BAC-browser: the tool for synthetic biology

Abstract

Background

Currently, synthetic genomics is a rapidly developing field. Its main tasks, such as the design of synthetic sequences and the assembly of DNA sequences from synthetic oligonucleotides, require specialized software. In this article, we present a program with a graphical interface that allows non-bioinformatics to perform the tasks needed in synthetic genomics.

Results

We developed BAC-browser v.2.1. It helps to design nucleotide sequences and features the following tools: generate nucleotide sequence from amino acid sequences using a codon frequency table for a specific organism, as well as visualization of restriction sites, GC composition, GC skew and secondary structure. To assemble DNA sequences, a fragmentation tool was created: regular breakdown into oligonucleotides of a certain length and breakdown into oligonucleotides with thermodynamic alignment.

We demonstrate the possibility of DNA fragments assemblies designed in different modes of BAC-browser.

Conclusions

The BAC-browser has a large number of tools for working in the field of systemic genomics and is freely available at the link with a direct link https://sysbiomed.ru/upload/BAC-browser-2.1.zip.

Introduction

Currently, synthetic genomics is an actively developing field of science. Synthesis of a nucleotide sequence with a length of several thousand base pairs (kbp) is already a routine method of molecular biology. DNA synthesis methods can also be used by scientists to create DNA libraries, for example, synthetic promoters and terminators [1] enhancers of promoters [2], intergenic regions in operons [3] or cell evolution to optimize the pathway of biosynthesis [4]. At the current stage of technology development, synthetic genomics makes it possible to synthesize complete genomes and chromosomes of viruses, bacteria, and lower eukaryotes [5, 6]. To date, several small bacterial genomes with a length of over 500 kbp [7,8,9,10] and Sc2.0 genome based on Saccharomyces cerevisiae with 6.5 synthetic chromosomes [11] have also been successfully synthesized from synthetic oligonucleotides.

A number of both commercial and freely distributed programs have been developed for gene synthesis, an overview of which and their functions can be found in [12]. Most of the programs for gene sequence design have the functions of optimizing the codon composition of the gene and searching for certain restriction sites, some also have the functions of optimizing GC composition, searching for secondary structures and repeats. The programs DNAWorks [13] Gene2Oligo [14], FastPCR [15] and POSoligo [16] which can generate thermodynamically aligned oligonucleotides from a given sequence, are currently the most accessible for assembling genes from synthetic nucleotides. However, all of these programs work with a given nucleotide sequence length comparable to the length of the average gene.

From the pioneer groups that created the first synthetic genomes, two have their own software. The yeast genome assembly group [17,18,19] has developed a set of separate utilities for genomic sequence design that have the functions of obtaining a nucleotide sequence from an amino acid sequence, optimizing the codon composition and removing restriction sites. The second group, that synthesized of Caulobacter ethensis-2.0 genome [9], uses a web utility for genomic sequence design, which automatically removes elements from a given sequence that prevent the assembly of a sequence of oligonucleotides (restriction sites, homopolymers, di- and trinucleotide repeats, hairpins and optimizes GC composition) [20]. To assemble extended genomic sequences, the first group uses a utility to split the sequence into regular segments with a length of 500–750 bp with overlaps flanked by service sequences, and then further splitting of segments into oligonucleotides with thermodynamic alignment [17,18,19]. The second group uses a utility to assemble extended sequences, which splits the original sequence into a set of hierarchical segments flanked by service sequences [21]. Also, an algorithm [17,18,19] was recently announced that divides a long DNA sequence into fragments, for each of which oligonucleotides are then generated for assembly.

We have developed a number of tools necessary for the design of synthetic sequences and the assembly of DNA sequences from synthetic oligonucleotides and included them in the software we developed earlier [22]. BAC-browser is a user-friendly genome browser designed for molecular biologists, microbiologists, and can be used by non-bioinformatics specialists. Historically, the software was known as the BAC (Bacterial Artificial Chromosome) browser, but currently, it is possible to operate with larger chromosomes, such as human ones.

The program has graphic user interface and is freely distributed. It allows not only to design oligonucleotides for gene synthesis, but also the design of the target sequence, the design of the assembly of extended DNA fragments up to the lengths of small genomes and to take into account additional parameters of the DNA sequence when splitting it into oligonucleotides.

Materials and methods

General information

BAC-browser software was written in VB.NET 9.0. The program loads genomes in Fasta, imports from full Genbank format and creates from text field. The program loads annotation in its own format.annot, imports from full Genbank format and creates from graphic user interface.

Applications for sequence design

Nucleotide sequence can be obtained by reverse translation of a protein sequence. For reverse translation the table of intended codon usage is required. The codon usage table is a TAB-separated file. The first column contains nucleotide triplet. The second column contains the corresponding amino acid in a single-letter code. The third column contains the codon usage in percent format. For example: [CTGL0.65]. When nucleotide sequence is generated, the codons will be used according to the specified frequency.

Standard regulatory sequences correspond to consensus E. coli promoter TTGACA[N]₁₇TATAAT, consensus Shine-Dalgarno sequence AGGAGG and strong promoter of Mollicutes TATGNTATAAT[N]₆G.

Applications for sequence analysis

The number of tools for molecular biology, including those for visualization of coding sequences (CDS), regulatory elements, restriction sites, and GC-content, as well as those for assessing secondary structure stability and searching for homologous sequences were described in our previous work [22].

Applications for generation of regular oligonucleotides

Regular design of the oligonucleotides implies the split of the sequence into a set of oligonucleotides of equal length, which is set by user. The oligonucleotides cover the sequence without gaps.

Applications for generation of thermodynamically aligned oligonucleotides

The thermodynamics of oligonucleotide annealing is calculated by SantaLucia method [23]. At first step the algorithm calculates target dG of the sticky ends. Currently target dG is calculated as average dG of sliding window. The sliding window length corresponds to the average sticky end length according to the current parameters. Next the algorithm designs sticky ends. The algorithm aims to make the sticky end dG as close as possible to the target dG. First sticky region starts at the sequence beginning. For ungapped oligonucleotides design the only parameter that the algorithm can vary to adjust the sticky end dG is its length. For gapped oligonucleotide design the spacer between the sticky ends can be added. The algorithm generates a set of allowed sticky ends and selects the one with dG closest to the target value. At the final step sticky ends are assembled into oligonucleotides. For ungapped design each oligonucleotide consists of two adjacent sticky ends. For gapped design it also includes a spacer between the sticky ends.

Applications for fragments generation

The algorithm can split the sequence into shorter fragments for assembly. The average fragment length is set by the user. The algorithm splits the sequence into the fragments of defined length. Then the algorithm searches the optimal areas for fragment assembly according to the defined thermodynamics parameters within ± 100 bp centered on the border between the fragments for each pair of adjacent fragments. Further the algorithm moves the borders of fragments accordingly.

Gene assembly using constructed oligonucleotides by BAC-browser

LCR (Ligase Chain Reaction) assembly. 10 µL of oligonucleotides mix with 1 µM concentration of each oligo was phosphorylated using T4 Polynucleotide Kinase (T4 PNK) (Thermo Fisher Scientific) according to the manufacturer's protocol. 2 µL of phosphorylated oligonucleotides with 0.5 µM concentration of each oligo were mixed with 18 µL solution containing 1 µL Taq DNA Ligase (NEB) and 2 µL Taq DNA Ligase Reaction Buffer (10X) (NEB) and were cyclically ligated. The reaction was heated to 95 °C for 3 min, followed by 22 cycles of 95 °C for 15 s, annealing temperature for 5 min. The annealing temperatures for assembly condition variation were 53, 54.5, 57, 60.5, 63.5 or 65 °C. Full gene was amplified from 2 µL of previous reaction using flanking primers and Tersus polymerase (Evrogen) according to the manufacturer's protocol. The reaction was heated to 95 °C for 3 min, followed by 25 cycles of 95 °C for 30 s, 58 °C for 30 s, 72 °C for 2 min, this was followed by a final extension of 72 °C for 5 min. The flanking primers were TATAGGATCCATGTTATTGACTGGCAAATTATACAAAGAA and TATAGTCGACTCATTTTAAACTCTTTCTAAGCTG for gp49, TATAGGATCCATGGAAAAGAAAATTACAGG and TATACTGCAGTTAAGCACCACC for camR. Sequences of gp49 and camR genes are available in Supplementary 1.

PCR (Polymerase Chain Reaction) assembly. 1 µL of oligonucleotides mix with 1 µM concentration of each oligo were elongated using Tersus polymerase (Evrogen) according to the manufacturer's protocol, except 10 mM MgCl₂ concentration. The reaction was heated to 95 °C for 3 min, followed by 20 cycles of 95 °C for 30 s, annealing temperatures for 30 s, 72 °C for 1 min. The annealing temperatures for assembly optimization were 53, 54.5, 57, 60.5, 63.5 or 65 °C. Full gene was amplified from 1 µL of previous reaction using flanking primers and Tersus polymerase (Evrogen) according to the manufacturer's protocol. The reaction for gp49 or camR gene was heated to 95 °C for 3 min, followed by 18 cycles of 95 °C for 30 s, 58 °C for 30 s, 72 °C for 1 min, this was followed by a final extension of 72 °C for 5 min. The reaction for bseRI gene fragment was heated to 95 °C for 3 min, followed by 15 cycles of 95 °C for 30 s, 62 °C for 30 s, 72 °C for 1.5 min, this was followed by a final extension of 72 °C for 5 min. The flanking primers for gp49 and camR were the same as for LCR assembly, the primers for bseRI gene fragment were ATGAACAATAGTGAAAAGCAAGTTGAGCTAGCTAGAGAGTGTATAATCGCTAGTTTGG and AGCTGAGAAATTATAAGCCATAAGATCTTCCTGCATTTCTTGTGCCACAGGTATTATTTTTTCAAGA. For specificity of assembly experiment all conditions were the same except the cycles in second step of assembly were 20 for gp49 or camR genes and 18 for bseRI first gene fragment.

BseRI full-gene assembly. BseRI full gene were assembled from 3 overlapping fragment with 1, 1.1 and 1.3 kb in length. Beforehand, all these fragments were assembled under conditions similar to bseRI first gene fragment. 1 µL of each amplicon of bseRI gene fragment were mixed and elongated using Tersus polymerase (Evrogen) according to the manufacturer's protocol. The reaction was heated to 95 °C for 3 min, followed by 15 cycles of 95°C for 30 s, 62 °C for 30 s, 72 °C for 5 min, this was followed by a final extension of 72 °C for 5 min. Than flanking primers for full bseRI gene were added and the reaction was heated to 95 °C for 3 min, followed by 15 cycles of 95 °C for 30 s, 62 °C for 30 s, 72 °C for 5 min, this was followed by a final extension of 72 °C for 5 min.

Full sequences of the assembled DNA fragment are in Supplementary 1 from Supplementary file. Oligonucleotide sequences are in Supplementary 2 from Supplementary file.

Assembled DNA fragments were analyzed on 1% agarose gel. GeneRuler 100 bp DNA ladder (Thermo Fisher Scientific) was used as L1 and GeneRuler 1 kb DNA ladder (Thermo Fisher Scientific) was used as L2.

Amplicons were quantified with dsDNA BR Kit (Thermo Fisher Scientific) using fluorometer Qubit 2.0 (Thermo Fisher Scientific).

All assembled genes were cloned into BamHI/SalI-cut pET15 plasmid, and their sequences have been confirmed using Sanger sequencing.

Results

Sequence design and analysis

We enhanced the genomic browser we developed earlier for the design and assembly of extended genomic sequences. It implements a graphical interface that allows users to visualize the sequence of interest with an annotation. Earlier, in our work [22], we described the use of the browser we created for the purposes of molecular and systems biology. Descriptions of many widely used molecular biological tools can be found there.

The graphical shell of the browser can display the nucleotide sequence in a linear (default view) or circular (View—Circular map view) view with an annotation of coding, structural, regulatory, additional sequences, etc. The sequence with the annotation can be exported from the generally accepted full Genbank (.gb) file, and also created using the File—New menu.

For the design of artificial genomic sequences, there are also functions for adding a nucleotide sequence with an annotation from the context menu of markers, as well as separately adding regulatory elements (promoters) according to a template. Promoter templates for Escherichia coli and Mycoplasma gallisepticum organisms are supported. Single or double markers can be used to highlight any coordinate of the nucleotide sequence. In addition to adding, editing and deleting sequence functions are supported.

The genomic browser provides functions for generating nucleotide sequences from amino acid sequences according to the table for the frequency of occurrence of codons for the target organism (Tools—Sequence—Reverse translate), generation of random sequences of a certain nucleotide composition (Tools—Random sequence). The table for the frequency of occurrence of codons can be pre-generated by the Tools—Codon usage function. The Annotation—Add features function is available separately.

To optimize the nucleotide sequence before assembly, it is possible to visualize restriction sites (the left panel of the program, the Sites tab. By default, a large set of restriction enzymes is supported, and it can be expanded by the user). It is possible to visualize GC-composition, GC- and AT-skew, as well as to assess the possibility of formation and stability of the secondary structure (View—Function—the desired parameter). There is an algorithm for finding homologous sections inside the constructed sequence (the markers context menu of the selected fragment—Send sequence to—Repetitive sequences finder). Checking for the presence of such sites is necessary to avoid the possible appearance of non-target assembly products due to cross-hybridization of oligonucleotides with homologous sites.

Oligonucleotides generation for DNA assembly

We developed three algorithms for oligonucleotides generation for DNA assembly: Regular, which splits the sequence into equal fragments of a defined length; Ungapped and Gapped, which splits the sequence into fragments of a different calculated length (Figs. 1 and 2). Generated oligos with Ungapped and Gapped algorithms uses thermodynamics optimization of segments. Ungapped algorithms splits duplex DNA into oligonucleotides without gaps. In this case, the generated single-stranded oligonucleotides overlap the entire sequence of double-stranded DNA. The Gapped algorithm can introduce gaps between oligonucleotides in such a way that some fragments of double-stranded DNA are covered with two overlapping oligonucleotides, and some with only one. This saves the number of synthetic oligonucleotides and the cost of DNA synthesis by about a third. Oligonucleotides created with Regular and Ungapped algorithms can be assembled by both LCR and PCR assemblies, whereas oligonucleotides created with Gapped algorithm can be assembled by PCR assembly. The algorithms are available from the main menu (Tools—Synthetic biology—Design assembly) or from the context menu of the selected with markers DNA fragment.

Fig. 1

Splitting the nucleotide sequence into oligonucleotides without gaps in the BAC-browser. A The window of the browser with splitting into oligonucleotides. Oligonucleotides are shown in pink, and sticky ends optimized for the thermodynamics of duplex formation are highlighted in gray. B The Assembly design window with available parameters on the left side and thermodynamic parameters for splitting on the right side

Fig. 2

Splitting the nucleotide sequence into oligonucleotides with gaps in the BAC-browser. Also, this sequence was previously split on three overlapping fragments, each of which will be assembled from oligonucleotides separately. A The window of the browser with splitting into oligonucleotides. Oligonucleotides are shown in pink, and sticky ends optimized for the thermodynamics of duplex formation are highlighted in gray. Three fragments F1-F3 are shown in grey. B The Assembly design window with available parameters on the left side and thermodynamic parameters for splitting on the right side

For Ungapped and Gapped algorithms it is possible to visualize thermodynamic characteristics of generated oligo pool: sequence dG profile, sticky segment dG and Tm distribution (Figs. 1 and 2). For Regular split, Ungapped and Gapped algorithms generated oligos are visualized in annotation interface. If necessary, the operator can make manual edits to the DNA splitting.

For assembly long DNA sequences, we developed algorithm that split the sequence into shorter overlapping fragments that can be assembled from synthetic oligonucleotides. The algorithm finds the thermodynamic optimal overlapping ends for fragments within the limits of the defined parameters. Then each fragment is split into oligonucleotides for synthesis by previously described algorithm. The design made in this way can be assembled by various methods of long-overlap-based assemblies [24,25,26]. In BAC-browser software this algorithm is realized in the function Fragments design that generated overlapping fragments of user-defined length and flanking primers for it. User can use this function from the main menu (Tools—Synthetic biology—Design assembly—Gapped assembly type) (Fig. 2). In the field Automatic fragment design choose Thermodynamic alignment and desired average length of fragments. For generation oligos without gaps between it, user can change Gap length parameter in Assembly parameter field to 0. Additionally, for generation of flanking primers user can check the box Primers in the field Automatic fragment design. User can add to the primers the additional sequences using for cloning in popular pTZ and pET plasmids, to do this, select the desired option in the drop-down menu next to the primer generation function. Additional parts of primers are not visualized in graphic interface but exported as a part of primers.

Export generated oligos is possible via Annotation—Export nucleotide sequences. In Sequence Exporter choose required assembly set in the Source of features field and preferred file format for save (Tab-separated table or Fasta format).

If user have already generated tab-separated file of oligonucleotides name and sequences, user can map it on reference using function Annotation—Map sequences.

Gene assembly using constructed oligonucleotides by BAC-browser

We tested developed software for gene synthesis. The wild-type phage endonuclease IV gene gp49 (495 base pairs) was split with Regular algorithm into oligonucleotides of 40 bases in length with 20 bp overlaps. The chloramphenicol resistance gene camR (680 base pairs) was optimized by codon usage for Mycoplasmas and then split with Ungapped algorithm into oligos with complementary segments with average lengths of 20 ± 5 bases. All manipulation using BAC-browser were made with default settings.

The two generated oligo pools without gap algorithms can be used for both, LCR and PCR gene assemblies. All these assemblies have two rounds: first is oligos hybridization and assembly using either ligation (with preliminary phosphorylation) or synthesis by polymerase, second is amplification of target fragment by polymerase using flanking primers [27,28,29]. Both of the aforementioned genes were successfully assembled using both methods (Fig. 3A, B).

Fig. 3

Genes assemblies using oligonucleotides generated with different algorithms by BAC-browser. The figure shows the varying annealing temperature at the first stage of assembly, L1 is a DNA ladder. The bottom panel under each EF gel shows the amount of DNA in the sample. A LCR assembly of gp49 with regular split and camR with thermodynamic ungapped split under different conditions; B PCR assembly of gp49 with regular split, camR with thermodynamic ungapped split and fragment of bseRI gene with thermodynamic gapped split under different conditions

The first 1 kb gene fragment of restriction endonuclease BseRI was edited by replacing restriction site of SalI and split into oligonucleotides with gaps using Gapped algorithm. The oligonucleotides lengths were 60 ± 7 bases in average, the average gap length was 10 bases with maximum gap was 19. These gene fragment was successfully assembled using PCR gene assembly (Fig. 3B).

To test the specificity of the gene assembly, we checked whether non-specific oligonucleotides interfere with the assembly of DNA fragments. We mixed oligonucleotides necessary to assembly of 3 fragments, gp49 and camR genes and first 1000 bp bseRI gene fragment in one reaction and carried out first step of PCR assembly. The second steps of assembly with flanking primers were carried out separately for each gene. We showed that the sequences were assembled in wide range of conditions with slightly lower efficiency that in single gene reactions (Fig. 4). Thus, gene assembly using generated by BAC-browser oligonucleotides is specific.

Fig. 4

Specificity of genes assemblies by PCR methods under different conditions. The figure shows the varying annealing temperature at the first stage of assembly, L1 and L2 are DNA ladders. Oligos for three genes were mixed together in first round of assembly, and the gene of interest was amplified using flanking primers in second round

We assembled full bseRI gene with a length of 3.5 kb from 3 overlapping fragment 1, 1.1 and 1.3 kb. Second and third gene fragment were designed and assembled from oligonucleotides as the first fragment (Fig. 5A–C). All three fragment were mixed together and successfully assembled using PCR overlap extension methods (Fig. 5 D).

Fig. 5

BseRI full-gene assembly, L2 is a DNA ladder. A PCR assembly of bseRI first gene fragment with thermodynamic gapped split under different conditions; B PCR assembly of bseRI second gene fragment with thermodynamic gapped split under different conditions; C PCR assembly of bseRI first gene fragment with thermodynamic gapped split under different conditions; D PCR assembly of bseRI full gene from three gene fragments assembled under different conditions

Discussion

Currently, genome synthesis is not a routine procedure yet. As a result, software for synthesizing extended DNA fragments is still being developed. We offer a synthetic genomics program with a graphical interface based on genomic browser, where it is possible to annotate and visualize all the necessary sequence elements. We believe this is a friendly interface for genetic engineers and synthetic biologists to work with.

We have developed several tools necessary for the design of synthetic sequences and the assembly of DNA sequences from synthetic oligonucleotides. These tools have been incorporated into the browser. In the first part of Results, we described functions related to sequence design, and further we have focused on oligonucleotides generation for de novo DNA assembly and practical efficiency of assembly. We have showed that using our oligonucleotides design, model DNA fragments were assembled in wide range of conditions. In our other recent work in the field of synthetic genomics, we have optimized the conditions for the PCR assembly of de novo DNA synthesis reaction [30].

BAC-browser v.2.1 is our suggestion for ergonomic and comfortable combination of function, that necessary for synthetic genomics for our experience. It is freely distributed, has graphic user interface and can be run on Windows. It visualizes genome with annotation and allows edit sequence and annotation as some commercial programs and, for example, UGENE [38]. BAC-browser can generate nucleic acid sequences from amino acid sequences according to the table for the frequency of codons. Currently, separate programs with advanced algorithms are being developed to solve this task [39]. At the current stage of the design of synthetic genomes, we have not started to develop automated sequence improvement for assembly, as this may lead to the loss of some biological, for example, regulatory functions. However, there are programs that optimize the DNA sequence [20, 40]. Designed sequences can be split into oligonucleotides for DNA assembly with Regular, and thermodynamic Ungapped and Gapped algorithms. Analogous software with their own algorithm are DNAWorks [13] Gene2Oligo [14], FastPCR [15] and POSoligo [16]. If the length of required sequence exceeds few kbp or reaches the length of the genome, BAC-browser generated overlapping fragments of user-defined length that can be assembled from oligonucleotides. Analogous function exists in programs [19, 21] or unpublished program GeneCut (http://genecut.unipro.ru).

Using described tools of BAC-browser our group have constructed Tn4001-based vector with codon-optimized for Mycoplasma gallisepticum transposase and tetracycline resistance protein TetM. This vector we have used to make a knockout mutants of M. gallisepticum [31, 32], mutants with overexpression regulator protein MraZ [33] and WhiA [34], and mutants with CRISPRi system for the repression of gene expression in mycoplasmas [35]. We have obtained a recombinant large fragment of Bacillus stearothermophilus DNA polymerase (Bst-pol) after codon optimization in Escherichia coli strain Bl21-gold [36]. Recently, using these tools, we have assembled full genome of Vibriophage N4 against Vibrio cholerae [37].

Conclusion

BAC-browser is a multifunctional genome browser that combines functions for molecular and systems biology and synthetic genomics. In this work we describe the extensions of BAC-browser version 2.1 for synthetic genomics. The program provides tools for both design of nucleotide sequences and oligonucleotides generation for oligo-based DNA assembly.

We have developed three methods of oligonucleotides generation for in vitro DNA assembly. We have shown that they can be used for LCR and PCR DNA assemblies. The genes from the oligonucleotides obtained using the BAC browser were assembled specifically and under a wide range of conditions. For long DNA sequences we have developed an algorithm that splits them into shorter overlapping fragments that can be assembled from synthetic oligonucleotides. We have demonstrated that using this algorithm it is possible to assemble a DNA fragment with a length of 3.5 kb.