Determination of Zinc in Millimicrogram Q BO G. MALlMSTRONI AND DAVID GLICK Medical School, University of Minnesota, Minneapolis
I
i'. T H E pest few decades, quantitative histoehemieel methods
been developed for the determination of B great number of organic and inorganic constituents of tissues, hut little attention has been given to biological trace metals. I n view of the close relation between zinc and the ensyme carbonic anhydrase (S), a method for the quantitative determination of this element in single microtome sections would seem particularly worth while. The method described here permits the determination of quantit.ies of Binc from 2 to 13 millimicrogram with an accuracy of 0.5 millimicrogram. The range of the amount of zinc to be determined can he extended. This is a modification of the method developed by Vallee and Gibson (Is),which has B sensitivity of l microgram and employs dithizone as a reagent. The desired sensitivity has been achieved by the use of microequipment, allowing the volumes of the liquids used to be reduced about 1000 times. While the method was developed with the biological application in mind, i t can also he employed for nonbiological samples. An example of the use of the method is an application t o the histological distribution of zinc in the dog gastric mucosa (8). Serial sections of the tissue were separately analyzed and their zinc content was correlated d h the histology, as revealed by stained control sections. Human whole blood was also analyzed. ;.we
- ~ ,._.".
ml. portions of the dithisone solution until the latter remsined pure green. APPARATUS
Tissue Borers. Borers of the same type as described by Linderstrp[m-Lang and Holter (7) were made, but Lucite was used instead of metal t o avoid Contamination of the sample with traces of ilinc.
.M
,....~C
REAGENTS 3
The water used was distilled twice, the second time in a n allborosilicate glass still. All solutions (except dithiaone) were stored in polyethylene bottles. Dithizone Solution, 0.7 mg. yo (w./v.). Dithizone (0.7 mg.) was dissolved in tetrachloroethylene, and the solution was made up to 100 ml. This solution was stored in the refrigerator in a dark borosilicatr glass bottle. It was made up fresh every week. Hydrochloric Acid, 2 N . Hydrogen chloride gas, generated by dropping concentrated sulfuric acid on sodium chloride was dissolved in zinc-free water. The normality of the solution was determined by titration of an aliquot, and i t was diluted with rincfree water to make it approximately 2 N .
Figure 2.
o
Detail
Extraction
hpya.ars.a
A.
8.
C. D, E .
F.
c. Reaction Tubes. The reaction tubes usea were maae of quarts. because this material can wit.hstand the high temperature mquired for ashing and is also relatively free of heavy metale. The tubes had a n inside diameter of 4 mm. and a length of 25 mm. and their tops were externally ground and fitted with ground borosilicate glass female oaps wit.h flat tops t o facilitate clamping in the extraction apparatus. As each cap matched only a, given tub,.all.tt+es ?nd caps were numbered. iuDe noiaer tor Iishiig. In a stainless steel block ( 5 X 7.5 x 1.3 om.) 35 holes ( 5 mm. in diameter and 10 mm. .deep) were drilled. Extraction Appar atus. An apparatus for quantitatiIve extraction of a solution wi t h a n immiscible liquid described b: i SchmidtNielsen ( 2 1 ) has bw?n modified as follows:
raetion A p p a r a t u s for Volumes of Liquid
The apparatus consists o f a centrifuge which rotates the reaction tubes in such a. manner that the denser liquid is thrown back and forth through the other liauid. This is accomnlished bv the make it aooroximatelv iN. from one ~
~
~
~~~~~~~
~ . ~ .
?he s o h i o n w a s brouiht to pH 5.5 with 1% N acetic acid using a glass eleotrode and made up to 250 ml. with zinc-free water. It wsa then shaken in a borosilicate glass separatory funnel with 5-
of the tubes to the other (Firmre 1). ' The n r e s e 2
~~~. --- -- - ---- -. The tubes lie in the &ooves of the removable part shown in ~~~~
1699
~~~
ANALYTICAL CHEMISTRY
1700 Figure 3, and these grooves are wider a t one end t o enable the caps to fit. This grooved part fits between the movable sides, BC (Figure 2), where it is held in place by pins D and E. The sides, B , are brought together with the screws, A, t o clamp against the ends of the capped tubes. C represents pieces of sponge rubber which permit the tubes to be held securely without danger of breaking. F indicates a gear Theel which is driven by the centrifuge so that F rotates on the ball bearing post, G . The gear ratio is such t h a t wheel F makes one revolution for every 5 l I 3 of the centrifuge. Parts A, D , E, and F are made of brass; the other metal parts are of aluminum. By changing the tube bed shown in Figure 3, tubes of different sizes or shapes can be used. The counterweight on the end of the centrifuge arm opposite that bearing the tube holder (Figure 1) is adjustable t o permit balancing. A knifeedge for balancing the centrifuge arm before it is set on the l crn I axle of the centrifuge motor is shown in Figure 4. A semicirFigure 3. Detail of Base cular ringstand base was used Plate for Tube Holder in to hold a short post t o which Extraction Apparatus the knife-edge was soldered.
graphic adapter (E. Leitz, Inc., S o . 72.005) provided Kith a shutter, an observation eyepiece, and a 10 Xmicroscope ocular. The prism in unit F , which reflects the light into the eyepiece, ran be moved out of the light path with a cable releape. Adjustable arm H supports the photocell, G, which is part of the photometer (Photovolt Corp., Model 512, photocell E). The photocell used has an appreciable spectral response between 360 and 700 mM, and is most sensitive a t 500 mp. Above 900 mp its response is zero. Photocells for other ranges of wave length may be substituted in the Photovolt instrument. Figure 6 shows the lens system ( B in Figure 5 ) in detail. It can be adjusted in a vertical direction by means of rack and pinion operated by wheel A and adjusting screw B permits it t o be tilted a t different angles. The lens system, J, has a focal length of 15 mm. and it consists of a mounted projection lens unit (Edmund Salvage Co , Barrington, X. J., S o . 4045) n-hich
Microcolorirneter. -4 diagram of the microcolorirneter is shown in Figure 5. The design is similar to that described by Holter and LPlvtrup ( 2 ) and by Krugelis ( 4 ) ,the difference being in the illuminating system. The earlier workers employed the illumination unit of a Zeiss Stufenphotometer. I n the present design, A represents a microscope illuminator (Bausch and Lomb Optical Co., Type 31-33-86) provided with a rlbbon filament lamp (General Electric Co., l8-ampereJ 6-volt, Figure 4. Knife-Edge for Balancing Centrifuge Arm Type T 10). A 110-volt alternating current source of current, of Extraction -4pparatus stabilized with a voltage regulator (Raytheon, VR 6113), and a transformer were used to supply the lamp. .Alternatively, a 6-volt storage battery nas used with a charger sup...... .. ...___ G plied by a 110-volt alternating current source stabilized as above. I n this \\ arrangement current mas fed into the battery a t the same rate that it a a s delivered t o the lamp, so that, in effect, the battery served only to flatten any ripple in the stabilized and transformed electrical supply. Although it requires a storage battery and rectifier in addition t o the apparatus used in the former setup, the latter assures a more constant illumination than is possible with the former unless the 100-volt alternating current supply is devoid of any but minor fluctuations. Storage batteries alone could be used to supply the current if they were kept in a well-charged condition, but the srrangrment described is more convenient. To obtain a narrow concentrated beam of light, the lens system, B , was introduced. This lens system is not necessary if the microscope illuminator of the Central Scientific Co., Chicago, No. 85262 with auxiliary lens and diaphragm No. 85271, is used as a light source, as the light then can be made t o converge sufficiently. The latter was found t o be more difficult t o adjust, because it lacked setscrews t o hold the mounted lenses rigidly. C represents a nonsilvered righeangle prism which was introduced t o give a single reflecting surface; D , a microFigure 5. Microcolorirneter Assembly scope condenser which has its top half E. Objective (5 X ) A . Microscope illuminator removed t o reduce the convergence; B . Lens system F. Microphotographic adapter E, a 5 X objective with a numerical C. Right-angle prism G. Photocell aperture of 0.16; and F , a microphotoD. Microscope condenser with top half removed H . Adjustable support ~
0
0
1701
V O L U M E 23, NO. 11, N O V E M B E R 1 9 5 1 L
G
K
I cm
3
Figure 6. A. B.
H
a
I
A
CUVETTE FILLING.The cuvette was attached t o a microscope elide with silicone stopcock grease. It was filled without introducing air bubbles by placing the tip of a narrow-stem noncalibrated pipet (outer diameter about 0.5 mm.) near the bottom of the cuvette and delivering the liquid with gentle air pressure applied by mouth through a rubber tube; when the liquid neared the top of the cuvette the tip of the pipet was raised carefully as the final addition of liquid was made. The liquid was allowed to overflow the top surface of the cuvette and a cover slip was put on. ADJUSTMENT OF COLORIMETER.The lens system, J in Figure 6, was adjusted so that an image of the lamp filament was formed in the focal spot of lens E. I n this way a narrow parallel beam of light was formed. The microscope condenser converged the beam somewhat. It was adjusted in such a way that the narrowest part of the beam-Le., the focal point-would fall in the middle of the cuvette on the microscope stage. The microscope was focused in such a manner that the top surface of the cuvette could be seen in the observation eyepiece. This allowed the cuvette t o be adjusted on the mechanical stage so that the light passed through the center of the lumen without striking the walls. The diameter of the light beam on entering and leaving the cuvette was around 0.5 mm., while in the (*enter it was around 0.2 mni.
Lens System for Illumination of Microcolorinieter
Vertical adjustment Horizontal adjustment
C. Collimator
E . Double convex lens (15.2-mm. focal length) F. G, J . Mounted projection lens unit (15-mm. focal length) H . Setscrews K. Post 1.. Mounting tube
can he adjusted to focus the beam on the focal point of the double convex lens, E, whose focal length is 15.2 mm. From E , the relatively parallel light is collimated by passing through the pin holes of unit C. E is permanently mounted in tube I,, while the positions of projection unit J and collimator C can be adjusted by means of screws H . The cuvettes used are pieces of capillary tubing with flat ground ends, 7 mm. in length and 0.9 mm. in inner diameter (obtained from 0. Dich, 33 C Lykkesholmsalle, Copenhagen, Denmark; also available from Fischer C Porter Co., Hat'boro, Pa.). METHOD
The theoretical foundation of the dithizone method has been treated by many aut,hors-e.g., Sandell (10). Cleaning of Glassware. The reaction tubes were cleaned as follows : They were filled with ordinary distilled water, which was then shaken out, immersed in 2 .V nitric arid which was heated to boiling and allowed to cool and after 12 hours the acid was shaken out. The tubes were rinsed 6 times with zinc-free water. Finally they were partly filled with dithizone solution, capped, and rinsed internally by whirling t h r m in the extraction apparatus. If the green dithizone color changed in any tube it was not clean enough. The solution was shaken out of the tubes and they were then used without drying. .I11 other glassware (borosilicate glass) was washed with soap and water, rinsed tv-ith ordinary distilled water, immersed in 2 S nitric. acid for 12 hours, rinsed 6 times with zinc-free water, and alloived to dry in a dust-free cabinet. Colorimetry. CHOICE OF T V . ~ V E LESGTH. The solvents commonly employed in dithizone methods, chloroform and carbon tetrachloride, were found to be too volatile for the present purpose, as a vapor bubble formed on top of the liquid in the cuvette even though a cover glass was placed on the filled cuvette. Tetrachloroethylene (boiling point 119-121 " ) n-as therefore employed. Both dithizone and zinc dithizonate \yere readily soluble in this solvent, and the absorption spectra of these solutions are given in Figure 7. Colorimet,ric measurements were made a t 620 mp, \\.here dithizone has its absorption maximum but zinc dithizonate does not absorb a t all. I t is thus the dithizone remaining after a certain amount has reacted xyith zinc that has actually been measured. A Farrand interference filter with a transmittance maximum a t 619 nip and a half-band width of 12 mp was used.
6 00
700
Wavelength ( m p )
Figure 7.
Absorption Spectra of Dithizone and Zinc Dithizonate in Tetrachloroethylene
COLORIMETRIC READIA-G.The instrument was set to a galvanometer reading of zero x i t h the shutter in unit F (Figure 5 ) closed. A cuvet,te containing the solvent, tetrachloroethylene, alone was adjusted with respect to the light beam by means of the mechanical st,age, t,he prism in unit F was moved out of the light path with the cable release, the shutter was opened with the other cable release, and the instrument was set to a galvanometer reading of 100. The unknown sample was then t,reated in t,he samr way as the solvent hlank and the per cent transmittance a t 620 mp was read. -4measure of the initial dithizone conemtration was obtained by detcwnining the tian$mitt,ance of a reagent tilank. OUTLINE OF hIETIIOD
1. Punch out a cylinder of the tissue to lie analyzed with the plastic borer of the diameter desired. Push out the cylinder from the borer with a plastic plunger and freeze the tissue t o the freezing head of the microtome with the aid of a drop of zincfree physiological saline. 2 Cut sections of desired thickness with a glass microtome knife (6). After each freezing, discard the first section because it will have a n indeterminate thickness due to contraction of the metal of the microtome head. Transfer a section to be analyzed to a quartz reaction tube v ith the tip of a drawn-out borosilicate glass rod. Keep the tip of this glass rod in a dithizone solution betiyeen transfers to prevent contamination by zinc in dust, etc. 3. Put the tubes in the stainless steel block and ash the srctions in a muffle furnace (3 hours a t 475' C.). After the
A N A L Y T I C A L CHEMISTRY
1702 ashing, the tubes should be protected with borosilicate glass caps except when reagents are added. 4. Add 20 PI. of 2 N hydrochloric acid with a constriction pipet (7). (This type of pipet is used for all reagents.) Let the tubes stand for half an hour with occasional shaking. 5. Add 2 N ammonium hydroxide to bring the p H to 5.5 (determine the volume required by previous titration). Shake the tubes (if this is not done a sulfur precipitate is often formed when the buffer is added, and this causes the analytical results to be low). 6. Add 25 PI. of buffer solution. 7. Add 20 p1. of 0.7 mg. % ' dithizone solution nith constant pressure as described by LinderstrZm-Lang and Holter (6). If the sample contains more than 13 my of zinc, a stronger solution of dithizone will be required. 8. Extract in the mixing apparatus for 5 minutes at a speed of 650 r.p.m. (35 times gravity). 9. Transfer an aliquot of the organic phase t o a cuvette and determine the transmittance of 620 m r as previously described. 10. Determine the transmittance of a blank carried through eteps 3 to 8.
Table I. Zinc Added,
Zinc Recovered, m-, Standard solution, Standard solution, ashed without tissue ashed with tissueD
Standard solution
my
2.6
2.0 2.0 2.0 4.0 4.0 4.0 6.0 6.0 6.0 8.0 8.0 8.0
2.0
1.7
10.0
2.0 1.7
2.0 1 7 3.6
1.7 3.6 4.0 4.0 5.3 6.0 5.6 8.3 7.3 7.6 10.3 10.6
4:0
4.0 4 6 6 3
3.6
6:6 6.0
6 0
.. ..
..
..
10:3 10.0
10.0 .. 10.0 10.0 .. .. Dog gastrir niucosa section\, 4 n i m . i n diameter X 3 0 p thick.
Deternilliatioiia on human \ ~ h o I cblood \+erein agreement \vith the data of Vallee and Gibson (IS). Therefore, the quantities of other metals present in gastric mucosa and blood do not interfere with the zinc determination. If the method is to be applied to other tissues, similar control experiments should be performed. Photometric Error. The variation of the photometric error with transmittance in the present procedure, where a color difference is measured, differs from that found in a Beer's law syst,em (9). An equation for the photometric error as a function of the transmittance for a given dit,hizone concentration was derived by taking the derivative of m , Equation 1 , with respect to T, where 1' is the transmittance of the sample in per cent of the incident light:
0.4
0.3
0
',0.2 n
0. I
din
-
-/n =
0.0
0.0
Recovery of Zinc in Presence and Absence of Tissue
2.0
4.0
6.0
8.0
10.0
mp/g Z n
Figure 8. Standard Curve for Determination of Zinc Standardization of Method. . 4 standard solution was made up by dissolving 8.79 mg. zinc sulfate heptahydrate in zinc-free 0.1 N hydrochloric acid and making the solution up to 100 ml. with the acid. This solution was st,able for several weeks and was used as a stock standard solution. Just before use it was diluted 50 times to yield a solution containing 2 my of zinc per 5 PI. Known quantities of zinc in the form of this solution ranging from 2 to 10 my were added to the reaction tubes and carried through steps 4 to 9 described above. .4 determination wit'hout added zinc was carried through step8 4 to 9 as a blank experiment. When (Do- D ) , where Dois t'he optical density of t,he reagent blank and D that of the sample, was plott,ed against the amount of zinc, a straight line was obtained. In Figure 8, each point is an average of three to six determinations. From bhe line plotted, the following equation was obtained: m = 33 (Do- D) (1) where nt ie the amount of zinc in millimicrograms.
x
100
dT
% rel. anal. error 43.4 l%abs. photom. error (Do- D )X
Equation 2 shows that it is advant,ageous to employ the Ion-est dithizone concentration that will still permit all the zinc in the sample to react and thus bring the value of T up to 100. HOTever, a too dilute dithizone solution n-ill give poor extraction; 0.7 mg. $c (w./v.) was found to be the most Satisfactory concentiation in the range 2 to 13 my of zinc, but a higher concentration of 30
I
1
1
1
1
I
dT
(Do-D)xT
,I
0.7 mg.% dithizone ,
light path
7
Accuracy of Method. The maximum probable error calculated from the standardization experiments was 0.5 my. Recovery experiments, in the absence and presence of biologicnl material, were carried out as follows: Known quantities of zinc in the form of the solution used for standardization were carried through the ent'ire procedure, including the ashing. Analyses were also carried out after known quantities of zinc were added to sections of gastric mucosa taken in the chief cell region where the material is histologically homogeneous, and the zinc content was found to be practically constant, (8); and for each of these sections, an adjacent section, without, added zinc, was analyzed to obt,ain the blank value for calculation of the recovery. The greatest deviation from the added amount was 0.7 my (Table I). It is clear that the ashing did not significantly increase the error. The largest single error is the photometric error, as shown herenith.
(2)
C
50
60
70
80
% transmittance
Figure 9. Photometric Error as a Function of Transmittance
V O L U M E 23, NO. 1 1 , NOVEMBER 1 9 5 1 dithizone will be required for greater quantities of zinc. Do is 0.42 at 620 mp for 0.7 mg. dithkone with a 7-mm. light path. The magnitude of the photometric error for different transmittance values can be seen in Figure 9. The photometric reading error with the instrument employed was found to be 1%. Thus, above 54% transmittance ( 5 my of zinc), the analysis error due to photometry is less than 5%. LITERATURE CITED
Glick, David, J . Natl. C a n c e r l n s t . , 10,321 (1949). Holter, H., and Lpivtrup, S., Compf. rend. trao. lab. Carlsberg,
1703 (7) LinderstrBni-Lang, K., and Holter, H., “Die enaymatische
Histoohemie,” in Bamann, E., and hfyrback, K., “Die Methoden der Fermentforschung,” Leipaig, Thieme, 1940. (8) Malmstrom, B. G., and Glick, D., Ezptl. Cell Research, in press. (9) Ringbom, A., 2.anal. Chem., 115, 332 (1939). (10) Sandell, E. B., “Colorimetric Determination of Traces of Metals,” pp. 72-93, Xew York, 1nt.erscience Publishers, 1944. (11) Schmidt-Sielsen, K., C‘onipt. re7zd. traz. lab. Carlsherg, Skr. chim.,24, 233 (1942). (12) Vallee, B. L., and Gibson, J . G., 11, J . H i d . Chem.. 176, 435 ( 1948). (13) Ihid., p. 445
Sir. chz?n., 27, 27 (1949).
Keilin, D., and Mann, T., Biochem. J . , 34, 1163 (1940). Krugelis, E. J., Compt. rend. Zap. traa. C a d s t e r g , S6r. chim., 27, 273 (1950).
Latta, H., and Hartmann, J. F.,Pvoc. SOC.Ezptl. B i d . M e d . , 74, 436 (1950).
Linderstr@m-Lang,K., and Holter, H., Conipf. r e n d . trar. lab. Ciirlsberg, S b . chin?.,19, 1 (1931).
R E C E I V EApril D 16, 1961. S o . S X I in the series, Studies in Histochemistry. KO.XX appeared in ( 1 ) . Work aldrd by grants from the National Cancer Institute, U. S. Public Health Ser\-ice, and the Research F u n d of the Gradua t e School, University of Minnesota. D a t a in this paper were presented in a thesis submitted by B. 0.Malmstroni t o the Graduate Faculty of t h e Cnirersity of hlinnesota in partial fulfillinent of the requirements lor the degree of doctor of philosophy.
Simplified Semimicro Aniline Point Test JENNIE &I. CHENET Humble Oil and Re$ning Co., Baytown, Tex. aniline point method (6)of test can be of great value in the characterization of hydrocarbon mixtures. This “property” is defined as the minimum equilibrium solut,ion temperature for equal volumes of aniline and sample ( 1 ) . Use of this test on narrow cuts from laboratory superfractionations and on other liquid products from bench scale units has been limited, however, by the rather large amounf of sample (10 ml.) required by the standard ASTM procedure (1). Elimination of the UP? of this inforniat,ive trst on many samples is not desirable and in ansn-rr to this problem a “scaled-doivn” version of an ASTM procedure (2) requiring only 0.2 ml. of sample has been developed. A 20-gage B & S copper-constantan thermocouple with clay separators is substituted for the usual ASTN aniline point’ thermometer; it also serves as a stirrer. The modified technique may be used on any hydrocarbon product for which an aniline point determination is desired, provided that t,he sample is light enough in color t o allow the chance from condition of niiscihilit#yto turbidity to be observed satisfactorily through the test. tube. Observation of the aniline point of slightly darker samples is facilitated h y placing n ~nia11light behind the apparatus, APPARATUS
The apparatus for the modified test is easily constructed and inexpensive (Figure 1).
It consists of a test tube, B , 9 to 10 mm. in outside diameter and 8 to 9 cm. in length, fitted by means of a slit cork, E, into a larger tube, A , about 13 111111. in outside diameter and 10 cm. in length. The ends of the thermocouple wires, which also serve as a manual stirrer, form a ring, I),which is slightly smaller than the inside diameter of the inner test tube. Immediately above the ring the wires form a right-angle bend and are held firmly by means of the clay insulation pipes, C. In the present work, niillivolts of thermocouple output measured on a Leeds & Northrup potentiometer (Catalog No. 8662) were converted to degrees centigrade by use of the Leeds & Northrup table (4). -4reference junction a t 0’ C. waa med. PROCEDURE
Before each series of determinations, the aniline point value is obtained on pure n-heptane by both the standard .46TM procedure (1) and the modified method, using aniline from the same container. Although the difference betxeen results on n-heptane Kith the two methods is usually small, applying this difference as a blank correction to the values from the modified test results in more accurate determinations..
The determination of the aniline point on a given sample calli: first for pipetting precisely equal volumes of dry aniline and sample (about 0.2 in]. each) into the clean dry apparatus, which is clamped on a ring stand. The thermocouple ring is then centered in the apparatus and the outer test tube is slowly heated by use of an open flame. Throughout the operation, the thermocouple ring is lifted and loir-ered by means of the clay pipes, causing a gentle stirring action. Care is taken that the thermocouple is not lifted above the surface of the liquid. The heating and stirring action is continued until complete miscibility is obtained, and then the mixture is allowed to cool dowly but with continued stirring. The potentiometer reading at which the mixture becomes cloudy t’hroughout with complete separation of aniline and sample phases is recorded. This heating and cooling procedure and record of temperature of complete cloudiness are repeated until a constant value is D - Thermocovple LOOP obtained. Not more than 5 cycles are usually required. Figure 1. hpparatue for Semimicrodeterminations of Aniline Point
ACCURACY AND PREClSlON
The accuracy and reproducibility of the small scale method were tested by a series of determinations on compounds of a nominal purity of 95 to 99%, as well ae on a refinew sample. Results on the pure compounds are compared with those given in t,he literature (3) anti those obtained on the same samples by the APT11 method ( 2 )in Table I. Deviations from theoretical Bureau of 1Iines values average about 0.4’ C. high, and deviations from the ASTU procedure average about 0.4” C. high. A maximum deviation from theoretical of 0.9” C. was observed for methylcyclohexanr. Results on t’he pure compounds, on the averagr, check the theoretical and A4STMmethod values ( 1 ) with about the same order of reproducibility ae the other ASTM method (Wj,which has a “reproducibility (different operator and different apparatus)” of ~ k 0 . 4C.~
,
