Single‐Particle Studies to Advance the Characterization of Heterogeneous Biocatalysts

Immobilized enzymes have been widely exploited because they work as heterogeneous biocatalysts, allowing their recovery and reutilization and easing the downstream processing once the chemical reactions are completed. Unfortunately, we suffer a lack of analytical methods to characterize those heterogeneous biocatalysts at microscopic and molecular levels with spatio‐temporal resolution, which limits their design and optimization. Single‐particle studies are vital to optimize the performance of immobilized enzymes in micro/nanoscopic environments. In this Concept article, we review different analytical techniques that address single‐particle studies to image the spatial distribution of the enzymes across the solid surfaces, the sub‐particle substrate diffusion, the structural integrity and mobility of the immobilized enzymes inside the solid particles, and the pH and O2 internal gradients. From our view, such sub‐particle information elicited from single‐particle analysis is paramount for the design and fabrication of optimal heterogeneous biocatalyst.


Introduction
Organicc hemists and chemicali ndustries are lastly embracing biocatalysis as ak ey enabling technology to access more complex synthetic schemes in am ore sustainable manner. [1,2] However,t he use of enzymes as isolated soluble catalysts suffers the limitations of the homogeneous catalysis where product purification is arduous and the catalysts are hardly reused. [3] Furthermore, the enzymes have been majorly evolved to work under mild conditions (room temperature, neutral pH, atmospheric pressure) within the crowded environment of the cell milieu, thus their stabilityb ecomes an issue when the chemical process demands harsh and diluted reactionc onditions. [4] Enzymei mmobilization is an old solutionf or both the solubility and the stabilityo ft he enzyme. Paraphrasing DiCosimo et al., [3] the immobilizede nzymes are essentially as pecialized form of heterogeneous catalysts -heterogeneousb iocatalyststhat can be recovered and reused,o ften retain activity forl ong periodsa nd are amenable to aw ide variety of reactor designs, including flow-reactor for continuous processes. Nevertheless, the immobilization of enzymes on solid materials poses some limitations such as mass transfer issues,i nternal gradients (pH, temperature, substrates, etc.) and negative effects that the immobilization itself causes on the enzyme properties. [3,5] To overcome those drawbacks, advanced analytical tools shoulds upport the better comprehension of the effects causedb yt he immobilization on the enzymep erformance. In this context, like in heterogeneousc atalysis, [9][10][11] spatio-temporal characterization of heterogeneous biocatalysts contributes to their optimization.H ence, understanding how enzymes work inside as olid particlem akes the fabrication of heterogeneous biocatalysts more rational. Unfortunately,e nzyme immobilization is as till trial-and-error approachw here characterization studies are still based on observable parameters at macroscopic level. [5,12] These averaged measurements are collected from information gathered at the liquid bulk solution without considering the spatialh eterogeneity of the sample. Thus, these macroscopics tudies hardly reveal the spatio-temporal performance and the intraparticle environments of the heterogeneous biocatalysts.L ike inside the living cells, the enzymatic systemss upported on porous materials are dynamics, therefore analytical studies with spatio-temporal resolution can provide essential information about the properties of the heterogeneous biocatalysts at single-particle level.T hese techniques can reveal the structural rearrangements undergone by the immobilized enzymes with spatio-temporal resolution, the 3D-organization of cell-free biological systems, the mass transport fluctuationso ft he reactants and the reactionk inetics inside the microstructure of the porousm aterials under operando conditions ( Figure 1).
As in heterogeneousc hemical catalysis, [13] single-particle studies reveal how the spatial heterogeneityo ft he solid Immobilized enzymes have been widely exploitedb ecause they work as heterogeneous biocatalysts,a llowing their recovery and reutilization and easing the downstream processing once the chemical reactions are completed. Unfortunately,w e suffer al ack of analytical methods to characterize those heterogeneousb iocatalysts at microscopica nd molecular levels with spatio-temporal resolution, which limits their design and optimization. Single-particles tudies are vital to optimize the performance of immobilized enzymes in micro/nanoscopic environments. In this Concept article,w er eview different analytical techniques that address single-particle studies to image the spatial distribution of the enzymes across the solid surfaces, the sub-particle substrate diffusion, the structural integrity and mobility of the immobilizedenzymes inside the solid particles, andt he pH and O 2 internal gradients. From our view, such sub-particlei nformatione licited from single-particle analysis is paramountf or the design and fabrication of optimal heterogeneous biocatalyst. materials (material defects, [14] size dispersion, [15] functional patterning, [16] etc.) influence the final properties of the immobilized enzymes. Therefore, insights in protein conformation, mass transfer effects and enzyme kinetics at the micro/nanoscale within as ingle-particle contribute to understand the observable productivity and stability of the heterogeneousb iocatalysts determined by macroscopic analysis (Figure 1). [17,18] In the last two decades, single-molecule [19] and single-cell [20] studies have paved the wayt ob etter understand the dynamics of biological processes and aided synthetic biology to succeed in differentb iotechnological applications. These studies have not only released fundamental biological and biochemical knowledge but also provided analyticalt ools that shine light on protein localization,p rotein conformations and metabolite transport within the cellular microenvironment. These studies show the immense technical possibilities existing for microscopic characterization of porous particles,w hich are by far simpler and better controllable than living cells. Althought he intraparticle characterization of heterogeneous biocatalysts has received much less attention, single-particle studies are gaining popularity between the biocatalysis community to characterize the internal properties of ready-to-use heterogeneous biocatalysts just before their operational evaluation. [17] Particularly, these studies are rather important in flow-biocatalysis;atopic that is intensively investigated nowadays. [20] Unfortunately, methods to imaging the performance of the immobilized enzymes in flow-reactors [21] (monoliths, microfluidic channels, packed-beds) are scarce, since looking inside the reactors while they are operating is rather more challenging than studying the isolated particles under the microscope.
In this Concepta rticle, we outline some of the most relevant advances for the characterization of heterogeneous biocatalysts at single-particlel evel. We will review different analytical techniques that elicit the spatialo rganization, the kinetics and the stability of immobilized enzymes within solid particles, as well as the mass transport of reagents and the pH and oxygen gradients inside the porous microstructure during the bioprocesses. This revision ultimately aims to stress the importance of single-particlea nalysis as driving force to optimize enzyme immobilization proceedings. From our view,t he information elicited fromt he interior of the particles is paramount for the design and fabrication of optimal heterogeneous biocatalysts.

Single-Particle Spatial Distribution of Immobilized Enzymes Across Solid Carriers
The three-dimensional distribution of immobilizede nzymes on solid carriers plays ac ritical role in the study of the biocatalyst activity and its reactionk inetics ( Figure 1). However,t he enzymesh ave been traditionally assumed to be homogeneously distributed across the surfaceo fsolid carriers. In fact, mosto f immobilization procedures lack the control of both the orientation and the spatiald istribution of the enzymes, resultingi n the misinterpreting effectsu nderlying the conventional characterization of heterogeneous biocatalysts.I nt his context, the preparation of immobilized enzymes demands new methodologies to elicit the spatialo rganization of enzymes across the microstructure of the solid carriers. [20,23] Over the past few decades, atomic force microscopy (AFM), [24,25] low-temperature field-emission scanning electron microscopy (Cryo-FESEM), [26] spherical aberration (Cs)-corrected STEM, [27] infrared and fluorescence spectroscopya mong others, have emerged for inquiring into the spatiald istribution of immobilized proteinsa cross the solid surfaces. Micro-Infrared and micro-RAMAN spectroscopiesi mage the spatial distribution of label-free enzymes acrosssolidcarriers (particles 20-200 mm), unveiling microscopic informationo fr eady-to-use heterogeneous biocatalysts. [28,29] However, fluorescence studies using confocal laser scanning microscopy (CLSM) are most widely preferred, owing to its high resolution and versatility. [30,31] In this field, many techniques are availablet os tudy the distribution of fluorophore-labeled proteins,f rom as inglem olecule to am ulti-enzymatic system.
All proteins, and consequently all enzymes,s how intrinsic fluorescencei nt he UV region of the spectrum, owing to the presence of three fluorescent amino acids (tyrosine, tryptophan and phenylalanine) in their primary sequence. [32] This strategyi sm ainly exploited to study protein conformational changes of the both soluble and immobilized enzymes. [33] Mappingt he proteins by recording their intrinsic fluorescence would be ideal because this techniquew ould be universal and not require anyl abeling, however, as far as we know,n onee xample hasbeen reported yet.
The main strategy to analyze the spatiald istribution of enzymes across solids urfaces is based on fluorophore-labeled proteins.T he fluorophores can be either genetically encoded (e.g. fluorescent proteins) [34] or organic labels (e.g. fluoresceine). [35] Bolivar et al. exploited fluorescent proteins to modulate the protein distribution acrossp orousa garose beads (50-150 mm). [36] Thisw ork demonstrates that the immobilization rate controls the protein distribution, observing uniforma nd not uniform protein distributionsw hen the immobilization was slow and rapid, respectively.T he immobilization rate can be easily modulated by controlling the immobilization chemistry (nature and density of the reactive groups) and the immobilization conditions (presence of competitors, pH, temperature). On the other hand, the single-particles tudies of green fluorescent protein tagged with al ectin domain allowed monitoring the spatial distribution of the immobilized proteins over the time. [37] This study revealed that protein is primarily immobilized on the outer surfaceo fp orous agarose beads, but then gradually colonizes the whole microstructure of the agarose particle( 50-150 mm) along 30 minutes.T hese results evidence that the interactions between the agarose surface and the lectin domain are dynamic and suggest an intraparticle association/dissociation equilibrium between the lectin and the sugars forming the agarose fibers. Such equilibrium seems to enable the reorganization of the spatial distribution inside the particles upon the immobilization.
However,w hen using non fluorescent proteins, chemical labeling with chemical fluorophores is an excellent solution to study the spatial distribution of immobilized enzymes. [38] The chemicall abelingo fs everal enzymes with different fluorophores have been successfully exploitedt od etermine the spatial organization of multi-enzymesystems immobilizedo ns olid carriers with different architectures. [39][40][41] The understanding of the spatial organization of multi-enzymes ystems immobilized on solid carriers is required to optimize the multi-functional heterogeneous biocatalystsi no rder to perform cascade reactions more efficiently.R ocha-Martíne tal. [40] tuned the subparticles patial distribution of two alcohol dehydrogenases by modulating their immobilization rates accordingt oaprevious work [36] (Figure 2).
As result,t hey observedt hat uniform co-localization of two enzymese nhances the in situ NADH recycling efficiency during the operation of ab io-redoxc ascade. The sub-micrometric proximity of the two dehydrogenases presumably exerts some cooperation effects that explain the low apparent K M values towards NADH, even lower than the ones presented by the soluble system, pointingo ut ac oncentration effect in the NADH pool within the porous surface. [40] These results revealt he high relevance of single-particle analysiso fp rotein distribution on the performance of heterogeneous biocatalysts. Therefore, the control of enzyme spatialo rganization is clue for improving the biotechnological processes catalyzed by heterogeneous systems.

Single-Particle Reaction Kinetics of Immobilized Enzymes
Monitoring and imaging the enzymea ctivity during the operational process (in operando studies) at the nano/microscale can directly provide information thatw ould be inaccessible otherwise. In ar ecent paper,H arada et al. exploitedh ighspeed atomicf orce microscopy to measure in operando the conformational changes of the cellobiosed ehydrogenase immobilized on flat gold surfaces functionalized with heme groups. [42] This work sheds light on the mechanism of this multi-domain protein at nanometric scale. The highly specialized infrastructure required for these studies makes that spatial resolution studies monitoring the catalytic activity of the immobilized enzymes are dominated by fluorescence spectroscopy;aworldwide and highly accessible technique. In the last decades, many advances [43][44][45][46][47][48] in development of fluorescence-based reporter systems have boosted the spatio-temporal resolution and sensitivity of analytical methods to study ab road range of dynamic processes during the catalytic reactions. In this context,many fluorogenic substrates are commercially available as useful reporter systems. [48] These fluorogenic substrates can be enzymatically transformed into fluorescent products that can be read out to assess the system kinetics ( Figure 3). [45,[49][50][51][52][53] Particularly,C LSM is av ery powerful technique for enzymology when enzymes are immobilized and allowsr ecording the local production of products. [54] One of the first studies to evaluate the kinetics of immobilized enzymes using fluorogenic substrates was the real time DNA sequencing. [55] Thew ell-known real-time DNA sequencing was performed using aD NA polymerase immobilized on a zero-mode waveguide (ZMW) pore (100 nm) of ag lass surface and fluorescently labeled nucleotides. The growth of the DNA chain was monitored by fluorescent bursts after incorporation of the fluorescent nucleotides. More recently,V elonia et al. demonstrated that single-particle enzyme kinetics can be monitored at real-time using CLSM as well. They used ai ndustrially relevant enzyme (lipase from Candida antarctica B; CALB) [56] immobilizedo nag lass surface derivatized with Figure 2. Effect of enzyme distribution inside agarose particles (50-150 mm) on the cofactor-recycling frequency.Alcohol dehydrogenase (Tt27-ADH2) and Glutamated ehydrogenase (Tt27-GDH) labeled with fluorescamine (green)a nd rhodamine B( red) respectively, performedabio-redoxcascade. The maindehydrogenase Tt27-ADH2 catalyzed the asymmetric bio-reduction of 2,2,2-trifluoro-acetophenone into (S)-a-(trifluoromethyl) benzyl alcohol with NADH consumption. Orthogonally,NADH-recyclingr eaction was performedb yT t27-GDH by using l-glutamate as sacrificing substrate. A) The two enzymeswereseparately immobilized on different agarose-Ni 2 + /Glioxil particles( AG-Ni 2 + /G):Tt27-GDH was immobilized in one batch by using IMAC chemistry,whereas Tt27-ADH2w as immobilized in ad ifferentb atch by using glyoxyl chemistry.The two batchesweremixed to perform the reaction( circles). B) The two labeled enzymes weresequentiallyco-immobilized onto the same AG-Ni 2 + /G particle throughd ifferentb onding chemistries:Tt27-GDH presented an ot uniform distributionwhile Tt27-ADH2 showed au niform one (squares). C) The two enzymes were homogeneously co-immobilizedo nto the sameAG-Ni2 + /G particle (triangles). In this case,Tt27-GDH was first immobilized in the presence of 0.2 m imidazole, which hindered the immobilization and enabled am ore homogeneousdistributiono fthe enzyme. Second, Tt27-ADH2w as sequentially immobilized through aldehydec hemistry.Reproduced from [40] with permissionofW iley. hydrophobic groups. Af luorogenics ubstrate was hydrolyzed by the immobilized CALB yieldingafluorescent product. These studies revealed that the enzymes absorbed on ah ydrophobic surfacep resentedakinetic variability attributedt ot he existence of different active enzyme conformations across the solid surface as result of the random immobilization mechanism. This sort of studies may provide spatial resolution to the kinetic behavior of the immobilized enzymes by imaging those locationsw here enzyme performance is altered.
Fluorogenic substrate-based strategies have served as a springboard to investigate more complex biological systems where multi-enzyme cascades are involved. Many factorsasenzymaticc ooperation, reactants transport, spatiald istribution, etc.,m ust be considered when studying multi-enzyme systems due to their intrinsic molecular complexity.T he reaction kinetics of the immobilizedmulti-enzyme system formed by glucose oxidase( GOX)a nd horseradish peroxidase( HRP) are the most studied at single-particlel evel by far. [41,45,[57][58][59] These two enzymes have been co-immobilized on al arge variety of surfaces throughadiversity of immobilization chemistries. Normally,a fluorogenic substrate (i.eA mplex red) is oxidized by HRP yielding af luorescent product (i.e Resorufin) only in presence of the hydrogen peroxide that is in situ produced as by-product from the oxidation of glucose catalyzed by GOX. Li et al. co-immobilized these two enzymes on copper-phosphate particles (10 mm) with different spatiald istributions. [41] The cascade workedm ore efficiently when the HRP was spatially confined inside the particles and the GOX was attached to the outer surface ( Figure 4A-C). As imilare nzymec ascade was co-immobilized on porous polymer monoliths by ap hotopatterning method, demonstrating again the importance of the spatial organization of the enzymes for continuouso peration under flow conditions. [21] By using two fluorogenic substrates, they observed that the fluorogenic glucose is firstly concentrated and oxidized in the outer surfacew here the GOX is located, whilet he fluorescent resorufin initially appears inside the particlesw here HRP was located and then gradually diffused out to the bulk (Figure 4D-F).
In am ore sophisticated architecture, severale nzymes are compartmentalized into different polymersomes (100-300nm) as sub-compartments within larger particles (60 mm) or oilwater droplets (80-1000 mm). [57,60] This approachw as applied for different multi-enzyme systemst hat catalyze sequential cascade reactions whose final product is fluorescent. The single-particle studies elicited the spatial distribution of the multi-enzyme system by imaging, after 3D computational reconstruction,t he fluorescences pots where the final product is primarily accumulatedb efore diffusing across the microstructure of the larger particle ( Figure 5). Those spots correspond to the polymersomes where the enzyme that catalyzes the last reaction of the sequential cascade is sub-compartmentalized. [57,60] The accumulationo ft he fluorescencei nto the polymersomei s explainedb yt he electrostatic entrapmento ft he product that preventsi ts rapid diffusion to other compartments within the micrometric polymersomeand to the reaction bulk. Additionally,t he accumulationo ft he final product in the compartments where the last enzyme is located also demonstrates an external transport of the substratea nd inter-compartment diffusion of the intermediates.T hese single-particle studies point out that the ordered transport of the substrates is the consequenceo fthes patial distribution of the multi-enzymes ystem co-immobilized on the solid materials.  . On line monitoring of single-particle kineticsb yfluorogenics ubstrates. (A-C) Confocal microscopy images of the co-immobilized HRP (labeledw ith rhodamine B, red) and GOx (labeled with FITC, green)i nnanocrystal complexes. A) GOx was firstly immobilized by precipitation with metal complexes. Then, HRP was absorbed to the precipitate surrounding the nanocrystal complex. B) HRP was precipitated andGOx was superficially absorbed to the nanocrystal complex. C) Both enzymes were precipitated with metals obtaining arandomdistribution. D) The cascade reaction catalyzed by GOx and HRP using fluorogenic substrates.E)Confocal images showingthe transport of af luorescentg lucose analogue (6-NBDG) from outside to inside in the biocatalytic system GOx@HRP. F) Confocal images showing the transportofr esorufin from inside to outside of GOx@HRP. Adapted from [41] with permission of Royal SocietyofC hemistry.
Cofactor-dependent enzymes( e.g. NADH-dependent reductases and oxidases, FAD-dependent oxygenases) catalyze many interesting reactions in industrial biocatalysis. [58][59][60] Many of these cofactors presenta uto-fluorescence that can be monitored during the enzymatic activity.U nlike the soluble enzymes where single-enzyme dynamics in presence of cofactors have been reported more extensively, [61] monitoring the cofactor utilization by immobilizede nzymes with spatialr esolution is underexploited. FADisanexcellent cofactorfor these studies because the cofactor-enzyme electron transfer causes fluorescence fluctuationst hat can be attributed to different cofactor conformations ands tates. Nevertheless mapping of those fluctuations remains unknown. [65] These single-molecule studies have served to advance in the understanding of the kinetics of cofactor-dependente nzymes observed in bulk studies. [65,66] In addition, the effect of the reaction conditions and mass transport restrictions on the enzyme activity can be studied with spatialr esolutionb ym apping the cofactor utilization within the solid phase. [67][68][69] In this regard, nicotinamide cofactors whose fluorescence intensity is higherf or the reduced form (NAD(P)H) than for the oxidized one (NAD(P) + )h ave been very useful to evaluate the activity of different immobilizeda lcohol dehydrogenases (ADH) under different operational conditions. Using soluble NAD + ,O 'Brien [70] et al. created artificially local pH gradients with microelectrodesa nd monitored the local activity of an immobilized alcohol dehydrogenase. Imaging the cofactor fluorescence, the authors demonstratedt hat only those enzymes surrounded by alkaline pH environment were able to efficiently reduce NAD + to NADH using ethanol as substrate.T his effect was observed with enzymes immobilizedo n both glass surfaces andp orous beads.
On the other hand, cofactor and enzymes can be co-immobilized to fabricate self-sufficient heterogeneousb iocatalysts that do not require exogenous supply of cofactor.E ven thought he concepto fs elf-sufficient heterogeneous biocatalyst has been exploited for ad ozen of cascade reactions, [40,[71][72][73][74] the cofactoru tilization inside the solid particles has been rarely studied. Velasco-Lozano et al. [69] have mapped the cofactor utilization within porous agarose particlest hat co-immobilize the main enzyme (alcohol dehydrogenase), the cofactor recycling enzyme (formate dehydrogenase) and the cofactor (NAD + ). In operando studies by means of fluorescencem icroscopy demonstrate that the co-factor is catalytically available for the coimmobilized enzymes andi ns itu recycled inside the porous surfacewithout diffusing out to the reaction bulk ( Figure 6).
Finally,s ingle-particle fluorescence studies have also provided valuable information about the kinetics of cell-free protein synthesis (CFPS) systems; ah ighly complex biological machinery.I nt hese systems, both transcriptional and translational machineries have been encapsulated together with DNA molecules in efforts to build am inimal cell. [75] To monitor the protein synthesis reactionw ithin the solid particles, fluorescent proteins (green fluorescent protein (GFP), [76,77] Venus, [78] etc.) has been in vitro synthesized and simultaneously measured by fluorescencem icroscopy.K ato et al., [15] monitored the fluorescence of nascent GFP over the time inside oil droplets, these experiments allowed to quantify the cooperative action of this complex multi-enzyme system in real-time ( Figure 7). These studies revealed as ignificant effect of particles ize on the reaction kinetics;t he protein was synthesized 5times faster in smaller particles (20 mm) than in larger ones (71 mm). In this complex scenario, the translation rate of isolated immobilized ribosomes synthesizingG FP was locally recorded, and the observations demonstrated that local and bulk protein synthesis kinetics were similar. [79] These single-particle studies of immobilized systems open new pathwayst os tudy the dynamics and the kinetics of biomanufacturing cascades spatially confined. Also, these techniquess howasuccessful in operando monitoringo fe nzyme kinetics in consecutives reactions by using immobilized enzymes. Hence, the information provided shinesl ight on the factors that locally affect the activity of immobilized enzymes. From these studies, we can evaluate the functional uniformity of one heterogeneous biocatalyst and understand the causes  (1) is catalyzed by phenylacetone monooxygenase(PAMO) with one unit of NADPHb eing consumed, to yield ester (2), which is subsequently hydrolyzed by the enzymel ipase Bf rom Candida antarctica (CalB) or the enzyme alcalase to provideaprimary alcohol (3). Alcohol dehydrogenase (ADH) oxidizes the alcohol, by using the cofactor NAD + ,t og ive aldehyde (4) which then undergoes spontaneousb eta-eliminationt obec onverted into resorufin (5) as the final fluorescentproduct.C )Enhanced spinningd isk confocal fluorescence imagingats ingle particle level of acellmimicafter production of resorfurin. The fluorescent product is easily observable like spots in the compartmentalized organelles. D) The 3D representationo ft he polymersome where the organelles producing resorfurin are highlighted. Adapted from [60] with per-missionofR oyal SocietyofC hemistry.

Single-Particle Stability of Immobilized Enzymes and Proteins by using Spectroscopic Methods
Beside easing the downstream processing, the immobilization of enzymes also may stabilizet hem against certain reaction conditions underlying the industrial processes. [6,80,81] Normally, the immobilization provokes as tructuralr igidification in the enzymes which may prevent them from the structural distortions triggered by inactivating agents such as high temperature, organic solvents, [82] etc. Thereupon, ab etter understanding of molecular processes that lead the protein rigidification may forecast the resulting stability and activity of enzymes upon the immobilization.T hisc omprehension would be helpful for the fabrication and optimization of heterogeneous biocatalysts.
Conventionally,s tability of immobilized enzymes is determined through bulk studies where the samples are incubated in presence of the corresponding inactivating agent, and the averaged enzymatic activity of each sample is measured at differentinactivation times. [83] Additionally,microcalorimetric analysis, Trp-fluorescence studies and circulard ichroism (CD) [33,[84][85][86] can providei nformation aboutt he structural stability of the immobilized enzymes. Nonetheless, thesea pproaches lack the spatialr esolution needed to grasp how the enzymes are stabilized at single-particle level. Fluorescence spectroscopy has lastly emerged as ap owerful tool for single-particle studies of immobilized enzymes,a ided by the outstanding advances in optical microscopy. [69,[87][88][89] Orregoe tal.,c ombined fluorescence lifetimei maging and spatial-resolved fluorescence polarization to determinet he fluorescence anisotropy of immobilized proteins as innovative strategy to analyze the stability [90] of immobilized proteins at single-particlelevel ( Figure 8).
In this study,s everalE GFP variantsw ere immobilized on porousa garose microbeads through different immobilization chemistries. The fluorescenced ata allowed determining the anisotropy of EGFP inside the beads with spatialr esolution. Based on that information,t he authors found ac orrelation Figure 6. In operando analysisoft he NAD + /H utilizationa ts ingle-particle level. Alcohol dehydrogenase from Thermus thermophilus (Tt-ADH2)a nd Formate dehydrogenasefrom Candida boidinii were irreversibly co-immobilized with the cofactorN AD + ionically absorbed on agarose beads (50-150 mm). The cofactore stablishes an association/dissociationequilibrium that allows its internaldiffusioni nside the pores but avoidsi ts external diffusiontot he bulk. A) Single-particle monitoring during the redox biotransformationof the substrate formic acid (red~)and incorporating the NAD + recycling with 2,2,2-trifluoroacetophenone(1) (green *).Thereaction was monitored and the averagef luorescence was quantified by measuringt he autofluorescence of NADH at 460 nm for 15 minutes in 10 microbeads.B)Analysisof the fluorescence intensity of single beadsf rom fluorescence microscopy images with (below)and without NAD + recycling (above), before (left) and after (right) the bioredox reactionswere accomplished.A dpated from [69] with permission of Wiley. between the protein flexibility and the thermal stabilityo ft he immobilized proteins.A sc onclusion, the fluorescencea nisotropy reveals that proteins immobilized through chemistries that reduce the protein flexibility (high anisotropy values) are more thermostable than proteinst hat remainm ore flexible upon the immobilization.
On the other hand, atomic force microscopy (AFM) and spectroscopy (AFS) are broadly used to probe protein-surface interactions. [91][92][93] AFM is the only microscopic technique capable of creatingt opographical maps where visualizing biomolecules at the single-molecule level with sub-nanometer resolution in liquid. [93] Aissaoui et al. [94] have exploitedA FM to image the aggregation of the enzymes immobilized on planar silanized surfaces (1 1 mm 2 )t hrough different immobilization chemistries. They found ac orrelation between the size of the aggregates and the catalytic properties of the immobilizedg lucose 6-phosphate dehydrogenase. The smaller aggregatese xhibit highers pecific activity likely owing to less transport restrictionso ft he substrates, whereas the larger aggregatesi ncrease the thermal stability of the immobilized proteins, suggesting that such aggregation packs the 3D protein structure (Figure 9). Beyond imaging, AFM can also perform spectroscopic studies to reveal the stiffness of those materials conjugated with biomolecules.T his technique has been highly informative about the mechanical properties of the cell walls when displaying different proteins. [95] In the fieldo ft he immobilized enzymes, Gregurec et al. [96] applieds ingle-bead atomic force spectroscopy to predict the thermals tability of several immobilizedp roteins. Here, the authors used ac olloidal probe (1 mm) to indent9 mm 2 surface of the single microbeads (120 mm) upon the immobilization of different oxidoreductases. The immobilization could be on line monitored in the microscope chamber because the conjugation of the proteins to the solid surface increased the stiffness of the whole bead. The immobilization chemistrya ffects differently to the mechanical properties of the beads upon the immobilization of the same protein. The resultss howed that irreversible andm ultivalent immobilization chemistries increase the stiffness of the microbeads and promote significant thermal stabilization of the immobilized enzymes. Remarkably,f orce spectroscopy studies arrivedt ot he same conclusion that the studies based on fluorescence anisotropy of immobilized enzymes. [90] Infrared (IR) techniques, including Fouriert ransform infrared (FTIR),a ttenuatedt otal reflectance coupled with FTIR (ATR-FTIR) and sum-frequency generation spectroscopies have been widelye xploited to evaluate the structurali ntegrity of the enzymes upon the immobilization. [97][98][99] Monitoring the signal corresponding to the Amide Iband of proteins,IRs tudies have revealed structural distortionso ft he immobilized enzymes, owing to the hydrophobicity of the surface, the crowding of the immobilizede nzymesa nd their orientation;a ll these factors directly affect the thermal stabilityo ft he immobilized enzymes. [99] To gain spatial resolution, the immobilized enzymes have been analyzed with micro-FTIR spectroscopy;apowerful technique for in operando studies in heterogeneousc hemical catalysis [10,100] althoughi ts applicationsf or heterogeneousb iocatalysts are still on its infancy.T he group of Prof. Lepore has exploited micro-FTIR to analyze the local conformation of glucose oxidase entrapped into as ol-gel matrix. [101][102][103] Beside the spatiald istribution of the enzymea cross the solid material, micro-FTIR unveils some local patches of structurally distorted glucoseo xidase upon the immobilization, although the enzyme activity was macroscopically preserved. Additionally, micro-ATR/FTIR analysis of the same immobilized enzyme shows at emporal evolution of the infrared spectra within the solid particles that can be correlated to the enzyme inactivation over the time. Therefore, mapping properties such the aggregation and crowding states,t he orientation, the flexibility and the stiffness of the immobilized proteins unmaskr elevant information that remains hidden in bulk experiments. Such information is vital to understandw hy immobilization often stabilizes enzymesa nd to furthero ptimize the fabrication of heterogeneous biocatalysts. We envision the application of singleparticles tudieso ver the time under different reactionc onditions to better understand the inactivation kinetics of the immobilized enzymes during the operational process.  (50-150 mm). AG10 is plain agarose, while AG10-Ni 2 + and AG10-G are agarose beads activatedwith nickel chelates and aldehyde groups,r espectively.LSL-EGFPi sal ectin tagged enhanced fluorescentp roteint hat is immobilized on AG10t hrough the lectin domain as spacer arm. His-EGFP is the same fluorescentprotein tagged with ap oly-histidine tagthat allowsi ts univalent and reversible immobilizationo nA G10-Ni 2 + and EGFP is the untagged protein that is multivalently and irreversibly immobilized on AG10-G.A )F luorescence intensity( grey scale) and B) fluorescence anisotropy (rainbow-like scale) of horizontal sections (80 80 mm) at the equator plane of the representativebeads where EGFPvariants are immobilized through different chemistries. The heterogeneous biocatalysts with higher anisotropy values werem ore thermally stablet han those onesw ith lower anisotropyv alues. Adapted from [90] with permission of AmericanC hemical Society.

Spatio-Temporal ResolutionofO 2 and pH Within ImmobilizedEnzymes
Enzymes immobilizedo np orous carriers suffer from diffusional effects, exhibitings ubstrate and product concentration gradients between the bulk liquid and the carriers urface; [18] pH and O 2 concentrationa re two variables particularly affected. [17,104] Both variables are important in biocatalysis since many enzymatic reactions create proton and O 2 gradients during the operational processes. TheO 2 limitation lies in its lows olubility in aqueous mediumt hat originates as mall driving force for O 2 supply into the solid biocatalyst;consequently,the low transfer rate comparedt ot he enzymatic reactionr ate leads to the dramatic depletion of the intraparticle O 2 concentration. [17,104] In this scenario, O 2 -dependent heterogeneous biocatalysts offer a very low apparent activity.O nt he other hand, the pH gradients are created due to either partitione ffects within the solid material [105,106] or higher protonr elease/consumption enzymatic rates compared to the protont ransfer rate. As consequence, the existence of internal pH gradients influences on the activity,s tability and selectivity of the heterogeneous biocatalyst. In both cases, dissecting whether apparent catalytic properties are due to immobilization effectso rt oc ertain intraparticle environment [17,18] is necessary for the biocatalyst optimization.
Real-time determination of concentration gradients between the internal environmentofthe immobilized enzymes and the bulk solution is the best way to assess the significance of the diffusional limitations and to offer ab etter biocatalyst characterization. [17,18] There are many available solutionst om easure O 2 andp Hi nh omogeneous liquid phase with excellent spatio-temporal resolution. [107][108][109][110] However, the measurements near-surface or within solids particles are scarce since they present ahigh technological complexity and requires pecializeds et-ups. The basic principle involves al uminescence indicator embedded within an analyte-permeable polymeric layer.T his is furthers ystem-integrated by controlled deposition as in sensor spots and optical fiber tips, enablingdifferent and often contactless optical readouts. However,t he methodology in the current form cannotb eu sed to analyze the pH within solid porous. [110,111] The application for characterization of internal gradientm easurements has two specific requirements. First, the luminescence dye and the enzyme should be properly co-immobilized within the same porous particle. Second, ar ead-outs et-up should be established to provide measurements with suitable spatio-temporal resolution. Ad etailed description can be found in recent reviews. [17,18]

Resolution of IntraparticleO 2 -Concentration
The quantification of spaced-averaged intraparticle O 2 -concentrationsh as been applied to different enzyme porous carriers containing immobilizedo xidases. [104,112,113] Polymethacrylate porousc arriers were made O 2 sensitive by labeling with an O 2 sensitive luminophore and the analytical principle was the lifetimem easurements based on applicationo ft he phase modulation technique. [104,112] The authors observed the formation of al arge O 2 concentrationg radient between the bulk and the intraparticle environment, which clearly indicates the O 2 -supply limitations within the solid carrier. [104] The determination of oxygen gradients between homogeneous liquid phase and internal catalytic environment was performed simultaneously to the bulk determinationo fc atalytic activity of immobilized oxidases) [104,112,113] (Figure 10). These studies showedt hat the internally available O 2 concentration controls the catalytic effectiveness of heterogeneous biocatalyst. [104,112,113] Thus,aclear distinction between the effect of immobilization and substrate limitationsw as made possible. [18,112] Hence, the application of internal sensing enabledt he optimization of geometricalp roperties (particle and pore sizes) of porous silica carriers to obtain biocatalytic process intensificationt hrough enhanced mass transport [113] (Figure 10 B-C).
Furthermore, internal measurements have provedt he possibility of enhancement of intraparticle environmentt or each conditions not achievable in liquid phase. [114] Intraparticle monitoring has allowed observing the releaseo fi nternal oxygen from H 2 O 2 using immobilizedc atalase, and conditions of

Resolution of IntraparticlepH
Space-averaged intraparticle pH has been measured for the characterization of heterogeneous biocatalystsi nvolving systematicb iotransformationso ptimization, reaction modelling, reactioncontrol and determination of kinetic and mass transfer parameters. [17,18,[115][116][117][118][119][120] The hydrolysiso fb-lactams ubstrates (which results in net proton formation) was studied by using FITC-labeled immobilized amidase in order to quantify the extento ft he overall carriera cidification during the operation. [121] The internal pH was determined from pH-sensitive fluorescencei ntensity of FITC, showing ap Hd ifferenceo f3units between the bulk and the interioro ft he particles. This principle was applied to af ixed-bed reactor to monitor the internal pH across the reactor length. Combination of internal data ande xternald ata for reaction modelling was then possible and facilitates process understanding and targeted reactor selection. [121] Moreover,t he determination of intrinsic kinetic parametersa nd mass transfer coefficients within porousm aterials is strongly supported when intraparticlec oncentrations are used instead of uniquely externald ata. [120] Recently,l ifetimem easurements (dual-lifetime referencing method, DLR) have improvedt he pH resolution [116,122] and served to study the influence of the carrier properties on the pH drop andm onitort he reaction time course.
Moreover,D LR has been extended to apply new control strategies based on pH measurement within catalytic environment. [115] Finally,t he internal pH has been used as key parameter to increase the lifetime of the immobilized enzyme under operational conditions. [117]

Internal Sensing at Single Particle Level
Opto-chemical sensing in combination with microscopy has the potentialt od etermine internal pH and O 2 concentration in real-timea nd with spatial resolution (Figure 11 A). However, only ah andfulo fe xamples have been reported.S pieß and colleagues performed ap ioneer study to characterized ifferent catalysts of immobilized penicillin Ga midase by using referenced fluorescencei ntensity measurements of internal pH in aC LSM. [123] Their study was seminal in demonstrating the importance of internal pH to optimize the enzymei mmobilization. They showed that internal pH alters the selectivity of the immobilized amidase, implying the need to select carriers and immobilization chemistries that providea noptimum internale nvironment. Huanga nd colleagues applieds imilar analytical techniques to determine pH gradients in biocatalytic membranes containing immobilized glucose oxidase. Ap Hd rop resulted in this case from the oxidation of d-glucose into d-gluconic acid. [124] The application of fluorescencel ifetime [125] or multiphoton laser scanningm icroscopy [126,127] solve some of the known limitations of measurements in CLSM. [17,18] Consequently,t hese techniquesh ave been applied to the spatial resolution of internal pH inside hydrogels particles (1.5 mm) where the events of substrate diffusion and enzymatic reactions were resolved by elucidating the local concentrationg radients through the catalytic particle. [127] Unfortunately,l ifetimeo rreferenced   [17] and [128]w ith permision of Cell Press and American Chemical Society. ChemCatChem 2018, 10,654 -665 www.chemcatchem.org 2018 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim measurements in CLSM depends on high-cost instrumentation that cannotb ea dapted to real-lifer eactor configurations and has limited throughput capacity, this has probablyl imited their application for heterogeneousb iocatalyst at single-particle level.
The spatio-temporal mapping of O 2 has been restricted due to the difficulty of applying lifetimei maging in the range of microsecond. Recently,anew method based on variable excitation time determined by the scanningv elocityw as implemented in aC LSM (Figure 11). The method allows phosphorescence lifetime imaging and thuss patio-temporal resolution within porouse nzyme carriers. [128] It was applied for the study of the oxygen depletion within particles containing immobilized lactate oxidaseunder packed-bed reactor configuration.

Summary and Outlook
Immobilized enzymes have been widely exploitedb ecause they work as heterogeneous biocatalysts,a llowing their recovery and reutilization and easing the downstream processing once the chemical reactions are completed. Although immobilized enzymes have been utilized since decades, they are still considered as a" black box" where the effects of the surfaceon the enzyme properties are poorly understood at microscopic level. The lack of that informationh as limited the rational design and optimization of more efficient heterogeneousb iocatalyst. In this Concepta rticle, we have reviewed how singleparticles tudies provide fundamentali nformation about the functionality,t he structural integrity and the microenvironments of the enzymes immobilized on solid materials. In the last three decades, the single-particles tudies of immobilized enzymesh ave advanced in the characterization of heterogeneous biocatalysts eliciting information that is masked in macroscopic studies based on bulk experiments. The latest advances in spectroscopict echniques with spatial resolution have boosted the characterization of immobilized enzymesa ts ingle-particle level, whicha llows better understandingt he operational performance of the enzymes bound to solid materials. In this article, we have given examples of how single-particle studies image the spatial distribution of the enzymes,t he substrate diffusion and the reaction kinetics across the solid surfaces, the structural integrity and mobility of the immobilized enzymes inside the solid particles, andt he pH and O 2 internal gradients. All these data provide vital information to unveil the optimal localization of immobilized enzymatics ystems to more efficiently catalyze chemical reactions, the optimal attachment between the solid surface and the enzymet oy ield more thermostable heterogeneous biocatalysts,a nd the optimal carriera rchitecturet or educe the mass transport limitations for reactants and oxygen. All this information contributes to develop more rational, reliable and reproducible proceedings when fabricating heterogeneous biocatalysts.
Nevertheless,t here is still al ong wayt oe licit structural and functional properties of immobilizede nzymes on porous materials at the nanometric scale, an even longera ta tomic-scale. Additionally,i noperando studies at single-particle level must gain in temporalr esolution to studyt he reactants utilization and the protein conformationalc hanges during the enzyme catalysis at very short-times (ns-ms). Likewise, we need to expand theset echniques to monitore nzyme properties in more complex chemical process, including chemo-enzymatic ones, and more sophisticated operational designl ike flow-(micro)r eactors. Therefore, the fabrication of extremelyo rdered materials and the emergence of high-resolution techniquesf or the solid-state will open new windows of understandingf or this specialized form of biocatalysis.