Immobilized heparinase: In vitro reactor model

Heparinase immobilized to agarose has previously been shown to be useful in degrading heparin and thereby preventing thromboembolytic complications when this anticoagulant has been used in extracorporeal perfusions. The current study examined the kinetics of this immobilized enzyme. When heparinase is covalently bound to 8% agarose, the partition coefficient of heparin in the catalytic particle is 0.36 +/- 0.048 (N = 10). The immobilized enzyme has a K(m) of 0.15 +/- 0.03 mg/mL and an activation energy of 10.3 +/- 0.57 kcal/gmol (N = 5). These values are statistically indistinguishable from the values for the free enzyme. The immobilized enzyme showed a pH activity optimum between 7.0 and 7.4, compared to the optimum pH of 6.5 for the soluble enzyme. The activity optimum of immobilized heparinase with respect to salt concentration was between 0 and 0.1M. A reactor containing immobilized heparinase recirculating internally at 1300 mL/min behaved as a continuously stirred tank reactor (CSTR) when solutions at a flow rate of 120 mL/min were passed through the device. The residence time distribution was determined using blue dextran (molecular weight 2 x 10(6) daltons), which is sterically excluded from the agarose catalyst. A model of the heparinase reactor based on ideal CSTR behavior and the immobilized enzyme kinetic parameters was developed. It accurately predicted experimental conversions over a range of catalyst volumes, enzyme loadings, and substrate concentrations to within 7% in most cases and with a maximum deviation of 13%.     

Investigation of a whole blood fluidized bed Taylor-Couette flow device for enzymatic heparin neutralization

The use of clinical bioreactors will increase as more therapeutic proteins are being cloned, expressed, and produced at a reduced cost. The proposed use of an immobilized heparinase I reactor to make heparin anticoagulation a safer therapy is an example of how the specificity and high activity of an enzyme could be incorporated into a system to ultimately benefit a patient. However, the development of a safe and efficient bioreactor is important for the use of immobilized heparinase I and other therapeutic proteins designed for use in medical extracorporeal procedures. This study examined the possibility of using Taylor-Couette flow and "flow-induced" recirculation of the agarose beads as a way to fluidize agarose-bound heparinase in whole blood. Heparinase I was immobilized onto agarose beads via cyanogen bromide activation. A reactor based on Taylor-Couette flow was designed and modified with a tangential recirculation line. The reactor was tested for efficacy and safety in vitro in human blood. Visualization studies in water and 42% glycerol were used to determine the minimum rotation rate for efficient fluidization. The strategic placement of the recirculation line allowed recirculation of the agarose without the use of an external pump. The device removed 90% of the heparin activity within 2 min from 450 cc of human blood at a blood flow rate of 100 mL/min. Furthermore, the device maintained inlet and outlet clotting times of 269 +/- 10 and 235 +/- 6 s, respectively, demonstrating the potential for regional heparinization. Blood damage was a function of gel volume fraction and rotation rate of the inner cylinder. Hemolysis of the red cells is an important issue when Taylor vortices are combined with macroscopic solid particles such as agarose beads. A modified Taylor-Couette flow device was developed to treat whole blood and operational criteria were established to minimize hemolysis.     

The influence of bond chemistry on immobilized enzyme systems for ex vivo use

The ideal derivatized support for the clinical use of an immobilized enzyme system should irreversibly bind active enzyme. We have investigated the behavior of heparinase and bilirubin oxidase immobilized via cyanogen bromide, tresyl chloride, epoxide, or carbodiimidazole activated natural and synthetic matrices. The protein bound to each activated support was 90% for cyanogen bromide (CNBr) activated agarose, 50-80% for tresyl chloride activated agarose, and 50% for oxirane activated acrylic (Eupergit C). The activity retention of immobilized heparinase was greatest (50%) with CNBr activated agarose while for the immobilization of bilirubin oxidase, the activity retention was greatest (25-30%) with tresyl chloride activated agarose and oxirane activated acrylic.The stability of the different covalent bonds was studied in vitro with radioiodinated enzymes. The leaching profiles showed the same trends for each support and chemistry. A plateau in portein leaching was reached after a few hours of incubatttion and the transient leaching period was well represented byu a logarithimic function of time. The amount of enzyme released from the least stable support (CNBr activated agarose) in 24 h was injected intravenously in New Zealand white rabbits. Using an indirect enzyme-linked immunnosorbant assay (ELISA), no immune responce was detected. The transient leaching profile was shortenend by washingthe enzyme-support conjugate with 1M hydroxylamine, pH8.5 intermolecular cross-linking with glutaraldehyde also improves the enzyme-support stability. Tresyl chloride and oxirane activated supports produce bonds with improved stability without adversely affecting enzymatic activity.     

An approach for the stable immobilization of proteins

An approach is presented for the stable covalent immobilization of proteins with a high retention of biological activity. First, chemical modification studies were used to establish enzyme structural and functional properties relevant to the covalent immobilization of an enzyme to agarose based supports. Heparinase was used as a model enzyme in this set of studies. Amine modifications result in 75-100% activity loss, but the effect is moderated by a reduction in the degree of derivatization. N-hydroxysuccinimide, 1,1,1-trifluoroethanesulfonic acid, and epoxide activated agarose were utilized to determine the effect of amine reactive supports on immobilized enzyme activity retention. Cysteine modifications resulted in 25-50% loss in activity, but free cysteines were inaccessible to either immobilized bromoacetyl or p-chloromercuribenzoyl groups. Amine reactive coupling chemistries were therefore utilized for the covalent immobilization of heparinase. Second, to ensure maximal stability of the immobile protein-support linkage, the identification and subsequent elimination of the principal sources of protein detachment were systematically investigated. By using high-performance liquid chromatography (HPLC), electrophoresis, and radiolabeling techniques, the relative contributions of four potential detachment mechanisms-support degradation, proteolytic degradation, desorption of noncovalently bound protein, and bond solvolysis-were quantified. The mechanisms of lysozyme, bovine serum albumin, and heparinase leakage from N-hydroxysuccinimide or 1,1,1-trifluoroethanesulfonic acid activated agarose were elucidated. By use of stringent postimmobilization support wash procedures, noncovalently bound protein loss. An effective postimmobilization washing procedure is presented for the removal of adsorbed protein and the complete elimination of immobilized protein loss.     

Immobilized enzyme cellulose hollow fibers: II. Kinetic analysis

Immobilized enzyme hollow fibers may be useful in the purification or treatment of whole blood under clinical conditions. In this study, catalytically pure heparinase was immobilized to cellulose to analyze the feasibility for the removal of heparin's anticoagulant activity from whole blood. The kinetics of catalytically pure heparinase immobilized to regenerated cellulose hollow fibers were quantified with respect to mass transfer coefficient and enzyme loading. The kinetic analysis showed that increases in the mass transfer coefficient of heparin in the fiber lumen decreased the apparent Michaelis constant while increases in enzyme activity immobilized to the fiber lumen increased the apparent Michaelis constant. The apparent Michaelis constant was an order of magnitude greater than the intrinsic K(m) value for the system. The intrinsic K(m) value for heparinase-cellulose is 0.4 +/- 0.3 mg/mL (N = 6) and it is the same order of magnitude as the K(m) value for soluble heparinase.     

Regional heparinization via simultaneous separation and reaction in a novel Taylor-Couette flow device.

The development of a safe and efficient bioreactor design has remained a challenge for the clinical application of immobilized enzymes. Specifically, the use of immobilized heparinase I has been the target of many studies to make heparin anticoagulation therapy safer for the critically ill patient with kidney failure or heart disease. We have investigated the use of Taylor-Couette flow for a novel type of bioreactor. In a previous study, we showed that the fluidization of agarose immobilized heparinase within Taylor vortices in whole blood can lead to extensive blood damage in the form of cell depletion and hemolysis. Based on these findings, we designed and developed a reactor, referred to as vortex-flow plasmapheretic reactor (VFPR), that incorporated plasmapheresis and fluidization of the agarose in the reactive compartment, separate from the whole-blood path. In the present study, immobilized heparinase I was tested as a means of achieving regional heparinization of a closed circuit. This is a method in which heparin is infused into the extracorporeal circuit predialyzer and neutralized postdialyzer. Saline studies were performed with an immobilized heparinase I-packed bed and with the VFPR. An in vitro feasibility study was performed with the VFPR using human blood. The VFPR achieved heparin conversions of 44 +/- 0.5% and 34 +/- 2% in saline and blood, respectively. In addition, the VFPR caused no blood damage. We report a novel method to achieve fluidization which depended on secondary, circumferencial flow, and was independent of the primary flow through the device.     

Escherichia coli dnaB protein. Affinity chromatography on immobilized nucleotides.

The purification of the Escherichia coli dnaB protein by affinity chromatography on nucleotides bound to agarose is described. The dnaB protein, which contains an associated ribonucleoside triphosphatase activity (Wickner, S., Wright, M., and Hurwitz, J. (1974) Proc. Natl. Acad. Sci. U. S. A. 71, 783-787) binds to immobilized ATP, ADP, and UDP, but not to AMP. The type of linkage of ATP to agarose influences the adsorption, elution, and purification of the enzyme. Optimal purification is achieved using ATP bound to agarose via its oxidized ribose moiety. By this means, the dnaB protein can be obtained at least 95% electrophoretically pure after only three purification steps. The enzyme can be eluted from immobilized nucleoside-5'-di- and -triphosphates by ATP, ADP, and pyrophosphate, but not by AMP or orthophosphate. ADP and pyrophosphate, as well as the substrate ATP in high concentration are at the same time inhibitors of the ribonucleoside triphosphatase. The dnaB complementing and ribonucleoside triphosphatase activities could not be separated from each other by affinity chromatography, supporting the finding of others that they both reside on the same protein complex, namely a dnaB multimer. The results indicate that the dnaB protein binds to immobilized nucleotides by means of its ribonucleoside triphosphatase, and that at least the pyrophosphate moiety is essential for adsorption as well as elution of the enzyme.     

Evaluation of CNBr, FMP and hydrazide resins for immunoaffinity purification of factor IX.

The American Red Cross has developed an immunoaffinity chromatography method to purify human coagulation Factor IX to high levels of purity for therapeutic treatment of hemophilia B. The resin currently used in this process is Sepharose CL2B, a cross-linked 2% agarose, which is activated with cyanogen bromide to immobilize an anti-Factor IX monoclonal antibody. This study evaluated two alternative resins and coupling chemistries, a synthetic polymer bead activated by 2-fluoro-1-methyl-pyridinium toluene 4-sulfonate (FMP) and a cross-linked 2% agarose bead with free hydrazide groups for site-specific coupling. The cyanogen bromide and FMP chemistries immobilize the monoclonal antibody in a random orientation. In hydrazide coupling, the monoclonal antibody is immobilized by the non-antigen-binding part of the molecule which, theoretically, should increase the amount of immobilized monoclonal antibody able to bind antigen. To examine this, the capacity of the resins to bind Factor IX and the purity and recovery of Factor IX eluted from the resins were measured. The FMP-activated resin exhibited the lowest capacity, binding only 2% of the Factor IX feed. Sepharose CL2B bound 87% of the loaded protein, while the hydrazide resin bound 43%. These results suggest that (a) hydrazide activation may be insufficient to orient monoclonal antibody and (b) other factors such as steric hindrances and diffusional resistances during immobilization may be important. Neither of the other resins tested demonstrated improved performance compared with cyanogen bromide-activated Sepharose CL2B for the immunoaffinity purification of Factor IX.     

Immobilized enzyme cellulose hollow fibers: I. Immobilization of heparinase

The immobilization of heparinase to tresyl-chloride-activated cellulose hollow fibers for the removal of heparin from the bloodstream was examined. Whole blood can be circulated through cellulose hollow fibers without hemolysis and the tresyl chloride chemistry provides a strong linkage which limits the release of the enzyme from the support. The tresylation and immobilization methods were modified and optimized to improve the heparinase activity retained by cellulose. Pretreatment of the hollow fibers with 0.05/V sodium hydroxide increased the degree of tresylation and the immobilization yield by a factor of five. The use of triethylamine as the organic base in the tresyl chloride activation resulted in threefold greater activity retention by the support than when pyridine was used. Together, sodium hydroxide pretreatment and triethylamine enhanced the activity retained by cellulose to 26.2 +/- 7.0% of that bound to the support. The activity retention was also a function of the technique used for immobilization. The best results were achieved when the enzyme was applied to the activated fibers once every 12 to 24 h for a total of four times. The active enzyme loading on the fibers was 0.3 mg heparin degraded/h cm(2) when 4.5 mug protein/cm(2) was bound to the fibers.     

Characterization of type XI collagen-glycosaminoglycan interactions

Using competitive binding experiments, it was found that native type XI collagen binds heparin, heparan sulfate, and dermatan sulfate. However, interactions were not evident with hyaluronic acid, keratan sulfate, or chondroitin sulfate chains over the concentration range studied. Chondrocyte-matrix interactions were investigated using cell attachment to solid phase type XI collagen. Pretreatment of chondrocytes with either heparin or heparinase significantly reduced attachment to type XI collagen. Incubation of denatured and cyanogen bromide-cleaved type XI collagen with radiolabeled heparin identified sites of interaction on the alpha1(XI) and alpha2(XI) chains. NH(2)-terminal sequence data confirmed that the predominant heparin-binding peptide contained the sequence GKPGPRGQRGPTGPRGSRGAR from the alpha1(XI) chain. Using rotary shadowing electron microscopy of native type XI collagen molecules and heparin-bovine serum albumin conjugate, an additional binding site was identified at one end of the triple helical region of the collagen molecule. This coincides with consensus heparin binding motifs present at the amino-terminal ends of both the alpha1(XI) and the alpha2(XI) chains. The contribution of glycosaminoglycan-type XI collagen interactions to cartilage matrix stabilization is discussed.     

Detoxification of organophosphate pesticides using an immobilized phosphotriesterase from Pseudomonas diminuta

A purified phosphotriesterase was successfully immobilized onto trityl agarose in a fixed bed reactor. A total of up to 9200 units of enzyme activity was immobilized onto 2.0 mL of trityl agarose (65 mumol trityl groups/mL agarose), where one unit is the amount of enzyme required to catalyze the hydrolysis of one micromole of paraoxon in one min. The immobilized enzyme was shown to behave chemically and kinetically similar to the free enzyme when paraoxon was utilized as a substrate. Several organophosphate pesticides, methyl parathion, ethyl parathion, diazinon, and coumaphos were also hydrolyzed by the immobilized phosphotriesterase. However, all substrates exhibited an affinity for the trityl agarose matrix. For increased solubility and reduction in the affinity of these pesticides for the trityl agarose matrix, methanol/water mixtures were utilized. The effect of methanol was not deleterious when concentrations of less than 20% were present. However, higher concentrations resulted in elution of enzyme from the reactor. With a 10-unit reactor, a 1.0 mM paraoxon solution was hydrolyzed completely at a flow rate of 45 mL/h. Kinetic parameters were measured with a 0.1-unit reactor with paraoxon as a substrate at a flow rate of 22 mL/h. The apparent K(m) for the immobilized enzyme was 3-4 times greater than the K(m) (0.1 mM) for the soluble enzyme. Immobilization limited the maximum rate of substrate hydrolysis to 40% of the value observed for the soluble enzyme. The pH-rate profiles of the soluble and immobilized enzymes were very similar. The immobilization of phosphotriesterase onto trityl agarose provides an effective method esterase onto trityl agarose provides an effective method for hydrolyzing and thus detoxifyuing organophosphate pesticides and mammalian acetylcholinesterase inhinbitors.     

Immobilization of penicillin G acylase on aminoalkylated polyacrylic supports

A procedure is described for the immobilization of penicillin G acylase (PA) on Amberlite XAD7 modified by transamidation with 1,2-ethylenediamine and activated with glutaraldehyde. Reduction with sodium borohydride of the Schiff's bases formed between the amino groups of the protein and glutaraldehyde results in a dramatic improvement of the operational stability of the immobilized enzyme without affecting the catalytic activity. The enzyme kept in presence of the substrate, penicillin G, displays an increased stability with respect to that stored in pure phosphate buffer solution. The inactivation kinetics of the immobilized preparations of PA, determined in a continuous fixed bed reactor, as well as a discontinuous batch reactor, are reported.     

Preparation and characterization of an enzymatically active immobilized derivative of myosin.

Purified skeletal muscle myosin (EC 3.6.1.3) has been covalently bound to Sepharose 4B by the cyanogen bromide procedure. The resulting complex, Sepharose-Myosin, possesses adenosine triphosphatase activity and is relatively stable for long periods of time. Under optimal binding conditions, approximately 33% of the specific ATPase activity of the bound myosin is retained. Polyacrylamide gel electrophoresis of polypeptides released from denatured Sepharose-Myosin indicates that 85% of the myosin is attached to the agarose beads through the heavy chains and the remainder through the light chains, in agreement with predictions of binding and release based upon either the lysine contents or molecular weights of themyosin subunits. The adenosine triphosphatase of the immobilized myosin has been investigated under conditions of varying pH, ionic strength, and cation concentration. The ATPase profiles of immobilized myosin are quite similar to those for free myosin, however subtle differences are found. The Sepharose-Myosin ATPase is not as sensitive as myosin to alterations in salt concentration and the apparent KM is approximately two-fold higher than that of myosin. These differences are probably due to chemical modification in the region of the attachment site(s) to the agarose beads and hydration and diffusion limitations imposed by the polymeric agarose matrix.     

Immobilization of heparin with trichloro-s-triazine: Purification of mouse antithrombin

A method for covalently attaching heparin to agarose in which the bound heparin is resistant to detachment is described. A controlled amount of 4-aminophenethylamine is linked to trichloro-s-triazine-activated agarose via the primary amino group. The resulting immobilized arylamine is then diazo coupled to an appropriately modified heparin. The immobilized heparin was used in a two-step procedure for the purification of mouse antithrombin. Some of the properties of the mouse antithrombin are described.     

Enzyme immobilization on tritylagarose

A method is described for the immobilization on tritylated agarose or Sepharose columns of a wide spectrum of enzymes, including types useful in contemporary biochemistry/molecular biology, many of which have never before been reported as immobilized. The method involves the formation of noncovalent hydrophobic bonds between the enzymes and trityl groups which are attached to the agarose by means of ether bonds. The immobilization of calf intestinal and E. coli alkaline phosphatases to tritylagarose is reported in detail. Their binding strength, binding capacity, and long-term stability (greater than six months) are described as a function of the salt concentration, pH, buffer type, and degree of agarose substitution. Homologies are noted between tritylagarose-bound and membrane-bound phosphatases. This method compares favorably with other methods, covalent or otherwise, reported to date, in terms of the enzyme immobilization yield (ca. 100%), the mildness of conditions, resulting, in most cases, in the retention of a high degree of activity, the ease and speed of the manipulations, and the long-term stability of the immobilized enzyme. Further, it is noted that highly tritylated and crosslinked Sephadex G10 selectively and mildly removes detergents from enzyme solutions.     

Immobilization of recombinant heparinase I fused to cellulose-binding domain.

Immobilization of biologically active proteins is of great importance to research and industry. Cellulose is an attractive matrix and cellulose-binding domain (CBD) an excellent affinity tag protein for the purification and immobilization of many of these proteins. We constructed two vectors to enable the cloning and expression of proteins fused to the N- or C-terminus of CBD. Their usefulness was demonstrated by fusing the heparin-degrading protein heparinase I to CBD (CBD-HepI and HepI-CBD). The fusion proteins were over-expressed in Escherichia coli under the control of a T7 promoter and found to accumulate in inclusion bodies. The inclusion bodies were recovered by centrifugation, the proteins were refolded and recovered on a cellulose column. The bifunctional fusion protein retained its abilities to bind to cellulose and degrade heparin. C-terminal fusion of heparinase I to CBD was somewhat superior to N-terminal fusion: Although specific activities in solution were comparable, the latter exhibited impaired binding capacity to cellulose. CBD-HepI-cellulose bioreactor was operated continuously and degraded heparin for over 40 h without any significant loss of activity. By varying the flow rate, the mean molecular weight of the heparin oligosaccharide produced could be controlled. The molecular weight distribution profiles, obtained from heparin depolymerization by free heparinase I, free CBD-HepI, and cellulose-immobilized CBD-HepI, were compared. The profiles obtained by free heparinase I and CBD-HepI were indistinguishable, however, immobilized CBD-HepI produced much lower molecular weight fragments at the same percentage of depolymerization. Thus, CBD can be used for the efficient production of bioreactors, combining purification and immobilization into essentially a single step.     

Effect of calcium on immobilization of rice (<Emphasis Type="Italic">Oryza sativa</Emphasis> L.) peroxidase for bioassays in sodium alginate and Agarose gel

The direct immobilization of soluble peroxidase isolated and partially purified from shoots of rice seedlings in calcium alginate beads and in calcium agarose gel was carried out. Peroxidase was assayed for guaiacol oxidation products in presence of hydrogen peroxide. The maximum specific activity and immobilization yield of the calcium agarose immobilized peroxidase reached 2,200 U mg⁻¹ protein (540 mU cm⁻³ gel) and 82%, respectively. In calcium alginate the maximum activity of peroxidase upon immobilization was 210 mU g⁻¹ bead with 46% yield. The optimal pH for agarose immobilized peroxidase was 7.0 which differed from the pH 6.0 for soluble peroxidase. The optimum temperature for the agarose immobilized peroxidase however was 30°C, which was similar to that of soluble peroxidase. The thermal stability of calcium agarose immobilized peroxidase significantly enhanced over a temperature range of 30∼60°C upon immobilization. The operational stability of peroxidase was examined with repeated hydrogen peroxide oxidation at varying time intervals. Based on 50% conversion of hydrogen peroxide and four times reuse of immobilized gel, the specific degradation of guaiacol for the agarose immobilized peroxidase increased three folds compared to that of soluble peroxidase. Nearly 165% increase in the enzyme protein binding to agarose in presence of calcium was noted. The results suggest that the presence of calcium, ions help in the immobilization process of peroxidase from rice shoots and mediates the direct binding of the enzyme to the agarose gel and that agarose seems to be a better immobilization matrix for peroxidase compared to sodium alginate.     

Hexasaccharides from the histamine-modified depolymerization of porcine intestinal mucosal heparin

Specific sequences in heparin are responsible for its modulation of the biological activity of proteins. As part of a program to characterize heparin-peptide and heparin-protein binding, we are studying the interaction of chemically discrete heparin-derived oligosaccharides with peptides and proteins. We report here the isolation and characterization, by one- and two-dimensional 1H NMR spectroscopies, of ten hexasaccharides, one pentasaccharide, and one octasaccharide serine that were isolated from depolymerized porcine intestinal mucosal heparin. Hexasaccharides were chosen for study because they fall within the size range, typically tetra- to decasaccharide in length, of heparin sequences that modulate the activity of proteins. The depolymerization reaction was catalyzed by heparinase I (EC 4.2.2.7) in the presence of histamine, which binds site specifically to heparin. Histamine increases both the rate and extent of heparinase I-catalyzed depolymerization of heparin. It is proposed that oligosaccharides produced by heparinase I-catalyzed depolymerization can inhibit the enzyme by binding to the imidazolium group of histidine-203, which together with cysteine-135 forms the catalytic domain of heparinase I. The increased rate and extent of depolymerization are attributed to competitive binding of the oligosaccharides by histamine.     

Stabilization of a highly active but unstable alcohol dehydrogenase from yeast using immobilization and post-immobilization techniques

The alcohol dehydrogenase (ADH) from Baker's yeast is very active but extremely unstable under several different conditions. Mild immobilization methods such as one-point attachment to agarose activated with cyanogen bromide groups or ionic adsorption to agarose activated with charged groups allow high activity recoveries (80–100%) but do not promote protein stabilization. In contrast, immobilization methods that force the enzyme to be covalently attached at multiple points on the support fully inactivate the enzyme. Herein, we propose an interesting solution to address the dichotomy between activity and stability. We have developed a protocol in which the enzyme is immobilized on agarose activated with glyoxyl groups in the presence of acetyl cysteine, which results in the recovery of 25% of the enzyme activity but increases the thermal stability of the soluble enzyme 50-fold. However, this immobilization technique does not stabilize the enzyme quaternary structure. Hence, a post-immobilization technique using functionalized polymers has been used to cross-link all enzyme subunits. In this method, polycationic polymers (polyethylenimine) cross-link the quaternary structure with a negligible effect on catalytic activity, which results in a derivative that is 5-fold more stable than non-cross-linked derivatives under very dilute and acidic conditions that highly favor subunit dissociation. Therefore, the stability was increased 500-fold for this optimal derivative compared to diluted soluble enzyme, although the relative expressed activity was low (25%). However, the low expressed activity may be overcome by designing immobilized biocatalysts with high volumetric activities.     

Three-dimensional spatial patterning of proteins in hydrogels

The ability to create three-dimensional biochemical environments that mimic those in vivo is valuable for the elucidation of fundamental biological phenomena and pathways. To this end, we designed a system in which proteins can be photochemically patterned in three dimensions within hydrogels under physiological conditions. Fibroblast growth factor-2 (FGF2) was immobilized within agarose hydrogels that were modified with two-photon labile 6-bromo-7-hydroxycoumarin-protected thiols. Two different methods were developed for FGF2 immobilization. The first procedure relies on the protein containing free cysteines for the formation of disulfide bonds with photoexposed agarose thiols. The second procedure takes advantage of the femtomolar binding partners, human serum albumin (HSA) and albumin binding domain (ABD), which have K(D) values of ~10(-14) M. Here HSA-maleimide was chemically bound to photoexposed agarose thiols, and then the FGF2-ABD fusion protein was added to form a stable complex with the immobilized HSA. The use of orthogonal, physical binding pairs allows protein immobilization under mild conditions and can be broadly applied to any protein expressed as an ABD fusion.