A proposed update for the classification and description of bacterial lipolytic enzymes

Introduction

Bacterial lipolytic enzymes are involved in a diverse range of microbial functions from metabolism of fat for energy (Drouault et al., 2002) to pathogenesis (Canaan et al., 2004). These enzymes are also of interest to the biotechnology industry for production of biodiesel and biopolymers (Jaeger & Eggert, 2002). The alpha-beta hydrolase family contains both lipases (Enzyme commission (EC): 3.1.1.3) and carboxylesterases (EC: 3.1.1.1), both of which can catalyse the hydrolysis of 4-nitrophenyl-bound fatty acids (Guo et al., 2010; Rao et al., 2011). Lipases are defined as those which display some ability to catalyse the release of free fatty acids from long-chain triacylglycerols, such as those found in plant oil and animal fat, however activity may be highest against short or medium chain fatty acids (Jaeger & Eggert, 2002; Levisson, Van der Oost & Kengen, 2007; Glogauer et al., 2011). This separates the lipases from the esterases, which may have activity against p-nitrophenyl (pNP)-bound fatty acids (a standard test for substrate specificity) but are unable to act upon triacylglycerol molecules (Yu et al., 2010; Masomian et al., 2016). A further subset of lipases are the true lipases which are defined as lipases which display preferential activity against long-chain fatty acids (>C10) (Sorokin & Jones, 2009). The gold standard classification system of bacterial lipolytic enzymes was originally published by Arpigny & Jaeger (1999) and defined eight families based on the biochemical properties and sequence similarity of 53 proteins.

Under the Arpigny & Jaeger (1999) classification system, true lipases were placed within a singular family (family I), within which sub-families are defined by unique lipase groups (Arpigny & Jaeger, 1999; Jaeger & Eggert, 2002). Originally, six sub-families of true lipases were described based on shared features such as source genus, requirement of a chaperone for folding, motifs and sequence similarity (Arpigny & Jaeger, 1999). Jaeger & Eggert (2002) redefined these families by splitting family I.5 into two new families, family I.5 and family I.6, and moving the previous members of family I.6 into family I.7 (Jaeger & Eggert, 2002). In the 20 years since this bacterial lipolytic enzyme classification system was published, more than twice as many ‘novel’ families have been proposed (Handrick et al., 2001; Ewis, Abdelal & Lu, 2004; Lee et al., 2010; Lee et al., 2006; Levisson, Van der Oost & Kengen, 2007; Bender et al., 2009; Kim et al., 2009; Hu et al., 2010; Rao et al., 2011; Jeon et al., 2011; Bassegoda, Pastor & Diaz, 2012; Charbonneau & Beauregard, 2013; Zarafeta et al., 2016; Castilla et al., 2017; Parapouli et al., 2018). A recent update to this system expanded the system to include some (11 of 30) of the recently published novel lipases within 19 families and eight true lipase sub-families, however many lipolytic proteins remain unclassified (Kovacic et al., 2019).

Many claims of identifying novel families of lipolytic enzymes have been made, however no exhaustive comparison has been conducted to verify and combine these results. This has led to an inconsistent characterisation system and multiple publications claiming discovery of the same family (Parapouli et al., 2018). Examples of this are family X (Levisson, Van der Oost & Kengen, 2007; Bassegoda, Pastor & Diaz, 2012) and family XV (Charbonneau & Beauregard, 2013; Zarafeta et al., 2016), both of which have been coined twice. Rao et al. (2013) attempted to incorporate previously published families into their analysis, however their literature search did not include multiple families (Chu et al., 2008; Bender et al., 2009; Hu et al., 2010; Lee et al., 2010; Jeon et al., 2011; Nacke et al., 2011; Bassegoda, Pastor & Diaz, 2012). Ferrer et al. (2015) also attempted to integrate additional families into the Arpigny & Jaeger (1999) classification system. Both previous attempts at including additional families relied on the authors claims of ‘novelty’ without reanalysis of these claims and, as discussed above, incomplete literature searches mean additional ‘novel’ families have been missed from the analysis. The possibility exists that multiple publications may describe members of the same ‘novel’ family under different names, artificially inflating the number of families. Therefore, before additional families can be incorporated into the system and provided with a family number, previous claims of novelty must be confirmed and integration of these proteins into a singular classification system must occur.

Alternative classification systems for bacterial lipolytic enzymes have been published alongside online databases dedicated to these new classification systems, including MELDB (Kang et al., 2006) and LIPABASE (Messaoudi et al., 2011). Both websites have since been deactivated preventing their continued use. The simplicity of the Arpigny & Jaeger (1999) classification system is that by providing type proteins for each family, the system can continue to be utilised by the research community. Additional databases such as the Lipase Engineering Database (Pleiss et al., 2000) and ESTHER (Lenfant et al., 2013) classify similar groups of enzymes, however these are not specific to bacterial lipolytic enzymes and as with MELDB and LIPABASE, their continued availability is not guaranteed.

In this paper, the novelty of lipolytic enzymes was confirmed via pairwise sequence comparison to the current type proteins. This allowed us to expand the current classification system and remove spurious claims of novelty and provide a consistent nomenclature to these proteins.

Methods

Results and Discussion

Conclusions

In this paper, we have significantly expanded the classification of lipolytic enzymes to 35 families and 11 true lipase subfamilies. Defining the biochemical features of each family along with providing type proteins allows for a greater number of potentially novel lipases to be assigned to pre-described families, preventing inconsistency within the literature. The lack of multiple members for each family with proven activity prevented comparative analysis of motifs and biochemical features.

N-terminal single transmembrane alpha-helices were detected within type proteins for four families/sub-families. Experimental evidence gained via deletion of the N-terminal transmembrane alpha-helix of LipA identified that this helix lies within the 3D structure of the protein allowing for opening of the lipase cap (Zha et al., 2014). It is likely that the transmembrane alpha-helices detected within the newly described families play a similar role in the structure and functionality of the proteins and do not represent membrane-bound lipases.

Previous research has shown that sequence identity does not guarantee shared functionality (Punta & Ofran, 2008). The average pairwise similarity between the lipolytic enzymes was 31.3% with the minimum being 14.2%. Whilst these enzymes do not share identical functionality, they have demonstrated the ability to hydrolyse the release of pNP-bound fatty acids. The low level of sequence similarity between lipolytic enzymes suggests that in vitro examination of proteins sharing low homology with known lipolytic enzymes may lead to identification of additional families.

Cultivation-independent investigation of environmental samples led to the identification and characterisation of multiple lipolytic families suggesting that a greater number of enzymes that have yet to be characterised likely exist. The inability of many families to be assigned to a host via genome search suggests that not only do further lipolytic enzymes remain uncharacterised, but their host species have yet to be cultured. We hope that with the recent increased effort in cultivation of novel species from environmental samples additional enzymes of biotechnological importance will be identified.

Supplemental Information