On this page, we present the detailed Design-Build-Test-Learn (DBTL) cycles for our engineering efforts, categorized into four distinct parts.

Cascade recording system

Cycle 1: CRISPReporter over plasmids in-trans

Design

Our initial endeavour revolves around constructing a model aimed at reaching the fundamental objective of "recording in response to a stimulus" distributed across multiple plasmids (Fig 1a). In line with our design, the "knock-out" event has dual functions: recording this event and triggering the subsequent event (Fig 1b).
In detail, every plasmid is categorized as a "level", corresponding to the sequence of its response to the stimulus. The knock-out of a specific level's target accomplishes two things: it leaves a mark of the stimulation, and the product itself becomes the sgRNA for the next level. To make it easy to quantify, we incorporated a rationally designed qPCR sequence into the region to be deleted (Fig 2).

Build

To validate the first-level knockout, we designed a system consisting of three plasmids: pCas, level 1, and level 2.
The level 1 plasmid is constructed by inserting the fragments of homologous arm and N20 sequence into the designed vector and the level 2 plasmid is constructed by inserting the qPCR fragment into the designed vector. We outsourced the synthesis of these fragments to Atentares and Tsingke, with the addition of BsaI recognition sites at specific locations. After receiving the synthesized sequences, we performed Golden Gate Assembly to assemble these plasmids.

Test

During the initial attempt to construct level 1, we faced a challenge where the N20 sequence was absent in the plasmid obtained through the assembly of three fragments using Golden Gate Assembly. Therefore, we improved the N20 insertion method in the second round of experiments. We synthesized two complementary primers and annealed them to form a double-strand N20 with sticky ends, and then inserted it into the vector. In the end, we successfully constructed the level 1 plasmid.

In the first round of level 2 construction, we attempted to assemble these fragments using Golden Gate Assembly. Regrettably, we were unable to obtain the recombined plasmids, possibly due to the shared antibiotic resistance between the vector plasmid and the donor plasmid.

In the second round of experiments, we adopted enzymatic digestion to separately cleave the inserted fragments and the vector plasmids. After purifying the fragments through gel extraction, we performed ligation. However, the concentration of the gel-purified target fragments was relatively low. Sequencing results indicated that successful fragment connection had not yet been achieved.

Learn

• In light of plasmid incompatibility and metabolic stress, the recording capacity is significantly constrained by the number of plasmids. To attain a higher degree of recording, it is imperative to curtail both sequence length and the quantity of plasmids employed.
• The vector should undergo optimization for greater compatibility.

Cycle 2: Attempt to the cassette recorder

Design-CRISPReporter over plasmids in-cis

Inspired by the work of Choi et al. (Fig 7, Choi, J., Chen, W., Minkina, A. et al., Nature 608, 98-107 2022), we attempt to set CRISPReporter as a cassette recorder in cis. We attempted to assemble all the components into one sequence. In the frame of Cas9, we found that D. Perli et al.'s and Zhao, D.'s methods can respectively achieve self-targeting and self-recombination, as we have mentioned in the design.

To validate the proof-of-concept, we made two sets of in-cis designs considering the self-recombination efficiency. In the first set, each cassette has a unique barcode to ensure proper recombination (Fig 8a). In the second set, the cassettes share a common segment of promoter, eliminating the need for barcode accumulation and resulting in a more compacted and extendable device (Fig 8b,c). As the previous cassettes block the expression of later ones, the cassettes are expected to operate sequentially (Fig 8d). As for experimental design, we intend to test the single cassette ones, to get initial data on the mechanism.

Build

We synthesized the required DNA fragments in collaboration with Atantares and Tsingke, two reliable biotech companies. Then we successfully co-transformed our working plasmid with pCas, the vector for Cas9 and Lambda Red, into Escherichia coli DH5α (Fig 9).

Test for the "stgRNA-barcode-cassette 1"

After amplification culture, the bacteria solution was diluted into 3ml of LB media with appropriate antibiotics and cultivated to reach an OD600 of 0.3~0.5. We added a variable dosage of arabinose and IPTG close to the working concentration. After cultivating 22hrs for arabinose and 5hrs for IPTG induction, we diluted the media and spread plate to halt induction and get single colonies (Tbl 1). Then, the relevant sequence was amplified and sequenced (Fig 10a), confirming a successful knock-out. Large-scale screening revealed that group 2 exhibited the highest efficiency, with efficiencies of 70%, 10%, and 38.1% for groups 2, 3, and 4, respectively (Fig 10b). Non-induction controls substantiated that induction is the prerequisite for recording (Fig 10c).

Learn from the stgRNA-cassette-barcode 1

The results demonstrated that the cassette, our building block of the recorder, can function as expected under induction. Furthermore, a relatively suitable induction condition for proof-of-concept is 2g/L for Arabinose and 3g/L for IPTG, for 22hrs and 5hrs, respectively. However, the condition data should be measured in detail for further multi-level induction, and the knock-out events should be supported by quantifiable data.

Test for the "stgRNA-cassette 1"

Using the same procedure but a different induction condition (Tbl 2), we also tested the function of a single cassette. To our disappointment, the target sequences remained all intact (Fig 11).

Learn from the stgRNA-cassette 1

Comparing the two cassette designs, the homologous arm was set downstream of each cassette in "stgRNA-cassette-barcode" ones, while the upstream of the next cassette in "stgRNA-cassette" ones (Fig 4a, b). As a result, the single cassette we have tested has no inherent homologous arm. However, the integration of another cassette can solve the problem. Based on the insights gained from previous experiments, we designed to insert an additional cassette downstream of the two devices, and then apply the appropriate induction condition to the stgRNA-cassette (1+2+3) (BBa_4630114) (Fig 4c) and stgRNA-cassette-barcode (1+2) (BBa_4630102).

Cycle 3: Integrated device test

Design

In line with the previous experiment, we'd like to first test the inherent capability of the recorder without barcode, and then test the cascade knock-out event of both. We set special restriction sites flanking the cassette to make it extendable. Using the principle of BioBrick Assembly (3A Assembly) , we introduced an XbaI site upstream of the segment to be inserted, and SpeI & EcoRI downstream of each cassette. By cutting the backbone with SpeI and EcoRI and cutting the segment with XbaI and EcoRI, the new cassette could be inserted downstream of the previous one(s). Besides, the remaining restriction sites make the system fully extendable to add additional cassettes.

Build

The subcloning of the recorder without a barcode, stgRNA-cassette (1+2), stgRNA-cassette (1+2+3) was executed by Atantares. Meanwhile, we carried out the subcloning of the stgRNA-barcode-cassette (0+2) and stgRNA-barcode-cassette (1+2) by ourselves (Fig 12). The stgRNA-barcode-cassette (0+2) is the ligated product of the knock-out product of stgRNA-barcode-cassette 1 and the stgRNA-barcode-cassette 2, which can be used in verification of the second editing.

Test-Attempt

We organized our experiment into groups. Employing the same induction condition (2g/L L-Arabinose and 3g/L IPTG, for 22hrs and 5hrs respectively) with the same tester, stgRNA-cassette (1+2), stgRNA-cassette (1+2+3) have been successfully knocked out while stgRNA-cassette 1 was intact (Fig 13). It is noted that all the parts remain one last cassette intact, indicating the significant role played by the inherent homologous arm in the recombination, and confirming the occurrence of multi-level knockout. Also, the induction carried out on stgRNA-cassette-barcode (1+2) got the expected outcome (Fig 14).

Learn

From the results, we conclude with greater confidence that the incapability of stgRNA-cassette 1 is due to the lack of a homologous arm. Also, all the other multi-level ones were knocked out to their fullest, showing no tiered recording results. We assume that the induction time is so long that permits the cassettes to be knocked out twice. As a result, we intend to make a comprehensive measurement of the building block and to carry out a time-dependency experiment of the stgRNA-cassette (1+2+3).

Test-Large scale

The solution to the double knock-out issue lies in better control of the induction conditions. That is, to gain more information about the induction strength, and determine the refractory phase of the part. Maintaining other variables at typical levels (3g/L L-Arabinose, 2g/L IPTG, inducing 22hrs and 5hrs respectively), we carried out a concentration matrix test (Tbl 3) and two time-gradient tests.

Concentration Matrix

It's noted that the induction of arabinose inhibited the growth of bacteria strikingly (Fig 15), and the Lac promoter exhibited significant leakage expression (Tbl 5, Fig 16b). Despite the missing data due to failure of sequencing, E6, E5, C2, B5 performed better (Tbl 4, Fig 16a, Tbl 6). However, a quality test based on electrophoresis provided parallel data for randomly picked groups, and the two data access showed a significant correlation, with paired t-test P = 0.7602, no significant difference (Tbl 7, Tbl 8, Fig 17). Given the substantially larger amount of data from the electrophoresis test for E6 (N = 20) compared to sequencing (N = 4), we adjusted the knock-out ratio of E6 to 60% (fig S2).
Surprisingly, we notice that at a very low rate, recombination might happen between the two plasmid-inherent Lac promoter sequences, instead of the set barcode sequence. In the first matrix, the overall unexpected knock-out ratio is 2.76%, significantly lower than the normal rate of 42.5% (Tbl 9, Fig 16b).
Further, to ensure data reliability, we tried the matrix test again. Growth inhibition and leakage expression still occur (Fig 18 a, b). Also, an integrated heatmap is plotted based on the mean value of the two matrices (Fig 18 c), and the variation pattern of the two matrices showed some kind of correlation. Paired t test result of the two matrices showed a significant difference, with P = 0.0054(**) and mean of differences = -0.2071(Fig 18 d). Two-way ANOVA of the second matrix and the average matrix indicate that the IPTG concentration is the main variation factor (Tbl 10).

Learn from the concentration matrix

From the data we conclude that the preferred induction concentrations are 8g/L L-Arabinose and 3g/L IPTG. The unexpected recombination proves the concept that DSB is the prerequisite of the recombination, and the recombination is preferred to happen between two near homologous arms. The data from electrophoresis and sequencing can compensate for each other.
The two parallel concentration matrix tests uncover that there is a huge variation in the measurement, and more experimental data are needed. Moreover, the variation pattern of the samples showed similarity, indicating probability of logical regression. And, the statistical analysis proves that IPTG concentration is the leading factor controlling the result, fulfilling part of our proof-of-concept.
However, the leakage of pLac makes it difficult to measure the multi-level recording. We'd like to change it to other promoters- check the next cycle, the EL222 part, and the application part.

First time-gradient test

We did the two tests subsequently. The first test of knock-out showed a bell-type-like shape (Tbl 11, Fig 19a). The increasing trend ends on dot 4h and the extension of IPTG induction time does not help the editing afterwards. Consequently, we set up another time-gradient within 2 hours on stgRNA-cassette (1+2+3), and all of them achieve full knock-out, at a ratio of 100% (Tbl 12, Fig 19b).

Learn from the time-gradient tests

In line with our PI's advice, long-term induction of IPTG proved harmful to the cells. However, the time-gradient within 2 hours did not really set up because of inconsistency with the first time-gradient and the small data amount of both. Not surprisingly, the test carried out on stgRNA-cassette (1+2+3) showed some kind of leakage expression. It's necessary for us to substitute the Lac promoter with a more reliable one.

Cycle 4: Promoter Substitution

Design

We designed many promoter substitution plans (See also Applications). Here, we demonstrate the substitution based on the plasmid backbone. There is a T7 promoter upstream of the constructed cassettes, so by simply omitting the Lac promoter and relevant sequences, we can let the T7 promoter take control (Fig 20).

Build

We successfully amplified the sequence using stgRNA-cassette (1+2+3) and underwent Golden Gate Assembly. However, due to the length preference in PCR, we only got one cassette (Fig 21). And due to time constraints, we haven't finished the co-transformation and further induction test.

EL222 directed evolution

Cycle1: Characterization of EL222 with eGFP Reporter

Design

In the initial phase, our objective was to evaluate the function of EL222 as a blue-light-inducible promoter. To assess this, the eGFP gene was employed as a reporter due to its widespread use as a fluorescent marker. In the design, pEL222 and EL222 were inserted upstream of the eGFP gene (Fig 1). The expression of EL222 was regulated by a constitutive promoter.

Build

We synthesized and performed subcloning of EL222 and pEL222 utilizing restriction cloning into pET-28a-[eGFP]. Then the plasmid was transformed into E. coli DH5α for amplification and BL21(DE3) for testing.
Colony PCR was performed to make sure the plasmid was constructed just like we designed (Fig 2).

Test

Bacteria were cultured, and exposed to 400 Lux blue light around 475nm for 20 hours (Fig 3). Then the dishes were observed under UV light to confirm the expression of eGFP proteins. Regrettably, no discernible green fluorescence was observed

Learn

A comparative analysis revealed that the excitation wavelength of eGFP is approximately 488nm, closely aligned with the blue light employed in our study. The absence of observable green fluorescence was attributed to a quenching effect induced by the blue light.

Cycle2: Characterization of EL222 with mRFP Reporter

Design

In the second iteration, we replaced eGFP with mRFP1 on the pET-28a-[EL222-eGFP] plasmid (Fig 4) to mitigate the fluorescence quenching effect.

Build

mRFP1 sequence was obtained via PCR and subsequently cloned into the pET-28a-[EL222-eGFP] vector utilizing Golden Gate assembly, to replace the eGFP sequence. Then the plasmid was transformed into E. coli DH5α for amplification and BL21(DE3) for testing.
Sequencing was performed to make sure the plasmid was constructed just like we designed (Fig 5).

Test

Bacteria was spread on plates and first cultured under 37℃ for 10 hours, then exposed to 400 Lux blue light around 475nm for 20 hours. We then performed observation on dishes and slides we made from colonies under UV light, to confirm the expression of mRFP1 proteins. There was no visible red fluorescence observed from our dishes, while it can be observed from the slides we made (Fig 6). However, there was no difference between the samples before and after exposure to blue light (Fig 6a,b).

Learn

Examination of the EL222 protein sequence unveiled a gap of 16 amino acids at the N-terminus of EL222, along with a missing initiation codon (Fig 7). This gap was found to be the cause of the mRFP1 leakage.

Cycle3: Modification of the EL222

Design

Upon identifying the differences in EL222 proteins, we opted to modify the EL222 sequence through PCR by adding the missing sequence (Fig 8a). Furthermore, a newly synthesized EL222 sequence was introduced, and a T7 terminator was placed downstream of EL222 (Fig 8b).

Build

The new EL222 sequence and the modified one are obtained via PCR (Fig 9) and subsequently cloned into the pET-28a-[EL222-eGFP] vector utilizing Golden Gate Assembly to replace the former eGFP sequence. Then the plasmid was transformed into E. coli DH5α for amplification and BL21(DE3) for testing.
Sequencing was performed to make sure the plasmid was constructed just like we designed (Fig 9).

Test

We conducted the induction experiment and observation with the same protocol as in cycle 2. There is no significant difference before and after exposure of the EL222-modified transformant, while the pET-28a- [EL222 new-mRFP1] transformant exhibits stronger red fluorescence compared to the negative control (Fig 10).

Learn

EL222-modified cannot work either, but EL222-new can work as expected, which may be attributed to the codon optimization of EL222-new. However, the leaky expression of pEL222 is still considerable.

Cycle4: Directed evolution of EL222

Design

We decided to first perform directed evolution on EL222 to reduce leakage and increase downstream expression. All the new sequences of pEL222 will be tested on the former plasmid by measuring the fluorescence intensity of the colony before and after induction.

Build

We acquired different sequences of pEL222 by error-prone PCR. Then, the generated sequences were incorporated into pET-28a-[EL222-mRFP1] by Golden Gate Assembly, replacing the original sequence of pEL222. The plasmids were transformed in BL21(DE3) directly.

Test

Firstly, we sequenced the PCR products (Fig 12) to confirm the success of error-prone PCR. Then we use colony PCR to confirm the changes in pEL222 cloned in pET-28a plasmids. To measure the expression of mRFP1, we cultured all the bacteria in PCR tubes and measured their absorbance at 580nm and 600nm, which present for mRFP1 and bacteria concentration respectively.
The OD580/OD600 at a specific time for a sample culture was determined after subtracting from each of the technical triplicate readings of the negative control cultures (fluorescence free) at the same time. The fluorescence synthesis rate of any sample at time t, was calculated by taking the difference of Fluo/OD600 values from two time points and dividing the result by the time interval.

Colonies with different pEL222 show differencet expression of mRFP1, resulting in disparate colors. IPTG is added to ensure the expression of EL222 protein.

Learn

The sequence of pEL222 is related to mRFP1 expression. Using different pEL222, we can change the rate of expression.

sgRNA screening

Cycle1: Blue-White Plaque Screening

Design

In the initial phase of our study, we aimed to employ the blue-white plaque screening technique to selectively identify sgRNA sequences exhibiting high editing efficiency. To accomplish this, we developed a screening plasmid designated as pUC57-N20s-gRNA+HA (Fig 1a).
This plasmid encompasses a sgRNA scaffold, and a sgRNA targeting sequence (named N20s), which consists of 50 individual N20 sequences concatenated, each accompanied by its respective PAM sequence. This N20s sequence is inserted into the LacZ gene, unbaling to synthesize the functional β -galactosidase. The scaffold can load a specific N20 sequence (20nts specificity-determining sequence) by Golden Gate Assembly and conduct targeting fragment knockout (Fig 1b), thereby driving correct LacZ expression when induced. Functional β- Galactosidase can decompose X-gal to produce blue products, making colonies appear blue. While the colonies that didn't correctly knock out the N20s sequence were white.
We randomly loaded 50 different N20 sequences onto the plasmid and identified efficient ones by sequencing blue colonies. When the colony count is sufficiently high, our method will examine all N20 sequences with near certainty.

Build

The plasmid synthesis was executed in collaboration with Atantares. We also synthesized various N20 sequences via primer annealing (Fig 2a) and subsequently inserted each of them into the plasmid employing Golden Gate Assembly (Fig 2b). The assembled products are first transformed into E. coli DH5α for amplification, then co-transformed with pCas (pRed_cas9_△poxb300) (Fig 2c). Of note, these plasmids with different N20 sequences are co-transformed individually but not together in one tube, as we want to first determine the feasibility of the system.

Test

Following a 24-hour induction with 2g/L arabinose, we determined the edited colonies via colony PCR. Our analyses revealed successful N20s sequence deletions for some of the tested N20 sequences, exemplified by NO. 11, 13, 14, and 15 (Fig 3), demonstrating the system's feasibility.

Sequence alignment inspections corroborated the integrity of the LacZ coding sequence in N20s-knocked-out plasmids, devoid of any mutations. However, despite employing various IPTG induction methods, these successfully edited colonies did not exhibit a blue phenotype in the substrate-supplemented medium.

Learn

• Golden Gate Assembly succeeds easily when the experiment is conducted with the correct protocol.
• Knockout of fragments on plasmids is feasible.
• After consultation with professors, we understood the intricate nature of the blue-white plaque screening technique's challenges. Consequently, we received advice to consider alternative markers for N20s knockout assessments.

Cycle2: Green fluorescence Screening

Design

In light of the failure of blue-white plaque screening, we adopted EGFP as the reporter of successful editing in the second DBTL cycle. EGFP is also a kind of widely used fluorescent protein in E. coli. We positioned it under the control of a constitutive promoter to save from the trouble of IPTG induction. We designed to insert two fragments into the pUC57 vector, HA+ gRNA & EGFP-N20s, to construct the plasmid pUC57-EGFP-N20s-HA+gRNA (Fig 4). If the N20s sequence is knocked out, EGFP would be expressed properly, resulting in colonies manifesting a green phenotype.

Build

We employed Gibson assembly to insert HA+ gRNA (Fig 5) into the vector and Restriction cloning to insert EGFP-N20s (Fig 6). The construction of pUC57-EGFP-N20s-HA+gRNA went successfully as we designed.

After successful construction, we loaded those N20 sequences described in the first cycle onto the plasmid in one tube and performed co-transformation with pCas into E. coli.

Test

We have run through different culture procedures to identify the optimal screening protocol (Chart 1). However, none of them achieved N20s knockout in any colony. After an extensive investigation, it was found that the preserved plasmid had undergone fragment loss. We then asked Genescript for synthesis and subcloning of the plasmid. Unfortunately, when we received it, there was no time for further testing.

Learn

We surmise that the concentration of Golden Gate products is too low. To solve this problem, we cultured E. coli in liquid LB media to expand the clone population. However, it led to false positive results: colony PCR indicates that plasmids harbored by the bacteria were origin vectors rather than Golden Gate products. Consequently, the batch random experimentation method remains to be optimized.

pCas optimization

Cycle1: Characterization of pRed_cas9_△poxb300

Design

As the material we got, pCas (pRed-Cas9-△poxb300) in Zhao, et al. 's work (Fig 1a, Zhao et al. Microb Cell Fact 2016, 15:205) is capable of conducting genome deletion and will be cured when cultivated at 37°C or higher. We first tested its genome editing efficiency. When induced by arabinose, the plasmid can express Cas9, Lambda-Red, and a poxb-targeting sgRNA, causing a 513bp deletion of poxb in the genome.

Build

We chose Escherichia coli DH5α as the cloning host and MG1655 to conduct editing. After transformation, we performed induction following the method offered by Zhao et al..

Test

We used 2g/L arabinose to induce editing. After induction, the medium was diluted 2000-fold and spread on the plate. Then, using colony PCR, we identified the editing efficiency of different induction durations by calculating the ratio of the shortened PCR fragments (Fig 1b).

We first found that 6h induction is not sufficient for 100% editing. Therefore, we conducted a gradient experiment on the arabinose induction time to figure out the optimal duration of induced editing. As it is said that the induction should last at least 6h, gradient induction was set up from 6h to 30h. Although double bents always existed (Fig 1c), which means that bacteria are not completely edited in this colony, 24h was considered as the optimal induction time with minimal double bents and almost 100% editing efficiency (Fig 1d).

Learn

• In the experiment, 24h induction can lead to optimal editing efficiency.
• E. coli grows slower after being transformed with pCas
  
  • The culture temperature is limited to 30 °C.
  • Expression of Cas9 and Lambda-Red increases the metabolic burden of bacteria.
  • DSB in the Genome may cause death to bacteria.

Considering these limitations of pCas, we intend to optimize it from the following aspects:

• Removal of the genome-targeting guide RNA and the homologous arm for poxb
• Removal of the temperature-sensitive site and the addition of restriction sites for Golden Gate assembly.

In this case, the pCasop (short for pCas_optimized) plasmid suits our project better.

Cycle2: Typical Site-directed, Ligase-Independent Mutagenesis (SLIM)

Design

Initially, we obtained a comprehensive set of experimental procedures from the literature. The method utilizes two long primers with overhang and two short primers in a single reaction for reverse PCR amplification of the template, meanwhile, the 5' end of each primer incorporates the desired sequences on the complementary strand at the end of the PCR product. Then, heteroduplexes are formed between mixed PCR products, resulting in the desired plasmids (Fig 2).

Correspondingly, we designed 4 primers to replace the original specificity-determined sequence (SDS, or N20 in the literature) with BsmBI restriction site. Then the original pCas template was digested using DpnI enzyme, to prevent false positives.

Build

Long fragments were successfully cloned using PCR. After denaturation and renaturation, the DNA product was obtained. Then it was transformed into E. coli DH5α and spread on Kanamycin plates. Positive transformants were verified and selected. To our disappointment, it turned out to be the original pCas, concluding from sequencing results (Fig 3).

Learn

We assumed that the original pCas had not been fully digested by DpnI, and pCas has better capability to be transformed, as it was less modified. Also, the PCR product has a similar length to the original pCas, making it hard to purify pCasop segments using a DpnI-independent way. As a result, we intend to amplify shorter and more fragments for better purification, and then use Gibson Assembly to get the final plasmid.

Cycle3: Three-segment assembly

Design

The whole plasmid was separated into three parts (Fig 4) and each part was amplified using well-designed primers. In addition, the Gibson primers were designed to clone fragments using pCas as the template, remove the poxb homologous arm sequence, and mutate the original 30°C sensitive sequence of pCas to enable rapid E. coli growth at 37°C.

Build

The designed three fragments were successfully PCR amplified using Gibson primers (Fig 5). These fragments were then ligated using Gibson assembly. But we found it hard to get positive transformants of E. coli DH5α.

Learn

Judging from the results and advice from our PIs, we realized that simply ligating two long fragments and one short fragment using Gibson Assembly was of low efficiency. Also, we are not so familiar with the technical details of Gibson Assembly. Consequently, we intend to jump to another path.

Cycle4: Simple removal of sgRNA

Design

As the length of the segments varies, we decided to remove redundant parts step by step, especially the functional sgRNA. We designed primers for the backbone to delete the sgRNA sequence using Gibson Assembly and Golden Gate Assembly (Fig 6).

Build

We successfully amplified the target sequence of pCas (Fig 7a). Subsequently, we employed Golden Gate Assembly. Then sequencing result unveiled that the sgRNA part of pCas had been removed (Fig 7c).

Test

Set the original pCas as control, we transformed the pCasop into E. coli DH5α and induced (Fig 8a-d). The bacteria amount of pCasop-containing group is much larger than pCas group, indicating an increase in bacteria viability (Fig 8e). The sequencing results prove that there is no targeting deletion in the pCasop group (Fig 8f).

Learn

The data demonstrated that we have got a capable pCasop, without knocking out the genome. Also, the prefix of this set of Golden Gate Assembly primers is quite useful, to omit a segment in a plasmid using single fragment self-assembly is feasible.

Cycle5: Integrated recording machine

Design

Though we get a useful pCasop, we intend to make the entire device more compact and better contribute to synthetic biology. In detail, we design Golden Gate Assembly primers for ligation of the Cas9 coding sequence and the Lambda Red, araC sequence (Fig 9a). Considering the multi-cloning site on the pCDF-Duet1 backbone, we attempt to add XhoI and NotI prefixes on the ligated Cas-Lambda sequence (Fig 9b). Then the segment can be ligated to the plasmid by restriction cloning. We re-organize them in one integrated plasmid (Fig 9c).

Build

We successfully amplified the target sequence and finished the Golden Gate assembly (Fig 10). Unfortunately, due to the limitation of experimental materials, we can only get the two enzymes from different companies, causing the buffer incompatibility. We found it hard to ligate the segment and backbone by restriction cloning.

Learn

From the plasmid construction, we confirm that the set of Golden Gate Assembly primers is quite versatile and the risk of restriction buffer incompatibility should be noticed.

Design-Three-segment Golden Gate Assembly

In this turn, we decided to amplify the two segments and the backbone using Golden Gate Assembly primers (see also Expeiments, primer list) (Fig 11).

Build

The three segments were successfully amplified, and we performed Golden Gate Assembly (Fig 12a-c). The sequencing results suggested that we have got the expected product (Fig 12d).

Test

Due to the limitation of time, we have not finished the induction of the integrated plasmid.

Supplementary data for cascade system