Team:Bielefeld-CeBiTec/Project/unnatural base pair/unnatural base pairs

Unnatural Base Pairs

Figure 1: Unnatural bases.


All amino acids are encoded by codons, which are defined by three base pairs. This information is encoded in the genome of an organism and since the origin of life every natural genome has consisted of the two-base-pair genetic alphabet dA-dT (adenine-thymine) and dG-dC (cytosine-guanine). There are strong efforts to replace a canonical base pair or expand the genetic code by a third unnatural base pair (UBP) (Martinot and Benner, 2004; Jiang and Seela, 2010; Kwok, 2012; Zhang et al., 2017; Yamashige et al., 2012; Seela et al., 2005; Switzer et al., 1989; Yang et al., 2011).
So far the modification of sugars and phosphates for nucleotides with important applications have been explored before. First experiments with unnatural bases extended the nucleotide alphabet by replacing thymine with 5-chlorouracil in E. coli over a period of 25 weeks (Dunn and Smith, 1957; Marlière et al., 2011). But for an UBP two modified nucleobases are needed. A. Rich discussed the extension of the DNA by two additional bases already in 1962 (Rich, 1962). An additional UBP can be interesting for physiochemical properties if the nucleobases can be site-specifically derivatized with linkers for chemical groups. Furthermore, the availability of an UBP in vivo would be a milestone in the field of synthetic biology. This would mean the creation of a semi-synthetic organism with distinguished storage capabilities for genetic information that leads to new and useful functions and applications (Malyshev and Romesberg, 2015).

UBPs with hydrogen bonding

Engineering an UBP is facing several challenges. At first it was focused on orthogonal pairing and realizing of in vitro replication. For this purpose UBPs with complementary hydrogen bonding were explored. Rapport and Benner laboratories independently investigated the UBP disoG-disoC, which is constitutional isomer of dG-dC. Main problems concerning this UBP are deaminiation and tautomerization that lead to mispairing with natural bases predominantly dT/U. Those problems resulted in further derivates of disoG-disoC, like the latest UBP dZ (6 6-amino-5-nitro-3-(1‘-β-D-2‘-deoxyribofuranosyl)-2(1H)-pyridone)-dP (2-amino-8-(1‘-β-D-2‘-deoxyribofuranosyl)-imidazol[1,2-α]-1,3,5-trizan-4(H)-one) from Benner laboratories that showed high-fidelity amplification by PCR (Yang et al., 2010). A Taq DNA polymerase was modified to accept the new ATCGPZ-DNA, resulting in a retention rate of 98.9% (Laos et al., 2014; Chen et al., 2011). The six-nucleotide genetic alphabet will lead to DNA with a B-form as well as an A-form, with the major groves being 1 Å wider than the natural G:C pair (Georgiadis et al., 2015). Also transcription as well as reverse transcription and even translation was successfully performed in vitro (Bain et al., 1992; Leal et al., 2015). Another UBP based on complementary hydrogen bonding is ds-dy, which are analogs to purine and pyridine developed by Hirao in 2000. in vitro transcription and translation was achieved using this UBP but the derivate dz with lower mispairing rates were insufficiently recognized by DNA and RNA polymerases as a triphosphate (Hirao et al., 2002; Hirao et al., 2004).

Other UBPs

Besides hydrogen bonding further research directing towards UBPs with metal-depending pairing, hydrophobic forces and ring stacking forces has been done (Malyshev and Romesberg, 2015). d5SICS – dMMO2 and d5SICS-dNaM are two promising candidates using hydrophobic interactions, which allowed transcription (Seo et al., 2009). The first demonstration in E. coli was based on one plasmid encoding the nucleoside triphosphate transporter for dNaM and d5SICS and the other plasmid encoding a gene sequence using the extended genetic code (Malyshev et al., 2014). Uptake of the synthetic bases as well as a stable plasmid replication over 24 generations was demonstrated (Malyshev et al., 2014). In 2017, the Romesberg group presented a new version of their semi-synthetic organism. The most important advances were an optimized transporter with improved uptake of unnatural triphosphates and better retention of XNA with dNaM-dTPT3. Furthermore, they used a CRISPR-Cas system to eliminate plasmids that lost the XNA (Zhang et al., 2017).

Our approach

The challenging part about using XNA is the need for synthetic or evolved proteins that allow for replication, transcription, and packaging of the XNA (Schmidt, 2010). For our approach to expand the genetic code we decided on the UBP disoG-disoCm (5-methyl-isocytosine). The 5-methyl derivative showed more stability towards hydrolysis than isoC (Tor and Dervan, 1993). The disoCm-disoGTP system also has an improved behavior concerning the in vitro transcription with T7 RNA polymerase. The presence of 5-methyl possibly results into a better contact between the template and the polymerase (Tor and Dervan, 1993).
Another aspect is the similarity of the unnatural bases isoG und isoCm to the natural bases guanine and cytosine while being an orthogonal system at the same time. Due to the structural similarity, there is better chance for compatibility with interacting enzymes. In 1992 the Benner laboratory showed, that the in vitro translation of mRNA containing disoC worked with a non-standard tRNA containing the purine complementary disoG inside the anticodon (Bain et al., 1992). Their cell free experiments showed a high specifity for the incorporation of a non-canonical amino acid by the ribosome using this unnatural base. With these stereoisomer of the natural bases it is more likely to achieve an optimized replication, transcription or translation with less adaption of the correspondent enzymes than with hydrophobic UBPs. On top of that, the hydrophobic UBPs are very expensive, because of their complex synthesis. Looking forward to create an autonomous synthetic organism it seems to be impossible to create a biosynthetic pathway for unnatural bases that differ a lot from natural bases. Whereas isoG is already known to be metabolic substance of the plant L. Croton tiglium. Revealing this metabolic pathway can make it usable for any synthetic organism and therefore stepping forward towards a fully autonomous synthetic organism.

De novo Synthesis of Purine and Pyrimidine Bases

De novo Synthesis of Pyrimidine Bases

The de novo synthesis of pyrimidines starts with the assembly of the orotate ring which is then converted into a pyrimidine nucleotide by binding the ring to a ribose phosphate (Berg et al., 2012). The first step in the synthesis of the pyrimidine ring is the formation of carbamoyl phosphate, which is formed from hydrogencarbonate and ammonia. The reaction is catalyzed by carbamoyl phosphate synthase (CPS) and requires two molecules of ATP. Glutamine is the main source for ammonia, which is produced by the hydrolysis of glutamine. This reaction is also catalyzed by CPS and yields ammonia and glutamate. Carbamoyl phosphate is converted into carbamoyl aspartate by aspartate carbamoyltransferase (ACT) through a reaction with aspartate. In turn, carbamoyl aspartate is oxidized to orotate, closing the ring structure. 5-phosphoribosyl-1-pyrophosphate (PRPP) reacts with orotate, a reaction that is catalyzed by orotate phosphoribosyltransferase (OPRTase). This reaction yields orotidylat, which in the next step is decarboxylated by oridylate decarboxylase (OCD) to uridylate (UMP). UMP acts as a precursor for the synthesis of cytidine. The first step in the synthesis of cytidine from UMP involves the phosphorylation of UMP to UTP . This reaction involves two steps. In the first step, UMP is converted to UDP by a specific nucleoside monophosphate kinase, the UMP kinase. ATP acts as a phosphate donor and is converted to ADP. UDP can now be converted to UTP by nucleoside diphosphate kinases, which are relatively unspecific. In the second step, UTP is converted to CTP in a reaction catalyzed by the cytidine triphosphate synthetase through the exchange of a carbonyl group with an amino group. CTP is subsequently converted into dCTP, a reaction that is catalyzed by ribonucleoside-triphosphate reductases (RTPR). RTPR also catalyzes the reaction of UTP to dUTP, which is then converted to dTMP through three consecutive reactions. dTMP is phosphorylated by dTMP kinases, yielding dTDP, which is then phosphorylated by nucleoside diphosphate kinases to dTTP.

Figure 1: De novo Synthesis of Pyrimidine Bases
.

De novo Synthesis of Purine Bases

Purine bases are produced de novo directly on the ribose (Berg et al., 2012). The synthesis starts with the replacement of the pyrophosphate of PRPP with an amino group, yielding phosphoribosylamine (PRA). This reaction is catalyzed by amidophosphoribosyltransferase (ATase) and also uses the ammonia from a glutamine side-chain as the donor of the amino group. The conversion of PRPP to PRA is a committing step in the purine biosynthesis. The synthesis of the purine ring involves nine additional steps with the first six reactions being relatively similar. In every reaction, an oxygen atom which is bound to a carbon atom is activated by phosphorylation and a subsequent substitution by ammonia or an amino-group, which act as a nucleophile agent. These subsequent reactions lead to the formation of inosinate (IMP), which acts as a key intermediate in the purine synthesis. Inosinate is converted into either AMP or GMP. AMP is synthesized by a substitution of the C-6 carbonyl oxygen with an amino group by adenylosuccinate synthase (ASS). In this reaction, GTP instead of ATP is used as a donor of the phosphoryl group. The conversion of IMP to GMP is catalyzed by the GMP synthase and starts with the oxidation of IMP to xanthylate (XMP) and the subsequent addition of an amino group. In a second step, XMP is converted into GMP, a reaction that requires ATP as a donor for an AMP group. GMP and AMP are again phosphorylated to GTP and ATP by specific kinases.

Figure 1: De novo Synthesis of Purine Bases

Conversion of Ribonucleosid Diphosphates to Deoxyribonucleotides

Deoxyribonucleotides are synthesized from ribonucleotides by substitution of the 2´-hydroxyl group of the ribose by a hydrogen. The reaction is catalyzed by the enzyme ribonucleotide reductase, which is strongly conserved in all living organisms (Berg et al., 2012). In E. coli, two main types of ribonucleotide reductases exist. Ribonucleoside-triphosphate reductases can convert ribonucleoside-triphosphates into deoxyribonucleoside-triphosphates, while ribonucleoside-diphosphate reductases convert ribonucleoside-diphosphates to deoxyribonucleoside-diphosphates (Kanehisa and Goto, 2000).

Salvage Pathways

Both purine and pyrimidine bases can be recycled and converted into the corresponding nucleotides through salvage pathways. Adenine can be recycled through conversion into AMP, a reaction that is catalyzed by the adenine phosphoribosyltransferase and requires PRPP . AMP can then be subsequently converted into ATP or dATP as described above. Hypoxanthine guanine phosphoribosyltransferase (HGPRT) catalyzes the recycling of guanosine, a reaction that also requires PRPP as a donor for a phosphate. HGPRT also catalyzes the conversion of hypoxanthine to IMP which again is a precursor of GMP and AMP. The recycling of thymine involves two steps: in the first step, thymine is converted to thymidine by the thymidine phosphorylase. In a second step, thymidine is converted to TMP by thymidine kinase. Cytosine can be recycled by conversion to uracil, a reaction that is catalyzed by cytosine deaminase. Following the conversion to UTP, CTP is produced by CTP synthase. The recycling of bases saves intracellular energy, since the de novo synthesis requires large amounts of ATP. Therefore, the recycling of bases through salvage pathways is usually favored by cells.

References

Berg, J.M., Tymoczko, J.L., and Stryer, L. (2012). Biochemistry 7th Edition. (Springer-Verlag: Berlin Heidelberg).
Kanehisa, M. and Goto, S. (2000). Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res. 28: 27–30.