Innovative Protein Structural Analysis with Microfluidic Modulation Spectroscopy

Matrix metalloproteinases (MMPs) are a key area of study. Exploring them increases awareness of protein biology and spurs interest in creating specific MMP inhibitors for medicinal use.

Details concerning their structure are crucial for supporting the aims of this research. This study used microfluidic modulation spectroscopy (MMS) to assess the structure of MMP-2, MMP-8, and MMP-9. It used first- and second-generation RedShift BioAnalytics tools and compared data quality.

The research also assessed the structural similarities and differences among the proteins. Although two of the three MMPs shared similar functions, all three structures were different.

Image Credit: RedShiftBio

Introduction

MMPs are proteases responsible for breaking down extracellular matrix (ECM) proteins. Various conditions, including cancer, multiple sclerosis, and strokes, are associated with the overexpression of specific MMPs.¹

Elevated MMP levels have been identified in patients with COVID-19 due to their critical role in lung physiology.^2,3 There is thus a significant research interest in using MMPs as biomarkers for estimating COVID-19 severity and developing inhibitors for targeted therapeutic approaches in severe cases of the condition.⁴

Awareness of the structural variations among different MMPs is critical for meeting this research aim. Different MMPs play diverse roles in ECM tissue remodeling—MMP-8, also known as collagenase 2, digests collagen, while MMP-2 and MMP-9, typically known as gelatinases A and B, target gelatin.

All these MMPs demonstrate increased expression levels in COVID-19 patients. ^2,3 Therefore, assessing their structural similarities and differences could be beneficial.

MMP-2, 8, and 9 have undergone significant research and structural characterization, particularly in their catalytic forms. The crystalline structures of the catalytic domains of such proteins are displayed in Figure 1.^6-8 It is demonstrated that all three proteins’ structures are very similar.

The catalytic domains have a secondary structure comprising three α-helices and five strands of β-sheets. Figure 2 demonstrates the full-length proteins, or proenzymes, of MMP-2 and 9.

The pro-domains and the catalytic domains demonstrate significant structural similarities. However, the fibronectin domains bear significant differences and comprise random coils (unordered structures). The crystalline structure of the full-length MMP-8 has yet to be resolved.

This study utilized MMS to define and compare MMP proenzyme structure.

Figure 1. X-ray crystal structures of the catalytic domains of MMP-2 (PDB: 1QIB), MMP-8 (PDB: 2OY4), and MMP-9 (PDB: 1GKC). Image Credit: RedShiftBio

Figure 2. X-ray crystal structures of the full-length proenzymes of MMP-2 (PDB: 1CK7) and MMP-9 (PDB: 1L6J). The C-terminal hemopexin-like domain is not shown. Image Credit: RedShiftBio

MMS assesses the amide I band of the infrared (IR) spectrum to determine protein structure. MMS continuously modulates against the reference buffer for precise and instantaneous background subtraction. The approach’s sensitivity, which is particularly critical for quality control, is compatible with diverse formulation buffers.

For this study, the first-generation MMS system, Apollo, and the second-generation tool, Aurora, were utilized, with the latter needing considerably less volume. Both instruments have a high-power Quantum Cascade Laser that is significantly more intense than traditional Fourier transform IR (FTIR) light sources.

This increased light intensity, alongside modulating background subtraction, renders MMS around 30-fold more sensitive than FTIR and five-fold more sensitive than circular dichroism for distinguishing discrete differences in protein structure.⁵

Methods

SinoBiological (Wayne, PA) supplied 1 mg each of MMP-2, MMP-8, and MMP-9. As for buffer matching, MMP-2 and MMP-9 faced dialysis against their different formulation buffers of 50 mM tris, 150 mM NaCl, 10 mM CaCl₂, and 0.05 % Brij35 at pH 7.5. MMP-8 was dialyzed against phosphate buffer saline.

Researchers diluted samples to 1 mg/mL and a volume of 1 mL. The dialyzed samples then underwent triplicate analysis utilizing a first-generation Apollo MMS system. Identical samples were additionally loaded on the second-generation Aurora system for comparison.

While the Aurora tool required just 50 µL of sample, Apollo required around 700 µL.

The two systems applied a backing pressure of 5 psi to move the samples into the flow cell. Inside the flow cell, samples were adjusted to 1 Hz between the sample and reference buffer, utilizing the same buffer used in dialysis for background subtraction.

The differential absorbance was calculated in the range of 1588–1711 cm^-1. Replicates were averaged, and the samples were normalized to obtain the absolute absorbance spectra.

Data processing in this research followed procedures in earlier studies. The raw differential absorbance data was converted into absolute absorbance, normalized by pathlength and concentration. The second derivatives of the absolute absorbance spectra were calculated following this to magnify spectral properties.

The resultant plot underwent inversion and baselining, making a “similarity plot” that qualifies the overlap area when compared against a control. This produced a similarity measure between samples.

Eleven Gaussian curves were fitted using Gaussian curve fitting, and the higher-order structure (HOS) was quantified relying on the detection of varying secondary structural elements across the amide I band (displayed in Table 1).

Table 1. Gaussian curve fit settings and HOS structural element designations. Source: RedShiftBio

Results and Discussion

The MMS results among the three MMP subtypes, MMP-2, MMP-8, and MMP-9 (as shown in Figure 3), underscored key structural differences.

While all three proteins demonstrate a combination of α-helix (1653–1658 cm^-1) and β-sheet (1636–1640 cm^-1) structures, their peak positions and intensities vary greatly, as depicted in Figure 3A.

The three spectra additionally exhibit less intense peaks at 1683–1687 cm^-1 and 1618–1624 cm^-1, suggesting β-turn and intermolecular β-sheet structures, respectively.

Regardless of the functional similarity that may indicate more structural resemblance between MMP-2 and MMP-9, the MMS results suggest otherwise. MMP-2 and MMP-9 show the largest shifts, around 5 v, for the α-helix and β-sheet peaks together, while MMP-8 lies between the other two.

Regarding intermolecular β-sheet peaks, MMP-9 demonstrates a considerably more intense peak at 1624 cm^-1, 6 v larger than MMP-2 and MMP-8, located around 1618 cm^-1.

The HOS bar chart in Figure 3B presents the quantified relative abundance of these secondary structural factors in each protein, calculated via Gaussian curve fitting utilizing the delta software.

Among all three proteins, β-sheet was the most abundant secondary structure, with β-turn and α-helix trailing after.

MMP-8 has the largest proportion of β-sheet structures among the three proteins, in accordance with the similarity spectra. MMP-2 demonstrates the most unordered structures, although MMP-9 has the most intermolecular β-sheet structures (also called “β-”).

Despite their similar biological functions, MMP-2 and MMP-9 do not share as much structural commonality as MMP-8.

Figure 3. Apollo data showcasing the secondary structural comparison of MMP-2, MMP-8, and MMP-9. (A) Similarity plots (baselined second derivative spectra) illustrating the peaks within the amide-I band for each MMP. (B) HOS plot displaying the relative abundance of each secondary structural element for the MMPs. The error bars represent +/- the standard deviation (N=3). Image Credit: RedShiftBio

For easier comparison, all three MMP samples were assessed using RedShift BioAnalytics’ second-generation tool, Aurora. As displayed in Figure 4A, the unique peak locations and intensities are well-matched with Apollo data.

The Aurora spectra may seem less smooth, especially around 1680 cm^-1. This discrepancy stems from a heightened resolution, shifting from 4 cm^-1 (Apollo) to 1 cm^-1 (Aurora), which decreased interpolation in the Aurora spectra.

The HOS bar chart in Figure 4B further validates the breakdown of HOS and testifies to Aurora’s data quality. The error bars and repeatability detailed in Table 2 suggest marginally better data quality with Aurora, attributable to RedShift BioAnalytics’ improved signal averaging with lower volume consumption for each replicate.

Figure 4. MMP secondary structure plots from the second-generation Aurora instrument. (A) Similarity plots (baselined second derivative spectra) illustrating the peaks within the amide-I band for each MMP. (B) HOS plot displaying the relative abundance of each secondary structural element for the MMPs. The error bars represent +/- the standard deviation (N=3). Image Credit: RedShiftBio

Calculating structural similarity among MMP proteins requires assessing the area of overlap (AO) between the similarity plots, as displayed in figures 3A and 4A. Table 2 contains data on the replicability of the triplicate determination inside all samples and the sample-to-sample similarity shown as a percentage of AO.

For a detailed comparison, each MMP protein served as the control sample, marked with 100 % similarity. The first column utilized MMP-2, the second utilized MMP-8, and the third utilized MMP-9 as the reference.

This emphasizes the idea that proteins are “similarly different” from one another. As seen in this study, two proteins might have the same percent similarity with a third protein, although they still possess unique structures.

Table 2 shows that regardless of the protein under comparison, the other two proteins continuously demonstrate relatively low similarity scores (84–87 % sample-to-sample similarity with 96–97 % replicate repeatability and > 98 % for Aurora). This suggests that all three proteins have varying structures.

Table 2. Repeatability of measurement and sample-to-sample similarity (the control for each similarity comparison is set at 100 %). The top portion of the table is data from Apollo, and the bottom is comparing data from Aurora. Source: RedShiftBio

Most MMP isoforms, such as those assessed in this research, have a C-terminal hemopexin-like domain in their full-length protein following expression (proenzyme). Figure 5, displaying MMP-9 as a dimer, shows that this domain has a four-bladed β-propeller structure comprising symmetrical blade-shaped β sheets.

Such β-propeller structures are usually identified due to their critical role in protein-protein interactions.

MMP-9 can consequently often exist in dimeric form. The information gathered from this research suggests the presence of a discrete quantity of intermolecular β-sheet signals at 1618 and 1624 cm^-1, indicating the potential for dimerization or protein-protein interactions in all solution samples.

This demonstrates the potential of MMS to assess protein-protein interactions that involve the formation of intermolecular β-sheets.

Figure 5. The hemopexin-like domain of MMP-9 in dimeric form (PDB: 1ITV). Image Credit: RedShiftBio 

Conclusions

This study assessed the proenzymes of three MMPs (MMP-2, 8, and 9) via two MMS systems: the Apollo and the Aurora. This is additionally the first side-by-side comparison of data obtained from these two systems.

Despite all three proteins originating from the same MMP family and both MMP-2 and MMP-9 being gelatinases, the MMS results underscored varying secondary structures in these three proteins. Strikingly, the structures of MMP-2 and MMP-9 were as different from each other as they were from MMP-8, contrasting what the activity would otherwise indicate.

MMP-9 demonstrated the most intermolecular β-sheet structure, providing insights into the hemopexin-like dimers. The identified structural variation in the proenzymes corresponds with each of their own crystalline structures.

These findings provide crucial data regarding the structural relationships between proenzymes and active MMP enzymes under preparation conditions.

References and Further Reading

1. Verma, R. P., & Hansch, C. (2007). Matrix metalloproteinases (MMPs): Chemical–biological functions and (Q) SARs. Bioorganic & medicinal chemistry, 15(6), 2223-2268.
2. D Avila-Mesquita, C., et al. MMP-2 and MMP-9 levels in plasma are altered and associated with mortality in COVID-19 patients. Biomed Pharmacotherapy. 2021;142:112067.
3. da Silva-Neto, P. V., et al. Matrix Metalloproteinases on Severe COVID-19 Lung Disease Pathogenesis: Cooperative Actions of MMP-8/MMP-2 Axis on Immune Response through HLA-G Shedding and Oxidative Stress. Biomolecules. 2022;12(5):604.
4. Gelzo, M. et al. Matrix metalloproteinases (MMP) 3 and 9 as biomarkers of severity in COVID-19 patients. Sci Rep 12, 1212 (2022).
5. Kendrick, Brent S. et al. "Determining spectroscopic quantitation limits for misfolded structures." Journal of pharmaceutical sciences 109.1 (2020): 933-936.
6. Morgunova, E., Tuuttila, A., Bergmann, U., Isupov, M., Lindqvist, Y., Schneider, G., & Tryggvason, K. (1999). Structure of human pro-matrix metalloproteinase-2: activation mechanism revealed. Science, 284(5420), 1667-1670.
7. Bertini, I., Calderone, V., Fragai, M., Luchinat, C., Maletta, M., & Yeo, K. J. (2006). Snapshots of the reaction mechanism of matrix metalloproteinases. Angewandte Chemie International Edition, 45(47), 7952-7955.
8. Elkins, P. A., Ho, Y. S., Smith, W. W., Janson, C. A., D'Alessio, K. J., McQueney, M. S., ... & Romanic, A. M. (2002). Structure of the C-terminally truncated human ProMMP9, a gelatin-binding matrix metalloproteinase. Acta Crystallographica Section D: Biological Crystallography, 58(7), 1182-1192.

About RedShiftBio

RedShiftBio is redefining the possibilities for analyzing protein structure and concentration.

RedShiftBio has developed a proprietary life sciences platform combining our Microfluidic Modulation Spectroscopy (MMS) and expertise in high-powered quantum cascade lasers that provide ultra-sensitive and ultra-precise measurements of molecular structure. These structural changes affect critical quality attributes governing the safety, efficacy, and stability of biomolecules and their raw materials. This combination of technologies is available to researchers in our fully-automated Aurora and Apollo systems and is backed by a global network of sales, applications, service, and support teams to address all market needs.

Alongside our commitment to further innovation in the formulations and development space, RedShiftBio also supports biopharmaceutical manufacturing with HaLCon, our bioprocess analytics platform, purpose-built to measure protein titer at time of need.

Led by an experienced management team with a proven track record of success in both large instrumentation companies and commercializing disruptive technologies, RedShiftBio is here to support your research, development, and manufacturing goals. Our instruments can be found in the majority of the leading biopharmaceutical companies and CDMOs in the world. We also run product demonstrations and process samples in the StructIR Lab, located in our Boxborough, MA headquarters, as well as at partner sites including the Wood Centre in Oxford, UK, Spectralys/UCB in Brussels, Belgium, and at Sciex laboratories in Redwood Shores, CA.

RedShiftBio is backed by Waters Corporation, Illumina Ventures, Technology Venture Partners, and one undisclosed leading life science company.

Sponsored Content Policy: AZoLifeSciences publishes articles and related content that may be derived from sources where we have existing commercial relationships, provided such content adds value to the core editorial ethos of AZoLifeSciences which is to educate and inform site visitors interested in life science news and information.