Dinitrophenyl-human serum albumin conjugate

Example 1

Two solutions were prepared.

The first comprised 10 μg of Fab prepared as described in the example 1a in 2.5 μl of 100 mM phosphate buffer at pH 6.50.

The second comprised 8⁻¹⁰ moles of succinimidyl ester of the fluorophore CF568 in 200 μL of anhydrous DMF and the concentration of fluorophore of the solution was checked by means of spectrophotometry in the visible (Principles of Fluorescence Spectroscopy Third Edition, Joseph R. Lakowicz). The two solutions were slowly mixed and incubated at 37° C. in the dark for an hour.

To the resulting mixture were added 80 μl of an aqueous solution at pH 8.6 containing 2M ethanolamine and 0.1% by volume of Tween 20, which was first kept in the dark at 37° C. for 15 minutes and then at 4° C. throughout the night. The separation of the fluorescent bio-conjugate from the excess of free fluorophore was carried out by means of molecular-exclusion chromatography (Sephadex G25 Fine, GE Healthcare), and the identification of the eluted fractions containing the fluorescent conjugate was carried out by means of thin layer chromatography using laser scanning, the absence of the free probe was verified by means of laser scanning. In detail, the product purified by chromatography was eluted in fractions of 50 microliters. 450 nanoliters of each fraction were loaded onto reverse phase TLC (C18) and the plate developed in H₂O/CH₃CN eluent (1:1 vol/vol) with the addition of 0.01% by volume of an ammonia-saturated aqueous solution. The plate was dried in the dark and at room temperature and then scanned with an Amersham Typhoon Imaging Systems reader.

The acquisition conditions were as follows: laser line at 561 nm or 633 nm; acquisition bands: 580 nm with 30 nm bandwidth, 680 nm with 30 nm bandwidth. The scan proceeded over an area of 5 cm×5 cm, with a pixel size of 100 micron² and the filtered emission was detected by means of a phototube set at 450 volts. The fractions in which spots at RF=0 were detected were collected and processed for the next step. The fractions with spots at RF=0 and RF=0.8 or only at RF=0.8 were discarded.

The Fab's degree of marking, quantized by visible UV spectrophotometry, was found to be between 1.8 and 2.2. The algebraic expression used for the calculation of the degree of marking (DOL) is as follows:

$\mathrm{DOL}=\frac{A_{\max }{\varepsilon }_{\mathrm{prot}}}{\left(A_{280}-A_{\max }{C}_{280}\right){\varepsilon }_{\max }}$

where the values A_max and A₂₈₀ indicate the maximum intensity recorded in the fluorophore absorption region and the maximum intensity recorded in the protein region respectively: such values are directly obtained from the UV-VIS spectrum for each conjugate. Diversely, the value ε_prot is estimated at around 71000 M⁻¹cm⁻¹, while C₂₈₀ and ε_max are provided by the producer of the fluorescent molecule.

The conjugate was stabilized by the addition of 20 μL of an aqueous solution of 0.1% albumin by weight and 5 mM sodium azide, and brought to a storage and use concentration of 0.1 mg/mL using a vacuum centrifugal evaporator.

where the values A_max and A₂₈₀ indicate the maximum intensity recorded in the fluorophore absorption region and the maximum intensity recorded in the protein region respectively: such values are directly obtained from the UV-VIS spectrum for each conjugate. Diversely, the value ε_prot is estimated at around 71000 M⁻¹cm⁻¹, while C₂₈₀ and ε_max are provided by the producer of the fluorescent molecule.

The conjugate was stabilized by the addition of 20 μL of an aqueous solution of 0.1% albumin by weight and 5 mM sodium azide, and brought to a storage and use concentration of 0.1 mg/mL using a vacuum centrifugal evaporator.

To mark the Fab obtained in the example 1a by means of fluorophore AF647, the procedure was that described in the example 3a.

To mark the Fab obtained in the example 1b by means of fluorophore CF568, the procedure was that described in the example 3a.

To mark the Fab obtained in the example 1b by means of fluorophore AF647, the procedure was that described in the example 3a.

To mark the Fab obtained in the example 1c by means of fluorophore CF647, the procedure was that described in the example 3a.

To mark the Fab obtained in the example 1b by means of fluorophore AF 532, the procedure was that described in the example 3a.

To mark the marked Fab obtained in the example 1b by means of fluorophore Aberrior star 635P, the procedure was that described in the example 3a.

Example 3

To mark the marked Fab obtained in the example 3c by means of fluorophore AlexaFluo488, the procedure was that described in the example 3a.

To mark the marked Fab obtained in the example 3d by means of fluorophore AlexaFluo488, the procedure was that described in the example 3a.

Example 4

Confocal microscopy experiments were conducted using the fluorescent Fabs obtained in the example 3a (INV), 3b (INV), 3c (INV) and 3e (CONTROL) respectively.

In particular, HeLa were fixed in formaldehyde, neutralized and permeabilized with a solution of 50 mM ammonium chloride, saponin of 0.05% by weight and incubated with a bovine albumin solution of 0.1% by weight. The samples were then left to hybridize with 15 μL of a phosphate buffer containing 0.05% saponin and 0.1% albumin by weight and 150 ng of the aforementioned fluorescent Fabs. The excess marker was washed with 1 mL of saline phosphate buffer (PBS) at pH 6.80-7.20 and mounted on 20 μL of Mowiol®.

The microscope used is an inverted confocal Leica SP5-II (Leica Microsystems, Milan, Italy). Cellular samples were included in Mowiol polymer on fluorescence slides, 0.7 mm thick. The images were taken from a 100× oil-immersed lens with a numerical aperture of 1.40 (Leica Microsystems). The excitation of the fluorophores was obtained by means of a laser line at 561 nm and 647 nm and by means of a white pulsed emission laser source (SuperK, Leica) respectively. The fluorescent emissions of each fluorophore were filtered by means of AOBS in a range from 560 nm to 650 nm.

In the images obtained with Fab from rabbit polyclonal directed against the BARS protein marked with CF568 fluorophore, prepared as described in the example 3a (INV), it is possible to observe nuclear and Golgi organelle staining, and drastically lower cytosolic staining (see FIG. 1A and FIG. 1Abis, which refer to images obtained with Fab marked by reaction with 1-pyrrolidinil, 2.5-dione and 1-pyrrolidinil succinimide ester, 3 sulfonyl 2.5-dione, respectively). The location of the fluorescent fragment is in perfect agreement with the data obtained by indirect immunofluorescence.

In the images obtained with Fab from rabbit polyclonal directed against the BARS protein marked with fluorophore AF647, prepared as described in example 3b (INV), it is possible to observe nuclear and Golgi organelle staining, and a drastically lower cytosolic staining (see FIG. 1B and FIG. 1Bbis, which refer to images obtained with Fab marked by reaction with 1-pyrrolidinil, 2.5-dione and 1-pyrrolidinil succinimide ester, 3 sulfonyl 2.5-dione, respectively). Furthermore, the intensity of the signals is lower than that obtained in the example 3a (see FIG. 1A and FIG. 1Abis), congruently to the lower quantum yield of the fluorophore used. The localization of the fluorescent fragment is in this case also in perfect agreement with the data obtained by indirect immunofluorescence.

In the images obtained with Fab from mouse monoclonal directed against the alpha tubulin protein marked with CF568 fluorophore, prepared as described in example 3c (INV), a cytoskeleton staining can be observed, specifically a staining of the microtubules (see FIG. 2A and FIG. 2Abis, which refer to images obtained with Fab marked by reaction with succinimide esters 1-pyrrolidinil, 2.5-dione and 1-pyrrolidinil, 3 sulfonyl 2.5-dione, respectively). The location of the fluorescent fragment is in perfect agreement with the data obtained by indirect immunofluorescence.

In the images obtained with Fab from rabbit polyclonal directed against the protein AKAP9 marked with CF647 fluorophore, prepared as described in the example 3e (COMPARISON), it is possible to observe an incorrect localization of the fluorescent conjugate, as nuclear staining is observed in addition to cytosolic and/or mitochondrial staining (see FIG. 2B and FIG. 2Bbis, which refer to images obtained with Fab marked by reaction with the succinimide ester 1-pyrrolidinil, 2.5-dione and 1-pyrrolidinil, 3 sulfonyl 2.5-dione, respectively). The localization of the fluorescent fragment is in total disagreement with the data obtained from literature and by indirect immunofluorescence, which instead report a localization of the target protein only on the Golgi organelle.

As can be seen from the results of the above examples, Fabs conjugated through a linker with specific length and flexibility characteristics to CF568 and AF647 fluorophores, having specific characteristics such as: KierFlex between 7.5 and 15; A log P below or equal to 0; 3DpolarSASA equal or higher than 300, are stable, have uniform characteristics in terms of positioning of fluorophore and maintain avidity for the antigen. In particular, the 3D polar SASA parameter, which provides for the degree of hydration of the molecule, is responsible for greater stabilization of the conjugate, less tendency to aggregate because it is hydrated and better emissive behavior due to increased solvation. Furthermore, the decreased tendency to interact with parts of the fragment also translates into a significant increase in avidity because the fluorophores do not collapse on the recognition regions of the Fab.

Example 5

Conventional marking was performed with mouse monoclonal antibody fragment, type IgG1, in HeLa human cells fixed, permeabilized and subjected to a common immunofluorescence procedure.

HeLa human cells were fixed, permeabilized and subjected to an immunofluorescence procedure as described below. The sample preparation procedure is completely similar to indirect immunofluorescence. The cells grew on a 1 cm×1 cm quartz slide, in RPMI medium (Dulbecco) supplemented with 10% in volume of bovine serum, penicillin, glutamine. The cells are fixed by treatment with 4% formaldehyde in buffered saline solution at pH 7.00 for 10 minutes. The excess aldehyde was neutralized with a 50 mM ammonium chloride solution, buffered at pH 7.00. The fixed samples undergo washing in saline phosphate buffer and permeabilized with 0.05% by weight of vegetable saponin solution. The samples were incubated for one hour with Fab directed against the alpha-tubuline protein, conjugated with fluorophore CF568 prepared as described in example 3c (samples 1-3) or fluorophore AF647 prepared as described in example 3d (samples 4-5) and the excess present was removed by 3 washes in saline phosphate buffer. Any non-specific adsorption sites of markers (primary and secondary) in the cell samples were prevented by incubation with 0.1% by weight bovine albumin solution in saline physiological solution.

To each sample was added a commercial probe consisting of the AlexaFluor488 marker conjugated to rabbit polyclonal antibody directed against mouse monoclonal antibody of the IgG1 type.

The signal emitted by the fragment, to which the present invention relates, was detected by means of the acquisition of emission of fluorescence in the wavelength range specific for the type of fluorophore: in the case of CF568, the acquisition range is within the range from 575 nm to 620 nm; while in the case of AF647, the acquisition range is within the range from 660 nm to 700 nm. The control fluorophore, AlexaFluor bound to the secondary antibody directed against the Fab, can be excited separately from the other two, so its emission, collected in the range from 490 nm to 450 nm, if collected only for the time in which it is individually excited, can be considered independent and therefore free of light contamination between the channels. It is therefore possible to make a quantitative and functional comparison between a conventional marking (indirect immunofluorescence) and that produced by fluorescent Fabs, to which the present invention relates.

3 samples were prepared marked with the CF568 probe and with the AlexaFluor488 commercial probe and 2 samples marked with the AF647 probe and with the AlexaFluor488 commercial probe.

Comparing the FIG. 3A with FIGS. 3B, 4A with 4B, 5A with 5B, 6A with 6B and 7A with 7B respectively, we can see not only the complete functionality of the proposed marker, consistently with the calculated data, both for the CF568 probe and for the AF647 probe, but from this analysis also emerges an aspect of considerable importance as regards the functional superiority of the invention.

As can be seen from FIGS. 3B, 4B, 5B, 6B and 7B representing the images in which the signal, emitted by the Fab to which the present invention relates, was detected by means of fluorescence emission acquisition in the wavelength range specific to the CF568 or AF647 fluorophore, the reduced dimensions of the probe to which the present invention relates, about one sixth compared to those of the conventional probe, are not only able to generate a high staining density, but also allow the localization of the probe in cell regions to which the conventional marker does not have access. The area of interest lies in the midbodies.

On the contrary, as can be seen from FIGS. 3A, 4A, 5A, 6A and 7A representing the images in which the signal emitted by the Fab to which the present invention relates was detected by means of fluorescence emission acquisition in the wavelength range specific to the commercial fluorophore AF488, the conventional probe does not have access to the area where the midbodies are located.

These regions are connections between cells which are reduced to a thin filament in the final stages of the cytokinesis process. Responsible for this bottleneck are the actin filaments forming a contractile filament which force the cytoskeleton and, by creating a bottleneck, cause the two daughter cells to separate. In the examples shown below, it can be seen that only the probe to which the patent relates is able to penetrate and indicate, in detail, the morphological state of the above midbodies.

Cell division is currently a process of considerable interest because it is involved in metastatic and neoplastic mechanisms, so the use of the Fab according to the invention to visualize cell compartments which are not visible by traditional marking can be an ideal example of the functional superiority of such reagents in the field of cell imaging.

Therefore such fluorescent Fabs are useful in microscopy techniques and permit obtaining a high level of resolution and visualizing cell compartments not visible with traditional marking.

In this regard, the FIGS. 8A, 8B, 8C, 8D represent a comparison between the images obtained with probes currently on the market and probes in accordance with the present invention.

The samples were incubated for one hour with Fab directed against the α-tubuline protein, conjugated with the fluorophore AF 647 prepared as described in the example 3d.

At the same time, to other samples was added a probe consisting of a commercial anti-α-tubuline antibody conjugated randomly with the fluorophore AF 647.

The signal emitted by the invented fragment (FIGS. 8A, 8B) was compared with the signal emitted by conventionally marked antibodies (FIGS. 8C, 8D) in order to quantify the uncertainty limit by which the α-tubuline is localized and, consequently, to measure the diameter of the microtubule.

As shown in the FIGS. 8C, 8D, the size of the microtubular section evidenced by the use of conventional antibody is larger than the actual size. This increase is associated with the greater uncertainty of localization caused by the diffusion of the emitted light signal and corresponding to an indefinite region of space within which the uncertainty on the real position of the microtubule is distributed.

On the contrary, as can be seen in FIGS. 8A, 8B which represent the images in which the signal emitted by the Fab to which the invention relates was detected, the smaller microtubular diameter is indicative of a smaller dispersion of the fluorescence emission and, therefore, of a greater resolution limit.

In detail, the distribution of the fluorescent signal along the intersection of the microtubule represents the indetermination of the marking which is significantly lower when the Fab to which the present invention relates is used.

Furthermore, the fact that the Fab to which the present invention relates is positioned at about 1-2 nm from α-tubuline results in a lower dispersion of the fluorescent signal.

From the FIG. 9, the functional superiority can be seen of the present invention in line with the above data. In fact, the mean diameter value for each segment measured is about 2 nm away from the actual value with Fab in accordance with the present invention. On the contrary, localizations obtained with conventional probes (in the figure indicated with Ab) oscillate by about 8 nm around the mean value.

In detail, the mean value recorded through the use of Fab in accordance with the present invention is equal to 28 nm, while the mean value recorded through the use of probes of the conventional type is 42 nm.

Considering the actual diameter of a microtubule equal to 25 nm, it appears evident that the resolution limit has been significantly increased by using probes in accordance with the present invention.

At the same time, as can be seen from the FIG. 10, screening has been developed using 10 fluorophores conventionally used in super-resolution microscopy and having a different structure.

In detail, the loss of affinity of the conjugated Fab towards its antigen can be directly verified by confocal microscopy, evaluating known cellular morphologies.

This analysis showed that four of the ten structures tested prevented the conjugate from binding the antigen; this corresponds to the loss of the expected morphology (in this case the cytoskeleton), as the presence of the fluorophore totally prevents its binding to the antigen.

At the same time, in order to evaluate the influence exerted by the pH reaction values on the possible overlapping of the Fabs behavior with regard to the derivatization of the two categories of amino groups previously described, a study was carried out on the in silico prediction of the relative pKa values (FIG. 10).

As is known, the mean values of the pKs of the terminal amino and ε-amino groups, are around 7.5 and 13 respectively.

Such difference would formally allow discrimination between the various amino groups by simply working on the operating pH. It can be approximately calculated that 3 units of pK of difference (e.g., pK 7.5 against pK 10.5) between a lysine terminal amino group and an ε-amino group, results in a high percentage difference in the state of protonation between the two compared amino groups.

For example, at a pH=7.4 which is close to the pK of the terminal NH₂ group, 44.3% of these are deprotonated and available, while only 0.08% of the ε-amino groups are available.

In any case, the simultaneous presence of amino groups (terminal against ε-amine) which compete with regard to a fluorophore functionalized with a succinimide group, requires a realistic evaluation of the percentages of deprotonated amino groups available for conjugation at each pH value around which the request for adequate specificity is satisfied.

The prediction of pKa values of amino groups was performed with a hybrid quantum mechanics/molecular mechanics method (ab initio QM/MM), using two applications, Qsite (Murphy, R. B.; Philipp, D. M.; Friesner, R. A., “A mixed quantum mechanics/molecular mechanics (QM/MM) method for large-scale modeling of chemistry in protein environments,” J. Comp. Chem., 2000, 21, 1442-1457; Philipp, D. M.; Friesner, R. A., “Mixed ab initio QM/MM modeling using frozen orbitals and tests with alanine dipeptide and tetrapeptide,” J. Comp. Chem., 1999, 20, 1468-1494) and Protein Titration Curve, contained in the program suites Small-Molecule Drug Discovery Suite (Small-Molecule Drug Discovery Suite 2018-2, Schrödinger, LLC, New York, NY, 2018) and Biologics Suite (Biologics Suite 2018-2, Schrödinger, LLC, New York, NY, 2018), respectively.

In terms of prediction of pKa values, a first analysis was centered in particular on two antibodies widely used in experimental practice and for which the relative 3-D structure is available:

• • i) the monoclonal FAB IgG1 isotype k mouse antibody fragment (anti-c-myc clone 9E10, PDB ID: 2OR9—Table 1) and
  • ii) the monoclonal FAB M204 rabbit antibody fragment (Table 2).

In both cases, the present inventors observed some lysine residues characterized by pKa values which were not between 10 and 13, as expected, but which had lower values below 9, and were therefore not far from those of the terminal amino groups the mean value of which is around 7.5.

A second important observation comes from the analysis of mouse antibody, whose light chain terminal amino group shows, on the contrary, a pKa significantly higher than expected (7.5) at the value of 9.96. These observations have highlighted conditions the occurrence of which is of clear importance to us. The analysis was then extended to predict the pKa values of the amino groups for further 18 antibodies (a total of 20, i.e. approximately 20% of the mouse and rabbit antibodies with known amino acid sequences), the 3D structure of which was determined. These antibodies were selected according to the criterion of maximizing the diversity of the amino-acid sequence, so as to exclude a possible aminoacidic-structural bias (Table 3).

This analysis shows that in most of the FABs considered, the respective pKas of the two chains (H, L) show mean values of around 7.5. However, in a smaller number of cases, as for the above-mentioned anti-c-myc clone 9E10 antibody, one of the two NH2 terminal groups (that of the light chain, L), shows significantly higher pKa values.

When this occurs, it is obvious that the terminal amino group actually available for selective derivatization, e.g. at pH≤7.4, is only 1 (that of the heavy chain).

As for the selectivity of the reaction, this is related to the particular operating pH.

The following example illustrates the case of an FAB characterized by a terminal amino group with pKa=7.8 and an ε-amino group with pKa=8.89, potentially in competition. FIG. 10 shows the titration curves from which the percentage fractions can be evaluated of the free amino groups (in the illustration, NH_2-ter) and of the ε-amino groups (in the illustration indicated as NH_2-lysine) at every pH.

At the same time, a curve (NH_2-ter/NH_2-lysine) is shown which indicates the variations in the NH_2-ter/NH_2-lysine ratio and from which it is possible to select an optimum pH for its maximization.

In accordance with the object of the present invention, it is clear how it is necessary to identify reaction conditions such as to ensure the selectivity of the derivatization reaction of the Fab fragment with an activated fluorophore.

In this regard, the conjugation reaction comprises numerous steps of which only the final one is irreversible and, on the contrary, the others are reversible and governed by equilibrium conditions and by the availability of the reactive substrate, the latter depending on the pH values.

From the study of the reaction of aminolysis of ester, it is clear that the latter proceeds by means of general basic catalysis, making it evident that dynamic reversibility exists only for the first reaction steps (consisting in the deprotonation of the amino group and in the formation of an intermediate tetravalent). The latter, however, irreversibly evolves in the final product by expulsion of the hydroxysuccinimide group followed by a substantially simultaneous formation of the amide link with the fluorophore.

Ultimately, the operating pH can be chosen so as to ensure selectivity with respect to the terminal amino groups without affecting the total yield of the reaction.