Golden EGG, a simplified Golden Gate cloning system to assemble multiple fragments

Abstract

The Golden Gate method is an efficient tool for seamless assembly of multiple DNA fragments, which uses Type IIS restriction endonucleases, cleaving the DNA outside of their recognition site to release DNA parts from PCR fragments or entry clones, thus allowing the design of overhangs for ligation at will. However, the construction of the entry clones requires the use of other restriction enzyme(s) or cloning techniques and different entry vectors for the individual overhangs. Here, we present a simplified Golden Gate cloning approach termed Golden EGG. It features (1) a single entry vector with a specific cloning site to host the DNA parts; (2) a unique primer design to create the restriction enzyme recognition site to release the fragments with the overhangs at will; (3) the use of a single Type IIS enzyme for the construction of both the entry and destination clones; (4) a specific temperature profile during the digestion-ligation reaction. Our user-friendly, streamlined method retains the key attributes of the Golden Gate technique, while offering the potential to generate compatible parts with any existing Golden Gate toolkit and to be accessible to a wide user base without the need for extensive acquisition of new vectors or expensive enzymes.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-77327-4.

Introduction

Large-scale systems and molecular biology investigations have necessitated the development of high-throughput methods to clone, store and assemble large number of DNA parts. The progression of cloning methodologies has remained continuous in recent years, marked by the continual publication of optimized systems and novel ideas. While traditional restriction enzyme digestion–ligation-based cloning is still almost inevitable, at least once in a while, numerous methods and complete toolkits were developed that can make certain types of cloning tasks easier, faster, and more efficient. Besides the well-known original methods such as Gateway cloning¹, Gibson assembly², Golden Gate cloning³, InFusion recombination and TOPO-cloning⁴, dozens of variations or completely different alternatives have been developed over the years, i.e., GoldenBraid⁵, SLiCE⁶, TEDA⁷, Cas9-based auto-cloning⁸. As most of these technologies require specific enzymes, vectors, and primer design⁹, the low level of compatibility between these techniques often hinders the switching to a new method, even if it would be desirable. Since most of the advanced cloning techniques are fast and have high efficiency, they compete in convenience, flexibility, reusability, and ultimately the cost.

Golden Gate (GG) cloning possesses almost all the desired features: it is fast, efficient, flexible, with high reusability of parts, and very cost-effective, making it increasingly popular. However, its complexity hinders its spread beyond synthetic biology communities¹⁰. In the case of the GG technique, DNA parts such as promoters, CDS, etc., can be assembled in a standardized way in a single-pot digestion-ligation reaction in the desired destination vectors. In this sense, it is similar to the Gateway system, but the GG method is superior owing to its flexibility, since the number of fragments that can be assembled into a single vector is almost unlimited¹¹. It also does not result in the insertion of foreign sequences, such as the att(achment) sequences required for and formed in the Gateway reactions. The possibility of seamless cloning is an appealing feature of GG cloning. Its advantage over other seamless cloning methods, which are based on long overlapping ends such as Gibson assembly, In-Fusion recombination cloning, is that the individual parts can be stored in the entry clones and can be checked by sequencing. This allows for the sharing and reusing of the DNA parts in unlimited assembly reactions. The 4-base long sticky end generated by the most commonly used Type IIS restriction endonucleases, such as BsaI, allows for the design of 240 different, non-palindromic DNA end motifs on the ends of the fragments³, however, this flexibility can make the designing process more complicated¹⁰. The disadvantage of this method is that to produce different overhangs for the assembly, the fragments have to be cloned into different entry vectors by using other Type IIS or conventional Type IIP restriction endonucleases or another previously mentioned method.

Over the years, many attempts have been made to simplify and standardize the GG cloning method^5,11–13, but the creation of entry clones and their cloning process remain one of the least standardized steps. Each of the mentioned methods is complicated by the use of multiple restriction enzymes and/or entry vectors making the design process convoluted and costlier.

As with all the restriction enzyme based cloning methods, the cloning strategy and design can be further complicated by the presence of internal restriction endonuclease recognition sites of the enzymes used in any step of the GG assembly. It is generally recommended to eliminate the internal recognition sites of these enzymes before using them in the GG reaction¹⁰ to maintain the high efficiency of the reaction. This is because in the single-pot reaction, the final product is stable only if it does not contain the recognition site of the used enzyme, as it would re-open the correctly assembled plasmids. Cloning and subcloning DNA parts with internal recognition sequences or with recognition sites remaining on the vector backbone are also possible in a single-pot reaction by stopping the digestion-ligation reaction with the heat inactivation of the enzymes followed by an extra ligation step³. This step not only increases the workload and the cost, but also re-circularizes the remaining empty vectors in the reaction. One approach to solve the high empty vector background in such cases is to use a negative selection marker cloned in the empty vector, such as the gene coding for the CcdB toxin, which is lost upon cloning the fragment(s) of interest¹³. However, the most common toolboxes use simple blue-white selection to distinguish empty vectors from clones with inserts^11,12.

Here we present a simplified method for the construction of GG entry clones and their assembly, which is based on a special primer design, a universal entry vector for all DNA parts, and the use of a single Type IIS endonuclease producing four nucleotide overhangs in conjunction with T4 DNA ligase in the digestion-ligation reaction performed at a special temperature profile. Since the same enzyme is used both for the cloning of the DNA fragments into entry vectors and for their release before assembly, and because strict domestication of DNA parts is not required, this method simplifies the experimental design and decreases the overall workload and costs. Moreover, this cloning strategy is either compatible with all existing GG toolkits, or adaptable through switching the restriction enzyme to the one that is used to release the DNA parts in the assembly reaction.

Results

Novel vector and primer designs allow the use of a single type IIS enzyme for the construction of entry clones and the assembly of multiple fragments in destination vectors

To simplify and make the Golden Gate procedure cost-effective, we developed an approach using a universal entry vector, a novel primer design (detailed in Supplementary file 2: GoldenEGGcloningSchemeOligoTool.xlsx) and a single Type IIS restriction endonuclease producing four-nucleotide overhangs (Eco31I/BsaI as an example) in conjunction with T4 ligase (Fig. 1). The new universal entry vectors described here contain a cassette with the ccdB and cat genes for negative selection which is flanked by outward directed Eco31I (BsaI) recognition sequences with identical overhangs, collectively termed as the EGG (Entry for Golden Gate cloning) site. The unique feature of the new pEGG vectors is that after digestion with Eco31I, inward directed Eco31I recognition sites incomplete at their 3’ end remain as overhangs on the vector backbone on both sides, with opposite orientation compared to the original Eco31I recognition sites. On the other hand, both ends of the PCR fragment to be cloned contain an inward directed Eco31I recognition sequence and their cleavage leaves inward directed Eco31I recognition sites incomplete at their 5’ end as overhangs. Ligation of the vector and the PCR fragment results in the formation of inward directed Eco31I sites that flank the sequence of interest. As each PCR primer contains a NGGTCTCHGTCTCNn1n2n3n4 extension 5’ to the target-specific sequence, where n1-n4 is one of the possible 240 (4⁴-4²) non-palindromic four nucleotide-long overhang sequences, the newly ligated insert can be released with any kind of desired sticky ends, allowing the seamless cloning of one or more fragments into any compatible destination vector. The insert fragments to be released from the entry clones and to be assembled in the destination vector are designed to have overhangs complementary to the neighbouring fragments (Fig. 1), i.e. the left end and right end 5’ overhangs of a fragment produced by the enzyme upon digestion are complementary to the overhang at the right end of the fragment on its left and to the overhang at the left end of its right, respectively. The compatible destination vectors should contain two outward directed Eco31I recognition sites producing left and right overhangs (OH_L and OH_R) that are complementary to the left overhang of the first fragment and the right overhang of the last fragment, respectively. To allow antibiotic selection at the entry level that is different from that used at the destination assembly, high copy number ColE1-based phagemid and plasmid entry vectors with different antibiotic selections were created.

Fig. 1.

The simplified Golden Gate cloning system. The pEGG entry vectors contain a cassette with the ccdB and cat genes for negative selection that is flanked by outward directed Eco31I (BsaI) recognition sequences, collectively named as EGG (Entry for Golden Gate cloning) site. After digestion with Eco31I, an inward directed Eco31I recognition site (green) ‒ incomplete at its 3’ end ‒ remains on the vector backbone on both sides as an overhang, i.e. with opposite orientation compared to the original Eco31I recognition sites (orange). Both ends of a PCR fragment contain an inward directed Eco31I recognition sequence (orange) and their cleavage leaves inward directed Eco31I recognition sites (green) incomplete at their 5’ end as overhangs. Ligation of the vector and the PCR fragment results in the formation of inward directed Eco31I sites (green) that flank the sequence of interest in the pEGG-clones. As each PCR primer contains a NGGTCTCHGTCTCNn1n2n3n4 extension 5’ to the target-specific sequence, where n1-4 is one of the possible 240 non-palindromic overhang (OH) sequences, the newly ligated insert can be released with any kind of desired sticky ends, allowing the seamless cloning of one or more fragments into any Eco31I compatible destination vector. Note that no temperature shift was applied during the assembly of the destination clone because no new Eco31I sites are formed (and cleaved) in the reaction.

Highly efficient single-tube cloning by cold treatment of the digestion-ligation reaction

In classical single-tube digestion-ligation GG reactions for cloning a fragment into an entry or destination vector, only the desired ligation products, the entry or destination clones lacking the restriction enzyme recognition site, are stable, while all other products are re-digested with the enzyme. In the approach we devised, however, two recognition sites for the Type IIS enzyme are formed in the resulting entry clones, which can be cleaved by the enzyme making the ligation product unstable. For maximizing the ratio of circularized entry clones before transformation, it would be possible to heat inactivate both enzymes and restart the ligation by adding ligase and ATP to the reaction mix. However, to make the whole process simpler and cheaper, we tested whether the reaction kinetics could be effectively shifted towards ligation only by lowering the reaction temperature. The reasoning behind this approach was that several restriction enzymes were shown to have much lower activity while T4 ligase was demonstrated to retain as high as > 50% activity at 0 °C¹⁴. Moreover, some companies also recommend 4 °C as reaction temperature for ligation to obtain the highest number of colonies. First, we investigated the effect of cold treatment on the digestion-ligation reaction performed on a linearized and dephosphorylated pGEM-T Easy DNA, which was either digested with Eco31I or left undigested. The reaction mixtures containing the recommended buffer, ATP and the enzymes as well as either 100 ng Eco31I-digested DNA (two fragments) or 100 ng undigested linear DNA were incubated at 37 °C for 5 or 15 min. They were then either stopped by heat treatment at 80 °C or were cold treated at 4 °C for 15 min. The fragments were separated in agarose gel revealing that the cold treatment shifted the reactions towards ligation independent of the input type (Fig. 2a). To further confirm the effect of the cold treatment, 25–50 ng of the different pEGG entry vectors and a PCR fragment with the gene coding for mCherry were mixed in a 1 to 3 ratio, incubated at 37 °C for 1 h and the reaction was either stopped at 80 °C or further incubated at 4 °C for 15 min, then transformed into homemade chemically competent cells of E. coli strain MDS42recAblue. As shown in Fig. 2b, the extra incubation at low temperature increased the cloning efficiency by a factor of four to ten times and provided a high number of clones.

Fig. 2.

The effect of 4 °C cold treatment on the Golden Gate reaction. (a) An EcoRI-digested and dephosphorylated linear pGEM-T Easy DNA was added either un-digested (2997 bp) or pre-digested with Eco31I (1596 and 1401 bp fragments) as input (shown in the control lanes) to the reactions. The Golden Gate digestion-ligation reactions contained 100 ng DNA, 10 units of Eco31I and 5 Weiss units of T4 ligase that were incubated first at 37 °C for 5 or 15 min then placed to 80 °C or 4 °C for 15 min. Notice that after stopping the reaction with heat inactivation of the enzymes, approximately equal amounts of digested and ligated products are present indicating the balance of the opposing reactions, while after cold treatments, the majority of products was ligated DNA independent of the starting material showing that mainly the ligase was active at low temperature. GeneRuler 100 bp Plus DNA Ladder (ThermoFisher Scientific) was used as DNA size marker (first lane). (b) Cloning of an mCherry fragment into different pEGG vectors in one-tube digestion-ligation reactions was carried out by keeping the reaction at 37 °C for 60 min then placing the tubes to 80 °C to inactivate the two enzymes or to 4 °C for 15 min to favor the ligation reaction. After transformation into E. coli strain MDS42recAblue, four to ten times more colonies were obtained when the cold treated reactions were used. Significance levels were determined by Welch’s t test (* P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001) from four to eight technical replicates.

For the optimization of the cloning reaction and to determine its efficiency, the vector and the mCherry PCR fragment were mixed in different ratios, treated the same way as earlier (37 °C for 60 min, 4 °C for 15 min). The reactions were then divided equally and the two halves were transformed into homemade competent cells of E. coli strains MDS42recAblue and DH5α (Fig. 3a, b). As a control, the PCR fragment was blunt-end ligated into the pJET1.2 vector (Thermo Fisher Scientific) and transformed into the competent cells of both strains. Similar to the conventional restriction enzyme-based cloning, vector insert ratios 1:2 to 1:4 were the most effective, though the differences in the 1:1 to 1:16 ratio range were not significant. Based on their fluorescence, ~ 95% of the colonies carried the mCherry insert (Fig. 3c), which could be released by Eco31I digestion and confirmed by Sanger sequencing (category I). Approximately 50–70% of the non-fluorescent clones released fragments with correct sizes after Eco31I digestion and carried point mutations or small indels based on Sanger sequencing indicating PCR error during amplification or spontaneous mutations during plasmid replication (category II). The remaining clones either carried truncated or irrelevant sequences with correct ends (category III), or they were mutated in both coding sequences and the Eco31I recognition sequence at one or both ends (category IV), as determined by Eco31I digestion and Sanger sequencing.

Fig. 3.

Highly efficient cloning through a wide range of vector to insert ratio. In each reaction, 50 ng pEGG1a vector was used and the amount of an mCherry PCR fragment was adjusted to reach the desired vector to insert molar ratio. 50–50% of the reactions was transformed into homemade competent cells prepared from E. coli strains (a) DH5α and (b) MDS42 ΔrecA Blue. Each experiment was repeated at least three times. The colonies were counted under a fluorescent stereomicroscope. As a control, the mCherry fragment was ligated with the pJET1.2 vector in a three to one molar ratio. (c) Plasmid DNA was isolated from ten colonies of both the fluorescent (red background) and non-fluorescent bacteria (blue background) and the inserts were checked by Eco31I digestion, then the digestion results were confirmed by Sanger sequencing in a few cases. Sanger sequencing column show the number of reads/number of sequenced clones that are in agreement with digestion results (NT: not tested). The tested fluorescent colonies carried the correct fragment and could be released by digestion (category I). ~60% of the plasmids (5 and 7 out of 10) isolated from non-fluorescent colonies carried a mutated mCherry insert of correct size that was flanked by Eco31I recognition sites (category II). The rest of the non-fluorescent colonies harboured plasmids with either Eco31I-releasable fragments with correct ends but of incorrect size (3 and 2 out of 10) and sequence (category III) or with unrelated inserts that could not be released (2 and 1 out of 10) by Eco31 digestion (category IV).

To address whether larger DNA fragments can be efficiently cloned, other PCR products of different sizes were used in the single-tube reaction instead of the mCherry fragment (Fig. 4). The cloning efficiency of EGFP sequences and of an ~ 1800 bp GUS reporter gene fragment was similar to each other, while cloning of the ~ 2500 bp Medicago truncatula Enod11 promoter, responding to Sinorhizobium inoculation and to the diffusible sinorhizobial signal molecules termed Nod Factors¹⁵, resulted in ~ 10-times less colonies. Cloning of the entire Enod11 gene on a 6.5 kbp fragment, which also contained and internal Eco31I recognition site, resulted in 50–100 times less but still sufficient number of colonies (Fig. 4b).

Fig. 4.

Cloning efficiency of PCR fragments with varying sizes. (a) PCR fragments (to be assembled in a plant transformation vector as shown in Fig. S1) carrying the ~ 1800 bp GUS (uidA) reporter gene (blue); the ~ 2500 bp MtEnod11 promoter (mustard) and the ~ 6500 bp MtEnod11 gene with an internal Eco31I recognition site (yellow) as well as the ~ 750 bp EGFP gene (green) with or without nuclear localization signal (NLS) as control were added in three to one molar ratio to 50 ng of the pEGG1a entry vector in one-tube digestion-ligation reactions. The overhangs used for the assembly or generated after the Eco31I digestion of the MtEnod11 gene are indicated. Note that the GUS encoding fragment was amplified with overhangs to be ligated either to the promoter (AATG) or to the EGFP sequence (GAGG) as shown in Fig. S1. (b) The number of colonies formed (shown in logarithmic scale) after transforming the ligation reactions with fragments of different sizes into E. coli MDS42 ΔrecA Blue competent cells.

Construction and use of destination vectors for plant transformation and bacterial genome manipulation

The cloned reporter genes coding for the (nuclear-localized) EGFP and β-glucuronidase (GUS) proteins along with the MtEnod11 promoter were combined into the pKNGG_mCR destination vector in a demonstrative experiment. This binary vector, tailored for plant transformation, was created by substituting the yellow fluorescent transformation marker of the pKNGG_YR plasmid to an mCherry reporter gene under the control of the AtUBI10 promoter. The designed MtEnod11 transcriptional reporter constructs were assembled in single tube reactions, in which the destination vector and two or three entry clones carrying the Mtenod11 promoter as well as the reporter genes were digested and ligated as described for the construction of the entry clones (Fig. 4a, Fig. S1.). Each construct was transformed into M. truncatula genotype A17 using Agrobacterium rhizogenes-mediated hairy root transformation, then, the transgenic roots were either treated with Nod Factors or inoculated with Sinorhizobium meliloti, the symbiotic partner of the plant. As expected and shown in Fig. S1, the transgenic roots express the reporter genes in response to both Nod Factor treatment and rhizobium inoculation.

In rhizobia, which have a 6–8 Mbp sized genome residing in a chromosome and 2–6 (mega-)plasmids, defined genomic deletions are usually created by cloning two flanking fragments of equal sizes for homologous recombinations into a mobilizable vector either with or without an antibiotic cassette between them to replace the gene(s) to be deleted. This vector can replicate only in E. coli and carries the sacB gene coding for the levansucrase enzyme producing a toxic compound from sucrose for counter-selection. The clone is then transferred into the Sinorhizobium cells and integrated via homologous recombination into their genome through one of the flanking sequences. The loss of the plasmid via a second recombination, which results in the allelic exchange if the recombination happens through the other flanking sequence (theoretically in 50% of cases), is selected for by growing the bacteria on sucrose-containing media. Bearing in mind to create serial genome deletions in rhizobia to investigate the role the strain-specific genomic regions during legume-rhizobium interaction, a pair of destination vectors of pK18mobSacB¹⁶ origin with determined overhang pairs were created in such a way that a single fragment cloned into a pEGG entry vector can be used for the deletion of two neighbouring genome regions. In this scheme, odd and even numbered fragments are flanked in the entry clones by Eco31I sites producing overhangs OH_A-OH_B and OH_C-OH_D, respectively (Fig. 5). When the deletion is to be formed between an odd and an even numbered fragment (left and right borders of the deletion), they are cloned into the destination vector with Eco31I sites producing OH_A and OH_D overhangs and flank an antibiotic resistance cassette released from its entry clone with OH_B and OH_C overhangs. However, when the deletion is to be formed between an even and an odd numbered fragment as left and right borders, respectively, they are cloned into the destination vector with Eco31I sites producing OH_C and OH_B overhangs and flank an antibiotic resistance cassette released from its entry clone with OH_D and OH_A overhangs (Fig. 5).

Fig. 5.

Deletion of large neighbouring regions from a bacterial genome. (a) Sequences (FR1-4) flanking the genome fragments to be deleted (indicated with Roman numbers in the lower part of the panel) are cloned into pEGG1a (orange backbone at the top of the figure) between A-B (odd) and C-D (even) overhang type Eco31I sites. Antibiotic resistance cassettes Ab^R1 and Ab^R2 to replace the deleted fragments are cloned into pEGG1a between either B and C or D and A overhang types. Destination vectors (black backbone) pK18MSB2S1DvAD with the A and D overhang types and pK18MSB2S1DvCB with the C and B overhang types accept the even-Ab^R1-odd and odd-Ab^R2-even constructs to form clones pDestAD and pDestCB, respectively (middle part).

(b) Representation of the pSymB section in S. meliloti strain Rm41 that carries the rkp-3 region responsible for the synthesis of the strain-specific K-antigen (green), a strain-specific (Lps1 in purple) and a species-specific (Lps2 in orange) gene cluster putatively involved in the production of the O-antigen and the core oligosaccharide of the LPS, respectively, before any deletion (top), or after replacing the Lps1 (middle) and Lps2 (bottom) regions with a gentamycin resistance gene (grey). A star indicates the position of the mutated rns1 gene. The fragments used for the homologous recombination reactions (red boxes) as well as the PCR primers used for diagnostic PCRs (arrows) are indicated. (c) Diagnostic PCR reactions detect the presence of the wild-type sequences (PCR2, PCR4) and the deletions (PCR1, PCR3) in the created strains.

To put the method into practice, an approximately 55 kbp (Lps1) and a neighbouring ~ 20 kbp (Lps2) region carrying strain-specific and species-specific genes, respectively, involved most probably in the synthesis of the O-antigen and the core of the lipopolysaccharide (LPS), respectively, were aimed at. Using this approach and three fragments as homologous recombination targets, the regions were deleted from the symbiotic megaplasmid pSymB of S. meliloti strain 41 (Fig. S2) mutated in the rns1 gene. The rns1 gene in the strain-specific region codes for a protein predicted to be involved in polysaccharide modifications and causes the inhibition of symbiotic nodule development and bacterial invasion of the plant cells in the M. truncatula F83005 ecotype¹⁷. With the help of the two deletions, we investigated whether the LPS is needed for the rns1-mediated incompatibility. Both deletion derivatives either carrying mutation in or devoid of the rns1 gene could establish effective symbiosis, while their derivatives harbouring a plasmid with the wild-type rns1 gene were unable to infect the F83005 ecotype, indicating that not the modification of the LPS core or O-antigen structure by the Rns1 protein, rather the presence of Rns1 results in the incompatible interaction. Similarly, the rkp-3 region¹⁸ on the other side of the Lps1 region, which is required for the synthesis of the strain-specific K-antigen and for the sensitivity towards phage 16 − 3, was also deleted with the method. As expected, the deletion-derivative strain became resistant for phage 16 − 3 and sensitive for the S. meliloti strain Sm1021 phage M10 (Fig. S3).

Discussion

Golden Gate cloning represents a highly versatile molecular cloning technique renowned for its applicability across a wide range of cloning tasks, spanning from basic to intricate procedures. The effectiveness, speed, and cost-efficiency of the GG method contribute to its increasing popularity. Recent technological advancements in GG cloning include innovative approaches, such as loop cloning mechanisms, enabling the simultaneous cloning of fragments in theoretically infinite steps. A big advantage of GG cloning is its enhanced capacity of cloning multiple fragments into a destination vector in a single step. A former study¹⁹ showcased its efficiency with the assembly of up to 52 fragments with high accuracy in a single step. Another advantage is that in contrast to methods like Gibson assembly and InFusion cloning, which are based on the overlaps between the different parts, GG cloning offers the unique capability of storing DNA fragments in entry vectors. This feature improves cost-effectiveness and long-term efficiency by eliminating the need for repetitive cloning and sequencing of frequently used DNA parts. However, it has disadvantages, e.g. the requirement for an additional step to clone these fragments into the entry vectors, often involving a distinct cloning approach using TA or TOPO cloning methods or another Type IIS or “classical” Type IIP restriction enzymes; and the need of an entry vector set.

Our approach simplifies and standardizes the entire cloning process through the implementation of specific design elements in the primers used to amplify the fragments of interest and in our novel and single entry vector where the crucial element is termed the EGG (Entry for Golden Gate) site. Both ends of the EGG site and of the PCR fragments contain a complete and an incomplete recognition site of a Type IIS enzyme producing four nucleotide overhangs. The arrangements of the sequences ensure that after digestion-ligation, the fragment in the created entry clone is flanked by inward-directed recognition sites newly formed from the incomplete recognition sequences and can be released with overhangs defined during the design of the oligonucleotide primer sequences. We propose to call this cloning mechanism the Golden EGG system. The devised design enables the use of only a single Type IIS endonuclease, such as Eco31I (BsaI), BbsI (BpiI) or BsmBI (Esp3I) along with T4 ligase, both to create the entry clones and then, to release and assemble the fragments with the desired overhangs in an appropriate, compatible destination vector in a seamless way. Thus, this innovation streamlines the cloning process, offering a more convenient and efficient alternative to traditional methods. In addition, our method can eliminate the numerous drawbacks of the DNA-FACE™ technology^20,21, which was developed for constructing artificial concatemeric DNA, RNA, and proteins with the help of a TypeIIS enzyme such as SapI. These drawbacks include a multistep cloning procedure, random number and order of fragments in the assembled clones ‒ thus, a need to check a lot of colonies ‒ as well as the addition of a single amino acid scar between all building blocks.

The novel recognition sites formed during the ligation of the PCR fragments and the entry vectors seemingly make the entry clone products unstable due to the opposing activity of the restriction enzyme opening the clones circularized by the T4 ligase in the GG reaction mixture, resulting in the formation of numerous linear DNA molecules. However, we took advantage of the known feature of a lot of restriction enzymes, namely that they have much lower activity at temperatures lower than 37 °C, while T4 ligase retains close to 50% of its activity at 4 °C¹⁴, which even prompted some suppliers to recommend this temperature for ligation in their protocol. Indeed, the number of circular products ‒ mirrored in the number of colonies after transformation ‒ was at least four-fold higher if a short low temperature incubation was applied after the digestion-ligation reaction at 37 °C as compared to when the reaction was stopped by heat inactivation of the enzymes.

Using this method, high cloning efficiency could be achieved: (i) Even at 1:1 or 1:16 vector to insert ratios, the number of formed colonies were only approximately three times lower than the number of colonies obtained when two to four-times more fragment than plasmid was added to the reaction. (ii) 94–96% of the colonies harboured the correct insert and even the majority of the incorrect inserts was the result of PCR or replication errors introducing mutations demolishing gene function. (iii) Although insert size and the presence of internal recognition site(s) negatively affect cloning efficiency, cloning of the ~ 2500 bp MtEnod11 promoter fragment and the ~ 6.5 kbp MtEnod11 gene with an internal Eco31I site resulted in 10 times and 50–100 times less colonies, respectively, than the ligation of the inserts with ~ 750 (EGFP) and ~ 1800 bp (GUS) sizes, but still produced sufficient number of correct clones. Single step assembly of the MtEnod11 promoter with one or both reporter protein coding genes from the entry clones into a binary plant transformation vector, equipped with Eco31I sites and appropriate overhang sequences, highlights the ease of creating constructs from any number of fragments stored with custom-tailored overhangs in our universal entry vector. The assembled reporter constructs were introduced into Medicago truncatula plants by Agrobacterium rhizogenes-mediated hairy root transformation where they showed the expected expression pattern.

Strains of Sinorhizobium meliloti and S. medicae are the symbiotic partners of legume plants belonging to the Medicago, Melilotus and Trigonella genera. The strains of the two Sinorhizobium species have genome sizes ranging from 6.2 to 7.8 Mbp with ~ 4800 to ~ 8900 predicted genes and it was reported that besides the core genome, they have an accessory genome of considerable size with up to 840 unique genes not present in any other strains²². These accessory genomes enable the strains to adapt to different environmental niches and affect either positively or negatively their interactions with the numerous natural variants of the host plants. To demonstrate the simplicity of our cloning method and facilitate the genetic study of the rhizobial genomes, we developed an Eco31I/BsaI compatible vector pair, which enables us to sequentially delete genomic regions in rhizobia. The classical method for deleting sequences involves cloning two approximately equal-sized flanking fragments of the desired deletion into a mobilizable vector using Type IIP enzymes either without or with an antibiotic cassette between them to replace the gene(s). Alternatively, more expensive recombination-based methods are employed. The construct will integrate via homologous recombination through one of the flanking sequences into the genome, then, after forcing the loss of the vector by a second recombination through either of the flanking sequences, theoretically half of the bacteria will carry the deletion. To clone the two fragments into the same vector, the careful design of the PCR primers with recognition sites are needed but usually, a fragment to flank a deletion on its other side cannot be re-used because the number of usable recognition sequences and multiple cloning sites with the appropriate order of recognition sites is very limited. In our approach, two vectors are used for the cloning of the two flanking fragments and a resistance cassette between them in the order of even fragment - casette -odd fragment and odd fragment - cassette- even fragment, respectively. This, along with a single enzyme, simplifies the design as well as decreases the workload and costs to a great extent because a fragment can be used for the generation of deletions on both of its sides. To validate the usability of this system, we deleted three neighbouring regions of 10–55 kbp in size from the pSymB replicon of S. meliloti strain Rm41, which carry genes predicted to be involved in the production of the strain-specific capsular polysaccharide, the strain-specific O-antigen, and the species-specific core of the lipopolysaccharide (LPS), respectively.

Conclusions

Our novel method provides a streamlined alternative not only to the conventional restriction-ligation-based approach but also an efficient and cost-effective GG cloning solution. The user-friendly nature of our system, encompassing primer design, reaction mixture composition, and reaction conditions, makes it easy to adopt by researchers without the need for extensive acquisition of new vectors or expensive enzymes. Thus, this method aims to make Golden Gate cloning accessible to a wide user base while it retains the key attributes: effectiveness, the preservation of reusable parts in entry vectors, and the capacity to assemble a high number of inserts into a destination vector in a single step. This method not only holds promise for seamless cloning but also offers the potential to generate compatible parts with any existing GG toolkit.

Materials and methods

Molecular biology reagents, bacteria and phage tests

Restriction endonucleases and T4 DNA ligase were purchased from Thermo Fisher Scientific. DNA fragments were amplified using Q5 High-Fidelity DNA Polymerase (New England Biolabs) according to the manufacturer’s suggestions. DNA isolation from agarose gel and plasmid preps were carried out using the QIAquick Gel Extraction Kit and the QIAprep Spin Miniprep Kit (Qiagene), respectively. Escherichia coli was grown in LB medium, while for the culturing of Sinorhizobium meliloti and Agrobacterium rhizogenes, LB medium was supplemented with 2.5 mM MgSO₄ and 2.5 mM CaCl₂. Homemade chemically competent cells of E. coli strains (Supplementary Table 2) were prepared by using the Mix & Go E. coli transformation buffer set (Zymo Research), while A. rhizogenes ARqua1 Freeze/Thaw-Competent cells were prepared using the protocol of Wise et al.²³. Phage tests were carried out as described by Putnoky et al.²⁴.

Restriction-ligation kinetic test

EcoRI digested and phosphatase treated pGEM-T Easy vector (Promega) was isolated from agarose gel and used as linearized DNA in the digestion-ligation kinetic assay. Each reaction mix contained 1x Green restriction endonuclease buffer, 0.5 mM ATP, 0.5 µl T4 ligase (5 Weiss units/µL), 0.5 µl Eco31I (10 U/µL) and either 100 ng undigested linear DNA (2997 bp) or fragments (1596 and 1401 bp) obtained with Eco31I digestion were added to the reactions. After 5/15 minutes incubation at 37 °C and 15 min at 4/80°C, samples were collected at -20 °C. Before gel electrophoresis, samples were thawed and mixed with 2,5 µl 6x DNA Gel Loading Dye (Thermo Fisher Scientific) supplemented with 0.8 µl of 10% SDS per sample.

Vector constructions

Construction of pEGG1a

In a blunted and circularized pGEM-T Easy vector, the Eco31I/BsaI site in the ampicillin resistance gene was removed by introducing a single nucleotide change resulting in a silent mutation. Part of the ampicillin resistance gene was amplified using AmpRmut-R - pGTe1-F primers. The mutated fragment was reinserted after Csp6I-Eco31I double digestion of the vector and the fragment. This vector was further modified by the insertion of the multiple cloning site amplified from pBluescript SK with BScBsai-F and BScBsai-R primers, and placed between PaeI and MunI restriction sites. Finally, the 1897 bp long SalI fragment of pDONR221 was inserted into the SalI site of the hybrid vector with same gene orientations as the ampicillin resistance gene and transformed into E. coli DB3.1 competent cells.

Construction of pEGG1s

A 2284 bp AatII-AseI digested fragment of pEGG1a was cloned into vector pL0¹¹ between the AatII and AseI sites and transformed into E. coli DB3.1 competent cells.

Construction of pEGG2a

A HindIII and MunI digested ccdB - chloramphenicol acetyltransferase fragment amplified from pEGG1a using CcdCmBsaF and CcdCmBsaR primers was ligated into a EcoRI and HindIII digested pUC57 vector backbone devoid of a Eco31I/BsaI recognition site and transformed into E. coli DB3.1 competent cells.

Construction of pKNGG_mCR

Vector pKNGG_YR²⁵ was digested with EcoRI and BcuI then the ends were filled in by using Phusion polymerase, finally, the purified backbone was circularized. The resulting intermediate vector (pKNGG_0) was opened with HindIII and then three digested PCR fragments were assembled in it in a single step reaction: (I) prAtUBQ10 (amplified from pKNGG_YR using primers UBIp_NotI_HindIII and GG_UBIp_R1, then digested with HindIII and Eco31I); (II) mCherry (amplified from pDONRP2R-P3-mCherry²⁶ using primers GG_mCherry_F1 and GG_mCherry_R1, then, digested with Eco31I) (III) tNOS (amplified from pKNGG_YR using primers GG_NOSt_F1 and NOS_AatII_Hind_R2, then digested with Eco31I and HindIII). The orientation of the mCherry coding gene was opposite of the streptomycin resistance gene. The generated plasmid was maintained in E. coli DH5α.

Construction of pK18MSB2S1DvAD and pK18MSB2S1DvCB

PagI and AjiI double digested oriT region amplified from pJQ200SK using oriTpag and oriTsac primers was ligated into a PagI and AjiI digested pK18mobSacB backbone to replace its original oriT region and to obtain pK18MSB2. An I-Sce site was introduced by ligating the annealed I-SceF and I-SceR oligonucleotides into the NheI site of pK18MSB2 in such a way that the NheI recognition sequence was re-created between the I-SceI and PvuI sites, resulting pK18MSB2S1. Subsequently DvAB and DvCB fragments were amplified from pEGG1a using DV1Bsa1AF - DV1bsa1DR and DV2Bsa1BR - DV2Bsa1CF primers, respectively, then were inserted between the HindIII and SacI sites of pK18MSB2S1 resulting in pK18MSB2S1DvAD and pK18MSB2S1DvCB, respectively.

Simplified single-tube Golden Gate reactions

PCR fragments were cloned into pEGG vectors in 10 µl final volume using the reaction mix containing 1x Green buffer, 0.5 mM ATP, 0.5 µl T4 ligase (5 Weiss/µL), 0.5 µl Eco31I (10 U/µL), maximum 50 ng pEGG vector (~ 15 fmol), and the insert fragment in the amount which was adjusted according to the planned insert to vector molar ratio. The reaction mix was incubated at 37 °C for 1 h, then shifted to 4 °C for 15 min before transformation into E. coli MDS42RB or E. coli DH5α competent cells. The assembly into destination vectors was carried out similarly but generally in higher volume adding ~ 40 fmol destination vector and ~ 120–120 fmol from each entry clone. These reactions were kept at 37 °C for a prolonged time period of 5–16 h before transformation. No temperature shift was applied because no new Eco31I sites are formed after ligation.

Generation of deletions

The selection marker used during the creation of the deletions was combined from an mCherry coding sequence and the gentamycin-3-acetyltransferase gene amplified from pMP7605 using mCh_B/mCh_D –CherryEcoR and GMEcoR –GmR_C/GmR_A primers, respectively, then inserted into Eco31I digested pEGG1a after cleaving the fragments with EcoRI and Eco31I enzymes. After cloning the selection marker and the fragments to flank the planned deletions into the pK18M2S1DvAD or pK18M2S1DvCB vectors by the simplified GG method, the clones were introduced into the S. meliloti strain Rm41 carrying a spectinomycin resistance cassette in the rns1 gene¹⁷ via tri-parental mating, then the deletions were created by the method of Schafer et al.¹⁶. A plasmid carrying the wild-type rns1 gene¹⁷ was introduced into the deletion derivatives by tri-parental mating.

Plant transformation, inoculation and transgene visualisation

Medicago truncatula cv. Jemalong A17 was transformed as described by Boisson-Dernier et al.²⁷. Transgenic A17 and wild-type F83005 plants were inoculated with S. meliloti strains 1021 and Rm41rns1::SpcR derivatives, respectively, as described earlier²⁸. β-glucuronidase activity was visualized by immersing the excised root segments in a staining solution²⁸. Microscopic analysis to detect the activity of the reporter genes was carried out using a Leica MZ10-F fluorescence stereo microscope (Leica Microsystems GmbH).

Primers

The used oligonucleotides are listed in Supplementary Table 1.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

JBB, PK and AK conceived and designed the study. JBB, KK, ZS, BG and TW performed the experimental work. JBB and KK carried out the statistical analyses. JBB, PK and AK wrote and edited the manuscript.

Funding

Hungarian National Research, Development and Innovation Office grants K134841 and K146663 (to AK) and K132646 (to PK).

Data availability

All data supporting the findings of this study are available within the paper and the corresponding supplementary materials published online.

Declarations

Competing interests

The authors declare no competing interests.

Statement regarding research involving plants

The model plant Medicago truncatula cv. Jemalong A17 is available for and widely used by the M. truncatula community and we have a permission to collect seeds. All the methods were performed in accordance with the relevant institutional, national, and international guidelines and legislation.