Active and inactive forms of biotin synthase occur in Heterodera glycines

Two different DNA fragments HgBioB-avr (avirulent; accession number MH251885) and HgBioB-vir (virulent; accession number MH251886) that code for HgBioB in inbred SCN strains TN10 (avirulent; Hg-type 0; grown on Roma tomato) and TN20 (virulent; Hg-type 1, 2, 3, 4, 5, 6, 7; grown on plant introduction 437654), respectively, were synthesized with codons optimized for E. coli expression (Life Technologies, Carlsbad, CA). Also, two different mutagenized forms of HgBioB-vir, HgBioB-vir-M1 (virulent, mutagenized allele 1; accession number MH251887) and HgBioB-vir-M2 (virulent, mutagenized allele 2; accession number MH251888), were prepared by generating (i) an alanine to proline mutation (M1) at amino acid position 24 (A24P); and (ii) a glutamine to arginine mutation (M2) at amino acid position 44 (Q44R), respectively. M1 was produced via site-directed mutagenesis following the manufacturer’s instructions for In-Fusion cloning (Clontech, Mountain View, CA), using HgBioB-vir as the template. A pair of inverse polymerase chain reaction (PCR) primers (Table 1), BioB-mut1-F and BioB-mut1-R, was designed to share a 15-bp homology with each other and include a single-base substitution to generate an A24P mutation. Unlike M1, M2 was prepared by custom-ordering the synthetic DNA fragment containing the desired Q44R mutation (Life Technologies).

To perform In-Fusion cloning (Clontech), two pairs of PCR primers (Table 1) were designed for the insert and the vector so that they would share a 15-bp homology with each other. All four DNA fragments were amplified using the PCR primers BioB-F and BioB-R; and pBAD.LIC.8A cloning vector (Addgene plasmid No. 37501) was amplified with PCR primers BioB-Vec-F and BioB-Vec-R. All PCRs were done using an iProof HF Mastermix (Bio-Rad, Hercules, CA) following the manufacturer’s recommended PCR conditions. The resulting PCR products were gel-purified using a QIAquick Gel Extraction Kit (Qiagen, Valencia, CA), joined together using an In-Fusion HD Cloning Kit (Clontech), and transformed into Stellar Competent Cells (Clontech) following the manufacturer’s instructions. Cells containing successful transformants were screened on Luria-Bertani (LB) agar plates with ampicillin (100 mg/ml) and a single colony was picked to inoculate 5 ml of LB broth with ampicillin (100 mg/ml) which was incubated overnight at 37°C. Recombinant plasmid DNA was extracted from the overnight culture using a QIAprep Spin Miniprep Kit (Qiagen) and the resulting plasmids were sequenced for verification at the University of Illinois Roy J. Carver Biotechnology Center.

A kanamycin-resistant E. coli vitamin B₇ auxotroph (K-12 Keio Collection: JW0758-1) was obtained from the Coli Genetic Stock Center at Yale University and used to prepare electrocompetent cells as described by Dower et al. (1988) and Hanahan (1983). Four different recombinant plasmids – pBAD.LIC.8A:: HgBioB-avr, pBAD.LIC.8A:: HgBioB-vir, pBAD.LIC.8A:: HgBioB-vir-M1, and pBAD.LIC.8A:: HgBioB-vir-M2 – were generated and independently transformed into the kanamycin-resistant ΔbioB mutant electrocompetent cells. For control purposes, an empty vector (pBAD.LIC.8 A) lacking any inserts was also transformed into the electrocompetent cells. The resulting transformants were screened on LB agar plates with ampicillin (100 mg/ml) and kanamycin (50 mg/ml), and a single colony was picked to inoculate 10 ml of LB broth with ampicillin (100 mg/ml) and kanamycin (50 mg/ml) for overnight incubation at 37°C. The overnight cultures of these four strains were thoroughly washed in liquid M9 minimal media to remove trace amounts of biotin that may be present, and then streaked onto M9 minimal media plates containing ampicillin (100 mg/ml), kanamycin (50 mg/ml), and either avidin or biotin. Plates containing avidin (1 mg/ml) were prepared to sequester trace amounts of biotin in the M9 media. Positive control plates were M9 media supplemented with biotin to a final concentration of 0.2 mM, and a negative control plate consisted of ΔbioB E. coli transformed with an empty pBAD.LIC.8A vector plated on M9 media lacking biotin. Some plates were also treated with 2% arabinose to induce HgBioB expression and monitor its potential effect on E. coli growth. All plates were initially incubated at 28°C for 48 hr, and then allowed to grow at room temperature for 10 d. E. coli growth was monitored daily.

A collection of soil samples containing diverse SCN field populations representative of the H. glycines diversity was randomly chosen and prepared for nematode culture. The majority of these soil samples was from Iowa, a kind gift from Dr Gregory Tylka at Iowa State University (Ames, IA), and some of them were from Illinois. At least two to three different field populations originating from the same geographical region (e.g., at least two to three field populations from the east central region) were randomly selected for cyst extraction. Cysts from each field population were collected as described by Niblack et al. (1993) and crushed over a 0.25-mm (60 mesh) sieve to release eggs. Eggs were suspended in 2% carboxymethyl cellulose solution for an easy delivery of an approximately equal number of eggs during inoculation. This egg suspension was used to inoculate soybean cv. Essex plants, previously pre-germinated and planted into 50-ml Falcon tubes with the absorbent wick described in Bekal et al. (2015). Inoculated plants were grown for 8 wk to allow eggs to hatch and infect the plants. The resulting cysts from each field population, approximately 2 to 200 cysts from each plant, were harvested in a 2-ml cryotube vial, and stored at −80°C until use. The cyst-containing vials were placed in a pre-cooled rack in liquid nitrogen and two stainless steel beads (3.2 mm diameter) were added into each vial. These frozen cysts were then freeze-fractured with a bead beater homogenizer and total nucleic acid was extracted using an RNeasy Mini Kit (Qiagen).

In addition to 30 DNA samples from field populations, 4 DNA samples from inbred SCN strains – 1 avirulent (OP25) and 3 virulent inbred strains (OP20, OP50, and TN20) – were also included in the analysis. The OP strains were previously maintained as described in Dong and Opperman (1997) by Dr Terry Niblack (Ohio State University, Columbus, OH), and TN20 was maintained on PI 437654. From each extracted 50-μl DNA sample from a single population, 2.5 μl was used as a template for amplification using a GenomiPhi Kit (GE Healthcare, Piscataway, NJ) following the manufacturer’s protocols. The amplified DNA was treated with ExoSAP-IT (Affymetrix, Santa Clara, CA) for the degradation of primers and nucleotides that may interfere with downstream applications. A set of PCR primers (Table 1), BioB-flank-F1a and BioB-flank-R1a, flanking the polymorphic region of the HgBioB gene was designed for PCR amplification and subsequent sequencing analysis. All PCRs were done using a CloneAmp Hi-Fi PCR Premix (Clontech) following the manufacturer’s recommended conditions. The resulting PCR product was gel-purified using a QIAquick Gel Extraction Kit (Qiagen) and then sent for sequencing at the University of Illinois Roy J. Carver Biotechnology Center. The sequences were aligned to the original HgBioB sequence using Sequencher (Gene Codes Corporation, Ann Arbor, MI), and the relative chromatogram peak heights corresponding to amino acid positions 24 and 44 were measured with a 15-cm ruler on the computer screen to estimate the percentage of each polymorphic nucleotide(s).

Peak heights were statistically analyzed using both descriptive and inferential methods. First, a box plot was generated to compare the median values and understand the distributional characteristics of these alleles at both polymorphic regions, and then a Wilcoxon signed rank test was performed using the SAS Proc Univariate procedure (SAS Institute, Cary, NC) to test if the differences between avirulent and virulent alleles were statistically significant. For the signed rank test, a new variable ‘diff’ was set as the difference between the virulent and avirulent alleles (diff = vir − avir) at each amino acid for each SCN sample prior to running the Proc Univariate procedure on that variable. Finally, a Spearman rank-order correlation was conducted using the SAS Proc Corr procedure (SAS Institute) to calculate the coefficients for the relationship between the two variables, amino acid positions 24 and 44, for each set of virulent and avirulent pairs.

We had observed differences in the predicted amino acid sequences between avirulent and virulent HgBioB alleles (Bekal et al., 2015), and thus hypothesized that such differences could lead to an altered enzyme function. To test for HgBioB function, we opted to use genetic complementation of a BioB-deficient E. coli mutant. Specifically, we wanted to learn if amino acid(s) changes, P24A or R44Q, the only two amino acids that are different between the HgBioB alleles, would alter BioB enzyme function.

To conduct this experiment, four plasmid constructs were generated to express the different forms of HgBioB; the virulent allele (HgBioB-vir, A24/Q44), the avirulent allele (HgBioB-avr, P24/R44), and the two single mutant alleles (HgBioB-vir-M1, P24/Q44; and HgBioB-vir-M2, A24/R44) (Fig. 1A). These two single mutants prepared from the virulent allele were included in the complementation studies to identify which amino acid location(s) were responsible for a functional enzyme activity. These plasmids, along with an empty vector control, were transformed into mutant E. coli lacking endogenous BioB function and were then allowed to grow on minimal agar with and without supplemented biotin.

Within 48 hr after streaking the E. coli containing the BioB constructs onto minimal media plates, all five biotin-supplemented plates (minimal media +biotin; positive controls) showed vigorous growth of E. coli as expected. Among the five plates lacking biotin (minimal media alone), however, growth only occurred for the bacteria containing the avirulent allele containing the P24/R44 amino acid combination (HgBioB-avr) (Fig. 1B). The other four constructs, E. coli carrying the empty vector (negative control), the virulent allele (HgBioB-vir), and the two mutagenized forms of the virulent allele (HgBioB-vir-M1 and HgBioB-vir-M2) failed to complement the growth of the mutant E. coli (Fig. 1B).

Our complementation test showed that avirulent SCN produces an active biotin synthase while virulent SCN contains an inactive form of the enzyme. Additionally, site-directed mutagenesis experiments confirmed that changing either of these two amino acids caused the HgBioB protein to become inactive, losing its ability to complement the mutant; therefore, a proline at position 24 (P24) and an arginine at position 44 (R44) are required for a functional BioB activity (Fig. 1A,B). This experiment was repeated and produced the same results.

Figure 1

Since our previous studies were based on just two inbred SCN strains, we were interested to learn if sequence variation in the HgBioB gene was common in different field populations of SCN. To assess how common HgBioB-vir alleles are in different populations of SCN, the HgBioB gene was PCR-amplified and sequenced from a series of SCN populations. Sequencing analysis of four inbred SCN strains (OP25, OP20, OP50, and TN20) showed that these strains were more likely to be uniform (i.e., have 100% allelic frequencies) carrying single homozygous forms (Table 2). In contrast, sequencing results from 30 SCN field populations showed that in most SCN populations, the HgBioB gene was not present in a single homozygous form, but rather in a diverse mixture consisting of both SNPs resulting in both active and inactive forms of HgBioB (Table 2). Statistical analysis of the field sequencing data was conducted by generating a box plot and comparing the median values. This analysis showed that the HgBioB-vir alleles were present in a higher percentage than the HgBioB-avr alleles at both polymorphic regions (i.e., more A24 than P24; more Q44 than R44) (Fig. 2A,B, Table 3). This trend was also supported by conducting a Wilcoxon signed rank test, the nonparametric alternative to the paired t-test, which was chosen because the polymorphic nucleotide percentage data for avirulent and virulent alleles were not normally distributed. The data were considered paired since we compared the two amino acid forms at each polymorphic region (i.e., A24 vs P24 and Q44 vs R44). The Wilcoxon signed rank test results showed that HgBioB-vir allele ranks were significantly higher than HgBioB-avr allele ranks for both amino acid positions 24 (Z=105, p<0.05) and 44 (Z=121.5, p<0.01). Therefore, from both descriptive and inferential statistical results, we concluded that the virulent alleles are more prevalent than the avirulent alleles in the field sites tested.

Figure 2

In addition, it is interesting to note that these SNP changes occur simultaneously in pairs, at about the same frequency, between amino acid positions 24 and 44. For example, if a SNP in HgBioB results in a P24 then it would be paired with an R44; if it were an A24 then it would be paired with a Q44. This linkage of SNP frequency was supported by testing the correlations between the virulent allele pair (A24 and Q44) as well as the avirulent allele pair (P24 and R44) using a Spearman rank-order correlation, the nonparametric alternative to the Pearson product-moment correlation, chosen because the assumptions of normality could not be met. The Spearman correlation results showed that there was a positive correlation for both virulent (r=0.8683, n=30, p<0.001) and avirulent (r=0.9184, n=30, p<0.001) pairs. Scatterplots for both avirulent and virulent pairs also summarize that there is a positive linear relationship between amino acid positions 24 and 44 (Fig. 3A,B). The strong, positive correlation between these positions suggests that these amino acid changes must both occur in the same protein to exert their desired, but unknown function. Although a rarity, alternative SNPs that result in a threonine at position 24 (T24) and a proline at position 44 (P44) were detected as well, suggesting an important but unknown function of these two amino acid positions in the HgBioB enzyme.

Figure 3