Sumner’s legacy: A century of urease crystals and recent structural advances

Abstract

In 1926, James B. Sumner crystallized jack bean urease—the first enzyme to be obtained in crystalline form—thus demonstrating that enzymes are proteinaceous. To honor the 100-year anniversary of that momentous event, this review highlights critical findings leading up to Sumner’s efforts, explains the significance of his results, and describes subsequent experimental findings related to urease. For example, nearly five decades after crystals became available Burt Zerner and colleagues identified urease as the first known nickel-containing enzyme. The surprising discovery of nickel in urease raised questions about the structure of the metal-containing active site, the enzyme mechanism, and pathway by which the catalytic center is synthesized – each of which is addressed here. Finally, I reflect on remaining open questions related to this remarkable enzyme and potential experimental directions that could be employed to provide corresponding insights.

Introduction

The enzyme urease is present in plants, selected fungi and algae, and a wide assortment of bacteria where it catalyzes the hydrolysis of urea with important agricultural and medical consequences (Mobley and Hausinger 1989; Ligabue-Braun and Carlini 2024; Khan et al. 2025). Urease is essential for the utilization of urea-based fertilizers; however, the prevalence of this enzyme in soil can lead to the loss of nitrogen through volatilization of ammonia and to phytotoxicity due to excessive increases in pH that accompany urea degradation. These deleterious consequences have stimulated interest in the development of urease inhibitors for agricultural purposes (Modolo et al. 2018). Inhibitors of urease also are of interest to combat the pathogenic aspects of the enzyme. For example, urease activity plays an important role in the ability of Helicobacter pylori to colonize the stomach where it can lead to peptic ulcer disease and gastric cancer (Sousa et al. 2023). Similarly, microbial ureases are implicated in the development of a significant proportion of urinary stones, catheter encrustations, pyelonephritis, and other medical dysfunctions (Mobley et al. 1995). The continuing quest for urease inhibitors (Deutch 2024; Hausinger 2024; Babaei et al. 2025) is facilitated by detailed structural and mechanistic characterization of the enzyme as reviewed below. Another intriguing aspect of urease that is not further described here is its toxicological properties that are separate from its catalytic activity (Carlini and Ligabue-Braun 2016). Because urease is an important enzyme in agriculture and medicine, investigations of its structure, catalytic mechanism, and assembly pathway merit careful assessment.

This critical review covers several topics related to urease. It first briefly summarizes the historical aspects of urea and urease leading to the crystallization of the enzyme by J. B. Sumner in 1926 (Sumner 1926). The significance of Sumner’s work to establish the makeup of enzymes is then detailed, followed by a description of key experimental findings related to urease during the ensuing 100 years. Remarkably, Burt Zerner’s group identified urease as the first known nickel-containing enzyme in 1975 (Dixon, Gazzola, Blakeley, et al. 1975), almost five decades after crystals were available. Subsequent topics in this compilation of research explore the tremendous advances in defining urease structures, highlight the properties of the urease metal-containing active site (i.e. the metallocenter) and the enzyme mechanism, and illustrate structural aspects of the multi-component urease biosynthetic pathway. Finally, this contribution reflects on the remaining open questions related to this remarkable enzyme.

Urease related history prior to Sumner’s contributions

Urease and its substrate, urea, had a notable history well before James B. Sumner initiated his crystallization trials (Figure 1). In ~1727, the Dutch botanist Herman Boerhaave first isolated urea from urine, although the French chemist Hilaire Marin Rouelle is often credited with the discovery of this substance in 1773 (Kurzer and Sanderson 1956). During their studies of crystalline urea in 1799, the French scientists Antoine Francis comte de Fourcroy and Louis Vauquelin coined its name (Fourcroy and Vauguelin 1799). In 1828, the German chemist Friedrich Wöhler synthesized urea, the first organic molecule ever prepared from inorganic materials (Wöhler 1828). During 1860, Louis Pasteur of France reported that a living organism is responsible for the alkaline fermentation that transforms urea in urine into carbonate and ammonia (Pasteur 1860). Four years later, the French botanist Philippe van Tieghem isolated a bacterium, later called Micrococcus ureae, that catalyzes this decomposition reaction (van Tieghem 1864). In 1890, Pierre Miquel in Paris described several additional “ureolytic” microorganisms and termed the transforming agent “urease” (Miquel 1890). The physical component conferring urease activity remained unknown, but in 1909 the Japanese scientists T. Takeuchi and R. Inone provided a key to address this question by their discovery that soybean, among other plants, possesses large amounts of urease activity (Takeuchi and Inone 1909). In 1916, John Mateer and Eli Marshall, Jr. from Johns Hopkins University identified jack bean as containing 16-fold greater levels of urease than soybean (Mateer and Marshall 1916). These findings set the stage for Sumner’s seminal efforts to identify the jack bean component associated with urease activity.

Figure 1.

Timeline of notable events related to urea and urease. The figure depicts a series of firsts in the history of urease and its substrate.

What is an enzyme? The contributions of James Batcheller Sumner

At the beginning of the 20^th century, the chemical nature of an enzyme was actively disputed. Richard Willstätter, a notable German chemist who synthesized many natural products and was the winner of the 1915 Nobel prize in chemistry for his work on chlorophyll and other plant pigments (Trauner 2015), was a proponent of the “Träger” or carrier theory in which an inert colloidal substance possessed unidentified small molecules that conferred reactivity (Willstätter 1922). Willstätter’s evidence arguing against proteins being responsible for enzyme activity included his group’s purification of peroxidase to such a high level of purity that the existing tests failed to detect any protein in the sample (Willstätter and Weber 1926). They also failed to isolate a protein when examining the esterase called chlorophyllase that catalyzes the first step of chlorophyll degradation (Willstätter and Stoll 1913; Trauner 2015).

James B. Sumner, a Cornell University biochemist, joined the discussion in 1926 when he reported the successful isolation and crystallization of urease as a protein from jack bean meal (Sumner 1926). Sumner later described the crystallization process as being “absurdly simple”, albeit after nine years of preliminary efforts, and involved stirring 100 g of jack bean meal into 0.5 L of 32% acetone, filtering at 2-2.5 °C, standing overnight, and centrifuging to collect the crystals shown in Figure 2 (Sumner 1926; Sumner and Somers 1953). His claim of isolating a crystalline protein as the enzyme initially was disputed by Willstätter, his students, and many other biochemists (Sumner 1937), even after John H. Northrup had succeeded in generating protein crystals of pepsin and trypsin in 1930 and 1932 (Northrup 1930; Northrup and Kunitz 1932). For example, Willstätter credited that work as “brilliant discoveries”, but went on to write “it remained doubtful whether the protein content of these enzymes really forms an indispensable constituent” and whether indeed “the proteolytic activity is a property of the protein molecule itself” (Willstätter 1933). Refuting such comments, Sumner’s ground-breaking paper on urease had clearly enumerated reasons why his protein crystals corresponded to the urease enzyme: microscopic examination of the crystals revealed no other contaminants, the dissolved crystals exhibited high urease activity, solvents incapable of dissolving the crystals also failed to extract any urease activity, other jack bean globulins purified from the starting material possessed little urease activity, the protein crystals had a unique octahedral appearance and were denatured by adding acid, and protein alone was detected in the urease crystals with no iron, manganese, or phosphorous being present (Sumner 1926). This milestone represents the first crystallization of any enzyme and played a central role in demonstrating the proteinaceous nature of enzymes. The 1946 Nobel prize in Chemistry was awarded to Sumner “for his discovery that enzymes can be crystallized” and to Northrup and Wendell Stanley “for their preparation of enzymes and virus proteins in a pure form”.

Figure 2.

James B. Sumner and his crystals of jack bean urease.

Urease is more than just protein. Contributions of Burt Zerner

We now recognize that nearly all enzymes are comprised of protein [exceptions are catalytically active RNA molecules, known as ribozymes, that were discovered in 1982 (Kruger et al. 1982)], but we also know that approximately half of these catalysts require for their activity an organic or inorganic cofactor (Richter 2013; Hausinger 2019). Indeed, five decades ago (i.e., 49 years after jack bean urease was crystallized) Burt Zerner and colleagues showed that this enzyme possesses tightly bound nickel ions (Dixon, Gazzola, Blakeley, et al. 1975). The discovery by the Australian research group was motivated by their observation of a weak absorption spectrum in concentrated jack bean urease solutions. A later study detailed the electronic spectrum after correction for light scattering and noted four features (λ_max of 407, 745, 910, and 1060 nm with the latter three transitions having ε_max of 46, 14, and 10 M⁻¹·cm⁻¹, respectively) associated with the bound metal (Blakeley et al. 1983). These authors also characterized the more intense thiolate → Ni(II) charge-transfer and other electronic transitions (λ_max of 324, 380, and 420 nm with ε_max of 1270, 580, and 460 M⁻¹·cm⁻¹) that were generated upon addition of β-mercaptoethanol (Blakeley et al. 1983). The metal ion content was initially reported as 2.0 g-atom of nickel per 105,000 g of protein (Dixon, Gazzola, Blakeley, et al. 1975), with subsequent investigations providing the slightly revised nickel content of 2.0 Ni/96,600 g of protein (Dixon, Blakeley, et al. 1980a).

Zerner and colleagues also investigated several other features of jack bean urease. For example, they demonstrated the enzyme has a subunit size of ~95 kDa and a native size of ~590 kDa, consistent with a hexameric structure (Dixon, Hinds, et al. 1980). They established that acetamide and methylurea (Dixon, Riddles, et al. 1980) along with semicarbazide (Gazzola et al. 1973) are poor substrates of the enzyme, joining the previously reported examples of N-hydroxyurea, N,N-dihydroxyurea, and formamide (Fishbein et al. 1965; Fishbein 1969, 1977). In addition, they characterized the inhibition of jack bean urease by hydroxamic acids, β-mercaptoethanol, fluoride, phosphoramidate, and selected other amides and esters of phosphoric acid (Blakeley et al. 1969; Dixon, Gazzola, Watters, et al. 1975; Dixon, Blakeley, et al. 1980b; Dixon, Hinds, et al. 1980; Andrews et al. 1986), which in some cases also are poor substrates of the enzyme (Andrews et al. 1986). On the basis of their spectroscopic and kinetic results, Zerner’s group suggested that the enzyme-bound nickel ions were essential for catalytic activity and they proposed a dual metal ion catalytic mechanism for jack bean urease (Dixon, Riddles, et al. 1980). They hypothesized that the two nickel ions are coordinated at opposite sides of the active site, with the carbonyl group of urea interacting with and being polarized by one nickel ion, whereas hydroxide coordinates to and is activated for use as a nucleophile during catalysis by the second metal ion. Carboxylate and thiol protein side chains were posited at the active site along with an unidentified general base.

Structural investigations have subsequently uncovered the protein’s quaternary and tertiary architecture as well as the metal environment at the active site. As will be seen below, a few aspects of the Zerner model for the active site are supported by the enzyme crystal structures, whereas other features of the proposal are incorrect.

Crystal structures of urease

Although jack bean urease was crystallized in 1926, the first diffraction studies of the enzyme were not reported until 1992 using both jack bean and Klebsiella aerogenes sources (Jabri et al. 1992). The plant urease diffracted to near 3.5 Å resolution, whereas the bacterial enzyme diffracted to better than 2.0 Å resolution, thus highlighting the importance of the bacterial enzyme for further structural studies. Indeed, the first urease structure reported was obtained using the bacterial protein from K. aerogenes (Table 1) (Jabri et al. 1995). This enzyme possesses a (UreA·UreB·UreC)₃ configuration with subunits of 100, 106, and 567 amino acid residues (Figure 3A), like the situation for most bacterial ureases. The enzyme structure revealed the presence of three dinuclear nickel sites with a carbamylated lysyl residue (i.e., in this case Lys217 that has reacted with carbon dioxide) and a solvent molecule bridging the metals, Ni-1 coordinated by two imidazole residues (His246 and His272), Ni-2 bound by two imidazoles and an aspartyl residue (His134, His136, and Asp360), and a singly coordinated solvent molecule present on each metal. Two non-coordinated histidyl residues also are present at the active site; His219 was well-positioned to stabilize urea binding to Ni-1 whereas His320 was suggested to function in catalysis. His320 is located on a highly mobile helix-loop-helix protein flap (residues 308-336) that covers the active site. Conformational restrictions of this lid region due to the crystal lattice were suggested to account for the near lack of urease activity within the crystals (Moncrief et al. 1995). This metallocenter architecture (Figure 4A) and the pair of nearby histidyl residues are common to all ureases.

Table 1.

Urease structures

Source	Protein	Samples,a (PDB ID), resolution	Reference
Klebsiella aerogenes	(UreABC)3	WT (1FWJ), 2.20 Å; WT (2KAU), 2.00 Å	(Jabri et al. 1995)
		Apoprotein (1KRA), 2.30 Å; H219A (1KRB), 2.50 Å; H320A (1KRC), 2.50 Å	(Jabri and Karplus 1996)
		WT at 298 K (1EJW), 1.90 Å; WT at 100 K (1EJX), 1.60 Å	unpublished
		Mononuclear Ni H134A (1FWI), 2.00 Å	(Park et al. 1996)
		C319A at pH 7.5 (1FWA), 2.00 Å; C319A at pH 6.5 (1FWB), 2.00 Å; C319A at pH 8.5 (1FWC), 2.00 Å; C319A at pH 9.4 (1FWD), 2.00 Å; C319A + AHA (1FWE), 2.00 Å; C319D (1FWF), 2.00 Å; C319S (1FWG), 2.00 Å; C319Y (1FWH), 2.00 Å	(Pearson et al. 1997)
		K217E apoprotein (1A5K), 2.20 Å; K217C/C319A apoprotein (1A5L), 2.20 Å; K217A apoprotein (1A5M), 2.00 Å; K217A + formate (1A5N), 2.40 Å; K217C/C319A + formate (1A5O), 2.50 Å	(Pearson et al. 1998)
		Dinuclear Mn (1EF2), 2.50 Å	(Yamaguchi et al. 1999)
		D221A (1EJR), 2.00 Å; H219N (1EJS), 2.00 Å; H219Q (1EJT), 2.00 Å; H320N (1EJU), 2.00 Å; H320Q (1EJV), 2.40 Å	(Pearson et al. 2000)
		WT examined for radiation damage: 100 K at start (4EP8), 1.55 Å; 100 K at end (4EPB), 1.75 Å; 300 K at start (4EPD), 1.70 Å; 300 K at end (4EPE), 2.05 Å	(Warkentin et al. 2012)
Sporosarcina pasteurii	(UreABC)3	WT + β-ME (1UBP), 1.65 Å	(Benini S. et al. 1998)
		WT (2UBP), 2.00 Å; WT + PPD (3UBP), 2.00 Å	(Benini S et al. 1999)
		WT + AHA (4UBP), 1.55 Å	(Benini S. et al. 2000)
		WT + phosphate (1IE7), 1.85 Å	(Benini S. et al. 2001)
		WT + borate (1S3T), 2.10 Å	(Benini S. et al. 2004)
		WT + citrate (4AC7), 1.50 Å	(Benini S. et al. 2013)
		WT (4CEU), 1.58 Å; WT + F− (4CEX), 1.59 Å	(Benini S. et al. 2014)
		WT + sulfite (5A6T), 1.65 Å	(Mazzei, Cianci, Benini, et al. 2016)
		WT + BQ (5FSE), 2.07 Å; WT + BQ + sulfite (5FSD), 1.75 Å	(Mazzei, Cianci, Musiani, et al. 2016)
		WT + catechol (5G4H), 1.50 Å	(Mazzei, Cianci, Musiani, et al. 2017)
		WT + NBPT (5OL4), 1.28 Å	(Mazzei, Cianci, Contaldo, et al. 2017)
		WT + AgNO3 (6G48), 1.91 Å	(Mazzei et al. 2018)
		WT + F− + urea (6QDY), 1.42 Å	(Mazzei, Cianci, Benini, et al. 2019b)
		WT + Au(MPB)Cl2 (6I9Y), 2.14 Å	(Mazzei, Wenzel, et al. 2019)
		WT + NBPTO at pH 7 (6H8J), 1.45 Å	(Mazzei, Cianci, Contaldo, et al. 2019)
		WT + NBPTO at pH 6.5 (6RP1), 1.49 Å; WT + NBPTO at pH 7.5 (6RKG), 1.32 Å	(Mazzei, Cianci, Benini, et al. 2019a)
		WT + 3MC (6ZNY), 1.50 Å; WT + 4MC (6ZNZ), 1.89 Å; WT + 3,4DMC (6ZO0), 2.23 Å; WT + 3,5DMC (6ZO1), 1.61 Å; WT + 4,5DMC (6ZO2), 1.65 Å; WT + 3,6DMC (6ZO3), 1.55 Å	(Mazzei, Contaldo, et al. 2021)
		WT + Ag(PEt3)Cl (7B58), 1.72 Å; WT + Ag(PEt3)Br (7B59), 1.63 Å; WT + Ag(PEt3)2NO3 (7B5A), 1.97 Å	(Mazzei, Cirri, et al. 2021)
		WT + Au(PEt3)I (7P7N), 1.80 Å; WT + [Au(PEt3)2]+ (7P7O), 1.87 Å	(Mazzei, Massai, et al. 2021)
		WT + 1,2-benzisoselenazol-3(2H)-one (7ZCY), 1.54 Å	(Macegoniuk et al. 2023)
		WT + HQ (8A18), 1.63 Å	(Mazzei et al. 2022)
		WT + TMTD (8Q2E), 1.68 Å	(Mazzei, Paul, et al. 2024)
		WT (9GML), 3.12 Å; WT + NBPTO (9GNR), 2.92 Å	(Mazzei, Tria, et al. 2024)
Mycobacterium tuberculosis	(UreA)3	UreA, no UreB or UreC (2FVH), 1.80 Å	(Habel et al. 2010)
Brucella melitensis	(UreA)3	UreA, no UreB or UreC (4FUR), 2.10 Å	unpublished
Helicobacter pylori	[(UreAB)3]4	WT (1E9Z), 3.00 Å; WT + AHA (1E9Y), 3.00 Å	(Ha et al. 2001)
		WT + β-ME (6QSU), 2.40 Å; WT + SHA (6ZJA), 2.00 Å	(Cunha et al. 2021)
Helicobacter mustelae	[(UreAB)3]4	Dinuclear Fe (3QGA and 3QGK), 3.00 Å	(Carter et al. 2011)
Yersinia enterocolitica	[(UreABC)3]4	WT (6YL3), 1.98 Å	(Righetto et al. 2020)
		WT (4Z42), 3.01 Å	unpublished
Canavalia ensiformis (jack bean)	[(Ure)3]2	WT + phosphate (3LA4), 2.05 Å	(Balasubramanian and Ponnuraj 2010)
		WT + AHA (4H9M), 1.52Å; WT + phosphate (4GY7), 1.49 Å	unpublished
		WT + F− (4GOA), 2.20 Å	unpublished
		Apparently mononuclear Ni + phosphate (7KNS), 2.77 Å	(Feathers et al. 2021)
Cajanus cajan (pigeon pea)	[(Ure)3]2	WT (4G7E), 2.20 Å	(Balasubramanian et al. 2013)

^aWT indicates samples coordinating a dinuclear Ni metallocenter to the native protein sequence. Alternative designations indicate apoprotein, other changes to the metallocenter, variants, and enzyme with bound inhibitors.

Inhibitors: AHA, acetohydroxamic acid; BQ, 1,4-benzoquinone; 3,4DMC, 3,4-dimethylcatechol; 3,5DMC, 3,4-dimethylcatechol; 4,5DMC, 4,5-dimethylcatechol; 3,6DMC, 3,6-dimethylcatechol; HQ, 1,4-hydroquinone; 3MC, 3-methylcatechol; 4MC, 4-methylcatechol; β-ME, β-mercaptoethanol; MPB, 1-methyl-2-(pyridin-2-yl)-benzimidazole); NBPT, N-(n-butyl)thiophosphoric triamide; NBPTO, N-(n-butyl)phosphoric triamide; PEt₃, triethylphosphine; PPD, phenylphosphorodiamidate; SHA, 2-[1-(3,5-dimethylphenyl)-1H-imidazol-2-yl]sulfanyl-N-hydroxyacetamide; TMTD, tetramethylthiuram disulfide.

Italicized samples were examined by cryo-electron microscopy, whereas all others were studied by x-ray crystallography.

Figure 3.

Urease crystal structures illustrating four distinct quaternary architectures. (A) Structure of a typical (UreA·UreB·UreC)₃ bacterial urease as represented by that of the Klebsiella aerogenes enzyme (PDB ID:1FWJ) shown in cartoon mode. Three UreA subunits are colored red, three UreB subunits are blue, and three UreC subunits are orange, olive, and yellow. (B) The [(UreA·UreB)₃]₄ urease structure from Helicobacter pylori (PDB ID:1E9Z). One (UreA·UreB)₃ unit is shown in cartoon mode with its three small UreA subunits shown in purple (a fusion of the red and blue subunits in A) and its large UreB subunits shown in the same colors as the UreC subunits in A. The other three (UreA·UreB)₃ units are shown in transparent white surface view. (C) Structure of [(UreA·UreB·UreC)₃]₄ urease from Yersinia enterocolitica (PBD ID:6YL3). One (UreA·UreB·UreC)₃ is shown as in A, and the other three units are shown in transparent white surface view. (D) The [(Ure)₃]₂ jack bean urease structure (PDB ID:3LA4). Each plant enzyme subunit equates to a fusion of the three bacterial subunits. One trimer is depicted in cartoon mode using the same color as for the larger subunits in A-C. The second trimer is shown in transparent white surface mode. Two views are shown in all cases, rotated by 90°, and nickel ions are indicated as green spheres.

Figure 4.

The dinuclear nickel site of urease and its interaction with selected inhibitors. (A) The active site structure of all nickel-containing ureases is represented by that of the enzyme from Sporosarcina pasteurii (PDB ID:2UBP). Comparable residue numbers are known for the enzymes from K. aerogenes (H134, H136, K217, H246, H272, and D360), H. pylori (H136, H138, K219, H248, H274, and D362), Y. enterocolitica (H139, H141, K222, H251, H277, and D365), and jack bean (H407, H409, K490, H519, H545, and D633). Two additional histidine residues do not coordinate the nickel ions but lie near the metallocenter and facilitate catalysis (H222 and H323 in S. pasteurii and H219/H320, H221/H322, H224/H325, H492/H593 in ureases from the other sources). Urease metallocenter interactions with (B) β-mercaptoethanol (cyan carbon atoms) (PDB ID:1UBP) and (C) acetohydroxamic acid (yellow carbon atoms) (PDB ID:4UBP). The result of incubation of enzyme with (D) phenylphosphorodiamidate (or N-(n-butyl)-phosphoric triamide) yielding diamidophosphoric acid (orange carbon atom) bound to the metallocenter [PDB ID:3UBP (or 6H8J)] and (E) N-(n-butyl)thiophosphoric triamide yielding monoamidothiophosphoric acid (orange carbon atom) bound to the active site (PDB ID:5OL4). Urease metallocenter interactions with (F) fluoride (PDB ID:4CEX) and (G) fluoride plus urea (green carbon atom) (PDB ID:6QDY). All coordinating ligands and inhibitors are depicted in stick view except for fluoride (light blue spheres). The nickel atoms are shown as green spheres with Ni-1 on the upper right and Ni-2 on the lower left, and solvent molecules are red spheres.

Additional structural studies of K. aerogenes urease included the characterization of variants with substitutions of selected metal ligands and nearby residues (Jabri and Karplus 1996; Park et al. 1996; Pearson et al. 1997; Pearson et al. 1998; Pearson et al. 2000). For example, structural and kinetic analyses of His219 and His320 variants support the proposals that the former residue helps polarize the urea carbonyl group and the latter residue serves as a general acid (Pearson et al. 2000), as detailed in the next section on the catalytic mechanism. The H134A variant was shown to be inactive and bound a mononuclear metallocenter at the Ni-1 site (Park et al. 1996). The carbamylated lysyl ligand was replaced in the K217A and K217C variants, but partial activity was restored for each by incubating with formic acid; the resulting structures provided prominent examples of chemically-rescued proteins (Pearson et al. 1998). The structure of the C319A variant enzyme with bound acetohydroxamic acid, a significant urease inhibitor, was also defined (Pearson et al. 1997). Also noteworthy is the structure of Mn-substituted enzyme that retains partial activity (Yamaguchi et al. 1999). Finally, an analysis of radiation damage to the protein as measured by B-factor increases was consistent with high mobility of the flap covering the protein active site (Warkentin et al. 2012).

An analogous structure was resolved for the three-subunit enzyme from Sporosarcina pasteurii (Benini S et al. 1999). Conspicuously, this enzyme was used to structurally characterize the protein’s interactions with an extensive range of inhibitors (Table 1). S. pasteurii urease with bound β-mercaptoethanol (Figure 4B) revealed how the inhibitor sulfur atom replaces the metal-bridging solvent molecule, consistent with the previously noted thiolate → Ni(II) ligand-to-metal charge-transfer transitions observed in jack bean and K. aerogenes ureases (Blakeley et al. 1983; Todd and Hausinger 1989), while its hydroxyl group binds to Ni-1 (Benini S. et al. 1998). For bound acetohydroxamic acid (Figure 4C), the keto group coordinates Ni-1, hinting at the site of interaction for the urea carbonyl group, whereas the hydroxyl group replaces the metal-bridging solvent molecule (Benini S. et al. 2000), as what was previously noted for the inhibited K. aerogenes enzyme (Pearson et al. 1997). Incubation of S. pasteurii urease with phenylphosphorodiamidate resulted in hydrolysis of the compound with the resulting inhibited protein structure (Figure 4D) possessing diamidophosphoric acid bound to the active site (Benini S et al. 1999). Significantly, this complex is consistent with nucleophilic attack on the substrate by the solvent molecule that bridges the metal ions accompanied by release of phenol. An analogous situation was noted for urease incubated with N-(n-butyl)thiophosphoric triamide, resulting in bound monoamidothiophosphoric acid (Figure 4E) and release of butylamine (Mazzei, Cianci, Contaldo, et al. 2017). Incubation of the enzyme with N-(n-butyl)-phosphoric triamide similarly yielded bound diamidophosphoric acid while releasing butylamine (Mazzei, Cianci, Contaldo, et al. 2019). A pair of fluoride-inhibited S. pasteurii urease structures provide keen mechanistic insights into catalysis. Incubation of the enzyme with fluoride alone leads to substitution of both the bridging solvent and the solvent on Ni-1 by this anion (Figure 4F) (Benini S. et al. 2014). Inclusion of urea with the fluoride-inhibited enzyme yielded a structure with the urea carbonyl coordinated to Ni-1, one urea nitrogen atom bound to Ni-2, and the metal-bridging position occupied by fluoride (Figure 4G) (Mazzei, Cianci, Benini, et al. 2019b). Replacing this fluoride with solvent would position the nucleophilic hydroxyl group in an ideal alignment to affect hydrolysis of the coordinated urea.

Additional structural studies of S. pasteurii urease have focused on enzyme samples for which Cys322 is covalently inactivated by various reagents including 1,4-benzoquinone (Mazzei, Cianci, Musiani, et al. 2016), 1,4-hydroquinone (Mazzei et al. 2022), catechols (Mazzei, Cianci, Musiani, et al. 2017; Mazzei, Contaldo, et al. 2021), and disulfides (Mazzei, Paul, et al. 2024). In addition, structures have been resolved showing how this side chain interacts with complexes containing silver (Mazzei et al. 2018; Mazzei, Cirri, et al. 2021) or gold (Mazzei, Wenzel, et al. 2019; Mazzei, Massai, et al. 2021). Cys322 is adjacent to His323 (equivalent to Cys319 and His320 of the K. aerogenes enzyme) and located on the protein flap covering the active site. The histidyl residue was suggested to increase the reactivity of the cysteinyl residue during the inactivation reactions, and the catalytic dyad was proposed to protonate the non-coordinated urea amine during catalysis (Mazzei, Paul, et al. 2024). Cryo-electron microscopy was used to demonstrate the presence of multiple conformations for the helix-turn-helix protein flap region containing these two residues of the S. pasteurii enzyme (Mazzei, Tria, et al. 2024).

The crystal structure of urease from H. pylori (Figure 3B) highlights two significant distinctions from the organization of the K. aerogenes or S. pasteurii enzymes (Ha et al. 2001). First, the two small subunits noted in most three-subunit bacterial ureases are fused together so that only two subunits (238 and 569 residues) are present. Second, four (UreA·UreB)₃ units are joined together into a supramolecular spherical biological assembly containing twelve active sites. Regardless of the large changes in overall appearance, the active site components are unchanged. In addition to the structural characterization of the native enzyme, structures were obtained for H. pylori urease inhibited by β-mercaptoethanol, acetohydroxamic acid, and 2-[1-(3,5-dimethylphenyl)-1H-imidazol-2-yl]sulfanyl-N-hydroxyacetamide (Table 1) (Ha et al. 2001; Cunha et al. 2021).

The [(UreA·UreB)₃]₄ quaternary structure of the H. pylori enzyme (Figure 3B) also was observed for the dinuclear iron-containing urease from Helicobacter mustelae (Table 1) (Carter et al. 2011). H. mustelae additionally synthesizes a second urease with a prototypical dinuclear nickel site that likely resembles the H. pylori enzyme, but its structure has not been reported. The iron-containing urease is described in a later separate section.

Urease from Yersinia enterocolitica is of interest because it contains three types of subunits, but forms a macromolecular assembly resembling that seen in the Helicobacter enzymes; i.e., an [(UreA·UreB·UreC)₃]₄ architecture with twelve dinuclear active sites (Table 1, Figure 3C) (Righetto et al. 2020).

Finally turning to jack bean urease, its crystal structure was elucidated in 2010—84 years after the enzyme was first crystallized (Table 1) (Balasubramanian and Ponnuraj 2010). The plant protein possesses a single type of subunit of 840 amino acids, corresponding to a fusion of all bacterial urease subunits. The trimer of these subunits is stacked back-to-back with a second trimer to form a hexameric protein containing six dinuclear active sites (Figure 3D). An analogous structure has been reported for pigeon pea urease (Balasubramanian et al. 2013).

The catalytic mechanism of urease

In neutral solutions, urea decomposes by an elimination reaction to produce ammonium and cyanate ions with an estimated rate constant of 5.5 x 10⁻¹⁰ s⁻¹ (i.e., a half-life of ~40 years) at 25 °C (Shaw and Bordeaux 1955; Alexandrova and Jorgensen 2007; Wolfenden 2011). Compared to the elimination reaction, the spontaneous hydrolysis of urea, k_non, is much slower and has never been directly measured, whereas the hydrolysis of 1,1,3,3,-tetramethylurea (incapable of the elimination reaction) has been estimated as 4.2 x 10⁻¹² s⁻¹ at this temperature (Callahan et al. 2005). Quantum mechanical calculation of the rate constant for urea hydrolysis yielded 2.5 x 10⁻²⁷ s⁻¹ at 37 °C (Estiu and Merz 2004). The latter investigators used a published k_cat/K_m value for urease of 1.4 x 10⁶ M⁻¹·s⁻¹ at this temperature to estimate the catalytic proficiency, (k_cat/K_m)/k_non, of the enzyme to be 5.6 x 10³² M⁻¹. This remarkable value suggested that urease was the most proficient enzyme identified to date (Estiu and Merz 2004), highlighting its unique catalytic mechanism.

Insights derived from urease structures, interactions with inhibitors, analysis of enzyme variants, and other results were used to evaluate a series of reaction mechanisms that have been proposed for the enzyme (Mazzei et al. 2020). The current simplified view of the urease catalytic mechanism is shown in Figure 5. A protein flap covering the active site of urease is known to be positioned differently in the various urease crystal structures and cryo-electron microscopy views, consistent with it being a flexible lid. The lid must open to allow the substrate to enter or the products to depart the active site, and it closes during catalysis. The urea carbonyl oxygen atom displaces a solvent molecule and coordinates to Ni-1 with hydrogen bond stabilization provided by His219 (K. aerogenes numbering) as demonstrated by the large increases in K_m for variants at this site (871, 73, and 95-fold for H219A, H219N, and H219Q, respectively) (Pearson et al. 2000). Closure of the lid results in one of the substrate nitrogen atoms displacing a solvent molecule and coordinating Ni-2, as shown in the fluoride-inhibited, urea-bound structure (Mazzei, Cianci, Benini, et al. 2019b). The same structure as well as those derived after incubating enzyme with phenylphosphorodiamidate, N-(n-butyl)thiophosphoric acid, and N-(n-butyl)phosphoric acid [resulting in urease with bound diamidophosphate or monoamidothiophosphoric acid (Benini S et al. 1999; Mazzei, Cianci, Contaldo, et al. 2017; Mazzei, Cianci, Benini, et al. 2019a; Mazzei, Cianci, Contaldo, et al. 2019)] are consistent with the bridging solvent molecule serving as the nucleophile which attacks the electrophilic carbonyl carbon atom to generate a tetrahedral intermediate. Kinetic studies detailing fluoride inhibition of K. aerogenes urease also were interpreted in terms of the bridging hydroxide being the nucleophile in catalysis (Todd and Hausinger 2000). Decomposition of the tetrahedral intermediate to release product ammonia was suggested early on to be aided by His320 acting as a general acid (Karplus et al. 1997). Substitution of this residue by site-directed mutagenesis to create H320A, H320N, and H320Q variants led to more than four orders of magnitude decreases in k_cat, confirming its importance in catalysis (Pearson et al. 2000). An alternative view had argued that this histidyl group was present in its neutral form and only stabilized the ammonia product (Mazzei et al. 2020); however, more recent work suggested that this histidyl group and its neighboring cysteinyl group influence each other’s reactivity and participate in the protonation step (Mazzei, Paul, et al. 2024; Mazzei, Tria, et al. 2024). Cys319 is not essential for K. aerogenes urease activity as shown by nearly 50% activity remaining in the C319A variant (Martin and Hausinger 1992), so the histidyl residue is most likely to act in protonation.

Figure 5.

A proposed reaction mechanism of urease. Urea displaces a solvent molecule from Ni-1 and coordinates this site with its carbonyl oxygen atom, stabilized by hydrogen bonding with a nearby histidyl group (H219 in the K. aerogenes enzyme). Flap closure leads to solvent dissociation from Ni-2 and binding by a urea nitrogen atom. The metal-bridging solvent molecule attacks the electrophilic carbonyl group of bound urea forming a tetrahedral intermediate. Expulsion of ammonia is facilitated by a nearby histidyl residue (H320 in K. aerogenes urease) and accompanies reformation of the carbonyl group. Opening of the protein flap allows for dissociation of products and binding of three solvent molecules.

Characterization of iron-containing urease

As mentioned earlier, H. mustelae, a pathogen of the ferret, possesses both nickel-containing and iron-dependent ureases. The nickel enzyme is associated with an H. pylori-like ureABIEFGH gene cluster (encoding the two urease subunits, a proton-gated urea permease, plus four accessory genes), whereas the iron enzyme is encoded by the distantly located ureA2B2 cluster of urease structural genes (O’Toole et al. 2010). Notably, the transcription of ureABIEFGH is induced by the addition of nickel ions, whereas the transcription of ureA2B2 is up-regulated by iron and downregulated by nickel ions due to repression via NikR, a nickel-responsive transcriptional regulator (Stoof et al. 2008). The UreA2B2 form of urease was purified from both H. mustelae and E. coli cells that overexpressed the H. mustelae genes. The protein was demonstrated to contain a dinuclear iron metallocenter that was free of nickel, found to have greatly reduced activity compared to the conventional nickel-containing enzyme encoded in the large gene cluster, and shown to be oxygen labile (Carter et al. 2011). The active diferrous H. mustelae enzyme converts to the diferric metallocenter when exposed to oxygen, and the properties of this form of the enzyme have been extensively examined (Proshlyakov et al. 2021). Although the iron-containing urease is much less active than the nickel-containing enzyme, UreA2B2 is produced in copious amounts in the low-nickel gastric environment, providing sufficient activity to allow survival of H. mustelae within the carnivore host.

Iron-containing ureases are thought to exist in several other microorganisms (e.g., Helicobacter acinonychis, Helicobacter felis, and Helicobacter cetorum that are gastric pathogens of cats or dogs, big cats, and dolphins or whales, respectively) (Proshlyakov et al. 2021). None of these proteins have been structurally characterized and they will not be further discussed. The iron-containing ureases have the disadvantage of oxygen sensitivity, but they have the advantage of not requiring an assortment of urease accessory genes for their biosynthesis, as typically observed with the nickel-containing ureases (see next section).

Biosynthesis of the dinuclear nickel site of urease

In bacteria, the genes encoding the two or three distinct urease subunits are typically located adjacent to required auxiliary genes (Mobley et al. 1995), although the Bacillus subtilis urease genes are a notable exception that lack those encoding the known accessory components (Kim et al. 2005). The contiguous genes encode UreD (sometimes called UreH in Helicobacter species), UreE, UreF, and UreG that are needed for assembly of the dinuclear nickel metallocenter (Mulrooney and Hausinger 1990; Lee et al. 1992; Farrugia, Macomber, et al. 2013; Tsang and Wong 2022). Homologs to UreD, UreF, and UreG are made in plants (Myrach et al. 2017) and ureolytic fungi (Singh et al. 2013) where they are encoded distant from the genes encoding the corresponding single-subunit hexameric ureases.

The following paragraphs first describe structural analyses by x-ray crystallography of the accessory proteins and then summarize cryo-electron microscopy characterization of how these components interact with the urease apoprotein (Table 2). As detailed below, UreE is a metallochaperone that delivers nickel to UreG. The GTP-dependent enzyme UreG transfers the metal ion into a tunnel that is formed within UreF and UreD to allow passage to the nascent urease active site. As also discussed, a series of protein complexes are formed among these accessory proteins and the urease apoprotein, with enzyme activation resulting in release of the holoenzyme from the auxiliary components.

Table 2.

Structures of urease accessory proteins and their complexes with urease

Source	Protein	Samples,a (PDB ID), resolution	Reference
Klebsiella aerogenes	(UreE)2	H91A/H144* Se-Met (1GMU), 1.50 Å; H91A/H144* Se-Met + Cu(II) (1GMW), 1.50 Å; H110A/H144* (1GMV), 2.80 Å	(Song et al. 2001)
Sporosarcina pasteurii	(UreE)2	Form 1 + Zn (1EB0), 1.85 Å; Form 2 + Zn (1EAR), 1.70 Å	(Remaut et al. 2001)
		UreE + Ni + Zn (4L3K), 1.88 Å	(Zambelli et al. 2013)
Klebsiella pneumoniae	(UreF)2	UreF (6JC4), 2.30 Å	(Liu et al. 2022)
	(UreG)2	UreG + GMPPNP + Ni (5XKT), 1.80 Å	(Yuen et al. 2017)
	(UreABC)3(UreD)3	Apoprotein + UreD (8HCN), 2.70 Å	(Nim et al. 2023)
Helicobacter pylori	(UreE)2 or (UreE)4	Apoprotein (3L9Z), 2.08 Å; UreE + Cu (3NXZ), 2.70 Å; UreE + Ni (3NY0), 3.09 Å; UreE + Me (3LA0), 2.86 Å	(Shi et al. 2010)
		Apoprotein (3TJA), 2.00 Å; UreE + Ni (3TJ8), 1.59 Å; UreE + Zn (3TJ9), 2.52 Å	(Banaszak et al. 2012)
	(UreF)2	A233* to Q236* Se-Met (3CXN), 1.55 Å	(Lam et al. 2010)
		UreF (2WGL), 2.00 Å	unpublished
		UreF (3O1Q), 1.85 Å	(Fong et al. 2011)
	UreD(UreF)2(UreD)	UreD + UreF (3SF5), 2.50 Å	(Fong et al. 2011)
	UreD(UreFG)2UreD	UreD + UreF + UreG + GDP (4HI0), 2.35 Å	(Fong et al. 2013)
	[(UreAB)3]4(UreDF)12	Apoprotein + E140A UreD + R179A/Y183D UreF (8HC1), 2.30 Å	(Nim et al. 2023)

^aVariants are show with * indicating a truncation after the specified residue;

Se-Met, indicates Se substitution for S; GMPPNP, guanylyl imidodiphosphate; Me, metal of unknown identity.

Italicized samples were examined by cryo-electron microscopy, whereas all others were studied by x-ray crystallography. H. pylori UreD was formerly known as UreH.

UreE proteins have been structurally resolved from K. aerogenes (Song et al. 2001), S. pasteurii (Remaut et al. 2001; Zambelli et al. 2013), and H. pylori (Shi et al. 2010; Banaszak et al. 2012). In each case, the native protein is a homodimer with a critical metal binding site at the subunit interface. The structure of nickel-bound H. pylori UreE (Figure 6A) illustrates metal coordination by using His102 residues from each subunit and His152 from one subunit (Musiani et al. 2004; Banaszak et al. 2012). Some UreE proteins possess additional metal-coordinating sites; e.g., the K. aerogenes protein contains the interfacial His96 residues, His110 and His112 residues in the peripheral domains, and a His-rich carboxyl region with ten histidine residues among the last 15 amino acids of the protein (Colpas et al. 1999; Song et al. 2001). Secondary nickel interaction sites may be used to increase the local concentration of this metal ion to enhance occupancy of the crucial interfacial site. UreE is termed a metallochaperone because it functions in metal delivery, in this case allocating nickel to UreG.

Figure 6.

Structural view of urease accessory proteins involved in the biosynthesis of the enzyme’s dinuclear nickel center. (A) Nickel-bound UreE metallochaperone from H. pylori (PDB ID:3TJ8) with the two subunits of the dimer shown in different shades of cyan, metal-coordinating residues indicated by sticks, and nickel at the subunit interface illustrated as a green sphere. (B) Nickel-bound UreG GTPase from K. pneumoniae (PDB ID:5XKT) with the two subunits of the dimer in different shades of magenta, two GMPPNP illustrated as sticks with yellow carbon atoms, and three nickel ions (two bound to GMPPNP and one to Cys-x-His motifs from each subunit at an interfacial site) are indicated by green spheres. (C) UreF from H. pylori (PDB ID:3O1Q) with the two subunits of the dimer in different shades of yellow. (D) UreD·(UreF)₂·UreD from H. pylori (PDB ID:3SF5) with the UreF subunits colored as in C and UreD subunits illustrated in different shades of green. (E) UreD·(UreF·UreG)₂·UreD from H. pylori (PDB ID:4HI0) is presumed to bind nickel forming UreD·(UreF·Ni·UreG)₂. Its protein chains are colored as above and GDP is shown as sticks with yellow carbon atoms. (F) (UreA·UreB·UreC)₃·(UreD)₃ from K. pneumoniae (PDB ID:8HCN) with UreA subunits in red, UreB in blue, and UreC in orange, and UreD in green. (G) [(UreA·UreB)₃]₄·(UreD·UreF)₁₂ from H. pylori (PDB ID:8HC1) with the UreA subunit in blue, UreB in orange, UreD in green, and UreF in yellow. (H) Model of the nickel transfer tunnel (gray mesh) that extends from the Ni-binding site in UreG (gray), through UreF (green) and UreD (gold) to UreB (purple) of H. pylori urease apoprotein [adapted with permission from supplementary Figure S12 (Nim et al. 2023)]. All proteins are shown in cartoon mode.

In H. pylori, UreE is thought to acquire nickel from HypA, a metallochaperone of nickel that is more commonly utilized for biosynthesis of [NiFe] hydrogenases (Benoit et al. 2007; Yang et al. 2014). The two proteins form a HypA·Ni(II)·(UreE)₂ complex (Hu et al. 2018) that has been studied by molecular modeling and hyperfine-shifted nuclear magnetic resonance spectroscopy, isothermal titration calorimetry, and X-ray absorption spectroscopy methods (Zambelli et al. 2023), but a structure is still not available. Of interest, HypA and UreG compete for binding to UreE and cannot both bind simultaneously (Benoit et al. 2012).

UreG occurs as a monomer, dimer, or an interconverting mixture, depending on its source and the presence of metal ions and GTP, with the only structure reported for the dimeric protein from Klebsiella pneumoniae (Figure 6B) (Yuen et al. 2017). This protein was crystallized with nickel and GMPPNP, a nonhydrolyzable analog of GTP—a substrate of all UreGs. One nickel ion is coordinated in square-planar geometry by Cys66 and His68 residues from each subunit, with additional metal ions substituting for the native magnesium associated with the two GMPPNP anions. The nickel-bound UreE dimer is known to interact with UreG in order to transfer the metal ion, with transient formation of (UreE)₂·Ni·UreG and (UreE)₂·Ni·(UreG)₂ species (Bellucci et al. 2009; Merloni et al. 2014). The nickel-bound UreG dimer next associates with dimeric UreF or with UreF within a larger complex.

The structures of (UreF)₂ from H. pylori (Lam et al. 2010; Fong et al. 2011) and from K. pneumoniae (Liu et al. 2022) (Figure 6C) were crystallographically characterized, as were the structures of complexes for H. pylori UreD·(UreF)₂·UreD (Figures 6D) (Fong et al. 2011) and UreD·(UreF·UreG)₂·UreD (Figures 6E) (Fong et al. 2013). In contrast, no characterization of the isolated UreD/H protein has been reported from any microorganism. Both the H. pylori UreF and the maltose-binding protein fusion of UreD from K. aerogenes were reported to coordinate nickel ions (Carter and Hausinger 2010; Zambelli et al. 2014), but their primary role is to form a scaffold and portal for nickel transfer from UreG to the apoprotein of urease. Neither the UreF dimer nor the UreD·(UreF)₂·UreD tetrameric complex exhibit interior channels; however, CAVER (Chovancova et al. 2012) analysis of the UreD·(UreF·UreG)₂·UreD hexameric complex from H. pylori revealed a tunnel that extends from the metal-binding site of UreG through UreD and UreF (Zambelli et al. 2014; Musiani et al. 2017; Masetti et al. 2021). The structural analyses of the accessory proteins were significant achievements, but these structures left unknown how the auxiliary proteins were bound to the apoprotein of urease for completing metallocenter biosynthesis.

Cryo-electron microscopy was used to unveil the structures of urease apoprotein in complex with its accessory proteins, specifically for samples of K. pneumoniae (UreA·UreB·UreC)₃·(UreD)₃ and H. pylori [(UreA·UreB)₃]₄·(UreD·UreF)₁₂ (Figures 6F,G) (Nim et al. 2023). UreD attaches near the UreB subunit vertices of the triangular K. pneumoniae protein, consistent with results from chemical cross-linking and small angle x-ray scattering studies of the K. aerogenes counterpart (Chang et al. 2004; Quiroz-Valenzuela et al. 2008). The H. pylori protein complex made use of an E140A variant of UreD that enhanced stability of the complex and an R179A/Y183D variant of UreF that disrupted its dimerization interface; however, the dimer form of UreF was suggested to be required for activation of the H. pylori urease apoprotein (Nim et al. 2023). The urease·UreD·UreF interaction sites noted for the H. pylori proteins agree well with the predictions made using the K. aerogenes proteins based on chemical cross-linking and small angle x-ray scattering results (Chang et al. 2004; Quiroz-Valenzuela et al. 2008). Although a high-resolution structure is not available for urease in complex with UreD, UreF, and UreG, this protein species from K. aerogenes was characterized by ion mobility mass spectrometry (Farrugia, Han, et al. 2013; Eschweiler et al. 2018). The results suggest the binding of one subunit of each accessory protein extending out from each UreA·UreB·UreC unit despite the occurrence of an overall dimeric UreD·(UreF·UreG)₂·UreD precursor being synthesized in this microorganism. These findings indicate the hexameric species must dissociate to UreD·UreF·UreG during formation of the activation complex with the urease apoprotein.

The structure of the H. pylori [(UreA·UreB)₃]₄·(UreD·UreF)₁₂ complex combined with that of the UreD·(UreF·UreG)₂·UreD complex allows one to model a 100 Å-long tunnel (Figure 6H) extending from the nickel-binding site of UreG through UreF and UreD to the nascent active site in UreB (Nim et al. 2023). Molecular dynamics simulations suggest that nickel passage through the tunnel occurs in a “bucket brigade” mode via a series of weak nickel-binding sites within UreF and UreD (Masetti et al. 2021). Mutagenesis studies of K. aerogenes ureD showed that the substitution of residues lining the proposed tunnel with larger amino acids blocks the ability of the protein to activate urease apoprotein, consistent with nickel trafficking through this tunnel (Farrugia et al. 2015).

The overall proposed pathway for urease activation begins with the dimeric metallochaperone UreE in its nickel-bound form handing off the metal ion to the UreG GTPase, which is monomeric or dimeric depending on the protein source and solution conditions. The nickel- and GTP-bound form of UreG associates as a dimer with (UreF·UreD)₂, possibly already bound to urease apoprotein, to generate UreD·(UreF·Ni·UreG)₂·UreD or the analogous complex with urease apoprotein. In at least some systems, the UreD·(UreF·Ni·UreG)₂·UreD complex dissociates to release UreD·UreF·UreG while leaving the second nickel- and GTP-containing UreD·UreF·UreG complex connected to each unit of urease apoprotein. GTP hydrolysis disrupts the metal binding site (Yuen et al. 2017) and promotes the release of nickel into the internal space of UreF. The released nickel ion traverses the tunnel through UreF and UreD to reach the nascent active site within the urease apoprotein containing the previously carbamylated lysyl residue. This process must be repeated to provide the second nickel ion of the dinuclear site, and the accessory proteins then dissociate from the holoenzyme for possible recycling.

Conclusion and Remaining Questions

In the century that has elapsed since James B. Sumner successfully crystallized jack bean urease, thus proving that enzymes are proteinaceous, the presence of a dinuclear nickel site within the protein was demonstrated and urease structures were elucidated for the enzyme from several sources. Surprisingly, one structurally characterized example was shown to be an oxygen-sensitive, iron-containing urease. Insights derived from the structures, especially for inhibitor-bound ureases, were combined with kinetic analyses and results from site-directed mutagenesis studies to uncover the catalytic mechanism of this remarkably proficient enzyme. The pathway for biosynthesis of the dinuclear nickel-containing urease was shown to require a series of accessory proteins, and structural studies have played critical roles in understanding how the metallochaperone UreE delivers nickel to the GTPase UreG which inserts the metal ion into a tunnel that passes through UreF and UreD to reach the nascent active site in urease apoprotein.

Many questions remain to be addressed about urease catalysis and its mechanism of biosynthesis. What factors associated with the metallocenter and the protein allow ureases to be among the most proficient enzymes? Computational simulations are perhaps the most fruitful approach to shed light on this question. What governs the different levels of activity for ureases containing nickel, iron, manganese, and other metal ions? Again, computational methods along with analysis of biomimetic inorganic complexes may provide useful insights for uncovering the metal dependence of the enzyme. What are the evolutionary and environmental drivers that direct the use of nickel versus iron in ureases? Further characterization of iron-containing ureases along with detailed bioinformatic analyses are needed to probe this question. What role is played by second-sphere residues at the active site (i.e., those which do not coordinate the metals), such as the histidyl and cysteinyl groups on the protein flap? Additional characterization of site-directed variants and computational approaches will play a critical role in answering this query. Why has Nature selected a carbamylated lysine residue for coordinating the metallocenter and how is the carbamylation achieved? Analyses of variants and investigation of carbamylated, but not metalated, forms of the enzyme (e.g., by ¹³C nuclear resonance spectroscopy of protein mixed with labeled carbon dioxide) my offer insights into this point. Where does UreE acquire nickel in non-Helicobacter species? Direct interactions between UreE and nickel transport proteins need to be studied. Why doesn’t UreG directly donate nickel to urease apoprotein without the intermediacy of UreD and UreF? The timing of carbamylation versus metallation should be further investigated to define whether the tunnel is important to prevent premature nickel incorporation. What are the structural consequences of GTP hydrolysis by UreG that promote nickel transfer into the UreF/UreD tunnel? Additional structural studies of urease complexes that include UreG variants may address this question. How is the process controlled to ensure two nickel ions are delivered to each urease activity site? The structural and activity consequences of using substoichiometric metal levels need to be clarified. How did such a complicated “Rube Goldberg-type” process evolve to achieve urease activation? Additional investigation of urease activation using the enzyme from B. subtilis, which does not possess any of the four known accessory proteins, may generate insights. Many of these and other questions likely will be addressed as we move into the second century since urease was crystallized.