A simplified semiautomated assay for plasma triglycerides

A simplified semiautomated assay for plasma triglycerides

ANAI,YTIC!& BIOCHEMISTRY 29, A Simplified for MAX 401r116 (1969) Semiautomated Plasma Assay Triglycerides E. ROYER AND HOWARD KO Pharmaceu...

663KB Sizes 0 Downloads 30 Views

ANAI,YTIC!&

BIOCHEMISTRY

29,

A Simplified for MAX

401r116 (1969)

Semiautomated Plasma

Assay

Triglycerides

E. ROYER

AND

HOWARD

KO

Pharmaceutical Research and Development Division, The Upjohn Company, Kalamazoo, Michigan 49001 Received September 30, 1968

Serum triglyceride levels which are higher than “normal” levels have been suggested as one of the biological indicat’ors of coronary artery disease (5). The semiautomated procedures of Lofland, Kessler, and Lederer, and Antonis (8-10) for measuring triglycerides require more or less extensive sample preparation. A new, simpler, extraction procedure has therefore been developed and the fluorometric procedure of Kessler and Lederer has been modified. MATERIALS

Sodium methylate: 0.1 M sodium methylate (Matheson Coleman & Bell) in isopropanol. The isopropanol solubilizes the sodium alkoxide in the nonane sample stream during transesterification. Sodium periodate (meta) : NaIO, (Mallinckrodt) ,0.025 M, dissolved in 5% acetic acid. Acetylacetone reagent: 4oJ0, 2,4-pentanedione (puriss. grade, Aldrich Chemical Co.) in an aqueous solution of 2 M ammonium acetate (Mallinckrodt) . Use 1 hr after preparation and only for 2 days. Nonane: distilled in vacua by Burdick and Jackson Laboratories, Muskegon, Michigan. Isopropanol: distilled by Burdick and Jackson Laboratories. Tripalmitin standard: 1 mg/ml in nonane. Florid: 100-200 mesh (Floridin Co., Hancock, West Virginia), heated lOO-120°C overnight, used for the phospholipid removal experiment. APPARATUS

FZuorometer:l Turner fluorometer, model 111, equipped with ,#405 (405 rnp peak) Turner filter (primary) and a #2A-12 (510 rnp sharp cutoff) Turner filter (secondary) . ‘A preliminary experiment indicates that a Technicon N tubular flow calorimeter equipped with a 420 mp filter and a 15 mm flow cell may also be used; however, a slower sampling rate was necessary for adequate resolution and the extent of interferences was unknown. 405

406

ROYER

AND

KO

Flow cell: Aminco-Bowman, 2 mm i.d. cyclindrical cell. To increase sensitivity the cell holder was altered to accept mirrors. This cell gave short washout times. Technicon Sampler II module. Technicon diSPo beakers: 2 ml .#B-2714 or 1 dram glass vials were used in the sampler. Technicon proportioning pump. Heating bath: 50°C. Recorder: Texas Instruments Servo/riter, 10 mv. Reaction coils: #l and ,#2 consisted of polyethylene

tubing, 0.062 in. i.d., wound in a 7 cm diameter circle (Bel-Art Products, Pequannock, New Jersey). Stretched out, tubing ,#l was 4 ft long and #2 was 20 ft. The coils were suspended in the heating bath with the coil axes horizontal (i.e., the planes of the coils were vertical). diSPo pipets: 9 in. (Scientific Products). Thin-layer plates (for the TLC phospholipid detection experiments) : 20 x 20 cm silica gel F,,, precoated (E. Merck AG., manufacturer, Brinkmann Instruments, distributor). METHODS

The Turner fluorometer is turned on 60 min prior to beginning triglyceride analysis to stabilize the electronics. A. Sample Treatment Fresh heparinized blood is rapidly cooled to ice-water temperature and centrifuged to obtain plasma (serum can also be used). To a 15.0 ml glass centrifuge tube are added: 2.0 ml nonane, 3.5 ml isopropanol, 1.0 ml 0.08 N H,SO,, and 0.5 ml plasma. The mixture is vigorously mixed 20 set by a Vortex stirrer and the phases are allowed to separate (usually 2-3 min is required). Approximately 1.5 ml of the nonane phase (upper) is transferred by a diSPo pipet to sampler cups2 in the sampler module. B. Sample Analysis The modules, transmission, and proportioning pump tubing are set up as in Figure 1. The acetylacetone-periodate-sample “Dl” junction should be kept horizontal; otherwise erratic responses may occur (such variability may arise from additions of reagents to the sample stream). Erratic response is also reduced by joining the two streams of periodate and ‘To check whether Florisil was required to remove any remaining phospholipid, about 80 mg of Florisil was added to the nonane phase with a spatula and the cups were vibrated and rotated for 20 min. The sampler tube was adjusted so that it was above the level of the Florisil in the bottom of the cup.

SEMIAUTOMATED

Reaction CobI 1 4-Foot Polyethrlene Tubing (0.062 in. I.D.) (Connected by N5 Nipple)

TRIGLYCERIDE

035”Pump I

Reaction CotI :2 20-h. 0 062 in. I.D Polyethylene Tubing

Tube White 0.60 ml. mtn.

Acetyhcetone

White

Periodote

Reqcnt

1 0.60 ml. min

D2 Orange-White

(Salvoflex)

Reagent

0.23 ml. min. Sample 50shr. 2:l Sample 10 Wash Ratio

I Blue (Acid flex) Red

1.19 ml.,‘min.

0.80 ml..‘min.

Yellow

I

407

ASSAY

Sodium methylate Reagent

Air

(Solvoflex)

1.06 ml.,‘min.

4

I

1. Flow

FIG.

diagram

of assay

for plasma

triglycerides.

acetylacetone reagents before joining the combined reagent’s to the sample stream. The use of polyethylene coils instead of standard glass coils eliminated the breakup of the bubble pattern which was observed in glass coils. The samples are analyzed at a rate of 50 per hour with a sample-towash ratio of 2/l using nonane as the wash. The assay response is standardized by tripalmitin or triolein, which is ‘dissolved in nonane, and which is then so distributed against isopropanol and aqueous H&SO, that the final proportions of the solvents are the same as in extracting plasma. The recorder assay responses of samples and unknowns above baseline are entered onto IBM cards. Dr. Carl Metzler and Ernest Markovich3 have programmed the IBM 360/30 digital computer to fit the assay responses (y) to concentrations (2) of triglyceride standards to the equation l/y = A/X + R. This equation was found to fit (the lack of fit was not significant, p > 0.05 by the F test) the standard curve for this assay. Estimates of the triglyceride levels of unknowns are calculated together with their standard deviations and 95% confidence limits. C. Thin-Layer

1. TLC #1:

Chromatography Phospholipid

(TLC) for Detecting Interference

Plasma

1 ml of the nonane phase extract was dried with N?. Then (2/l) was added to each tube and 20 ,~l was

0.1 ml of CHCl,/MeOH 31nformation

and

Computer

Services

Division

of The

Upjohn

Company.

408

ROYFiR

AND

HO

spotted on a TLC plate. The TLC plate was developed with heptane/ EtOEt/HOAc (50/50/2) (see reference 12) and the zones visualized by iodine vapor. 2. TLC #S: (a) 1.0 ml of rat plasma was extracted; 1.5 ml of the nonane phase was treated 20 min with 80 mg of Florisil and 1.0 ml of the nonane phase ‘was dried on the Biichi Rotavapor (Rinco Instrument Co., Inc.). (5) A second 1.0 ml plasma sample was treated in the same manner except that the Florisil treatment was left out. (c) Folch extract-3.0 ml of rat plasma was extracted by CHCI,/ MeOH (2/l) according to Folch et al. (7). Seven aliquots of the lower phase (equivalent to 0.56 to 57% of the dried nonane extracts if nonane were able to extract plasma lipids with complete efficiency) were dried with N,. Each of the dried residues (a), (b) , and the seven aliquots (c) were dissolved in aliquots of 0.05 ml of CHCI,/MeOH (2/l) ; 20 jul of the solutions were spotted on a TLC plate. The plate was developed by MeOH/ CHCl,/H,O (35/65/4) solvent (13) and the zones visualized by iodine vapor. RESULTS

AND

DISCUSSION

The automated analyses of triglyceride in plasma involve manual extraction of plasma and automated analysis of triglycerides in the extract which includes: release of glycerol by transesterifying triglycerides (4)) periodate oxidation of glycerol to formaldehyde (2)) reaction of formaldehyde with acetylacetone and ammonia to form the fluorescent product 3,5-diacetyl-1,bdihydrolutidine (3)) measurement of fluorescence, and calculation of the amount of triglyceride. Extraction of Plusmu for Triglycerides. The extraction system was initially patterned after the Dole (14), using heptane, isopropanol, and water. However, the heptane extract evaporated at a rate of 3 to 4y0 per hour while on the sampler in spite of being covered. Nonane was found to be sufficiently nonvolatile and was therefore employed. Using the working hypothesis that triglycerides are extracted efficiently over a wide range of nonane/isopropanol/0.033 N H$O, ratios, we sought ratios that minimize recovery of phospholipids (Table 1). After thin-layer chromatography (TLC .#l) of phospholipid recovered in the nonane layer, the resolved components were visualized by iodine vapor. It was found that (a) the zone for phospholipid w.as least dark when extraction systems ,#5 or #6 were employed (i.e., when sufficient isopropanol was present) ; (b) emulsions difficult to break were formed when no isopropanol was used; (c) the amount of H&SO, between 16 and 406 meq did

SEMIAUTOMATED

Extraction

Mixtures

Miz3Y NOFlrtllE!

TRIGLYCERIDE

TABLE 1 Tested for Efficiency of Phospholipid

Extraction

Sample (ml)

Hz0

(ml)

(ml)

m

Isopyynol m

1

2.0

1.0

3.5

0.5

0.0

2 3

0.5 0.5

1.5

6

2.0 2.0 2.0 2.0 2.0

7 8

2.0 2.0

4 5

0.16

r

ff?SOa

409

ASSAY

1.0

2.0

1.0 1.0

1.5 1.0

1.0

0.5

0.5 0.5

1.0

0.0

0 .i

3.5

1.0

2.4

0.1

1.0

0.0

2.5

1.5 1.5

from Rat Plasma Remarks

Emulsion: phases would not separate

2.0 2 .5 3.0

Gave minimal phospholipid recovery

not markedly affect the recovery of phospholipid. The extraction mix (,#6) providing minimal phospholipid recovery was therefore chosen. To provide some estimate of phospholipid recovered in the nonane phase, the extraction efficiency of system #6 was compared by thin-layer chromatography (TLC ,#2) with that of Folch’s method (7), which extracts 99+% of tissue lipids into the chloroform phase. The phospholipid zones visualized by iodine after TLC of the residue from nonane layer (extraction system #6) were compared on the same plate with the zone corresponding to various amounts of Folch extracts of the same plasma. The intensity of the phospholipid zone (phosphatidylcholine) from the nonane extract was bracketed by the intensities of the phospholipid spots that were from the Folch extract and that corresponded to 0.56 and 1.4% as much plasma; thus only about 1% of the phospholipid available in rat plasma was extracted in this system. If the phospholipid in the nonane layer is absorbed with 80 mg of activated Florisil, much less’ than 0.56% of the phospholipid is recovered. The extraction procedure could be used without the additional Florisil adsorption provided the following positive bias from phospholipid could be tolerated: If the phospholipid content remains under five times the triglyceride level, the positive bias from phospholipid should be less than 5% (relative). For man, cattle, sheep, dog, rabbit, guinea pig, rat, and chicken, the positive bias would be less than 3 mg % since the “normal” plasma levels of phospholipid are usually less than 300 mg $I (1). When the complete assay system was set up, the efficacy of phospholipid removal by the extraction system was tested in another way: In an experiment on 31 samples of human serum, the triglyceride values obtained with Florisil treatment of the nonane (upper) phase of the serum extract

410

ROYER

AND

KO

did not differ significantly (p > 0.05, F test, Table 2) from those obtained w.ithout such treatment. Under different conditions’, e.g., pathological ones or sera from different species, where positive bias from phospholipid is unacceptable, absorption of the nonane phase with Florisil will reduce the biases to negligible values. Maximization of Assay Response. The parameters (viz., concentrations of sodium methylate, sodium metaperiodate, acetic acid, actylacetone, and ammonium acetate and temperature) in the assay were varied simulOaneously by the method of steepest ascents (6) until maximal responses

Relative

Fluorometric Treatment

TABLE Response of Extracts

Average

relative

2 with and without of Human Serum fluorometric

response

Florisil

(number Florisil

Sample

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

20 21 22 23 24 25 26 27 28 29 30 31

No.

No Florisil

49.0 (1) 77.0 (1) 144.0 (1) 204.0 (1) 112.0 (1) 138.0 (1) 197.0 (1) 189.0 (1) 120.0 (1) 120.0 (1) 140.0 (1) 163.0 (1) 177.0 (1) 191.0 (1) 175.0 (1) 146.0 (1) 124.0 (1) 116.0 (1) 98.0 (1) 154.0 (1) 178.0 (1) 128.0 (1) 170.0 (1) 130.0 (1) 79.0 (1) 89.0 (1) 147.0 (1) 135.0 (1) 105.0 (1) 153.0 (1) 96.0 (1)

of replicates)

(80 mg) treatment of nonam phase

47.5 78.0 144.5 205.0 114.5 137.5 195.5 189.5 120.5 120.0 140.5 165.0 175.0 191.5 177.0 146.0 123.5 115.5 99.0 155.0 180.0 129.0 168.0 129.0 80.0 88.0 147.0 134.0 106 .O 152.5 93.5

(2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2) (2)

SEMIAUTOMATED

TRIGLYCERIDE

411

ASSAY

were obtained. The parameters corresponding to the maximal responses were then reinvestigated one at a time about their maximal values (“regular concentration”) to check that plateau values had indeed been reached (Fig. 2). With the concentrat,ions of reagents providing maximal response, no difference (p > 0.05) in response between a heating bath temperature of 50” and 55°C was observed.

SODIUM METHYLATE

.L

921

FIG.

2. Maximization

8 l/2 x

of fluorometer

lX=Regular CONCENTRATION

response

2X

to reagent

concentration.

Interferences. In addition to the phospholipids already mentioned, compounds which lead to aldehyde formation by means of the reactions used in the assay or which are aldehydes themselves are possible sources of interference. Such compounds are vicinal hydroxyls, a-amino alcohols, cr-ketols, or aldehydes. Representative compounds such as glucose, glycerol, and serine could be expected to interfere if they were extracted into the nonane phase. As seen in Table 3, these interferences occur only to the extent of less than 0.2% for the amounts’ usually present in plasma. Adsorption of the nonane phase with 40 mg of Florisil reduced the interferences even more. Since we did not have plasma data on the amount of free ethanolamine or octanal present, the interference from these substances was estimated on the basis that these substances comprised as

412

ROYER

AND

HO

TABLE 3 Degree of Interference from Plasma Constituents

Compound

Tripalmitin Blank Glycerol Glucose Serine Ethanolamine Octanal Blank Glycerol Glucose Serine Ethanolamine Octanal

Amount added/ml nonane in extraction mixture

*y~~l~f adsorbent

1.0 mg 0.0

0 0

0.1 ml

0

10.0 mg 10.0 mg O.lml 0.1 ml 0.0 O.lml 10.0 mg 10.0 mg 0.1 0.1

0 0 0 0 40.0 40.0 40.0 40.0 40.0 40.0

mg mg mg mg mg

mg

Response in arbitrary units f S.E.M. (Iv’ = 4)

Approximate expected level in plasma

591 0 51 f 5 9.25 f 1.3 5.0 + 0.1 28.5 k 1.5 6.5f0.3 1.3kO.5 8.3 * 1.2 -0.25 f 0.25 -0.25 f 0.25 1.0 + 0.67 9.75 f 1.55

1 mg/ml

a Per cent response to 1 mg/ml of triglyceride amine or octanal in 1.0 ml plasma.

10 a/ml 1 w/d 2 mg/ml 10 m/ml 1 mg/ml 2 mg/ml -

Expected ‘$5 interference in 1 ml human plasma

-

0.00068% 0.16% 0.17% (4.8%)” (1.1%)” 0.0001% ~0.00008% SO.O0016’% (0.17%)”

(1.6%)”

if there were as much as 0.1 ml ethanol-

much as 10% of the plasma volume. Even so, the interference from these substances was relatively minor. Other possible sources of interference in plasma that could occur are those compounds which fluoresce in themselves or which arise during the presence of one or more assay reagents (i.e., sodium methylate (M), sodium metaperiodate (P) , and acetylacetone (A) ) (Table 4). The type of reactions or product associated with the various reagent combinations are indicated. No fluorescent serum blanks were observed, however, as a result of extracting human serum and using the automated procedure, where one or more of the reagents were omitted. In the case in which all the reagents are present (MPA), any extracted phospholipids would int,erfere and, as previously discussed, this source of interference has been minimized through appropriate choice of extraction conditions. Another feature for reducing potential interference is provided by the automated analytical scheme. Triglyceride fatty acids (and, to a large extent, free fatty acids) are transesterified to esters which are nonanesoluble. Nonane from the sample stream is carried along through all the reactions and can serve as a wash. Thus, any nonane-soluble interferences that occur will be extracted from the aqueous phase. Since the nonane phase is removed by the debubbler, any fluorescence interference from these sources would be reduced. Recovery Studies. The presence of protein during extraction of plasma triglycerides conceivably could lead to lower recovery than that when the

SEMIAUTOMATED

TRIGLYCERIDE

413

ASSAY

TABLE 4 Test for Fluorescent Blanks Arising from Assay Reactions with Compounds Extracted from Serum Hypothetical Reaction(s) form Reagents present,J

Transesterification

- (none) M--P-A MPM-A -PA MPA

+

conditions

under

which

(one or more) occur fluorescent productsb Oxidation

fluorescence

could

Compound intrinsically fluorescent

Hantzsch reaction

+ + + + +

+ +

f

Observationc on human Serum extract

+ + + + + + + +

+

+ f

occur

to

a M 3 sodium methylate; P E sodium metaperiodate; A = acetylacetone; solvent only. b (+) reactions or conditions can occur to form fluorescent products. c 0 = no fluorescence above blank values observed; F = fluorescence observed.

-=

distribution is made in the absence of protein. The extraction system (,#6, Table 1) was therefore tested. However, no significant difference (Table 5, p > 0.05) was found between the recoveries of tripalmitin (initially dissolved in nonane) when the distribution was made against rat plasma (plus isopropanol/aqueous H,SO,) as compared with that against water (plus isopropanol/aqueous H,SO,) . Triglycerides could also be poorly recovered if the level of triglycerides in serum exceeded their solubility in the nonane phase. To test this, different volumes of a pool of human lipemic sera were extracted. The nonane phase of each extract was correspondingly diluted so that, if the extraction efficiency was the same for all the levels of triglycerides, the TABLE 5 Recovery of Triplamitin when Distributed vs. Rat Plasma as Measured by the Automated Assay Sample

Observed

Tripalmitin alone Rat plasma alone

equivalents

per ml

1.0303 + 0.0049 (N = 10) 0.8199 + 0.0046 (N = 10’1

Sum Rat plasma + triplamitin Difference

Recovery * S.E.M.

m g of triplamitin

=

1.8601

added

f 0.0067

1.8415 f 0.0074 (N = 0.0087 (99.57

It O.OlCO f 0.87)s

10)

A S.E.M.

414

ROYER

Fklative Extraction

AND

RO

TABLE 6 Efficiency of Different Amounts of Triglyceride

Extra;&~~~nm

l.Oml 0.5 ml 0.2 ml E Level of triglyceride

Dilution of the nonane (upper) phase

5.0x 2.5X 1.0x 484.5 mg ‘% (tripalmitin

in Human Serum

Fluorometric response zt S.E.M.

173.3 zk 1.0 (N = 8) 174.8 f 0.8 (N = 8) 174.8 + 1.1 (N = 8) equivalents).

same fluorometric response would result. We see in Table 6 that there is no significant difference in the fluorometric responses (p > 0.05). This supports the hypothesis that extraction efficiencies of the triglycerides’ are the same between 96.9 to 484.5 mg of tripalmitin equivalents per 100 ml of serum. Comparison with Other Semiautomated Assays. No significant difference was found between the values obtained by the semiautomated chromotropic acid assay (modified and measured by Dr. John Flokstra, personal communication; cf. (10)) and the fluorometric assay described in this paper (Table 7). The procedure described here, however, is simpler because the steps common to the method above and some other semiautomated procedures are made unnecessary, viz., absorption of phospholipid, one or more drying steps, and/or centrifugations (S-10). For example, oleic acid has been reported to interfere in the chromotropic acid end point after saponification, periodic acid oxidation, and reaction with chromotropic acid; thus fatty acids released by triglyceride hydrolysis were eliminated by extraction (11)) an additional step. A larger number of samples can also be processed through the automated portion of the assay (50 per hour vs. 20-30 per hour) since no sample blanks are obtained. SUMMARY

A simpler, semiautomated, fluorescence method for the determination of triglycerides has been developed for the AutoAnalyzer (Technicon). The method involves manual plasma extraction and then t.he following of triglycerides (sodium automated procedure: (a) t ransesterification methylate in isopropanol) to release free glycerol, (5) oxidation of glycerol to formaldehyde, and (c) reaction (Hantzsch) of formaldehyde with acetylacetone and ammonia to form the fluorescent product 3,ELdiacetyl1,4-dihydrolutidine. Triglyceride is extracted from human serum under conditions that maximize extraction (quantitative) of triglyceride, minimize extraction of glycerol, glucose, and phospholipid, and provide good separation of phases. Interferences from phospholipid in human serum are sufficiently low that the usual step (in other methods) of adsorbing

SEMIAUTOMATED

Comparison

TRIGLYCERIDE

TABLE Assay

of Fluorometric Fluorometric

assay

7 with

Number replicates

430

2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 2

1 2 3 4 435 6 7 440 1 2 3 4 445 6 7 8 9 450 2 3 4 455 6 7 8 9 460 3 464

of

obtained obtained

7’ z 0

127.4 121.6 150.8 143.8 228.0 218.7 129.6 129.2 154.2 185.0 200.1 221.5 201.0 161.3 133.6 124.0 104.1 172.4 205.0 139.7 190.7 140.5 82.5 92.2 162.6 146.7 112.2 169.8 99.1 140.1

Molecular weight ratio tripalmitin = 1.08 Q Values b Values

Esjim,“t+ m

of corn

Chromotropic

Acid

Assay

(this paper)5 Aprx$n&te

Ident;$ation

415

ASSAY

estimate

0.7 0.6 0.6 0.7 0.8 0.7 0.6 0.6 0.6 0.6 0.7 0.7 0.7 0.6 0.6 0.6 0.6 0.6 0.7 0.6 0.6 0.6 0.5 0.6 0.6 0.6 0.6 0.6 0.6 0.7

oil to

using tripalmitin by Dr. J. Flokstra

Chrpmotropic acld asseyb bg %)

126 127 162 152 243 241 134 136 162 197 247 260 223 182 149 138 122 183 220 156 207 166 97 110 182 163 128 181 101 148 (N

as standard. using corn

= 30) Av.

Ratio of chomrotropic to fluorometric rtssay

0.988 1.043 1.073 1.056 1.065 1.101 1.033 1.052 1.050 1.064 1.234 1.173 1.109 I. 127 1.114 1.112 1.171 1.061 1.072 1.116 1.085 1.180 1.174 1.191 1.119 1.110 1.140 1.065 1.018 1.055 1.098

+ 0.010 (S.E.M.)

oil as standard.

phospholipid (onto silicic acid, Florisil, aeolite, or Doucil) is not required. To minimize evaporation, nonane is used as nonpolar solvent in the extraction mixture. The amounts of reagents, the temperature and time required to carry out the reaction steps (a) through (c) were varied stepwise until maximal or near-maximal response was obtained. The automated portion of the assay is run at 50 samples an hour. With

416

ROYER

AND

HO

“normal” human serum, the relative standard deviation of triplicate determinations ranged from 0.71% for 83 mg % of triglyceride (milligrams of tripalmitin equivalents per 100 ml of serum) and 0.35% for triplicate determinations of 228 mg $I. The lowest detectable amount of triglyceride for which the 95% confidence limits did not include zero was 1.6 mg %. ACKNOWLEDGMENTS Thanks are due to Dr. John Flokstra and Paul Schurr for materials, and Dorothy Porten for rat blood. Dr. Flokstra also provided serum in which triglyceride had been measured. Appreciation is due to Barbara Cronk, who punched the data on IBM cards, and Dr. Carl Metzler for statistical analyses. REFERENCES 1. LONG,

C., ed., Biochemists

Handbook.

Van Nostrand,

Princeton,

N. J., 1961.

2. VORIS, L., ELLIS, G., AND MAYNARD, L. A., J. Biol. Chem. 133, 491 (1940). 3. BELMAN, S., Anal. Chim. Acta 2Q, 120 (1963). 4. L~DDY, F. E., BARFORD, R. A., AND RIEMENSCHNEIDER, R. W., J. Am. Oil Chemists’ Sot. 37, 447 (1960). 5. FRIEDMAN, M., J. Am. Med. Assoc. 190, 85 (1964). 6. COCHRAN, W. G., AND Cox, G. M., Experimental Designs, 2nd ed., p. 357. Wiley, New York, 1957. 7. FOLCH, J., LEES, M., AND SLOAN STANLEY, G. H., J. Biol. Chem. 226, 497 (1957). 8. ANTONIS, A., J. Am. Oil Chemists’ Sot. 44, 333 (1967). 9. KESSLER, G., AND LEDERER, H., in “Automation in Analytical Chemistry” Technicon Symposia 1965 (L. T. Skeggs, Jr., ed.). Mediad Incorporated, New York, 1966. 10. LOFLAND, H. B., Anal. Biochem. 9,393 (1964). 11. CARLSON, L. A., AND WADSTROM, L. B., Clin. Chim. Acta 4, 197 (1959). 12. VOGEL, W. C., DOIZAKI, W. M., AND ZIEVE, L., J. Lipid Res. 3, 138 (1962). 13. H.~BERMANN, E., BANDTLOW, G., AND KRUSCHE, B., Klin. Wochschr. 39, 516 (1961). 14. DOLE, V. P., J. Clin. Invest. 35, 150 (1956).