Escherichia coli ykfE ORFan Gene Encodes a Potent Inhibitor of C-type Lysozyme*

The complete nucleotide sequences of over 37 microbial and three eukaryote genomes are already publicly available, and more sequencing is in progress. Despite this accumulation of data, newly sequenced microbial genomes continue to reveal up to 50% of functionally uncharacterized “anonymous” genes. A majority of these anonymous proteins have homologues in other organisms, whereas the rest exhibit no clear similarity to any other sequence in the data bases. This set of unique, apparently species-specific, sequences are referred to as ORFans. The biochemical and structural analysis of ORFan gene products is of both evolutionary and functional interest. Here we report the cloning and expression ofEscherichia coli ORFan ykfE gene and the functional characterization of the encoded protein. Under physiological conditions, the protein is a homodimer with a strong affinity for C-type lysozyme, as revealed by co-purification and co-crystallization. Activity measurements and fluorescence studies demonstrated that the YkfE gene product is a potent C-type lysozyme inhibitor (K i ≈ 1 nm). To denote this newly assigned function, ykfE has now been registered under the new gene name Ivy (inhibitor ofvertebrate lysozyme) at the E. coligenetic stock center.

The complete nucleotide sequences of over 37 microbial and three eukaryote genomes are already publicly available, and more sequencing is in progress. Despite this accumulation of data, newly sequenced microbial genomes continue to reveal up to 50% of functionally uncharacterized "anonymous" genes. A majority of these anonymous proteins have homologues in other organisms, whereas the rest exhibit no clear similarity to any other sequence in the data bases. This set of unique, apparently species-specific, sequences are referred to as ORFans. The biochemical and structural analysis of ORFan gene products is of both evolutionary and functional interest. Here we report the cloning and expression of Escherichia coli ORFan ykfE gene and the functional characterization of the encoded protein. Under physiological conditions, the protein is a homodimer with a strong affinity for C-type lysozyme, as revealed by co-purification and co-crystallization. Activity measurements and fluorescence studies demonstrated that the YkfE gene product is a potent C-type lysozyme inhibitor (K i Ϸ 1 nM). To denote this newly assigned function, ykfE has now been registered under the new gene name Ivy (inhibitor of vertebrate lysozyme) at the E. coli genetic stock center.
Despite the accumulation of sequence information from a large diversity of species and phyla, newly sequenced bacterial genomes continue to reveal a high proportion of genes of unknown function (1), including a significant subset of "ORFans" (2), i.e. putative open reading frames (ORFs) 1 without significant similarity to any previously encountered protein (or conceptual translation) sequences. Most genes found in data bases have only been predicted by computer methods and never experimentally validated. It is thus expected that some annotated ORFs, in particular among the ORFans, might not correspond to real genes. In a previous study, we verified the existence of a cognate transcript for 25 Escherichia coli ORFans with a surprising rate of success (92%) (3). Given that most ORFans appear to be transcribed, we have now initiated a systematic expression and structure determination program for the proteins encoded by these (apparently) unique genes. Because three-dimensional structures are more resilient to evolution and change than amino acid sequences, it is expected that some ORFans should exhibit structural similarity to previously described protein families, hence providing some functional hints. Alternatively, targeting ORFans for structure determination is also a suitable strategy to optimize the discovery of original protein folds, one of the goals of structural genomics.
In a pilot study involving five ORFan genes, we succeeded in producing four of them in E. coli as soluble proteins, and we report here the most advanced project, ykfE. YkfE (Swiss-Prot accession number P45552; b0220 in the Blattner data base (4)) is a 474-nucleotide-long uncharacterized ORF. It is part of a single gene operon and was found to exhibit a high level of expression during the exponential and stationary phases of E. coli growth (3). The ykfE ORF exhibits an N-terminal signal peptide cleaved to produce the mature protein (5,6). Initial purification steps and biochemical analyses suggested a strong interaction between this protein and hen egg white lysozyme (HEWL). The existence of a stable complex was confirmed by biophysical analyses, and enzymatic studies revealed the capacity of ykfE to inhibit hen and human C-type lysozymes through a specific interaction. The x-ray structure determination of ykfE, both in isolation (7) and in a complex with HEWL, is currently in progress and should allow us to understand the molecular basis of the ykfE-lysozyme interaction at atomic resolution. To denote its newly assigned function, ykfE has been registered under the new gene name Ivy (for inhibitor of vertebrate lysozyme) at the E. coli genetic stock center.
Expression and Purification of the YkfE Gene Product-The ykfE gene product (Ivy) was expressed by culturing E. coli XL1-Blue carrying the plasmid pQE-0220 in LB ϩ Amp medium. After initial growth at 37°C, temperature was set at 30°C when A 600 reached 0.4, and Ivy expression was induced by adding 1 mM isopropyl-1-thio-␤-D-galactopyranoside. Cells were harvested at A 600 around 2-2.5 and resuspended in Buffer A (20 mM sodium phosphate, pH 8.0, 300 mM NaCl) containing 1.5% Triton X-100, 1.5% glycerol, and 1 mg⅐ml Ϫ1 HEWL before sonication. Protein extraction was also performed in the absence of exogenous lysozyme to obtain lysozyme-free ykfE protein after the existence of a complex had been recognized. In both cases, purification was achieved by nickel affinity chromatography. The cleared lysate was applied to a 5-ml HiTrap chelating column (Amersham Pharmacia Biotech) charged with Ni 2ϩ and was washed with 10 column volumes of Buffer A, followed by 10 column volumes of Buffer A containing 25 mM imidazole, and 5 column volumes of Buffer A containing 70 mM imidazole at a flow rate of 1 ml⅐min Ϫ1 . Elution was performed with a linear gradient over 8 * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ To whom correspondence should be addressed. column volumes from 70 to 500 mM imidazole. The recombinant protein was eluted with 150 -200 mM imidazole, and fractions were pooled and desalted against 20 mM Tris, pH 8.0, on a fast desalting column HR 10/10 (Amersham Pharmacia Biotech) at a flow rate of 5 ml⅐min Ϫ1 . Protein concentration was determined by UV absorption at 280 nm using extinction coefficients calculated on the basis of tyrosine and tryptophan contents (8). Protein purity was assessed by SDS polyacrylamide gel electrophoresis and isoelectrofocusing (IEF) using 3 to 10 pH gradient pre-cast gels (Novex). Preliminary molecular weights for the purified proteins were estimated by gel filtration using a calibrated Superdex 75 HR10/30 column (Amersham Pharmacia Biotech) equilibrated with a 20 mM sodium citrate buffer, pH 6.5, at a flow rate of 0.5 ml⅐min Ϫ1 . The purified proteins were characterized by mass spectroscopy (matrix-assisted laser desorption ionization/time of flight, Voyager DE-RP; PerSeptive Biosystems) and by N-terminal Edman sequencing (473A; Applied Biosystems).
Interaction Measurements-To assay the interaction between Ivy and HEWL, 50 mg of HEWL were loaded on a nickel column at a flow rate of 1 ml⅐min Ϫ1 of 20 mM sodium phosphate, pH 8.0, 300 mM NaCl in the presence or absence of 3 mg of pure Ivy protein. After extensive washing with 20 mM sodium phosphate, pH 8.0, 1 M NaCl, elution was performed with a linear gradient over 5 column volumes to 1 M imidazole Intrinsic protein fluorescence was measured with a Spex Fluorolog3 photon-counting spectrofluorimeter (Jobin Yvon-Spex, Longjumeau, France) equipped with a 450-watt Xenon source and a cooled photomultiplier. Tryptophan fluorescence emission spectra were recorded between 290 and 450 nm from solutions containing the individual proteins and from solutions containing a mixture of proteins excited with 280 nm of light. The degree of protein-protein interaction was determined from the extent of fluorescence quenching observed at 344 nm when spectra of a mixture of proteins were compared with the sum of the individual protein spectra at the same concentration. Interaction-dependent fluorescence quenching was determined in 10 mM Tris-HCl buffer, pH 7.0, 8.0, or 9.0, containing 100 mM NaCl at protein concentrations varying from 1 M to 0.5 nM.
Determination of the Apparent Dissociation Constant (K i )-HEWL activity assay was performed at 25°C in 100 mM potassium phosphate, pH 6.4, using 0.125 mg⅐ml Ϫ1 Micrococcus lysodeikticus (Sigma) as substrate. Inhibition studies were carried out by monitoring the change in turbidity associated with the lysis of M. lysodeikticus cells as described previously (9). One unit of HEWL activity was defined as the amount of enzyme causing a decrease in extinction of 0.001 per min at 450 nm. K i value was determined according to the slow tight binding competitive inhibition model (with no conformational change) (10,11). The following equation was used, where K i is the apparent dissociation constant, Et is the total enzyme (HEWL) concentration, It is the total inhibitor (Ivy) concentration, V Ivy is the inhibited velocity for a given concentration of Ivy, and V 0 is the velocity in the absence of inhibitor. HEWL (70 nM) was pre-incubated with Ivy (0 -200 nM) at room temperature for 15 min prior to the addition of the M. lysodeikticus substrate (0.125 mg⅐ml Ϫ1 ). The K i value was determined by fitting the experimental data onto the V Ivy /V 0 theoretical curves computed from the above equation (see Fig. 4).
Analysis of the Specificity of Ivy-The effect of Ivy on C-type lysozyme activity was determined in the presence of 0.5 g⅐ml Ϫ1 of HEWL (Sigma), with increasing concentrations of Ivy from 0 to 100 g⅐ml Ϫ1 . The effect on phage lysozyme activity was assayed in 20 mM Tris, pH 8.0, at 25°C, according to Soumillion et al. (12) and using chloroformtreated E. coli K-12 MG1655 cells as substrate. Activity was determined by measuring the decrease of turbidity over time at 570 nm in the presence of 0.04 g⅐ml Ϫ1 of phage lysozyme and concentrations of Ivy ranging from 0 to 6 g⅐ml Ϫ1 . The chitinase assay was performed in a 200 mM potassium phosphate buffer, pH 6.0, and 2 mM CaCl 2 using crab shell chitin covalently linked with remazol brilliant violet 5R (Sigma) as substrate, according to Hackman and Golberg (13). The effect of Ivy on chitinase activity was determined in the presence of 0.5 g⅐ml Ϫ1 of Streptomyces griseus chitinase (Fluka) at Ivy concentrations ranging from 0 to 5 g⅐ml Ϫ1 . The enzymatic activity was monitored at 575 nm.
Human saliva was also used as a source of lysozyme. 10 l of fresh saliva were mixed with 1 ml of 100 mM potassium phosphate, pH 6.4, containing 0.125 mg⅐ml Ϫ1 M. lysodeikticus. Concentrations of Ivy ranging from 0 to 20 g⅐ml Ϫ1 were used for this assay.

RESULTS AND DISCUSSION
Exogenous HEWL is usually added prior to sonication to help the disruption of the E. coli cell wall according to the usual extraction protocol for recombinant proteins. HEWL is then removed during the subsequent purification steps. In the case of Ivy, a succession of anomalies led us to suspect a strong interaction between the two proteins. After the purification step by metal chelating chromatography on a nickel resin, SDS polyacrylamide gel electrophoresis analyses of the eluted proteins revealed the presence of two bands of nearly identical molecular mass, around 15 kDa, thus close to the predicted value for the mature form of Ivy (without signal peptide). Mass spectrometry and N-terminal sequencing clearly indicated that these fractions consisted of a mixture of two proteins present in equivalent quantities and identified one of them as the mature Ivy protein (molecular mass ϭ 15.04 kDa) and the other as the exogenous HEWL (molecular mass ϭ 14.3 kDa).
Preliminary results suggested the existence of a specific interaction between the two proteins. During metal chelating chromatography on a nickel resin, HEWL could only be retained if Ivy (extracted in the absence of lysozyme) had first been trapped on the column (see "Experimental Procedures"). The SDS polyacrylamide gel electrophoresis analysis of the eluted fractions confirmed the co-elution of Ivy and HEWL. Finally, the effect of the increase of HEWL concentrations on the IEF behavior of Ivy also suggested a strong interaction (Fig.  1). In the absence of HEWL, the IEF migration of Ivy exhibited two close bands at pI ϭ 7.0 and pI ϭ 6.7 (for a theoretical pI of 6.74). These two bands most likely correspond to the dimeric and monomeric forms of Ivy in solution. HEWL alone was found to migrate at pIϾ10. The increase of HEWL concentrations resulted in a decrease of intensity of the Ivy bands at pI ϭ 7.0 and pI ϭ 6.7 (Fig. 1) and the appearance of an extra band of material reverse migrating into the gel wells, at pI Ӎ 10.
The HEWL-Ivy interaction was further studied by fluorescence spectroscopy. The fluorescence spectrum of a 1 M lysozyme, 1 M Ivy mixture was found to differ significantly from that expected when adding the fluorescence emission spectra of the individual proteins. An overall 20% quenching of the fluorescence was measured with maximal quenching at 344 nm (Fig. 2). The shape and maximum of the spectrum are consistent with at least one relatively exposed tryptophan being quenched in the lysozyme-Ivy complex. The examination of the concentration dependence of the quenching spectrum showed that the shape of this spectrum was independent of the protein concentration between 1 nM and 1 M. No reliable K d value measurement could be obtained because of the insufficient intrinsic fluorescence intensity in the nM concentration range. Finally, co-crystallization experiments and the subsequent analysis of the crystal content demonstrated the presence of both proteins, thus suggesting a specific and stable interaction between the two molecules (data not shown). The determination of the complex three-dimensional structure is currently in progress.
In addition to its specific physical interaction with HEWL, Ivy is also a potent inhibitor of HEWL enzymatic activity (Fig.  3). Preliminary experiments showed that in the presence of 1 g⅐ml Ϫ1 of Ivy, the addition of 1 g⅐ml Ϫ1 of HEWL produced a nonlinear kinetic with an upward concavity (Fig. 3a, curve b). In contrast, the pre-incubation of Ivy with HEWL for 15 min resulted in kinetic exhibiting a slight downward concavity (Fig.  3a, curve c). These results suggest a slow binding kinetic model for the Ivy-HEWL interaction. In addition, near-complete inhibition is reached for a range of Ivy concentrations comparable with the concentration of HEWL (see Figs. 3b and Fig. 4), indicating that Ivy behaves as a slow tight binding inhibitor. A K i value of about 1 nM was thus estimated by fitting the experimental data (Fig. 4)  The previous experiments demonstrated the potent inhibitory activity of Ivy on hen egg white lysozyme. We then explored the effect of Ivy on the related proteins of increasing evolutionary divergence. We selected a set of lysozyme and lysozyme-like proteins based on structural similarity using the MMDB data base (14). Using HEWL as initial query (MMDB accession number 1151), phage lysozyme (root mean square deviation, 1.3 Å; 21.1% identity), and chitinase (root mean square deviation, 1.9 Å; 11.1% identity) were selected as representatives of structural homologs with low sequence similarity. The inhibitory effect of Ivy was thus tested on the two proteins. Ivy was found to cause a weak inhibition of phage lysozyme (Fig. 5). The activity was only reduced by 15% at a molar ratio of 200:1, Ivy: phage lysozyme. We found no inhibitory effect of Ivy on chitinase from S. griseus. We then investigated the capacity of Ivy to inhibit other C-type lysozymes and tested human saliva, because this secretion was reported to contain 30 -55 g⅐ml Ϫ1 of lysozyme (15). Ivy was found to strongly inhibit the lysozyme activity in saliva (Fig. 6). Around 50 g⅐ml Ϫ1 of Ivy is sufficient to observe a decrease of 50% of the activity, which was fully abolished for an Ivy concentration of 0.5 mg per ml of saliva.
On a gel filtration column, Ivy is eluted with an apparent molecular mass of about 30 kDa, indicating that the predominant form in solution is a homodimer, as already suggested by IEF experiments. Fluorescence studies confirmed this model. The fluorescence emission spectrum of HEWL exhibits a broad peak with a maximum at 342 nm and a long wavelength tail typical of relatively exposed tryptophan residues. In contrast, the spectrum of Ivy shows a peak at 334 nm ϳ25% more intense on an absolute scale and 2.5 times more intense on a per tryptophan scale (Fig. 2). Such intense fluorescence and the relatively short wavelength of maximum emission both argue for tryptophans buried within apolar environments. Furthermore, the shape of the emission spectrum was found to be independent of the protein concentration in a broad 0.5 nM to 1 M range and appears insensitive to change in pH between 7.0 lysodeikticus as substrate. Activities are expressed as a ratio between V Ivy and V 0 , the initial velocities in the presence or absence of inhibitor, respectively. K i was determined as the value allowing the best fit of the experimental data onto a slow competitive tight binding inhibition model described by Eq. 1 (see text). The curve corresponds to the model for a K i value of 1 nM and 9.0. The order of magnitude of the dimerization K d appeared much lower than 10 Ϫ9 M, although no precise measurement could be made in this concentration range. Altogether, these biophysical results suggest that the Ivy homodimer is the physiologically active unit. C-type lysozyme is an ancient protein whose origin goes back about 500 million years (16). It has long been recognized that lysozymes (the family of enzymes hydrolyzing the 1,4-␤ linkages between N-acetyl-D-glucosamine and N-acetylmuramic acid in the peptidoglycan moiety of bacterial cell walls) are part of a nonimmunological ancestral bactericidal system in vertebrates. C-type lysozyme is found in the serum (17)(18)(19), in milk (20,21), in the digestive tract (22), in the airway (23), and in all mucosal surfaces and secretions (24 -31). There is multiple evidence that lysozymes play a significant role in the control of the host microflora to prevent infection (32)(33)(34)(35)(36). Bacterial Ctype lysozyme inhibitors might thus have emerged to balance the host defense. Indeed, an increase of anti-lysozyme activity has been linked to bacterial persistence in several systems (37,38). However, lysozyme inhibitors could also be directed against lysozyme activities of other microorganisms and play a role in ecological competition (39,40). Finally, these inhibitors could also have emerged as a protection against bacteriophageencoded lysozymes, the activity of which is essential to the release of mature virions (41). It thus makes evolutionary sense that bacteria might have evolved a resistance mechanism against the bactericidal activity of various lysozymes found in their environment. However, it must be noted that for Gramnegative bacteria such as E. coli, the presence of an outer membrane impermeable to molecules larger than 0.6 kDa should be sufficient to protect the peptidoglycan moiety from the lysozymes present in the medium.
We selected ykfE as an ORFan gene expressed by E. coli K12 (5). The presence of a signal peptide predicted a periplasmic location for its protein product, which is now consistent with its newly assigned function. Our biochemical and functional analyses of Ivy show that its predominant homodimeric form strongly interacts with hen and human, and probably all Ctype lysozymes, thereby abrogating their activity in a stoichiometric manner. Ivy does inhibit C-type lysozymes under physiological conditions, as tested in human saliva, the secretion where lysozyme is naturally found at the highest concentration. It is thus likely that at least one purpose of Ivy is to protect E. coli from its natural host lysozyme bactericidal activity, for instance in cases where the integrity of the outer membrane might be compromised (e.g. by chemically aggressive compounds in the medium or at the time of cell division).
As such a resistance mechanism against an ubiquitous bactericidal enzyme should be advantageous to all murein-containing bacteria, in particular Gram-positive bacteria, ykfE/ Ivy-like genes are expected to exist in many bacterial genomes, making its ORFan nature a paradox. Indeed, we detected a putative ortholog of E. coli Ivy within the recently published genome of Pseudomonas aeruginosa (1). However, the two protein sequences only share 30% of identical residues in their most similar region, indicating a fast divergence rate (Fig. 7). It is thus likely that genes of the ykfE/Ivy family are not detected in other bacteria because of their low sequence conservation. It is our hope that the knowledge of the three-dimensional structure of Ivy will allow the discovery of other Ivy homologues by the identification of a set of critical positions in the sequence beyond the twilight zone of sequence similarity. The depicted alignment has been produced using default parameters in FASTA (42). The underlined sequence corresponds to the known signal peptide in Ivy.