The pht4;1-3 mutant line contains a loss of function allele in the Fatty Acid Desaturase 7 gene caused by a remnant inactivated selection marker—a cautionary tale

A striking and unexpected biochemical phenotype was found in an insertion mutant line in the model plant Arabidopsis thaliana. One of two investigated insertion mutant lines in the gene encoding the phosphate transporter PHT4;1 demonstrated a prominent loss of trienoic fatty acids, whereas the other insertion line was indistinguishable from wild type in this aspect. We demonstrate that the loss of trienoic fatty acids was due to a remnant inactive negative selection marker gene in this particular transposon tagged line, pht4;1-3. This constitutes a cautionary tale that warns of the importance to confirm the loss of this type of selection markers and the importance of verifying the relationship between a phenotype and genotype by more than one independent mutant line or alternatively genetic complementation.


INTRODUCTION
Reverse genetics is a powerful tool in plant biology to establishing causal relationships between genotype and phenotype. There are numerous mutagenesis strategies, both targeted (Felippes, Wang & Weigel, 2012;Gaj, Gersbach & Barbas, 2013;Yin, Gao & Qiu, 2017) and untargeted (Alonso et al., 2003;Henikoff & Comai, 2003;McCallum et al., 2000) for generating the starting material for reverse genetics studies. Once the genetic location of a mutation has been confirmed, and the corresponding gene product is shown to be affected, the work in connecting genotype to phenotype can begin (O'Malley & Ecker, 2010). But importantly, the plant line must also be cleared of off-target mutations that may have arisen during mutagenesis. The common recommendation to circumvent this problem is to use several independent mutant lines when inferring phenotype from genotype, or alternatively, the mutant line can be functionally complemented with the wild-type copy of the gene of interest.
Under phosphate limiting conditions, plants can exchange a substantial part of the plasma membrane phosphoglycerolipids for galactolipids (Andersson et al., 2003). In light of these previous findings, we investigated whether pht4;1 mutants are affected in leaf lipid composition. To our surprise we identified a second mutation in the pht4;1-3 line located in the FAD7 locus that rendered plants defective in the synthesis of trienoic fatty acids. We propose that this phenotype, not connected to PHT4;1 per se, could potentially influence properties previously ascribed to absence of the phosphate transporter.
Plant DNA was extracted by heating a small leaf piece (approximately 10 mm 2 ) in 750 µL extraction buffer (200 mM Tris-HCl pH 7.5, 250 mM NaCl, 25 mM EDTA, 0.5% SDS) for 5 min at 95 • C. The solution was left in room temp for a few hours before the DNA was precipitated through the addition of 750 µl 2-propanol. The supernatant was removed after centrifugation and the pellet washed once in 70% ethanol. Samples were dried in room temperature and the pellet containing DNA was reconstituted in 150 µL H 2 O.

Extraction and quantification of lipids
The galactolipid species composition was analyzed as previously described (Nilsson et al., 2014) and total lipids fatty acid methyl esters was analyzed by GC-MS after direct transmethylation of leaf material. Leaf pieces were placed in boiling 2-propanol and supplemented with a known amount of di-nonadecanoyl phosphatidylcholine as internal standard. The samples were dried under a stream of nitrogen and transmethylated in 0.5 M sodium methoxide in dry methanol. The fatty acid methyl esters were extracted into heptane and analyzed by GC-MS as described (Najm et al., 2017).

RESULTS AND DISCUSSION
Since PHT4;1 has been suggested to regulate phosphate transport inside the chloroplast (Karlsson et al., 2015), and since phosphate starvation is known to trigger exchange of phospholipids for galactolipids (Andersson et al., 2003), we profiled membrane lipids in the two PHT4;1 mutant alleles pht4;1-2 and pht4;1-3, and compared them to the parent wild type Ler (Fig. 1). While there were no significant differences in the amount of the major membrane lipid classes, the pht4;1-3 line displayed an altered species composition of several membrane lipids. This was clear for both major thylakoid lipid classes MGDG and DGDG (Fig. 1A), which demonstrated a substantial reduction in the highly unsaturated species 34:6 and 36:6. The pht4;1-2 mutant line, on the other hand, did not differ from   wild type Ler in its membrane lipid composition. Analysis of fatty acid methyl esters from a total lipid extract of the pht4:1-3 mutant confirmed a decrease in linolenic-(18:3) and hexadecatrienoic (16:3) acids concomitant with an increase in the less saturated linoleic-(18:2) and hexadecadienoic (16:2) acids (Fig. 1C, Table 2). In particular, 16:3 was almost completely absent from the mutant line. This phenotype is very similar to that reported for mutants of the FAD7 gene (Browse, McCourt & Somerville, 1986;Iba et al., 1993), leading us to suspect that a second site mutation affecting FAD7 might be present in the pht4;1-3 line.
To investigate this, we performed PCR on genomic DNA obtained from pht4;1-3 and wild type Ler using primers spanning different regions of the FAD7 gene ( Figs. 2A  and 2B). A product containing the first exon and the 5 UTR of FAD7 was found to be absent in reactions from the mutant line (Fig. 2B). Further PCR with extended elongation time revealed an approximately 2.5 kbp insertion to be present in pht4;1-3 in this region (Fig. 2C). Sequencing (sequence available in Data S1) of the insertion revealed it to contain an inactivated copy of the indole acetic acid hydrolase (IAAH) negative selection marker   Table 2 Fatty acid profiles of pht4;1-lines. Glycerolipid fatty acids were analyzed by GC-MS from leaf pieces obtained from the indicated lines.
The Ds transposable element of the pht4;1-3 starter line DsG6 has previously been mapped to the FAD7 gene (Parinov et al., 1999). It thus appears that a part of the Ds element has remained in place in the pht4;1-3 line due to the inability to use the negative selection based on IAAH. Rearrangements of the Ds element after transposase activation and subsequent inability to select against progeny carrying the starter line cassette have previously been reported (Parinov et al., 1999).
To confirm that that the altered lipid profile of pht4;1-3 is indeed caused by the second insert in FAD7, a genetic test was conducted. The pht4;1-3 line was crossed to fad7-1 or wild type Ler and fatty acid profiles of the resulting F 1 and parental lines were analyzed (Table 2). This clearly shows that the fatty acid desaturation defect in pht4;1-3 is recessive and that the F 1 of pht4;1-3 and fad7-1 show an identical phenotype to their parents, supporting the notion of an inactivation of FAD7 in the pht4;1-3 line. Finally, pht4;1-3 was backcrossed to Ler and the resulting F 2 plants were genotyped using PCR primers for the insertion in FAD7 and the previously described insertion in pht4;1-3. Selected F 2 plants homozygous for the two different insertions were tested for acyl lipid fatty acid composition. This fully supported the inferred linkage between loss of trienoic fatty acids and the FAD7 insertion.
The pht4;1-3 mutant line has, as far as we could find, only been used in two publications (Karlsson et al., 2015;Wang et al., 2011). Careful scrutiny of these reports shows that in fact all but one experiments are based on the use of at least one additional allele of pht4;1. The one experiment only including the pht4;1-3 allele and wild type describes the spacing of thylakoids membranes in grana stacks (Karlsson et al., 2015). Since this is a phenotype that could at least partially be attributed to the physicochemical properties of the thylakoid lipids, it seems critical that this result is confirmed with a bona fide pht4;1 mutant.

CONCLUSIONS
To conclude, this study underlines the importance that every inferred phenotype-genotype relationship is confirmed by at least two independent genetic lines of evidence, i.e., through the use of several independent knock-out or silenced mutant lines, or by genetic complementation.