Connectivity between Catalytic Landscapes of the Metallo-β-Lactamase Superfamily

Highlights

• We analyzed evolutionary and functional connectivity within the MBL superfamily.

• Enzymes showed catalytic promiscuity, and they catalyze on average 2.5 reactions.

• Catalytically distinct functions are highly connected through promiscuous enzymes.

• Some evolutionary and functionally connected pairs are also observed in other superfamilies.

• New functions could evolve rapidly from the repertoire of existing enzymes.

Abstract

The expansion of functions in an enzyme superfamily is thought to occur through recruitment of latent promiscuous functions within existing enzymes. Thus, the promiscuous activities of enzymes represent connections between different catalytic landscapes and provide an additional layer of evolutionary connectivity between functional families alongside their sequence and structural relationships. Functional connectivity has been observed between individual functional families; however, little is known about how catalytic landscapes are connected throughout a highly diverged superfamily. Here, we describe a superfamily-wide analysis of evolutionary and functional connectivity in the metallo-β-lactamase (MBL) superfamily. We investigated evolutionary connections between functional families and related evolutionary to functional connectivity; 24 enzymes from 15 distinct functional families were challenged against 10 catalytically distinct reactions. We revealed that enzymes of this superfamily are generally promiscuous, as each enzyme catalyzes on average 1.5 reactions in addition to its native one. Catalytic landscapes in the MBL superfamily overlap substantially; each reaction is connected on average to 3.7 other reactions whereas some connections appear to be unrelated to recent evolutionary events and occur between chemically distinct reactions. These findings support the idea that the highly distinct reactions in the MBL superfamily could have evolved from a common ancestor traversing a continuous network via promiscuous enzymes. Several functional connections (e.g., the lactonase/phosphotriesterase and phosphonatase/phosphodiesterase/arylsulfatase reactions) are also observed in structurally and evolutionary distinct superfamilies, suggesting that these catalytic landscapes are substantially connected. Our results show that new enzymatic functions could evolve rapidly from the current diversity of enzymes and range of promiscuous activities.

Introduction

New enzymatic functions are thought to evolve through the recruitment and optimization of latent promiscuous functions of existing enzymes, leading to the functional expansion of superfamilies [1], [2], [3]. Closely related enzymes have shared promiscuous activities or crosswise promiscuity, as defined by each possessing a low level of catalytic activity against the other's native reaction [3], [4]. Thus, enzyme promiscuity can provide an additional layer of evolutionary connectivity between functional families in addition to their sequence and structural relationships. Systematic characterizations of substrate promiscuity among homologous and reconstructed ancestral enzymes have elucidated how substrate specificity has diverged within enzyme families [5], [6], [7], [8], [9], [10]. In addition, catalytic promiscuity, an enzyme's ability to catalyze distinct chemical reactions to their native one, provides functional connectivity between homologous enzymes [3], [11], [12], [13], [14], [15]. For example, lactonase, arylesterase and phosphotriesterase activities are shared among enzymes of the amidohydrolase superfamily [16], [17], [18]. Connectivity between phosphatase, phosphodiesterase, phosphonatase and arylsulfatase activities among members of the alkaline phosphatase superfamily has also been demonstrated [14], [19], [20], [21].

By extending Maynard-Smith's picture of a continuous network of functional proteins [22], we hypothesize that, through promiscuous enzymes, the shared and overlapping catalytic landscapes form a continuous network (Box 1). Therefore, within a given superfamily, functional changes can occur gradually and in a continuous manner [23]. Previous studies have focused on the connectivity between the native reactions of individual pairs of closely related homologues (~ 30% sequence identity) and included similar activities that share chemical properties such as transition-state geometry, hydrolyzable bond and bond charge [14], [16], [19], [20]. Hence, the extent to which catalytic landscapes are evolutionary and functionally connected, on a broad level that spans the vast sequence space of a functionally diverged superfamily, remains elusive.

The metallo-β-lactamase (MBL) superfamily served as our model enzyme superfamily. Members of the MBL superfamily are substantially diverged in sequence and function. Approximately 34,000 sequences are registered in the protein family database (Pfam) [24]. The amino acid sequence identity between members can be less than 5%, but members share structural features such as the αββα-fold (MBL fold) and a mononuclear or binuclear active-site center with a unique metal binding motif (H-X-H-X-D-H) [25] (Fig. 1 and Figs. S1 and S2). To date, at least 24 distinct functional families have been identified within the MBL superfamily, including DNA, RNA and nucleotide processing, detoxification, antibiotic resistance, quorum-quenching and pesticide hydrolysis [25], [26], [27], [28], [29], [30], [31], [32]. Most of these functions involve hydrolytic reactions and target diverse substrates with different chemical properties such as phosphodiester, phosphotriester, choline-phosphoester, thiol-ester, sulfate-ester, carbon-ester and β-lactam bond [25], [26], [27], [28], [29]. Other functions involve non-hydrolytic reactions such as nitric oxidoreduction [30] and sulfur dioxygenation [31], as well as non-enzymatic functions such as binding and transport [25], [32]. Many of the hydrolytic functions are amenable to characterization by simple colorimetric assays, which makes the MBL superfamily particularly well suited to analyze the connectivity of catalytic landscapes within a superfamily. In this study, we analyzed the sequence relationship within the MBL superfamily using sequence similarity networks, revealing evolutionary proximity among functional families. Furthermore, we assayed the activity profile of 24 MBL members from 15 different functional families for 10 catalytically distinct reactions. Finally, we relate the observed functional connections to evolutionary connections and chemical similarity between the reactions, and we discuss how the overlapped catalytic landscapes are related to evolutionary divergence and evolvability of the MBL superfamily.