Spectroscopic insights into multi-phase protein crystallization in complex lysate using Raman spectroscopy and a particle-free bypass

Abstract

Protein crystallization as opposed to well-established chromatography processes has the benefits to reduce production costs while reaching a comparable high purity. However, monitoring crystallization processes remains a challenge as the produced crystals may interfere with analytical measurements. Especially for capturing proteins from complex feedstock containing various impurities, establishing reliable process analytical technology (PAT) to monitor protein crystallization processes can be complicated. In heterogeneous mixtures, important product characteristics can be found by multivariate analysis and chemometrics, thus contributing to the development of a thorough process understanding. In this project, an analytical set-up is established combining offline analytics, on-line ultraviolet visible light (UV/Vis) spectroscopy, and in-line Raman spectroscopy to monitor a stirred-batch crystallization process with multiple phases and species being present. As an example process, the enzyme Lactobacillus kefir alcohol dehydrogenase (LkADH) was crystallized from clarified Escherichia coli (E. coli) lysate on a 300 mL scale in five distinct experiments, with the experimental conditions changing in terms of the initial lysate solution preparation method and precipitant concentration. Since UV/Vis spectroscopy is sensitive to particles, a cross-flow filtration (cross-flow filtration)-based bypass enabled the on-line analysis of the liquid phase providing information on the lysate composition regarding the nucleic acid to protein ratio. A principal component analysis (PCA) of in situ Raman spectra supported the identification of spectra and wavenumber ranges associated with productspecific information and revealed that the experiments followed a comparable, spectral trend when crystals were present. Based on preprocessed Raman spectra, a partial least squares (PLS) regression model was optimized to monitor the target molecule concentration in real-time. The off-line sample analysis provided information on the crystal number and crystal geometry by automated image analysis as well as the concentration of LkADH and host cell proteins (HCPs) In spite of a complex lysate suspension containing scattering crystals and various impurities, it was possible to monitor the target molecule concentration in a heterogeneous, multi-phase process using spectroscopic methods. With the presented analytical set-up of off-line, particle-sensitive on-line, and in-line analyzers, a crystallization capture process can be characterized better in terms of the geometry, yield, and purity of the crystals.

Keywords: E. coli lysate; Raman spectroscopy; capture process; partial least squares regression (PLS); principal component analyses (PCA); process analytical technology (PAT); protein crystallization; ultraviolet-visible light (UV/Vis) spectroscopy.

FIGURE 1

Experimental and analytical set-up of the protein crystallization experiments as a scheme. The desired product characteristics are listed on the right together with the respective analytical measurement method and output. The vessel contains the clarified lysate with the HCPs, nucleic acids, the molecule of target LkADH, and later during the process LkADH crystals while the installed Raman probe records in-line spectra. The CFF-based set-up facilitates the solid-liquid separation, thus enabling on-line VP UV/Vis measurement in the permeate stream. Both the retentate and permeate streams are directed back to the vessel. All off-line samples are analyzed with IMAC and microscopic imaging. Selected samples are further analyzed with automated ELISA to determine the HCP content.

FIGURE 2

Counted crystals in microscopic images of off-line samples. The microscopic images of off-line samples are analyzed with a ML-based image analysis tool (Bischoff et al., 2022) counting the crystals and providing information on the crystal geometry (see Supplementary Material 2.3). The mean crystal count per imaged well and its standard deviation of undiluted and diluted off-line samples are visualized over the experimental time with dark green squares and light green circles with dotted lines to guide the eye and with their respective error bars. The off-line samples with micro-crystals present are shaded in gray as they are difficult to detect due to the image resolution (A, D). The off-line samples showing larger crystals are shaded in a light green box (B, C, E). Exemplary, the results of the automated image detection are shown for an undiluted (F) and a diluted (G) off-line sample of Exp2 after 19.7 h. The crystallization conditions for the experiments Exp1 - Exp5 (A–E) are listed in Table 1.

FIGURE 3

UV/Vis analysis in the analytical bypass. The recorded VP UV/Vis slope $\frac{A_{280\text{nm}}}{d_{\text{path}}}$ (A) and the $\frac{A_{260\text{nm}}}{A_{280\text{nm}}}$ (B) ratio of Exp1 - Exp4 are shown over time by blue circles and turquoise crosses, respectively. For comparison and clearer visualization, each row emphasizes one experiment and shows the other three experiments in light gray.

FIGURE 4

Preprocessing of Raman spectra. The raw Raman spectra, used for regression modeling, are shown over the recorded wavenumber range from yellow to red to visualize the different samples of all experiments (A). The black line represents one specific spectrum to better visualize the preprocessing effects. Preprocessing techniques, such as the baseline correction (C) and the SG filter, are applied to the spectra with a KS (B) or manual data split (D) to enhance spectral differences. The gray boxes in (C, D) depict the Raman shift ranges that are used for the PLS model development.

FIGURE 5

PCA scores of Raman spectra. The scores of the first & second PCs are identified as PC1 and PC2 and are shown for the five experiments (A–E, see Table 1). The observations of each experiment are classified by investigation of the off-line microscopic images. Observations before the initial centrifugation step, after the centrifugation step until the first, visual occurrence of crystals and after the first detected crystals are shown in blue, orange, and yellow, respectively. Linear fits of all data in (A, C, D) and the main trend in (B, E) are visualized by black arrows. The arrow direction shows the observations versus the process time. The coefficients of determination R ² of the linear fits are included. The gray diamonds symbolize the center of the data in the experiments (A, C, D).

FIGURE 6

Chemometric regression model based on Raman spectra and effects of validation sampling techniques. The preprocessed Raman spectra are regressed on the IMAC derived LkADH concentration with PLS models. Two models differing in the choice of the validation data set are compared. The white circles, gray squares, and dashed line represent the calibration, validation data, and theoretical values, respectively. First, the measured concentrations are visualized versus the model-predicted concentrations in (A) for a model with KS data split. Analogous plots are shown in (B) for a model where Exp5 was chosen manually as the validation data set. High R ² and Q ², and low RMSECV and RMSEP values indicate an applicable model.

FIGURE 7

PLS model application to crystallization processes out of clarified lysate. The PLS model calculated with the manual data split predicts the LkADH concentration on the basis of the in-line Raman spectra in orange for the five conducted experiments (A–E). Off-line LkADH calibration and validation concentrations are calculated from the IMAC analysis and are depicted by green circles and squares, respectively. Analogous to Figure 2, the light green boxes (B, C, E) indicate the time range when crystals are expected in the crystallization vessel as crystals are detected in the microscopic images of the off-line samples. The off-line samples in the time range illustrated by the gray boxes (A, D) showed only micro-crystals which were difficult to distinguish from precipitate visually.