UniClo: scarless hierarchical DNA assembly without sequence constraint

Abstract

Type IIS restriction enzyme-mediated DNA assembly is efficient but has sequence constraints and can result in unwanted sequence scars. To overcome these drawbacks, we developed UniClo, a type IIS restriction enzyme-mediated method for universal and flexible DNA assembly. This is achieved through a combination of vector engineering, DNA methylation using recombinant methylases, and steric blockade using CRISPR-dCas9 technology to regulate this methylation. Type IIS restriction enzyme sites within fragments to be assembled are methylated using recombinant methylases, while the fragment-flanking outer sites are protected from methylation by a recombinant dCas9-sgRNA complex. During the subsequent assembly reaction, only the protected flanking sites are cut as only they are unmethylated. Fragments are correctly assembled, despite containing internal sites for the single type IIS restriction enzyme used for the one-pot assembly. The assembled plasmid can be used as a donor plasmid in a subsequent assembly round with the same type IIS restriction enzyme and the assembly vector engineering ensures removal of potential scars by a trimming process. This simplifies assembly design and only three vectors are required for any multi-round assembly. These vectors all use the same pair of overhangs. UniClo provides a simple scarless approach for hierarchical assembly of any sequence and has wide potential application.

Graphical Abstract

Figure 1.

Methylase and type IIS restriction enzyme sites and engineering for DNA assembly. (A) A switch methylase is a methylase whose recognition site overlaps partially, but not completely, with a type IIS restriction enzyme recognition site. The design of an original MetClo assembly vector includes two outer and two inner sites for the same type IIS restriction enzyme. The outer sites are “methylation-switchable,” because they are recognized by a switch methylase which can methylate them and so inhibit cutting of these sites by the type IIS restriction enzyme. In contrast, the inner sites are “always-cuttable” as they are not recognized by the switch methylase and so are never methylated by it and are always cut by the restriction enzyme. UniClo vectors also include an sgRNA target sequences which as shown. (B) A nonswitchable methylase is a methylase whose recognition site overlaps exactly with the type IIS restriction enzyme recognition site. Therefore, in the presence of the methylase all such restriction enzyme recognition sites will be methylated and so protected from cutting by the restriction enzyme. We term these sites “always-methylatable” and they can occur within any DNA such as in fragments to be assembled. We hypothesized that access of a nonswitchable methylase to an always-methylatable site could be sterically blocked using a CRISPR–dCas9 complex. We term this steric blockade approach “methylation-protection,” and we term the sites that can be protected from methylation “methylation-protectable” sites. Sites can be rendered methylation-protectable by generating an RNA-guided CRISPR–dCas9 complex that binds specifically to the restriction enzyme recognition site and the neighboring DNA. We term the recombinant methylation-protection RNA-guided CRISPR–dCas9 complex MP-dCas9–sgRNA A plasmid acting as a donor plasmid in a DNA assembly contains a fragment flanked by outer restriction enzyme sites which are methylation-protectable and within the fragment there may be one or more internal sites which will be always-methylatable. These sites are always-methylatable because they do not have targets sites for the MP-dCas9–sgRNA complex. The methylation-protection approach could be used to protect the outer sites from methylation while a nonswitchable methylase is used to methylate all the internal sites. (C) Before a UniClo assembly reaction the fragments to be assembled are methylated at any internal recognition site for the restriction enzyme used in the assembly. This methylation is carried out in the presence of the MP-dCas9–sgRNA which prevents methylation of the outer flanking restriction enzyme sites. These internal restriction enzyme sites are not cut during the assembly reaction because they are methylated. However, the outer methylation-protectable are cut during the assembly reaction releasing the intact fragments to be assembled. In parallel, during a UniClo assembly reaction, the assembly vector is cut from the outer sites which are not methylated. A scarless assembled plasmid is obtained which can be used as a new donor plasmid using the methylation-protection approach again.

Figure 2.

Methylation-protection. (A) The MP-dCas9–sgRNA is composed of a deactivated Cas9 nuclease (dCas9) bound to an sgRNA, which conducts the dCas9 to its a target DNA. (B) The sgRNA includes a variable target-specific spacer (20 nt) which is complementary to the target DNA and a crRNA–tracrRNA constant region. An sgRNA DNA template is first assembled by PCR amplification using standard and custom primers as illustrated. The sgRNA is then generated by in vitro transcription of the amplified DNA template. (C) “Methylation-protectable” sites are protected from methylation by the MP-dCas9–sgRNA and so can subsequently be digested by the type IIS restriction enzyme. The MP-dCas9–sgRNA allows the site-selective methylation-protection of a type IIS restriction site that has been suitable engineered with the target DNA sequence for the MP-dCas9–sgRNA and so is a methylation-protectable site. The MP-dCas9–sgRNA target sequence (purple) is engineered to be in close proximity to the nonswitchable methylase site, which overlaps completely with the BsaI site (GGTCTC, shown in green). The target sequence extends from the BsaI site outwards (that is away from where a fragment might be inserted. The PAM sequence (NGG, shown underlined) is positioned within the BsaI site (which is also the nonswitchable methylase site). The target sequence also includes part of recognition site for the switch methylase M.Osp807II (GACNNNGTC, shown in bold) which allows the methylation-protectable BsaI site to also be used as a methylation-switchable BsaI site in a DNA assembly vector. (D) “Always-methylatable” sites are not targets for MP-dCas9–sgRNA and so are not protected from methylation and will always be methylated and therefore not digested by the restriction enzyme. The “always-methylatable” site illustrated has a BsaI site (GGTCTC, green) that overlaps completely with a site for the nonswitchable methylase and so is always methylated in the presence of the methylase. The nucleotides (shown in orange) in the equivalent position to the target sequence of the methylation-protectable site are not recognized by the sgRNA around and the nucleotides (underlined) in the equivalent position to the PAM sequence do not form constitute an effective PAM sequence.

Figure 3.

Scar formation in a hierarchical assembly with Golden Gate/MoClo-based DNA assembly methods using type IIS restriction enzymes. (A) Schematic representation of a DNA fragment (Fragment 1) in a donor plasmid for an assembly reaction using BsaI, where the four letters aaaa and bbbb represent the overhangs of Fragment 1. (B) Schematic representation of two rounds of an assembly of nine fragments using BsaI. In the first round, the fragments are assembled in groups of three. The overhangs between fragments (bbbb, cccc, eeee, ffff, hhhh, and iiii) can be freely chosen from within the sequence to be assembled. Thus, no scars are formed within the assembled fragments. However, the overhangs at the end of the assembled fragments (aaaa, dddd, gggg, and jjjj) must match the overhangs from the assembly vectors. In the second round, these assembled fragments, each containing three original fragments, are assembled to produce the final assembled fragment. Here and in any subsequent rounds, since the overhangs at each end of the assembled fragments are derived from the assembly vectors, they become incorporated into the assembled DNA as short unwanted scars. In this round the scars will be dddd and gggg within the assembled fragment and aaaa and jjjj at the ends of the assembled fragment. It is necessary to have multiple assembly vectors because the sets of terminal overhangs in each assembly in any one round must differ from each other because they will determine the order of assembly in the next round.

Figure 4.

DNA assembly using a standard nontrimming assembly vector. The inner and outer BsaI sites are symmetrically distributed around the overhangs. In the first round of assembly, the outer site is methylated and inactive. During the first round of the assembly cutting from the inner site removes the stuffer fragment (brown) which contains LacZalpha and is replaced with the assembled fragment (pink). In a second round, the outer sites are not methylated and so are active. The assembled fragment (pink) that is released from the assembled plasmid, is identical to the fragment that was assembled in the first round. The overhangs at each end of this assembled fragment become incorporated as short scars between the assembled fragments in the second round (see also Fig. 3).

Figure 5.

DNA assembly using a scarless VM assembly vector that trims vector adaptor sequences at each end of an assembled fragment. The outer BsaI sites are nearer to the overhangs that form during the assembly than the inner sites. The inner sites remain in the same position in relation to the overhang as the inner sites in the standard vector shown in Fig. 4. The outer BsaI sites are themselves involved in the overhangs that form during the assembly and so are incomplete until the assembly occurs. The stuffer fragment (brown) which contains LacZalpha is removed during the assembly and replaced with the assembled fragment (pink). When the assembled fragment is assembled into the assembly vector, the outer sites are reconstituted in the assembled plasmid. In a subsequent second round of assembly, the fragment (dark pink) that is released from the assembled plasmid is not the same as the fragment that was assembled in the first round. Moreover, the vector adaptor sequences and their overhangs that were used in the first round have been trimmed off. Therefore, there are no scars formed between the fragments assembled in the second round.

Figure 6.

DNA assembly using a scarless VL assembly vector. The design of this vector is a hybrid of the standard vector and VM. At the left end, the inner and outer BsaI sites are symmetrically placed around the overhang. At the right end, the inner and outer sites are asymmetrically distributed with the outer site nearer than the inner site to the overhang. As a consequence, the assembled fragment (pink) from the first round of assembly differs from the fragment released during the second round (dark pink) in that the vector adaptor sequence from the right end has been trimmed whereas the left end has been retained.

Figure 7.

DNA assembly using a scarless VR assembly vector. The design of this vector is a hybrid of the standard vector and VM. At the right end, the inner and outer BsaI sites are symmetrically placed around the overhang. At the left end, the inner and outer sites are asymmetrically distributed with the outer site nearer than the inner site to the overhang. As a consequence, the assembled fragment (pink) from the first round of assembly differs from the fragment released during the second round (dark pink) in that the vector adaptor sequence from the left end has been trimmed whereas the right end has been retained.

Figure 8.

General scheme for scarless hierarchical DNA assembly of fragments with internal type IIS restriction sites using UniClo. (A) Illustration of how a fragment, Fragment 1, would be cut from a donor plasmid of VL, VM, or VR type. The fragments have the vector adaptors sequences CCTT and TGAGAC required for assembly using the scarless three-vector set. The four letters aaaa and bbbb represent overhangs that could be generated from within the sequence of Fragment 1. CTCC and AGAC are the overhangs generated from the vector adaptor sequences and are the required overhangs for the end of an assembled fragment to be compatible with the overhangs in the assembly vector (whether VL, VM, or VR type). (B) Schematic representation of two rounds of assembly of nine fragments using BsaI. In the first round, the fragments are assembled in groups (three in this example, but more can be used) in assembly vectors VL, VM, and VR. Using this scarless three-vector set, the overhangs of the vector adaptor sequences are trimmed off as appropriate according to the vector type used. In the second round, these assembled fragments are further assembled into the final assembled fragment. The vector adaptor sequences from the first round of assembly are trimmed off at both ends generating a fully scarless assembled fragment. The vector adaptor sequences are removed in each round of the assembly. This assembly design can be used for fragments with internal BsaI sites as these will be methylated by a recombinant nonswitchable methylases, while the outer BsaI sites are protected from methylation by recombinant MP-dCas9–sgRNA.

Figure 9.

Scarless hierarchical assembly of a 10.8 kb DNA fragment containing 39 internal BsaI sites using UniClo. (A) Scheme of the two-round assembly. In the first round, 11 fragments, each in a donor plasmid, were assembled in groups of 4, 3, 2, and 2 resulting in four assembled plasmids. In the second round, these four assembled plasmids were used as donor plasmids for the assembly of the final assembled fragment of 10.8 kb. (B) An example of one of the 11 donor plasmids from the first round. POC1545 containing the fragment 1_11 flanked by the outer methylation-protectable BsaI sites (purple circles) and including multiple internal always-methylatable BsaI sites (green filled circles). The fragment 1_11 is in a VR plasmid because it forms the right end in the first round of assembly. (C) An example of one of the four assembled plasmids from the first round which are also used as donor plasmids in the second round. POC1456 contains the assembled fragment 2_1 without any unwanted scars between the four fragments assembled in the first round. The assembled fragment 2_1 is flanked by outer BsaI sites which are methylation-protectable, allowing it to be cut from the plasmid in the second round of the assembly. The fragment 2_1 was assembled in a VL vector because it forms the left end of the next round of the assembly. (D) The assembled plasmid POC1550 from the second round of the assembly containing the assembled fragment 3_1 (10.8 kb) without scars. The assembled fragment 3_1 was assembled in a VM vector because it is the final assembled fragment and cutting by BsaI will trim the vector adaptor sequences off both ends generating the desired assembled fragment fully free of scars. MP-dCas9–sgRNA: methylation-protection dCas9–sgRNA molecule. I1–I11 represent fragments 1 to 11.