MIDAS: A modular DNA assembly system for ... - ACS Publications

termed MIDAS (for Modular Idempotent DNA Assembly System), can be used to ... a unique set of vectors, MIDAS can provide a flexible route to assemblin...
0 downloads 0 Views 3MB Size
Download PDF
Research Article Cite This: ACS Synth. Biol. 2018, 7, 1018−1029

pubs.acs.org/synthbio

MIDAS: A Modular DNA Assembly System for Synthetic Biology Craig J. van Dolleweerd,*,† Sarah A. Kessans,‡ Kyle C. Van de Bittner,‡,§ Leyla Y. Bustamante,‡,§ Rudranuj Bundela,‡ Barry Scott,∥ Matthew J. Nicholson,‡,§ and Emily J. Parker‡,§ †

Downloaded via UNIV OF SUNDERLAND on November 2, 2018 at 20:59:03 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Protein Science & Engineering, Callaghan Innovation, School of Biological Sciences, University of Canterbury, Private Bag 4800, Christchurch 8140, New Zealand ‡ Department of Chemistry, University of Canterbury, 20 Kirkwood Avenue, Christchurch 8041, New Zealand § Ferrier Research Institute, Victoria University of Wellington, Kelburn, Wellington 6012, New Zealand ∥ Institute of Fundamental Sciences, Massey University, Private Bag 11 222, Palmerston North 4442, New Zealand S Supporting Information *

ABSTRACT: A modular and hierarchical DNA assembly platform for synthetic biology based on Golden Gate (Type IIS restriction enzyme) cloning is described. This enabling technology, termed MIDAS (for Modular Idempotent DNA Assembly System), can be used to precisely assemble multiple DNA fragments in a single reaction using a standardized assembly design. It can be used to build genes from libraries of sequence-verified, reusable parts and to assemble multiple genes in a single vector, with full user control over gene order and orientation, as well as control of the direction of growth (polarity) of the multigene assembly, a feature that allows genes to be nested between other genes or genetic elements. We describe the detailed design and use of MIDAS, exemplified by the reconstruction, in the filamentous fungus Penicillium paxilli, of the metabolic pathway for production of paspaline and paxilline, key intermediates in the biosynthesis of a range of indole diterpenesa class of secondary metabolites produced by several species of filamentous fungi. MIDAS was used to efficiently assemble a 25.2 kb plasmid from 21 different modules (seven genes, each composed of three basic parts). By using a parts library-based system for construction of complex assemblies, and a unique set of vectors, MIDAS can provide a flexible route to assembling tailored combinations of genes and other genetic elements, thereby supporting synthetic biology applications in a wide range of expression hosts. KEYWORDS: synthetic biology, modular assembly, Golden Gate cloning, Type IIS, metabolic pathway engineering, indole diterpene

A

process using conventional (Type IIP) restriction enzymes, a set of design rules termed the BioBrick standard was developed,5 wherein the idea of idempotencywhere a reaction performed on a component produces a new structure, yet leaves that structure unchanged with respect to its ability to participate in further reactionswas proposed in the context of DNA assembly. More recently, modular DNA assembly techniques based on Golden Gate cloning,6 which utilizes the ability of Type IIS restriction enzymeswhich recognize non-palindromic sequences and cleave at one side of the recognition siteto seamlessly join multiple DNA fragments together in a single (one-pot) restriction−ligation reaction, have been described. Techniques such as GoldenBraid,7,8 the topologically equivalent TNT-Cloning,9 MoClo,10,11 Scarless Stitching,12 and the plasmid assembly system described by Binder et al.13 have used the idea of idempotency to extend the principles of Golden Gate cloning, thereby permitting modular and combinatorial

central requirement of synthetic biology is the ability to dissect biological systems into basic, reusable parts or modules and to reorganize and reassemble those parts, in a standardized way, to produce novel genetic modules, interacting proteins, control circuits, metabolic pathways, and high-value products. In recent years, numerous enabling technologies have emerged to address these requirements at the genetic level. Most of these technologies fall into four broad categories, surveyed extensively in recent years,1−4 each offering powerful features: (i) techniques which utilize site-specific recombinases to assemble or rearrange DNA molecules, (ii) techniques based on the in vitro assembly of DNA molecules containing overlapping sequences, (iii) in vivo techniques that utilize cellular, recombination-based mechanisms to assemble DNA molecules bearing homologous sequences, and (iv) techniques based on restriction enzymes. The ability to assemble devices from prefabricated parts or modules in a combinatorial fashion is a highly desirable feature of any DNA assembly system. Such modularity can provide flexibility in terms of mixing and matching of parts, as well as allowing reuse of assembled devices and exchange of parts among researchers. In an effort to standardize the assembly © 2018 American Chemical Society

Received: October 11, 2017 Published: April 5, 2018 1018

DOI: 10.1021/acssynbio.7b00363 ACS Synth. Biol. 2018, 7, 1018−1029

Research Article

ACS Synthetic Biology

Figure 1. Overview of MIDAS, demonstrating the modular and hierarchical nature of assembly. MIDAS permits basic functional modules such as promoters (ProUTR), coding sequences (CDS), and terminators (UTRterm) of genes of interest to be cloned, using BsmBI-mediated restriction− ligation reactions, into the Level-1 source vector. Cloned modules are then assembled into full length transcription units (TUs) in a Level-2 shuttle vector using BsaI-mediated restriction-ligation reactions. At Level-3, assembled TUs are combined, using alternating AarI- and BsmBI-mediated restriction−ligation reactions, to form multigene assemblies.

activity against a variety of cancer types including breast cancer17,18 and glioblastoma.19 The chemical diversity and associated bioactive properties of IDTs arise from the action of numerous enzymes that generate a core cyclic diterpene structure and decorate it with different substituents and stereochemistry. The genetic machinery and biosynthetic pathways for IDT production in several filamentous fungal species have been elucidated.20−32 In Penicillium paxilli the PAX cluster (Supporting Information, Figure S1a), containing genes involved in the synthesis of paxilline and other cyclic IDTs, has been described,23,24,33 with four of the genespaxG (encoding a geranylgeranyl diphosphate synthase), paxC (prenyltransferase), paxM (flavin-dependent monooxygenase), and paxB (indole diterpene cyclase)forming the core biosynthetic genes for the assembly of paspaline,22 a key intermediate in the biosynthesis of paxilline (Figure S1b). Two additional genes paxP and paxQ (both encoding cytochrome P450 monooxygenases)are required to catalyze the conversion of paspaline to paxilline.23 Here we demonstrate that MIDAS can be used to efficiently assemble these six pax genes from basic functional parts and show that these assembled genes can then be combined into a single multigene plasmid. We further show that the paxilline biosynthetic pathway can be reconstituted when a multigene plasmid harboring these six genes is used to transform a mutant P. paxilli strain deficient in the PAX cluster. In summary, MIDAS allows multigene assemblies to be constructed from libraries of basic parts in a hierarchical fashion, with control of gene order and orientation, as well as

assembly of genes from libraries of smaller fragments, and the subsequent assembly of multiple genes together in a single plasmid. Here we describe a new Golden Gate-based modular assembly system termed MIDAS (for Modular Idempotent DNA Assembly System), which comprises a set of vectors and design rules for assembling genes from basic, reusable parts and for assembling plasmids containing multiple genes. The modular nature of MIDAS cloning allows the user to create libraries of cloned genetic parts (e.g., libraries of promoters, coding sequences (CDSs) of genes of interest, transcriptional terminators, fusion tags, etc.) and to assemble these parts together to form full-length genes. MIDAS further allows the full-length genes to be assembled together in a single plasmid, with full user control over gene order and orientation, as well as control of the direction of growth of the multigene assembly, a feature that allows genes to be nested between other genes or genetic parts. Moreover, by incorporation of the appropriate parts (promoters, terminators, recombination sites, homology arms, insulators, tags, selection markers, etc.) the expression vectors built using MIDAS can be fully customized for different expression hosts. The MIDAS platform has obvious applications toward building and modifying biological pathways, e.g., for the efficient production of natural products of economic importance. One such class of products, the indole diterpenes (IDTs), are secondary metabolites that have attracted commercial interest because of their potent insecticidal properties,14,15 anti-H1N1 influenza activity,16 and biological 1019

DOI: 10.1021/acssynbio.7b00363 ACS Synth. Biol. 2018, 7, 1018−1029

Research Article

ACS Synthetic Biology

Figure 2. Principle of MIDAS Level-1 cloning. (a) The Golden Gate cloning cassette (lacZα flanked by divergent BsmBI sites) of the pML1 source vector. (b) A PCR product containing a promoter (ProUTR) module flanked by convergent BsmBI sites. The 4 bp sequences GGAG (blue) and AATG (red) represent, respectively, the prefix and suffix signatures unique to ProUTR modules (other modules will have differing prefix and suffix signatures; see Supporting Information, Figure S2). Following a BsmBI-mediated Golden Gate reaction between pML1 and the amplified module, a plasmid consisting of the module cloned into pML1 is obtained (c). Once cloned into pML1, each module becomes flanked by convergent BsaI recognition sites, and the prefix and suffix sequences form the BsaI-specific 4 bp overhangs when the module is released from pML1 during the subsequent (i.e., Level-2) BsaI-mediated Golden Gate assembly of the full-length TU.

containing a transcription unit (TU). Eight Level-2 vectors are employed by MIDAS to provide full user control over (i) the order in which TUs are added into the multigene assembly, (ii) the TU orientation (direction of transcription) in the multigene assembly and, (iii) uniquely to MIDAS, the direction of growth (polarity) of the multigene assembly, the latter feature allowing TUs to be nested between other TUs or sequence elements. At Level-3, the TUs that were assembled at Level-2 are released from the pML2 plasmids and are sequentially assembled together in a Level-3 destination vector (pML3), using alternating AarI- and BsmBI-mediated Golden Gate reactions, to form functional multigene constructs which can then be transformed into the desired expression host. Plasmids and full sequences of the 10 MIDAS toolkit vectors are publicly available from Addgene; their Addgene IDs are listed in Table S16. For simplicity, the description and exemplification of MIDAS presented here focuses on the assembly of TUs from basic parts and their incorporation into multigene assemblies. However, MIDAS is not restricted to assembly of TUs, and the same assembly principles can also be used to clone and assemble other sequence elements which do not form the basis of TUs. Such sequence elements could include origins of replication for propagation of plasmids (e.g., the yeast 2 μ replication origin), T-DNA left and right borders for Agrobacterium-mediated transformation of plant cells, homology arms for recombination-mediated gene replacement in the desired expression host, sequences involved in site-specific recombination, or indeed any other sequence element of interest.

control of the direction of growth of multigene assemblies. The ability to incorporate sequence elements besides transcription units allows the multigene expression vectors built using MIDAS to be customized for different expression hosts.

■

RESULTS AND DISCUSSION MIDAS Design Overview. The MIDAS toolkit is based on the Golden Gate assembly technique,6 which utilizes the ability of Type IIS restriction enzymes that generate user-defined overhangs upon cleavage to seamlessly join multiple DNA fragments together in a single reaction. MIDAS makes use of three Type IIS restriction enzymesAarI, BsaI, and BsmBI which generate user-defined 4 bp overhangs upon cleavage, with AarI having a rarer (7 bp) recognition sequence. As with other recently described Golden Gate-based modular assembly techniques,7−13 assembly of genes and multigene constructs using MIDAS is a hierarchical process (Figure 1). At the first level (MIDAS Level-1), functional modules (promoters, CDS, terminators, tags, etc.) are cloned into the Level1 source vector (pML1) using BsmBI-mediated reactions, where they form libraries of sequence-verified, reusable parts (Level-1 entry clones). The complementary design of the modules and source vector ensures that, once cloned into pML1, these modules can be released from the vector by digestion with BsaI. At the second level (Level-2), compatible sets of the sequence-verified Level-1 modules are released from the pML1 source vector and assembled into a Level-2 shuttle vector (pML2) using a BsaI-mediated Golden Gate reaction, leading to creation of a Level-2 plasmid (Level-2 entry clone) 1020

DOI: 10.1021/acssynbio.7b00363 ACS Synth. Biol. 2018, 7, 1018−1029

Research Article

ACS Synthetic Biology

Figure 3. Overview of MIDAS cassettes. The structures of the MIDAS cassettes in each of the eight pML2 shuttle vectors are depicted schematically and are shown classified according to the presence or absence of a lacZα marker (B and W, respectively) and according to polarity (+ and −). Within each MIDAS cassette, the dashed box shows the Golden Gate cloning cassette (comprised of divergent BsaI sites flanking a pheS negative selection marker) used for Level-2 assembly of TUs. See main text for a detailed description.

bases are incorporated as part of the PCR primer sequences (and not as part of the pML1 source vector), MIDAS is completely adaptable to the addition of new parts (i.e., new fusion site combinations) and part types. For example, using the module address standard described by Patron et al.,34 ProUTR parts can be broken down into smaller parts, fusion proteins can be constructed, and N- or C-terminal tags can be added to CDSs. Level-2: TU Assembly. At Level-2, compatible sets of cloned and sequence-verified Level-1 TUMs (for example, ProUTR, CDS, and UTRterm modules) are assembled into a pML2 shuttle vector using a BsaI-mediated Golden Gate reaction, leading to creation of a Level-2 plasmid (pML2 entry clone) containing a complete (i.e., full-length) transcription unit (TU). In MIDAS, there are eight Level-2 (pML2) shuttle vectors into which a TU can be assembled, the choice of which depends on the desired configuration of TUs in the multigene plasmid produced at Level-3, namely, (i) the order in which TUs are added to the multigene assembly, (ii) the orientation (direction of transcription) of each TU in the multigene plasmid, and (iii) the direction of growth (polarity) of the multigene plasmid. These features are discussed further below. The pML2 shuttle vectors are distinguished from one another by the arrangement of specific sequence features that are central to the operation of MIDAS. These sequence features, collectively called the MIDAS cassette (Figure 3), define the Level-2 assembly of TUs and govern the assembly of multigene constructs produced at Level-3. Each MIDAS cassette is defined by (i) having a Golden Gate cloning cassette with flanking, divergent BsaI recognition sites, (ii) differing arrangements of recognition sites for AarI and BsmBI, and (iii) the presence or absence of a lacZα scorable marker. These features are described in greater detail. In contrast to many published Golden Gate cloning cassettes, which typically harbor a lacZ scorable marker or the ccdB

By incorporation of the appropriate parts (promoters, terminators, recombination sites, homology arms, insulators, tags, selection markers, replication origins, etc.), the multigene expression vectors built using MIDAS can be fully customized for different expression hosts. In this work we have exemplified the principles of MIDAS design and assembly by reconstructing a metabolic pathway in the filamentous fungus Penicillium paxilli. Level-1: Module Cloning. At Level-1, functional transcription unit modules (TUMs), such as promoters, terminators, and coding sequences (CDSs) for the genes of interest, are generated, either as a PCR product or as a synthetic polynucleotide sequence (from a gene synthesis company), and are cloned into the Level-1 source vector (pML1) by BsmBImediated Golden Gate cloning (Figure 2). To enable subsequent (i.e., Level-2) assembly of full-length TUs, each TUM is designed to be flanked by four modulespecific, user-defined nucleotides at the 5′ end (prefix), and four different, module-specific, user-defined nucleotides at the 3′ end (suffix), which are included as part of the PCR primer sequences. As each TUM is defined by its flanking prefix and suffix nucleotides, these module-specific bases effectively form an address system for each TUM, and they determine its position and orientation within the assembled TU. The developers of MoClo and GoldenBraid2.0 have already worked in concert to develop a common syntax or set of standard addresses (referred to as “fusion sites” in the MoClo system and “barcodes” in GoldenBraid2.0) for a wide variety of TUMs to facilitate part exchangeability for plant expression,34 and this standard is also adopted here for MIDAS-based assembly of TUs for expression in filamentous fungi (see Figure S2). In this work we chose to assemble TUs for expression in P. paxilli from three basic parts: ProUTR, CDS, and UTRterm modules. Considerations for the design of PCR primers for amplifying ProUTR, CDS, and UTRterm TUMs are shown in the Supporting Information, Table S6. Since the module-specific 1021

DOI: 10.1021/acssynbio.7b00363 ACS Synth. Biol. 2018, 7, 1018−1029

Research Article

ACS Synthetic Biology

Figure 4. MIDAS principle of Level-3 (multigene) assembly. MIDAS multigene assembly involves the sequential addition of TUs into the Level-3 destination vector, pML3, to form a multigene construct. The Level-3 assembly proceeds by alternating Golden Gate reactions using TUs assembled in “White” and “Blue” pML2 shuttle vectors (left). The plasmid produced after each round of assembly (right) becomes the destination vector for the next round of assembly. For clarity, the vector backbones have been omitted, and only the TUs and Type IIS recognition sites are shown.

Level-2 Golden Gate assembly of TUs (it is reserved for the Level-3 cloning), but the choice of “Blue” or “White” vector into which a TU should be assembled must be made during Level-2 assembly of TUs as this will determine the order in which that TU can be added to the multigene construct at Level-3. Likewise, the AarI and BsmBI sites are also not used for Level-2 assembly of TUs; instead they are integral to the Level3 assembly of multigene constructs. These considerations, including the differences between the (+) and (−) vectors, are discussed further below, under the Level-3 description. The orientation (direction of transcription) of each TU can also be freely defined by assembling each TU in either a pML2 “Forward” vector (indicated by “F” in the plasmid name) or a pML2 “Reverse” vector (indicated by “R”), thereby allowing the researcher to create and test multigene assemblies harboring TUs in all possible relative orientations. Level-3: Assembly of Multigene Constructs. At MIDAS Level-3, TUs that were assembled in the pML2 plasmids are sequentially loaded (by binary assembly) into the Level-3 destination vector (pML3) to form the multigene construct. Assembly of multigene plasmids at Level-3 is crucially dependent on the relative configuration of the AarI and BsmBI restriction sites in the MIDAS cassettes located in the “Blue” and “White” pML2 shuttle vectors. As illustrated in Figure 4, the nesting and inversion of these restriction sites, which are switched between the “Blue” and “White” vectors, means that TUs assembled into “White” MIDAS cassettes can be inserted into “Blue” MIDAS cassettes using AarI-mediated Golden Gate reactions and, conversely, TUs assembled into “Blue” MIDAS cassettes can be cloned into “White” MIDAS cassettes using BsmBI-mediated Golden Gate reactions. This cycle of cloning (i.e., alternating between “White” and “Blue” pML2 entry

negative selection marker, those found in the pML2 shuttle vectors contain a mutant E. coli pheS gene (driven by the promoter of the E. coli gene for chloramphenicol acetyltransferase) and are flanked by divergent BsaI recognition sites. The Thr251Ala/Ala294Gly double mutant of the E. coli pheS gene used here confers high lethality to cells grown on LB media supplemented with the phenylalanine analogue 4-chloro-DLphenylalanine, 4CP.35 During BsaI-mediated Level-2 assembly of TUs, the mutant pheS gene is eliminated from the pML2 vectors and can therefore be used as a negative selection marker, so that only colonies containing a pML2 shuttle vector with an assembled TU will grow on the selection plates, and since it shows a dominant phenotype over the wild-type chromosomal allele, no specially designed strain is required, in contrast to vectors harboring the ccdB marker. The order in which each TU is loaded into the multigene plasmid (at Level-3) can be freely defined by assembling the TU in either a “Blue” (indicated by “B” in the plasmid name) or “White” (indicated by “W”) pML2 shuttle vector at Level-2. There are four “Blue” and four “White” pML2 shuttle vectors, defined by the presence or absence, respectively, of a lacZα gene in the MIDAS cassette (see Figure 3). The “Blue” and “White” vectors also differ in the relative configuration of the AarI and BsmBI restriction sites in their MIDAS cassettes. Thus, in the “Blue” vectors, the entire MIDAS cassette contains flanking, convergent BsmBI sites, and nested within is the lacZα gene flanked by divergent AarI sites. In the “White” vectors, the enzyme configuration is switched (the entire MIDAS cassette is flanked by convergent AarI sites, and nested within are two divergent BsmBI sites), and there is no lacZα gene. It is important to note that the lacZα chromogenic marker in the pML2 vectors is not used for blue/white screening during the 1022

DOI: 10.1021/acssynbio.7b00363 ACS Synth. Biol. 2018, 7, 1018−1029

Research Article

ACS Synthetic Biology

chosen for this work. Accordingly, a CDS module for the nptII gene was amplified, and, to drive expression of this gene, ProUTR and UTRterm modules from the Aspergillus nidulans trpC gene were also amplified (see Table S8). The exon/intron structure of each of the pax genes was left unchanged and, where necessary, PCR primers were designed to amplify module fragments for domestication purposes (i.e., removal of internal recognition sites for AarI, BsaI, and BsmBI). Domestication primers are listed in Table S8. The amplified full-length modules (and compatible sets of domesticated module fragments) were then cloned, by BsmBI-mediated Golden Gate assembly, into pML1. The results of these Level-1 cloning experiments are discussed further in the Supporting Information, Results section. The libraries of TUMs generated are shown in Table S10. Level-2 Assemblies. At Level-2, TUs for paxG, paxC, paxM, paxB, paxP, and paxQ were constructed by BsaIassembling each pax CDS module with its homologous (i.e., native) ProUTR and UTRterm modules in pML2 shuttle vectors (see Figure S6 and Table S11), and the assembled TUs are annotated with the name of the CDS they contain. For example, Level-2 entry clone pSK23, containing a paxB TU, was produced by assembling, in pML2(+)BR, the paxBCDS module (from plasmid pSK8) with the paxB promoter (i.e., the paxBProUTR module in pSK7) and paxB terminator (paxBUTRterm module in pSK9). Likewise, a paxP TU (pSK73) was assembled in pML2(+)BR from paxPProUTR (pSK75), paxPCDS (pSK69), and paxPUTRterm (pSK70) modules, while a paxQ TU (pSK74) was assembled in pML2(+)WR from paxQProUTR (pSK76), paxQCDS (pSK71), and paxQUTRterm (pSK72) modules. Since the desired order in which TUs are assembled into a Level-3 multigene assembly dictates whether it is assembled in a Blue or White pML2 shuttle vector, and because we wanted to test some TUs in different multigene arrangements, some TUs were assembled in both a White and a Blue pML2 shuttle vector. For example, a paxM TU was assembled in both pML2(+)WR and pML2(+)BR. Both paxM Level-2 entry plasmids (pSK22 and pRB1) were assembled from the same Level-1 entry clones (pSK4, pSK5, and pSK6). Two paxG TUs were assembled; in one case (plasmid pSK21), a paxG TU was produced by assembling, in pML2(+)BR, the paxGCDS module (from plasmid pSK2) with its native promoter (i.e., the paxGProUTR module in pSK1) and terminator (paxGUTRterm, pSK3). To demonstrate the versatility of MIDAS for combinatorial assembly, a second paxG TU (plasmid pSK47) was also assembled in pML2(+)BR using the same paxG CDS (paxGCDS, pSK2) and terminator (paxGUTRterm, pSK3), but with expression driven by the heterologous paxB promoter (paxBProUTR, pSK7), and the TU structure in pSK47 is shown as PpaxB-paxG-TpaxG. Three paxC TUs were assembled; in two cases (plasmids pSK59 and pKV29), the paxC TUs were assembled from the same Level-1 plasmids, by combining the paxC CDS (pSK11) with its native promoter (pKV28) and terminator (pSK12), albeit in different pML2 shuttle vectorspML2(+)WF in the case of pSK59, and pML2(+)BF in the case of pKV29. Once again, to demonstrate the versatility of MIDAS for combinatorial assembly, a third paxC TU (pSK61) was assembled in pML2(+)WF using the same paxC CDS, but using heterologous promoter (the trpCProUTR in pSK17) and terminator (trpCUTRterm, pSK15) modules. To distinguish this paxC TU from the other paxC TUs assembled from native

clones) can be repeated indefinitely and each plasmid generated by cloning a TU into the multigene construct becomes the destination vector for the next cycle of TU addition. Following each cloning cycle, positive clones are identified by blue/white screening. Thus, for TUs assembled into the multigene construct using AarI-mediated Golden Gate reactions, white colonies are picked for analysis, while TUs assembled into the multigene construct using BsmBI-mediated Golden Gate reactions are analyzed by picking blue colonies (see Table S7). In addition to control over the order and orientation of TUs in the multigene plasmid, the polarity (direction of growth) of the multigene assembly, i.e., the direction in which incoming TUs are loaded into the multigene plasmid, can be also be freely defined, by assembling each TU in a pML2 shuttle vector of either “plus” (+) or “minus” (−) polarity. The asymmetric arrangement of the AarI and BsmBI sites around the pheS marker, which differs between the (+) and (−) vectors (Figure 3), means that when a pML2(+) entry clone is used for Level-3 assembly, the TU loaded next in the multigene construct will be added to the right (downstream), as illustrated in Figure S3a,b. In contrast, use of a pML2(−) entry clone for Level-3 assembly forces the next TU to be added to the left (upstream). If entry clones of both polarity (i.e., both pML2(+) and pML2(−) entry clones) are used to build the multigene construct, then this confers MIDAS with the ability to switch the direction in which new TUs are loaded into the Level-3 assembly. This facility permits TUs to be nested between other TUs or other genetic elements, and for the hypothetical assembly shown in the Supporting Information, Figure S3c, all subsequently added TUs will be nested between TU3 and TU2. The nesting feature is an important aspect of MIDAS; for example, in the hypothetical assembly shown in Figure S3c, if TU1 and TU2 had, instead, been left and right homology arms, then this nesting feature would allow all subsequently added TUs to be placed between the homology arms and would permit recombination-mediated integration of those TUs into the chromosome of a suitable expression host. Metabolic Pathway Engineering. Gene reconstitution experiments using Penicillium paxilli strain PN2250 (CY2), which has a deletion of the entire PAX locus, have shown that four genes (paxG, paxC, paxM, and paxB) are required for the production of paspaline,22 a key intermediate in the P. paxilli biosynthetic pathway for paxilline and other cyclic IDTs, with two additional genes (paxP and paxQ) being required for the biosynthesis of paxilline.23 To demonstrate the utility of MIDAS in synthetic biology applications, we decided to test the system for its ability to restore the P. paxilli pathway for paspaline and paxilline biosynthesis in this PAX-deficient strain. Accordingly, we tested the ability of MIDAS to (i) assemble each of these six pax gene TUs from basic modules, (ii) assemble multigene plasmids containing up to six pax TUs (with each TU having the same relative position and orientation as found in the native PAX cluster), and (iii) in a series of complementation experiments, determine whether such plasmids could reconstitute paspaline and/or paxilline production in the PAX-deficient strain. Level-1 Assemblies. Transcription unit modules (ProUTR, CDS, and UTRterm) were amplified for each of the six pax genes using genomic DNA from P. paxilli strain PN2013 as template (Table S8). Since the aim was to produce multigene plasmids for transformation of P. paxilli strains, a suitable selectable markerneomycin phosphotransferase (nptII), conferring resistance to G418 (Geneticin)was 1023

DOI: 10.1021/acssynbio.7b00363 ACS Synth. Biol. 2018, 7, 1018−1029

Research Article

ACS Synthetic Biology

Figure 5. MIDAS Level-3 multigene plasmids produced in this work for reconstructing the paspaline (pSK63 and pSK64) and paxilline (pSK79) biosynthetic pathways in P. paxilli strain PN2250 (CY2). Plasmids were produced by sequentially loading the TUs for nptII, paxG, paxM, paxB, paxC, paxP, and paxQ into the pML3 destination vector. The order in which each TU was loaded is shown by the numerals above each TU. For clarity, the vector backbones have been omitted.

promoters and terminators, the paxC TU structure in pSK61 is shown as PtrpC-paxC-TtrpC. A nptII TU (conferring resistance to G418) was prepared by assembling the nptIICDS module (pSK16) with the trpCProUTR (pSK17) and trpCUTRterm (pSK15) modules in pML2(+)WF, giving rise to the Level-2 entry clone pSK26 (and the TU structure is shown as PtrpC-nptII-TtrpC). The results of these Level-2 cloning experiments are discussed further in the Supporting Information, Results section. Level-3 Assemblies. The TUs assembled at Level-2 were used to generate a variety of multigene plasmids at Level-3 (see Figure 5, Figure S8, and Table S12). The first step in creating multigene plasmids suitable for transformation of P. paxilli strains was to clone in the fungal selectable marker gene. Therefore, following Level-2 assembly, the nptII TU (harbored by pML2 entry clone pSK26) was loaded into pML3 using an AarI-mediated Golden Gate reaction. The resultant plasmid (pSK33) then served as the destination vector for sequentially assembling multigene plasmids harboring up to six pax genes. The results of these Level-3 cloning experiments are discussed further in the Supporting Information, Results section. Fungal Transformations and Analysis of Indole Diterpene (IDT) Phenotypes. In a series of complementation experiments, a selection of the Level-3 plasmids produced in this work were transformed into P. paxilli strains harboring appropriate genetic backgrounds. Paspaline and/or paxilline phenotypes of G418-resistant fungal transformants were determined by an initial thin-layer chromatography (TLC)

screen of mycelial extracts and confirmed by LC-MS. For these purposes, paspaline and paxilline reference standards were prepared from extracts of wild-type P. paxilli (strain PN2013) by semi-preparative HPLC. The HPLC peaks were analyzed by high-resolution mass spectrometry, which identified [M+H]+ masses of 422.3055 m/z at 17.6 min and 436.2485 m/z at 5.3 min, corresponding to the masses of paspaline (calcd [M+H]+ 422.3054 m/z) and paxilline (calcd [M+H]+ 436.2482 m/z), respectively (Figures S12 and S13, respectively). NMR was used to confirm the structures of the purified reference standards (see Table S15 and Figures S27−S36). Complementation of Single Gene Knockout Mutations. Single gene knockout (KO) mutants of the PAX cluster provide a useful resource for individually testing whether MIDAS-assembled pax TUs contain all the necessary cis-acting elements to successfully transcribe and express the pax CDSs. When Level-3 plasmids harboring individual pax TUs were transformed into appropriate P. paxilli KO strains, production of paspaline and paxilline was restored, indicating successful complementation of the single gene knockouts and showing, in turn, that each of the amplified promoters used to drive TU expression were functional (see Supporting Information, Results section). Restoration of Paspaline Production in a ΔPAX Deletion Mutant. Since genetic reconstitution experiments have shown that complementation using just four genes (paxG, paxM, paxB, and paxC) can restore paspaline production in P. paxilli strain PN2250 (CY2), which has a deletion of the entire PAX locus,22 we decided to test whether MIDAS-assembled 1024

DOI: 10.1021/acssynbio.7b00363 ACS Synth. Biol. 2018, 7, 1018−1029

Research Article

ACS Synthetic Biology

Figure 6. HPLC analysis of P. paxilli transformants. HPLC analysis (271 nm) was used for identification of IDTs extracted from fungal mycelia. (a) Overlaid HPLC traces of paspaline 4b and paxilline 6 reference standards. (b) HPLC trace of wild-type strain PN2013, showing the presence of paspaline 4b and paxilline 6. An HPLC trace of the ΔPAX mutant PN2250 (CY2) is shown (ci) and HPLC traces following transformation of this mutant with plasmids pSK64 (cii), pSK63 (ciii), pSK37 (civ), or pSK79 (cv). The colored arrows to the left of the HPLC traces show the genotypes of each P. paxilli strain, with the genotype of each transforming plasmid shown underneath. A red cross (X) indicates the pax gene(s) that have been deleted from the genome of the KO strains. The yellow, vertical dashed lines indicate the diagnostic elution peaks for paspaline 4b and paxilline 6 across each of the HPLC traces. As expected, paspaline 4b and paxilline 6 are absent from the parental ΔPAX KO strain PN2250 (CY2) as assessed by HPLC (trace ci) and EIC (Figure S15). For transformants pSK64:PN2250 and pSK63:PN2250 (traces cii and ciii, respectively), peaks corresponding to paspaline 4b were identified, albeit at low levels. Their identity was confirmed by the corresponding 422.305 ± 0.01 m/z EIC (Figures S23a and S24a, respectively). For transformant pSK79:PN2250 (trace cv), peaks corresponding to paspaline 4b and paxilline 6 were identified. Their identities were confirmed by the corresponding 422.305 ± 0.01 m/z and 436.248 ± 0.01 m/z EIC (Figure S25). Transformant pSK37:PN2250 served as a negative control and, as expected, no peak corresponding to paxilline was identified by HPLC (trace civ), nor in the EIC analysis Figure S26b). An apparent minor peak with a similar elution time to paspaline was identified by HPLC (trace civ), but the EIC analysis (Figure S26a) confirmed the absence of paspaline, as expected. The large peak eluting just prior to the paspaline 4b peak is not an indole diterpene, as it does not have the same absorption spectrum nor the same MS fragmentation pattern that corresponds to an indole-containing compound.

multigene plasmids harboring these four core genes could also restore production of this IDT in this ΔPAX strain. Accordingly, TUs for paxG (pSK21), paxM (pSK22), and paxB (pSK23), each under the control of their native promoters and terminators, were loaded sequentially into pSK33, followed by a paxC TU from either pSK59 (native promoter) or pSK61 (heterologous trpC promoter), to produce pSK64 and pSK63, respectively (see Figure 5 and Table S12). Because pSK64 and pSK63 differ only in the final added TU, it was possible to construct them from the same precursor plasmid, pSK37. When plasmids pSK64 or pSK63 were transformed into P. paxilli strain PN2250 (CY2), 8/18 (approximately 44%) of the G418-resistant colonies screened by thin layer chromatography (TLC) showed the presence of paspaline in their extracts (Figure S9). LC-MS analysis of transformant pSK64:PN2250#14 identified a peak with a retention time (see Figure 6, trace cii) and extracted-ion chromatogram (EIC) mass (Figure S23a) corresponding to that of paspaline. In the case of transformant pSK63:PN2250#8, a peak with the retention time of paspaline was barely detectable in the HPLC trace (Figure 6, trace ciii), but the EIC 422.305 ± 0.01 m/z (Figure S24a) confirmed the presence of paspaline. While it is interesting to speculate whether the apparent difference in

paspaline yield between the two constructs might be related to the use of the different promoters and terminators controlling expression of the paxC genes in the two plasmids (paxC expression in pSK63 was controlled by the Aspergillus nidulans trpC promoter and terminator, while paxC expression in pSK64 was driven by its native promoter and terminator), a proper assessment of the effect of these transcriptional control elements on IDT yield would require assessment of gene copy number and normalization of the amounts of IDTs produced against an internal measurement (such as fresh weight of cells or total soluble protein content) over a much larger sample of independent transformants, something we did not undertake in this work. Nevertheless, a qualitative assessment of these results demonstrated successful restoration of the paspaline biosynthetic pathway in this KO mutant when the four core pax genes (paxGMBC) were introduced on either of these MIDAS-assembled multigene plasmids. Due to the extensive deletion of the PAX cluster in parental strain PN2250 (CY2), which includes paxP and paxQ, as expected no paxilline was identified in the LC-MS analyses of the pSK64 and pSK63 transformants (see Figures S23b and S24b). Also as expected, neither the parental strain, PN2250 (CY2), nor PN2250 (CY2) transformed with plasmid pSK37 (which harbors paxGMB, but not paxC), showed evidence of 1025

DOI: 10.1021/acssynbio.7b00363 ACS Synth. Biol. 2018, 7, 1018−1029

Research Article

ACS Synthetic Biology

assembling TUs in pML2 shuttle vectors of either “plus” (+) or “minus” (−) polarity. The basic principles described for MIDAS Level-1 (module cloning) and Level-2 (TU assembly) are like those described for other Golden Gate-based modular assembly techniques.7−10,13 It is at the stage of multigene assembly (Level-3 in MIDAS) that the techniques effectively diverge, and it is this different approach to multigene assembly that influences the design of the plasmids used at the earlier levels, leads to major differences in the assembly processes and differences in the overall architecture, and confers each technique with its own advantages and drawbacks. In MIDAS, the direction of growth (polarity) of multigene assemblies is controlled not only by the compatibility of Type IIS overhangs on neighboring fragments and vector ends (as is the case with the other techniques), but also, and uniquely to MIDAS, by the asymmetric arrangement (switched between the (+) and (−) pML2 shuttle vectors) of Type IIS restriction sites (AarI and BsmBI) surrounding the marker (pheS). Use of both (+) and (−) pML2 shuttle vectors for multigene assembly has the advantage of imparting MIDAS with the ability to switch the direction of growth (polarity) of the multigene assembly. This feature allows TUs to be nested between other TUs or genetic elements and, for example, allows TUs to be loaded between homology arms for recombination mediated chromosomal integration. Another novel feature is that, at Level-3, components can be added iteratively into a single construct indefinitely (in theory), rather than having to move into higher levels to make longer and more complex constructs. A further advantage of MIDAS is the relatively small vector suite (10 plasmids), which is comparable to GoldenBraid2.0 (9 plasmids) and compares favorably to MoClo (minimum of 47 plasmids for full capability). One of the disadvantages of MIDAS is the potential requirement to assemble a transcription unit in a different Level-2 plasmid (White versus Blue) should the user wish to change the position of a TU from an odd to even location (or vice versa) within the multigene assembly. Another limitation relates to the binary nature of MIDAS multigene assembly (only one TU is added at a time to the growing multigene plasmid), which means that MIDAS is slower than some of the other techniques. We have illustrated the utility of MIDAS by constructing a series of multigene plasmids from basic TUMs. The largest and most complex plasmid produced in this workpSK79 (25.2 kb)was assembled from 21 different modules (seven TUs, each composed of three basic parts), and we showed that this plasmid could reconstitute the metabolic pathway for paxilline production when transformed into a P. paxilli strain devoid of the entire PAX cluster. Tagami et al. previously showed36 that the four core pax genes involved in paspaline biosynthesis in P. paxilli (paxGCMB) could be used to reconstruct the paspaline pathway when transformed into Aspergillus oryzae. Subsequent addition of paxP and paxQ resulted in the conversion of paspaline to paxilline. This organism has also been used for production of the IDTs aflatrem,29 penitrem,30 and shearinine.32 Saccharomyces cerevisiae has also been used as a heterologous host for production of novel IDTs,31 demonstrating that this yeast has the requisite genetic machinery to produce the early precursors that can feed into the pathway for IDT biosynthesis. In these previous studies, conventional restriction enzyme cloning and/or In-Fusion cloning37 was used to assemble plasmids harboring up to two genes, while

paspaline or paxilline in their extracts as assessed by HPLC (see Figure 6, traces ci and civ, respectively) and EIC analysis (Figures S15 and S26, respectively). This contrasts with the wild-type strain (PN2013), which shows LC-MS peaks for both paspaline 4b (17.6 min, 422.3055 m/z) and paxilline 6 (5.3 min, 436.2485 m/z) (see Figure 6, trace b, and Figure S14). Restoration of Paxilline Production in a ΔPAX Deletion Mutant. Since gene disruption and chemical complementation experiments have shown that paxP and paxQ are required for the biosynthetic conversion of paspaline to paxilline,33 we also tested whether MIDAS-assembled multigene plasmids harboring paxP and paxQ, in addition to the four core genes (paxGMBC), could restore production of paxilline in P. paxilli strain PN2250 (CY2). Therefore, plasmid pSK64 (harboring the four core pax genes involved in paspaline biosynthesis, all under the control of their native promoters) was loaded sequentially with a paxP TU (from pSK73) to produce pSK78, and then with a paxQ TU (from pSK74) to produce plasmid pSK79, which harbors six pax genes in total. Plasmid pSK79 was transformed into P. paxilli PN2250 (CY2) and, of the 15 G418 resistant lines screened by TLC, nine showed the presence of both paspaline and paxilline in their extracts (Figure S10). LC-MS analysis of transformant pSK79:PN2250#14 identified peaks with retention times (Figure 6, trace cv) and EIC masses (Figure S25) corresponding to those of paspaline and paxilline, thereby confirming successful restoration of the paxilline biosynthetic pathway in this KO strain.

■

DISCUSSION We have developed a highly flexible, modular DNA cloning system, termed MIDAS (for Modular Idempotent DNA Assembly System), that makes use of Golden Gate cloning6 for the ordered and hierarchical assembly of multigene constructs from basic standardized parts or modules. The MIDAS platform consists of an integrated suite of vectors designed to give maximum user control over the construction of transcription units (TUs) or other genetic elements from basic modular parts, as well as control of TU order and orientation, and the polarity of assembly of TUs in a multigene construct. Since the basic parts are physically stored (in the pML1 source vector), they form readily accessible libraries of defined and validated modules, from which the more complex structures (i.e., transcription units and multigene plasmids) can be assembled. Moreover, by incorporation of appropriate genetic elements, such as recombination sites, homology arms, insulators, tags, selection markers, replication origins, etc., the expression vectors built using MIDAS can be fully customized for different expression hosts. For Level-2 assembly, eight pML2 shuttle vectors are available to choose from. In its simplest configuration, MIDAS can achieve multigene assembly using only two pML2 shuttle vectors: one “White” vector and one “Blue” vector (Figure S3a). The full set of eight pML2 vectors are provided to enable maximum user control over (i) the order in which each TU is added to the growing multigene construct (by choice of “White” or “Blue” pML2 shuttle vector into which the TUs are assembled), (ii) the desired orientation (that is, the direction of transcription) of each TU (through the choice of “Forward” or “Reverse” pML2 shuttle vector), and (iii) the polarity of assembly, i.e., the direction in which incoming TUs are loaded into the multigene construct, by 1026

DOI: 10.1021/acssynbio.7b00363 ACS Synth. Biol. 2018, 7, 1018−1029

Research Article

ACS Synthetic Biology

HST08 Stellar cells (Clontech Laboratories, Inc.) were used for routine transformations and maintenance of all other plasmids (including source, shuttle, and destination vectors and all plasmids assembled at Level-1 and Level-2). Penicillium paxilli strains used in this study are shown in Table S1. Details of the media and reagents used for the fungal work are provided in the Supporting Information. Golden Gate Assembly Reactions. Protocols for Level-1, Level-2, and Level-3 Golden Gate assembly reactions are given in the Supporting Information. Fungal Protocols. Detailed protocols for the preparation and transformation of P. paxilli protoplasts are provided in the Supporting Information. Indole Diterpene Production and Extraction. Fungal transformants were grown in 25 mL of CDYE medium with trace elements for 7 days at 28 °C in shaker cultures (≥200 rpm), in 125 mL Erlenmeyer flasks capped with cotton wool. Mycelia were isolated from fermentation broths by filtration through nappy liners and transferred to 50 mL centrifuge tubes (Lab Serv, Thermo Fisher Scientific), and IDTs were extracted by vigorously shaking the mycelia (≥200 rpm) in 2-butanone for ≥45 min. Normal-Phase Thin-Layer Chromatography (TLC). The 2-butanone supernatant (containing extracted IDTs) was used for thin-layer chromatography (TLC) analysis on solid phase silica gel 60 aluminum plates (Merck). IDTs were chromatographed with 9:1 chloroform:acetonitrile or 8:2 dichloromethane:acetonitrile and visualized with Ehrlich’s reagent (1% (w/ v) p-dimethylaminobenzaldehyde in 24% (v/v) HCl and 50% ethanol). Reverse-Phase Liquid Chromatography−Mass Spectrometry (LC-MS). Samples were prepared for liquid chromatography−mass spectrometry (LC-MS) from selected transformants that were identified by TLC. Accordingly, a 1 mL sample of the 2-butanone supernatant (containing extracted IDTs) was transferred to a 1.7 mL microcentrifuge tube, and the 2-butanone was evaporated overnight. Contents were resuspended in 100% acetonitrile and filtered through a 0.2 μm membrane into an LC-MS vial. LC-MS samples were chromatographed on a reverse-phase Thermo Scientific Accucore 2.6 μm C18 (50 × 2.1 mm) column attached to an UltiMate 3000 Standard LC system (Dionex, Thermo Fisher Scientific) run at a flow rate of 0.200 mL/min and eluted with aqueous solutions of acetonitrile containing 0.01% formic acid using a multistep gradient method (Table S14). Mass spectra were captured through in-line analysis on a maXis II quadrupole-time-of-flight mass spectrometer (Bruker).

plasmid co-transformations and feeding of intermediate IDT metabolites were frequently used to achieve heterologous production of the final IDTs. In contrast, using MIDAS, we were able to reconstruct an entire IDT pathway in a single plasmid, and then, without recourse to plasmid co-transformations or supplementation of the growth media with precursor metabolites, we showed that this plasmid could reconstitute paxilline production when transformed into a suitable P. paxilli strain. MIDAS thus offers an efficient approach for biosynthetic pathway reconstruction, and, in a further demonstration of its power, we have recently used MIDAS to reconstruct, in P. paxilli, the biosynthetic pathway for production of nodulisporic acid F using genes from the filamentous fungus Hypoxylon pulicicidum.38 Since MIDAS easily allows new modules to be added to the libraries of basic parts, and permits rapid construction of new TUs by combinatorial shuffling of modules, it offers the potential for speeding up the construction of novel metabolic pathways and optimizing gene expression in heterologous hosts. Indeed, the ability to rapidly trial different combinations of basic parts and different combinations of genes using a simple set of design rules is a highly desirable feature of any synthetic biology platform. We envisage MIDAS to be both a versatile research tool for elucidating the genetic componentry of metabolic pathways and, through the efficient and flexible assembly of pathway elements, providing a powerful platform for production of economically important compounds.

■

MATERIALS AND METHODS Molecular Biology. Restriction endonucleases were purchased from New England Biolabs (NEB), except AarI, which was purchased from Thermo Fisher Scientific. T4 DNA Ligase, 10× T4 DNA Ligase buffer, and 10 mM ATP were from NEB. Primers and gBlocks were synthesized by Integrated DNA Technologies (IDT). Other synthetic polynucleotides were synthesized by Epoch Life Science, Inc. Plasmids pBB528 and pBB535 were a gift from Bernd Bukau (Addgene plasmids #27390 and #27392, respectively). Kits for purification of plasmid DNA and PCR products using spin-column protocols were purchased from Macherey-Nagel. Genomic DNA from Penicillium paxilli was isolated using the ZR Fungal/Bacterial DNA MicroPrep Kit from Zymo Research. All PCRs for the construction of the MIDAS source, shuttle, and destination vectors and for amplification of MIDAS modules were performed using Phusion High-Fidelity PCR Master Mix with HF Buffer (NEB). Construction of the MIDAS vectors is described in the Supporting Information. Isopropyl β-D-1thiogalactopyranoside (IPTG) was purchased from Calbiochem, 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (XGal) from PanReac AppliChem, and 4-chloro-DL-phenylalanine (4CP) from Sigma. Antibiotics used in this work were Geneticin (G418, from Sigma), kanamycin (PanReac AppliChem), and spectinomycin (Gold Biotechnology). The vector toolkit developed in this work is publicly available via the community repository Addgene (http://www.addgene. org). Addgene ID numbers for the plasmids are listed in Table S16. Bacterial and Fungal Strains. Routine growth of Escherichia coli was performed at 37 °C in LB broth. Chemically competent E. coli XL10-Gold Ultracompetent cells (Agilent Technologies) were used for transformation and maintenance of plasmids assembled at Level-3. Chemically competent E. coli

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acssynbio.7b00363. Additional results, materials and methods, Figures S1− S41, and Tables S1−S16, including oligonucleotide sequences (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Craig J. van Dolleweerd: 0000-0002-1380-6930 Emily J. Parker: 0000-0002-9571-9987 1027

DOI: 10.1021/acssynbio.7b00363 ACS Synth. Biol. 2018, 7, 1018−1029

Research Article

ACS Synthetic Biology Author Contributions

Multigene Expression, Gene Silencing and Silencing Rescue in Plants. PLoS One 9, e88218. (14) Hensens, O. D., Ondeyka, J. G., Dombrowski, A. W., Ostlind, D. A., and Zink, D. L. (1999) Isolation and structure of nodulisporic acid A1 and A2, novel insecticides from a Nodulisporium sp. Tetrahedron Lett. 40, 5455−5458. (15) Ondeyka, J. G., Helms, G. L., Hensens, O. D., Goetz, M. A., Zink, D. L., Tsipouras, A., Shoop, W. L., Slayton, L., Dombrowski, A. W., Polishook, J. D., et al. (1997) Nodulisporic Acid A, a Novel and Potent Insecticide from a NodulisporiumSp. Isolation, Structure Determination, and Chemical Transformations. J. Am. Chem. Soc. 119, 8809−8816. (16) Fan, Y., Wang, Y., Liu, P., Fu, P., Zhu, T., Wang, W., and Zhu, W. (2013) Indole-diterpenoids with anti-H1N1 activity from the aciduric fungus Penicillium camemberti OUCMDZ-1492. J. Nat. Prod. 76, 1328−1336. (17) Sallam, A. A., Houssen, W. E., Gissendanner, C. R., Orabi, K. Y., Foudah, A. I., and El Sayed, K. A. (2013) Bioguided discovery and pharmacophore modeling of the mycotoxic indole diterpene alkaloids penitrems as breast cancer proliferation, migration, and invasion inhibitors. MedChemComm 4, 1360. (18) Sallam, A. A., Ayoub, N. M., Foudah, A. I., Gissendanner, C. R., Meyer, S. A., and El Sayed, K. A. (2013) Indole diterpene alkaloids as novel inhibitors of the Wnt/β-catenin pathway in breast cancer cells. Eur. J. Med. Chem. 70, 594−606. (19) Abdullaev, I. F., Rudkouskaya, A., Mongin, A. A., and Kuo, Y.-H. (2010) Calcium-activated potassium channels BK and IK1 are functionally expressed in human gliomas but do not regulate cell proliferation. PLoS One 5, e12304. (20) Nicholson, M. J., Eaton, C. J., Starkel, C., Tapper, B. A., Cox, M. P., and Scott, B. (2015) Molecular Cloning and Functional Analysis of Gene Clusters for the Biosynthesis of Indole-Diterpenes in Penicillium crustosum and P. janthinellum. Toxins 7, 2701−2722. (21) Nicholson, M. J., Koulman, A., Monahan, B. J., Pritchard, B. L., Payne, G. A., and Scott, B. (2009) Identification of two aflatrem biosynthesis gene loci in Aspergillus f lavus and metabolic engineering of Penicillium paxilli to elucidate their function. Appl. Environ. Microbiol. 75, 7469−7481. (22) Saikia, S., Parker, E. J., Koulman, A., and Scott, B. (2006) Four gene products are required for the fungal synthesis of the indolediterpene, paspaline. FEBS Lett. 580, 1625−1630. (23) Scott, B., Young, C. A., Saikia, S., McMillan, L. K., Monahan, B. J., Koulman, A., Astin, J., Eaton, C. J., Bryant, A., Wrenn, R. E., et al. (2013) Deletion and gene expression analyses define the paxilline biosynthetic gene cluster in Penicillium paxilli. Toxins 5, 1422−1446. (24) Young, C., McMillan, L., Telfer, E., and Scott, B. (2001) Molecular cloning and genetic analysis of an indole-diterpene gene cluster from Penicillium paxilli. Mol. Microbiol. 39, 754−764. (25) Young, C. A., Bryant, M. K., Christensen, M. J., Tapper, B. A., Bryan, G. T., and Scott, B. (2005) Molecular cloning and genetic analysis of a symbiosis-expressed gene cluster for lolitrem biosynthesis from a mutualistic endophyte of perennial ryegrass. Mol. Genet. Genomics 274, 13−29. (26) Zhang, S., Monahan, B. J., Tkacz, J. S., and Scott, B. (2004) Indole-diterpene gene cluster from Aspergillus f lavus. Appl. Environ. Microbiol. 70, 6875−6883. (27) Motoyama, T., Hayashi, T., Hirota, H., Ueki, M., and Osada, H. (2012) Terpendole E, a Kinesin Eg5 Inhibitor, Is a Key Biosynthetic Intermediate of Indole-Diterpenes in the Producing Fungus Chaunopycnis alba. Chem. Biol. 19, 1611−1619. (28) Saikia, S., Takemoto, D., Tapper, B. A., Lane, G. A., Fraser, K., and Scott, B. (2012) Functional analysis of an indole-diterpene gene cluster for lolitrem B biosynthesis in the grass endosymbiont Epichloe festucae. FEBS Lett. 586, 2563−2569. (29) Tagami, K., Minami, A., Fujii, R., Liu, C., Tanaka, M., Gomi, K., Dairi, T., and Oikawa, H. (2014) Rapid reconstitution of biosynthetic machinery for fungal metabolites in Aspergillus oryzae: total biosynthesis of aflatrem. ChemBioChem 15, 2076−2080.

C.J.v.D. conceived MIDAS, constructed the MIDAS vector suite, and wrote the manuscript. B.S., M.J.N., and E.J.P. conceived the fungal study. S.A.K., K.C.V.d.B., L.Y.B., and R.B. performed the experiments. C.J.v.D.,. K.C.V.d.B. and E.J.P. edited and revised the manuscript. All authors reviewed and analyzed the data. Notes

The authors declare the following competing financial interest(s): A patent application (AU2017903955) has been filed for the vector system described in this article.

■

ACKNOWLEDGMENTS We would especially like to thank Arvina Ram for her invaluable training and expert advice in the preparation of protoplasts and transformation of P. paxilli. The authors would also like to thank Dr. Amelia Albrett for expert assistance with the LC-MS, and Alistair Richardson for assistance with NMR. K.C.V.d.B. is a recipient of a Fulbright New Zealand Scholarship and a University of Canterbury Doctoral Scholarship. This work was supported by the University of Canterbury and funding from the New Zealand Ministry of Business, Innovation, and Employment (UOCX1405).

■

REFERENCES

(1) Kelwick, R., MacDonald, J. T., Webb, A. J., and Freemont, P. (2014) Developments in the tools and methodologies of synthetic biology. Front. Bioeng. Biotechnol. 2, 60. (2) Patron, N. J. (2014) DNA assembly for plant biology: techniques and tools. Curr. Opin. Plant Biol. 19, 14−19. (3) Chao, R., Yuan, Y., and Zhao, H. (2015) Recent advances in DNA assembly technologies. FEMS Yeast Res. 15, 1−9. (4) Casini, A., Storch, M., Baldwin, G. S., and Ellis, T. (2015) Bricks and blueprints: methods and standards for DNA assembly. Nat. Rev. Mol. Cell Biol. 16, 568−576. (5) Knight, T. F. (2003) Idempotent Vector Design for Standard Assembly of Biobricks, DSpace@MIT, http://hdl.handle.net/1721.1/ 21168. (6) Engler, C., Kandzia, R., and Marillonnet, S. (2008) A One Pot, One Step, Precision Cloning Method with High Throughput Capability. PLoS One 3, e3647. (7) Sarrion-Perdigones, A., Falconi, E. E., Zandalinas, S. I., Juarez, P., Fernandez-del-Carmen, A., Granell, A., and Orzaez, D. (2011) GoldenBraid: An Iterative Cloning System for Standardized Assembly of Reusable Genetic Modules. PLoS One 6, e21622. (8) Sarrion-Perdigones, A., Vazquez-Vilar, M., Palaci, J., Castelijns, B., Forment, J., Ziarsolo, P., Blanca, J., Granell, A., and Orzaez, D. (2013) GoldenBraid 2.0: A Comprehensive DNA Assembly Framework for Plant Synthetic Biology. Plant Physiol. 162, 1618−1631. (9) De Paoli, H. C., Tuskan, G. A., and Yang, X. (2016) An innovative platform for quick and flexible joining of assorted DNA fragments. Sci. Rep. 6, 19278. (10) Weber, E., Engler, C., Gruetzner, R., Werner, S., and Marillonnet, S. (2011) A Modular Cloning System for Standardized Assembly of Multigene Constructs. PLoS One 6, e16765. (11) Werner, S., Engler, C., Weber, E., Gruetzner, R., and Marillonnet, S. (2012) Fast track assembly of multigene constructs using Golden Gate cloning and the MoClo system. Bioengineered 3, 38−43. (12) Smanski, M. J., Bhatia, S., Zhao, D., Park, Y., Woodruff, L. B. A., Giannoukos, G., Ciulla, D., Busby, M., Calderon, J., Nicol, R., Gordon, D. B., Densmore, D., and Voigt, C. A. (2014) Functional optimization of gene clusters by combinatorial design and assembly. Nat. Biotechnol. 32, 1241−1249. (13) Binder, A., Lambert, J., Morbitzer, R., Popp, C., Ott, T., Lahaye, T., and Parniske, M. (2014) A Modular Plasmid Assembly Kit for 1028

DOI: 10.1021/acssynbio.7b00363 ACS Synth. Biol. 2018, 7, 1018−1029

Research Article

ACS Synthetic Biology (30) Liu, C., Tagami, K., Minami, A., Matsumoto, T., Frisvad, J. C., Suzuki, H., Ishikawa, J., Gomi, K., and Oikawa, H. (2015) Reconstitution of biosynthetic machinery for the synthesis of the highly elaborated indole diterpene penitrem. Angew. Chem., Int. Ed. 54, 5748−5752. (31) Tang, M. C., Lin, H. C., Li, D., Zou, Y., Li, J., Xu, W., Cacho, R. A., Hillenmeyer, M. E., Garg, N. K., and Tang, Y. (2015) Discovery of Unclustered Fungal Indole Diterpene Biosynthetic Pathways through Combinatorial Pathway Reassembly in Engineered Yeast. J. Am. Chem. Soc. 137, 13724−13727. (32) Liu, C., Minami, A., Dairi, T., Gomi, K., Scott, B., and Oikawa, H. (2016) Biosynthesis of Shearinine: Diversification of a Tandem Prenyl Moiety of Fungal Indole Diterpenes. Org. Lett. 18, 5026−5029. (33) McMillan, L. K., Carr, R. L., Young, C. A., Astin, J. W., Lowe, R. G., Parker, E. J., Jameson, G. B., Finch, S. C., Miles, C. O., McManus, O. B., et al. (2003) Molecular analysis of two cytochrome P450 monooxygenase genes required for paxilline biosynthesis in Penicillium paxilli, and effects of paxilline intermediates on mammalian maxi-K ion channels. Mol. Genet. Genomics 270, 9−23. (34) Patron, N. J., Orzaez, D., Marillonnet, S., Warzecha, H., Matthewman, C., Youles, M., Raitskin, O., Leveau, A., Farre, G., Rogers, C., et al. (2015) Standards for plant synthetic biology: a common syntax for exchange of DNA parts. New Phytol. 208, 13−19. (35) Miyazaki, K. (2015) Molecular engineering of a PheS counterselection marker for improved operating efficiency in Escherichia coli. BioTechniques 58, 86−88. (36) Tagami, K., Liu, C., Minami, A., Noike, M., Isaka, T., Fueki, S., Shichijo, Y., Toshima, H., Gomi, K., Dairi, T., and Oikawa, H. (2013) Reconstitution of biosynthetic machinery for indole-diterpene paxilline in Aspergillus oryzae. J. Am. Chem. Soc. 135, 1260−1263. (37) In-Fusion HD Cloning Kit User Manual (2014), Clontech Laboratories, Inc. (38) Van de Bittner, K. C., Nicholson, M. J., Bustamante, L. Y., Kessans, S. A., Ram, A., van Dolleweerd, C. J., Scott, B., and Parker, E. J. (2018) Heterologous Biosynthesis of Nodulisporic Acid F. J. Am. Chem. Soc. 140, 582−585.

1029

DOI: 10.1021/acssynbio.7b00363 ACS Synth. Biol. 2018, 7, 1018−1029