“Results! Why, man, I have gotten a lot of results. I know several thousand things that won't work.” - Thomas A. Edison


In our project we had five main aspects:

  1. Compete
  2. Attack
  3. Protect
  4. Integrate
  5. New chassis development

The highlights of our work relating to Lactobacillus crispatus are summarized at the new chassis development page.


As mentioned in our compete page, we wanted to allow L. crispatus to secrete a truncated FimH in order to compete with UPEC, and we wanted to test our system in Bacillus subtilis as a model organism of L. crispatus.

We had initially restricted and ligated our insert into the pBE-S plasmid and transformed it into E. coli.

colony PCR FimH-GFP

Figure 1: Colony PCR after transformation of FimH-GFP insert. Expected size of 1172bp was achieved in the marked lanes (colonies 6,7,8 and 10). The forward primer was designed to anneal within the inserted FimH sequence, therefore negative colonies should show no amplification.

As mentioned on our engineering success page, we had several failed transformation attempts into B. subtilis, before identifying the issue.

Eventually, we were able to perform a successful transformation of this genetic circuit into B. subtilis, however we did not have enough time to assess the functionality of the FimH expressed by B. subtilis, so we ceased working on this aspect and focused our attention on the other plans we had.

Naturally, this is a test we would like to do in the future by utilizing cell-lines that present the same receptors as those found in the urinary tract such as HEK293.


As mentioned in our protect page, we wanted to ensure user safety and develop a killswitch system. We had successful transformations into E. coli:

colony PCR E. coli parR paraE

Figure 2: Colony PCR results for transformations of the killswitch promoters- paraE and paraR (both with mCherry downstream). There was a mistake and a ladder wasn't loaded but it is estimated that the amplification is of the expected size (414bp for paraR and 478bp for paraE). Again, primers were designed that one anneals inside the insert, so with no insert there should be no amplification.

Fluorescence testing

As mentioned on our engineering success page, we found that paraR facilitated mCherry expression in E. coli, despite this promoter coming from B. subtilis, and so we proceeded with testing in E. coli as well.

Initially we inoculated a starter from the positive colonies, introduced L-arabinose and measured the fluorescence after two hours, based on information from the literature and iGEM registry [1], [2].

We found that in the morning, some of the paraR starters had already turned pink, despite no arabinose induction, while all the paraE starters remained yellow.

After the two-hour incubation, some of the paraE starters, that weren’t induced, had also turned pink indicating that the accumulation of mCherry is somewhat time dependent regardless of induction.

We proceeded with the measurement anyway and found that paraE had lower expression levels in the presence of L-arabinose, while paraR had no change.

Figure 3: Normalized fluorescence level of mCherry expressed in E. coli under the control of paraE, after two hours with and without L-arabinose induction.

Figure 4: Normalized fluorescence level of mCherry expressed in E. coli under the control of paraR, after two hours with and without L-arabinose induction.

The results for paraR were not surprising in the technical sense: the starters were already visibly pink in the morning, meaning there was high expression and the inducer had little room to make an effect. They were also not surprising because paraR was suspected to be leaky, as elaborated in the protect page (and as mentioned, was already observed to be leaky when the E. coli colonies turned pink).

The results we received for paraE were contrary to those expected of B. subtilis, however, they were based on three biological repeats, each with four technical repeats, indicating the validity of the results. The explanation we found, as elaborated in the registry for paraE (BBa_K4633006), is host incompatibility. In short, E. coli has a native L-arabinose utilization system with a repressor protein araC [3]. AraC is self-regulated, meaning that in the presence of arabinose there are higher levels of araC in the bacterial cell [4]. It is possible that since araC and araR have different binding sequences, that araC doesn’t detach well from the araR binding sequence present in paraE (and paraR). This would lead to greater quantities of a repressor which cannot detach easily, resulting in a decrease in fluorescence [4]–[6].

We had simultaneously worked on characterizing these same parts in B. subtilis.

We couldn’t draw conclusions from the experiment since too many of the treatments failed: the paraR plasmid used for transformation in B. subtilis appeared to be dysfunctional despite being sequenced and there were no mismatches between it and the intended sequence. It is known that Sanger sequencing makes mistakes from time to time, it is possible that a critical base was sequenced and shown as expected despite it being a mistake and was mutated.

The experiment was repeated, along with testing for additional conditions. Despite the fact that we couldn’t draw conclusions, we noticed that the fluorescence levels in B. subtilis were significantly lower than those in E. coli, by one-two orders of magnitude. We decided to test the expression in B. subtilis in the recommended arabinose dose (0.4% w/v) based on literature as well as double the concentration (0.8% w/v) [2]. We also decided to test at the established time from the literature (two hours) and twice as long (four hours) [2].

Again, too many of our treatment failed to draw statistically significant conclusions and no clear trend could be seen for paraE:

Figure 5: Normalized fluorescence level of mCherry expressed in B. subtilis under the control of paraE, after two hours and four hours, with and without L-arabinose induction in concentrations of 0.4% (w/v) and 0.8% (w/v).

We understood that our approach was problematic, and we decided to change the testing protocol. Previously, a starter culture would be inoculated and after an undefined amount of time, the induction with arabinose would commence. The updated protocol included incoulation of the starter and immediate induction with arabinose and fluorescence measurement after 24 hours. The test following 24 hours was recommended to us by our advisors, since mCherry takes 90 minutes to fold, so the previous waiting period might not have been sufficient to observe changes in expression levels [7], [8].

We then executed the experiment with the additional incubation time, both for E. coli and B. subtilis.

For E. coli we saw a downward trend in paraE:

Figure 6: Normalized fluorescence level of mCherry expressed in E. coli under the control of paraE, after 24 hours with and without L-arabinose induction.

And for paraR it is difficult to determine if there is a downward trend or no change due to the large error bars:

Figure 7: Normalized fluorescence level of mCherry expressed in E. coli under the control of paraR, after 24 hours with and without L-arabinose induction.

These results align with the results obtained using the previous protocol and the explanation for these results is as mentioned before, most likely host incompatibility and a leaky promoter.

For B. subtilis we saw the expected upwards trend for paraE:

Figure 8: Normalized fluorescence level of mCherry expressed in B. subtilis under the control of paraE, after 24 hours with and without L-arabinose induction.

In paraR, we had two treatments that indicated an upwards trend and two treatments that indicated a downward trend, so we couldn’t draw any conclusions.

This experiment would need to be repeated in the future in order to draw clear conclusions and to properly characterize paraR in B. subtilis, which we were not able to do unfortunately due to time constraints.

Microscopic imaging

We had also imaged E. coli under a fluorescent microscope. As mentioned in the registry, this experiment was cut short due to the war breaking out in Israel, and we only managed to image a few of our planned treatment groups (most notable in its absence is B. subtilis).

The planned treatment included WT bacteria, bacteria with the basic pBE-S plasmid, bacteria with paraR and bacteria with paraE. All in E. coli and B. subtilis, all with and without L-arabinose 24 hours' induction.

Each treatment was imaged twice, at 550nm for mCherry, and with CID (whitelight):

Figure 9: E. coli under fluorescent microscope. From left to right: overlay of 550nm reading and CID, reading at 550nm only, CID only. Top half is paraE starter induced with arabinose, bottom half is WT starter induced with arabinose. Each starter was imaged twice.

As can be seen for paraE with L-arabinose induction, despite the low fluorescence and that the starter not visibly becoming pink, some of the bacteria did produce mCherry. We believe that images of B. subtilis would have shown a similar trend.

Figure 10: E. coli under fluorescent microscope. From left to right: overlay of 550nm reading and CID, reading at 550nm only, CID only. From top to bottom: paraR starter without arabinose, paraR starter induced with arabinose, WT starter induced with arabinose. Each starter was imaged twice.

As can be seen for paraR, the results are similar to those seen in the plate-reader test: both treatments have high levels of mCherry production, but the treatment without arabinose has higher levels. This supports the conclusion that the trend we saw is based in biology and not merely a statistical error.


The genetic components, for unknown reasons, were stuck in customs (this didn’t happen with any of the other material we ordered) and arrived very shortly before the wiki freeze so we unfortunately, were unable to perform even the initial cloning stages.

Because this is an important part of our multi-pronged approach in tackling UTIs we hope to be able to continue this work, even if it will only be after the grand jamboree.


As explained in the integrate page, we wanted to ensure that our solution would be long lasting and in order to achieve that aimed to integrate our system into L. crispatus’ genome. Our goal was to use L. crispatus’ native CRISPR system- Cas3. To this end, we needed a target gene, that we would damage during the integration, chosen to be the tetracycline resistance gene naturally found in the genome of our strain, and a donor DNA, containing the gene we wished to integrate and homological areas for the area flanking the insertion point, to allow the integration.

InFusion method

The sequence of the donor DNA we had wished to use was too long to be produced as one fragment. To solve this issue, we opted to use the InFusion method from the commercial kit by Takara, which functions based on similar principles to the Gibson assembly method.

We performed the transformation of these parts several times because each time the transformation seemed to fail when incubating our transformed E. coli overnight, as per our protocol. With typical restriction cloning we would see colonies on the plate in the morning, however this did not occur with the bacteria that received the InFusion products. Colonies appeared only after about 24 hours and they were very small and took several more hours to grow to a suitable size. In one case a wait time of about 48 hours was required.

Figure 11: E. coli colonies after transformation with InFusion products. The image was taken about 24 hours after transformation. As can be seen, there are very few colonies, and they are very small.

Figure 12: E. coli colonies after transformation with InFusion products. The image was taken about 48 hours after transformation. As can be seen, there are very few colonies, and they are very small.

We tested our colonies by colony PCR process several times. The primers were designed to anneal outside the insert, meaning that distinguishing between self-ligation and desired products was based on the fragment size that appeared in the gel. No amplification was observed:

Figure 13: Colony PCR of E. coli from the first transformation. While no amplification can be observed, in one lane marked in red there is a weak signal outside the range of the ladder meaning it is larger than 10kb.

Figure 14: Colony PCR of E. coli from the second transformation. While no amplification can be observed, in a few lanes there is a weak signal marked in red outside the range of the ladder meaning it is larger than 10kb.

After consulting again with the advisor that recommended the kit, we learned that InFusion products often don’t respond well to colony PCR and that these results are typical.

We opted to extract the plasmid from a few colonies, including the ones that presented a weak signal in the gel, and test the size of the extracted plasmid to confirm the success of the process. We observed that even in liquid medium bacteria took longer to grow than usual.

Plasmids were extracted and restricted for testing, and gel electrophoresis was planned for the next day however we were unable to perform it due to the war breaking out in our country. We are eager to learn if the process succeeded or if different transformation methods, such as Gibson assembly, need to be considered. We hope to be able to complete this experiment in the future.

MIC test

Minimal inhibitory concentration (MIC) experiments were conducted to support the integration of DNA into the genome of L. crispatus ATCC33280. To achieve this, it was imperative to pinpoint a specific target gene within the bacterium's genomic structure. After an exhaustive review of relevant literature, a gene endowing resistance to tetracycline was identified within this bacterial strain, you can read more about it in our integrate page. To ascertain the presence of this resistance gene and assess the extent of tetracycline resistance exhibited by L. crispatus ATCC33280, a meticulous MIC test was executed. All bacterial cultures were grown in liquid MRS medium, containing various tetracycline concentrations. Notably, as we found that the bacterium can thrive without the need for an anaerobic system, this experiment was conducted under aerobic conditions.

In the context of antibiotic resistance, a bacterium is generally classified as resistant when it can withstand exposure to antibiotic concentrations exceeding a specific threshold. According to the literature, tetracycline resistance within L. crispatus was found to be discernible at a concentration of 6 (µg/ml).

The results obtained, as presented, exhibit an absence of a consistent pattern. Contrary to expectations, the data does not distinctly delineate the bacteria's response to increasing antibiotic concentration over the same time frame. Nevertheless, it is unequivocally apparent that bacterial growth is observed up to a tetracycline concentration of 64 µg/ml, beyond which a significant decrease in growth becomes pronounced. This intriguing observation underscores the bacterium's robust resistance to tetracycline, up to the 64 µg/ml concentration threshold within the established time frame, and expression of the tetracycline resistance gene that we targeted for genomic integration. As we can see in the graph:

Figure 15: Time-Dependent MIC Variations in Tetracycline for L. crispatus ATCC 33280. This graph presents the Minimal Inhibitory Concentration (MIC) of tetracycline over time (24, 48, 72, 96, 120 hours) and various concentrations (4 µg/ml to 256 µg/ml) for L. crispatus ATCC 33280.

As can be seen in figure 15, for all antibiotics concentrations, there was initial minimal growth, and in the lowest concentrations (4,6,8 and 10 µg/ml) there was extensive growth within 24 hours up to an OD of around 0.7, which is considered high.

It can also be observed that at the highest concentrations which showed growth (32 and 64 µg/ml), not only is there an upwards trend in the OD value at the end of the testing period, but also that in both of these treatments the bacteria have exceeded an OD value of 0.3, which according to Professor Per Saris indicates that L. crispatus is in its exponential growth phase [9].

Figure 16: L. crispatus exhibiting growth after 24 hours from inoculation in tetracycline concentrations of 4, 6, 8 and 10 µg/ml, from left to right. The bacterial growth is identified as a white sediment at the bottom of the tube.

Figure 17: L. crispatus after 48 hours in tetracycline concentrations of 128 and 256 µg/ml, from left to right. According to figure 15, at this time was the highest OD for these treatments, yet no bacterial growth can be visually identified.

Figure 18: L. crispatus after 120 hours in tetracycline concentrations of 32 and 64 µg/ml, from left to right. The bacterial growth is identified as a white sediment at the bottom of the tube.

New chassis development

Numerous endeavors were made to achieve successful transformation of the L. crispatus bacterium, yet lamentably, none of these endeavors yielded favorable results. Initially, transformations were conducted following a prescribed protocol outlined in a published research article [9]. These transformations involved the use of various plasmids, closed pTRKH2 (4963 bp), closed pMG36E (3610 bp), and pMG36E with inserts. We had two inserts: amyA+linker+GFP (861 bp) and amyL+linker+GFP (861 bp). AmyA and amyL are two secretion peptides we wished to test, as can be seen in our parts page. For the sake of control, transformations were also performed using ultrapure water (UPW) in lieu of DNA, and the resulting transformants were cultured on MRS agar plates with and without antibiotics to assess the efficiency of the transformation process. It is worth noting that in all transformation experiments, the bacterial cells exhibited growth on plates lacking antibiotics, signifying their survival during the procedure. The plates clearly depict a significant presence of bacterial growth, characterized by smears (figure 19).

In light of the initial lack of success, a meeting was convened with Prof. Per Saris, an expert in the transformation of L. crispatus and the author of the article from which we derived our protocol, to gain insights into optimizing the transformation process. Prof. Saris offered several suggestions, including the adoption of smaller plasmids and adjustments to the glycine concentrations. Glycine plays a critical role in compromising the integrity of the bacterial cell wall, allowing greater DNA uptake. In accordance with these recommendations, a transformation experiment was conducted, where L. crispatus cells were incubated in growth medium with varying glycine concentrations, namely 0.5%, 0.8%, 1.1%, and 1.4%. The selection of these concentrations was based on a thorough review of the literature, aimed at identifying the ideal threshold for achieving positive outcomes. Regrettably, these modified transformation attempts also proved to be unsuccessful. The plate clearly illustrates the absence of bacterial growth in the given experimental conditions (figure 19).

Furthermore, the use of the compact plasmid pMG36E, with a small size of 3610 base pairs, did not yield successful results.

Figure 19: Transformation of L. crispatus with pMG36E+AmyL on Erythromycin MRS Plate. This figure visually documents the outcome of the transformation attempt involving L. crispatus using the pMG36E+AmyL plasmid on Erythromycin-containing MRS agar plates.

Figure 20: Transformation of L. crispatus with pMG36E+AmyL on MRS Plate. This figure offers a visual representation of the transformation process applied to L. crispatus utilizing the pMG36E+AmyL plasmid on standard MRS agar plates.

Subsequently, an alternative transformation protocol from a separate research article was adopted. A notable difference between this protocol and the previous one pertained to the growth conditions of L. crispatus. In the first protocol, bacterial cells were initially cultivated for 8 hours in sterile MRS (de Man, Rogosa, and Sharpe) medium, followed by transfer to a medium containing 0.8% glycine. In the second protocol, cells were grown in MRS medium containing 0.8% glycine for approximately 10 hours. However, the adoption of this second protocol also led to unfavorable results. We can assume that several factors contributed to this lack of success:

First and foremost, one of the transformations resulted in contamination with a bacterium that was not L. crispatus, due to human error. Notably, the plates visually portray the presence of numerous colonies on the plate, some of which exhibit distinct characteristics that are not typical of L. crispatus. These atypical colonies indicate a contamination issue within the transformation process (figures 21,22,23).

Figure 21: Transformation of L. crispatus on Erythromycin-Containing MRS Plate. Transformation of L. crispatus with closed pMG36E on Erythromycin MRS Plate. This figure visually documents the outcome of the transformation attempt involving L. crispatus using the pMG36Eplasmid on Erythromycin-containing MRS agar plates.

Figure 22: Transformation of L. crispatus with UPW on MRS Plate. This figure visually documents the outcome of the transformation attempt involving L. crispatus using the UPW on MRS agar plates.

Figure 23: Close Examination of Atypical Colonies. This figure offers a magnified view of the colonies identified as atypical in Figure 22. The colonies in question exhibit characteristics distinct from the typical appearance of L. crispatus and have raised concerns of contamination. This close examination allows for a more detailed scrutiny of these colonies, aiding in the effort to discern their unique features and origin

This was confirmed by culturing the bacteria in liquid MRS broth from the colonies that emerged on the transformation plates. The growth of these colonies displayed characteristics distinct from typical L. crispatus growth. Notably, the contaminated bacterium did not settle at the bottom of the test tube, resulting in turbidity throughout the culture, providing definitive evidence of contamination (figure 24).

Figure 24: Liquid MRS Medium inoculated with an atypical Colony. This figure displays a liquid MRS medium inoculated with an atypical colony derived from the transformation plate suspected to be a contamination identified in Figures 22 and 23.

The unsuccessful transformation of L. crispatus can be attributed to several critical factors. Firstly, the Gram-positive nature of L. crispatus presented a significant challenge. Gram-positive bacteria typically have thicker cell walls, making it more difficult for plasmids to penetrate and successfully enter the cells. Despite the efforts to minimize plasmid size, the barrier posed by the cell wall might have still impeded effective plasmid uptake. This could be a key reason for the transformation's failure.

Moreover, the transformation method involved electroporation, a technique demanding a high level of laboratory expertise, precision, and rapid execution, often necessitating specialized, temperature-controlled equipment. Given the inexperience of our team in performing this method, there exists a possibility that errors were introduced during the electroporation process, further complicating the issue.

Given the challenges with transformation, we are considering an integration of genetic modifications into the bacterium genome using the CRISPR-Cas3 system. For more information on our approach, please visit our genome integration page.

Regrettably, due to time constraints imposed by the project, we were unable to conduct further transformation experiments that might have yielded successful outcomes. Nonetheless, we firmly believe in the future potential for achieving a successful transformation in L. crispatus. Such a successful transformation would mark a significant milestone in advancing genetic engineering within this critical bacterium.


  1. “Part:BBa I0500 -” (accessed Sep. 16, 2023).
  2. L. J. Mota, L. Morais Sarmento, and I. De Sá-Nogueira, “Control of the Arabinose Regulon in Bacillus subtilis by AraR In Vivo: Crucial Roles of Operators, Cooperativity, and DNA Looping,” J. Bacteriol., vol. 183, no. 14, p. 4190, 2001, doi: 10.1128/JB.183.14.4190-4201.2001.
  3. I. Sá-Nogueira, T. V. Nogueira, S. Soares, and H. De Lencastre, “The Bacillus subtilis L-arabinose (ara) operon: Nucleotide sequence, genetic organization and expression,” Microbiology, vol. 143, no. 3, pp. 957–969, Mar. 1997, doi: 10.1099/00221287-143-3-957/CITE/REFWORKS.
  4. S. A. Bustos and R. F. Schleif, “Functional domains of the AraC protein.,” Proc. Natl. Acad. Sci., vol. 90, no. 12, pp. 5638–5642, Jun. 1993, doi: 10.1073/PNAS.90.12.5638.
  5. I. Sá-Nogueira and L. J. Mota, “Negative regulation of L-arabinose metabolism in Bacillus subtilis: characterization of the araR (araC) gene,” J. Bacteriol., vol. 179, no. 5, pp. 1598–1608, 1997, doi: 10.1128/JB.179.5.1598-1608.1997.
  6. A. Brunelle and R. Schleif, “Determining residue-base interactions between AraC protein and araI DNA,” J. Mol. Biol., vol. 209, no. 4, pp. 607–622, Oct. 1989, doi: 10.1016/0022-2836(89)90598-6.
  7. E. Balleza, J. M. Kim, and P. Cluzel, “Systematic characterization of maturation time of fluorescent proteins in living cells,” Nat. Methods 2017 151, vol. 15, no. 1, pp. 47–51, Nov. 2017, doi: 10.1038/nmeth.4509.
  8. A. Doerr, “(Protein) timing is everything,” Nat. Methods 2012 98, vol. 9, no. 8, pp. 783–783, Jul. 2012, doi: 10.1038/nmeth.2120.
  9. S. S. Beasley, T. M. Takala, J. Reunanen, J. Apajalahti, and P. E. J. Saris, “Characterization and Electrotransformation of Lactobacillus Crispatus Isolated from Chicken Crop and Intestine,” Poult. Sci., vol. 83, no. 1, pp. 45–48, Jan. 2004, doi: 10.1093/PS/83.1.45.