A Simplified Ultra-Micro Kjeldahl Method for the Estimation of Protein and Total Nitrogen in Fluid Samples. of Less Than 1·0μ1

In adding to the already swollen list of micro-modifications of the Kjeldahl method our object has been the simplification of apparatus and procedure without loss of sensitivity. The method which we have evolved requires no apparatus which cannot easily be made from ordinary materials without special facilities, and we have introduced a new feature in the separate estimation of protein nitrogen. The quantities of nitrogen which can be estimated range from 1·0 to 10·0 μ.g. contained in 0·5-1·0 μl. with a standard deviation of 0·02-0·04 μg. N. We have thus been able to make duplicate determinations of protein and total nitrogen, as well as of chloride by Wigglesworth’s micro-method and of osmotic pressure by Baldes’ thermoelectric technique, using only a fraction of the blood of an insect larva (such as Sialis lutaria or Aeschna cyanea) from which further samples can be taken later.

Most of the previous methods of adapting procedure to this scale have been reviewed by Conway (1947). Their main features are set out in Table 1. Especial difficulties are encountered in: (1) accurate delivery of the sample into the digestion vessel; (2) transference of the digest to the diffusion vessel; (3) adjustment of volumes and concentrations of reactants and design of diffusion vessel to ensure complete absorption of the ammonia in a reasonable time with no danger of direct contact between alkalinized digest and absorbing standard acid; (4) design of diffusion vessel so that the acid can afterwards be titrated in situ; (5) evolving a sufficiently sensitive titration; (6) avoidance of contamination by extraneous nitrogen.

Of the methods listed in Table 1 we have the practical experience of Needham & Boell (1939) and Briiel, Holter & Linderstrom-Lang (1946). That of Needham & Boell entails specially constructed combined digestion and diffusion vessels and a rocking device. The digestion, which involves the adjustment of micro-burners, we have found difficult and tedious, and not suitable for the simultaneous treatment of more than a few samples. Owing to the small surface of acid exposed, the absorption time is as much as 12 hr. at 37° C. Boell’s method (Table 1e), though simple, does not attain the required sensitivity. Tompkins & Kirk (1942) used a digestion-diffusion vessel somewhat similar to that of Needham & Boell (1939), but reduced the diffusion time and increased the accuracy by exposing the standard acid in a spoon-shaped cup attached to the stopper. The method of Brüel et al. (1946) is the simplest as regards apparatus and we have adopted their technique for the digestion. But their diffusion procedure depends upon the subsequent method of titration, which we have simplified, and in general we have found the method tedious and requiring constant practice.

The technique described in this paper involves the following features not entailed in the methods listed in Table 1 : (1) waxed capillary pipettes modified from the Wigglesworth (1938) type for quantitative delivery of the sample and of the standard acid; (2) separation of the protein sample; (3) a type of diffusion vessel designed to house a hanging drop of standard acid on a waxed slide; (4) final titration with an easily constructed and simply operated microburette,^* the drop on the waxed surface being stirred with an air jet as in Wiggles worth’s micro-chloride method (1938).

In the dirt-laden atmosphere of Newcastle it was found necessary to keep all apparatus and reagents under dust covers and to perform all operations under a canopy which excludes falling particles.

The distilled water was prepared free of ammonia by redistillation from glass after addition of sulphuric acid and potassium permanganate. Blank determinations showed that these precautions were adequate.

The waxed sample pipette is modified from the Wigglesworth type in that the capillary protrudes some 5 cm. from the end of the supporting tube in order that the bottom of the narrow digestion tube can be reached (Fig. 1 a). The sample (0·7-0·8 μl.) extends about 3 cm. from the tip, the upper end of the sample being marked by a fine spot of black paint. These pipettes have the advantage that the whole of the section holding the sample, not merely the tip, can be rewaxed by immersion and blowing under a wax bath. We have even succeeded in removing the wax film by drawing up xylol several times to just beyond the mark, followed by acetone, concentrated nitric acid to remove protein particles, acetone, xylol and finally rewaxing. In this way a pipette can be renovated after the wax film has broken down or particles of organic matter have become attached to it which is not possible with the original model. These modified pipettes are not in practice more fragile than the original and the renovation has so far made no detectable difference to capacity as determined by calibration for the micro-chloride method, for which we also use them. The drop of standard acid is delivered from a waxed pipette of the normal Wigglesworth type holding approximately 7 μl. The wax used in these operations is the B.D.H. ‘so-called’ paraffin wax (clearing point 65/71° C.).

Other pipettes are simple drawn-out tubes of suitable dimensions and not waxed. All are operated by rubber suction tubes with glass mouth pieces packed with filter-paper.

The pyrex digestion tubes are of dimensions 55 × 2 mm. (Fig. 1 b).

For total nitrogen the sample is placed at the bottom of the tube followed by one washing of distilled water. For protein nitrogen a column about 4 mm. long of 20 % trichloracetic acid is placed about half-way down the tube by means of a fine unwaxed pipette and the sample is added to this with one washing. After centrifuging to the bottom the supernatant acid is sucked from the precipitate and the tube is filled with 5 % trichloracetic acid with careful washing down the sides of the tube. After further centrifuging and removal of acid this washing is repeated.

The procedure is now common to both samples.

A satisfactory digestion mixture was found in 50% sulphuric acid saturated with K₂SO₄ and containing 2% CuSO₄ and o·6% SeO₂. About 12 μl. are placed as a column at the top of each tube and shaken to the bottom. The tubes are then left overnight in an oven at 105° C. to remove water.

The digestion is completed by the method of Brüel et al. (1946), the tubes being heated at 280-300° C. for 5 hr. in a bath of sulphuric acid saturated with K₂SO₄. The bath is a pyrex tube of dimensions about 6×2 cm. with a constriction as shown in Fig. i b, and is embedded to a depth of about 1 cm. in a heated metal block. By this method there is no spluttering and the operation requires little attention as it is comparatively easy to adjust a bunsen under the block to maintain the temperature within this range. Several baths, each containing 6-12 tubes, can be held in one block, and a large number of digestions can thus be done simultaneously and with little trouble.

The diffusion chambers, 13 mm. internal diameter and 5 mm. deep (Fig. ic, d) are cut from a brass plate, the bottom being of thin sheet-brass and soldered on. They could be made of other materials, but since the heat of neutralization and dilution might melt the wax coat there is possibly some advantage in using a good heat conductor. We have not experienced any trouble of this sort. The diffusion chambers are coated by total immersion in molten wax (B.D.H. clearing point 65/71° C.). Glass micro-slides are similarly coated with pure paraffin wax (m.p. 48° C.). The rim of each chamber is then greased with Conway’s fixative (5° g-paraffin mp. 48° C. mixed by heating with 80 ml. of liquid paraffin).

A saturated solution of potassium metaborate (50-70 μl.) is placed in each chamber to one side, and on the other the contents of a digestion tube with about 80 μl. of distilled water used as washings. The transfer of the digest is made with a fine pipette followed by careful washing down the sides of the tube with three lots of distilled water. The technique is in fact that of Brüel et al. (1946), except that the transfer is direct to the chamber and not via a wax block.

About 7 μl. of 0·1 N-HCI delivered quantitatively from a waxed pipette is placed as a drop in the centre of each waxed cover-slide. This volume is suitable for the Estimation of protein and total nitrogen influid samples 19 estimation of 3-7 pg. nitrogen. The indicator for the final titration is incorporated in this acid as well as in the standard alkali in the burette. It is made up according to Conway (1947, p. 55)—1·3 ml. 0·1 % methyl red and 0·66 ml. 0·1 % bromocresol green per 100 ml. of standard solution.

Each cover is then inverted and fixed in position immediately after shaking the chamber sufficiently to bring the digest and metaborate together (Fig. ic). We have found that no significant ammonia is lost in this way, whereas the hanging drop of acid is apt to be displaced if the shaking is done after the cover is in position.

The time allowed for diffusion is 7 hr. at room temperature (see section on diffusion time).

The burette is made from the lower end of a broken thermometer tube about 10 cm. long. The tube which we have in use holds about 7 μl. for 140 divisions on the thermometer scale and tests have shown that the bore is sufficiently uniform to enable us to use these divisions. In fact a good quality thermometer of suitable bore is required. The previous bulb end is fused to another tube and drawn out to a capillary bent at right angles near its origin (Fig. 2).

The calibrated tube is fixed horizontally with the drawn-out capillary directed vertically downwards. The length of the capillary and the bore at the tip are so adjusted that, when the burette is filled with a mouth-suction tube, there is no discharge until the tip touches the drop of acid on the waxed slide raised from below on an adjustable screw stand (Fig. 2). An air jet is fixed relative to the tip of the burette so as to stir the drop when raised into position. The horizontal setting of the upper calibrated portion of the burette ensures that there is no change in height of the liquid column and thus the rate of delivery is uniform, which is essential for accurate quantitative delivery.

The burette is filled with 0·1 N-NaOH with added indicator (see above) from a drop placed on a freshly waxed slide. The end-point (red to green) is sharp when seen against a white and well-illuminated background. The titration is controlled by raising and lowering the screw-stand.

No solutions are accurately and independently standardized except the acid used as ammonia absorbent. From this the 0·1 N-NaOH is repeatedly standardized by titration in terms of scale units on the burette.

It is also unnecessary to determine accurately the volume of any of the pipettes. It is required only to know the ratio-volume of standard acid pipette/volume of blood-sample pipette (A/B). This can be found by filling each with standard add and titrating against alkali in the burette.

If a₁ scale units of alkali are required to neutralize one pipette load of standard add, and a₂ are needed after bsorption of the ammonia, then the concentration of nitrogen in the sample (as normality) is

With the procedure and quantities detailed above it was found that within 2 hr. about 95% of the sample had been digested. Thereafter the process was much slower and more than 4 hr. were required to complete it. Hence 5 hr. was chosen as a suitable time.

The curves in Fig. 3 demonstrate the course of diffusion and absorption of ammonia at room temperature from a 0·33% solution of ammonium sulphate released by potassium metaborate as already described. The progressive addition of the other liquids required for the full procedure—water for washing the digestion tubes and the acid digestion mixture—greatly lengthen the diffusion time (Fig. 3). This is presumably due to the lowered alkalinity and perhaps also to the reduction in surface/volume ratio of the mixture. On this basis 7 hr. was chosen for the diffusion time at room temperature.

• (a)

  Titration only. Table 2 summarizes three sets of titrations of 7 μl. N/10 HCl against approx, N/10 NaOH in the burette. The standard deviation was 0·2-0·4%. The standard deviation on the burette readings is of the same order when the titration is done in conjunction with the rest.of the procedure (see Tables 3 and 4).

  Table 2

  7·0μl. 0·1026 N-HCI titrated against approx, N/10 NaOH.

  

  

  View Large
  Table 3

  7·0 μl. 0·329 % (NH₄)₂SO₄. Standard acid 0·1026 N

  

  

  View Large
  Table 4

  Egg albumin solutions. Standard acid 0·1026 N

  

  

  View Large

• (b)

  Diffusion-titration. The results of three sets of nitrogen estimations on 7 μl. samples of 0μ33% ammonium sulphate are given in Table 3, the standard deviation being 0·02-0·04 μg-N.

• (c)

  Digestion—diffusion—titration. Samples were taken from solutions of eggalbumin containing about 2·5 /μg. N per /μl., the exact weight of nitrogen being assumed to be dry weight of albumin/6·25. The results tabulated in Table 4 give a standard deviation of 0·02-0·04 /μg. N.

• (d)

  Protein separation. An attempt to assess the efficiency of the precipitation technique was made by estimating the nitrogen in albumin and ammonium sulphate solutions, separately and after mixing, with and without precipitation with trichloracetic acid. The following results were obtained:

  

  

  View Large

The apparent loss of nitrogen from the albumin sample after precipitation is not unexpected in view of the probable presence of amino nitrogen not precipitable by trichloracetic acid. But a more important point is the practical identity of the last two figures, which proves that the non-precipitable nitrogen is effectively removed by the washing procedure.

For our purpose it was not necessary to reduce the scale to this range. But, since we have used the same digestion technique as Brüel et al. (1946), who estimated quantities of this order, it was of interest to find whether the rest of our procedure could be adapted. We therefore did four diffusions with 7 μl. samples of 0·033 % ammonium sulphate using N/i 00 acid and alkali. A standard deviation of 0·006 /μg. N was obtained. It is probable therefore that the complete procedure would give about the same degree of accuracy as the method of Brüel et al. (Table 1).

To illustrate the performance with actual blood samples, Table 5 shows results obtained from the blood of two specimens of each of the insect larvae Sialis lutaria and Aeschna cyanea. The large individual variation in protein nitrogen is a characteristic feature.

The basis of this method was worked out in the Zoophysiological Laboratory in Copenhagen in September 1947. We are very much indebted to the Director, Prof. P. Brandt Rehberg, and to Dr E. Zeuthen for their hospitality and encouragement and to Dr H. Holter of the Carlsberg Laboratory, for giving us the opportunity for practical experience of their micro-Kjeldahl technique (Brüel et al. 1946). One of us (L. C. B.) was in receipt of a grant from the Travelling Fund of the Royal Society, and the research of which this forms a part is supported by a grant from the Leverhulme Trust.