TECHNICAL FIELD

The invention relates to a method of determining protein concentration in a selected material, such as a multi-component material comprising food. The invention also relates to a system for determining protein concentration in a selected material.

BACKGROUND ART

Traditionally protein determinations have been determined using wet chemistry methods such as Kjeldahl and Dumas digestion methods.

Generally, such wet chemistry methods are very time demanding, expensive, and relies on total nitrogen content then recalculated to protein content using assumed Jones factors (e.g. for milk the Jones factor is 6.38 gram protein per gram nitrogen).

Methods using IR instruments has also been applied, particular Foss MilkoScan. Such IR instruments are fast, but requires careful, regularly calibration.

US 2005/0270026 discloses a method for determining the content of at least one component e.g. protein, of a sample by means of a nuclear magnetic resonance pulse spectrometer. The method comprises the steps of initially saturating the magnetization of the sample, influencing the magnetization by a sequence of radio-frequency pulses such that the signal amplitude to be observed can be determined, wherein the signal amplitudes which are determined at each time by the longitudinal and transverse relaxation time T1 and T2 and/or T2 ^* and/or T 1 p, from which value for the content of the at least one component is determined, are measured at the same time in a cohesive experimental procedure. The content of the at least one component in the sample is, determined by measuring different relaxation influences. This method is rather complicated and has never been applied in practice.

DISCLOSURE OF INVENTION

The objective of the present invention is to provide a method of determining protein concentration in a selected material, which is fast, relatively simple and with a high accuracy even where the selected material comprises additional components such as water, small organic compounds, carbohydrates and/or fat.

In an embodiment, an objective of the present invention is to provide a system for use in determining protein concentration in a selected material, which is fast, relatively simple and with a high accuracy even where the selected material comprises additional components such as water, small organic compounds, carbohydrates and/or fat.

These and other objects have been solved by the invention or embodiments thereof as defined in the claims and as described herein below.

It has been found that the invention or embodiments thereof have a number of additional advantages, which will be clear to the skilled person from the following description.

The method of determining protein concentration in a selected material according to the invention has been found to be very fast and accurate.

The method and systems of the invention and preferred embodiments thereof will be described further below.

It should be emphasized that the term “comprises/comprising” when used herein is to be interpreted as an open term, i.e. it should be taken to specify the presence of specifically stated feature(s), such as element(s), unit(s), integer(s), step(s) component(s) and combination(s) thereof, but does not preclude the presence or addition of one or more other stated features. Reference made to “one embodiment” or “an embodiment” means that a particular feature, structure, or characteristic described in connection with an embodiment is included in at least one embodiment of the subject matter disclosed. Thus, the appearance of the phrases “in one embodiment” or “in an embodiment” in various places throughout the specification is not necessarily referring to the same embodiment. Further, the skilled person will understand that particular features, structures, or characteristics may be combined in any suitable manner within the scope of the invention as defined by the claims.

The term "substantially" should herein be taken to mean that ordinary product variances and tolerances are comprised.

Reference made to “one embodiment” or “an embodiment” means that a particular feature, structure, or characteristic described in connection with an embodiment is included in at least one embodiment of the subject matter disclosed. Thus, the appearance of the phrases “in one embodiment” or “in an embodiment” in various places throughout the specification is not necessarily referring to the same embodiment. Further, the skilled person will understand that particular features, structures, or characteristics may be combined in any suitable manner within the scope of the invention as defined by the claims.

The method of the invention of determining protein concentration in a selected material, comprises

• preparing a liquid sample of the selected material to provide that it has a preselected concentration of a selected ion,

• applying the sample in an NMR apparatus, e.g. by inserting the sample into the NMR apparatus or by pumping the sample into the NMR apparatus,

• performing an NMR reading of at least one NMR signal of an isotope of the selected ion and generating time domain data representing signal dependence on T1 (longitudinal) or T2 (transverse) relaxation times, • determining an isotope relaxation rate (R1 and/or R2) from the time domain data,

• correlating the determined isotope relaxation rate (R1 and/or R2) to a standard curve of isotope relaxation rates relative to known protein concentrations and

• determining the concentration in the selected material.

The selected ion is an ion of an alkali metal element, an alkaline earth metal element or a halogen element.

Thanks to the inventors of the present invention, a new and very effective method and system for protein determination as been provided. The inventors have found that the relaxation rate of an isotope of the selected ion depends on the concentration of protein in the sample. Hence by determining the relaxation rate and correlating to a standard curve of isotope relaxation rates relative to known protein concentrations, the protein concentration in the liquid sample can be determined, and hence the concentration in the selected material may be determined by correlating the determined concentration in the liquid sample to the amount of the selected material per unit liquid sample.

Without being bound by the theory, it is believed that the protein interacts with the selected ion in the liquid sample. This interaction enhances the relaxation rate of isotopes of the selected ion, and the more protein, the faster relaxation. This effect has been found for ions of alkali metal elements, alkaline earth metal elements and halogen elements.

The interaction between protein and selected ion depends on the selected ion and some ions have been found to have strong interaction with the proteins in the sample and thus a relaxation rate that is dependent on the protein concentration. Such ions includes Li ⁺ , Na ⁺ , K ⁺ , Rb ⁺ , Mg ²⁺ , F, Cl , B r and G, which accordingly are preferred to apply as the selected ion. It should be observed that the selected ion could also include a mixture of two or more selected ions, in such case a standard curve of isotope relaxation rates for isotopes of each of the selected ion are provided and the determined isotope relaxation rate for the liquid sample is correlated to both of the standard curved and the concentration is determined as the average. Hence using two or more selected ions may provide an even more accurate result. However, generally the results determined from the two or more standard curves are very close or practically identical.

When using one or more of the above preferred ions, it is preferred that the isotope comprises one or more corresponding isotope selected from ⁶ Li, ⁷ Li, ²³ Na, ³⁹ K, ⁸⁵ Rb, ⁸⁷ Rb, ²⁵ Mg, ¹⁹ F, ³⁵ CI, ³⁷ CI, ⁷⁹ Br , ⁸¹ Br, and ¹²⁷ I.

In a preferred embodiment the selected ion is Na ⁺ , K ⁺ , P, or Cl and the isotope is the corresponding ²³ Na, ³⁹ K, ¹⁹ F, or ³⁵ CI.

In an embodiment, there are two or more selected ions and the determined relaxation rate is determined for corresponding isotopes of the two or more ions and these are correlated to respective standard curve of isotope relaxation rates for the corresponding isotopes.

In an embodiment, there are two or more selected ions selected from Li ⁺ ,

Na ⁺ , K ⁺ , Rb ⁺ , Mg ²⁺ , F , Cl , B r, and G and the determined relaxation rate is determined for corresponding isotopes of the two or more ions e.g. the isotopes mentioned above and these are correlated to respective standard curve of isotope relaxation rates for the corresponding isotopes.

Nuclear magnetic resonance - abbreviated NMR - is a phenomenon, which occurs when the nuclei of an isotope with a nuclear spin in a magnetic field absorb and re-emit electromagnetic radiation. The emitted electromagnetic radiation has a specific resonance frequency, which depends on the strength of the magnetic field and the magnetic properties of the isotope. NMR allows the observation of specific quantum mechanical magnetic properties of the atomic nucleus. Many scientific techniques exploit NMR phenomena to study molecular physics, crystals, and non-crystalline materials through NMR spectroscopy. NMR is also routinely used in advanced medical imaging techniques, such as in magnetic resonance imaging (MRI).

An NMR measurement is performed by NMR spectrometer or NMR relaxometer, here also referred to as an analyzer.

The terms "spectroscopy" and "spectrometry" are used interchangeable and in the same way a spectroscope is the same as a spectrometer.

NMR spectroscopy is well known in the art and has for many years been applied for laboratory measurements in particular where other measurement methods could not be used. NMR spectroscopy is performed using an NMR spectrometer. Examples of spectrometers are e.g. described in US 6,310,480 and in US 5,023,551. The term NMR spectrometer also includes a NMR relaxometer.

General background of NMR formation evaluation can be found, for example in U.S 5,023,551.

A general background description of NMR measurement can be found in "Understanding NMR Spectroscopy" by James Keeler, John Wiley & Sons Ltd, 2005 or in a practically oriented setting in, e.g., "NMR Logging Principles and Applications" by George R. Coates et al, Halliburton Energy Services, 1999. See in particular chapter 4.

The terms 'NMR reading' and "NMR measurement" are used interchangeable. It should be observed that used in singular for also includes the plural form i.e. a plurality of NMR readings unless other is specified. Often many NMR readings are performed and an average of the readings is used for the further analysis.

The term "relaxation" describes processes by which nuclear magnetization excited to a non-equilibrium state return to the equilibrium state. In other words, relaxation describes how fast spins "forget" the direction in which they are oriented. Methods of measuring relaxation times T1 and T2 are well known in the art.

The relaxation time T2 is herein used to include "apparent T2" (sometimes also called T2*). Apparent T2 includes a contribution caused by instrumental effects, such as magnetic field inhomogeneity. Instrumental effects (e.g. large magnet inhomogeneity) may cause that measured T2 relaxation times reflect apparent T2 relaxation times rather than pure natural T2 relaxation times. However, such instrumental effects may for example be minimized using a proper echo-train pulse sequence (e.g. CPMG) and may often be ignored (at least for the intensity determination), specifically where the same instrument is used for generating the standard curve and for performing the measurement.

The NMR spectrometer advantageously comprises an integrated or an external computer associated with a memory.

T2 relaxation is also called the transverse relaxation. Generally, T2 relaxation is a complex phenomenon and involves decoherence of transverse nuclear spin magnetization. T2 relaxation values are substantially not dependent on the magnetic field applied or the NMR frequency applied during excitation of the *H nuclei. Hence, it is preferred that the generated data comprises T2- dependent time-domain data. When using Tl-dependent time-domain data, it is preferred that the magnetic field applied and/or the NMR frequency applied for generating the standard curve is the same or within +/- 20 % from the magnetic field applied and/or the NMR frequency applied when performing the protein concentration determination.

In an embodiment, the magnetic field applied and/or the NMR frequency applied for generating the standard curve is the same or within +/- 10 %, such as within +/- 5 %, such as within +/- 1 %, from the magnetic field applied and/or the NMR frequency applied when performing the protein concentration determination. A standard technique for measuring NMR signals and obtaining information about the spin-spin relaxation time T2 utilizing CPMG (Carr-Purcell-Meiboom - Gill) sequence is as follows. As is well known after a wait time that precedes each pulse sequence, a 90-degree exciting pulse is emitted by an RF antenna, which causes the spins to start processing in the transverse plane perpendicular to the external magnetic field. After a delay, a first 180-degree pulse is emitted by the RF antenna. The first 180-degree pulse causes the spins, which are dephasing in the transverse plane, to reverse direction and to refocus and subsequently cause an initial spin echo to appear. A second 180-degree refocusing pulse can be emitted by the RF antenna, which subsequently causes a second spin echo to appear. Thereafter, the RF antenna emits a series of 180-degree pulses separated by a short time delay. This series of 180-degree pulses repeatedly reverse the spins, causing a series of "spin echoes" to appear. The train of spin echoes is measured and processed to determine the spin-spin relaxation time T2.

In an embodiment, the refocusing RF pulse(s) is/are applied after the exciting RF pulse with an echo-delay time-period between the exciting RF pulse and the subsequent refocusing RF pulse. In the case of multiple echoes, the refocusing RF pulses are typically separated by twice the delay from the exciting RF pulse to the first refocusing RF pulse. The echo-delay time (also called echo time TE) is preferably of about 500 ps or less, more preferably about 150 ps or less, such as in the range from about 50 ps to about 100 ps depending on the homogeneity of the magnetic field applied (here assuming an inhomogeneity of the applied magnetic field of about 500 ppm, while longer an echo-delay time is suitable if a more homogenous magnetic field is applied).

This method is generally called the "spin echo" method and was first described by Erwin Flahn in 1950. Further information can be found in Flahn, E.L. (1950). "Spin echoes". Physical Review 80: 580-594, which is hereby incorporated by reference. A typical echo-delay time is from about 10 ps to about 50 ms, preferably from about 50 ps to about 200 ps. The repeat delay time (also called wait time TW) is the time between the last CPMG 180° pulse and the first CPMG pulse of the next experiment at the same frequency. This time is the time during which magnetic polarization or T1 recovery takes place. It is also known as polarization time. The repeat delay time, typically in the order of 2 s, should typically be sufficiently long to ensure full recovery of the polarization, but may also be shortened to obtain Tl-dependent data.

This basic spin echo method provides good results for obtaining Tl-modulated data by varying TW and T2-modulated data by varying the echo-delay time or by using plurality of refocusing pulses.

The delay between refocusing pulses is also called the Echo Spacing and indicates the time identical to the time between adjacent echoes. In a CPMG sequence, the TE also reflects the time between 180° pulses.

The data representing signal dependence on T2 (T2-dependent data) may advantageously be acquired using a spin echo train experiment (e.g. the CPMG pulse sequence) or a series of spin echo experiments. The acquisition of T1 information may advantageously comprise one or more acquisitions with the saturation recovery or inversion recovery, or modified experiment versions based on these experiments.

This CPMG method is an improvement of the spin echo method by Hahn. This method was provided by Carr and Purcell and provides an improved determination of the T2 relaxation values, which again allows for better quantitative determination of the signal intensity via more precise consideration of T2 effects obtained from single or multi curve fitting for most precise envelope of spin echo amplitudes.

Further information about the Carr and Purcell method (which is a basic echo-train method and the fundament for the CPMG method) can be found in Carr, H. Y.; Purcell, E. M. (1954). "Effects of Diffusion on Free Precession in Nuclear Magnetic Resonance Experiments". Physical Review 94: 630-638, which is hereby incorporated by reference.

The preparation of the liquid sample depends largely on the selected material and whether or not it is in liquid form itself. Liquid sample means any sample comprising free liquid. Advantageously, at least about 50 % by weight is in liquid form, such as at least about 60 %, such as at least about 70 %, such as at least about 80%, such as at least about 90 % by volume is in liquid form. In an embodiment, the liquid sample is free of solid material. The selected material may in principle be any kind of material suspected of containing protein. If the selected material is solid, a sufficient amount of solvent is advantageously added and the sample may be comminuted e.g. using a blender or similar equipment.

If the selected material is a liquid with solid parts, the solid parts may advantageously be comminuted.

It is desired that at least some and preferably a substantially amount such as 50 % by mole or more of the protein is dissolved in the liquid sample.

It has been found that the protein specifically interact with selected ion when the protein is dissolved in the liquid sample. The selected ion as well as other ions of dissolved salts in the liquid sample may enhance solution of the protein.

Specifically, it is desired that the liquid sample comprising the selected material is prepared by the same method as preparation of the reference liquid samples used for generating the standard curve. Thereby the amount of dissolved protein for a given protein concentration may be substantially identical. In an embodiment, the preparation of the sample comprises withdrawing sample material from the selected material and providing that at least a part of the protein is in dissolved condition

The solubility of the protein may e.g. be increased by adjusting the pH value of the sample and/or by adding salts, such as NaCI, Na2SO, (NH4)2SC>4 or Na2SC>4, or by adding a detergent, such as sodium dodecyl sulfate (SDS). It should be observed that the pH value also may have influence on the protein interaction with the selected ion and hence on the relaxation rate as described further below. Generally, it is desired that the pH value is 4 or higher, such as 5 or higher and more preferably 6 or higher e.g. up to 9 or 10.

The liquid sample may be shaked or stirred for a desired time to ensure a good solubility. In addition the preparation of the liquid sample, may be comprise heating the liquid sample to increase protein solubility, preferably heating the liquid sample to a temperature above 30 °C, but less than coagulation temperature, such as to a temperature of from about 40 °C to about 50 °C.

In an embodiment, where the selected material is solid or comprises solid parts, the preparation of the sample comprises comminuting the material and providing that at least a part of the protein in dissolved condition.

In an embodiment, the sample subjected to the NMR reading is a fraction of a mother sample, which mother sample is prepared to have the preselected concentration of the elected ion. The mother sample is advantageously a homogenous sample. In an embodiment, the mother sample is inhomogeneous e.g. by being a two phases sample (e.g. a polar phase and a non-polar phase) resulting in a different protein and/or ion concentration in the sample subjected to the NMR reading relative to the average concentrations in the mother sample : in such case it is desired that the method comprises compensating for inhomogeneity of protein and/or ion in the mother sample. In an embodiment, the preparation of the liquid sample comprises adding a solvent for the protein, preferably, the solvent is an aqueous solvent.

Advantageously, the preparation of the liquid sample comprises adjusting the pH value, preferably by adding a buffer, adding an acid and/or adding a base. The prepared liquid sample may advantageously have a pH value between 6 and 9, such as between 7 and 9, such as about 8.

To ensure a very good influence of the protein on the selected salt it has been found that a substantially amount of the protein advantageously is in dissolved condition in the liquid sample. Advantageously, at least about 40 %, such as at least about 50 %, such as at least about 60 %, such as at least about 80 %, such as at least about 90%, such as substantially all of the protein in the sample is dissolved or the method comprising providing that at least about 40 %, such as at least about 50 %, such as at least about 60 %, such as at least about 80 %, such as at least about 90%, such as substantially all of the protein in the liquid sample is dissolved.

In an embodiment, the preparation of the liquid sample comprises digestion of proteins by enzymatic digestion or by chemical hydrolysis, optionally catalysed by acidic or alkaline conditions. After the digestion the pH value may be adjusted if desired. The digestion of protein is in particular desired where the selected material comprises large proteins, such as about 40.000 Dalton or larger or large quantities of other macromolecular species, such as carbohydrates. The protein digestion may increase the solubility or accessibility of the proteins.

Where the selected material comprises protein complex(es), the preparation of the liquid sample may advantageously comprise extracting proteins from the one or more protein complexes. This extraction may preferably comprise adding a detergent and/or buffer solution, such as sodium dodecyl sulfate (SDS) and/or Triton-X. Alternatively or in addition the extraction may comprise a heat treatment and/or pressure treatment, such as a pulsed pressure treatment.

The protein complex may e.g. comprise two or more associated polypeptide chains linked by non-covalent protein-protein interactions. The protein complex may for example have a quaternary structure, such as hemoglobin.

Where the selected material comprises cell bound proteins, referred to a cell protein complexes, the extraction of proteins may advantageously comprise subjecting the cells to cell lysis.

In an embodiment, the preparation of the liquid sample comprises denaturation of proteins using detergent, such as sodium dodecyl sulfate (SDS) and/or chelating agents such as Ethylenediaminetetraacetic acid (EDTA). The detergent may ensure that at least a part of the protein remains dissolved. In an embodiment, the method comprises adding urea to increase solubility. In an embodiment, the preparation of the liquid sample comprises addition of at least one molecular compound adapted for binding to one or more proteins expected to be in the in the sample. The molecular compound may be specific for a selected polymer or type of polymers, such as a specific ligand for a polymer. The molecular compound may for example be a dodecyl sulfate, such as sodium dodecyl sulfate (SDS).

The molecular compound, such as SDS may result in a modified interaction with the preselected ion. The molecular compound, such as SDS may be a surfactant and/or may work by disrupting non-covalent bonds in the proteins, and so denaturing them. In an embodiment, the concentration of the selected ion is on a desired level without adding further salt, which can dissolve to provide the selected ion. In an embodiment, salt, which can dissolve to provide the selected ion, may be added to increase solubility of the protein. In an embodiment, the preparation of the sample comprises adding a salt or a diluted salt to the sample to reach the preselected concentration of the selected ion.

Advantageously, the preselected concentration of the ion is at least about 0.001 mM (mmol/liter), such as at least about 0.01 mM, such as at least about 0.1 mM, such as at least about 1 mM, such as at least about 10 mM, such as at least about 100 mM.

It has been found that good determinations of protein concentration may be obtained even where the preselected ion concentration is relatively low. A lower ion concentration may require a longer measuring time. In addition, a desired preselected ion concentration depends on the selected ion because the NMR sensitivity varies from ion to ion. For example, the NMR sensitivity is approximately 20 times larger for ²³ Na than for ³⁵ CI, approximately 100 times larger for ¹⁹ F, and approximately 5 times less for ³⁹ K.

Flence, a preselected ion concentration is advantageously at least about 0.2 mM for ³⁵ CI, at least about 0.01 mM for ²³ Na, at least about 0.002 mM for ¹⁹ F, or at least about ImM for ³⁹ K. A higher preselected ion concentration is generally preferred, because the method is faster and more accurate. The preferred preselected ion concentration is thus about 100 mM or higher, such as about 500 mM or higher, such as about 850 mM or higher, e.g. up to about 5 M.

The preselected ion concentration may in principle be up to saturation, however a lower concentration is desired. Where the preselected ion is at its saturation, the liquid sample may comprises salt(s) of the ion. The salt(s) may for example be precipitated and optionally held outside the measuring site of the NMR spectrometer.

In addition, since the relaxation rate for salt bound ions are practically not affected by protein, signals originating from salt bound ions may be omitted from the generated time domain data. Advantageously, the preselected concentration of the ion is up to about 1 M, such as up to about 5 M, such as up to about 10 M, such as up to saturation.

In an embodiment, the selected material is a multi-component material comprising at least two different components including small organic compounds, carbohydrates, proteins and/or fat. Small organic compounds may for example be or comprise C1-C5 organic compounds, such as hydrocarbons and/or organic compounds comprising nitrogen, silicon, halogen or combinations thereof.

It has been found to be particularly beneficial to use the method of the invention for protein determination in food. In many countries, it is required to inform about the protein content of food product. Also, the content of protein may be applied as a quality parameter. Further, to ensure an adequate nutrition of individuals, it may be desirable that the protein content of the food it determined with a high reliability and accuracy.

In practice, the NMR apparatus may be movable, e.g. incorporated into a carriage e.g. corresponding to a suitcase trolley in size and hence the determination of protein concentration may be determined anywhere, e.g. anywhere in a factory or in a shop, such as a supermarket, e.g. for control of the protein content.

In an embodiment, the selected material comprises a food product, such as a dairy product, meat, fish, vegetables or any fragments or combinations thereof.

In an embodiment, the selected material and/or the liquid sample comprises at least one of water, carbohydrates and/or fat in addition to the protein, preferably the sample comprises at least one of carbohydrates and fat.

The various component of the selected material need not be separated from the liquid sample. This makes the method highly advantageous.

Advantageously, the liquid sample is an aqueous sample. In an embodiment, where the food product comprises a liquid food product, the liquid sample may a dilution or a concentration of the liquid food product.

In an embodiment, the liquid sample comprises a solid food product dissolved and/or suspended in liquid, such as water. In an embodiment, the preparation of the sample comprises adding a buffer to the sample to stabilize and/or adjust the pH value of the prepared liquid sample.

As mentioned above, the inventors have found that the pH value may in some embodiments have influence on the protein determination and generally, it is desired that the standard curve is generated using measurements at same pH value as the liquid sample in the protein determination.

Adding buffer may be important for controlling pH as it may be important to get accurate measurements with method where the liquid sample without the buffer has very variable pH and it may accordingly be beneficial to ensure a uniform pH after mixing with salt (saline) as the pH can affect the interaction between ion and protein.

The optimum pH value will vary according to the ion used - and may also slightly depend on the sample type measured. For example, if the liquid sample comprises milk it may not be desirable to introduce coagulation because this may affect the interaction between the selected ion and the protein.

In an embodiment, the method of the invention is applied as a test in a production process, such as in a fermentation process. For example, where a protein containing food or drug is produced during a fermentation process, the method of the invention may be applied as a process control and/or as a quality control.

In an embodiment, an NMR apparatus is connected to a fermentation tank, such that the preparation of the liquid sample comprising withdrawing a sample from the fermentation and optionally adding a salt to provide the preselected ion (this is not required if the Is already comprise the predetermined amount of the selected ion). Thereafter the NMR reading is performed and the liquid sample may be returned to the fermentation tank or the liquid sample may be discharged.

In an embodiment, the selected material comprises a slurry of organic material and the method comprises correlating the determined protein concentration to a content of organic nitrogen.

The slurry of organic material may for example be an animal slurry.

The liquid sample may be prepared from the slurry as described for food material above.

The content of organic nitrogen may for example be determined as an organic-N fraction estimating that a preset percentage of the organic nitrogen is in the form of protein bound nitrogen. The preset percentage of the organic nitrogen in the form of protein bound nitrogen may preferably be 75 % or more, such as 90 % or more, such as 100 %. Where a corresponding slurry in various dilutions is used as reference samples for the standard curve, the method preferably comprises determining the protein concentration correlated to a content of organic nitrogen.

In an embodiment, the isotope relaxation rate is the transverse relaxation rate (R2).

In an embodiment, the isotope relaxation rate is the longitudinal relaxation rate (Rl).

It has been found that where the isotope relaxation rate is the transverse relaxation rate (R2), the protein determination may be even more accurate.

The standard curve of isotope relaxation rates relative to known protein concentrations may advantageously be a standard curve of isotope relaxation rates relative to known protein concentrations in reference materials. The determination of the concentration in the selected material may be derived directly from the standard curve.

In an embodiment, the standard curve of isotope relaxation rates relative to known protein concentrations is a standard curve of isotope relaxation rates relative to known protein concentrations in prepared reference samples and the determination of the concentration in the selected material is derived from the standard curve optionally compensating for dilution of the selected material during the preparation thereof.

The NMR reading may advantageously be performed in a magnetic flux density of from 1 mT to 6 T. It has been found that even when using a low- resolution NMR apparatus e.g. configured for performing the NMR reading in a low field magnetic field, such as a magnetic field with a magnetic flux density up to 2 T, such as up to 1.5 T, the accuracy of the protein determination may be highly accurate.

In an embodiment, the preparation of the sample comprises preparing a liquid sample of the selected material to have a preselected concentration of at least one additional ion and the method comprises generating time domain data representing dependence on T1 and/or T2 relaxation time, NMR reading of at least one NMR signal of an isotope of the at least one additional ion, determining an isotope relaxation rate (R1 and/or R2) from the time domain data, correlating the determined isotope relaxation rate (R1 and/or R2) to a standard curve of isotope relaxation rates relative to known protein concentrations and performing and additional determination of the concentration in the selected material.

In an embodiment, the method comprises generating the standard curve.

The standard curve may advantageously be generated using reference liquid samples with known content of protein

The generation of the standard curve may advantageously comprise • preparing a plurality of liquid reference samples with different and known protein concentrations to have the preselected concentration of the selected ion,

• for each reference sample, performing a NMR reading of at least one NMR signal of an isotope of the selected ion and generating time domain data representing signal dependence on T1 and/or T2 relaxation time,

• for each reference sample, determining an isotope relaxation rate (R1 and/or R2) from the time domain data,

• performing a linear regression including points of respective pairs of determined isotope relaxation rate and corresponding, known protein concentration,

The liquid reference samples with different, known protein concentrations may advantageously be liquid reference samples with directly known protein concentrations or the liquid reference samples with different, known protein concentrations may be liquid reference samples prepared from selected material(s) with known protein concentrations.

Advantageously the standard curve is provided by a method comprising extrapolating from at least one point of a pair of the liquid reference samples with different, known protein concentrations, such as from at least one point of a pair representing a highest protein concentration and/or of a pair representing a lowest protein concentration. Thereby a very accurate protein determination may be ensured.

The number of references used in the generating of the standard curve may be from a very few to hundreds or more. Normally 5 to 25 reference samples are desired.

In an embodiment, the plurality of liquid reference samples with different and known protein concentration comprises at least, 3, such as at least 4, such as at least 5, such as at least 8 reference samples. For ensuring an even more accurate protein determination, it is desired that the plurality of liquid reference samples with different and known protein concentration comprises a reference sample with a preselected highest protein concentration, wherein the preselected highest protein concentration is preselected to be larger than an estimated protein concentration of the liquid sample.

Advantageously, the plurality of liquid reference samples with different and known protein concentration are adjusted to have same pH value as the liquid sample.

In an embodiment, the plurality of liquid reference samples with different and known protein concentration has a percentage of dissolved protein relative to total protein, which is substantially the same percentage as the liquid sample.

In an embodiment, substantially all protein in the plurality of liquid reference samples are in dissolved condition.

In an embodiment, the preparation of the plurality of liquid reference samples comprises adding a salt or a diluted salt to the sample to reach the preselected concentration of the selected ion.

In an embodiment, the plurality of liquid reference samples with different and known protein concentrations are prepared by the same method as the preparation of the liquid sample.

The method may be performed for two or more ions and the determination may be performed as an average of determinations of the various ions.

In an embodiment, the method comprises preparing the liquid sample of the selected material to provide that it has a preselected concentration of at least two different selected ions, • performing NMR readings of at least one NMR signal of an isotope of each of the selected ions and generating time domain data representing signal dependence on T1 or T2 relaxation times,

• determining an isotope relaxation rate from the time domain data for each of the isotopes

• correlating the determined isotope relaxation rates to respective standard curves or a combined standard curve of isotope relaxation rates relative to known protein concentrations and

• determining the concentration in the selected material.

The invention also comprises a system for determining protein concentration in a selected material.

The system of the invention for determining protein concentration in a selected material comprises an NMR apparatus comprising a computer associated with a memory. The NMR apparatus is configured for generating a standard curve comprising

• receiving protein concentration data for each of a plurality of liquid reference samples

• for each of the plurality of liquid reference samples, performing a NMR reading of at least one NMR signal of an isotope of the selected ion and generating time domain data representing signal dependence on T1 and/or T2 relaxation time,

• for each reference sample, determining an isotope relaxation rate (R1 and/or R2) from the time domain data,

• performing a linear regression including points of respective pairs of determined isotope relaxation rate and corresponding, known protein concentration, and

• storing the generated standard curve of isotope relaxation rates relative to known protein concentrations in the memory.

Advantageously, the NMR apparatus further is configured for • subjecting a liquid sample prepared from the selected material to an NMR reading of at least one NMR signal of the isotope of the selected ion and generating time domain data representing signal dependence on T1 (longitudinal) or T2 (transverse) relaxation times,

• determining an isotope relaxation rate (R1 and/or R2) from the time domain data,

• correlating the determined isotope relaxation rate (R1 and/or R2) to the standard curve of isotope relaxation rates relative to known protein concentrations and

• determining the concentration in the selected material,.

The selected material, the preparation of the liquid sample, the ion, the NMR reading, the selected ion, the isotope and the determination of the relaxation rate and determination of protein concentration and any other steps of the method, which the NMR apparatus is configured for, may be as describe above.

In an embodiment, the system of the invention is adapted for performing the method of the invention as described above.

Advantageously the NMR apparatus is adapted for generating a standard curve according to the method described above.

The reference liquid samples may advantageously be as described above

The invention also comprises a system for determining protein concentration in a selected material, therein the system comprises a standard curve.

This system comprises an NMR apparatus comprising a computer associated with a memory, wherein the memory comprises data representing a standard curve of isotope relaxation rates of an isotope of a selected ion relative to known protein concentrations. Then the NMR apparatus is configured for subjecting a liquid sample prepared from the selected material to an NMR reading of at least one NMR signal of the isotope of the selected ion and generating time domain data representing signal dependence on T1 (longitudinal) or T2 (transverse) relaxation times,

• determining an isotope relaxation rate (R1 and/or R2) from the time domain data,

• correlating the determined isotope relaxation rate (R1 and/or R2) to the standard curve of isotope relaxation rates relative to known protein concentrations and determining the concentration in the selected material, The selected material, the preparation of the liquid sample, the ion, the NMR reading, the selected ion, the isotope and the determination of the relaxation rate and determination of protein concentration and any other steps of the method which the NMR apparatus is configured for may be as describe above.

In an embodiment, the system of the invention is adapted for performing the method of the invention as described above.

The NMR apparatus is advantageously adapted for performing the NMR reading in a magnetic flux density of from 1 mT to 6 T, such as in a magnetic flux density up to 2 T, such as in a magnetic flux density up to 1.5 T.

As it will be realized by the skilled person, the method of the invention may be combined with additional NMR measurements involving NMR enhancement such as enhancement involving polarization transfer, e.g. DEPT (Distortionless Enhancement by Polarization Transfer) or INEPT (Insensitive nuclei enhanced by polarization transfer). By using DEPT and/or INEPT combined with ¹³ C NMR readings concentrations of the presence of primary, secondary and tertiary carbon atoms (CH, CH2 and CH3 groups) may be determined. This determination may be combined with the determination of the method of the present invention and thereby further refine the determination of protein concentration. All features of the inventions including ranges and preferred ranges can be combined in various ways within the scope of the invention, unless there are specific reasons not to combine such features.

BRIEF DESCRIPTION OF THE EXAMPLES AND DRAWING The invention is being illustrated further below in connection with a few examples and embodiment and with reference to the drawings in which:

Fig. 1 shows a standard curve of ³⁵ CI isotope relaxation rates.

Figs. 2a and 2b contain lists of prepared liquid samples for the standard curve in Fig, 1. Figs. 3a and 3b show ³⁵ CI isotope relaxation rates as a function of diluted protein-N content (protein nitrogen content calculated from the protein content using the Jonas factor).

Fig. 4 shows ²³ Na isotope relaxation rates as a function of protein content.

Fig. 5 shows ³⁵ CI isotope relaxation rates as a function of pH value. Fig. 6 shows ²³ Na isotope relaxation rates as a function of pH value.

Fig. 7 shows ^ isotope relaxation rates as a function of pH value.

Fig. 8a shows ³⁵ CI isotope intensity as a function of dissolved NaCI (salt).

Fig. 8b shows ³⁵ CI isotope relaxation rate as a function of dissolved NaCI (salt). Fig. 9a shows ²³ Na isotope intensity as a function of dissolved NaCI (salt).

Fig. 9b shows ²³ Na isotope relaxation rate as a function of dissolved NaCI (salt).

Fig. 10 shows ²³ Na and ³⁵ CI relaxation rates versus sugar (sucrose) concentration in a clean solution of NaCI, NH4CI, and KCI with added sugar. To demonstrate the feasibility of the method, Figure 1 show measured ³⁵ CI R2 values plotted versus the protein contents for a variety of different milk materials as listed in figures 2a and 2b. The resulting curve demonstrates a standard curve of ³⁵ CI isotope relaxation rates.

The standard curve of ³⁵ CI isotope relaxation rates was provided as described in example 1.

Example 1

Liquid reference samples were prepared from the milk materials (supermarket milk products) listed in figures 2a and 2b. To some of the samples sucrose was added to get a milk with 9.7% sugar whereas normally 4.7%). All samples were analyzed after dilution and addition of NaCI to a salt concentration of 5%.

Some samples were prepared in duplicates, and some samples were prepared using dilution factors as indicated by the concentration factors (cone factors) given in figures 2a and 2b (here the concentration factors are normalized to a dilution by a factor of 6. This means that a concentration factor of 1 corresponds to dilution by a factor 6, a concentration factor of 0.4 corresponds to dilution by a factor of 15, etc.).

Figures 2a and 2b give an overview of samples used in this example and analyzed with the method using sodium and chloride measurements, with reference values for fat, carbohydrates and protein contents (declarations from cartons for supermarket product, calculated for milk powder solutions). Additionally the N contents in the digested samples (the samples prepared for measurement) are given. The N contents are calculated according to the reference values and using the Jonas factor of 6.38 as well as the actual used mass of the original sample.

A sample from each milk material was diluted by a factor 2.5-15 with water and NaCI solution to yield a final concentration of dissolved NaCI of about 5% prior to measurement. Generally, the sodium naturally present in milk products are almost similar and accordingly the sodium originally present in the milk materials was disregarded.

In addition, the concentration of dissolved NaCI in the original products were insignificant relative to the added concentration (e.g., 0.08-0.12 % in original products results in less than 0.05 % in the final dilutions prepared with 5 % added NaCI).

Each sample was applied in an NMR apparatus and was temperature stabilized. NMR readings of ³⁵ CI relaxation rates were performed and time domain data representing signal dependence on T2 relaxation times were generated.

Each sample was analyzed with 4 or 8 subsequent ³⁵ CI measurements for all samples, the first ³⁵ CI measurement was discarded and only 3 ³⁵ CI measurement were used for each sample to have a homogeneous sampling pattern. The ³⁵ CI measurements were each conducted with 128 scans, 300 ms recycle delay (50 s measuring time).

Since temperature-stabilization was included prior to all measurements the total measurement time was about 25 minutes. Temperature-stabilization could have been performed prior to applying in the NMR apparatus.

The results obtained from the acidified milk products (‘buttermilk’ and ‘tykmaelk’ (junket), marked with “ ^* ” in figure 2b)) is slightly inaccurate relative to the calibration, which is ascribed the low pH and coagulation in these samples. It is believed that these issues may be solved by using a modified preparation protocol including addition of detergent, buffer solution, chelating agent, and/or pH adjustment. However, in the present case, these results were therefore not used in generation of the standard curve.

The remaining results were used for generating the standard curve by performing a linear regression of respective pairs of determined isotope relaxation rate and corresponding, known protein concentration. The result is shove in figure 1. Example 2

The liquid samples used in example 1 were subjected to ²³ Na measurements and ¹ FI measurements. The ²³ Na measurements were conducted with 16 scans and 400 ms recycle delay.

Results for the ³⁵ CI and ²³ Na relaxation rates (R2=1/T2) from example 1 and example 2 are shown in figures 3a, 3b and 4.

For both isotopes ³⁵ CI and ²³ Na an excellent correlation is observed between R2 and the protein content (however, for Na results for samples with protein content lower than 3% are not shown. For the samples with protein content below 3% the correlation was not linear using the applied preparation and dilution scheme, this may be compensated by modifying the preparation and dilution scheme e.g. by less dilution). Also, correlation with the ¹ FI relaxation rates was previously shown (however, the contributions to ¹ FI R2’s are more complex). Using the present preparation procedure, we found that the response on protein concentration for ³⁵ CI relaxation times was superior to ²³ Na, which was superior to ¹ H.

Figure 3a and 3b show ³⁵ CI relaxation rate vs. protein-N content in actual sample. Figure 3a shows three subsequent analysis of each sample. The three blue lines indicate the linear correlation obtained using only the first, second or third evaluations for all samples. Figure 3b shows the mean values for each sample. The ‘s’ and ‘t’ samples were analyzed in duplicates. With the applied dilution, the protein content in the original milk product may be calculated as 38.3 ^* protein-N (e.g. 1 g/l on the axis corresponds to 3.83 % protein).

Figure 3b shows the same data as shown above in figure 1 , just recalculated to protein N content on the x axis and with labels on the sample points.

Figure 4 shows ²³ Na relaxation rate vs. protein content in the original (non- diluted, no additives, no pH regulation) samples. Based on the ³⁵ CI results (shown above), we obtained here an average absolute standard deviation (STD) of 0.10% given in protein-percent in the original sample. The corresponding STD on the difference between reference values and the calibrated NMR measurements is 0.23%. We note that approximately the same STD’s are obtained no matter if only dilution series or regular samples are included in the analysis.

By analyzing dilution series (individual sample prepared independently) of low-fat milk, reduced-fat milk, and whole milk, we found overlapping curves, which indicates that the measurements are independent of fat content.

From analysis with different added sugar contents, we found R2’s increasing with about 0.2 1/s per g/l sugar for Na, and 0.03 1/s per g/l sugar for Cl. Recalculated to the R2-increase for 4.7% sugar, this is 1.2 1/s for Cl and 9 1/s for Na, which again corresponds to protein content in the original sample of 0.12% for Cl and 2.4% for Na. Thus, Cl seems relative independent of sugar content, while Na R2 depend on the sugar content. This may be addressed by using same or similar (such as +/- 10 % relative) sugar content in reference liquid samples. Alternatively, sugar content may be determined and compensated for.

However, for Na, this results in an offset in R2 for samples with higher sugar contents, which can also be observed from the results shown in figure 4.

Example 3

Evaluation of pH effects

The pH value in the prepared sample may influence the obtained results. For instance, a lower pH may result in increased protonation of the amino groups in the proteins and a stronger interaction with the chloride ions leading to faster chlorine relaxation. On the other hand, lower pH may induce protein precipitation, which results in less contact between dissolved chloride and protein (and thus a much slower chlorine relaxation). For evaluation, we performed the following experiment to test the effects of different pH in the samples. Using a citric acid buffer solution, we adjusted sample pH to different values. These measurements were performed on skimmed milk, which was mixed with a NaCI solution and buffer (fixed content of citric acid and varied content of KOH) to give samples with -10% dissolved NaCI concentration and pH values from 3 to 7.

The samples at pH 6 and 7 were clear colorless liquids (with only very minor visible particles) while the others were milky with some precipitation.

The samples were analyzed for both ³⁵ CI, ²³ Na, and ¹ H relaxation times and intensities. The intensities were fairly independent of pH (intensities within roughly +/- 3%). The obtained relaxation rates show clear pH dependence and are shown in the three figures 5, 6, and 7.

For chlorine, the experimental results support that the Cl-to-protein interaction is stronger at lower pH, but only if pH is above the ‘coagulation point’. Also, we note that the pH dependence seems stronger for chlorine than for sodium and hydrogen. In any case, these tests indicate that it may be beneficial to control pH in the sample (for example by adding buffer during the preparation steps).

Figure 5 shows ³⁵ CI relaxation rates versus pH. Slow relaxation is observed at pH 3-4, while a much faster, declining, rate is observed from pH 5 to 7. We ascribe this to the existence of the two different regimes with coagulation occurring between pH 4 to 5. These results also indicate that a very accurate protein quantification may require a pH in a certain range (e.g. away from pH 4-5 for milk samples, since at pH 4-5 the measurement seems to be very sensitive to pH).

Figure 6 shows ²³ Na relaxation rates versus pH. Slightly faster relaxation with pH from 3 to 5, where after relatively stabilized from pH 5 to 7. We note that in this case the difference from highest to lowest is much less than for ³⁵ CI results (approx. x1.5 vs. x5).

Figure 7 shows ¹ H relaxation rates versus pH. In this case a decreasing relaxation rate with pH is observed, which could however also be seen as two relative stable regimes. Interestingly, we note that the difference between slowest and fastest R2 is only about 15%, which is much less than for ³⁵ CI (-500%) and ²³ Na (-50%).

Example 4 Evaluation of response at different salt (diluted NaCI) concentrations

To find an optimal salt concentration, it may be relevant to evaluate the relaxation rates observed at different salt concentrations. Using a skimmed milk sample (sample ‘t6a/b’ in figure 2b) we evaluated the relaxation rates for different salt concentrations. The results are shown in figures 8a/b and 9a/b. For both Na and Cl, the intensity plots show a good linear correlation as expected.

For chlorine, higher salt concentration leads to faster relaxation in the evaluated range, whereas for sodium, a minimum in relaxation rate is observed. Most likely, the minimum also exist for chlorine but at a lower concentration. These observations we ascribe to relaxation rates being faster when salt molecules interact more with each other (this concentration dependence is observed with high concentrations, while on the other hand also being faster when the concentration is lower so a larger fraction interact with the proteins and sugar in the milk in a given time. The figures 8a/b and 9a/b include results for samples with the milk diluted only with salt solution (red) and the same milk+salt samples diluted to 75% with additional water (blue) which lowers both the salt and protein concentrations. Since we have previously shown a linear correlation between protein concentration and relaxation rate, the latter results are scaled with the dilution factor (x0.75) to account for the changes due to protein concentration. In this way, the diluted samples are found at approximately the same curve as the original data and the effect of salt concentration seems to be additive to the effect from protein concentration. Figure 8a shows ³⁵ CI intensity versus salt concentration in the prepared sample.

Figure 8b shows ³⁵ CI relaxation rate versus salt concentration in the prepared sample.

Figure 9a shows ²³ Na intensity versus salt concentration in prepared sample.

Figure 9b shows ²³ Na relaxation rate versus salt concentration in prepared sample.

To find an optimal salt/ion concentration for protein measurements it may be important to consider and balance a number of aspects including the following:

1 ) Higher concentration of the selected ion improves the signal-to-noise ratio (SNR) resulting in shorter measuring times.

2) Longer relaxation may improve resolution at lower protein concentrations (may be controlled be adjustments of dilution factors in the sample preparation scheme).

3) Lower concentrations of the selected ion may reduce the relaxation- rate dependence on the concentration of the selected ion itself. Example 5

Evaluation of NMR response at different sugar concentrations

The concentration of carbohydrates may also affect the observed relaxation rates, see figure 10. However, the changes in relaxation rates seem much smaller than the changes following the protein concentration. It is believed that the dependence on sugar concentration occurs due to the change in viscosity induced by the sugar and/or increasing interactions between molecules. We note that changes in macro-viscosity are not necessarily equal to changes in the micro-viscosity. Figure 10 shows ²³ Na and ³⁵ CI relaxation rates versus sugar (sucrose) concentration in a clean solution of NaCI, NFUCI, and KCI with added sugar.

Estimating a linear trend, we found that R2 is increasing with about 0.2 1/s per g/l sugar for Na, and 0.03 1/s per g/l sugar for Cl. This seems to have a significant effect for protein measurements using ²³ Na, whereas the effect is much less significant using ³⁵ CI measurements as discussed above (see example 2).

In general, effects of various components may be evaluated prior to measurements and conduction of standard curves and optimization of sample preparation schemes.