High-efficiency homologous recombination in the oil-producing alga Nannochloropsis sp.

Abstract

Algae have reemerged as potential next-generation feedstocks for biofuels, but strain improvement and progress in algal biology research have been limited by the lack of advanced molecular tools for most eukaryotic microalgae. Here we describe the development of an efficient transformation method for Nannochloropsis sp., a fast-growing, unicellular alga capable of accumulating large amounts of oil. Moreover, we provide additional evidence that Nannochloropsis is haploid, and we demonstrate that insertion of transformation constructs into the nuclear genome can occur by high-efficiency homologous recombination. As examples, we generated knockouts of the genes encoding nitrate reductase and nitrite reductase, resulting in strains that were unable to grow on nitrate and nitrate/nitrite, respectively. The application of homologous recombination in this industrially relevant alga has the potential to rapidly advance algal functional genomics and biotechnology.

Research on eukaryotic algae has provided fundamental insights into many basic cellular processes, particularly photosynthesis. Recently, algae have reemerged as potential next-generation feedstocks for advanced biofuels, such as biodiesel and other hydrocarbons (1–4), and have attracted considerable interest from both the private and public sectors. Many algae have high photoautotrophic growth rates and can accumulate more than half their dry weight biomass as lipids (5), including triacylglycerol and a number of high-value pharmaceutical and nutraceutical products (6). Compared with agricultural plants, algae do not require arable land, and many species can be grown using wastewater or salt water. However, because unimproved algae are unlikely to possess all of the traits necessary for economic production of biofuels (1), robust molecular biology tools are required for strain optimization (7).

Eukaryotic algae comprise a diverse, polyphyletic group of organisms with members in four of the five eukaryotic supergroups (8). One supergroup, the Plantae, includes rhodophytes (red algae), glaucophytes, chlorophytes (green algae), and plants, which all contain plastids derived from primary endosymbiosis of a cyanobacterium (9). Photosynthetic plastids have been transferred to other eukaryotic supergroups through secondary endosymbiosis of green algae and red algae. For example, chromalveolate algae such as diatoms, brown algae, eustigmatophytes, and most dinoflagellates contain a plastid derived from a red alga (10).

The green alga Chlamydomonas reinhardtii is currently the eukaryotic model alga of choice, because of its genetic tractability (11), array of molecular tools (12, 13), and comprehensively annotated nuclear genome sequence (14). Genome sequences of several algae of ecological, economic, and/or phylogenetic significance have been determined recently (15). Transformation methods have been developed for several algal species, most notably the model diatom Phaeodactylum tricornutum (16, 17); however, the molecular genetic tools available for these species are generally quite limited in comparison with Chlamydomonas.

Although Chlamydomonas is often regarded as “the green yeast” (18, 19), the lack of efficient homologous recombination (HR) in the nuclear genome of this alga (20–22) has been a major limitation for algal biology research. By contrast, the implementation of HR in the yeast Saccharomyces cerevisiae has enabled precise genetic manipulation and systematic functional analysis of its genome (23) and has led to pioneering discoveries in eukaryotic biology.

Although it resembles green algae of the genus Chlorella (24), Nannochloropsis sp. is a eustigmatophyte that is related to diatoms and brown algae (25). Like Chlorella but unlike Chlamydomonas, Nannochloropsis is a robust industrial alga that has been extensively grown in outdoor ponds and photobioreactors for aquaculture (26, 27). Under nitrogen-starvation conditions, Nannochloropsis can accumulate oil exceeding 60% of its biomass on an ash-free dry weight basis (27), making it an excellent candidate for biofuel production. In addition, Nannochloropsis is a rich source of high-quality protein (28) and eicosapentaenoic acid (29), an omega-3 fatty acid with numerous health benefits (30).

Here we describe the development of a highly efficient transformation system for Nannochloropsis. Moreover, we demonstrate that Nannochloropsis sp. exhibits very efficient HR, allowing us to reliably and rapidly perform targeted genetic manipulation of this organism. These findings open the door for systematic functional genomics and biotechnological improvement of this industrial alga.

Results

Development of an Efficient Transformation System for Nannochloropsis.

Nannochloropsis sp. (strain W2J3B) grows rapidly on solid or liquid media, with a doubling time of ∼14 h under laboratory conditions (Materials and Methods). To establish a transformation system for Nannochloropsis sp., we used endogenous promoters and 3′ untranslated regions (3′ UTRs) from two unlinked violaxanthin/chlorophyll a-binding protein (VCP) genes, VCP1 and VCP2 (Fig. 1). The protein encoded by the VCP1 gene is nearly identical to a previously characterized light-harvesting antenna protein from Nannochloropsis sp. (31). The VCP2 locus consists of two identical copies of the VCP2 gene driven by a bidirectional promoter (Fig. 1, SI Materials and Methods, and Fig. S1). Using transformation constructs containing the Sh ble selectable marker gene, which confers resistance to zeocin (32), we developed and optimized an efficient transformation method based on electroporation of intact Nannochloropsis sp. cells (SI Materials and Methods and Fig. S1). The highest transformation efficiency (∼2,500 transformants/μg of DNA) was obtained using a very high electric field strength within the electroporation cuvette (∼11,000 V/cm field strength, 50 μF capacitance, and 500 Ohm shunt resistance; Materials and Methods, SI Materials and Methods, and Fig. S2). The number of transformants obtained was approximately linear with respect to the amount of DNA used, over a range from 200 ng to 3 μg (SI Materials and Methods and Fig. S3). Transformation of a modified version of the C2 transformation construct (Fig. 1) in which the promoter had been flipped (construct C1; Fig. S1) generated a similar number of transformants as the C2 construct, showing that the VCP2 promoter indeed drives transcription in both directions (SI Materials and Methods and Fig. S4). When we tried to transform a plasmid containing the C2 cassette without prior linearization, we obtained only a single zeocin-resistant colony in five independent transformation experiments, indicating that efficient stable transformation of Nannochloropsis sp. requires linear DNA fragments (SI Materials and Methods).

Fig. 1.

Composition of transformation constructs. The VCP1 gene from Nannochloropsis sp. consists of a promoter (Prom), the coding region, and a 3′ untranslated region. The VCP2 locus consists of two identical coding sequences (dubbed VCP2 and VCP2′) driven by a central, bidirectional promoter element (VCP2-Prom bidirectional) and each followed by a 3′ UTR. In the C2 transformation construct, the Sh ble gene conferring resistance to zeocin is fused to the bidirectional promoter from VCP2 and the 3′ UTR from VCP1. In the NT7 cassette, the bidirectional promoter is truncated.

We also constructed transformation cassettes based on selectable markers conferring resistance to hygromycin B and blasticidin S (Fig. S1). To determine the frequency of cotransformation, we performed experiments in which a zeocin-resistance cassette was transformed together with an excess of the hygromycin B- and blasticidin S-resistance constructs, and transformants were initially selected by plating the cells on zeocin only (SI Materials and Methods). Subsequent replating of zeocin-resistant colonies on hygromycin B- or blasticidin S-selective media revealed total cotransformation frequencies of 50 and 55% for at least one unselected marker and 23 and 31% for both unselected markers in two independent experiments (Table S1).

Gene Knockouts by Homologous Recombination.

To test for the occurrence and frequency of HR in Nannochloropsis sp., we designed gene knockout (KO) constructs based on a modified zeocin-resistance cassette (NT7) with a truncated VCP2 promoter (Fig. 1). Nannochloropsis sp. is able to grow on medium containing nitrate, nitrite, or ammonium as the sole nitrogen source, and we chose the nitrate reductase (NR) and nitrite reductase (NiR) genes as targets for HR, because successful disruption of these genes can easily be detected by growth on different nitrogen sources. Within the algal cell, nitrate is first reduced to nitrite by the action of NR and then further reduced to ammonia by NiR. Thus, cells lacking NR can grow on nitrite but not nitrate, whereas cells without NiR are not able to use nitrate or nitrite. Ammonia assimilation does not require NR or NiR, and thus cells with a defective NR or NiR gene can be grown on ammonium as a nitrogen source.

We performed transformation with KO constructs containing ∼1-kb flanking sequences that target the NR (Fig. 2A) and NiR (Fig. 2B) genes. KO constructs were designed to replace 242 bp of the NR gene encoding a part of the molybdenum cofactor-binding domain (cd02112) or 793 bp of the NiR gene encoding the terminal nitrite/sulfite reductase ferredoxin-like half-domain (pfam03460) with the NT7 selection marker cassette. In both cases, we obtained zeocin-resistant transformants on medium containing ammonium as the sole nitrogen source. Replating of these colonies on media containing either nitrate or ammonium as the sole nitrogen source revealed that 25–94% of the NR-KO and 11–22% of the NiR-KO transformants bleached on nitrate, whereas all transformants grew on ammonium (Fig. 3 and Table S2).

Fig. 2.

KO of nitrate reductase and nitrite reductase genes by HR in Nannochloropsis sp. W2J3B. Structures of NR-KO (A) and NiR-KO (B) transformation constructs (TC), wild-type (Wt) genes, and HR products. Each KO construct consists of a left flank (red) and a right flank (orange) separated by the NT7 selection marker cassette (green). A target region (purple) of each gene is replaced by the NT7 cassette when HR occurs. Primer positions (arrows) for PCR analysis and expected PCR product sizes for Wt and KO mutants are indicated.

Fig. 3.

Analysis of transformants obtained with NR-KO and NiR-KO transformation constructs. (A, B, D, and E) Sixteen randomly selected zeocin-resistant transformants obtained with each TC were resuspended in medium lacking a nitrogen source and spotted on plates containing ammonium (A and D) or nitrate (B and E) as a sole nitrogen source. Putative NR-KO and NiR-KO mutants bleach, because they cannot use nitrate as a nitrogen source. (C and F) PCR analysis of the transformants shown in B and E. Genomic DNA was amplified by PCR with the primer pairs indicated in Fig. 2. Transformants shown by PCR analysis to have undergone HR at the NR or NiR locus are marked with an asterisk in B and E, respectively.

PCR analysis of the genomic DNA isolated from NR-KO and NiR-KO transformants using one or both primers outside of the homologous genomic DNA-flanking regions in the KO constructs (Fig. 2) revealed that the NT7 cassette had successfully inserted into the genome and replaced part of the target genes in the transformants that bleached on nitrate (Fig. 3). We obtained single PCR products for the NR-KO and NiR-KO transformants and, as expected, these products were longer in the cases of successful HR than those obtained for transformants that did not exhibit bleaching on nitrate (Fig. 3). DNA sequencing of the longer PCR products confirmed a perfect exchange of the targeted genomic DNA with the NT7 cassette. The presence of a single PCR product and the absence of the shorter wild-type allele in the KO strains strongly suggest that Nannochloropsis sp. W2J3B is haploid. This was further supported by PCR analysis of two nontargeted insertion sites obtained with transformation constructs lacking additional flanking sequences (SI Materials and Methods). In both cases, we also obtained a single PCR product, indicating the presence of a single allele (Fig. S5).

Liquid growth analysis of two NR-KO and two NiR-KO transformants that bleached on nitrate demonstrated that NR-KO transformants could not use nitrate for growth, but were able to grow on nitrite as a nitrogen source (Fig. 4). NiR-KO transformants could not grow on either nitrate or nitrite as a nitrogen source (Fig. 4).

Fig. 4.

Growth of wild-type, NR-KO (NR1 and NR2), and NiR-KO (NiR1 and NiR2) strains with different nitrogen sources. Cells in mid-log phase were washed in nitrogen-free medium, resuspended in media with the indicated nitrogen sources, and allowed to grow to early stationary phase. Results are expressed relative to the wild type in 1 mM NH₄Cl, and SDs for three independent cultures of each strain are indicated.

Discussion

We have established a highly efficient transformation protocol for Nannochloropsis sp., a photoautotrophic alga of high importance for the production of animal feed, biofuels, and nutraceuticals. Transient transformation of Nannochloropsis sp. using Agrobacterium has been described (33). Transformation of N. oculata after enzymatic digestion of the cell wall has also been reported (34), but the lack of a selectable marker gene in the transformation construct necessitated PCR screening of hundreds of colonies to identify a few stable transformants. In contrast, our protocol for the genetic transformation of Nannochloropsis sp. is simpler and can be directly applied to wild-type cells without prior removal of the cell wall, while yielding high numbers of transformants. Successful transformation of Nannochloropsis sp. via electroporation requires unusually high electric field strengths within the electroporation cuvette, which might be explained by the very robust cell wall of this alga (35).

Efficient transformation of Nannochloropsis sp. enables research on this alga using a wide range of approaches that are commonly used with Chlamydomonas (12, 13). For example, the C2 transformation construct might be useful for protein expression by fusing a gene of interest to the other end of the bidirectional VCP2 promoter (Fig. 1). Additional, unselected expression constructs or other DNA elements can be introduced by cotransformation, which occurs efficiently (Table S1). Because the C2 and NT7 transformation constructs use a 5′ promoter and 3′ UTR from different genes, integration of these cassettes (without additional flanking sequences) into the nuclear genome of Nannochloropsis sp. can occur by nonhomologous recombination (Fig. S5), resulting in insertional mutagenesis that could be exploited for forward genetics. The C2 construct with the bidirectional VCP2 promoter could also potentially be used for activation tagging (36).

Based on our results, we conclude that Nannochloropsis sp. is a haploid alga that is amenable to targeted gene manipulation by HR. Addition of flanking DNA sequences to the NT7 selectable marker cassette (Fig. 2) allowed for targeted integration of the cassette into the nuclear genome of Nannochloropsis sp. by HR, and we demonstrated the occurrence of efficient HR by knocking out the NR and NiR genes (Figs. 3 and 4 and Table S2). In the NR-KO and NiR-KO strains, the NT7 cassette had integrated into the NR and NiR genes, respectively, and the corresponding wild-type allele was undetectable by PCR (Fig. 3). Successful KO of NR or NiR in a single transformation step suggests that Nannochloropsis sp. is haploid. Diploid organisms that integrate transforming DNA by HR generally require successive rounds of transformation to obtain homozygous KO strains (37). The conclusion that Nannochloropsis sp. is haploid is further supported by the results of mutant isolation experiments (38) and the recent genome sequencing of a N. oceanica strain (39).

Although HR is routine in several cyanobacterial species (40) and in the chloroplast genome of Chlamydomonas (41), the dominant route of integration of transforming DNA into the nuclear genome of most photosynthetic eukaryotes is by nonhomologous recombination. DNA integration by HR is a relatively rare event in most model plants (42) and algae (7), including Arabidopsis, maize, rice, Chlamydomonas, and diatoms. Indeed, efficient targeted insertion of DNA constructs into the nuclear genome has been demonstrated in very few photosynthetic eukaryotes. One well-established example is the moss Physcomitrella patens, in which the use of HR has helped to establish this organism as a powerful model organism for studying the evolution of land plants (43). The only eukaryotic alga previously demonstrated to exhibit efficient HR is the red alga Cyanidioschyzon merolae (44, 45); however, this organism is a rather divergent red alga and an extremophile that is not generally viewed as a suitable model species for investigating biofuel production by algae. In contrast, Nannochloropsis sp. is a robust industrial alga that is considered to be a candidate feedstock species for the generation of biomass, biofuels, and valuable bioproducts.

The identification of a photoautotrophic alga that (i) is haploid, (ii) can be easily transformed with high efficiency, (iii) is amenable to high-frequency HR, and (iv) can be readily grown in mass culture opens the door to new opportunities for basic research and biotechnological applications using algae. The systematic generation and characterization of targeted KOs and promoter replacements by HR are likely to greatly accelerate gene function analysis and lead to rapid strain improvements for algal biotechnology. In summary, we believe that Nannochloropsis sp. W2J3B and the molecular tools that we have established for this organism will provide the scientific community with a new “green yeast” that has the potential to rapidly advance algal biology.

Materials and Methods

Growth Conditions.

Nannochloropsis sp. W2J3B was grown photoautotrophically in F2N medium, which is identical to F/2 medium (46) except that trace metals, vitamins, and phosphate solutions were added in fivefold higher concentrations, 10 mM Tris⋅HCl (pH 7.6) was added to maintain the pH, and 5 mM NH₄Cl was included as a nitrogen source. All chemicals were obtained from Sigma as reagent-grade. Agar plates were prepared with 0.8% Bacto agar (Difco) in F/2 medium (46) with 50% artificial seawater, except that 2 mM NH₄Cl was used as a nitrogen source. Zeocin, blasticidin S, or hygromycin B, if needed, was added to a final concentration of 2 μg/mL, 50 μg/mL, or 300 μg/mL, respectively. Liquid cultures were generally maintained in F2N medium at a photon flux density of 85 μmol photons⋅m⁻²⋅s⁻¹ and bubbled with CO₂-enriched air (3% CO₂) at 28 °C. Agar plates were maintained at the same light intensity at 26 °C.

Nucleic Acids Used for Transformation.

For PCR we used the Takara LA Taq polymerase. Two overlapping PCR products containing the Sh ble gene were amplified from pTEF1/Zeo (Invitrogen) via primer pair 5′-ATGGCCAAGTTGACCAGTGCCGT-3′ and 5′-TTAGTCCTGCTCCTCGGCCACGAA-3′ and primer pair 5′-ATGGCCAAGTTGACCAGTGCCGT-3′ and 5′-ACAGAAGCTTAGTCCTGCTCCTCGGCCACGAA-3′ (phosphorylated). The resulting products with different lengths were gel-purified (QiaEx II; Qiagen), mixed in equimolar amounts, denatured, and allowed to anneal at room temperature. Similarly, two overlapping products containing the 3′ UTR of the VCP1 gene (GenBank accession no. JF957601) were amplified from genomic DNA of Nannochloropsis sp. W2J3B with primer pair 5′-CTGATCTTGTCCATCTCGTGTGCC-3′ and 5′-GCTTCTGTGGAAGAGCCAGTGGTAG-3′ and primer pair 5′-CTGATCTTGTCCATCTCGTGTGCC-3′ and 5′-GGAAGAGCCAGTGGTAGTAGCAGT-3′. These products were also gel-purified, mixed in equimolar amounts, denatured, and allowed to anneal at room temperature. The products of the two annealing reactions were ligated for 1 h with T4 ligase (Fermentas) to generate the product ble^UTR, which was then gel-purified and amplified with primers 5′-ATGGCCAAGTTGACCAGTGCCGTTCC-3′ (phosphorylated) and 5′-CTGATCTTGTCCATCTCGTGTGCC-3′ and gel-purified. Primers 5′-ACTTAAGAAGTGGTGGTGGTGGTGC-3′ and 5′-ACTTGAGAGAGTGGTGGAGTTGACT-3′ were used to amplify the bidirectional VCP2 promoter (VCP2^Prom; GenBank accession no. JF946490). The VCP2^Prom and ble^UTR products were blunt-ligated, gel-purified, cloned into the pJet1 vector (Fermentas), and transformed into Escherichia coli DH5α cells. After reisolation of plasmids and sequencing, we obtained vectors pJet-C1 and pJet-C2, driving expression of the Sh ble gene from one side or the other of the bidirectional VCP2 promoter. The selection marker cassettes C2 or NT7 were amplified from pJet-C2 with primer pair 5′-ACTTAAGAAGTGGTGGTGGTGGTGC-3′and 5′-CTGATCTTGTCCATCTCGTGTGCC-3′ or 5′-AAGCAAGACGGAACAAGATGGCAC-3′ and 5′-CTGATCTTGTCCATCTCGTGTGCC-3′, respectively. The difference between NT7 and C2 is that C2 contains the entire bidirectional promoter, whereas NT7 contains only the part driving expression of the Sh ble gene.

For the nitrate reductase (GenBank accession no. JF946488) KO construct, we amplified two ∼1-kb parts of the NR gene separated by 242 bp within the genome as recombination flanks with the primers 5′-AGTCGTAGCAGCAGGAATCGACAA-3′ and 5′-GGCACACGAGATGGACAAGATCAGTGGAATAATGAGGCGGACAGGGAA-3′ (NR left flank), and 5′-GTGCCATCTTGTTCCGTCTTGCTTGCGCAAGCCTGAGTACATCATCAA-3′ and 5′-ATGACGGACAAATCCTTACGCTGC-3′ (NR right flank). Flanks were constructed for the nitrite reductase (GenBank accession no. JF946489) gene by amplifying left and right flanks (separated by 793 bp within the genome) with the primers 5′-TGACATGGACCAGCGGCTTAAGTA-3′ and 5′-GTGCCATCTTGTTCCGTCTTGCTTGCCGTATTTGGCATTGGTCTGCAT-3′ (NiR left flank), and 5′-GGCACACGAGATGGACAAGATCAGAGGCCGCATATGACATTCCTCAGA-3′ and 5′-ACGGTGGAAGAGATGGTGAGAGAA-3′ (NiR right flank). Flanks derived from the NR or NiR gene were fused to the NT7 transformation cassette by fusion PCR using the primers 5′-AGTCGTAGCAGCAGGAATCGACAA-3′ and 5′-ATGACGGACAAATCCTTACGCTGC-3′ or 5′-TAACGGGCTACTCACATCCAAGCA-3′ and 5′-AGTATCGCGTGGCAATGGGACATA-3′, respectively. The resulting PCR products (NR-KO and NiR-KO, respectively) were gel-purified before transformation.

Nuclear Transformation of Nannochloropsis sp. W2J3B.

Cells were grown in F2N medium to mid-log phase (∼2 × 10⁶ cells/mL) and washed four times in 384 mM d-sorbitol. Cell concentration was adjusted to 10¹⁰ cells/mL in 384 mM d-sorbitol, and 100 μL cells and 0.1–1 μg DNA were used for each electroporation within an hour. Electroporation was performed with a Bio-Rad Gene Pulser I electroporator in 2-mm cuvettes. The electroporator was adjusted to exponential decay, 2,200 V field strength, 50 μF capacitance, and 500 Ohm shunt resistance. After electroporation, cells were immediately transferred to 15-mL conical Falcon tubes containing 10 mL F/2 medium and incubated in low light overnight. Cells (5 × 10⁸) were plated the next day on F/2 square agar plates (500-cm²) containing 2 μg/mL zeocin. Colonies appeared after 2 wk and could be further processed after 3 wk.

Screening and Analysis of Knockout Mutants.

Initial screen.

Zeocin-resistant colonies obtained by transformation with either NiR-KO or NR-KO were resuspended in 50 μL F/2 medium lacking any nitrogen source, and 6 μL was spotted on agar plates containing 1 mM KNO₃ or 1 mM NH₄Cl as a sole nitrogen source. Many clones started to bleach on plates containing nitrate, indicating starvation for a nutrient, whereas no signs of starvation were visible on plates containing NH₄Cl (Fig. 3).

PCR screen.

Genomic DNA was isolated, and PCR with primers 5′-AGTAGGCGTAGCCTTGGAGTTTGT-3′ and 5′-TCTGAAGCACAAGCGAAGCACT-3′ on NR-KO mutants or with primers 5′-ACGGTGGAAGAGATGGTGAGAGAA-3′ and 5′-AAGCTTAAGAAGGACGGCTCGGTA-3′ on NiR-KO mutants was used to amplify the genomic DNA around the NR or NiR gene, respectively. PCR on genomic DNA isolated from the wild type was used as a control.

Growth test.

Wild type and two clones each of confirmed NR- and NiR-KO mutants (NR1, NR2, NiR1, and NiR2) were grown to mid-log phase in F2N medium containing 1 mM NH₄Cl. Cells were washed three times with 50% artificial seawater by centrifugation (5 min, 3,000 × g) and subsequent resuspension of the cells. Beakers with a clear lid containing 100 mL of F2N medium with no nitrogen source, 1 mM KNO₃, 1 mM NaNO₂, or 1 mM NH₄Cl were inoculated in triplicate with washed cells to a concentration of 4 × 10⁵ cells/mL and allowed to grow under 3% CO₂ atmosphere at 200 μmol photons⋅m⁻²⋅s⁻¹ for 4 d under constant shaking (80 rpm). At this time, wild-type cultures supplemented with 1 mM NH₄Cl reached stationary phase after exhausting the nitrogen source. Cells were counted with an Accuri C6 flow cytometer equipped with an Accuri C6 sampler in duplicate. Growth was estimated as % cells compared with wild-type cultures grown in F2N medium containing 1 mM NH₄Cl.