DOI: 10.1002/cbic.201500061
Highlights
Engineering a Bacterial Tape Recorder Alexander Prokup and Alexander Deiters*[a] A method has been developed to produce and integrate single-stranded DNA into genomic locations in bacteria in response to exogenous signals. The system functions similarly to a cellular tape recorder by writing information into DNA and reading it at a later time. Much like other cellular memory platforms, its operation is based on DNA recombinase function.
However, the scalability and recording capacity have been improved over previous designs. In addition, memory storage was reversible and could be recorded in response to analogue inputs, such as light exposure. This modular memory writing system is an important addition to the genomic editing toolbox available for synthetic biology.
Nucleic acids, specifically DNA, have a significant capacity for information storage. DNA is the universal data storage device, providing the genetic information for all living organisms. However, utilizing DNA to record actual memory states remains a challenge. Information storage and retrieval have been achieved through highly parallelized DNA synthesis and sequencing, respectively.[1] Impressively, over 600 000 bytes have been recorded and accessed almost perfectly.[2] However, fabrication of these databanks relied on synthetic DNA, and the information is static. The DNA of living systems can also be used for memory storage; however, the size of storable information much smaller compared to synthetic information banks.[3] An advantage of using living systems is the ability to interface the memory state with biomolecules and cellular processes. Recently, recombinases have been used to write digital information into genomic memory[3, 4] and create reversible set and reset switches.[5] Recombinases are a popular choice for constructing memory-recording devices due to their precise and programmable insertion or excision of DNA. However, previous recombinase-based storage technologies depended upon preinserting recognition sites in the genome, which is laborious and limits storage capacity. Addressing the limitations of previous live-cell memory systems, Farzadfard and Lu developed a genetic system they termed SCRIBE (synthetic cellular recorders integrating biological events).[6] Their system is capable of producing and inserting small oligonucleotide inputs into the genomic DNA of bacterial cells to store memories. In order to create the required oligonucleotides, a gene cassette, known as a retron, that encodes a reverse transcriptase[7] and two RNA transcripts, msr and msd, was used (Figure 1 A). The msd transcript acts as a template, and the msr transcript acts as a primer for the retron RT, which in turn generates a single-stranded DNA (ssDNA) and RNA hybrid known as multicopy single-stranded DNA (msDNA). No ssDNA was produced when the entire msd sequence was replaced with the sequence of interest. There-
fore, a portion of the msd sequence was retained to enable transcription of ssDNA by the RT. The expressed customizable ssDNA enabled specific targeting of loci in the bacterial genome and their mutagenesis, this allowed information to be stored in DNA. The ssDNA oligomer is bound by b-recombinase and directs a site-specific genome editing event. In contrast to more common site-specific recombinases, such as Cre recombinase, b-recombinase does not require recognition sites to be genetically engineered into the target DNA and has been used to insert short synthetic oligonucleotides of 30–90 base pairs.[8] As a proof-of-concept experiment, a chemical input, isopropyl b-d-1-thiogalactopyranoside (IPTG), was used to initiate a DNA writing event. (Figure 1 B). The result of the writing event was the mutation of two stop codons in a kanamycin (kan) resistance gene back to the wild-type sequence, thus enabling a readout through bacterial growth on kan-containing plates. This confirmed that a record of the writing event was stored in the cell population using SCRIBE. The writing of a DNA sequence was also shown to be reversible. The authors used two different chemical inputs to sequentially initiate two separate writing events in the same cell. Mutations were encoded in two ssDNAs produced by the individual inputs. Both mutations targeted the same gene, thus enabling the bacteria to selectively metabolize specific carbon sources. One mutation turned the gene off by incorporating stop codons; the other mutation turned the gene on by turning the stop codons back to wild-type sequences. Sequential addition of inputs to the same bacterial cell population turned the gene ON!OFF and later OFF!ON. Reversing the memory state demonstrated that DNA events are rewritable and could be triggered in response to environmental changes (e.g., the presence of small molecules). Additionally, memory storage is not limited by the number of loci, as multiple memories can be targeted to the same locus. The input detection, write, and read operations of the SCRIBE system could also be decoupled from each other. After an input was recorded by the writing mechanism discussed above, the memory was read out later with the addition of a second input. As shown in Figure 2 A, the input, acyl homoserine lactone (AHL), initiated a writing event that reverted
[a] A. Prokup, A. Deiters Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260 (USA) E-mail: [email protected]
ChemBioChem 0000, 00, 0 – 0
These are not the final page numbers! ÞÞ
1
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Highlights
Figure 2. A) A decoupled SCRIBE system enabled independent control of the write and read operations. When the SCRIBE cassette was induced by AHL, an ssDNA was produced by the input, which was written into the lacZ gene by aTc-triggered b-recombinase. At a later stage, the information was read out enzymatically through IPTG-induced LacZ expression. B) The SCRIBE system was also engineered to allow for light as an external input signal. The duration of the light exposure was directly proportional to the recombination frequency, thereby allowing the storage of analogue memory in cell populations. Modified with permission from ref. [5]. Copyright AAAS 2012.
Figure 1. A) Transcription of a SCRIBE gene cassette to create RNA–ssDNA molecules. After induction with IPTG, the msr (black line) and msd (blue line) RNAs fold and bind to complementary regions. A reverse transcriptase (RT) begins transcription at a conserved guanosine residue (G). A customizable sequence (red line) can be inserted within msd, and thus be generated as ssDNA in bacterial cells. The single-stranded sequence is part of an RNA and ssDNA hybrid known as multicopy single-stranded DNA (msDNA). B) Writing memory into a kanamycin gene (kanR) by using a SCRIBE cassette. Induction with the IPTG input initiated transcription of the custom msr–msd sequence and production of ssDNA (red line). A second induction with anhydrotetracycline (aTc) allowed for expression of b-recombinase (blue ovals), which inserted the ssDNA into the genome, mutating the kanamycin resistance gene. Black stars represent stop codons. Modified with permission from ref. [5]. Copyright AAAS 2012.
ration or amount of input, thereby storing an analogue signal. Light was used as an exogenous input upstream of the SCRIBE cassette by interfacing ssDNA expression with the yf1/fixJ system[9] through inhibition of the transcription of a repressor protein of the SCRIBE cassette in the presence of light (Figure 2 B). The cumulative amount of light received by the bacteria was directly proportional to the number of writing events that occurred. The DNA was stored in the genomic memory culminating in the mutation of a gene. Thus, modification of the SCRIBE cassette with photo-responsive genes enabled optogenetic genome editing. Much like other DNA-based memory storage systems,[3, 5] SCRIBE is able to detect an exogenous input and store a record of the event in DNA. However, one of the important technical barriers that the SCRIBE system overcomes is the ability to record analogue information. Storing the magnitude of an input enables specific details about an input to be recorded, such as its concentration or duration of exposure, other than its mere presence.
stop codons in a lacZ gene. Beta-recombinase was expressed after induction with anhydrotetracycline (aTc) and incorporated the newly produced ssDNA into the genome. In order to read the event, a second induction with IPTG induced expression of lacZ. The decoupled read and write SCRIBE system allowed for “sample-and-hold” logic, in which the writing and reading events are controlled independently of input detection. Memory states can be saved until reading is induced at a later time. Delaying the writing event might facilitate integration of the decoupled SCRIBE system with other computation circuits by providing temporal control. In addition to detecting the presence or absence of an input, the analogue SCRIBE system was able to record the du-
&
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
2
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Highlights lation. Additionally, the current design only functions in bacterial cells and would need to be modified for use in eukaryotic systems. This could potentially be achieved by translating the concepts developed here to the CRISPR/Cas9 system[10] as an alternative to the ssDNA-guided b-recombinase. However, the ability to write and rewrite analogue information into bacterial DNA has numerous potential applications, from dynamic and conditional genome editing in response to different input signals to the engineering of cells that can remotely interrogate their environment and memorize it for a later readout.
In many ways, the SCRIBE system contains many of the same functions as a tape recorder, but utilizes chemicals and cellular biomolecules rather than electricity and magnetism. Instead of recording events onto a magnetic tape, the SCRIBE system uses genomic DNA. Similar to pressing the “record” button, a chemical inducer can be added to induce the expression of b-recombinase, which will integrate the ssDNA into the genome. To “play” back the recorded information, a reporter gene is expressed, and its activity is measured. Just as a tape can be erased or recorded over, the previously written information can be rewritten by incorporating a different DNA oligomer at the same genomic location. Much like the dynamic range of an audio recording on tape, the analogue SCRIBE system will record the magnitude of an input, not just the mere presence of the input. Thus, SCRIBE is able to operate similarly to a tape recorder in contrast to a computer hard drive. Direct application of the SCRIBE system still requires consideration of certain limitations. One limitation, discussed by the researchers is the finite number of orthogonal inducible promoters available as inputs. Thus, careful attention to circuit design might be necessary for larger circuits, which typically utilize multiple inputs and outputs. Writing the genomic memory at more than two loci could also prove challenging. The recombination frequency for two writing events within a single cell was 1000 times less efficient than a single writing event. This decrease in efficiency could restrict the total number of writable inputs at the single-cell level. However, multiple signals can be recorded into genomic memory and distributed across distinct subpopulations of a bacterial popu-
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
These are not the final page numbers! ÞÞ
Keywords: genomic memory · recombinases · synthetic biology
[1] N. Goldman, P. Bertone, S. Chen, C. Dessimoz, E. M. LeProust, B. Sipos, E. Birney, Nature 2013, 494, 77 – 80. [2] G. M. Church, Y. Gao, S. Kosuri, Science 2012, 337, 1628. [3] L. Yang, A. A. Nielsen, J. Fernandez-Rodriguez, C. J. McClune, M. T. Laub, T. K. Lu, C. A. Voigt, Nat. Methods 2014, 11, 1261 – 1266. [4] P. Siuti, J. Yazbek, T. K. Lu, Nat. Biotechnol. 2013, 31, 448 – 452. [5] J. Bonnet, P. Subsoontorn, D. Endy, Proc. Natl. Acad. Sci. USA 2012, 109, 8884 – 8889. [6] F. Farzadfard, T. K. Lu, Science 2014, 346, 1256272. [7] H. M. Temin, Nature 1989, 339, 254 – 255. [8] H. M. Ellis, D. Yu, T. DiTizio, D. L. Court, Proc. Natl. Acad. Sci. USA 2001, 98, 6742 – 6746. [9] A. Mçglich, R. A. Ayers, K. Moffat, J. Mol. Biol. 2009, 385, 1433 – 1444. [10] L. Cong, F. A. Ran, D. Cox, S. Lin, R. Barretto, N. Habib, P. D. Hsu, X. Wu, W. Jiang, L. A. Marraffini, F. Zhang, Science 2013, 339, 819 – 823. Received: February 3, 2015 Published online on && &&, 0000
3
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
HIGHLIGHTS A. Prokup, A. Deiters*
A DNA tape recorder: A DNA-based storage platform, termed SCRIBE, was developed to record analogue information in bacterial cells. The system functions similarly to a tape recorder by utilizing different chemical or physical inputs to write, read, and rewrite information stored in genomic DNA.
&& – && Engineering a Bacterial Tape Recorder
&
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
4
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Highlights
Engineering a Bacterial Tape Recorder Alexander Prokup and Alexander Deiters*[a] A method has been developed to produce and integrate single-stranded DNA into genomic locations in bacteria in response to exogenous signals. The system functions similarly to a cellular tape recorder by writing information into DNA and reading it at a later time. Much like other cellular memory platforms, its operation is based on DNA recombinase function.
However, the scalability and recording capacity have been improved over previous designs. In addition, memory storage was reversible and could be recorded in response to analogue inputs, such as light exposure. This modular memory writing system is an important addition to the genomic editing toolbox available for synthetic biology.
Nucleic acids, specifically DNA, have a significant capacity for information storage. DNA is the universal data storage device, providing the genetic information for all living organisms. However, utilizing DNA to record actual memory states remains a challenge. Information storage and retrieval have been achieved through highly parallelized DNA synthesis and sequencing, respectively.[1] Impressively, over 600 000 bytes have been recorded and accessed almost perfectly.[2] However, fabrication of these databanks relied on synthetic DNA, and the information is static. The DNA of living systems can also be used for memory storage; however, the size of storable information much smaller compared to synthetic information banks.[3] An advantage of using living systems is the ability to interface the memory state with biomolecules and cellular processes. Recently, recombinases have been used to write digital information into genomic memory[3, 4] and create reversible set and reset switches.[5] Recombinases are a popular choice for constructing memory-recording devices due to their precise and programmable insertion or excision of DNA. However, previous recombinase-based storage technologies depended upon preinserting recognition sites in the genome, which is laborious and limits storage capacity. Addressing the limitations of previous live-cell memory systems, Farzadfard and Lu developed a genetic system they termed SCRIBE (synthetic cellular recorders integrating biological events).[6] Their system is capable of producing and inserting small oligonucleotide inputs into the genomic DNA of bacterial cells to store memories. In order to create the required oligonucleotides, a gene cassette, known as a retron, that encodes a reverse transcriptase[7] and two RNA transcripts, msr and msd, was used (Figure 1 A). The msd transcript acts as a template, and the msr transcript acts as a primer for the retron RT, which in turn generates a single-stranded DNA (ssDNA) and RNA hybrid known as multicopy single-stranded DNA (msDNA). No ssDNA was produced when the entire msd sequence was replaced with the sequence of interest. There-
fore, a portion of the msd sequence was retained to enable transcription of ssDNA by the RT. The expressed customizable ssDNA enabled specific targeting of loci in the bacterial genome and their mutagenesis, this allowed information to be stored in DNA. The ssDNA oligomer is bound by b-recombinase and directs a site-specific genome editing event. In contrast to more common site-specific recombinases, such as Cre recombinase, b-recombinase does not require recognition sites to be genetically engineered into the target DNA and has been used to insert short synthetic oligonucleotides of 30–90 base pairs.[8] As a proof-of-concept experiment, a chemical input, isopropyl b-d-1-thiogalactopyranoside (IPTG), was used to initiate a DNA writing event. (Figure 1 B). The result of the writing event was the mutation of two stop codons in a kanamycin (kan) resistance gene back to the wild-type sequence, thus enabling a readout through bacterial growth on kan-containing plates. This confirmed that a record of the writing event was stored in the cell population using SCRIBE. The writing of a DNA sequence was also shown to be reversible. The authors used two different chemical inputs to sequentially initiate two separate writing events in the same cell. Mutations were encoded in two ssDNAs produced by the individual inputs. Both mutations targeted the same gene, thus enabling the bacteria to selectively metabolize specific carbon sources. One mutation turned the gene off by incorporating stop codons; the other mutation turned the gene on by turning the stop codons back to wild-type sequences. Sequential addition of inputs to the same bacterial cell population turned the gene ON!OFF and later OFF!ON. Reversing the memory state demonstrated that DNA events are rewritable and could be triggered in response to environmental changes (e.g., the presence of small molecules). Additionally, memory storage is not limited by the number of loci, as multiple memories can be targeted to the same locus. The input detection, write, and read operations of the SCRIBE system could also be decoupled from each other. After an input was recorded by the writing mechanism discussed above, the memory was read out later with the addition of a second input. As shown in Figure 2 A, the input, acyl homoserine lactone (AHL), initiated a writing event that reverted
[a] A. Prokup, A. Deiters Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260 (USA) E-mail: [email protected]
ChemBioChem 0000, 00, 0 – 0
These are not the final page numbers! ÞÞ
1
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Highlights
Figure 2. A) A decoupled SCRIBE system enabled independent control of the write and read operations. When the SCRIBE cassette was induced by AHL, an ssDNA was produced by the input, which was written into the lacZ gene by aTc-triggered b-recombinase. At a later stage, the information was read out enzymatically through IPTG-induced LacZ expression. B) The SCRIBE system was also engineered to allow for light as an external input signal. The duration of the light exposure was directly proportional to the recombination frequency, thereby allowing the storage of analogue memory in cell populations. Modified with permission from ref. [5]. Copyright AAAS 2012.
Figure 1. A) Transcription of a SCRIBE gene cassette to create RNA–ssDNA molecules. After induction with IPTG, the msr (black line) and msd (blue line) RNAs fold and bind to complementary regions. A reverse transcriptase (RT) begins transcription at a conserved guanosine residue (G). A customizable sequence (red line) can be inserted within msd, and thus be generated as ssDNA in bacterial cells. The single-stranded sequence is part of an RNA and ssDNA hybrid known as multicopy single-stranded DNA (msDNA). B) Writing memory into a kanamycin gene (kanR) by using a SCRIBE cassette. Induction with the IPTG input initiated transcription of the custom msr–msd sequence and production of ssDNA (red line). A second induction with anhydrotetracycline (aTc) allowed for expression of b-recombinase (blue ovals), which inserted the ssDNA into the genome, mutating the kanamycin resistance gene. Black stars represent stop codons. Modified with permission from ref. [5]. Copyright AAAS 2012.
ration or amount of input, thereby storing an analogue signal. Light was used as an exogenous input upstream of the SCRIBE cassette by interfacing ssDNA expression with the yf1/fixJ system[9] through inhibition of the transcription of a repressor protein of the SCRIBE cassette in the presence of light (Figure 2 B). The cumulative amount of light received by the bacteria was directly proportional to the number of writing events that occurred. The DNA was stored in the genomic memory culminating in the mutation of a gene. Thus, modification of the SCRIBE cassette with photo-responsive genes enabled optogenetic genome editing. Much like other DNA-based memory storage systems,[3, 5] SCRIBE is able to detect an exogenous input and store a record of the event in DNA. However, one of the important technical barriers that the SCRIBE system overcomes is the ability to record analogue information. Storing the magnitude of an input enables specific details about an input to be recorded, such as its concentration or duration of exposure, other than its mere presence.
stop codons in a lacZ gene. Beta-recombinase was expressed after induction with anhydrotetracycline (aTc) and incorporated the newly produced ssDNA into the genome. In order to read the event, a second induction with IPTG induced expression of lacZ. The decoupled read and write SCRIBE system allowed for “sample-and-hold” logic, in which the writing and reading events are controlled independently of input detection. Memory states can be saved until reading is induced at a later time. Delaying the writing event might facilitate integration of the decoupled SCRIBE system with other computation circuits by providing temporal control. In addition to detecting the presence or absence of an input, the analogue SCRIBE system was able to record the du-
&
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
2
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Highlights lation. Additionally, the current design only functions in bacterial cells and would need to be modified for use in eukaryotic systems. This could potentially be achieved by translating the concepts developed here to the CRISPR/Cas9 system[10] as an alternative to the ssDNA-guided b-recombinase. However, the ability to write and rewrite analogue information into bacterial DNA has numerous potential applications, from dynamic and conditional genome editing in response to different input signals to the engineering of cells that can remotely interrogate their environment and memorize it for a later readout.
In many ways, the SCRIBE system contains many of the same functions as a tape recorder, but utilizes chemicals and cellular biomolecules rather than electricity and magnetism. Instead of recording events onto a magnetic tape, the SCRIBE system uses genomic DNA. Similar to pressing the “record” button, a chemical inducer can be added to induce the expression of b-recombinase, which will integrate the ssDNA into the genome. To “play” back the recorded information, a reporter gene is expressed, and its activity is measured. Just as a tape can be erased or recorded over, the previously written information can be rewritten by incorporating a different DNA oligomer at the same genomic location. Much like the dynamic range of an audio recording on tape, the analogue SCRIBE system will record the magnitude of an input, not just the mere presence of the input. Thus, SCRIBE is able to operate similarly to a tape recorder in contrast to a computer hard drive. Direct application of the SCRIBE system still requires consideration of certain limitations. One limitation, discussed by the researchers is the finite number of orthogonal inducible promoters available as inputs. Thus, careful attention to circuit design might be necessary for larger circuits, which typically utilize multiple inputs and outputs. Writing the genomic memory at more than two loci could also prove challenging. The recombination frequency for two writing events within a single cell was 1000 times less efficient than a single writing event. This decrease in efficiency could restrict the total number of writable inputs at the single-cell level. However, multiple signals can be recorded into genomic memory and distributed across distinct subpopulations of a bacterial popu-
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
These are not the final page numbers! ÞÞ
Keywords: genomic memory · recombinases · synthetic biology
[1] N. Goldman, P. Bertone, S. Chen, C. Dessimoz, E. M. LeProust, B. Sipos, E. Birney, Nature 2013, 494, 77 – 80. [2] G. M. Church, Y. Gao, S. Kosuri, Science 2012, 337, 1628. [3] L. Yang, A. A. Nielsen, J. Fernandez-Rodriguez, C. J. McClune, M. T. Laub, T. K. Lu, C. A. Voigt, Nat. Methods 2014, 11, 1261 – 1266. [4] P. Siuti, J. Yazbek, T. K. Lu, Nat. Biotechnol. 2013, 31, 448 – 452. [5] J. Bonnet, P. Subsoontorn, D. Endy, Proc. Natl. Acad. Sci. USA 2012, 109, 8884 – 8889. [6] F. Farzadfard, T. K. Lu, Science 2014, 346, 1256272. [7] H. M. Temin, Nature 1989, 339, 254 – 255. [8] H. M. Ellis, D. Yu, T. DiTizio, D. L. Court, Proc. Natl. Acad. Sci. USA 2001, 98, 6742 – 6746. [9] A. Mçglich, R. A. Ayers, K. Moffat, J. Mol. Biol. 2009, 385, 1433 – 1444. [10] L. Cong, F. A. Ran, D. Cox, S. Lin, R. Barretto, N. Habib, P. D. Hsu, X. Wu, W. Jiang, L. A. Marraffini, F. Zhang, Science 2013, 339, 819 – 823. Received: February 3, 2015 Published online on && &&, 0000
3
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
HIGHLIGHTS A. Prokup, A. Deiters*
A DNA tape recorder: A DNA-based storage platform, termed SCRIBE, was developed to record analogue information in bacterial cells. The system functions similarly to a tape recorder by utilizing different chemical or physical inputs to write, read, and rewrite information stored in genomic DNA.
&& – && Engineering a Bacterial Tape Recorder
&
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
4
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
