Validating a Promoter Library for Application in Plasmid-Based Diatom Genetic Engineering

Abstract

While diatoms are promising synthetic biology platforms, there currently exists a limited number of validated genetic regulatory parts available for genetic engineering. The standard method for diatom transformation, nonspecific introduction of DNA into chromosomes via biolistic particle bombardment, is low throughput and suffers from clonal variability and epigenetic effects. Recent developments in diatom engineering have demonstrated that autonomously replicating episomal plasmids serve as stable expression platforms for diverse gene expression technologies. These plasmids are delivered via bacterial conjugation and, when combined with modular DNA assembly technologies, provide a flexibility and speed not possible with biolistic-mediated strain generation. In order to expand the current toolbox for plasmid-based engineering in the diatom Phaeodactylum tricornutum, a conjugation-based forward genetics screen for promoter discovery was developed, and application to a diatom genomic DNA library defined 252 P. tricornutum promoter elements. From this library, 40 promoter/terminator pairs were delivered via conjugation on episomal plasmids, characterized in vivo, and ranked across 4 orders of magnitude difference in reporter gene expression levels.

Keywords: diatom; episomal gene expression; forward genetics; genetic engineering; parts registry; promoter characterization.

Figure 1

Validation of GUS marker gene expression as a reporter for promoter activity when driven from a diatom episome. Neg is the negative control strain containing a genomically integrated selection marker gene but lacking the GUS gene. PtGG represents the hairpin forming DNA sequence-GUS-T_fcpA cassette delivered via pDEST and not expected to drive protein expression, whereas PB-fcpB represents a strain containing a P_fcpB-GUS-T_fcpA cassette integrated into the genome via particle bombardment with the error bars representing 10 independently picked transformants. The other lines were generated by conjugation and refer to the promoter upstream of the GUS reporter as follows: p₄₉ = hypothetical protein (Phatr3_J49202), nr = nitrate reductase (Phatr3_J54983), h4 = histone H4 (Phatr3_J34971), and fcpB = fucoxanthin-chlorophyll a–c binding protein B (Phatr3_J25172). P_nr represents pairing of the nr 5′UTR with the fcpA terminator, while P_nr2 utilized the native 3′UTR. Error bars for the conjugative vectors represent the standard deviation from assaying three (3) exconjugants in parallel, while the error bar for PB-fcpB is the result of assaying 10 biological replicates. Asterisks indicate a t-test P-value ≤0.05 (*) or ≤0.001 (***) when comparing the expression of each cell line to PtGG. For the conjugation strains, the difference in the expression of P_fcpB vs P_h4, P_p49, and P_nr2 was analyzed, and in agreement with transcriptomic data, the relative strength of this promoter set was ranked as P_p49 > P_nr2 > P_fcpB = P_h4. For the conjugation lines, samples of liquid cultures were taken at three different time points during the day cycle; within a promoter, the expression levels did not fluctuate in a statistically significant manner over the course of the day for this set of plasmids (t-test, p > 0.05). Cells were cultured in the presence of 8.8 mM NO₃. Raw RFUs were normalized to total protein content in each assayed lysate.

Figure 2

Venn diagrams representing the results of the forward genomic library promoter screen sequencing. (A) 252 sequencing hits were obtained from the forward genetic screening using the diatom conjugative plasmid driving the phleomycin resistance gene when exconjugants were isolated using the medium containing different nitrogen sources (nitrate vs urea). Of these 252 hits, 186 were recovered from NO₃-grown exconjugants, while 66 were recovered from exconjugants selected using urea as the nitrogen source. (B) When mapped to the Phatr3 genome assembly, it was shown that 150 unique hits were recovered from NO₃-grown colonies and 30 unique hits from colonies selected on urea. 36 duplicate hits were recovered from both media selections (total of 72). (C) Based on the criteria outlined in the text, 175 of the 252 sequencing hits were classified as “promoters” with 102 hits from NO₃ and 15 from urea-selected colonies. Twenty-nine promoters were identified as duplicates (found in both media conditions) for a total of 146 unique promoters identified in this screen. (D) 77 sequencing hits were labeled as “possible”, defined as DNA fragments recovered from colony selection not fitting the criteria for a canonical promoter but still able to drive expression of ShBle in vivo. NO₃-selected colonies produced 48 hits, while urea-selected colonies produced 15 hits unique to that selection condition. Seven duplicate (14 in all) unclear hits were found to overlap, for a total of 70 hits.

Figure 3

Validation of the forward genomics promoter screen workflow by cloning and analysis of GUS expression relative to the negative PtGG construct. In the top panel, the activities of seven putative P. tricornutum promoters were quantified based on GUS expression when cells were grown in L1 medium with either nitrate or urea as the sole nitrogen source. PtGG represents a negative control strain containing a plasmid harboring the PtGG DNA fragment cloned in front of GUS with no expected promoter activity. The h4 promoter is used as a positive control with known promoter activity. The error bars represent the standard deviation of three biological replicates for each cell line. Asterisks represent a t-test P-value ≤0.01 (**) or ≤0.001 (***) when comparing the expression to PtGG. Bottom panel, table of hit IDs for colonies tested with the GUS assay.

Figure 4

Expression profiles of pEG constructs active across a range of 4 orders of magnitude comparing the presence of native versus non-native terminators with putative promoters. Normalized beta-glucuronidase activity for diatom exconjugants expressing synthetic promoter–reporter–terminator constructs on episomes. The GUS activity is determined using a fluorescence assay on protein extracts and normalized to protein concentrations. Error bars are for 3–6 biological replicates with 1–2 technical replicates. Neg represents a negative control strain containing a plasmid harboring the PtGG DNA fragment cloned in front of GUS with no expected promoter activity. Mapping of promoter names to full protein IDs is provided in Supporting Information Table S6. As presenting 361 individual t tests would make this figure unreadable, we have included this information as a separate heatmap in Figure 5.

Figure 5

P-values from t tests for native and fcpA terminator pairs. (A) All vs all heatmap of the p-values for the promoter-native terminator pairs. (B) All vs all heatmap of the p-values for the promoter–fcpA terminator pairs. Comparisons that yielded a statistically significant p-value are highlighted in gray. The column on the right-hand side represents the mean expression value.

Figure 6

Expression of GUS from conjugated diatom episomes is stable over an extended period of culturing. P. tricornutum strains harboring plasmids with two high (P_nr and P_h4) (promoter/native terminator), one medium (P_dph1), and one low (P_gnef) (promoter/fcpA terminator) expressing promoter/terminator pairs driving the GUS marker gene were analyzed initially (T0) and after a period of 4 months (T4). No difference in expression strength (t-test, P > 0.05, within promoter t = 0 vs t = 4 tested) or loss of plasmid occurred when the cells were cultured appropriately and in the presence of antibiotic selection (phleomycin, 20 μg/mL). The error bars represent the standard deviation of expression of 3 biological replicates for each cell line.