and R1F-R2R primer combinations. PCR reactions were performed as described above and DNA patterns were visualized on 1.5% Agarose-TBE gels. The greatest sensitivity (the lowest limit of detection) was observed for the F1-R1 primer set, which was chosen and used in further validation assays.
After incubation, adhered rhizosphere substrates were collected by gently shaking root systems and DNA was extracted from 0.25 g of sample using QIAGEN DNeasy PowerSoil Kit (QIAGEN N.V., Düsseldorf, Germany), following manufacturer instructions. For treatments without plants, 0.25 g of substrate was used for DNA extraction. Root endosphere DNA was also extracted from 0.15 g of root tissue, as previously described [11], using Plant and Seed DNA Miniprep Kits (Zymo Research Inc., Irvine, CA, USA).
PCR was performed on all DNA samples using primer set F1-R1 (Table 2). DNA (15 ng) from Azospirillum sp. strain B510 was used as a positive control. Optimized PCR reactions were performed as described above and DNA fragments were visualized on 1.5% Agarose-TBE gels as described above. PCR products were also verified in a 5200 Fragment Analyzer (Agilent Technologies., CA, USA) using the DNF915 dsDNA Reagent kit following manufacturer instructions (35 to 5000 bp) (Agilent Technologies., Santa Clara, CA, USA).

Primer Design
The primer used for targeting the CRISPR locus NC_013854_8 of Azospirillum sp. strain B510 revealed amplification for all designed primer sets using the DNA polymerase manufacturer's standard protocol (Figure 1). Despite the size of the target locus (2494 bp), no defined bands were detected over 500 bp. Primer sets F1-R1, F2-R1 and R1F-R1R revealed similar banding patterns in terms of band number (four noticeable bands each), and molecular weight. However, the F1-R1 and R1F-R1R combinations presented banding variability between repetitions, with similar size among all reactions. In contrast, the R1F-R2R combination showed differential amplification among repetitions. In addition, the primer combinations F1-R1, F2-R1 and R1F-R2R presented characteristic primer dimer bands, requiring greater standardizations for all tested primer sets.

CRISPR loci -PCR Standardization
CRISPR loci -PCR standardization analyzes indicate that Promega GoTaq TM Flexi was the most effective in revealing a sufficient number and types of CRISPR loci in DNA from Azospirillum sp. strain B510 (Figure 2a). While a low concentration of primers (0.1 µM RP and 0.05 µM of FP) was found to be optimal to obtain proper and repeatable fingerprints of the targeted CRISPR loci (Figure 2b), a high concentration of dNTPs (as high as 1 mM) was required to reveal sufficient CRISPR loci bands in gels (Figure 2c).
Among the other parameters examined in the PCR reaction, an optimal annealing temperature of 60 • C provided the best amplification using all primers (Table 2). Interestingly, increasing amounts of DMSO also improved the efficiency of PCR reactions (Figure 2d). The use of 10% DMSO improved band definition and diminished the nonspecific amplification previously observed (Figure 2a-c).
Both the TBE and SB buffers were found to be suitable to reveal the desired CRISPR loci fingerprints (Figure 2e,f, respectively). Although TBE buffer has proven to produce better defined gel bands, SB buffer allowed electrophoresis to be run at twice the voltage in half the time (15 min), with a concomitant reduction in buffer heating.

Specificity of CRISPR loci -PCR
Both colony and endpoint PCR assays and template DNA from all the strains listed in Table 1 were used to examine the specificity of the developed CRISPR loci -PCR assay. Results of this analysis revealed amplification only for DNA from Azospirillum sp. strain B510. No bands were amplified using DNAs from Azospirillum sp. strain strains B4 and B506, isolated from the same source as Azospirillum sp. strain B510 (Figure 3a), or from seven native Azospirillum strains isolated from wheat plants in Chile. Similarly, no CRISPR loci -PCR bands were seen using DNAs from any of the other 38 strains tested in this study, representing different genera. CRISPRloci-PCR standardization analyzes indicate that Promega GoTaq TM Flexi was the most effective in revealing a sufficient number and types of CRISPR loci in DNA from Azospirillum sp. strain B510 (Figure 2a). While a low concentration of primers (0.1 µM RP and 0.05 µM of FP) was found to be optimal to obtain proper and repeatable fingerprints of the targeted CRISPR loci (Figure 2b), a high concentration of dNTPs (as high as 1 mM) was required to reveal sufficient CRISPR loci bands in gels (Figure 2c). Among the other parameters examined in the PCR reaction, an optimal annealing temperature of 60 °C provided the best amplification using all primers (Table 2). Interestingly, increasing amounts of DMSO also improved the efficiency of PCR reactions ( Figure  2d). The use of 10% DMSO improved band definition and diminished the nonspecific amplification previously observed (Figure 2a-c).
Both the TBE and SB buffers were found to be suitable to reveal the desired CRISPR loci fingerprints (Figure 2e,f, respectively). Although TBE buffer has proven to produce better defined gel bands, SB buffer allowed electrophoresis to be run at twice the voltage in half the time (15 min), with a concomitant reduction in buffer heating. Agarose gels run using colony PCR only showed bands when strain B510 was used as template, although resolution was remarkably lower when compared to that obtained when using purified DNA as template. The remaining 47 strains assayed did not show amplified CRISPR loci -PCR bands under any of the assayed PCR conditions tested. It should be noted, however, that gel analyses revealed a large diffuse band under 100 bp from the samples examined, as is typically characteristic of unutilized primer residue.

Detection Limit
The detection limit for the developed CRISPR loci -PCR assay was examined after 24 h of cell incubation. At this time, point cell numbers in culture were 7.49 × 10 6 cells ml −1 , as determined by flow cytometry, and this was equivalent to~15.1 ng µL −1 DNA. All four DNA amounts tested (10, 5, 2.5 and 1.25 ng) revealed amplified bands with all the designed primer sets (Figure 3b), although tubes with 2.5 and 1.25 ng of DNA showed non-discernible amplification bands. The lowest amount template DNA tested (0.625 µg) did not result in the production of any visible amplification products. Based on these results, we set the detection limit for the CRISPR loci -PCR assay at 5 ng (equivalent tõ 2.48 × 10 6 cells), accomplished using the primers and PCR conditions discussed above.
anisms 2021, 9, x FOR PEER REVIEW Both colony and endpoint PCR assays and template DNA from all the strai in Table 1 were used to examine the specificity of the developed CRISPRloci-PC Results of this analysis revealed amplification only for DNA from Azospirillum s B510. No bands were amplified using DNAs from Azospirillum sp. strain strains B506, isolated from the same source as Azospirillum sp. strain B510 (Figure 3a), seven native Azospirillum strains isolated from wheat plants in Chile. Similarly, n PRloci-PCR bands were seen using DNAs from any of the other 38 strains tested study, representing different genera. Agarose gels run using colony PCR only showed bands when strain B510 w as template, although resolution was remarkably lower when compared to that o when using purified DNA as template. The remaining 47 strains assayed did n amplified CRISPRloci-PCR bands under any of the assayed PCR conditions tested. I be noted, however, that gel analyses revealed a large diffuse band under 100 bp f samples examined, as is typically characteristic of unutilized primer residue.

Detection Limit
The detection limit for the developed CRISPRloci-PCR assay was examined a of cell incubation. At this time, point cell numbers in culture were 7.49 × 10 6 cells determined by flow cytometry, and this was equivalent to ~15.1 ng µL -1 DNA. DNA amounts tested (10, 5, 2.5 and 1.25 ng) revealed amplified bands with all signed primer sets (Figure 3b), although tubes with 2.5 and 1.25 ng of DNA show discernible amplification bands. The lowest amount template DNA tested (0.625 not result in the production of any visible amplification products. Based on these we set the detection limit for the CRISPRloci-PCR assay at 5 ng (equivalent to ~2 cells), accomplished using the primers and PCR conditions discussed above.

Method Validation Using Plant Assays
A two-week wheat plant colonization assay was also conducted to validate the CRISPR loci -PCR assay. The samples were collected from plants grown in nonsterile and sterile substrates, and rhizosphere and root endosphere compartments. At sampling time, plants did not show any physiological of phenological differences. Results in Figure 4a show that the CRISPR loci -PCR DNA fingerprint banding patterns obtained using plantderived materials were as expected and similar to those obtained previously. Moreover, Azospirillum sp. strain B510-specific DNA fingerprints were seen when using materials obtained from plants grown in sterile and nonsterile substrates (SSI and NSI), and these were comparable to fingerprints obtained when using pure DNA (Figure 4a). In contrast, treatment controls without inoculation (SS and US) did not reveal any amplification bands.

Discussion
Tracking the fate of biofertilizer-based microbial inoculants released into the environment is a fundamental aspect of their technological success. Assessing the colonization and in situ activity of an inoculated microbe has been the bottleneck of biofertilizer studies since the first inoculation trials [15]. The former needs to be addressed in order to reduce the significant yield differential between chemical fertilization and biofertilization.
Hereby, we used the CRISPR locus repeat sequence NC_013854 (5′-GCT TCA ATG AGG CCC AAG CAT TTC TGC CTG GGA AGA C-3′) to developed four primer sets that specifically targeted Azospirillum sp. strain B510. The specificity of such primers was tested in silico under NCBI BLASTn (http://blast.ncbi.nlm.nih.gov (accessed on 1 June 2017, Jan 2020, and Jan 2021), where the selected CRISPR-repeat sequence did not reveal similarity with other organisms or metagenomic datasets sequenced to date. The four primer sets demonstrated to be specific; however, the primer set F1-R1 (5′-CGA ACG CTT

Discussion
Tracking the fate of biofertilizer-based microbial inoculants released into the environment is a fundamental aspect of their technological success. Assessing the colonization and in situ activity of an inoculated microbe has been the bottleneck of biofertilizer studies since the first inoculation trials [15]. The former needs to be addressed in order to reduce the significant yield differential between chemical fertilization and biofertilization.
Hereby, we used the CRISPR locus repeat sequence NC_013854 (5 -GCT TCA ATG AGG CCC AAG CAT TTC TGC CTG GGA AGA C-3 ) to developed four primer sets that specifically targeted Azospirillum sp. strain B510. The specificity of such primers was tested in silico under NCBI BLASTn (http://blast.ncbi.nlm.nih.gov (accessed on 1 June 2017, January 2020, and January 2021), where the selected CRISPR-repeat sequence did not reveal similarity with other organisms or metagenomic datasets sequenced to date. The four primer sets demonstrated to be specific; however, the primer set F1-R1 (5 -CGA ACG  CTT TCT CAA ACC AC-3 and 5 -AGA AAT GCT TGG GCC TCA TT-3 , respectively) demonstrated the ability to produce the most reliable amplification at low concentrations of purified DNA (~1.2 ng µL).
After standardization, fingerprint analyses were performed using genomic DNA from other two culture collection Azospirillum sp. CRISPR-harboring strains, as well as nine other native Azospirillum spp., and 38 native strains of different taxa previously isolated from wheat rhizosphere and root endosphere from the Region de La Araucanía [10,11].
To achieve strain-specificity and enhance PCR performance in this study, several prior steps were considered. The identification of CRISPR loci by our method relies exclusively in the availability of a completely and properly assembled genome, which needs to be further analyzed. Currently, the CRISPR databases are in constant curation [16], nonetheless genome databases have a much faster growth rate and therefore external software such as PILER-CR [17] can be used to detect CRISPRs without the use of specific databases. During the experimental designs, however, it is important to note that the target CRISPR repeats usually present palindromic sequences [18], and thus several primer combinations are required for the design to be successful.
Comparable PCR-based approaches to those proposed in this study have also been reported. Three primer sets targeting sequence characterized amplified region (SCAR) markers were designed and implemented for endpoint and qPCR assays for the quality control of a biofertilizer consortium composed of Azotobacter chroococcum, Bacillus megaterium and Azospirillum brasilense strains [19]. While authors described this method as reliable for field applications, once the designed primer sequences are subjected to BLASTn in silico analyses, cross reactivity and false positive biases with other plant-soil associated bacteria are found. Similar results were observed on a qPCR SCAR markers assay for Pseudomonas brassicacearum MA250 [20] and Azospirillum. lipoferum CRT1 [21]. Despite the primer specificity for such SCAR genes, strain-specificity could not be assured. By contrast, our CRISPR loci-based method development first considered the strain specificity of the DNA primer sets prior to their evaluation in plant assays.
In this study, the selected Azospirillum sp. strain B510 was detected by conventional PCR in sterile and nonsterile substrate, as well as in the rhizosphere of wheat seedlings sown in sterile substrate, under growth chamber conditions. It needs to be addressed that, despite differences in band size detected between the two PCR validation assays, all assays using this methodology should be accomplished in concert with purified DNA from the inoculum purified DNA as a control. Complementarily, all novel PCR assays based in this approach must be appropriately standardized by using several combinations and brands of reagents for every targeted bacterial strain because strain-specific PCR reaction conditions might significantly vary. It has been widely reported the biases in PCR-based methods produced using complex environmental samples, such as soil and rhizosphere soils, may be due to a low efficiently of DNA extraction, PCR inhibitor (e.g., humic and fulvic acids), etc. [22] In addition, plant rhizosphere is recognized as a complex matrix that harbor DNA from thousands of microbial taxa and as one of the main hotspots for activity of microorganisms in terrestrial ecosystems [23]. This high complexity could have affected the efficiency and specificity of designed primer sets using CRISPR loci -PCR. This problem could have also been exacerbated in the root endosphere samples, considering that not only microbial DNA is extracted but also a higher portion of DNA belongs to plant tissues, producing unspecific PCR reactions. It is necessary to mention that the DNA extraction kit used for endosphere samples here was designed for plant DNA; however, specific methods for the extraction of endosphere bacteria are scarce and difficult to implement due to the plant tissue's internal biochemistry, as well as differences between plant species [24]. In addition, the fingerprinting differences we have seen may due to low levels of colonization, or possibly the absence of inoculated strain in this plant niche. In this sense, studies have revealed that plant endospheres recruit specific microbial communities which can significantly differ from those found in rhizospheres of the same plant [25]. There are multiple factors that could be involved, such as the number of cells that reside in the host plant root tissues, interferent plant DNA present in the sample or incompatibility between the inoculum and the host.
The strain specificity observed in the PCR reactions confirms that the designed primer set, as well as the target CRISPR loci could be used for tracking of microbial inoculants, but further and deeper studies are still required to use the CRISPR loci -PCR approach at field level. Moreover, it must be taken into consideration that environmental and ecological studies accomplished using CRISPR loci are scarce in samples obtained from agricultural soils. Among those, the nearest approximation to our design is a CRISPR-based typing study of 85 Erwinia amilovora strains [26].
Finally, in this study we used CRISPR loci to develop an inexpensive and easy-toapply method for a specific Azospirillum strain, one of the most used bacterial genera in agriculture. The CRISPR loci -PCR approach proposed here might be used as template for further initiatives directed to evaluate the quality of commercial biofertilizers, their colonization and persistence, once commercial Azospirillum strains are inoculated by farmers and technical laboratories. In its current state, this PCR approach may also be implemented via higher-throughput PCR variants (such as quantitative PCR and droplet digital PCR) and cultivation of bacteria (such as microcolonies counting), and therefore replace the use of conventional plate-counting methods, which are a time-consuming and strainunspecific approaches.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/microorganisms9071351/s1, Figure S1: Proposed strategy for the tracking of a specific Azospirillum strain based on combined use of PCR and CRISPR loci structure. (a) Primers designed for CRISPR loci -PCR approach using fixed primer (FP) as forward primer and repeat primer (RP; complementary to conserved sequence repeats) as reverse primer. (b) Primers designed for CRISPR loci -PCR approach using repeat primer (RP; complementary to conserved sequence repeats) as forward primer and fixed primer (FP) as reverse primer. Target bacterium is named as "Inoculant".  Table 1 are available in NCBI GenBank and Genome databases (http://ncbi.nlm.nih.gov (accessed on 1 June 2017 and January 2020, May 2021). The CRISPR-loci selected for primer design (NC_013854_8) is available in https: //crispr.i2bc.paris-saclay.fr (accessed on 1 June 2017 and January 2020). Commercial Azospirillum sp. strains B510, B4 and B506 are available for acquisition in JCM-RIKEN culture collection.