The first step on our roadmap was to identify the UCYN-A Transit Peptide (uTP) sequence, which is essential for importing host-encoded proteins into UCYN-A. This process is crucial because protein import into UCYN-A must be understood and replicated to develop a successful symbiotic relationship between the organelle and a new eukaryotic host. By enabling the import of specific host proteins into UCYN-A, this step lays the foundation for transplanting UCYN-A into a model organism.

Our dry lab team used bioinformatics tools, including multiple sequence alignment (MSA) and motif analysis, to identify conserved motifs in uTP sequences from the proteomics data of Braarudosphaera bigelowii (B. bigelowii) [1]. These C-terminal extensions exhibited characteristics similar to transit peptides in mitochondria and chloroplasts, and we hypothesized that these sequences could direct protein import into UCYN-A.

To investigate the in vivo behavior of the identified uTP sequences, we engineered plasmids containing uTP fused to fluorescent proteins. Our aim was twofold: first, to observe the localization of uTP-marked proteins in cells lacking UCYN-A, which could provide insights into potential relationships between the UCYN-A protein import system and other known mechanisms. Second, expressing a uTP-containing protein would enable us to purify it, inject it into live B. bigelowii, and monitor its localization in the native host, thereby confirming uTP's function as a targeting signal.

For C. reinhardtii, we used the plasmid backbone pOpt2-mVenus-Ble [6] encoding mVenus, a yellow fluorescent protein, as this protein has a lower background fluorescence in C. reinhardtii compared to green fluorescent protein (GFP), as noted by [2]. For S. cerevisiae, we chose the pUDE1311 [7] plasmid encoding mNeonGreen.

Additionally, we created control plasmids containing mitochondrial transit peptides (MTS1) [8] and chloroplast transit peptides (CTS) [9]. These controls would allow us to compare the localization of uTP-tagged proteins with those that are known to localize to specific organelles in S. cerevisiae and C. reinhardtii. This comparison lets us assess whether uTP shows similar or distinct localization patterns.

During the testing phase, we attempted to transform S. cerevisiae and C. reinhardtii with the plasmids to observe protein localization using fluorescence microscopy. Unfortunately, we were unable to successfully construct plasmids with the uTP sequence incorporated. However, we managed to construct and image plasmids containing MTS1 in S. cerevisiae, which showed correct localization to the mitochondria as expected.

From this initial round of testing, we learned that while we successfully confirmed mitochondrial localization for MTS1, the uTP plasmid construction did not succeed.

Although we lacked results for uTP, this partial success with MTS1 suggests that our experimental pipeline works for mitochondrial localization. The failure with the uTP constructs highlights the need for further refinement in plasmid assembly and transformation.

As a next step, we plan to reattempt the assembly of the plasmid with the uTP sequence and repeat the transformations. Once the uTP plasmids are successfully constructed, we aim to test their localization properties in eukaryotic hosts like S. cerevisiae and C. reinhardtii to confirm whether the uTP sequence can direct protein import into UCYN-A.

Our initial plan for identifying the uTP sequence was based on the results from Coale et al. (2024). We aimed to identify the conserved C-terminal region, find repeating motifs, and create a consensus sequence.

To implement our analysis plan, we developed Python code for sequence analysis and filtering. For motif analysis, we utilized the MEME suite, a powerful tool for discovering and analyzing sequence motifs.

We visualized the identified motifs and their co-occurrences to better understand the patterns within the uTP sequences.

Through this analysis, we discovered that there are multiple variations in which motifs occur, resulting in different uTP variants. Despite the variations, we observed some shared properties among these variants.

Based on our findings from the first cycle, we decided to investigate whether certain uTP variations correlate with specific protein sequences. This could provide insights into the relationship between protein structure and uTP variation.

To test this hypothesis, we implemented machine learning classifiers trained on the sequence data. These classifiers were designed to predict the corresponding uTP motifs from the protein's main body.

We evaluated the performance of our classifiers using 5-fold cross-validations and permutation tests. These rigorous testing methods helped ensure the reliability of our results.

Our analysis revealed that three of the classifiers indicated a statistically significant correlation between protein sequences and motif variations. Among these, the Logistic Regression classifier performed best. We used this top-performing classifier to design a uTP sequence for our fluorescent proteins.

To further evaluate the constructed fluorescent uTP proteins, we decided to investigate their 3D structure using prediction tools. This would allow us to compare the structural features of our designed uTP sequences with those of native uTP-containing proteins.

We used AlphaFold 3, a state-of-the-art protein structure prediction tool, to predict the structures of all uTP-containing B. bigelowii proteins and our uTP constructs.

The predicted structures were analyzed using PyMol and Python. We created a consensus structure from the native uTP sequences to highlight the structurally conserved regions.

Our analysis identified a region in the uTP sequences exhibiting a highly conserved structure. When we superimposed this region onto our construct structures, we observed a clear correspondence between the two. This finding suggests that our designed uTP sequences likely maintain the crucial structural features necessary for their function.

In this cycle, our goal was to construct plasmids that would express fluorescent proteins fused to uTP or control sequences in S. cerevisiae and C. reinhardtii. To ensure proper incorporation of the inserts, we designed primers with specific overhangs for PCR amplification. We planned to linearize the plasmid backbones (pOpt2-mVenusBle for C. reinhardtii and pUDE1311 for S. cerevisiae) at the C-terminal and N-terminal regions for precise insertion of the uTP, MTS1, and CTS sequences.

The assembly of plasmids containing the uTP, MTS1, and CTS sequences was then completed using the amplified backbones and inserts with overlapping ends. For this purpose, we used Gibson Assembly to combine the linearized plasmid backbones with the uTP, MTS1, and CTS inserts due to its efficiency in assembling multiple DNA fragments with overlapping ends. We aimed to confirm the correct plasmid assembly by transforming E. coli with the resulting constructs and performing diagnostic tests.

We amplified the plasmid backbones using KOD polymerase and incorporated the appropriate overhangs into our PCR products. The C-terminal and N-terminal regions of both the C. reinhardtii and S. cerevisiae plasmids were successfully linearized. We then carried out Gibson Assembly using the amplified plasmid backbones and inserts.

After excising the correct fragments and assembling the plasmids using Gibson Assembly, we ran additional diagnostic tests, including diagnostic PCR on the Gibson Assembled plasmid. In the case of the C. reinhardtii plasmid, we observed multiple fragments during gel electrophoresis, suggesting incomplete or incorrect assembly.

We also transformed E. coli cells with the assembled plasmids to test for successful construct formation. To confirm the successful incorporation of the inserts into the plasmids, we conducted colony PCR on the E. coli colonies. For the C. reinhardtii plasmid, the colony PCR yielded promising results, indicating successful assembly, which was later invalidated since the sequence was incorrect. However, for the S. cerevisiae plasmid, many colonies showed false positives, with only the plasmid backbone (without the insert) present.

Through this process, we learned that, in the case of the C. reinhardtii and S. cerevisiae N-terminal constructs, using primers with non-specific binding can lead to the generation of multiple fragments during gel electrophoresis, leading to complexities in plasmid construction. We excised the fragment of the correct size, but this approach raised concerns about potential DNA damage from UV exposure during gel excision, which could lower the yield or affect the quality of the DNA. For future plasmid construction, we plan to design new primers to avoid excision altogether and explore using restriction digestion instead of Gibson Assembly to increase the accuracy of our insertions.

Upon closer investigation, we also discovered that the false positives in the colony PCR of the S. cerevisiae constructs were likely due to leftover circular pUDE plasmid from the backbone amplification process. The ratio of Gibson product (with insert) to non-linearized backbone was too similar, leading to a higher proportion of colonies without the desired insert. This reduced the efficiency of Gibson Assembly. To address this issue, we decided to use DpnI digestion.

In this cycle, our goal was to transform E. coli with the resulting plasmid constructs and perform diagnostic tests. However, this time, we used DpnI for restriction digestion. DpnI is a restriction enzyme that cuts at methylated sites to eliminate any leftover circular plasmid from the backbone amplification.

We once again amplified the plasmid backbones using KOD polymerase and incorporated the appropriate overhangs into our PCR products. The C-terminal and N-terminal regions of S. cerevisiae plasmid were successfully linearized. We then carried out Gibson Assembly using the amplified plasmid backbones and inserts. Following this, we transformed E. Coli cells with the constructed plasmids.

To confirm the successful incorporation of the inserts into the plasmids, we conducted colony PCR on the transformed E. coli colonies.

After applying DpnI digestion, the transformation efficiency improved, and we obtained positive colonies verified by colony PCR. From this cycle, we learned that the leftover circular plasmid from the amplification process can significantly reduce the efficiency of Gibson Assembly and lead to false positives during colony PCR. By using DpnI digestion, we were able to remove the unwanted circular plasmid and increase the transformation efficiency. Moving forward, we plan to incorporate DpnI digestion as a standard step when using amplified plasmid backbones to ensure that only linearized DNA is used in the assembly.

In this cycle, we aimed to transform S. cerevisiae with the plasmids successfully assembled in the previous cycle. These plasmids contained fluorescent protein constructs (mNeonGreen) tagged with either the mitochondrial transit peptide (MTS1) or UCYN-A Transit Peptide (uTP2), alongside a control without transit peptide (mNeonGreen only) and a negative control without plasmid.

To select for successfully transformed yeast cells, we used the URA3 gene as a selective marker, allowing growth in uracil-deficient media. Our goal was to confirm successful transformation and integration of the plasmids into S. cerevisiae for subsequent localization studies.

We followed a yeast transformation protocol adapted from [3]. After transforming the yeast with the mNeonGreen-MTS1, mNeonGreen-uTP2, and control plasmids, we plated the cells on uracil-deficient media to select for successful transformants.

Colonies that formed on the selective plates were isolated, and diagnostic PCR was performed on 9 colonies from the mNeonGreen-MTS1 transformation to confirm the presence of the MTS1 transit peptide sequence. Out of the 9 colonies tested, 8 showed positive results for the MTS1 sequence, indicating successful transformation. However, none of the colonies transformed with the mNeonGreen-uTP2 construct showed positive results.

To test the results of the transformations, we repeated the transformation with the uTP2 construct to rule out low transformation efficiency as the cause of the negative results. After the second round of transformation, the colonies formed more slowly and did not grow as well as those transformed with the MTS1 construct.

We inoculated the cells in liquid media supplemented with the appropriate selective antibiotics, and after 3 days, the cultures grew successfully. Colony PCR confirmed the presence of the correct inserts in the colonies for the MTS1 construct, but microscopy imaging revealed no significant fluorescent signal in the cells transformed with the uTP2 construct.

We sequenced the Gibson assembly constructs using Plasmidsaurus, which revealed that the backbones of the uTP2 plasmid were incorrect, explaining the lack of fluorescence. This discovery highlighted the importance of sequencing the constructs before proceeding to transformation. Moving forward, we plan to sequence the plasmids immediately after assembly to avoid unnecessary transformation steps with incorrect constructs.

Our goal in this cycle was to design a protocol for fusing UCYN-A into a eukaryotic host cell. As a proof of concept, we aimed to replicate the fusion protocol described by [4], which uses polyethylene glycol (PEG) to make host cell membranes more permeable, allowing for the fusion of E. coli into S. cerevisiae. We adapted the protocol by varying the concentrations of zymolyase (an enzyme used to digest the yeast cell wall), and we used E. coli and S. cerevisiae strains that were auxotrophy-independent, meaning they did not need to rely on each other for survival.

To track the success of the fusion under the microscope, we used E. coli NCM3722 expressing PlsB-msGFP2, which would allow us to observe the GFP fluorescence inside the yeast cells if successful fusion occurred.

We executed our adapted PEG fusion protocol by treating S. cerevisiae cells with zymolyase to make the cell walls more permeable, followed by PEG treatment to facilitate the fusion with E. coli cells. The cells were then prepared for microscopy imaging to observe the fusion results.

After completing the protocol, we imaged the samples using scanning confocal microscopy to detect any fluorescent E. coli inside S. cerevisiae cells. Unfortunately, the results were inconclusive. While some fluorescence was observed, it was not localized to the yeast cells, and no clear evidence of E. coli inside S. cerevisiae was found. Additionally, some yeast cells exhibited strong autofluorescence, which interfered with the measurements.

From this initial round of testing, we concluded that several factors may have contributed to the lack of conclusive results:

1. The strong autofluorescence in S. cerevisiae cells could have masked the GFP signal from E. coli.
2. The zymolyase concentration used may not have been sufficient to adequately digest the yeast cell wall, preventing successful fusion.
3. The ratio of S. cerevisiae to E. coli cells may not have been optimal for fusion.

In light of these findings, we planned a follow-up experiment with control samples to determine the cause of the observed fluorescence. We also considered the use of additional staining methods, such as methylene blue to assess yeast cell viability and Concanavalin-A FITC to visualize whether the yeast cell wall had been fully digested.

After the initial PEG fusion experiments yielded inconclusive results, we designed a follow-up set of experiments to address the key issues we identified, such as autofluorescence in S. cerevisiae and insufficient zymolyase concentration. Our objective was to test the hypothesis that the observed fluorescence in the yeast cells might have been due to autofluorescence or improper cell wall digestion.

To test this hypothesis, we planned control samples and introduced additional staining methods to clarify the source of the fluorescence:

1. S. cerevisiae treated with PEG but without E. coli (to control for autofluorescence).
2. E. coli added to S. cerevisiae without PEG treatment (to control for random bacterial association).
3. Repetition of the original PEG fusion protocol with two different zymolyase concentrations.
4. Use of methylene blue staining to check yeast viability (live cells should not take up the stain).
5. Use of Concanavalin-A FITC to visualize whether the yeast cell wall had been successfully digested.

We prepared the control samples and repeated the PEG fusion protocol using the adjusted zymolyase concentrations. After the fusion process, the cells were stained with methylene blue and Concanavalin-A FITC to assess cell wall integrity and cell viability.

Once the staining was complete, the samples were prepared for imaging with confocal microscopy to evaluate fluorescence and compare the controls to the experimental samples.

We imaged the control and experimental samples using scanning confocal microscopy to detect GFP fluorescence and assess cell viability. Additionally, we performed optical microscopy to check yeast viability using methylene blue staining.

The results showed no significant difference between the control samples and the full protocol experiments. The same fluorescence observed in some S. cerevisiae cells appeared across all conditions, confirming that the signal was likely due to autofluorescence. Furthermore, E. coli was not visible in the yeast cells, which suggested that the GFP expression was too weak under the current conditions.

From this second cycle of testing, we learned that the autofluorescence signal in S. cerevisiae cells was likely masking the GFP signal from E. coli, as we observed similar fluorescence in the control samples. Additionally, the weak GFP expression in E. coli may have further contributed to the lack of observable fusion events.

To improve the fusion protocol, we plan to:

1. Use DAPI staining, which binds to DNA, to detect E. coli without interference from yeast autofluorescence.
2. Consider using a stronger promoter to enhance GFP expression in E. coli, ensuring that any fusion events are more easily detectable in future experiments.

The objective of this cycle was to transform Chlamydomonas reinhardtii with the mVenus-insert fused to different peptide sequences: the UCYN-A Transit Peptide (uTP2), mitochondrial transit peptide (MTS1), and chloroplast transit peptide (CTS). The goal was to test whether these peptides could direct the localization of the fluorescent proteins to their respective compartments within the Chlamydomonas cells.

We first designed a transformation protocol based on the T-sorbonne Chlamydomonas guide and a transformation protocol by [5]. The UVM4 strain of C. reinhardtii, obtained from the Planck Institute, was chosen for its high transformation efficiency. For the transformation, we selected the pOpt2-mVenus-Ble plasmid backbone, which contains specific introns in the mVenus coding sequence to prevent gene silencing in Chlamydomonas and strong promoters to drive expression.

We prepared the mVenus-insert by performing restriction digestion around the mVenus-insert fragment, followed by SV gel and PCR cleanup to isolate the insert. Due to inefficient cleanup methods that resulted in a low yield of the insert, we repeated the digestion and switched to ethanol precipitation, which improved the recovery of the insert for transformation.

Electroporation was performed using the protocol we designed, and the transformed cells were plated on zeocin-resistant agar plates. In the first transformation attempt, we used too much liquid and a high cell density, which left the plates too wet and caused poor colony formation. For the second attempt, we optimized the conditions by reducing the liquid volume and cell density, resulting in clearer plates with distinct colonies.

Following the second transformation and plating attempt, the optimized conditions led to the formation of distinct zeocin-resistant colonies. We selected 12 colonies for each construct (mVenus-CTS, mVenus-MTS1, mVenus-uTP2, mVenus) and screened them for fluorescence in a 96-well plate format to test whether the transformed colonies expressed mVenus with the peptide sequences attached.

From the initial screening results, we observed that only one colony transformed with the mVenus control showed fluorescence. The absence of fluorescence in the other constructs suggests that the transformation efficiency or expression of the peptide-fused fluorescent proteins may need further

In future experiments, we would focus on improving the expression of the peptide-tagged mVenus proteins. Specifically, verifying the quality and yield of the insert before transformation would help ensure successful expression in C. reinhardtii.