Journal of Pharmaceutical Sciences 105 (2016) 1851-1857

Contents lists available at ScienceDirect

Journal of Pharmaceutical Sciences

journal homepage: www.jpharmsci.org

Pharmaceutical Biotechnology

Protein Nitrogen Determination by Kjeldahl Digestion and Ion

Chromatography
Hsiaoling Wang, Nagarani Pampati, William M. McCormick, Lokesh Bhattacharyya*
Division of Biological Standard and Quality Control, Ofce of Compliance and Biologics Quality, Center for Biologics Evaluation and Research, Food and Drug
Administration, Silver Spring, Maryland 20993

a r t i c l e i n f o a b s t r a c t

Article history: We report development and validation of a simple, rapid, and accurate method for the quantitation of
Received 8 February 2016 protein nitrogen, which combines Kjeldahl digestion and ion chromatography with suppressed con-
Revised 22 March 2016 ductivity detection and requires nanomolar amount of nitrogen in samples (
10 mg protein). The
Accepted 30 March 2016
mechanism of suppressed conductivity detection does not permit analysis of samples containing copper
(present in Kjeldahl digestion solution) and aluminum (present in many vaccines as adjuvants) due to
precipitation of their hydroxides within the suppressor. We overcame this problem by including 10 mM
Keywords:
oxalic acid in Kjeldahl digests and in the eluent (30 mM methanesulfonic acid). The chromatography is
analytical biochemistry
biotechnology
performed using an IonPac CS-16 cation exchange column by isocratic elution. The method reduces the
proteins digestion time to less than 1 h and eliminates the distillation and titration steps of the Kjeldahl method,
vaccines thereby reducing the analysis time signicantly and improving precision and accuracy. To determine
protein nitrogen determination protein nitrogen in samples containing non-protein nitrogen, proteins are precipitated by a mixture of
Kjeldahl digestion deoxycholate and trichloroacetic acid and the precipitates are analyzed after dissolving in KOH. The
ion chromatography method is particularly useful for biological samples that are limited and can also be applied to food,
suppressed conductivity detection environmental, and other materials.
oxalate complexation with copper and
2016 American Pharmacists Association. Published by Elsevier Inc. All rights reserved.
aluminum
validation

Introduction technique is often unsuitable for biological and environmental

The determination of total protein in pharmaceutical, biological,
food, environmental, and other materials is performed widely by
the Kjeldahl method.1-5 The method indirectly quanties the total
protein content from nitrogen measurement and involves 3 major
steps: digestion, distillation, and titration. The original method has
been subject to many modications and improvements. In its most
recent form, the samples are digested in sulfuric acid/potassium
sulfate in the presence of a catalyst to quantitatively convert
nitrogen in proteins to ammonium sulfate. The solution is then
distillated in the presence of excess of sodium hydroxide and the
liberated ammonia is absorbed in an acid, which is titrated to
quantitate ammonia. The method is time consuming and requires
large sample quantities. For example, an experienced chemist
would spend about 10 h for determination of nitrogen using the
manual micro-Kjeldahl method, which typically requires about
0.1-0.7 mg nitrogen (0.6-4.0 mg protein) per sample.5 Thus, the
* Correspondence to: Lokesh Bhattacharyya (Telephone: 1-240-402-9399;
Fax: 1-301-595-1489).
E-mail address: lokesh.bhattacharyya@fda.hhs.gov (L. Bhattacharyya).
samples where only small sample quantities are available.
Ion chromatography (IC) has become a well-established tech-
nique for the quantitation of ammonium ion. The use of IC to
determine total nitrogen as ammonium ion after Kjeldahl digestion
can signicantly improve the speed, accuracy, and precision of
analysis by eliminating distillation and titration steps. In addition,
because only nanomolar amounts of ammonium are required for
analysis,6,7 the amount of sample required for the digestion can be
signicantly reduced, which is of critical importance for biological
and environmental samples. The IC method for the determination
of ammonium ion typically involves separation by cation-exchange
chromatography using a dilute solution of a strong acid followed by
suppressed conductivity detection using a Cation Electrolytically
Regenerated Suppressor (CERS) [formerly called Cation Self-
regenerating Suppressor].8 The efuent from the column passes
through CERS before entering into the conductivity detector. In
CERS, all anions, including the anion(s) present in the eluent, are
replaced by OH, which forms H2O with H present in the eluent
thereby essentially eliminating any background conductivity due to
eluent and increasing the signal intensity (because ion conductance
of OH- is greater than all other anions).9 Although this improves
sensitivity and accuracy considerably, the technique poses a

http://dx.doi.org/10.1016/j.xphs.2016.03.039
0022-3549/ 2016 American Pharmacists Association. Published by Elsevier Inc. All rights reserved.
1852 H. Wang et al. / Journal of Pharmaceutical Sciences 105 (2016) 1851-1857
signicant problem to analyze Kjeldahl digests, which typically use
copper sulfate as the catalyst. In the presence of hydroxide ion,
Cu2 forms sparingly soluble copper hydroxide that would pre-
cipitate to choke the eluent ow and clog CERS.
Another problem of using IC to analyze Kjeldahl digests is that
the digests typically contain large molar excess of copper,2-5 which
reacts with ammonium ion to form a blue cupramine complex in
neutral and acidic solutions.10 Because copper sulfate is used in
large excess, there will be no detectable free ammonium left in the
solution when eluted with a dilute acid during IC. In addition,
copper being a bivalent transition metal ion is expected to bind
strongly to the cation exchange stationary phase, which could
result in either irreversible damage to the column or require
frequent column clean up.
Analyses of Kjeldahl digest using IC have been reported.11-13 The
authors attempted to circumvent the above mentioned problems
by using catalysts other than copper. Jackson et al.11 reported
digestion with sulfuric acid-acetic acid mixture in the presence of
either hydrogen peroxide or Hg2 as catalysts, for determination of
total nitrogen in food and environmental samples. Kjeldahl digests
were analyzed by IC using indirect conductivity determination and
obtained 5%-14% difference in results from the Kjeldahl method in
samples containing 136-307 mg/mL ammonium ion. Although in-
direct conductivity determination does not require CERS, it requires
large amount of samples for accurate quantitation due to high
background conductance.9 The reported assay range is too high for
the analysis of most of the biological and environmental samples.
Furthermore, the use of mercury or hydrogen peroxide as
catalyst requires either signicant safety precaution due to the
toxicity of mercury or complex and expensive equipment because
of potential explosion due to the presence of H2O2.12 de Medina
et al.12 reported determination of total nitrogen in water samples by
oxidizing organic nitrogen to nitrate by potassium peroxodisulfate
under high temperature and pressure using a Parr-type bomb and
determining nitrate by IC with suppressed conductivity detection
and obtained results that are comparable with the standard Kjel-
dahl method for river water containing 0.5-5.3 mg/mL of total
nitrogen. Although the concentration range is suitable for analyzing
biological samples, operation and maintenance of a Parr-type bomb
is expensive and complex, and requires taking signicant safety
precautions.
Pontes et al.13 reported analyses of soil and sandstone samples
for the determination of total nitrogen, in which the traditional
distillation step in the Kjeldahl method is replaced by distillation
using a purge-and-trap ultrasonic system and ammonia in the
distillate is quantitated using IC. The method used a calibration
curve in the range of 0.03-0.80 mg/mL of ammonium. Although the
method requires small amount of sample, it does not provide any
advantage over the traditional micro-Kjeldahl analysis.
We report a simple and rapid method for the determination of
nitrogen or protein concentration in biological samples, including
vaccines, by a two-step procedure, Kjeldahl digestion followed by
IC analysis of the digest after dilution. The method provides accu-
rate and precise results in the range of 0.01-1.0 mg/mL of ammo-
nium and requires very small amount of protein. The method
eliminates the need for distillation and titration steps, thereby
reducing the run time considerably and improving accuracy. We
overcame the problem of having copper in the digest by simply
adding oxalic acid to the Kjeldahl digest and the eluent. In the
presence of oxalic acid, copper forms a stable copper bioxalate
complex, [Cu(OX)2]2.14,15 Copper bioxalate being an anion does not
bind to the column and is replaced by OH by CERS. In addition,
complexation with oxalate prevents formation of cupramine
complex and concomitant loss of ammonia. Furthermore,
aluminum present in many vaccines as adjuvants also forms
anionic complex with oxalate15 and is removed by CERS, which
permits application of the method to vaccines.
Many biological products contain non-protein nitrogen in
excipients.16 Because proteins constitute the active components, it
is important to determine protein nitrogen in such product. In the
food industry, it has also become critical to monitor protein nitro-
gen instead of total nitrogen after a few infamous food adulteration
cases.17 In this communication, we also report a simple procedure
for the selective determination of protein nitrogen in samples that
also contain non-protein nitrogen by precipitating protein with
deoxycholate (DOC) and trichloroacetic acid (TCA) before Kjeldahl
digestion and analyzing the precipitates after dissolving in KOH.
Materials and Methods
Proteins
Bovine serum albumin (BSA), human serum albumin (HSA), and
ovalbumin (OVA) were of the highest purity available from Sigma-
Aldrich (St. Louis, MO). Two mock lots of Anthrax vaccine were
prepared in our laboratory, each by mixing contents of multiple vials
from different lots of the vaccine from the approved manufacturer
thereby removing identity traces of the lots. Five mock lots of inu-
enza (Flu) vaccine antigens were prepared similarly by mixing
monovalent bulks of the vaccines from approved manufacturers
without noting the strain identications, identity of the manufac-
turers, or protein concentrations. Each of these monovalent bulks
contains inuenza proteins from a strain but no live or attenuated
viruses. The excipients present in these preparations are essentially
the same as those present in actual vaccines. Thus, these samples can
be considered as representatives of corresponding vaccines for our
method.
Reagents and Standards
Ammonia standard (100 ppm, NIST traceable) was from Ricca
Chemical Company (Arlington, TX). Methanesulfonic acid (MSA),
sodium DOC, 0.1 M oxalic acid, and concentrated sulfuric acid were
from Sigma Aldrich. Potassium sulfate, potassium hydroxide, TCA,
50% sodium hydroxide solution, copper (II) sulfate pentahydrate,
and 0.01 N HCl were obtained from Fisher Scientic (Waltham,
MA). The reagents needed for Lowry method are from Thermo
Scientic (Waltham, MA). Deionized water (18.2 MU$cm) was used
to prepare all prepared solutions.
Kjeldahl digestion solution is prepared as described previously2
by adding 10 mL of a saturated solution of copper (II) sulfate to 500
mL of concentrated sulfuric acid, incubating for 7 days at room
temperature so that excess copper sulfate is precipitated out, and
collecting the clear solution from the top.
Equipment
The chromatography was performed using an ICS-3000 Ion
Chromatography System, IonPac CS16 cation-exchange analytical
column (5  250 mm), IonPac CG16 guard column (5  50 mm), CERS-
500 (4 mm), and Chromeleon software for the system operation and
data analysis, all from Thermo Scientic (Dionex, Sunnyvale, CA). The
ion suppressor, CERS-500, is inserted in-line between the column
outlet and the detector. The Micro-Kjeldahl Digestor and RapidStill I
distillation unit were from Labconco (Kansas City, MO). An Agilent
(Santa Clara, CA) 8453 UV-Visible Spectrophotometer was used to
measure absorbance at 280 nm.
 H. Wang et al. / Journal of Pharmaceutical Sciences 105 (2016) 1851-1857
Micro-Kjeldahl Method
The micro-Kjeldahl analysis is performed by a modication of
the AOAC method 982.38.2 About 1 mg of protein is digested with
500 mg K2SO4 and 2.0 mL of CuSO4/H2SO4 Kjeldahl digestion
solution. The digestion is initiated by maintaining the temperature
at about 200 C until the fuming starts. The temperature is then
raised until the acid boils gently at slightly above 400 C. The sample
undergoes various color changes and nally becomes colorless and
the fuming ends. The digestion is continued further at this
temperature for an additional 30 min. The total digestion time is
3-4 h. The temperature was controlled by xed rheostat settings
during digestion. A sample of 2.0 mL of water is also digested
simultaneously to constitute the blank. Both sample and blank
digests are made alkaline with NaOH and steam distilled. Liberated
ammonia is absorbed in solutions of 2% boric acid and nitrogen
contents are determined by titration with 10 mN HCl after blank
correction.
Modied Kjeldahl Digestion for Ion Chromatography
In the nal method, a 250 mL protein solution was digested with
200 mL Kjeldahl digestion solution and 50 mg of K2SO4. Two hun-
dred fty microliter of water is also digested simultaneously, to be
used as the blank. After the fuming ends, the digestion is continued
for another 10-15 min at slightly above 400 C.
This method should be performed in a fume hood with pro-
tective gloves (2 pairs) and eyewear.
Ion Chromatography
Unless stated otherwise, Kjeldahl digests are quantitatively
transferred with water to 25-mL volumetric asks and 250 mL of 1.0
mM oxalic acid is added to make the nal oxalate concentration of
10 mM, and made up to the mark with water. Twenty microliter
solution is injected and the column is eluted isocratically with
30 mM MSA containing 10 mM oxalic acid at the ow rate of 1.2
mL/min with column and detector temperatures set at 40 C (sup-
pressor current: 106 mA). The nitrogen content is determined from
the area of the ammonium peak, after blank correction, using a
calibration curve of ammonium standard solutions. The standard
was not blank corrected.
Protein Concentration
Protein concentrations are determined by the micro-Kjeldahl
and Kjeldahl-IC methods from the nitrogen content using conver-
sion factors 6.07, 6.04, and 6.34 calculated from the published
sequences of BSA, HSA, and OVA, respectively.18 For Anthrax and Flu
antigens, a conversion factor of 6.25 is used.19 Protein concentra-
tions are also determined by Lowry20 and BCA21 methods using BSA
as the standard. In addition, the concentrations of pure proteins are
determined from absorption measurements at 280 nm using their
molar extinction coefcients calculated from the published amino
acid sequences, as described by Pace et al.22
Sample Preparation for the Protein Nitrogen Determination in
Nitrogen-Containing Formulation Matrix
This is done by a modication of a previously reported proce-
dure.20 To 1.0 mL each sample containing 100 and 250 mg BSA
without and spiked with 1.0 mM L-glutamine as a representative
nitrogen containing excipient, 0.1 mL of 0.15% DOC is added and the
mixture is incubated for 10 min at room temperature, after which
0.1 mL 72% TCA is added. After an additional 10 min incubation, the
1853
samples are centrifuged at 13.2 K rpm for 10 min. The pellet is
washed twice with 0.2 mL ice-cold acetone, dried in air at room
temperature for 30 min, and dissolved in 1.0 mL of 0.1 M KOH. Two
hundred fty microliter solutions (containing 25 and 62.5 mg of
protein) were digested for Kjeldahl-IC analysis.
Results and Discussion
Kjeldahl Digestion
Our preliminary results indicated that, for the analyses of
microgram quantities of proteins, it is critical to optimize Kjeldahl
digestion to prevent the loss of ammonia. Micro-Kjeldahl digestion
requires 3-4 h to assure quantitative conversion of protein to
ammonium.2 There is also an observable color change during
fuming, from essentially colorless to light yellow/brown/purple
(depending on protein concentration) and then back to clear and
colorless as the fuming ends. The color changes are not obvious for
our Modied Kjeldahl Digestion method described above, where
only a few micrograms of proteins were digested; however, the
fuming was clearly visible. We found that an additional 10-15 min
digestion after the cessation of fumes is adequate to get optimum
recoveries. Prolonging the digestion time resulted in solidication
of the mixture and loss of ammonia. The total digestion times are
about 30 min for pure proteins and about 45 min for proteins in
complex matrices, such as are present in vaccine preparations.
Ion Chromatography
Method Development
In preliminary experiments, we digested 300 mL of 1 mg/mL BSA
and diluted the digest to 5 mL with 10, 25, 50, and 100 mM oxalic
acid to optimize the concentration of oxalic acid necessary. The
diluted digests were chromatographed (20 mL injection) isocrati-
cally with 30 mM MSA containing the corresponding concentra-
tions of oxalic acid. The column and the detector were maintained
at 40 C. Ammonium concentrations in the digests are determined
from the area of the ammonium peak using a standard curve of
ammonium ion in the range 1-5 mg/mL. We observed excellent
linearity of the standard, R2
0.995 in the presence of 10 and 25 mM
oxalic acid, while the linearity is somewhat poor (R2 ~0.95) in the
presence of 50 mM oxalic acid. In addition, the chromatograms
show baseline noise in the presence of 50 mM or higher concen-
trations of oxalic acid; however, the baseline noise was not seen
with 10 and 25 mM oxalic acid. Although there were other problems
(see later), we did not observe excessive back pressure during
elution at any oxalic acid concentrations, including 10 mM oxalic
acid, indicating that copper was not precipitated in the CERS in the
presence of oxalic acid as low as 10 mM in the eluent. The results
show that 10 mM oxalic acid, when included in the Kjeldahl digest
as well as in the eluent, is sufcient to convert copper ions present
in the Kjeldahl digest to the negatively charged copper bioxalate,
which is then removed by the CERS. Had copper not been converted
to copper bioxalate completely, the unconverted copper would
precipitate as copper hydroxide in the CERS, as discussed above,
resulting in signicant increase in back pressure leading to stop-
page of the chromatographic run. Thus, inclusion of 10 mM oxalic
acid in the sample to be injected and in the eluent is sufcient to
completely chelate copper present in the Kjeldahl digest. All sub-
sequent experiments were performed with samples and eluents
containing 10 mM oxalic acid.
In our preliminary experiments, we found signicant problems
with the chromatogram when the Kjeldahl digests were diluted to
5 mL. The problems include (a) lack of baseline stability, the base-
line appeared to be constantly decreasing; (b) the resolutions
1854 H. Wang et al. / Journal of Pharmaceutical Sciences 105 (2016) 1851-1857

Table 1
Performance Characteristics of Ammonium Peak

Protein (Amount Digested) Kjeldahl Digest Resolution Between Resolution Between K Theoretical Tailing Factor of
Diluted to (mL) Na and NH 4 Peaks and NH 4 Peaks Plate (NH) 4 Peak

BSA (100 mg) 5 2.6 1.7-2.2 1250-1960 1.7

BSA (100 mg) 25 4.8-4.9 4.1-4.6 8720-8910 1.7
BSA (100 mg) 50 4.6-4.8 4.2-5.1 6720-7910 1.7
Flu antigen (~100 mg) 25 4.8 4.2-4.4 8270-8460 1.7
Flu antigen (~100 mg) 50 4.5-4.6 4.3-4.6 6970-7060 1.7
Anthrax vaccine (~50 mg) 50 4.6-4.7 4.5-4.9 7510-7600 1.7

between Na and NH
4 peaks and that between NH4 and K peaks when the Kjeldahl digests were diluted to 25 or 50 mL, we
were poor (Table 1); (c) the number of Theoretical Plate was low observed that (a) the baseline is stable (Fig. 1), (b) the resolution
(Table 1); and (d) the Na peak eluting at about 6 min show a between Na and NH 4 peaks and that between NH4 and K

shoulder peak, which we could not explain. In addition, the results
showed signicant variability for independent replicate digests
prepared on the same day and different days (poor precision).
The same problem was also observed when the Kjeldahl
digests were diluted to 10 mL, albeit to a lesser extent. However,
peaks improved signicantly (Table 1), (c) the number of
Theoretical Plate increased signicantly (Table 1), (d) the Na
peak showed no shoulder (Fig. 1), and (e) inter and intra-day
precision improved considerably (see under Method Validation
for details).

Figure 1. Chromatograms of Kjeldahl digests of (a) BSA, (b) mock Anthrax vaccine, (c) mock Flu antigen preparation, and (d) Blank. The experimental conditions are as described
in the text.
 H. Wang et al. / Journal of Pharmaceutical Sciences 105 (2016) 1851-1857 1855

IonPac CS16 is a carboxylate-based cation-exchange column. At Table 2

pH ~1.5 of 30 mM MSA, carboxylate groups of the stationary phase %Recovery of Pure Proteins

are mostly protonated and the capacity of the column is signi- Protein Amount Digested (mg) %Recovery
cantly reduced. Thus, the problems are most likely due to column BSA 10 103 (n 3, 1 day)
overload when the digest is diluted to 5 mL, which also explains the 15 97.3 (n 3, 1 day)
lack of baseline stability of the chromatogram and the low Theo- 25 100-106a
retical Plate (Table 1). 50 93.6-101b
100 95.4-98.3c
Based on these initial results, all of our subsequent results were
HSA 25 105 (n 3, 1 day)
obtained by diluting Kjeldahl digests to 25 mL, unless stated 100 98.2-105c
otherwise. OVA 25 102-105c
To optimize eluent concentration for IC, we chromatographed 100 97.6-101c
BSA digests using eluents that contain varying MSA concentrations a
Over 5 days, 3-6 replicates each day.
between 20 and 48 mM (containing 10 mM oxalic acid) and ow b
Over 12 days, 3-6 replicates each day.
c
rates between 1.0 and 1.2 mL/min. The chromatograms obtained Over 3 days, 3-5 replicates each day.

with 20 and 30 mM MSA are similar, with comparable peak reso-

lution, except that the elution time was 30-35 min in the presence
of 20 mM MSA. The chromatograms obtained with 48 mM MSA are
similar to those obtained with 30 mM MSA and presents no
particular advantage in terms of elution time or peak resolution.
We did not observe any difference in the chromatogram between
ow rates 1.0 and 1.2 mL/min, except that longer elution time is
necessary when the ow rate is 1.0 mL/min. Thus, we optimized the
elution conditions as described in the Materials and Methods
section.
Figure 1a shows a representative chromatogram of BSA using
the nal method. The ammonium peak is eluted at ~7.5 min
between the sodium peak eluting at ~6.0 min and a very large
potassium peak eluting at ~11.0 min. Similar chromatograms were
also obtained with HSA and OVA (not shown), as well as with the
formulated Anthrax vaccine (Fig. 1b) and Flu antigen (Fig. 1c). The
doublet peak at ~2.5 min is present in the chromatograms of all
Kjeldahl digests, including those of the blank (Fig. 1d), but not in the
chromatograms of the standard, which is not digested. Therefore,
we conclude that this peak is due to the digestion solution. We did
not investigate origin of this peak further because it does not
interfere with the assay.
We also observed a small peak at the elution time of ammonium
peak (~7.5 min) when the Kjeldahl digestion blank is analyzed by IC
(Fig. 1d) but not when water (without digestion) is injected. To
identify the source of this peak, we analyzed copper sulfate, po-
tassium sulfate, and sulfuric acid individually at the same respec-
tive concentrations as in the Kjeldahl digestion mixture (and
diluted to 25 mL to mimic sample preparation) (chromatograms
not shown). We found that a small peak was present at the elution
time of the ammonium peak in copper sulfate and potassium sul-
fate but not in sulfuric acid, indicating that the peak at ~7.5 min in
the Kjeldahl digestion blank was due to contaminants present in
the Kjeldahl digestion solution. Consequently, we included a blank
correction step for the analysis of proteins but not for standards
because standards are not digested. Similar blank correction has
been recommended for the traditional Kjeldahl and micro-Kjeldahl
methods.2-5
In order to evaluate the effect of volume of protein solutions
digested, 100 and 500 mL of solutions containing 25, 50, and 100 mg
BSA were digested, while keeping the amounts/volumes of diges-
tion solution and K2SO4 the same. With 100 mL containing 50 and
Method Validation
Linearity. Linearity of the standard was evaluated at 0.01-1.0, 0.1-
5.0, and 0.05-100 mg/mL ranges of ammonium. The results show
linear plots of peak areas versus ammonium concentrations with
R2
0.995 in the ranges 0.01-1.0 and 0.1-5.0 mg/mL. However, the
plot was found to deviate from linearity above 5 mg/mL. The results
in the range 0.05-100 mg/mL show excellent t to a quadratic
regression t (R2
0.995). Similar results for the analyses of
ammonium by IC have been reported and have been attributed to
the formation of a weak base (NH4OH) in the suppressor.6,7
Considering that the concentration of protein is low in biological
samples and that higher dilutions of the Kjeldahl digests were
found to give more robust results because of reduction of inter-
ference due to the potassium ion, we chose 0.01-1.0 mg/mL of
ammonium to construct the standard curve routinely.
Linearity of the method evaluated using BSA in the range 10-100
mg protein digested shows R2 0.999. The limit of quantitation of
the method for proteins estimated from this linearity data by the
standard deviation of the response and the Slope method, as
described in ICH Q2(R1) guideline,23 is 9.4 mg. Because the lowest
quantity of protein at which we evaluated precision and accuracy is
10 mg (Table 2), we set the LOQ of the method at 10 mg.
Precision. With 50 mg BSA digested, the relative standard deviation
values for intra-day (6 replicate digestions) and inter-day (12 days)
evaluations were 3.7% and 4.0%, respectively. The RSD for triplicate
analyses are 7.9% at 10 mg, 5.8% at 15 mg, and 1.6% at 100 mg of BSA
digested. The RSD for inter-day precision (5 days) is 9.2% when 25
mg BSA digested. The RSD for inter-analysts (2) precision is 4.8%
with 100 mg BSA digested.
Accuracy. The accuracy of the method was evaluated by comparing
(a) measured concentrations of pure proteins by the current
method with those determined from absorbance measurements at
280 nm using extinction coefcients (280) of the respective
Table 3
Comparison of Protein Concentrations Determined by Orthogonal Methods

100 mg of proteins, the results show 100.2% and 95.4% recoveries, Protein/Sample Concentration (mg/mL)a
respectively. With 500 mL, the recoveries are 103.6%, 93.6%, and
BCAb Lowryb Micro-Kjeldahl Kjeldahl-IC (Current Method)
98.2% for 25, 50, and 100 mg protein digested, respectively. The
BSA-1 101 102
recoveries are comparable to the recoveries shown in Table 2
BSA-2 86.7 92.2
(under Accuracy), indicating that the volume of protein solutions HSA-1 99.7 80.3 107 104
can be increased up to 500 mL, while keeping the amount of protein HSA-2 95.4 81.9 107 109
in the range 25-100 mg. Thus, the method permits digesting larger a
All results are average of 3 replicates each on 2 days.
volumes of protein solutions, if necessary, when the solution is b
BSA is used as the standard for the determination of concentration by BCA and
dilute. Lowry methods. Hence BSA was not assayed by BCA and Lowry methods.
1856 H. Wang et al. / Journal of Pharmaceutical Sciences 105 (2016) 1851-1857

Table 4
Protein Concentrations in Vaccine Samples by Orthogonal Methods

Vaccine Protein Concentration (mg/mL)a Ratio Kj-IC to Micro-Kj Ratio Kj-IC to Lowry

Amount Digestedb Average Micro-Kjeldahlb,d Lowryc

25 mg 50 mg 100 mg

Anthrax-1 165 172 0.96

Anthrax-2 142 151 0.94
Flu-1 929 1010 999 980 1050 1020 0.94 0.96
Flu-2 1779 1560 1580 1642 1610 1650 1.02 1.00
Flu-3 1846 2210 1960 2010 2130 2130 0.94 0.94
a
All results are average of 3 replicates each on 2 days.
b
Anthrax vaccine samples were digested without dilution (50-60 mg of protein based on the results of the traditional Kjeldahl assay).
c
The amounts Flu antigens digested are based on Lowry protein concentration of the respective samples using BSA as the standard. They were diluted with water such that
250 mL contained the stated amount.
d
About 1 mg of the protein was digested for traditional micro-Kjeldahl method based on Lowry protein concentration.

proteins (%Recovery) (Table 2) and (b) results obtained by the
current method and other methods (Table 3).
Our literature search showed that different authors have re-
ported different experimentally determined 280 values for the pure
proteins we analyzed by Kjeldahl-IC.21 Hence, we calculated 280 of
the proteins based on their amino acid sequences, as described by
Pace et al.22 Table 2 shows that the results obtained by the current
method for BSA, HSA, and OVA at different concentrations are in
good agreement with concentrations obtained from their absor-
bance measurement using calculated 280.
Table 3 shows a comparison of results obtained for 2 indepen-
dent preparations each of BSA and HSA at about 100 mg/mL nominal
concentrations by our method with other commonly used methods
of the determination of protein concentration, including the micro-
Kjeldahl method. The results obtained by the current method show
good agreement with those obtained by the BCA method for HSA,
and by the micro-Kjeldahl method for BSA and HSA. However, the
results obtained by the Lowry method is ~20% lower for HSA. BSA is
used as the standard for the determination of the concentrations of
HSA by the BCA and Lowry methods. It should be noted that Lowry
assay involves reactions of copper ions with the peptide bonds
under alkaline conditions with the oxidation of aromatic protein
residues. Thus, the method is biased toward aromatic amino acids
present in the protein.20 It is possible that BSA is not a suitable
standard for concentration determination of HSA by the Lowry
method because these two proteins contain different amounts of
aromatic amino acids.
Vaccine Samples
Protein Concentration. Figures 1b and 1c show the chromatograms
of one lot each of the mock anthrax vaccine (Anthrax-1) and Flu
antigen (Flu-1) preparations, respectively. Similar chromatograms
are also obtained with other mock anthrax vaccine and the Flu
antigen preparations. The only difference between chromatograms
of pure proteins (Fig. 1a) and vaccines (Figs. 1b and 1c) is the
presence of a large sodium peak in the latter. This is because we
used lyophilized materials for pure proteins but the vaccine sam-
ples (or monovalent bulks) were in liquid formulations, which
contain sodium salts as excipients. Thus, the large sodium peak in
vaccine samples are due to the sodium salts present in vaccine
formulations. However, the Na peak is well resolved from the
ammonium peak (Fig. 1 and Table 1). Table 1 also shows that per-
formance characteristics of the ammonium peaks are comparable
between BSA and the vaccines, even though the latter contains
large amount of the neighboring Na ion.
Table 4 shows that for mock anthrax vaccine samples (~50 mg of
protein digested), the protein concentrations obtained by Kjeldahl-
IC method are 95 1% of that obtained by the micro-Kjeldahl
method. For the 3 Flu antigen preparations, the results obtained
with 25, 50, and 100 mg of protein by our method are comparable to
the results obtained by the micro-Kjeldahl and Lowry methods,
with 94%-102% and 94%-100% recovery, respectively. It should be
noted that the micro-Kjeldahl method requires digestion of large
amount of proteins (about 1 mg), while our Kjeldahl-IC method
shows comparable results with 25-100 mg of proteins digested.
Anthrax samples contain aluminum as adjuvant. Our results
show that aluminum did not precipitate as hydroxide (with OH
ion generated by CERS) to choke the eluent ow in the presence of
10 mM oxalic acid, presumably because of the formation of a
negatively charged complex with oxalate in the same manner as
copper, which is removed by CERS and replaced by OH ion. For-
mation of stable and soluble anionic complex between aluminum
and oxalate has been reported.15
Solution Stability. The Kjeldahl digests of vaccine samples were
diluted to 50 mL and analyzed on the same day (day 0), next day
(day 1), and after 6 days (solutions stored at 2 C-8 C). The results
summarized in Table 5 show that the solutions are stable for 1 day
after digestion. Although the results obtained on day 1 are slightly
higher than those obtained on day 0 in all cases, the RSD between
the results on 2 days are 3.9% or less, which is consistent with the
inter-day precision of the assay.
Selective Protein Nitrogen Determination
To determine protein nitrogen in samples that also contain non-
protein nitrogen, samples containing 25 and 62.5 mg of BSA (in 250
mL) without and spiked with 1.0 mM L-glutamine as a representa-
tive nitrogen containing excipient were analyzed according to the
procedure described above. The results show 99.2% and 98.8%
recoveries of BSA, respectively, in the spiked samples compared to
the unspiked sample, with RSD of triplicate analyses of each sample
being 3.3%. Protein precipitation with 20% TCA alone did not work
well as it gave ~60% recovery in our hands. Because Kjeldahl digest
contains a large amount of K2SO4, we used KOH to dissolve the
precipitated proteins to increase K concentrations to evaluate a
worst-case scenario.
Table 5
Solution Stability of Kjeldahl of Vaccine Samples
Vaccine Day 0 Day 1 Day 6
Protein Concentration (mg/mL)
Anthrax-1 165 174 114
Anthrax-2 142 146 102
Flu-4 1210 1220 1028
Flu-5 1290 1366 1180
 H. Wang et al. / Journal of Pharmaceutical Sciences 105 (2016) 1851-1857
Conclusion
We present a simple, rapid, precise, and accurate method for the
determination of protein concentration in biological samples, which
combines Kjeldahl digestion and IC of the digest with suppressed
conductivity detection. Even though the method requires nano-
molar amount of nitrogen in samples, the results are comparable to
those obtained by the micro-Kjeldahl method. In addition, our
method reduces the digestion time to less than 1 h and eliminates
the need for the distillation and titration steps of Kjeldahl methods,
thereby reducing the analysis time signicantly while improving
accuracy and precision. The method is particularly useful for bio-
logical samples where sample limitation is an issue. The method
may be applicable also in food, environmental, and other areas.
Acknowledgments
The authors thank Dr. Arthur W. Fitchett of Thermo Scientic
(Dionex Corporation) for helpful discussions and suggestions.
References
1. United States Pharmacopeial Convention. USP 38eNF 33: General Chapters:
<1057> Biotechnology-derived Articles-total Protein Assay, Method 5. Rockville,
MD: United States Pharmacopeial Convention; 2015.
2. AOAC International. Ofcial Methods of Analysis. 18th ed. Arlington, VA: AOAC
International; 2007.
3. Persson J-A, Wennerholm M, O'Halloran S. Handbook for Kjeldahl Digestion: A
Recent Review of the Classical Method with Improvements Developed by FOSS.
Hilleroed: FOSS; 2008.
4. Egli H. Kjeldahl Guide. Switzerland: BCHI Labortechnik AG; 2008.
5. Labconco Corporation. A Guide to Kjeldahl Nitrogen Determination Methods and
Apparatus. Kansas City, MO: Labconco Corporation; 1998.
6. Small H. Ion Chromatography. New York. NY: Plenum; 1989:175.
7. Thermo Scientic (Dionex Corporation). Application Note 141: Determination of
Inorganic Cations and Ammonium in Environmental Waters by Ion Chromatog-
raphy Using the IonPac CS16 Column. Sunnyvale, CA: Thermo Scientic; 2001.
1857
8. Thermo Scientic (Dionex Corporation). Salt ConvertereCation Self-regenerating
Suppressor 300. Sunnyvale, CA: Thermo Scientic; 2008.
9. Srinivasan K, Bhardwaj S, Lin R, Pohl C. Suppressor design and detection for ion
chromatography. In: Bhattacharyya L, Rohrer JS, eds. Applications of Ion Chro-
matography for Pharmaceutical and Biological Products. New York, NY: Wiley;
2012:91-106.
10. Trevani LN, Roberts JC, Tremaine PR. Copper(II)-ammonia complexation equi-
libria in aqueous solutions at temperatures from 30 to 250 C by visible spec-
troscopy. J Sol Chem. 2001;30(7):585-622.
11. Jackson PE, Krol J, Heckenberg AL, Mientjes M, Staal W. Determination of total
nitrogen in food, environmental and other samples by ion chromatography
after Kjeldahl digestion. J Chromatogr. 1991;546:405-410.
12. de Medina HL, de Vargas MC, Marin J, Pirela D. Determination of total nitrogen
in water samples by means of high-pressure bombs and ion chromatography.
J Chromatogr A. 1994;671:287-293.
13. Pontes FVM, Carneiro MC, Vaitsman DS, et al. A simplied version of the total
Kjeldahl nitrogen method using an ammonia extraction ultrasound-assisted
purge-and-trap system and ion chromatography for analyses of geological
samples. Anal Chim Acta. 2009;632:284-288.
14. Cockerell LD, Woods PH. Equilibrium studies of the copper(II) oxalate complex
between an aqueous solution and an anion-exchange Resin. J Am Chem Soc.
1958;80:3856-3858.
15. Harrison WJ, Thyne GD. Predictions of diagenetic reactions in the presence of
organic acids. Geochim Cosmochim Acta. 1992;46:565-586.
16. Rowe RC, Sheskey PJ, Owen SC. Handbook of Pharmaceutical Excipient. London:
APhA/Pharmaceutical Press; 2009.
17. Moore JC, DeVries JW, Lipp M, Grifths LC, Abernethy DR. Total protein
methods and their potential utility to reduce the risk of food protein adul-
teration. Compr Rev Food Sci Food Saf. 2010;9:330-357.
18. Uniprot Consortium. Available at: http://www.uniprot.org. Accessed February
4, 2016.
19. Jones DB. Factors for Converting Percentages of Nitrogen in Food and Feeds into
Percentages of Proteins. USDA Circular No. 183. Washington, DC: USDA; 1931:1-
22.
20. Sigma. Technical Bulletin: Total Protein Kit, Micro Lowry, Peterson's Modication.
St. Louis, MO: Sigma; 2003.
21. Walker JH. The Protein Protocols Handbook. 2nd ed. Totowa, NJ: Humana Press
Inc; 2002.
22. Pace CN, Vajdos F, Lanette F, Grimsley G, Gray T. How to measure and predict
the molar absorption coefcient of a protein. Protein Sci. 1995;4:2411-2423.
23. International Conference on Harmonization. Q2(R1). Validation of Analytical
Procedures: Text and Methodology; 2005. Available at: http://www.ich.org/
leadmin/Public_Web_Site/ICH_Products/Guidelines/Quality/Q2_R1/Step4/
Q2_R1__Guideline.pdf. Accessed March 21, 2016.