WO2003006974A1

WO2003006974A1 - Method for quantifying different components in organic and biological material

Info

Publication number: WO2003006974A1
Application number: PCT/NO2002/000209
Authority: WO
Inventors: Tore Skjetne; Trond E. Singstad; Emil Veliyulin; Henrik W. Anthonsen
Original assignee: Leiv Eiriksson Nyfotek As
Priority date: 2001-07-09
Filing date: 2002-06-14
Publication date: 2003-01-23
Also published as: NO20013387D0; NO313899B1; NO20013387A

Abstract

Method for quantifying different components in organic and biologic material, by means of NMR and calibration curves, wherein pre-treatment of the samples is unnecessary. The method includes analysis in a NMR instrument where it for samples containing little water there is performed a spin echo pulse sequence and a solid echo pulse sequence, and for samples containing a lot of water it is performed a solid echo pulse sequence and a double quantum pulse sequence. The resultant NMR signals are recorded and the results are processed in a computer, where they are converted into amount of the different types of components, by means of calibration curves.

Description

Method for quantifying different components in organic and biological material.

The present invention relates to a method for quantifying different components in organic and biological material, according to the preamble of patent claim 1.

Background

In some processes, for example in the production of food articles for humans or animals, it is an advantage to quantify the amount of different components, such as water, fat, protein etc during the process. If this information is available immediately, the process may be reulated continuously so that it always is optimized, for example with regards to supply and use of raw material, and energy consumption. It is also an advantage to follow the product from production to consumer, in order to be able to adjust during the production for possible alterations during transport and storage.

There are several known methods for quantifying the said components. Diffusion filtered NMR techniques may be used online, and they provide a relatively quick answer. However, this method demands a lot during set up of the NMR instrument. The gradients must be well adjusted, and the diffusion will not necessarily always be sufficient differentiate between water and fat in all compositions, which may result in an inaccurate answer. An another known method is NIR, which also may be used online and provides an answer relatively quickly. In this method a lot of statistical handling is involved. However, the method is less accurate, and the answer depends on all the samples being fairly similar. The method also demands a lot of very accurate calibration.

Nowdays the most used method, is probably chemical analysis. This provides an accurate answer, but is ari offline method demanding a lot of processing of the sample, and thereby specially trained staff. It also uses solvents, which may present an environmental problem.

High resolution NMR spectroscopy may also be used, as it provides accurate answers with a lot of additional information regarding molecular and metabolic relations. However, it is an expensive method which is difficult to use on-line. It also demands specially trained staff.

NMR analysers give a number of fast and accurate analyses, without destroying the sample. With new automated NMR instruments, the analyses may be performed by nearly anyone. They fulfill nearly all needs for experimental determination of fat and water content. Determination of fat in samples containing small amounts of free water (for instance fish feed) using an NMR instrument, is based on the fact that the amplitude of the observed proton-NMR signal is directly proportional to the amount of hydrogen atoms being present in the sample. The unknown fat content is then calculated by calibrating the instrument to references with known fat content. The fat content of an unknown sample may simply and easily be decided in less than a minute, by using a calibration curve. One such known NMR technique is spin echo described among others by H. Firebolin, "Basic One- and Two-Dimensional NMR Spectroscopy", WILEY-VCH Verlag GmbH, 1998, p.173-175, incorporated herein by reference.

The said method is very sensitive, and well suited for systems such as fish feed, which normally contains very small amounts of free water. A small amount of water remaining after the dehydration is heavily related to the rest, and will thus not contribute to the amplitude of the observed signal, because the interaction between the water and the solid components gives the water a substantially shorter relaxation time than the fat. The said NMR technique is less suited to differentiate between fat and water in biological material, in cases where the amount of water is large.

Detection of NMR signals from large and immobilized molecules, such as proteins, has always provided challenging experimental work. These large molecules have extremely short NMR relaxation times, and the main part of the NMR instrument is not capable of providing reliable quantitative measurements of such signals, with the simple spin echo technique described above, mainly due to physical limitation in the detection electronics (receiver dead time).

With 'HNMR (Hydrogen Nuclear Magnetic Resonance) measurements in organic systems, water and fat will appear as dissolved components. This is probably because fat normally occurs in globules of fat. Proteins and carbohydrates however, occur as more solid components. Thus, magnetic polarization of water and fat will return to normal condition slower than magnetic polarization of protein and carbohydrates.

Another known technique for measuring single quantum transitions is solid echo, among others described by Paul T. Callaghan, "Principles of NMR Microscopy", Clarendon Press, Oxford, 1991, p.84-85, incorporated herein by reference. In solid echo two pulses are sent after one another as soon as the system allows. A NMR signal is then recorded. This technique provides a well defined "HNMR signal, where the height of the amplitude is directly proportional to the total amount of components in the sample (both solid and mobile). At the same time the shape of the curve of the observed NMR signal makes it possible to differenetiate between the response from more solid molecules such as proteins and carbohydrates, and the response from mobile molecules such as water and fat, and may thus be used to quantify both.

In 'HNMR measurements both hydrogen atoms in water will behave identically as long as the molecule is allowed to move freely (in molecular level). This means that water has only one quantum transition, and thus produce only signal at single quantum 'HNMR, which is the usual experiment. Several following excitations may induce transition to a higher quantum number. With molecules having hydrogen in different magnetic surroundings, such as CH₂-groups in for instance fat and proteins, it is possible to induce more quantum steps. The simplest technique is double quantum, or two changes of the spin quantum number in the same molecule system, which for instance is obtained with the "double quantum" technique, described among others by P. Granger and R. Harris, "Multinuclear Magnetic Resonance in Liquids and Solids - Chemical Applications", Kluwer Academic Publishers, 1990, p.103- 129, incorporated herein by reference. This technique precludes observations of single quantum transitions, and includes observations of all higher quantum steps. Single quantum signals or coherence have one speed after excitation, double quantum signals have two times this speed, and thus these signals will be separated in an organized way, which may be used to see one and not the other signal. Thus, it is possible to differentiate between the water-signal and the signal from the rest of the molecule in a number of compositions such as food articles (animal and vegetable), petroleum products, chemical products etc. The technique may also be used in vivo in humans and animals. However, not all molecules may give quantum coherence of a higher number. Molecules such as water, hydrogen gas, methane and ethane, wherein each hydrogen atom is chemically and magnetically identical, may only provide single or no quantum excitation. Molecules such as fat, oil, carbohydrates, proteins etc., wherein the molecules contain a number of different nucleus', may provide transition to a higher quantum number, and give observable multi-quantum 'HNMR signals, if the observation is correctly designed.

Since double quantum steps are more rare than single quantum steps, the signal/noise relationship in the technique is lower than at single quantum transitions, and the selectivity gives the technique great commercial potential. When the method is combined with dedicated recording equipment, this may be a powerful technique for industry and diagnosing.

Object The main object of the invention is to provide a method for fast and accurate quantifying of different components in organic and biologic material. It is a further object that the method not require specially trained staff, expensive equipment or a lot of space.

The invention The object is fulfilled with a method according to the characterizing part of patent claim

1. Further advantageous features are described in the dependent claims.

The invention will hereinafter be described with reference to the enclosed drawings, wherein - Figure 1 shows the 'HNMR solid echo pulse sequence, and the NMR-signal belonging thereato,

- Figure 2 shows 'HNMR pulse sequence and the thereto belonging response for double quantum excitation (double quantum technique) belonging thereto,

- Figure 3 shows a curve of the fat content in a number of milk samples, wherein 'HNMR double quantum measurements and conventional measurements are plotted against each other,

- Figure 4 shows a curve of the fat content in a number of samples, wherein 'HNMR spin echo measurements and conventional measurements are plotted against each other,

- Figure 5 shows a curve of the protein content in a number of samples, wherein 'HNMR solid echo measurements and conventional measurements are plotted against each other, and

- Figure 6 shows a curve of the water content in a number of samples, wherein the water content, calculated from a 'HNMR solid echo pulse sequence and a 'HNMR spin echo pulse sequence, are plotted against conventional measurements.

If a sample contains both solid protein and fat (which will occur as dissolved), the observed solid echo 'HNMR signal will represent a superposition, meaning a combination of the contributions. Both the "HNMR pulse sequence and such a continued response (both echo from "solid" and "dissolved") are shown schematically in Figure 1. The figure shows solid echo 'HNMR, comprising two (π/2) pulses at 0° and 90° phases, respectively, with a time τ between the pulses. Of course other pulses and phases may be used to obtain the same result as will be known to a person skilled in the art. The τ-time should be as short as possible, but not shorter than the dead time of the NMR instrument (τ = 28 s in an instrument used in the present invention). The dead time is the time one should wait after an RF pulse before a NMR signal may be detected. The detection of the signal is not possible at an earlier time due to fluctuations of the excitation pulse which continues to "live" in the dead time period, and this may ruin the observed signal. The dead time will vary with 'HNMR system.

The 'HNMR signal from the solid echo technique mentioned above, is a combination of the signal from solid (protein) and liquid (fat and water) components (see Figure 1). The signal will in most cases increase at first, then flatten off for a short period of time, and finally decline relatively quickly. After some time there is a change in speed wherein the signal declines. Thus there is a break 1 in the signal versus time curve. The amount of solid components is proportional with the height 2 of the signal from break to maximum value. Thhe amount of liquid components will be proportional with the height 3 of the signal at the break. In most samples the amount of liquid components will be the sum of fat and water, and the height 3 will thus be proportional to the sum of the amount of fat and water. The signal from solid components will often be the signal from proteins, but may also contain signals from other solid components, such as carbohydrates. To differentiate between proteins and other solid components may be necessary in order to effectively follow the production extra well in some production systems, and the results according to the present invention, have proven to be well suited to quantify proteins, for instance in fish feed.

Detection of 'HNMR signals from fat in a system with a lot of water, requires another technique. When a pulse sequence has suitable delays, a multi-quantum 'HNMR signal will only contain signals from mobile molecules with magnetically different nucleus'. For a number of products this will provide a very exact quantification of fat, even in the presence of water. Such a technique is double quantum, as mentioned above. This technique may be modified in two different ways. One modification uses cycling of phases of RF pulses (16 different combinations) to suppress the water signal from the sample. The other modification uses external magnetic gradients to obtain the same result. In cases where different 'HNMR signals are summed, the phases of the pulses must be changed in order to obtain the double quantum coherence without disturbance from other signals. If magnetic gradients are available, the number of 'HNMR signals may be reduced without affecting the result. A possible progress with magnetic gradients is shown in Figure 2. The technique that uses external magnetic gradients requires less of the equipments^' hardware, and thus it is more robust and simpler to implement into a low field NMR instrument.

The progress of the process, as shown in Figure 2, comprises two (π/2) RF pulses with an interval of d₀=l (4J) (J is the coupling constant in fat molecules, in Hz) with the following (π), (π/2) and (π) RF pulses, all having equal time interval d, one from the other, (all pulses at 0° phase). After time delay d,, a signal/echo is observed, wherein the height 4 of the amplitude is proportional to the fat/oil amount in the sample. Between the third and fourth pulse of the pulse sequence, a magnetic gradient is switched on and off. The same magnetic gradient is switched on and off between the last (fifth) RF pulse and the observed signal. A combination of the techniques, meaning spin echo 'HNMR, for quantifying the amount of fat and water (or just the amount of fat in samples containing little free water), double quantum 'HNMR for quantifying just fat, and solid echo 'HNMR for quantifying proteins and carbohydrates (and simultaneously the sum of water and fat), and the right choice of delays in the pulse sequences, makes it possible to quantify fat, water, protein and carbohydrates in a large number of products both before, during and after a process step in the production. A summary of this is shown in Table 1. Corresponding to other NMR systems, the height of the signal-amplitudes must be converted into amount of components by means of calibration curves.

As stated in Table 1 , both spin echo and double quantum techniques may be used to quantify fat in samples containing little water. However, the type of instruments able to perform double quantum technique are more complex in construction and more expensive. Thus it will be an advantage to only use solid echo instead of double quantum. It may be an advantage to use a combination of spin echo, solid echo and double quantum techniques to all samples, in order to secure the best result possible, but in practice, it will involve a lot of unnecessary cost. Further, both spin echo and solid echo may be used to quantify the sum of fat and water in samples containing a lot of water. Since the solid echo pulse sequence is necessary in order to quantify protein, the spin echo pulse sequence may be left out. However, in order to secure the best result possible, it is an advantage to perform it. For samples containing little water, the spin echo technique may Thus be used to quantify fat, while the solid echo may be used to quantify protein and the sum of fat and water. The amount of water may thus be calculated by subtracting the height of the amplitude of the signal in the spin echo pulse sequence, from height 3 (see Figure 1) of the signal from the solid echo pulse sequence (being proportional to the amount of liquid components, meaning water and fat), and then convert the difference into amount of water by means of a calibration curve. The amount of fat may be calculated from the height of the amplitude of the signal from the spin echo pulse sequence, and a fat calibration curve, while the amount of protein may be calculated from height 2 (see Figure 1) of the signal from the solid echo pulse sequence (being proportional with all solid components) and a protein calibration curve.

In samples containing a lot of water, the double quantum technique may be used to quantify fat, while the amount of protein and the sum of the amount of fat and water is quantified with the solid echo technique. In these samples, the amount of water may be calculated by subtracting amplitude height 4 (see Figure 2) of the signal from the double quantum pulse sequence, from height 3 (see Figure 1) of the signal from the solid echo pulse sequence (being proportional to the amount of liquid components), and convert the difference into amount of water by means of a calibration curve. The amount of fat may be calculated from amplitude height 4 (see Figure 2) of the signal from the double quantum pulse sequence by means of a fat calibration curve, and the amount of protein may be calculated as for samples containing little water.

A person skilled of in the art will in most cases know whether a sample contains a lot of water or not. If this is not certain, or there is doubt, the method for measuring samples containing a lot of water may be used, meaning a combination of double quantum and solid echo, because this method will provide the right answer no matter how much water the sample contains.

The method may be used on line, and has proven to be very robust and simple to use, when these applications are a part of the instruments. Upon operating these instruments one only has to press the start button, and a computer coupled to the system will keep order of the process parameters. The process may thus be optimized with regards to costs, efficiency and exactness/reproducibility.

The method according to the present invention is useful on any complex system, because signals from water in the sample may efficiently be suppressed with double quantum technique, the amount of fat may be quantified exactly. The unknown amount of fat in the sample is calculated by calibrating the 'HNMR signal relative to a set of reference samples, wherein the amount of fat is known. A number of such calibration procedures, performed at SINTEF Unimed MR senteret (Trondheim, Norway) has proven very good reproducibility of the results. Figure 3 shows a calibration curve based on 5 reference samples, 2 parallels of each. Adjustment of the experimental points provides a correlation factor of 0,9998, which proves an excellent conformity between the result from the method according to the present invention, and the known amount. By using the calibration curve, the amount of fat may be quantified in less than a minute. Separate calibration curves for materials of different sources, may be necessary.

The invention will in the following be described by means of a comparison analysis, wherein the results from known measurement methods are compared with the results from the method according to the present invention.

Comparison analysis The amount of fat, water and protein was quantified in 26 samples, by Felleskjøpet AS (Trondheim, Norway) and Norsk Matanalyse AS (Oslo, Norway) by means of known chemical methods: Bϋchi-Caviezel for quantifying fat and AOAC 981.10 (1983) for proteins. Buchi-Caviezel was performed on a BUCHI B-815/B-820, according to BUCHI Instructions, Fat determination B-820, Version C, 96571 (1996) and BUCHI Fat

Determination system B-815/B-820. Method, Version B, 95668 (1995). The quantification of proteins was performed on a Kjeltec 2400 (Kjeltab S tablets) instrument, with meat based certified reference material (SMRI 94-1, meat based). The results are shown in Table 2, and Figures 4, 5 and 6. The amount of water was determined with wet chemical standard method NMKL 23.

The 'HNMR signal for the amount of fat and water, and protein ('HNMR signal / sample weight) in these samples (with exactly known fat and protein content) was then measured according to the present invention, on a Bruker minispec mq20'HNMR instrument. The adjustments are enclosed as an appendix. The spin echo sequence, as implemented and tested, comprised a single π/2 RF pulse at

0° phase and a π pulse at 0° phase. The solid echo pulse sequence comprised two (π/2) RF pulses at 0° and 90° phase respectively. The interval between the pulses was longer than, or similar to, the dead time of the 'HNMR instrument. The double quantum pulse sequence as implemented and tested, comprised two (π/2) RF pulses with an interval of d₀=T/4J (J is the coupling constant in fat molecules, in Hz) with the following (π), (π/2) and (π) RF pulses, all having equal time interval d, from each other (all pulses at 0° phase). A magnetic gradient Gl was switched on and off, between the third and the fourth pulse of the pulse sequence. The same magnetic gradient was switched on and off between the last (fifth) RF pulse and the observed signal. After a time delay d,, a signal was observed as a measure of the amount of fat/oil in the sample. The spin echo, solid echo and double quantum sequences may comprise other pulses and methods according to known technique, which will be known to a person skilled in the art.

The results from both the chemical and the NMR measurements are shown in Table 2. and Figures 4, 5 and 6. The figures show measured [signal/weight] values for protein, fat and water, respectively, presented as a function of chemically determined amounts in the corresponding samples. All curves are adjusted to a linear function. The correlation factor for the fat curve was about 0,9986, for protein it was about 0,9677, and for water it was about 0,8209. This indicates conformity between the methods for measuring fat- protein and water content in the samples.

After such a calibration based on a set of samples with known protein, (fat + water) and fat content, the method according to the present

invention may be used to determine protein, fat and water content in unknown samples simultaneously, in less than a minute.

* NMR results for WATER (two last columns) are calculated by subtracting 'HNMR signal/weight for FAT from 'HNMR signal/weight for FAT + WATER

Above there is described use of three 'HNMR pulse sequences for quantifying fat, water and protein in heterogenous biological systems It will be obvious to a person skilled in the art to that with very simple modifications, the method may be used to quantify other components The invention comprises combinations and subcombinations of the described features, and modifications and variations of this being obvious to a person skilled in the art, and within the scope of the following claims

Claims

Patent claims

1. Method for quantifying different components in organic and biologic material, by means of NMR and calibration curves, wherein pre-treatment of the samples is unnecessary, characterized in that the method includes

- analysing in an NMR-instrument, wherein a spin echo pulse sequence and a solid echo pulse sequence are performed on samples containing little water, and a solid echo pulse sequence and a double quantum pulse sequence are performed on samples containing a lot of water, after which the resultant obtained NMR signals are recorded, and

- processing of the resultant signals in a computer.

2. Method according to claim 1 , characterized in that the method comprises for all samples a spin echo pulse sequence, a solid echo pulse sequence and a double quantum pulse sequence.

3. Method according to claim 1 or 2, characterized in that the spin echo pulse sequence comprises a single π/2 RF pulse at 0° phase, and a π pulse at 0° phase.

4. Method according to claim 1 or 2, characterized in that the solid echo pulse sequence comprises two (π/2) RF pulses at 0° and 90° phase, respectively, and that the time interval between the pulses is longer than or equal to, the dead time of the NMR instrument.

5. Method according to claim 1 or 2, characterized in that the double quantum pulse sequence uses cycling of phases on RF pulses with 16 different combinations.

6. Method according to claim 1 or 2, characterized in that the double quantum pulse sequence uses external magnetic gradients.

7. Method according to any one of the preceding claims, characterized in that the processing in the computer includes - for samples containing little water

- subtraction of the amplitude height of the signal from the spin echo pulse sequence from the height (3) of the signal from liquid components in the solid echo pulse sequence, and conversion of the difference into water amount by means of a water calibration curve,

- conversion of the amplitude height of the signal from the spin echo pulse sequence, into amount of fat, by means of a fat calibration curve, and

- conversion of the height (2) of the signal from the solid components in the solid echo pulse sequence, into amount of protein, by means of a protein calibration curve, - for samples containing a lot of water

- subtraction of the amplitude height (4) of the signal from the double quantum pulse sequence, from the height (3) of the signal from liquid components in the solid echo pulse sequence, and conversion of the difference into amount of water by means of a water calibration curve, - conversion of the amplitude height (4) of the signal from the double quantum pulse sequence, into amount of fat by means of a fat calibration curve, and

- conversion of the height (2) of the signal from solid components in the solid echo pulse sequence, into amount of protein by means of a protein calibration curve.