GPAT3

Fig. 29: Ali-KO and Ali-ADH-KO did not show an effect on phagocytosis. The same clones of the KOs as that investigated for TAG and SE formation were used for phagocytic analysis in the absence of PA (A) or presence of PA (B). Phagocytosis assay was performed and presented as in Fig. 24. Assays in both media were repeated at least three times.

2.4 Characterization of triacylglycerol biosynthetic enzymes found in the

Tab. 6: Putative orthologues to human GPATs in Dictyostelium

Human isoforms aa UniPROT ID Homologues in Dictyostelium

GPAT1 828 Q9HCL2 DDB0305750 (26% (140/524) identity, score 209) GPAT2 795 Q6NUI2 DDB0305750 (19% (43/218) identity, score 42) GPAT3 434 Q53EU6 DDB0235400 (44% (142/341) identity, score 276)

DDB0307026 (26% (33/125) identity, score 43.9) DDB0308174 (36% (23/63) identity, score 34.3) GPAT4 456 Q86UL3 DDB0235400 (36% (137/378) identity, score 241)

DDB0307026 (24% (52/214) identity, score 44.7)

Fig. 30: Domain description of GPAT3. SMART analysis predicted three transmembrane domains indicated in dark grey and a conserved PlsC domain marked in orange.

All mammalian GPATs and AGPATs proteins have 4 conserved motifs within the phosphate acyltransferase domains (Takeuchi and Reue, 2009). Motifs I and IV are responsible for catalytic function, while motifs II and III are important for glycerol-3-phosphate binding. The key amino acids in the four motifs are conserved in all human enzymes and Dictyostelium GPAT3 (Tab. 7). Again, comparing only these 4 conserved motifs with human GPATs, Dictyostelium GPAT3 shows better conservation with human GPAT3 and GPAT4 than with human GPAT1 and GPAT2.

Tab. 7: Comparison of acyltransferase motifs in human GPATs and Dictyostelium GPAT3

The consensus sequences of Dictyostelium GPAT3 (DdGPAT3) and human GPATs (hGPATs) are highlighted in green. Underlined amino acids are important for phosphate acyltransferase activity (Takeuchi and Reue, 2009).

Residues histidine (H) and asparate (D) in motif I, phenylalanine (F) and glycine (G) in motif III and proline (P) in motif IV are important for catalysis, while phenylalanine (F) and arginine (R) in motif II, glutamate (E) in motif III are crucial for binding glycerol-3-phosphate.

Motif I Motif II Motif III Motif IV

hGPAT1 227 LPVHRSHIDYLLL 274 FFIRRR 312 IFLEGTRSR 347 ILIIPVGISYG hGPAT2 202 LSTHKTLLDGILL 249 LFLPPE 287 IFLEEPPGA 323 ALLVPVAVTYD hGPAT3 226 VANHTSPIDVLIL 270 WFERSE 300 IFPEGTCIN 324 GTIHPVAIKYN hGPAT4 226 VANHTSPIDVLIL 289 WFERSE 319 IFPEGTCIN 343 ATVYPVAIKYD DdGPAT3 222 VANHTTVMDVVVL 266 WFDRAE 296 IFPEGVCVN 321 VIIYPVAIKYN

2.4.1.2 GFP-tagged GPAT3 is a LD-associated protein

The expression of GFP-tagged GPAT3 protein was confirmed using Western-blot (Fig. 31).

Similar with Ldp proteins, GFP-GPAT3 (GFP-tag at the N-terminus of GPAT3) is slightly larger than GPAT3-GFP because of the longer linker region between GFP and GPAT3 in GFP-GPAT3 fusion construct. It is worth to mention, that the weak expression of GPAT3-GFP resulted from the strong heterogeneity of expression levels within this cell population. There are only small amounts (below 20% in clone 3-23) of green cells of the whole cell population observed microscopically.

Fig. 31: Western-blot of GFP-tagged GPAT3 proteins. Total cellular proteins were extracted from 2 x10⁶ GFP-GPAT3 (#937) clones 1-2, 1-8, 2-7 and 2-21 as well as 4 x10⁶ GPAT3-GFP (#938) clones 3-23 and 4-2, so that GPAT3-GFP expression was visible in response to anti-GFP antibody in the Western-blot. Anti-severin served as loading control. AX2 was used as negative control for anti-GFP antibody. Protein marker components (M) are shown on the left.

Similar to its human homologues GPAT3 and 4, both N-terminal and C-terminal GFP-tagged GPAT3 from Dictyostelium was detected in the ER under normal cultivation conditions (Fig.

32A). Rather surprisingly, most of the protein translocates to the surface of LDs when the cells grew in medium supplemented with PA (Fig. 32B), which has not been reported in mammals before. The ring structure around LDs, typical for LD-associated proteins, was clearly visible and confirmed the proteomic result (Tab. 4).

Fig. 32: Cellular localization of GFP-tagged GPAT3 proteins. For confocal microscopy the following clones were used: GFP-GPAT3 (#937) 1-8 and GPAT3-GFP (#938) 3-23. GFP-tagged GPAT3 cells grown in normal medium (A) or in PA-containing medium (B) were fixed and incubated with anti-PDI antibody or LD540. Scale bar, 5 µm.

2.4.1.3 Generation of GPAT3-KO cells

The metabolic function of the GPAT3 protein was studied in GPAT3-KO Dictyostelium cells.

The floxed-Bsr cassette was inserted into the first exon of the GPAT3 gene in opposite direction with respect to the transcription of GPAT3 (Fig. 33A). KO cells were verified by

PCR according to the same principle specified previously for the Ldp-KO using the primer upstream of the start codon together with the Bsr primer and in-gene primer (Fig. 33 A and B).

Fig. 33: GPAT3-KO construction and its verification. A) Schematic view of GPAT3-KO construct (plasmid

#931) as present in the genomic region and the primers used for KO confirmation. Exons are coloured bright blue, the Bsr cassette is red. The amplification direction of each primer is indicated with arrow. I1-5 indicates the 5 introns. Thin lines indicate genome regions flanking the target gene. B) GPAT3-KO verification using PCR.

Fragments were amplified on gDNA from WT cells and 5 GPAT3-KO independent clones 1-19, 1-21, 1-35, 1-40 and 2-6. The 1.3 kb products were amplified using primers #456 Bsr-middle and #775 5’-GPAT3 KO in GPAT3-KO cells, whereas no amplification occurred in AX2 cells lacking the Bsr cassette. Another primer combination of #774 3’-GPAT3 new and #775 5’-GAPT3 KO generated a fragment of 2.5 kb in the GPAT3-KO and 0.9 kb in AX2 cells. The gene amplification of TRX using primers #216 5’-TRX and #217 3’-TRX serves as control for template amounts. DNA markers (M) are shown on the left of the images.

2.4.1.4 GPAT3 is required for TAG synthesis and its absence decreases TAG content and inhibits the LD formation

Since the GPAT3 enzyme is the initial acyltransferase of the glycerol-3-phosphate pathway of TAG biosynthesis, TAG content was measured by the enzymatic assay and lipid separation by TLC (Fig. 34). Both assays were performed on two independent GPAT3-KO clones in normal or PA-containing medium. TAG formation was decreased by 80% in both GPAT3-KO clones with PA-treatment.

C1-BODIPY-C12 is a fluorophore-labelled fatty acid analogue and can be incorporated into LDs of Dictyostelium cells (von Lohneysen et al., 2003). PA-treated GPAT3-KO and WT cells were pre-incubated in C1-BODIPY-C12 medium for 15 min prior to fixation. Confocal images showed that LDs were very rarely seen in GPAT3-KO cells in contrast to a profusion of LDs in AX2 (Fig. 35). The reduced LD number is in the same order of magnitude as the TAG content as estimated from TAG and TLC analysis.

Fig. 35: Greatly reduced numbers and size of LDs in PA-treated GPAT3-KO cells. LDs were visualized by incorporation of C1-BODIPY-C12 and observed by confocal microscopy in fixed cells. Scale bar, 10 µm.

Fig. 34: GPAT3-KO inhibits TAG formation in Dictyostelium. The TAG content was measured in 2 independent GPAT3-KOs, 1-21 and 2-6, as well as AX2 cells in normal medium (PA-) or in PA-containing medium (PA+) by the enzymatic TAG assay (bar diagram) and TLC lipid separation (lower panel). The enzymatic TAG assay was performed 3 times and the TAG contents are presented as mean ± SD. **

signifies p<0.01 in a pairwise student t-test comparing the TAG values between PA-treated AX2 cells and GPAT3-KO clones. TLC lipid separation and lipid standards (S) are described in Fig. 14A. The arrow indicates the position of TAG on the TLC plate. The two assays are shown correspondingly, thus the descriptions of cell strains and growth conditions at the bottom of the figure apply to the both assays.

2.4.1.5 Loss of GPAT3 promotes the particle uptake but not plaque formation on bacterial lawn

The same GPAT3-KO clones were used to perform phagocytosis assay and to measure plaque diameter. Lack of GPAT3 stimulated phagocytosis of particles in both normal and PA-supplemented medium (Fig. 36). The phagocytic rate of yeast particle uptake in both GPAT3-KO mutants increased by 21% within the first 90 min, then decreased gradually to reach WT level at 120 min. In PA-containing medium, GPAT3-KO showed an even larger effect on phagocytosis. The phagocytic rate of both KO mutants exceeded that of WT by 46%. The increased phagocytic rate suggests a negative correlation between TAG level and particle uptake. This negative correlation is consistent with the observation of LD formation.

Dictyostelium cells induced by adding PA into the growth medium decreases phagocytic activity by about 50%, which is represented by the difference of phagocytic values from AX2 cells in Fig. 36 panel A and B.

Fig. 36: Increased phagocytic rate in GPAT3-KO. Phagocytosis assay with two GPAT3-KO clones, 1-21 and 2-6, as well as AX2 was performed in normal medium (A) or in medium with PA (B). The assays were repeated for 3 times for each condition and presented as in Fig. 24.

Although the phagocytic rate of particles in medium is strongly elevated, both GPAT3-KO clones did not significantly alter the growth properties on a bacterial lawn (Fig. 37). As mentioned before (2.3.1.6), phagocytosis is not the only factor affecting the ability of Dictyostelium to grow on bacterial lawn, cell motility and division also contribute but these parameters were not tested further.

2.4.1.6 GFP-tagged GPAT3 restores the dramatic reduction of TAG content in GPAT3-KO strain

In order to examine whether GPAT3 overexpression rescues the markedly decreased TAG formation in PA-containing medium caused by the GPAT3-KO, both C-terminally and N-terminally GFP-fused GPAT3 enzymes (plasmids #937 and #838) were transformed into the KO 1-21 strain. The resulting cell lines were named GFP-GPAT3 rescue and GPAT3-GFP rescue respectively. A Western-blot showed the expression levels of tagged GPAT3 molecules in both rescued cell lines (Fig. 38). Like GFP-tagged GPAT3 in WT background (2.4.1.2), GPAT3-GFP shows rather heterogeneous expression in the GPAT3-KO background.

Clone 2-2 showed the highest fraction of green cells as compared to the other clones, and about 40% cells of the whole cell population were green as judged by microscopy.

0 20 40 60 80 100 120 140

3. day 4. day

diameter of plaque (%WT)

Growing time on E.coli B/2 lawn AX2

GPAT-KO 1-21 GPAT-KO 2-6

Fig. 37: GPAT3-KO showing normal growth on bacterial lawn.

The same GPAT3-KO clones 1-21 and 2-6 as used above were tested for plaque formation on E.coli B/2 spread on an SM-agar plate. At the 3rd and 4th day of cultivation, the diameters of plaques were measured. This assay was conducted for 3 times and values are shown as mean ± SD, relative to the AX2 wildtype. There is no statistical difference (p>0.05) between WT and GPAT3-KO.

Fig. 38: A Western-blot showing the expression level of GFP-tagged GPAT3 proteins in GPAT3-KO rescue Dictyostelium cells. Protein samples from each strain were extracted from 2x10⁶ cells. The overexpression level was demonstrated by anti-GFP antibody. An antibody specific for vacuolin was used as a loading control.

Molecular weights of protein marker (M) are indicated on the left.

GFP-GPAT3 rescue clone 1-12-6 and GPAT3-GFP rescue clone 2-2 were selected for further examination of TAG formation. Both GFP-tagged GPAT3 enzymes had the same cellular localization as in WT background (data not shown). Enzymatic TAG assay and TLC lipid separation were performed to quantify the TAG content in WT, GPAT3-KO and its two rescue cell lines (Fig. 39). Both GFP-tagged GPAT3 construct could restore the TAG content to normal level, similar to WT in PA-supplemented medium. Surprisingly, even the GPAT3-GFP rescue clone 2-2 with only weak overexpression could restore the TAG content to normal values, indicating that the level of the endogenous protein was reached.