Increasing lipid productivity in Chlamydomonas by engineering lipid catabolism using the CRISPR-Cas9 system

Currently, most of the attention in renewable energy industry is focused on the development of alternative, sustainable energy sources. Microalgae are a promising feedstock for biofuel production in response to the energy crisis. The use of metabolic engineering to improve yields of biofuel-related lipid components in microalgae, without affecting cell growth, is now a promising approach to develop more sustainable energy sources and to make this approach more economically feasible. Results The CRISPR-Cas9 system was successfully applied to generate a target-speciﬁc knockout of the ELT gene in Chlamydomonas reinhardtii . The target gene encodes an enzyme involved in lipid catabolism, in which the knockout phenotype impacts fatty acid degradation. As a result, the knockout mutants show up to 28.52% increased total lipid accumulation in comparison with the wild-type strain. This is also accompanied by a shift in the fatty acid composition with an increase of up to 27.2% in the C18:1 proportion. These changes do not signiﬁcantly impact cell growth. This study provides useful insights for the improvement of the oleaginous microalgae strain for biodiesel production. The acquired elt mutants showed improved lipid accumulation and productivity without compromising the growth rate.

biodiesel, through transesterification processes, or they can be refined into other fuel constituents [3]. Total lipids and other biomass compositions can be converted into crude oil alternatives through thermochemical processes, such as hydrothermal liquefaction [4]. However, the economic feasibility of using microalgae as a biofuel source is dependent upon making improvements in the entire production process [5]. One of the most influential improvements would be to increase lipid yields [6].
Increasing the lipid yield, without hindering growth, in microalgae is the key to achieving economic feasibility. In comparison to traditional approaches, which are mainly based on identifying high-lipidproducing strains [7], recently developed techniques for metabolic engineering offer alternative methods of improving lipid production in microalgal strains.
In recent years, various strategies for increasing lipid production have been proposed, in particular, random mutagenesis, genome editing, and metabolic pathway engineered to increase lipid biosynthesis, as well as blocking the competing pathways of carbohydrate metabolism and lipid degradation processes. However, as a result of disrupting cellular metabolic pathways, the engineered strains that yield high lipid content often suffer from a reduced growth rate [8,9]. When applied to large-scale economic cultivation, high growth rates and high biomass production are essential [10]. At the same time, engineering advancements that can improve lipid yield, without impacting on growth rate, can significantly decrease production cost and fortify the economic viability of microalgae-derived biofuels [6]. Among the considered strategies, the inhibition of competing pathways has been recognized as a potential alternative to enhance cellular lipid yield. Some have suggested that inhibition of lipid catabolism, more specifically the enzymes that catalyze the release of free fatty acids from lipids, may increase lipid accumulation. In microalgal cells, lipid catabolism facilitates membrane re-organization by providing acyl groups, which are required for remodeling membranes and reorganizing the photosynthetic system [11,12]. However, recent study has shown that unlike the disruption of carbohydrate pools, which play a role in primary carbon storage product in a wide range of microalgae [13], knockdown of genes involved in lipid catabolism may have less of an impact on the growth rate in cell culture [14].
Our study was performed on Chlamydomonas reinhardtii, a model microalgal species with a fully sequenced genome [15]. Microalgae have long been recognized as an ideal source of biofuels and recent technological developments have allowed for genomic manipulation of these microorganisms [16,17]. Using the optimized genetic editing tool, CRISPR-Cas9, here we report the knockout of Cre01.g000300, a gene encoding an enzyme which belongs to the multifunctional ELT family, and we characterize the resulting mutants. The knockout strains show an increase in cellular lipid accumulation during both nutrient-replete and nutrient-deplete cultivating conditions. Our results demonstrate that targeted knockout of this catabolic enzyme in eukaryotic microalgae may be an effective strategy to enhance the outcome of the desired products, while minimize the negative impacts on culture growth rate.

Results
Characterization of the Cre01.g000300 A Basic Local Alignment Search Tool (BLAST) search result indicated that Cre01.g000300 encodes an enzyme that is closely related to a member of the esterase/lipase/thioesterase 1 (ELT1) family of acyltransferases. ELT is composed of two conserved domains, one of which has hydrolase activity, while the other has acyltransferase activity [18]. Each domain has a relatively high sequence similarity compared to organisms (Fig. 1). A GXSXG motif in hydrolase domain and H(X)4D near acyltransferase domain were also found ( Fig. 1), and these are residues for catalytic activity domains in earlier studies [14,19]. Lipid catabolism in microalgae. New Phytologist. 218: 1340-1348). The alpha/beta hydrolase domain-containing proteins are structurally related to various catalytic activities. Acyltransferases catalyze acylation in the lipid synthesis pathway, allowing for transformation from a long-chain acyl-CoA to Coenzyme A [20]. Acyl-CoA molecules are used as substrates for the synthesis of membrane glycolipids, triacylglycerols, and other acylated molecules.
Acyl-CoA can also be directed to beta-oxidative catabolism [21]. Moreover, acyl-CoA can be utilized by a number of lipid metabolic enzymes including endogenous lipase/acyltransferase. Thus, the availability of acyl-CoA plays a fundamental role in determining the quantity and composition of membrane lipids and storage lipids. The manipulation of the gene encoding acyl-CoA can, therefore, be genetically manipulated to improve cellular lipid production.
Targeted knockout of ELT using CRISPR-Cas9 elt knockout mutants were generated in C. reinhardtii using CRISPR-Cas9 technology. The sgRNA used to generate the mutation in the target gene was designed by using CasDesigner (www.rgenome.net).
Gene editing was accompanied with the introduction of hygromycin resistance in the subsequent mutants allowing for the selection of successfully mutated colonies on hygromycin selection agar plates. Mutations were assessed by sequencing with the primers listed in Table 1. Three mutant strains, elt-1, elt-2, and elt-3, were selected by colony PCR and confirmed by Sanger sequencing.
Results show short random DNA fragment insertions with the cassette in these strains (Fig. 1c).
Southern blot analysis of genomic DNA digested with PstI and KpnI was carried out to confirm the number of inserted hygromycin resistance cassettes in the generated mutants. The result showed a single band in all mutants (Fig. 2b), which suggested that each mutant had a single copy of this cassette in their genome. Comparative growth properties of elt knockout mutants Enhancing the desired metabolic products, leading to uncompromised biomass accumulation is an essential factor when applying microalgae in large-scale cultivation for biofuels [22]. To this end, we evaluated the growth and photosynthetic efficiency of the generated mutant strains. The mutant and wild-type (WT) strains were cultured in optimal condition for growth and cell densities were measured every 24 h during the 3 days of growth. The growth of all three elt mutants were faster and reached the maximum density earlier than the WT strain (Fig. 3a). Additionally, elt-2 also had a higher cell density than the WT and the two other mutant strains ( Table 2). The calculated specific growth rate of 0.672, 0.540, 0.686, and 0.572 were recorded for WT, elt-1, elt-2, and elt-3, respectively. The biomass accumulation of elt-2 (  (Fig. 3b) which indicates that the knockout event did not impact the photosynthetic performance of the generated mutants. Table 2 Comparative growth properties of wild-type (WT) and mutant strains.  Table 3 Lipid content and productivity of wild-type (WT) and mutant strains in growth phase and nitrogendeprivation.

Lipid content (mg/g DCW)
Lipid productivity (mg/L/day) nitrogen-deprived phase, the fluorescence signal from the mutants was 19.4% higher than the WT strain ( Fig. 5a). BODIPY visualization also demonstrated an increase in the number and size of lipid droplets in mutant strains (Fig. 5b). Work from other groups has shown that nutrient deprivation causes a disturbance in cell growth resulting in decreased biomass accumulation. Under nitrogen limited conditions, WT C. reinhardtii drives its partial carbon flux from carbohydrate compounds into lipid biosynthesis [24]. Total lipid content was extracted for measurement during the three days of growth phase and three days of nitrogen starved lipid inducing phase. In both phases, all mutants showed a significantly higher value of lipid content than the WT strain, with mutant elt-2 accumulating the highest cellular lipid content (~ 28.52%), followed by elt-1 (~ 26.33%) and elt-3 (~ 25.87%; Fig. 5c). We also analyzed the activity level of enzyme encoded by Cre01.g000300 in both exponential growth phases and in the nitrogen starvation phase. Like other physiological parameters, however, there were slight differences between the strains. Despite the similar genotypes, elt-2 demonstrated the most equivalent growth profile to the WT strain as well as the highest increase in lipid content among three mutants.

Fatty acid composition analysis
The fatty acid (FA) profiles of the WT and mutant strains after three days of nitrogen starvation were analyzed using gas chromatography. The contents of different fatty acid classes are listed in the decrease compared to the WT. No significant changes were observed in the proportion of palmitic acid and saturated stearic acid (Fig. 6). Table 4 The content of fatty acid classes (mg/g) in wild-type (WT) and the representative mutant strain.

Discussion
We have successfully applied CRISPR-Cas9 technology to generate ELT knockout mutants in C.
reinhardtii [26]. This method is advantageous over traditional techniques because generation of knockdown or knockout mutants by random DNA mutagenesis is labor-intensive and time-consuming.
The targeted knockout of ELT was generated by recruiting the RNP complex co-operating with gene for selection by a knock-in with a hygromycin cassette at the site of interest. The insertion of the hygromycin resistance cassette was confirmed by Southern blot analysis (Fig. 2b). This insertion facilitates the selection of successfully mutated colonies on hygromycin selection plates. Such advantages allow this method to significantly reduce the time spent on screening and selecting for putative mutants [31].
There was little difference in the growth pattern among elt strains, and there were few differences in the maximum cell numbers between WT and elt mutants (Fig. 3). Notably, the density of elt mutant cells increased and reached the maximum more rapidly than the wild-type strain under nitrogen deprived conditions. This proves that the growth of elt was not compromised and, therefore, does not affect productivity under nitrogen deprived condition.
In a review, Park et al. (2019) proposed that improving lipid content, while maintaining efficient growth, holds the key to achieving economic feasibility of microalgae-derived biofuel production because biomass, from which lipids are extracted, is a critical factor for lipid productivity [32]. The inhibition of lipid catabolism, particularly of the enzymes catalyzing the release of free fatty acids from lipids, may increase lipid accumulation [14]. Previous studies have found that, unlike the disruption of carbohydrate pools, which play role as primary carbon storage product in wide range of microalgae (Chauton et al., 2013), the knockdown effect in lipid catabolism has a smaller impact on growth rate (Trentacoste et al., 2013). Another report tried to use a similar strategy to support this hypothesis. In a previous study on Chlamydomonas, a phospholipase knockout was generated by CRISPR-Cas9. The resultant mutant showed an increase of up to 62.25% in the total lipid content [33].
The lipase mutant strains showed uncompromised growth, as well as high biomass production, thereby suggesting that it may be possible to obtain a microalgal strain with simultaneous rich oil content and high lipid productivity. Such mutants would be applicable for efficient industrialized production of algal biofuels.
The lipase knockout effects ELT1 (Cre01.g000300) encodes an alpha/beta hydrolase, which shares a highly conserved domain with Thaps3_264297 (E-value of 1.3 × 10 − 46 ; [34]. Identification of the 58/alpha/beta hydrolase domain 5 (CGI-58/ABHD5), a homologue of Thaps3_264297, by comparative genomics showed that it has lipid hydrolase activity along with a signature motif, which is found in the ELT family [34]. In human, a CGI-58 mutation results in Chanarin-Dorfman syndrome, a rare autosomal recessive disease of lipid metabolism with the accumulation of an abnormally large number of cytosolic lipid droplets in various tissues, such as the skin, liver, and leukocytes [35]. The knockout of a similar homolog was obtained in A. thaliana and caused a Chanarin-Dorfman-like phenotype with a high lipid phenotype [36]. The elt knockout mutants showed a higher level of total lipid accumulation than the WT. In both phases of cultivation (Fig. 5c). To confirm the knockout effects on the target gene expression, we performed qRT-PCR. It seems that the mutated ELT gene is still transcribed in the knockout mutants, and regulated in the same manner as it is in the WT. However, proper processing of the mutated transcripts might be disturbed due to the inserted cassette in front of the protospacer adjacent motif (PAM) site in the first exon (Fig. 2C, Additional file 1: Figure S1). Nonetheless, we found that lipase activity in the mutant cells is significantly reduced (Fig. 4). Therefore, these result shows that ELT gene knockout in C. reinhardtii affected lipid catabolism, in which lipid triglycerides are hydrolyzed into glycerol and free fatty acids. Another study that tried the same strategy using RNA interference (RNAi), also shows the same effect. The knockdown of ELT led to a 3.3-fold increase in lipid content in normal growth condition, and a more than 4-fold increase in nutrition-deplete condition, in Thalassiosira pseudonana [14]. In terms of technical approach, compared to RNAi mediated knockdown, which depends on constant transcription of an RNAi coding transgene segment, targeted gene editing is advantageous because it permanently knocks out the expression of a functional gene product.
The fatty acid profile of mutants showed a noticeable shift in the FA proportion, from oleic acid (C18:1) to alpha-linoleic acid (C18:3) with no significant changes in saturated FA components and no changes in the length of carbon chains. A study on acyltransferases showed that this enzyme group determined the fatty acid composition of glycerolipids [37], this study also indicated that acyltransferases took part in the regulation of TAG biosynthesis. In the microalgae, Nannochloropsis oceanica, the changes in acyltransferase activity resulted in changes in C16:0 and C18:1 in the TAG sn-1/sn-3 positions, which could be explained by the difference in the substrate preference of this enzyme [38]. The elt mutants showed an increase in the proportion of unsaturated FA C18:1. Some studies have suggested that the C18:1 component is beneficial for the balance between oxidative stability and low-temperature properties. These characteristics help to promote biodiesel quality [39].
Modifying the fatty acid composition is a potential approach to improve biodiesel properties by enrichment of components with more favorable properties.

Conclusion
In this study, we present a knockout mutant successfully generated using the CRISPR-Cas9 system.
By targeting the multifunctional enzyme ELT, the acquired mutants demonstrated an improvement in lipid accumulation as well as lipid productivity without compromising the growth rate. The expression of the target gene was also tested to confirm the relation between the genotype and phenotype of the mutants. Our results show that disrupting lipid catabolism may be an efficient strategy to obtain microalgae strains with the desired characteristics for biofuel production.

Culture condition
Chlamydomonas reinhardtii strain CC-4349 [40] was cultured in Tris-acetate phosphate (TAP) medium and all data is indicated as the mean ± standard error.

Delivery of the RNP complex and selection marker into the cell
The RNP complex was delivered into microalgae cells following as described by [26] with a minor modification. The cells were harvested at the exponential phase by centrifugation at 3200 rpm for 5  [40]. The target sequences were amplified by PCR using the specific primers (Table 1). The corresponding band was then eluted and purified for Sanger sequencing (Macrogen, South Korea).

Southern blot analysis
The southern blot protocol followed as previously described [40]. The extracted genomic DNA of the identified mutants was digested with PstI and KpnI (Takara, Japan) and electrophoresed on a 0.8% where B is amount (nmole) of glycerol generated within a reaction time, V is sample volume (mL) added to each well. Lipase activity is reported as nmole/min/mL = milliunit/mL.

RNA extraction and qRT-PCR
Total RNA extraction was carried out by using the RNeasy Plant mini kit (QIAGEN) and the concentration of extracted RNA was measured using a Nanodrop spectrophotometer (Thermo Fisher Scientific, USA). The semi-quantified RNA was used as template to synthesize cDNA using reversetranscription master mix (Takara, Japan). The RACK1 gene of the WT strain was used for normalization and quantitative reverse transcriptase PCR (qRT-PCR) was performed with the specific primers listed in Table 1.

Total lipid extraction
Cells were harvested at two different stages: 1) cells at exponential phase in TAP medium and, 2) cells at stationary phase in nitrogen deprived TAP medium. Total lipid was extracted as previously described [41] with some modification. The harvested cells were resuspended in distilled water and then diluted with 15 mL of 2:1 (v/v) methanol/chloroform solution. The cells were incubated at room temperature while shaking at 120 rpm for 12 h. Next, 5 mL of distilled water and 5 mL of chloroform were added to the mixture, which was then vigorously vortexed. Two different phases were separated performed to visualize the accumulation of lipid droplets. The process was performed as previously described [42] and cells were observed under a fluorescence microscope (Nikon Y-TV55, Japan).

Fatty acid profile analysis
The FAME profile was analyzed using gas chromatography as previously described [43]       Percentage of total lipid content of the mutants and WT strains in the nitrogen replete and depleted conditions. The data indicates the mean ± standard error of the three biological replicates of each strain. Statistical analyses were performed using Student t-test, *P < 0.05.

Figure 6
Analysis of fatty acid profile in mutants and WT strains after three days of nitrogen starvation. The data are indicated as the mean ± standard error of the three biological replicates of each strain. Statistical analyses were performed using Student t-test, *P < 0.05.

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download. suppl_fig.docx