TxtH is a key component of the thaxtomin biosynthetic machinery in the potato common scab pathogen Streptomyces scabies.

Streptomyces scabies causes potato common scab disease, which reduces the quality and market value of affected tubers. The predominant pathogenicity determinant produced by S. scabies is the thaxtomin A phytotoxin, which is essential for common scab disease development. Production of thaxtomin A involves the nonribosomal peptide synthetases (NRPSs) TxtA and TxtB, both of which contain an adenylation (A-) domain for selecting and activating the appropriate amino acid during thaxtomin biosynthesis. The genome of S. scabies 87.22 contains three small MbtH-like protein (MLP)-coding genes, one of which (txtH) is present in the thaxtomin biosynthesis gene cluster. MLP family members are typically required for the proper folding of NRPS A-domains and/or stimulating their activities. This study investigated the importance of TxtH during thaxtomin biosynthesis in S. scabies. Biochemical studies showed that TxtH is required for promoting the soluble expression of both the TxtA and TxtB A-domains in Escherichia coli, and amino acid residues essential for this activity were identified. Deletion of txtH in S. scabies significantly reduced thaxtomin A production, and deletion of one of the two additional MLP homologues in S. scabies completely abolished production. Engineered expression of all three S. scabies MLPs could restore thaxtomin A production in a triple MLP-deficient strain, while engineered expression of MLPs from other Streptomyces spp. could not. Furthermore, the constructed MLP mutants were reduced in virulence compared to wild-type S. scabies. The results of our study confirm that TxtH plays a key role in thaxtomin A biosynthesis and plant pathogenicity in S. scabies.

Introduction

Over 580 species of Streptomyces have been identified to date (Garrity et al., 2007), of which only a very small number have the ability to infect living plant tissue and cause plant diseases (Bignell et al., 2010a). One of the best studied plant‐pathogenic species is Streptomyces scabies (syn. S. scabiei), which causes common scab disease of potato (Bignell et al., 2014; Loria et al., 2006). The main symptom associated with this disease is the formation of superficial, raised or deep‐pitted lesions on the tuber surface, and these lesions reduce the market value of affected potatoes, leading to economic losses for potato growers (Dees and Wanner, 2012). As S. scabies is neither tissue‐ nor host‐specific, it can cause scab disease symptoms on other economically important root crops such as radish, carrot, beet and turnip (Dees and Wanner, 2012). Also, the seedlings of model plants such as Arabidopsis thaliana and Nicotiana tabacum can be infected by S. scabies, resulting in root stunting, swelling, necrosis and seedling death (Loria et al., 2006).

The primary pathogenicity determinant produced by S. scabies consists of a family of specialized phytotoxic metabolites called the thaxtomins, which are cyclic dipeptides (King and Calhoun, 2009). Eleven different thaxtomins have been described, of which thaxtomin A is the predominant member produced by S. scabies and other scab‐causing pathogens such as Streptomyces turgidiscabies, Streptomyces acidiscabies, Streptomyces europaeiscabiei and Streptomyces stelliscabiei (King and Calhoun, 2009; King et al., 1989). A positive correlation was established between the pathogenicity of S. scabies strains and their ability to produce thaxtomin A (King et al., 1991), and disruption of thaxtomin biosynthesis in S. acidiscabies abolished the ability of the pathogen to cause necrotic lesions on potato tubers (Healy et al., 2000). Thaxtomin A targets the plant cell wall by functioning as a cellulose synthesis inhibitor (Scheible et al., 2003), and its production is induced by cellobiose and cellotriose, which are the smallest subunits of cellulose (Johnson et al., 2007). In A. thaliana, thaxtomin A has been shown to affect the expression of genes involved in cell wall synthesis, and it also reduces the number of cellulose synthase complexes in the plant cell plasma membrane (Bischoff et al., 2009). In addition, thaxtomin A elicits an early defence response in Arabidopsis by inducing the influx of Ca2+ and the efflux of H+ ions (Bischoff et al., 2009; Errakhi et al., 2008; Tegg et al., 2005).

The biosynthetic gene cluster responsible for the synthesis of thaxtomin A and related analogues is highly conserved in scab‐causing Streptomyces spp. and consists of seven genes: txtA, txtB, txtC, txtD, txtE, txtH and txtR (Fig. 1). txtD encodes a nitric oxide synthase that generates nitric oxide (NO) from l‐arginine, and txtE encodes a novel cytochrome P450 monooxygenase that nitrates l‐tryptophan using the NO to produce the intermediate 4‐nitro‐l‐tryptophan (4‐NTrp) (Barry et al., 2012; Johnson et al., 2009). Two nonribosomal peptide synthetases (NRPSs) encoded by txtA and txtB have been proposed to synthesize thaxtomin D using l‐phenylalanine and 4‐NTrp as substrates, respectively (Healy et al., 2000; Johnson et al., 2009; Loria et al., 2008). NRPSs are a family of large proteins that produce nonribosomal peptide molecules with diverse structures and activities. NRPSs consist of multiple enzymatic domains, of which the adenylation domain (A‐domain) is responsible for selecting and activating the amino acid substrate for incorporation into the peptide product (Süssmuth and Mainz, 2017). The txtC gene encodes a cytochrome P450 monooxygenase that introduces two hydroxyl groups onto the thaxtomin D backbone to generate the final thaxtomin A product (Fig. 1) (Healy et al., 2002), and txtR encodes a cluster‐situated regulator that activates the expression of the thaxtomin biosynthetic genes (Joshi et al., 2007). Additionally, a small gene called txtH is located between txtB and txtC and encodes a protein belonging to the MbtH‐like protein (MLP) family (Bignell et al., 2010a).

Figure 1

(A) Organization of the thaxtomin biosynthetic gene cluster in Streptomyces scabies 87.22. The block arrows represent the genes within the cluster, and the direction of each arrow indicates the direction of transcription. Biosynthetic genes txtA, txtB, txtC, txtD and txtE are represented in black, the regulatory gene txtR is grey, and the MbtH‐like protein (MLP)‐encoding txtH gene is orange. (B) The proposed biosynthetic pathway of thaxtomin A in S. scabies 87.22.

MLPs are small (~70 amino acids) proteins that are usually encoded within NRPS gene clusters (Baltz, 2011). Recent studies have shown that they play essential roles in promoting the proper folding and activity of the NRPS A‐domains, though for reasons currently unknown, not all NRPS A‐domains require an MLP for proper function (Boll et al., 2011; Felnagle et al., 2010; McMahon et al., 2012; Schomer and Thomas, 2017; Zhang et al., 2010). It has also been revealed that MLPs from different pathways can, in some instances, functionally complement each other with varying efficiencies (Boll et al., 2011; Lautru et al., 2007; Mori et al., 2018; Schomer and Thomas, 2017; Wolpert et al., 2007; Zhang et al., 2010).

In the current study, we used multiple approaches to investigate the requirement of TxtH and other MLPs during the biosynthesis of thaxtomin A in S. scabies.

Results and Discussion

Bioinformatics analysis of S. scabies TxtH

The TxtH amino acid sequence was aligned with that of other MLP homologues from the database, including some that have been previously characterized (Fig. 2A; Table S1). Pairwise comparisons revealed that TxtH shares the greatest degree of amino acid identity (100%) with the corresponding homologues from the thaxtomin biosynthetic gene clusters in the potato scab pathogens S. acidiscabies and S. europaeiscabiei (Table S2). In contrast, the TxtH homologue from the thaxtomin biosynthetic gene cluster of another scab pathogen, S. turgidiscabies, shares only 80% amino acid identity with the S. scabies TxtH. This is consistent with a previous phylogenetic analysis that suggested that the thaxtomin biosynthetic gene clusters from S. scabies and S. acidiscabies are more closely related to each other than to the S. turgidiscabies gene cluster (Huguet‐Tapia et al., 2016). As expected, the TxtH homologues from the pathogenic Streptomyces spp. formed a well‐supported clade in the constructed phylogenetic tree (Fig. 2B). Interestingly, an MLP (ACM01_RS10820) from the nonpathogenic species Streptomyces viridochromogenes also clustered together with the TxtH homologues from the pathogenic species and showed a very high degree of amino acid identity (97%) with TxtH from S. scabies, S. europaeiscabiei and S. acidiscabies (Fig. 2B and Table S2). An analysis of the S. viridochromogenes genome sequence (accession number PRJNA238534) revealed that the MLP is encoded in the vicinity of four other genes that show strong similarity to thaxtomin biosynthetic genes, though three of the genes appear to be pseudogenes. Two other MLPs encoded in the S. scabies 87.22 genome, SCAB_3331 (herein referred to as MLPlipo) and SCAB_85461 (herein referred to as MLPscab), both share only 52.3% amino acid identity with TxtH (Table S1) and cluster together in a separate clade as compared to the one containing TxtH (Fig. 2B). The MLPlipo‐encoding gene is located within an NRPS gene cluster that is responsible for the biosynthesis of a putative lipopeptide metabolite (Yaxley, 2009), whereas the MLPscab‐encoding gene is localized within the NRPS gene cluster that synthesizes the siderophore scabichelin (Kodani et al., 2013). An orphan MLP (MXAN_3118) from Myxococcus xanthus DK 1622 showed the least amino acid identity (33.8%) with TxtH in the pairwise comparison (Table S2). A recent study showed that MXAN_3118, which is not located within or near any NRPS biosynthetic gene clusters, can interact in vivo and in vitro with several different NRPSs in M. xanthus (Esquilin‐Lebron et al., 2018).

Figure 2

(A) Amino acid alignment of TxtH from Streptomyces scabies with other MbtH‐like protein (MLP) homologues. Highly conserved amino acids are highlighted as follows: black, 100% identity; dark grey, 80–99% identity; grey, 60–79% identity; light grey, <60% identity. The MLP signature sequence is indicated by the black line above the alignment, and the conserved residues subjected to mutation in the S. scabies TxtH are indicated with the asterisks. (B) Phylogenetic analysis of the MLP homologues. The tree was constructed using the maximum likelihood algorithm, and bootstrap values >50% for 1000 repetitions are shown. The scale bar indicates the number of amino acid substitutions per site. The S. scabies TxtH is highlighted in red, while the other MLPs encoded in the S. scabies genome are shown in blue. The non‐cognate MLPs used for the complementation experiments are indicated with the black asterisks. Ssc, Streptomyces scabies; Sac, Streptomyces acidiscabies; Stu, Streptomyces turgidiscabies; Seu, Streptomyces europaeiscabiei.

Baltz (2011) previously proposed a signature sequence [NxExQxSxWP(x)5PxGW(x)12L(x)6WTDxRP] consisting of multiple amino acid residues that are invariant in most MLPs, all of which are also conserved in TxtH and in other MLP homologues analysed here (Fig. 2A). Structural analysis of the PA2412 MLP from Pseudomonas aeruginosa PAO1 revealed that several of these residues, including the three highly conserved tryptophan residues, lie on one face of the protein, which is thought to interact with conserved components of the cognate NRPS (Drake et al., 2007). The structure of SIgN1, a 3‐methylaspartate‐adenylating enzyme with an MLP domain at its N‐terminus, revealed that two of the conserved tryptophan residues (W25 and W35) from the MLP domain are located at the interface between the MLP and the A‐domain and are important for this interaction (Herbst et al., 2013). Analysis of mutants defective in equivalent residues (W22A/W32A) in another MLP, PacJ, showed that they contribute to PacJ’s ability to form a complex with the cognate PacL NRPS to stimulate the adenylation activity of the synthetase (Zhang et al., 2010). Based on these studies, we predict that the conserved amino acid residues in TxtH play an important role in its interaction with the thaxtomin NRPS in S. scabies.

TxtH is required for promoting the solubility of the TxtA and TxtB A‐domains

Previously it was shown that YbdZ, an MLP encoded in the enterobactin biosynthetic gene cluster of Escherichia coli, can interact with adenylating enzymes from different NRPS biosynthesis pathways (Felnagle et al., 2010). Therefore, the TxtA and TxtB A‐domains were expressed as N‐terminal 6 × histidine (His6)‐tagged proteins in an E. coli ybdZ mutant [BL21(DE3)ybdZ:aac(3)IV] to avoid any potential interference caused by YbdZ during co‐expression studies using TxtH (Table 1). Each A‐domain (herein referred to as His6‐TxtA‐A and His6‐TxtB‐A) was expressed in the presence or absence of TxtH, which itself either contained or lacked an N‐terminal His6‐tag, to rule out any influence that the tag might have on the function of TxtH. The ability of TxtH to promote the solubility of each A‐domain was then determined by western blot analysis of isolated soluble protein fractions using an anti‐His6 antibody.

Table 1

Bacterial strains used in this study.

Strain	Description	Resistance†	Reference or source
Escherichia coli strains
DH5α	General cloning host	n/a	Gibco‐BRL
NEB5α	DH5α derivative, high efficiency competent cells	n/a	New England Biolabs
BL21(DE3)ybdZ:aac(3)IV	BL21(DE3) derivative, ybdZ replaced with an apramycin resistance cassette (aac(3)IV)	ApraR	Herbst et al., 2013
ET12567/pUZ8002	dam –, dcm –, hsdS –; non‐methylating conjugation host	KanR, CmlR	MacNeil et al., 1992
Streptomyces scabies strains
87.22	Wild‐type strain	n/a	Loria et al., 1995
87.22/Δmlplipo_int	Strain 87.22 containing plasmid pIJ12738/Δmlplipo inserted into the chromosome	ApraR	This study
∆mlplipo	mlplipo deletion mutant derivative of strain 87.22	n/a	This study
∆txtH	txtH deletion mutant derivative of strain 87.22	ApraR	This study
∆mlplipo/∆txtH	txtH deletion mutant derivative of strain Δmlplipo	ApraR	This study
ΔtxtH/Δmlpscab	mlpscab deletion mutant derivative of strain ΔtxtH	ApraR, HygR	This study
∆mlplipo/∆txtH/∆mlpscab	mlpscab deletion mutant derivative of strain Δmlplipo/ΔtxtH	ApraR, HygR	This study

n/a = not applicable.

ApraR, KanR, CmlR and HygR = apramycin, kanamycin, chloramphenicol and hygromycin resistance, respectively.

As shown in Fig. 3A, only trace levels of soluble His6‐TxtA‐A and His6‐TxtB‐A protein were detected in E. coli when expressed in the absence of TxtH, whereas both proteins were readily detectable in soluble form when co‐expressed with the MLP. The solubility promoting activity of TxtH was observed regardless of whether or not the protein contained an N‐terminal His6 tag (Fig. 3A), indicating that the tag did not interfere with the activity of the protein. Our results therefore suggest that TxtH likely functions as a chaperone that is essential for the proper folding of both A‐domains in the thaxtomin NRPSs, a role that is consistent with that proposed for other MLPs (Imker et al., 2010; Zhang et al., 2010; Zolova and Garneau‐Tsodikova, 2012).

Figure 3

(A) Western blot analysis of soluble His6‐TxtA‐A and His6‐TxtB‐A expressed in the presence and absence of His‐tagged and untagged TxtH. Lanes: 1, His6‐TxtB‐A co‐expressed with His6‐TxtH; 2, His6‐TxtB‐A expressed without His6‐TxtH; 3, His6‐TxtA‐A co‐expressed with His6‐TxtH; 4, His6‐TxtA‐A expressed without His6‐TxtH; 5, His6‐TxtB‐A co‐expressed with TxtH; 6, His6‐TxtA‐A co‐expressed with TxtH; 7, His6‐TxtB‐A expressed without TxtH; 8, His6‐TxtA‐A expressed without TxtH. (B) Western blot analysis of soluble His6‐TxtA‐A and His6‐TxtB‐A that was co‐expressed with wild‐type and mutant His6‐TxtH proteins. The lanes corresponding to the different His6‐TxtH point mutants are indicated, and lanes containing A‐domain produced in the presence (+) or absence (–) of wild‐type His6‐TxtH are also shown. (C) Western blot analysis of soluble His6‐TxtA‐A and His6‐TxtB‐A expressed in the absence of an MbtH‐like protein (MLP) (lanes 4 and 8) or co‐expressed with His6‐TxtH (lanes 3 and 7), His6‐MLPlipo (lanes 2 and 6) or His6‐MLPscab (lanes 1 and 5).

To further explore the role of the highly conserved amino acid residues in the MLP signature sequence of TxtH, we constructed several His6‐TxtH point mutants (N17A, Q21A, S23A, S23Y, L24A, W25A, W35A, W55A, T56A and D57A) and then co‐expressed each mutant protein with His6‐TxtA‐A and His6‐TxtB‐A. As shown in Fig. 3B, the solubility of both the His6‐TxtA‐A and His6‐TxtB‐A proteins was reduced or abolished when co‐expressed with all of the TxtH point mutants. Of particular note is the S23Y mutation, which resulted in complete loss of soluble protein for both A‐domains. Herbst and colleagues showed that the same mutation in the MLP domain of the SIgN1 hybrid adenylase resulted in a 5‐fold reduction in adenylation activity of the enzyme, most likely due to impairment of the interaction between the MLP and adenylation domains by the bulky tyrosyl residue (Herbst et al., 2013). In contrast, the S23A mutation in TxtH caused a drastic reduction of soluble His6‐TxtB‐A protein but did not lead to a complete loss of soluble protein, and it only slightly reduced the solubility of the His6‐TxtA‐A protein (Fig. 3B). This is possibly due to the fact that an alanine side chain is less bulky than a tyrosine side chain and may therefore cause less steric interference during the interaction of the MLP with the A‐domains. All three highly conserved tryptophan residues in TxtH (W25, W35, W55) (Fig. 2A) were found to be essential for promoting the solubility of His6‐TxtB‐A, whereas only W35 and W55 are essential for promoting His6‐TxtA‐A solubility (Fig. 3B). The W55 residue is part of the highly conserved WTDxRP motif, which in the P. aeruginosa PA2412 occurs between two alpha helices and was proposed to play a role in the proper orientation of the C‐terminal helix (Drake et al., 2007), whereas in MbtH from M. tuberculosis the motif lies within a disordered region (Buchko et al., 2010). Our results show that in addition to W55, two other residues within this motif (T56, D57) are critical for the ability of TxtH to promote the solubility of His6‐TxtB‐A, whereas neither residue is essential for obtaining soluble His6‐TxtA‐A, though His6‐TxtA‐A solubility was clearly affected in the presence of these point mutants. Other TxtH residues that were found to be essential for promoting the solubility of His6‐TxtB‐A are N17 and L24. Overall, our results show that all of the highly conserved amino acid residues found in the MLP signature sequence are important for the solubility‐promoting activity of TxtH. We anticipate that structural studies examining the interaction of TxtH with each A‐domain will provide further insight into the specific function of these residues during such interactions.

Loss of MLPs abolishes thaxtomin A production in S. scabies

To examine the in vivo role of txtH in the thaxtomin A biosynthetic pathway, we deleted txtH from the S. scabies chromosome (Fig. S1). Four mutant isolates were examined for thaxtomin A production, and all were found to produce significantly less thaxtomin A as compared to the wild‐type strain (Fig. 4A,B). Production in the ΔtxtH1 mutant isolate was partially restored when txtH was expressed from an integrative plasmid using the strong, constitutive ermEp* promoter (Fig. 4C). Notably, two other metabolites with retention times of 3.82 and 4.64 min were found to accumulate at very low levels in the ΔtxtH mutant isolates but not in wild‐type S. scabies (Fig. 4B, peaks ▼ and ∇). Liquid chromatography‐high resolution electrospray ionization mass spectrometry (LC‐HRESIMS) analysis of the ΔtxtH1 mutant culture extract in negative ion mode revealed a pseudomolecular [M‐H]– ion at m/z 421.1524 for peak ▼ and a pseudomolecular [M‐H]– ion at m/z 405.1577 for peak ∇, which is consistent with the accumulation of thaxtomin B and D, respectively (King and Calhoun, 2009). Thaxtomin D was  previously reported to accumulate in a ΔtxtC mutant of S. acidiscabies (Healy et al., 2002), suggesting that there may be some polar effects on the expression of txtC caused by the deletion of txtH, even though the orientation of the inserted apramycin resistance cassette was the same as the original txtH gene (Figs S1 and S2). Indeed, semiquantitative RT‐PCR analysis showed that the txtC transcription level was reduced in the ΔtxtH1 mutant compared to the wild‐type strain, though expression of txtC could still be detected in the mutant, especially at higher PCR cycle numbers (Fig. 5; data not shown).

Figure 4

Production of thaxtomin A by Streptomyces scabies strains. Shown are the mean thaxtomin A production levels (ng thaxtomin A/mg dry cell weight) from triplicate cultures of each strain, with error bars representing the standard deviation from the mean. Means with different letters (a, b, c, d) were determined to be significantly different (P ≤ 0.05). (A) Thaxtomin A production levels in S. scabies 87.22 and in the ΔtxtH mutant isolates 1–4. (B) HPLC chromatograms of culture extracts from S. scabies 87.22 (i), ΔtxtH1 (ii), Δmlp/ΔtxtH (iii), ΔtxtH/Δmlp (iv) and Δmlp/ΔtxtH/Δmlp (v). The peak corresponding to thaxtomin A in each chromatogram is indicated with the red asterisks, and the peaks corresponding to the thaxtomin B and thaxtomin D intermediates are indicated with ▼ and ∇, respectively. (C) Thaxtomin A production levels in the ΔtxtH1 mutant isolate following complementation with the txtH gene. (D) Thaxtomin A production levels in the Δmlp/ΔtxtH/Δmlp (ΔΔΔ) mutant following complementation with the txtH, mlp, mlp, cdaX and clav_p1293 genes. VC, vector control.

Figure 5

RT‐PCR analysis of gene expression in Streptomyces scabies 87.22 and the ΔtxtH1 mutant. Reverse transcription reactions containing (+) or lacking (–) reverse transcriptase enzyme were used as a template for the PCR, while control (C) reactions contained water in place of the cDNA template. The number of cycles used for each set of gene‐specific primers is indicated. The gyrA gene encoding the DNA gyrase subunit A was included as a loading control.

It has been reported that MLPs from different pathways can functionally complement each other (Boll et al., 2011; Lautru et al., 2007; Mori et al., 2018; Schomer and Thomas, 2017; Wolpert et al., 2007; Zhang et al., 2010). In organisms where multiple MLPs are encoded in a single genome, the deletion of a single MLP often does not abolish the production of the cognate metabolite, but instead metabolite production is eliminated only when all copies of MLP‐encoding genes are removed from the host genome (Lautru et al., 2007; Wolpert et al., 2007). As the S. scabies genome harbours two additional MLP‐encoding genes, mlp and mlp it is possible that either or both MLPs might be able to partially compensate for the loss of txtH in the ΔtxtH mutant. When we deleted mlp from the wild‐type S. scabies chromosome (Fig. S3), thaxtomin A production was similar in the Δmlp mutant as compared to the wild‐type strain (data not shown). Deletion of both txtH and mlp resulted in thaxtomin A production levels that are similar or slightly reduced as compared to the ΔtxtH single mutant, whereas deletion of txtH and mlp abolished thaxtomin A production completely, and similar results were observed when all three MLP genes were deleted (Fig. 4B). Both the mlp and mlp genes were shown to be expressed in wild‐type S. scabies and in the ΔtxtH1 mutant under thaxtomin‐inducing conditions (Fig. 5), suggesting that the lack of thaxtomin A production in the ΔtxtH/Δmlp mutant was not due to a lack of transcription of the mlp gene. Interestingly, both MLPlipo and MLPscab were able to promote the soluble expression of the TxtA and TxtB A‐domains in E. coli, though the solubility‐promoting activity of MLPscab was less efficient for the His6‐TxtB‐A protein (Fig. 3C). This suggests that despite the inability of the ΔtxtH/Δmlp mutant to produce detectable levels of thaxtomin A, both MLPlipo and MLPscab have the ability to functionally replace TxtH in its interaction with the thaxtomin NRPS A‐domains. Further investigations will be required to determine the reason for the lack of detectable thaxtomin A production in the ΔtxtH/Δmlp mutant.

Engineered expression of MLPs in wild‐type S. scabies and in the MLP triple mutant

To further explore the ability of MLPs from different biosynthetic pathways to promote thaxtomin A production in the absence of txtH, we constructed several plasmids that overexpress different MLP‐encoding genes using the ermEp* promoter and then introduced them into the Δmlp/ΔtxtH/Δmlp mutant. As shown in Fig. 4D, overexpression of mlp and mlp from S. scabies restored thaxtomin A production in the triple mutant to levels similar to that observed when txtH was overexpressed, confirming that both MLPs can functionally replace txtH in the thaxtomin biosynthetic pathway. We note that overexpression of txtH, mlp and mlp also led to accumulation of the thaxtomin B and D biosynthetic intermediates (Fig. S5), confirming that there are some polar effects of the ΔtxtH mutation on expression of the downstream txtC gene. In contrast, overexpression of the MLP‐encoding genes cdaX from Streptomyces coelicolor and clav_p1293 from Streptomyces clavuligerus did not restore thaxtomin metabolite production in the S. scabies triple MLP mutant (Fig. 4D), suggesting that neither MLP can exhibit functional cross‐talk with TxtH. Both CdaX and Clav_p1293 localize in different phylogenetic clades from TxtH (Fig. 2B), though CdaX is predicted to be closely related to MLPlipo and MLPscab, both of which can exhibit cross‐talk with TxtH (Fig. 4D). Interestingly, a recent study by Schomer and Thomas (2017) also showed that while some non‐cognate MLPs are able to functionally replace the YbdZ MLP in the E. coli enterobactin biosynthetic pathway, others cannot, and no apparent correlation between MLP functionality and sequence similarity could be identified (Schomer and Thomas, 2017).

Previously, it was reported that the overexpression of cognate and non‐cognate MLPs in vivo increases vancomycin production in the high‐producing strain Amycolatopsis orientalis KFCC10990P (Lee et al., 2016). We investigated whether overexpression of txtH, mlp, mlp, cdaX and clav_p1293 in S. scabies 87.22 enhances thaxtomin A production in this strain; however, none of the overexpression strains produced significantly higher levels of thaxtomin A compared to the control strain (data not shown). Other studies have shown that an A‐domain requires a 1:1 molar ratio with its MLP partner for the maximum enzyme activity, and increasing the amount of MLPs beyond this optimal ratio did not stimulate the adenylating activity beyond a point (Boll et al., 2011; Zhang et al., 2010). Our results suggest that a similar situation may exist with TxtH and its cognate NRPS, though further investigations into this are needed.

Plant pathogenic phenotype of the S. scabies MLP mutants

We conducted a potato tuber slice assay in order to compare the virulence phenotype of the different S. scabies MLP mutant strains. As expected, S. scabies 87.22 readily colonized the surface of the potato tuber tissue and caused significant necrosis of the tissue after 10 days post‐inoculation (Fig. 6). The ΔtxtH and Δmlp/ΔtxtH mutants also colonized the tissue and induced tissue necrosis, though both strains were less efficient at doing so than the wild‐type strain. In contrast, there was very little visible growth of the Δmlp/ΔtxtH and Δmlp/ΔtxtH/Δmlp mutant strains on the tuber tissue, and both strains caused very little necrosis of the tissue (Fig. 6). Given that a positive correlation has been noted between the production of thaxtomin A and the virulence of scab‐causing Streptomyces spp. (Healy et al., 2000; King et al., 1991), the observed virulence phenotype of the different MLP mutant strains is consistent with the corresponding thaxtomin A production profiles observed in liquid culture (Fig. 4B). It remains to be determined whether production of the putative lipopeptide metabolite and the scabichelin siderophore are also affected in the MLP mutant strains and whether these metabolites also contribute to the pathogenicity of S. scabies. As siderophore production is known to contribute to the virulence phenotype of plant‐pathogenic bacteria (Franza et al., 2005; Taguchi et al., 2010), it will be interesting to further investigate the role of scabichelin in S. scabies plant pathogenicity.

Figure 6

Potato tuber slice assay for assessing the virulence phenotype of Streptomyces scabies strains. Tuber slices were inoculated with wild‐type and mutant S. scabies strains and were incubated for 10 days. Uninoculated medium (YMSm) was included as a negative control. The bioassay was performed twice in total and representative results are shown.

Concluding remarks

This study demonstrated the importance of TxtH in the biosynthesis of thaxtomin A in S. scabies. In particular, TxtH is required for promoting the soluble expression of both A‐domains from the thaxtomin NRPS in E. coli, suggesting that it performs a chaperone‐like role to enable the proper folding of the NRPS in S. scabies. Amino acid residues that contribute to the solubility‐promoting activity of TxtH have been revealed in this study, and future structural investigations will provide important insights into the role of these residues in mediating interactions between TxtH and the thaxtomin NRPSs. We also showed that MLPlipo from the putative lipopeptide biosynthetic pathway and MLPscab from the scabichelin biosynthetic pathway can functionally replace TxtH in the thaxtomin biosynthetic pathway, whereas two MLPs from other Streptomyces species cannot. Further investigations are required to better understand the mechanisms behind MLP cross‐talk and why certain MLPs from different pathways can functionally complement each other while others are unable to do so. Finally, our study confirmed that TxtH is important for the plant pathogenic phenotype of S. scabies.

Experimental Procedures

Bacterial strains, cultivation and maintenance

The bacterial strains used in this study are listed in Table 1. Escherichia coli strains were cultivated at 37 °C unless otherwise stated. Liquid cultures were grown with shaking (200 – 250 rpm) in Luria‐Bertani (LB) Lennox broth (Fisher Scientific, Ottawa, ON, Canada), low salt LB broth (1% w/v tryptone; 0.5% w/v yeast extract; 0.25% w/v NaCl), super optimal broth (SOB) or super optimal broth with catabolite repression (SOC) medium (New England Biolabs, Whitby, ON, Canada), while solid cultures were grown on LB Lennox (or low salt LB) medium containing 1.5% w/v agar. When necessary, the growth medium was supplemented with 50 μg/mL apramycin (Sigma Aldrich, Oakville, ON, Canada), 50 μg/mL kanamycin or hygromycin B (Millipore Sigma, Oakville, ON, Canada), or 25 μg/mL chloramphenicol (Acros Organics, Geel, Belgium) (final concentration). Escherichia coli strains were maintained at 4 °C for short‐term storage or at −80 °C as glycerol stocks for long‐term storage.

Streptomyces scabies strains were cultured at 28 °C unless otherwise indicated. Liquid cultures were typically grown with shaking (200 rpm) in trypticase soy broth (TSB; BD Biosciences, Mississauga, ON, Canada) medium with stainless‐steel springs. Plate cultures were routinely grown on potato mash agar (PMA; Fyans et al., 2015), nutrient agar (BD Biosciences) and soy flour mannitol (SFM) agar (Kieser et al., 2000). When necessary, the growth medium was supplemented with 50 μg/mL apramycin (Sigma Aldrich), 60 μg/mL nalidixic acid (Fisher Scientific), or 25 μg/mL thiostrepton (Sigma Aldrich) (final concentration). Seed cultures for RNA extraction were prepared by inoculating 100 μL of a S. scabies spore stock into 5 mL of TSB in a 50 mL spring flask followed by incubation for 48 h until dense mycelial growth was obtained. The seed cultures (50 μL) were then spread onto the surface of cellophane discs (75 mm diameter) on oat bran agar (Johnson et al., 2007) containing 0.35% w/v cellobiose (OBAC), after which the plates were incubated for 42 h. Cultures for analysis of thaxtomin A production were prepared by inoculating 50 μL of TSB seed cultures into 5 mL of oat bran broth containing 0.35% w/v cellobiose (OBBC; Johnson et al., 2009) in six‐well tissue culture plates (Fisher Scientific) and then incubating at 25 °C and 125 rpm for 7 days. Strains used for potato tuber slice bioassays were cultured at 28 °C for 14 days on yeast extract‐malt extract‐starch (YMS) agar (Ikeda et al., 1987) that had been modified by replacing the malt extract with Bacto Malt Extract Broth (BD Biosciences).

Plasmids, primers and DNA manipulation

The plasmids and cosmids used in this study are listed in Table 2. Standard molecular biology procedures were implemented for all DNA manipulations performed in this study (Sambrook and Russell, 2001). Restriction enzymes were purchased from New England Biolabs unless otherwise stated. PCR was routinely performed using Phusion DNA polymerase (New England Biolabs) according to the manufacturer’s instructions, except that 5% v/v DMSO was included in the reactions. All oligonucleotide primers used for cloning, PCR, site‐directed mutagenesis and sequencing were purchased from Integrated DNA Technologies (Coralville, IA, USA) and are listed in Table S3. DNA sequencing was performed by The Centre for Applied Genomics (Toronto, Canada). Streptomyces genomic DNA was isolated from mycelia harvested from 1–2 day‐old TSB cultures using the QIAamp® DNA mini kit as per the manufacturer’s protocol (QIAGEN Inc, Toronto, ON, Canada).

Table 2

Plasmids and cosmids used in this study.

Plasmid or cosmid	Description	Resistance†	Reference or source
pGEM‐T	General cloning vector	AmpR	Promega Corporation
pGEM‐T EASY	General cloning vector	AmpR	Promega Corporation
pGEM‐T EASY/Δmlplipo	pGEM‐T EASY derivative containing a 3725 bp insert with a deletion of the mlplipo gene	AmpR	This study
pET28b	N‐ or C‐terminal 6 × histidine fusion tag protein expression vector with T7 promoter and lac operator	KanR	Novagen
pET28b/His6‐txtH	pET28b derivative containing a DNA fragment for expression of the His6‐TxtH protein	KanR	This study
pET28b/His6‐mlplipo	pET28b derivative containing a DNA fragment for expression of the His6‐MLPlipo protein	KanR	This study
pET28b/His6‐mlpscab	pET28b derivative containing a DNA fragment for expression of the His6‐MLPscab protein	KanR	This study
pET28b/txtH	pET28b derivative containing a DNA fragment for expression of the untagged TxtH protein	KanR	This study
pACYCDuet‐1	N‐ terminal 6 × histidine fusion tag expression vector with T7 promoter and lac operator	CmlR	Novagen
pACYCDuet‐1/His6‐txtA‐A	pACYCDuet‐1 derivative containing a DNA fragment for expression of the His6‐TxtA‐A protein	CmlR	This study
pACYCDuet‐1/His6‐txtB‐A	pACYCDuet‐1 derivative containing a DNA fragment for expression of the His6‐TxtB‐A protein	CmlR	This study
pIJ12738	Conjugative plasmid, non‐replicative in Streptomyces, containing MCS and I‐SceI site	ApraR	Fernández‐Martínez and Bibb, 2014
pIJ12738/∆mlplipo	pIJ12738 derivative containing two flanking regions of mlplipo	ApraR	This study
pIJ12742	Conjugative plasmid containing the temperature‐sensitive replication origin and the codon optimized I‐SceI gene under the control of the strong constitutive ermEp* promoter	ThioR	Fernández‐Martínez and Bibb, 2014
pIJ773	Template for PCR amplification of the aac(3)IV‐oriT cassette used for PCR targeting	ApraR	Gust et al., 2003a
pIJ10700	Template for PCR amplification of the hyg‐oriT cassette used for PCR targeting	HygR	Gust et al., 2003b
Cosmid 1989	SuperCos1 derivative containing the S. scabies thaxtomin A biosynthetic gene cluster	AmpR, KanR	Zhang et al., 2016
Cosmid 57	SuperCos1 derivative containing the S. scabies mlpscab gene	AmpR, KanR	This study
Cosmid 1989/ΔtxtH	Cosmid 1989 derivative containing the aac(3)IV‐oriT cassette in place of the txtH gene	AmpR, KanR, ApraR	This study
Cosmid 57/Δmlpscab	Cosmid 57 derivative containing the hyg‐oriT cassette in place of the mlpscab gene	AmpR, KanR, HygR	This study
pRLDB50‐1a	Overexpression plasmid containing the strong constitutive ermEp* promoter	ApraR, ThioR	Bignell et al., 2010b
pRLDB50‐1a/txtH	pRLDB50‐1a derivative containing the S. scabies txtH gene	ApraR, ThioR	This study
pRLDB50‐1a/mlplipo	pRLDB50‐1a derivative containing the S. scabies mlplipo gene	ApraR, ThioR	This study
pRLDB50‐1a/mlpscab	pRLDB50‐1a derivative containing the S. scabies mlpscab gene	ApraR, ThioR	This study
pRLDB50‐1a/clav_p1293	pRLDB50‐1a derivative containing the S. clavuligerus clav_p1293 gene	ApraR, ThioR	This study
pRLDB50‐1a/cdaX	pRLDB50‐1a derivative containing the S. coelicolor cdaX gene	ApraR, ThioR	This study

AmpR, ApraR, KanR, CmlR, ThioR and HygR = ampicillin, apramycin, kanamycin, chloramphenicol, thiostrepton and hygromycin resistance, respectively.

Construction of protein expression plasmids

Plasmids were constructed for overexpression of TxtH in E. coli with and without an N‐terminal His6 tag as well as for overexpression of N‐terminal His6‐tagged MLPlipo and MLPscab. The txtH gene was PCR‐amplified using cosmid 1989 as template and using primers PL150 and PL36 for construction of the untagged TxtH expression plasmid, and PL35 and PL36 for construction of the His6‐TxtH expression plasmid. The resulting PCR products were directly cloned into the expression vector pET28b via the NdeI/EcoRI and NcoI/EcoRI restriction sites to give pET28b/His6‐txtH and pET28b/txtH, respectively. mlp was PCR‐amplified from genomic DNA using primers PL163 and PL164, and mlp was PCR‐amplified from cosmid 57 using primers PL165 and PL166. The PCR products were directly cloned into the NdeI/EcoRI restriction sites of pET28b to give pET28b/His6‐mlp and pET28b/His6‐mlp

Plasmids were also constructed for overexpression of the TxtA and TxtB A‐domains as N‐terminal His6‐tagged proteins. The DNA sequences encoding TxtA‐A and TxtB‐A were PCR amplified using the primer pairs PL37/PL38 and PL40/PL41, respectively, and using cosmid 1989 as template. The products were cloned into the pGEM‐T vector as per the manufacturer’s instructions, after which the inserts were released by digestion with EcoRI and HindIII and were cloned into similarly digested pACYCDuet‐1 to give pACYCDuet‐1/His6‐txtA‐A and pACYCDuet‐1/His6‐txtB‐A. The cloned inserts in all constructed expression vectors were verified by DNA sequencing.

Site‐directed mutagenesis of TxtH

Site‐directed mutagenesis of TxtH was performed using the QuikChange II Site‐directed Mutagenesis Kit (Agilent Technologies Canada, Inc., Mississauga, ON, Canada) as per the manufacturer’s instructions. Mutagenic primers for the desired mutation was designed online with QuikChange® Primer Design Program (https://www.genomics.agilent.com/primerDesignProgram.jsp). The desired mutations were verified by DNA sequencing.

Co‐expression of His6‐TxtA‐A and His6‐TxtB‐A with MLPs

The BL21(DE3) ybdZ:aac(3)IV bacterial strain was used for co‐expression of His6‐TxtA‐A or His6‐TxtB‐A with tagged or untagged MLP proteins. Strains containing either pACYCDuet‐1/His6‐txtA‐A or pACYCDuet‐1/His6‐txtB‐A with and without pET28b/txtH, pET28b/His6‐txtH (wild‐type or point mutants), pET28b/His6‐mlp or pET28b/His6‐mlp were grown overnight in 3 mL of LB medium supplemented with 1% glucose, apramycin and chloramphenicol. Kanamycin was additionally included for strains containing the MLP expression plasmids. The overnight cultures were subcultured (1% v/v) into 50 mL of fresh LB containing appropriate antibiotics, and the cultures were incubated at 37 °C and 200 rpm until the OD600 was 0.4–0.6. Then, the cells were induced with 1 mM isopropyl β‐d‐thiogalactopyranoside (IPTG) and were further incubated at 16 °C and 200 rpm for 48 h. Cells from 1 mL of culture were harvested by centrifugation and were resuspended in 200 μL of 50 mM Tris‐HCl (pH 8.0) containing 1 × cOmplete EDTA‐free protease inhibitor (Roche Diagnostics, Laval, QC, Canada). The cells were then lysed by sonication for 25 s (10 s pulses alternating with 10 s pauses, 40% Amp) and the cell debris was removed by centrifugation (1 min at 16 000 rpm). The supernatants containing soluble proteins were collected and the protein concentration was quantified using a Bradford protein assay kit (Fisher Scientific).

Western blot analysis

Soluble protein extracts (10 μg) were subjected to standard sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS‐PAGE) before being transferred to Amersham™ Hybond™ ECL membrane (GE Healthcare Canada Inc., Mississauga, ON, Canada) as per the manufacturer’s instructions. Membranes were blocked overnight in TBS‐T buffer (50 mM Tris‐HCl pH 7.6, 150 mM NaCl and 0.05% v/v Tween 20) containing 5% w/v skim milk, and were then incubated with 6 × His Epitope Tag Antibody (Fisher Scientific) at a 1:2000 dilution. The membranes were washed several times with TBS‐T buffer and were then incubated with the secondary antibody (Fisher Scientific) at a 1:2000 dilution. The membranes were processed using ECLTM western blotting detection reagent (GE Healthcare) and were visualized an ImageQuant LAS4000 Biomolecular Imager (GE Healthcare).

Construction of an MLP‐deficient strain of S. scabies

A markerless deletion mutant of the mlp gene was generated using the meganuclease I‐SceI system (Fernández‐Martínez and Bibb, 2014). A 1766 bp region upstream of mlp (5′ mlp) was amplified using S. scabies 87.22 genomic DNA as template and using primers PL3 and PL4 to generate a DNA fragment with terminal XbaI and BamHI sites. A 1959 bp region downstream of the gene (3′ mlp) was separately amplified using the same template and primers PL5 and PL6 to generate a DNA fragment with terminal BamHI and EcoRI sites. These two flanking fragments were each cloned into pGEM‐T EASY (Table 2) generating pGEM‐T EASY/5′ mlp and pGEM‐T EASY/3′ mlp, the inserts of which were confirmed by DNA sequencing. The 3′ mlp insert was then released following digestion with EcoRI and BamHI and was cloned into similarly digested pGEM‐T EASY/5′ mlp to generate pGEM‐T EASY/Δmlp which contained a 3725 bp insert with a deletion of the mlp gene. Next, the 3725 bp insert was released by digestion with XbaI and EcoRI and was cloned into similarly digested pIJ12738 to give pIJ12738/Δmlp, which was then introduced into S. scabies 87.22 by intergeneric conjugation with E. coli as described before (Kieser et al., 2000). Apramycin‐resistant exconjugants (assigned as 87.22/Δmlp_int) were selected and verified by PCR using primers PL62 and PL63. Then, the delivery vector pIJ12742 containing the codon optimized I‐SceI gene under the control of the ermEp* promoter was introduced into verified S. scabies 87.22/Δmlpint by conjugation with E. coli. The exconjugants were cultured on PMA at 37 °C in order to promote the loss of pIJ12742, which was confirmed by screening for sensitivity to thiostrepton. Spores of thiostrepton‐sensitive exconjugants were then serially diluted in sterile water and were plated onto PMA plates to obtain single colonies, which were then screened for sensitivity to apramycin. Successful deletion of mlp was confirmed by PCR (Fig. S3).

The Redirect PCR targeting system (Gust et al., 2003a,2003b) was used to construct the ΔtxtH, Δmlp/ΔtxtH, ΔtxtH/Δmlp and Δmlp/ΔtxtH/Δmlp mutant strains. The txtH gene on cosmid 1989 was replaced with an extended apramycin resistance cassette [aac(3)IV‐oriT] that was PCR‐amplified using pIJ773 as template and using primers DRB627 and DRB628. The mlp gene on cosmid 57 was replaced with an extended hygromycin resistance cassette [hyg‐oriT] that was PCR‐amplified using pIJ10700 as template and using primers PL153 and PL154. The ΔtxtH and Δmlp mutant cosmids were verified by PCR (Fig. S2; data not shown) and were then introduced into S. scabies by intergeneric conjugation with E. coli. The resulting mutant strains were analysed by PCR to confirm replacement of the target genes (Figs S1 and S4).

Construction of MLP overexpression plasmids

The txtH, mlp and mlp genes from S. scabies, together with the clav_p1293 and cdaX MLP‐encoding genes from S. clavuligerus and S. coelicolor, respectively, were PCR‐amplified using cosmid 1989 (for txtH) or genomic DNA (for mlp, mlp, clav_p1293 and cdaX) as template and using gene‐specific primers with BamHI and XbaI restriction sites added (Table S3). The resulting products were digested with BamHI and XbaI, and were ligated into similarly digested pRLDB50‐1a (Bignell et al., 2010b) to generate pRLDB50‐1a/txtH, pRLDB50‐1a/mlp, pRLDB50‐1a/mlp, pRLDB50‐1a/clav_p1293 and pRLDB50‐1a/cdaX (Table 2). The expression plasmids along with the control plasmid (pRLDB50‐1a) were then introduced into S. scabies 87.22 and the Δmlp/ΔtxtH/Δmlp mutant by intergeneric conjugation with E. coli.

Quantification of thaxtomin A production

Thaxtomin A was extracted from S. scabies OBBC cultures as described by Fyans et al. (2016). Quantification of thaxtomin A in the culture extracts was by reverse‐phase HPLC using a standard curve that was constructed from known amounts of a pure thaxtomin A standard (Sigma Aldrich). The thaxtomin A production levels were normalized using dry cell weights (DCWs) as described before (Fyans et al., 2016) and were reported as ng thaxtomin A/mg DCW. Statistical analysis of the results was conducted in Minitab 18 using one‐way ANOVAs with a posteriori multiple comparisons of least squared means performed using the Tukey test. P values ≤ 0.05 were considered statistically significant in all analyses.

LC‐HRESIMS analysis of S. scabies culture extracts

LC‐HRESIMS analysis of S. scabies culture extracts was performed at the Memorial University Centre for Chemical Analysis, Research and Training using an Agilent 1260 Infinity HPLC system interfaced to an Agilent 6230 orthogonal time‐of‐flight mass analyser. Separation was achieved using a ZORBAX SB‐C18 analytical column (4.6 × 150 mm, 5 μm particle size) held at a constant temperature of 40 °C and an isocratic mobile phase consisting of 30% acetonitrile and 70% water at a constant flow rate of 1.0 mL/min. Metabolites were monitored by absorbance at 380 nm and by electrospray ionization MS in negative ion mode.

Potato tuber slice bioassay

The virulence phenotype of S. scabies strains was assessed using a potato tuber slice bioassay as described before (Loria et al., 1995). Streptomyces scabies strains were cultured on modified YMS agar for 14 days until well sporulated. Agar plugs were then prepared from the plates and inverted onto the tuber slices. The tuber slices were incubated at room temperature (~22–25 °C) in the dark in a moist chamber and were photographed after 10 days. The assay was performed twice in total.

Total RNA isolation

Streptomyces scabies mycelia (100–200 mg) from 42 h OBAC plates were placed into sterile 1.7 mL microcentrifuge tubes and were flash frozen in a dry ice/ethanol bath and then stored at −80 °C. Total RNA was isolated using an innuPREP RNA Mini Kit 2.0 and a SpeedMill PLUS tissue homogenizer (Analytik Jena AG, Jena, Germany) as per the manufacturer’s instructions. The resulting RNA samples were treated with DNase I (New England Biolabs) as directed by the manufacturer to remove trace amounts of genomic DNA, after which the DNase‐treated RNA samples were quantified using a NanoDropTM 1000 Spectrophotometer (Fisher Scientific). The integrity of the RNA was confirmed by agarose gel electrophoresis using a 1.2% w/v RNase‐free agarose gel in 1 × TBE (Tris‐borate‐EDTA) buffer. The RNA samples were stored at −80 °C.

Reverse transcription PCR

Reverse transcription (RT) was performed using SuperScript IV reverse transcriptase (Fisher Scientific) with 2 µg of DNase‐treated total RNA and random hexamer primers as per the manufacturer’s instructions. A negative control reaction lacking the reverse transcriptase enzyme was included to verify the absence of genomic DNA in the RNA samples. RNA was removed from the synthesized cDNA by adding 1 uL of RNAse H and incubating at 37 °C for 20 min. PCR was performed using 2 μL of the cDNA template. Amplification was conducted using Taq DNA polymerase (New England Biolabs) with 1 × Standard Taq Reaction Buffer, 250 µM dNTPs, 0.5 µM of gene‐specific primers (Table S1) and 5% v/v DMSO. The PCRs were initiated by denaturing at 95 °C for 2 min followed by 22 (txtA, txtB, txtC, txtH), 25 (gyrA) or 27 (mlp, mlp) cycles of 95 °C for 15 s, 60 °C for 30 s and 68 °C for 15 s. After the amplification, 10 μL of each PCR product was analysed on a 1% agarose gel by electrophoresis.

Bioinformatics analysis

Identification of the adenylation domain within the TxtA and TxtB amino acid sequences was performed using the Pfam database (http://pfam.xfam.org/) (Finn et al., 2016). TxtH homologues were identified using the NCBI Protein Basic Local Alignment Search Tool (BLASTP) (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Amino acid sequence alignment of TxtH and other MLPs was performed using ClustalW within the Geneious v. 6.1.2 software (Biomatters Inc., Newark, NJ, USA). The accession numbers for the protein sequences used in the alignment are listed in Table S1. The MLP phylogenetic tree was constructed by maximum likelihood with the MEGA 7 software (Kumar et al., 2016) using the Whelan and Goldman plus gamma (WAG + G) substitution model (Whelan and Goldman, 2001). Bootstrap analyses were performed with 1000 replicates.

Fig. S1. PCR verification of the Streptomyces scabies txtH deletion mutants. (A) Schematic diagram showing the annealing sites of primers (indicated by the red arrows) used for the PCR verification. The expected product sizes for S. scabies 87.22 (wild‐type) and the ΔtxtH or Δmlp/ΔtxtH mutant strains are indicated. FRT, flip recombinase recognition sites. (B) Agarose gel electrophoresis of the PCR products generated using genomic DNA from S. scabies 87.22 (lane 7) and from the ΔtxtH mutant isolates (lanes 3–6). A negative control reaction (lane 2) was conducted for each primer set using water in place of template DNA. The size (kb) of each product was estimated by comparison with the 1 kb ladder (lane 1). (C) Agarose gel electrophoresis of the PCR products generated using genomic DNA from S. scabies 87.22 (lane 7) and from the ΔmlpΔtxtH mutant isolates (lanes 3–6). A negative control reaction (lane 2) was conducted for each primer set using water in place of template DNA. The size (kb) of each product was estimated by comparison with the 1 kb ladder (lane 1).

Click here for additional data file.

Fig. S2. PCR verification of the orientation of the extended apramycin resistance cassette in the ΔtxtH mutant cosmid. (A) Strategy used to verify the orientation of the aac(3)IV‐oriT cassette in cosmid 1989/ΔtxtH. The primers used for PCR amplification are indicated by the red arrows, and the expected product size is also shown. Cosmid 1989 lacks the binding site for the Apra For primer and thus should not generate a product. FRT, flip recombinase recognition sites. (B) Agarose gel electrophoresis of the PCR products generated using cosmid 1989/ΔtxtH (lanes 3 and 4) and cosmid 1989 (lane 5) as template. A negative control reaction (lane 2) was conducted using water in place of template DNA. The size (kb) of the products was estimated by comparison with the 1 kb ladder (lane 1).

Click here for additional data file.

Fig. S3. PCR verification of the Streptomyces scabies Δmlpmutant. (A) Schematic diagram showing the annealing sites of the primers (indicated by the red arrows) used for the PCR verification. The expected product sizes for S. scabies 87.22 (wild‐type) and the Δmlpmutant are indicated. The blue and green shaded areas represent the upstream and downstream regions used to construct the Δmlpdeletion plasmid. (B) Agarose gel electrophoresis of the PCR products generated using genomic DNA from S. scabies 87.22 (lane 9) and from the Δmlpmutant isolates 1–6 (lanes 3–8). A negative control reaction was conducted using water in place of template DNA (lane 2). The size (kb) of each product was estimated by comparison with the 1 kb ladder (lane 1) and with the 100 bp ladder (lane 10).

Click here for additional data file.

Fig. S4. PCR verification of the Streptomyces scabies Δmlpmutants. (A) Schematic diagram showing the annealing sites of the primers (indicated by the red arrows) used for the PCR verification. The expected product sizes for S. scabies 87.22 (wild‐type), ΔtxtH/Δmlpor ΔmlpΔtxtH/Δmlpmutant isolates are indicated. FRT, flip recombinase recognition sites. (B) Agarose gel electrophoresis of the PCR products generated using genomic DNA from S. scabies 87.22 (lane 6) and from the ΔtxtH/Δmlpmutant isolates (lanes 3–4). A negative control reaction was conducted for each primer set using water (lane 2) in place of template DNA, and a positive control was included for the PL71/PL72 primer set using Cosmid 57/Δmlp as template (lane 5). The size (kb) of each product was estimated by comparison with the 1 kb ladder for the PL71/PL72 primer set (lane 1) and with the 100 bp ladder for the PL155/PL156 primer set (lane 1). (C) Agarose gel electrophoresis of the PCR products generated using genomic DNA from S. scabies 87.22 (lane 8, left image; lane 7, right image) and from the Δmlp/ΔtxtH/Δmlp mutant isolates (lanes 3–6). A negative control reaction was conducted for each primer set using water in place of template DNA (lane 2), and a positive control was included for the PL71/PL72 primer set using cosmid 57/Δmlp as template (lane 7, left image). The size (kb) of each product was estimated by comparison with the 1 kb ladder for the PL71/PL72 primer set (lane 1) and with the 100 bp ladder for the PL155/PL156 primer set (lane 1).

Click here for additional data file.

Fig. S5. Complementation of the Streptomyces scabies MLP triple mutant. HPLC chromatograms of culture extracts from wild‐type S. scabies 87.22 (i), the triple MLP mutant (Δmlp/ΔtxtH/Δmlp) (ii), the triple MLP mutant containing plasmid pRLDB50‐1a (iii), the triple MLP mutant containing the txtH expression plasmid (iv), the triple MLP mutant containing the mlp expression plasmid (v) and the triple MLP mutant containing the mlp expression plasmid (vi). The peak corresponding to thaxtomin A in each chromatogram is indicated with the red asterisks, and the peaks corresponding to the thaxtomin B and thaxtomin D intermediates are indicated with ▼ and ∇, respectively.

Click here for additional data file.

Table S1. Accession numbers of MLP protein sequences used for constructing the amino acid alignment and phylogenetic tree.

Click here for additional data file.

Table S2. Pairwise comparison of amino acid identity (lower tier) and similarity (upper tier) for the MLPs included in this study

Click here for additional data file.

Table S3. Oligonucleotide primers used in this study

Click here for additional data file.