A quantitative assay measuring the function of lipase maturation factor 1.

Newly synthesized lipoprotein lipase (LPL) and related members of the lipase gene family require an endoplasmic reticulum maturation factor for attainment of enzyme activity. This factor has been identified as lipase maturation factor 1 (Lmf1), and mutations affecting its function and/or expression result in combined lipase deficiency (cld) and hypertriglyceridemia. To assess the functional impact of Lmf1 sequence variations, both naturally occurring and induced, we report the development of a cell-based assay using LPL activity as a quantitative reporter of Lmf1 function. The assay uses a cell line homozygous for the cld mutation, which renders endogenous Lmf1 nonfunctional. LPL transfected into the mutant cld cell line fails to attain activity; however, cotransfection of LPL with wild-type Lmf1 restores its ability to support normal lipase maturation. In this report, we describe optimized conditions that ensure the detection of a complete range of Lmf1 function (full, partial, or complete loss of function) using LPL activity as the quantitative reporter. To illustrate the dynamic range of the assay, we tested several novel mutations in mouse Lmf1. Our results demonstrate the ability of the assay to detect and analyze Lmf1 mutations having a wide range of effects on Lmf1 function and protein expression.

Lipase maturation factor 1 (Lmf1) has been identifi ed as the protein affected by the combined lipase defi ciency ( cld ) mutation in the mouse ( 1 ). Mice homozygous for cld exhibit severe chylomicronemia due to the defi ciency of lipoprotein lipase (LPL) activity ( 2 ). LPL is synthesized and secreted by a wide variety of cell types, and it repre-Abbreviations: cld , combined lipase defi ciency mutation; hLmf1 wt , full length human Lmf1 (567 amino acid); hLmf1 Y439X , human Lmf1 truncated at position 439; Lmf1, lipase maturation factor 1; LPL, lipoprotein lipase; mLmf1 cld , mouse Lmf1 truncated at position 361; mLmf1 G181E , mouse Lmf1 containing a G/E substitution at position 181; mLmf1 N430A , mouse Lmf1 containing a N/A substitution at position 430; mLmf1 wt , full length mouse Lmf1 (574 amino acid); SEAP, secreted form of alkaline phosphatase; wt , wild type. 1 To whom correspondence should be addressed. e-mail: markdool@ucla.edu and mpeterfy@ucla.edu Roche). An LPL/SEAP plasmid master mix was prepared and used throughout all assay experiments performed. The master mix contained LPL and SEAP at a concentration of 0.4 and 0.1 mg/ml, respectively. All expression vector plasmids used in these studies were prepared using the EndoFree Maxi kit (Qiagen) according to the manufacturer's instructions. Diluted plasmid solutions were quantitated using a NanoDrop 2000 spectrophotometer (Thermo Scientifi c).

Cotransfection and cell harvesting
Each cotransfection experiment contained an Lmf1 target sequence and a mixture of two reporter constructs (LPL/SEAP master mix; see above). Transfection of cld mutant cells was achieved using the Effectene® transfection reagent (Qiagen) at a DNA-reagent ratio of 1:10. Each well of a 12-well plate was transfected with 0-20 ng of the Lmf1 target sequence and a mixture of 0.8 µg and 0.2 µg of LPL and SEAP, respectively; transfection was performed according to the manufacturer's instructions. At 24 h posttransfection, a sample of medium was taken for SEAP activity measurement, and then heparin was added to a fi nal concentration of 10 U/ml. At 48 h posttransfection, fresh heparin-containing medium was added ( t = 0) and at various times later (30 min-8 h), samples were removed for LPL activity measurements. At the end of the experiment, cells were washed twice with PBS and lysed in detergent-containing buffer (0.2% sodium deoxycholate, 10 U/ml heparin, 10 mM Tris-HCl, pH 7.5). After sonication and centrifugation, supernatants from the resulting lysates and medium samples were stored at Ϫ 80°C until use.

Detection and quantitation
SEAP activity was measured using the SEAP Reporter Assay kit (InvivoGen) according to the manufacturer's instructions. SEAP provides the basis of an assay to assess Lmf1 function. Theoretically, the ability of target Lmf1 sequences to restore LPL activity in cld cells can be used as a measure of Lmf1 function. We provided proof of principle for such an assay by showing that a nonsense mutation (Y439×) in a human Lmf1 expression construct failed to restore LPL activity in cld mutant cells ( 1 ). However, Y439× represents a complete loss-of-function mutation that is qualitative in nature (all or none) and thus relatively easy to detect. Our original assay did not have the quantitative dynamic range to detect mutations that cause only a limited loss of function, or those that only partially affect Lmf1 protein expression. In this report, we describe an optimized version of our original assay that has the ability to detect Lmf1 mutations that affect function or protein levels over a wide dynamic range. We demonstrate the utility of the assay using several induced mutations in mouse Lmf1.

The cld mutant cell line
The cld mutant cell line used in the assay has been described previously ( 7 ). It represents a hepatocyte-derived cell line derived from 16-to 18-day-old fetal cld / cld mice immortalized by transfection with the large SV40 T antigen [pSV3neo ( 8 )], and it is available upon request. The cells lines were maintained at split ratios of 1:20 in DMEM-10% FBS and included passage numbers 25-53. For the Lmf1 assay, a total of 1.8 × 10 5 cells were distributed in a 12-well plate (Falcon), each fl at-bottomed well containing 0.8 ml DMEM-10% FBS and having a surface area of 3.8 cm 2 . Transfection was initiated 24 h after plating when cells reached р 90% confl uence.

Expression constructs and Lmf1 target sequences
The LPL expression construct represents a full-length human LPL cDNA subcloned into the pcDNA6/V5-His expression vector (Invitrogen). The C terminus of the expressed LPL protein is fused with a V5 epitope tag as described ( 9 ).
All human and mouse Lmf1 sequences were subcloned into the pcDNA3.1 expression vector (Invitrogen); the expressed Lmf1 proteins contained an N-terminal c-myc epitope tag ( 1 ). The following human LMF1 target sequences were used: wildtype Lmf1 (hLmf1 wt ) encoding the full-length 567 amino acid protein, and the Y439× nonsense mutation, removing 127 amino acids from the C terminus of Lmf1 (hLmf1 Y439× ) ( 1 ). The following mouse Lmf1 target sequences were used: wild-type Lmf1 (mLmf1 wt ) encoding the full-length 574 amino acid protein; the naturally occurring cld mutation, removing 214 residues from the C terminus of the Lmf1 protein (mLmf1 cld ) ( 1 ); an induced missense mutation substituting Ala for Asn at amino acid 430 of Lmf1 that is part of the single Asn-Ala-Ser glycosylation site in wild-type Lmf1 (mLmf1 N430A ); and a missense mutation causing a Glu for Gly substitution at residue 181 of the Lmf1 protein (mLmf1 G181E ). The G181E mutation was identifi ed in a clonal library of ENUmutagenized ES cells maintained at the University of North Carolina Mouse Mutant Regional Resource Center. The Lmf1 open reading frame was PCR-amplifi ed from ENU-mutagenized ES cell DNA and was scanned for mutations using denaturing HPLC. The G181E mutation is due to a G-to-A substitution in exon 4 of the mouse Lmf1 gene.
To normalize for transfection effi ciency, we used a secreted human placental alkaline-phosphatase (SEAP) reporter gene subcloned into the pM1 expression vector (X-extremeGENE®, Fig. 1. Overview of the Lmf1 functional assay. The fi rst step of the assay is cotransfection of the cld mutant cell line with an Lmf1 target sequence and two reporter constructs, LPL and SEAP. The second step quantitates LPL activity, Lmf1 protein levels, and SEAP activity, and the latter used to correct for differences in transfection effi ciency. It is essential that the Lmf1 target sequence is expressed at levels that fall within the sensitive range of LPL activity (see text and Figs. 2 and 3 ).

RESULTS AND DISCUSSION
Our assay is based on the ability of transfected Lmf1 to restore LPL activity in cld mutant cells; thus, LPL activity becomes a reporter of Lmf1 function. Because these cells do not endogenously synthesize LPL, they are cotransfected with LPL along with a second reporter construct, SEAP ( Fig. 1 ). LPL restoration by an Lmf1 target sequence is compared with a range of values whose maximum is defi ned by wild-type Lmf1 expressing full function as an LPL maturation factor and a minimum defi ned by Lmf1 carrying a loss-of-function mutation. Thus, an Lmf1 target sequence to be tested requires comparison to both wildtype and mutant Lmf1 controls, each representing separate cotransfection experiments. To account for differences in transfection effi ciency occurring in separate experiments, LPL activity is normalized to SEAP activity, and this value is used as the readout of Lmf1 function. In a similar manner, Lmf1 protein levels are compared between the Lmf1 target sequence and the wild-type control ( Fig. 1 ). In such a way, the normalized LPL activity for any given Lmf1 target sequence can be assessed relative to wild-type Lmf1 function and protein level, as shown below.

Determining the sensitive range of the assay
The sensitive range of the assay ensures the detection of a wide range of Lmf1 function (full, partial, or complete loss of function) using LPL activity as a quantitative reporter. catalyzes the hydrolysis of p Nitrophenyl phosphate producing a yellow end product read spectrophotometrically at 405 nm. For each assay, OD 405nm readings were taken over 30-240 min to establish the linear range for the assay, which was then used for normalization of LPL activity levels.
LPL activity was measured using a lecithin-stabilized radiolabeled triolein (glycerol tri[9,10(n)- 3 H]oleate) substrate as described ( 10 ). Each mU of LPL activity is equivalent to 1 nmol free fatty acid released per minute. Secreted LPL activity is given as mU/ml of medium, normalized to SEAP activity per ml (i.e, mU/ SEAP). Because SEAP activity is based on OD 405nm measurements, normalized LPL activity is a relative value and thus given as an arbitrary number (see Figs. 2 and 3 ).
For detection of Lmf1 protein, a biotinylated rabbit anti-cmyc polyclonal antibody (Gene Tex) was used at a dilution of 1:2,500; LPL protein was detected using a biotinylated rabbit anti-LPL antibody (a kind gift from André Bensadoun) at 1:2,000. Biotinylated antibodies were detected using horseradish peroxidase-conjugated Streptavidin (Invitrogen) at 1:50,000 and visualized using a chemiluminescent substrate (ECL+®, GE Healthcare). Quantitation of Lmf1 band intensity was performed using the software NIH Image 1.63. The cld mutant cell line was transfected with increasing amounts of a human wild-type Lmf1 expression plasmid (hLmf1 wt ; see Methods and Materials); at 24 h posttransfection, SEAP activity was measured and heparin added to medium. Eight hours postheparin, secreted LPL activity (mU/ml) was measured in the medium. Initially, changes in LPL activity (normalized to SEAP) were dependent on the amount of Lmf1 plasmid transfected but then plateaued at ~10 ng; the sensitive range of the assay is illustrated. The normalized LPL activity value at 5 ng Lmf1 plasmid was arbitrarily set to 100. B: Mutant cld cells transfected with increasing amounts of Lmf1 plasmid (shown in A) were subjected to Western blot analysis to assess Lmf1 and LPL protein levels. The signals shown were obtained from equal loads of SEAP activity. C: Mutant cld cells were transfected with either an hLmf1 wt or hLmf1 Y439× expression construct within the sensitive range of the assay (5 ng Lmf1 plasmid). SEAP was measured 24 h posttransfection and heparin added immediately after (t = 0). Medium was collected through 120 min and assayed for LPL activity, which was then normalized to SEAP activity. Normalized LPL activity levels in A and C represent the average of three individual transfections (n = 3) ± SEM.  Fig. 2C , the secretion curves of LPL activity arising from hLmf1 wt and hLmf1 Y439× target sequences confi rms Y439× as a mutation causing loss of function. In general, we recommend that the protein level of an Lmf1 target sequence should be less than or equal to a wild-type Lmf1 control whose sensitive range has been predetermined (see below).

Analyzing the function of Lmf1 mutant proteins
To illustrate the dynamic range of the assay, we tested several novel mutations in mouse Lmf1. These included two induced missense mutations, G181E and N430A, and the naturally occurring cld truncation ( 1 ). N430A removes the single glycosylation site in mouse Lmf1, while G181E is adjacent to a putative transmembrane-spanning region. Both Asn and Gly at positions 430 and 181 are conserved amino acids; however, glycosylation occurs only in mouse Lmf1 (data not shown), suggesting that loss of N-linked glycosylation by N430A may be unimportant. The effect of G181E on Lmf1 function is unknown, but its juxtaposition near a transmembrane domain could disrupt proper membrane insertion, affecting Lmf1 membrane topology.
In an initial comparison, equivalent plasmid amounts were used for all transfections (5 ng) based on a wild-type Lmf1 control predetermined to be within the sensitive range of the assay ( Fig. 3A ). The secretion curves of the three Lmf1 mutant plasmids compared with the wild-type control are shown in Fig. 4A , whose slopes are replotted in Fig. 4B as LPL activity secreted per minute. While the In this range, the quantitative reporter (LPL activity) responds proportionally to changes in Lmf1 protein levels (or function), as shown in Fig. 2A . When increasing amounts (0-20 ng) of wild-type human Lmf1 (hLmf1 wt ) plasmid are cotransfected with a constant amount (0.8 µg) of LPL plasmid, levels of LPL activity are directly dependent on transfected Lmf1 plasmid levels until they become saturating at ~10 ng ( Fig. 2A ). Notably, increasing the amount of hLmf1 wt plasmid in the transfection increased Lmf1 protein levels proportionally ( Fig. 2B ). The assay is most sensitive when expressed Lmf1 protein levels fall within the linear portion of the LPL activity curve (highlighted in Fig. 2A ). A similar sensitive range was determined for mouse Lmf1 ( Fig. 3 ), showing that both human and mouse Lmf1 target sequences can be tested successfully in the assay. As expected, LPL protein levels are relatively unchanged, because the amount of LPL plasmid used in the transfection is held constant ( Fig. 3B ). Figures 2A and 3A indicate that the ratio of target Lmf1 to LPL plasmid should be chosen within the sensitive range of LPL activity, ensuring that the full dynamic range of the assay is utilized. Such a ratio was used to compare hLmf1 wt to hLmf1 Y439× , the latter containing a C-terminal deletion causing cld and hypertriglyceridemia in a human patient ( 1 ). In this case, 5 ng of both Lmf1 plasmids were used while the amount of LPL plasmid remained constant (0.8 µg), representing an hLmf1 wt -LPL ratio falling within the predetermined sensitive range ( Fig. 2A ). As shown in N430A mutation is comparable to the wild-type control, Lmf1 carrying either G181E (mLmf1 G181E ) or cld (mLmf1 cld ) exhibit a signifi cant decrease in their ability to restore LPL activity levels (Fig. 4B). Notably, the assay is able to detect a distinct difference between all three mutations by showing that G181E is intermediate in its severity when compared with N430A and cld . Regardless of the Lmf1 sequence tested, LPL protein expression remained relatively unchanged ( Fig. 4C ), indicating that G181E and cld affect LPL activity in a posttranslational manner.
Decreased Lmf1 protein levels and loss of Lmf1 function could account for the observed decrease in the ability of mLmf1 G181E and mLmf1 cld to restore LPL activity ( Fig.  4B ). Indeed, the G181E and cld mutations cause markedly reduced Lmf1 protein levels when compared with N430A and the wt control ( Fig. 4D ). Thus, both mutations likely cause increased Lmf1 protein turnover, possibly due to misfolding and/or improper membrane topology. To offset these low Lmf1 protein levels in the assay, increased amounts of plasmid for mLmf1 G181E and mLmf1 cld were used in the transfection while maintaining the preestablished plasmid levels for the mLmf1 wt control. As shown in Fig.  5A , LPL activity was restored by mLmf1 G181E in a dosedependent manner; in contrast, mLmf1 cld remained nonfunctional even when using highly elevated plasmid amounts in the assay. Based on these results, G181E clearly causes reduced Lmf1 protein levels, but it is diffi cult to assess whether it also affects Lmf1 function. To eliminate Lmf1 protein level as a variable, the readout of the assay was changed from LPL activity to Lmf1-specifi c activity, defi ned as LPL activity restored per unit Lmf1 protein. As shown in Fig. 5B , Lmf1-specifi c activity clearly shows that G181E does not affect Lmf1 function, which is in stark contrast to Lmf1 truncated by cld . Thus, G181E causes reduced Lmf1 protein expression without affecting function, whereas cld clearly represents a loss-of-function mutation. We hypothesize that G181E may decrease the effi ciency of proper Lmf1 membrane insertion, resulting in increased protein turnover. However, some fraction of mLmf1 G181E attains a native state, and these molecules remain stable and function normally as a lipase maturation factor. Importantly, the N430A, G181E, and cld mutations illustrate the ability of the assay to detect and analyze mutations having a wide range of effects on Lmf1 function and protein expression.  5. Functional analysis of Lmf1 mutants. A: Shown are LPL activity curves using an extended range of plasmid amounts for the two mutants, G181E and cld ; plasmid amounts for the wt control were identical to those used in Fig. 3A . B: Densitometric scans of Lmf1 band intensities after Western blot analysis were used as a measure of Lmf1 protein mass. Lmf1 mass was measured from transfected Lmf1 plasmid amounts of 2.5 ng ( wt ), 20 ng (G181E), and 300 ng (cld) as denoted by solid symbols in A, as they gave rise to similar Lmf1 band intensities of 38.8 ( wt ), 41.1 (G181E), and 60.4 (cld). These values lie within the sensitive range of the assay, whose upper limit is defi ned by the wt construct (5 ng), giving an Lmf1 mass of 100 arbitrary units. For each construct, the corresponding LPL activity levels (denoted by solid symbols) were used to calculate Lmf1 specifi c activity, defi ned as LPL activity per unit Lmf1 protein mass. The wt control was arbitrarily set to 100, and each value represents the average ± SEM of three separate transfections. The asterisk denotes a signifi cant difference between cld and wt ( P < 0.001).