Major Role of Human KLK14 in Seminal Clot Liquefaction*

Liquefaction of human semen involves proteolytic degradation of the seminal coagulum and release of motile spermatozoa. Several members of human kallikrein-related peptidases (KLKs) have been implicated in semen liquefaction, functioning through highly regulated proteolytic cascades. Among these, KLK3 (also known as prostate-specific antigen) is the main executor enzyme responsible for processing of the primary components of semen coagulum, semenogelins I and II. We have recently identified KLK14 as a potential activator of KLK3 and other KLKs. This study aims to elucidate the cascade-mediated role of KLK14 ex vivo. KLK14 expression was significantly lower (p = 0.0252) in individuals with clinically delayed liquefaction. Concordantly, KLK14 expression was significantly (p = 0.0478) lower in asthenospermic cases. Specific inhibition of KLK14 activity by the synthetic inhibitor ACTG9 resulted in a significant delay in semen liquefaction, a drop in the “early” (30 min postejaculation) “chymotrypsin-like” and KLK1 activity, and an increase in the “late” (90 min postejaculation) chymotrypsin-like activity. Conversely, the addition of recombinant active KLK14 facilitated the liquefaction process, augmented the early chymotrypsin-like activity, and lowered late chymotrypsin-like activity. Given that the observed chymotrypsin-like activity was almost completely attributed to KLK3 activity, KLK3 seems to be regulated bidirectionally. Accordingly, a higher level of KLK3 fragmentation was observed in KLK14-induced coagula, suggesting an inactivation mechanism via internal cleavage. Finally, semenogelins I and II were directly cleaved by KLK14. Semenogelins were also able to reverse KLK14 inhibition by Zn2+, providing a novel regulatory mechanism for KLK14 activity. Our results show that KLK14 exerts a significant and dose-dependent effect in the process of semen liquefaction.

Spermatogenesis, the process of differentiation of testicular stem cells into mature spermatozoa, is initiated in the seminiferous tubules of the testes, which produce immature sperm cells (1)(2)(3). Subsequent maturation occurs during epididymal transition, where immature spermatozoa acquire motility and fertilizing capacity (4,5). Sperm motility is particularly important at the time of fertilization, since it facilitates sperm penetration to the zona pellucida of the oocyte and fusion of the two cells (1,2,6).
Asthenospermia, described as impaired sperm motility, is considered as one of the main factors of male subfertility or infertility and may be caused by a number of conditions, including incomplete liquefaction, delayed liquefaction, or nonliquefaction of semen (7)(8)(9). Normally, human semen coagulates spontaneously upon mixing of its various glandular fractions in order to form a depository of spermatozoa in the rear vaginal cavity (10 -12). Subsequent liquefaction of coagulum within minutes (ϳ5-20 min after ejaculation) allows for a progressive release of motile spermatozoa (10,13). Liquefaction is achieved through a stepwise proteolytic cleavage of the gel proteins semenogelin I and II (SgI and -II) into soluble proteins, followed by their peptidic fragmentation. These peptides are eventually degraded into their constituent amino acid residues (14 -18).
Semen coagulation/liquefaction is under tight regulatory control. For instance, Sg proteins chelate with the excess of free Zn 2ϩ immediately after ejaculation and undergo structural modifications, inducing aggregate complex formation (19 -23). Sg degradation is mainly modulated through activation of KLK3 (kallikrein-related peptidase 3), also known as prostatespecific antigen (22,24,25). The enzymatic activity of KLK3 is tightly controlled through a number of endogenous inhibitors and regulatory feedback loops. For instance, along with Sg proteins, the serine protease inhibitor, protein C inhibitor, is secreted from the lumen of seminal vesicles (26). Recent evidence indicates that protein C inhibitor complexes with Sg, preventing its premature hydrolysis by active KLK3 (27). KLK3 activity is believed to be further inhibited by free Zn 2ϩ in prostatic secretions (25,28). Sg chelation with free Zn 2ϩ results in an immediate drop in the available Zn 2ϩ , which consequently leads to KLK3 activation. Conversely, Zn 2ϩ is released gradually as Sg proteins are fragmented by KLK3. The increased level of Zn 2ϩ serves as a negative feedback loop to prevent excessive proteolysis that may damage the integrity of spermatozoa (29). Recent evidence indicates an additional level of complexity in the regulation of the proteolytic cleavage of Sg proteins. For instance, in vitro data suggest that other proteases, particularly other members of the kallikrein-related peptidase (KLK) 2 fam-* 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. □ S The on-line version of this article (available at http://www.jbc.org) contains supplemental Table 1  ily (i.e. KLK2 and KLK5), are directly or indirectly involved in Sg processing (30 -32). In addition, emerging reports suggest a cascade-mediated protease activation mechanism, regulated by a number of positive and negative feedback loops. For example, KLK5 is suggested to autoactivate and, in turn, activate pro-KLK3 (33). Likewise, although still controversial, KLK2 has been suggested to activate pro-KLK3 (34 -36). Finally, our recent work implicates KLK14 as a potential activator of pro-KLK3 as well as several other seminal pro-KLKs (i.e. pro-KLK1 and pro-KLK11) (37). Characteristic to classic proteolytic cascades, we have previously proposed a bidirectional regulatory mechanism of KLK3, in which KLK14-mediated activation is followed by inactivation via internal cleavage of active KLK3 at position Lys 145 (37). KLK14 is a trypsin-like KLK, with preference over P1-Arg (38,39). Zinc ions have been shown to strongly inhibit KLK14 enzymatic activity (38), reinforcing the potential role of the protein in semen liquefaction. In an attempt to delineate a possible function of KLK14 in seminal plasma, this study examines the interaction between this enzyme and other potential components of the seminal proteolytic cascade involved in semen liquefaction.
Materials-Coagulated semen, having a normal liquefaction rate at room temperature, was collected, split into three fractions, and frozen immediately after ejaculation in liquid nitrogen. Samples were stored at Ϫ80°C until required. Liquefied semen was obtained from 95 subjects with normal and delayed liquefaction, under informed consent and approval by the Institutional Review Boards of Mount Sinai Hospital and the University Health Network. If required, semen coagula were artifi-cially emulsified, by the addition of a small amount of chymotrypsin enzyme at 37°C for up to 1 h.
Measurement of Clinical Parameters of Semen-Liquefaction rate was estimated by attempting to draw the specimen into a Pasteur pipette. Complete liquefaction was achieved when all of the fluid entered the pipette. In addition, liquefaction level was evaluated visually by a phase-contrast microscope as a measure of disappearance of the gel-like coagulum structure.
Overall sperm motility (percentage) was determined using automated computer-assisted semen analysis as (a ϩ b/(a ϩ b ϩ c)) ϫ 100, where a, b, and c represent the number of progressively motile sperm, sperms moving in random directions, and nonmotile sperms, respectively. Cases with a percentage of sperm motility equal to or less than 35% were considered as asthenospermic.
Cleavage of SgI and -II Proteins-500 ng of purified SgI and SgII were incubated individually with 56 ng of KLK14 in 30 l of KLK14 optimal assay buffer (100 mM phosphate buffer, 0.01% Tween 20, pH 8.0) at 37°C for various time points. Reactions were snap-frozen in liquid nitrogen and run on SDS-polyacrylamide gels under reducing conditions. Gels were silver-stained to visualize fragmentation.
Sg-mediated Reversal of Zn 2ϩ Inhibition-To examine Sgmediated reversal of Zn 2ϩ inhibition, 12 nM KLK14 was incubated with 0 and 120 nM of Zn 2ϩ (in the form of zinc acetate), at a final volume of 100 l, for 10 min at 37°C. Subsequently, the fluorogenic substrate QAR-AMC was added at a final concentration of 1 mM. Fluorescence release was measured on a Wallac Victor fluorometer (PerkinElmer Life Sciences), set at 355 nm for excitation and 460 nm for emission. Fluorescence was measured for a total of 20 min. Five minutes after initiating the read, a 0.05 M concentration each of SgI and SgII or 0.01 M EDTA was added to each well. Measurement was resumed as described above. Background fluorescence was subtracted from raw values. All experiments were performed in triplicate.
Enzyme Activity Assays-The "chymotrypsin-like" activity of seminal plasma samples (diluted 10 times) was kinetically examined, using 0.8 mM colorimetric substrate RPY-pNA in a final volume of 100 l of KLK3-optimized assay buffer (0.1 mM Tris, 3 mM NaCl, 0.01% Tween 20, pH 7.5). Absorbance was measured on a Wallac Victor Fluorometer at 405 nm. Background absorbance was subtracted from raw values of seminal plasma alone, and samples were treated with either active recombinant KLK14 or ACT G9 , described above. Reactions were repeated three times.
KLK1-specific activity was measured by fluorescence release of the pulled down KLK1 protein, as previously described (37). Briefly, 200 ng of KLK1-specific polyclonal antibody (HUK-IgG) were immobilized on a 96-well white polystyrene plate overnight. The plate was washed two times prior to the addition of reaction mixtures. KLK1 activity was measured as an increase in the fluorescence of PFR-AMC substrate after 2 h of sample incubation. Reaction rates (fluorescence units/min) correspond to the slope of the fluorescence release-time plot.
KLK3 Depletion from Seminal Plasma-1 mg of monoclonal anti-KLK3 antibody was immobilized on 1 ml of 50% N-Hydroxysuccinimide-activated Sepharose Fast Flow bead slurry, according to the manufacturer's protocol. Briefly, beads were equilibrated three times in 2 ml of ice-cold 1 mM HCl. They were then incubated with 1 mg of monoclonal anti-prostatespecific antigen antibody for 1 h at room temperature with endover-end mixing. Residual active groups of beads were subsequently blocked by washing beads sequentially three times with 2 ml each of buffer A (50 mM Tris⅐HCl, 1 M NaCl, pH 8.0) and buffer B (0.1 M acetate, 0.5 M NaCl, pH 4.0). Beads were further washed two times with buffer A and then incubated for 15 min at room temperature. Further blocking was achieved by sequential incubation of beads three times each with buffer B, A, and B. Beads were equilibrated for protein binding in TBS (50 mM Tris, 150 mM NaCl, pH 7.5). 20 l of seminal plasma were diluted in TBS in total volume of 1 ml and incubated with beads for 1 h at room temperature, with end-over-end mixing. The flow-through (depleted samples) was collected and further analyzed kinetically. Beads were washed five times in wash buffer (TBS with 2 M urea, pH 7.5) and eluted with 1 ml of elution buffer (0.1 M glycine with 2 M urea, pH 3.0). A mockdepleted sample was prepared in parallel, using beads alone. Percentage of depletion was estimated by measuring KLK3 in flow-through samples, using KLK3-specific ELISA. Collected flow-through samples were concentrated 10 times, using membranes with a molecular weight cut-off of 5000. 5 l of concentrated samples were diluted in 95 l of KLK3 optimal assay buffer. Enzymatic activity toward the tripeptide RPY-pNA substrate was measured as described above.
To ensure that the observed drop of enzymatic activity is due to exclusive depletion of KLK3, two identical reactions of 20 l of immunodepleted and mock-depleted elutions were run on SDS-PAGE under reducing conditions. One gel was silverstained, and the other was immunoblotted with anti-KLK3 antibody as described below.
Western Blotting for Identification of KLK3 Fragmentation in Seminal Plasma-To monitor KLK14-mediated fragmentation of KLK3 ex vivo, semen coagula were spiked for 1 h to various amounts of active recombinant KLK14 and were analyzed by Western blot. Similarly, KLK14-mediated fragmentation of pro-KLK3 was reconfirmed in vitro by incubating recombinant pro-KLK3 with active KLK14 at a 10:1 molar ratio for varying times, in a total volume of 30 l. Recombinant and seminal proteins were resolved by SDS-PAGE, using the NuPAGE Bis-Tris, with 4 -12% gradient polyacrylamide gels at 200 V for 45 min and transferred onto a Hybond-C Extra nitrocellulose membrane (GE Healthcare) at 30 V for 1 h. The membrane was subsequently blocked for 1 h with 5% milk/TBS-Tween (0.1 mol/liter Tris-HCl containing 0.15 mol/liter NaCl and 0.1% Tween 20) at 4°C and probed using rabbit anti-KLK3 polyclonal sera (diluted 1:1000) for 1 h at room temperature. The membrane was washed three times for 15 min with TBS-Tween and treated with alkaline phosphatase-conjugated goat antirabbit antibody (diluted 1:8000) for 45 min at room temperature. The membrane was rewashed as above, and fluorescence was detected on x-ray film using a chemiluminescent substrate.

RESULTS
Clinical Association between KLK14 Expression and Liquefaction Rate-KLK14 concentration in seminal plasma from 95 volunteers (including 34 normal cases and 61 patients with delayed liquefaction) ranged from 0.2 to 181.2 g/liter, with a mean of 13.2 g/liter and a median of 6.8 g/liter. The expression level of KLK14 had a median of 5.2 and 11.55 g/liter and mean of 12.38 and 12.99 g/liter in samples with delayed and normal liquefaction, respectively (Fig. 1A). We concluded that KLK14 levels were significantly decreased (p ϭ 0.0252) in the patient group with delayed liquefaction. In addition, KLK14 expression was found to be significantly (p ϭ 0.0478) lower in 70 asthenospermic patients (15 cases with undetermined or inconclusive percentage motility were excluded from the study) (Fig. 1B). The level of KLK14 was dropped to 9.8 g/liter (mean) and 7.9 g/liter (median) in asthenospermic cases, as compared with normal individuals with a mean value of 22.5 g/liter and a median of 13.4 g/liter.
Role of KLK14 as a Seminal Liquefying Protease-To further investigate the possible role of KLK14 in semen liquefaction, the proteolytic activity of the enzyme in seminal plasma was induced and reciprocally inhibited by using either active recombinant KLK14 or the highly specific KLK14 inhibitor ACT G9 , respectively. Complete liquefaction ranged from 10 to 20 min in normal samples. The addition of ACT G9 inhibitor to a split fraction of a normal ejaculate sample strongly delayed liquefaction (Ն30 min). As expected, the progression of liquefaction was also reduced in inhibitor-treated samples, since the gel-like coagulum structure persisted longer than their untreated control counterparts (Fig. 2). Conversely, liquefaction was accelerated upon the addition of active recombinant KLK14. The gel-like structure of semen coagula seemed to be less dense in KLK14-induced samples (Fig. 2). The coagula of normal liquefying ejaculates were dissolved too fast; for this reason, we could not determine the effect of KLK14 on liquefaction rate.
Cleavage of Sg Proteins by KLK14-Given the pronounced effect of KLK14 on semen liquefaction, we next examined whether any of the primary components of semen coagulum function as immediate downstream targets of KLK14. The ability of KLK14 to cleave purified SgI and -II proteins was tested.
SgI and -II were incubated with active recombinant KLK14 in separate reactions. KLK14 was able to almost fully cleave both SgI and -II, as quickly as 12 min of incubation (Fig. 3). New fragments were generated as early as 2 min after initiation of the reaction.
Reversal of Zn 2ϩ Inhibition by SgI and -II-Zn 2ϩ has previously been proposed to function as a cationic protease inhibitor of KLK14 (38). As mentioned previously, SgI and -II can indirectly regulate the activity of a number of KLKs by binding to Zn 2ϩ molecules, rendering them unable to inhibit KLK activity. To examine whether Sg proteins have the same effect on Zn 2ϩ -mediated inhibition of KLK14, KLK14 was incubated with a 10-fold molar excess of Zn 2ϩ . The enzymatic activity of KLK14 was monitored kinetically as above. The addition of SgII after 5 min of initiation of the reaction rapidly reversed the inhibition (Fig. 4), suggesting a common regulatory mechanism with several other seminal KLKs. No such effect was observed for SgI (data not shown).
Correlation between KLK14 and the Chymotrypsin-like Activity-Pro-KLK3 was previously proposed to function downstream of KLK14 in vitro (37). Unfortunately, there is no tool currently available to specifi-  cally quantitate KLK3 enzymatic activity in complex biological samples, such as seminal plasma. However, given, KLK3 shows preference to substrates with P1-tyrosine, P2-proline, and P3-arginine. Given that KLK3 is the major chymotrypsin-like enzyme in seminal plasma and its preferential substrate recognition, we reasoned that KLK3 activity could accurately be estimated in seminal plasma by measuring the chymotrypsin activity toward the RPY-pNA substrate. To corroborate this assumption, a series of ex vivo depletion experiments were performed. Seminal plasma with ϳ95% depleted KLK3 exhibited almost zero activity toward the RPY tripeptide substrate, as compared with the mock-depleted control (supplemental Fig.  1A). In addition, eluted samples of depleted and mock controls were examined by silver stain and Western blotting against KLK3 (supplemental Fig. 1B). All of the proteins eluted from the immunodepleted sample were successfully identified as full-length KLK3 or KLK3 fragments by Western blotting, verifying the specificity of pull-down.
Given that KLK3 activity could confidently be assessed by measuring the chymotrypsin-like activity against the tripeptide RPY-pNA (referred to here as "chymotrypsin-like" for short), KLK14-mediated regulation of KLK3 was next examined. The chymotrypsin-like activity of seminal plasma was dependent upon the level of KLK14 activity, since samples treated with active recombinant KLK14 exhibited ϳ78% higher "early" (30 min after ejaculation) chymotrypsin-like activity, compared with those treated with the KLK14 inhibitor (Fig. 5A). As previously suggested, the observed increase was rapid and transient, followed by a decrease in the chymotrypsin-like activity. The reaction rate declined following longer incubation (90 min postejaculation) of seminal coagula, resulting in a reversal of the activity pattern of treated samples versus the untreated controls (Fig. 5B). The "chymotrypsin-like" activity of ACT G9 -treated samples increased ϳ10%, whereas a drop of almost 78% in was seen in samples added to active recombinant KLK14 (Fig. 5B).
Fragmentation of Seminal KLK3 by KLK14-Our previous in vitro work suggests an inactivation mechanism of KLK3 through internal cleavage of the active protein. To confirm this, we compared degraded products of KLK3 in vitro and in seminal plasma by Western blotting, using rabbit anti-KLK3 polyclonal sera. All major fragments identified previously by silver staining (37) were detected by our antibody (Fig. 6A). A very similar fragmentation pattern was observed in seminal plasma spiked with various amounts of active recombinant KLK14 (Fig.  6B). As expected, fragmentation was dependent on the level of KLK14 activity. Interestingly, the prominent band generated following KLK14 induction has the molecular mass of the previously identified fragment produced uniquely by KLK14, after cleavage of KLK3 at the peptide bond Lys 145 -Lys 146 (37).
Activation of Seminal KLK1 by KLK14-KLK1 has been proposed as one of the downstream targets of KLK14, in vitro (37). To evaluate a possible KLK14-mediated activation mechanism of seminal KLK1, we examined KLK1-specific activity in ACT G9 -treated samples as compared with an untreated split fraction of the same ejaculate. The specific activity of KLK1 was attenuated ϳ20% upon treatment of the ejaculate with the ACT G9 synthetic inhibitor against KLK14 (Fig. 7A). This would suggest that KLK14 could activate pro-KLK1 in seminal plasma. Given the high abundance of trypsin-like KLKs with overlapping substrate specificity, it is critical to ensure pulldown specificity of the KLK1 antibody. In order to exclude the possibility of nonspecific pull-down of physiologically relevant KLKs, protein expressions of KLK1, -2, -4, -5, -11, -12, and -14 were measured using ELISAs developed in house (supplemental Table 1). The pull-down specificity of anti-KLK1 HUK IgG was evaluated using active recombinant KLK2, -4, -5, -11, -12, and -14 in their equivalent amounts found in seminal plasma (Fig. 7B). Although these KLKs are highly active when soluble (data not shown), almost no enzymatic activity was observed after they were pulled down with KLK1 antibody. Based on the information provided above, a novel cascade pathway for KLK14 function in semen liquefaction was developed (Fig. 8).

DISCUSSION
Human semen coagulates spontaneously after ejaculation and consequently liquefies within 5-20 min under normal physiological conditions (11). Although the mechanism is not fully understood, the process of semen coagulation/liquefaction is believed to be regulated through a series of enzymes, mainly proteases, and inhibitory factors (14,15).
More recently, a number of well known components of the blood coagulation and fibrinolysis systems, including protein C inhibitor, tissue and urokinase type plasminogen activator, tissue factor, tissue factor pathway inhibitor, and blood coagulation factor X, have been identified in seminal plasma and have been associated with male fertility (42)(43)(44)(45)(46). Given the overlapping regulatory components of the seminal and blood homeostasis, this emerging evidence suggests that analogous to fibrinolysis, semen liquefaction is regulated through highly orchestrated proteolytic cascades (7,10,31,47,48).
Classic proteolytic cascades consist of sequential activations of protease zymogens through three main phases of initiation, progression (or propagation), and execution (49). A proteolytic

Role of KLK14 in Semen Liquefaction
cascade is often initiated by an external stimulus, which in turn triggers the initiating protease or "initiator" to self-activate. Subsequently, active initiator converts downstream propagator proteases into their active form by limited proteolysis. Finally, the active propagator activates executor enzymes during the execution phase. Furthermore, active proteases often activate more of their initiator(s) via positive feedback mechanisms, which would result in a rapid amplification of proteolytic activity. To prevent unwanted protein degradation, a typical proteolytic cascade contains multiple regulation points, including inhibitors, autodegradation, and internal cleavage mechanisms.
Accumulating evidence suggests that several members of the KLK family participate in the seminal proteolytic cascade and are involved in the process of degradation of the semen coagulum (31). In vitro data by our group and others suggest that KLK14 might function as a key factor in the proteolytic cascade in seminal plasma, regulating major seminal KLKs, including KLK1, KLK3, and KLK11 (37,50). Furthermore, the enzymatic activity of KLK14 has recently been shown to be inhibited by Zn 2ϩ (38), strengthening the proposed function of the enzyme in seminal plasma and prostatic tissue. Here, for the first time, we propose a cascade-mediated role for KLK14 in seminal plasma, as one of the key trypsin-like regulatory proteases involved in liquefaction of the seminal coagulum.
Trypsin-like proteases are of main importance, since they can function as activators of KLKs that are unable to self-activate (32). A prime example of KLKs lacking autoproteolytic ability is the chymotrypsin-like enzyme KLK3. As mentioned previously, KLK3 has extensively been studied as a main executor KLK in seminal plasma, functioning through cleavage of gel-like proteins and initiating semen liquefaction (22,28). However, surprisingly, no significant difference was found in KLK3 expression level between normal and delayed liquefaction (51), suggesting possible aberration at the regulatory level of the protein, due to insufficient activation. Previously, we reported KLK14 as an activator of pro-KLK3. Interestingly, our clinical data indicate that there is a significant correlation between abnormal liquefaction and asthenospermia and the expression level of seminal KLK14. The physical constraint of retained coagula seems to adversely affect sperm motility, since we observed an ϳ70% drop in number of motile sperms in samples with delayed liquefaction (data not shown). Whether the observed reduced level of KLK14 is due to its abrogated expression in the prostate or its partially obstructed secretion to seminal plasma remains to be determined.
In addition, using targeted inhibition and reciprocal overactivation of KLK14 in seminal plasma, we demonstrated that KLK14 is vital for complete liquefaction of the seminal clot. The mutant inhibitor ACT G9 used in this study is highly potent and selective toward KLK14 (40). ACT G9 contains mutations at the reactive center loop of the biological inhibitor ACT, converting the natural reactive center loop to the phage display-selected KLK14 substrate G9 (40). This would confer an excellent inhibitory specificity toward KLK14; other major seminal KLKs,  including KLK2, -3, -4, -5, and -12, were not inhibited by this protein (40). 3 Although KLK14 inhibition considerably delayed semen liquefaction, it did not completely block the process. This suggests functional redundancy in activator components of the seminal proteolytic cascade, compensating for KLK14 function. The physiological relevance of other candidate activators of the cascade, such as KLK5 and KLK2, needs to be further investigated.
As mentioned previously, Sgs are the main effector components of the semen liquefaction cascade. Our in vitro data suggest that KLK14 cleaves Sg proteins with high efficiency. In addition, our previous studies have implicated KLK14 in the processing of fibronectin, another key component of the semen coagulum (22). Furthermore, Sg proteins play an instrumental role in seminal clot liquefaction through sequestration of Zn 2ϩ from active executors, thus modulating their proteolytic activity (29). Such a reversal effect of Sg has been shown for several members of the KLK family, including KLK3 and KLK5 (28,31). Our results suggest a similar regulatory mechanism for KLK14 in seminal plasma, at the physiologically relevant molar ratio of 10-fold excess Zn 2ϩ (28) to SgII protein.
Moreover, we previously demonstrated that KLK14 is able to regulate pro-KLK3 in vitro. At the astounding expression level of 10 mg/ml, KLK3 is the most abundant chymotrypsin-like enzyme in seminal plasma (41,52). However, the majority of  Samples were incubated at 37°C for 10 min. KLK1 was pulled down in 96-microtiter plates, coated with anti-KLK1 antibody, as follows. 200 ng of anti-KLK1 antibody were immobilized overnight on a microtiter plate. 100 l of each of the treated and untreated samples were loaded to each well in triplicates and incubated at room temperature for 2 h. Activity of the pulleddown KLK1 was monitored by cleavage of 0.5 mM of the PFR-AMC substrate. B, specificity of the KLK1 sandwich pull-down assay. Recombinant active KLK1, 2, 4, 5, 11, 12, and 14 were loaded on a KLK1 antibody-coated microtiter plate, at their physiologic level listed in supplemental Table 1. Enzymatic activity of pulled-down enzymes were measured as described in A. The reaction rate of all of the KLKs except KLK1 was almost zero.
active KLK3 is complexed with seminal inhibitors, such as protein C inhibitor and ␣ 2 -macroglobulin (27,53,54), rendering it inactive. Although a number of chromatographic and immunologic approaches have previously been proposed to measure active KLK3 (55-57), their low recovery rate limits their use as a sensitive comparative means in complex biological samples. Due to this technical limitation and given the substrate preference of KLK3 for the RPY tripeptide substrate, we examined the "chymotrypsin-like" activity of seminal plasma as a measure of KLK3 activity. To ensure that the majority of observed enzymatic activity against this substrate is due to KLK3 activity, we compared samples depleted from KLK3 with mock controls. As expected, upon 95% depletion of KLK3, almost no enzymatic activity was observed. Eluted proteins were identified as KLK3 or KLK3 fragments, excluding the possibility of simultaneous depletion of other chymotrypsin-like enzymes.
KLK14-mediated regulation of KLK3 activity seems to be bidirectional, since we observed a reversal in the correlation pattern between KLK14 and the "chymotrypsin-like" activity following longer incubation. Given the importance of chymotryptic proteolysis of Sg proteins during semen liquefaction, activation of KLK3 is most likely triggered within seconds postejaculation and continues until complete fragmentation of gel-like proteins. Aberrant proteolysis due to prolonged protease activity is prevented by subsequent inactivation of executor chymotryptic enzyme(s). This finding is in agreement with our in vitro observation of sequential activation and deactivation of pro-KLK3 by KLK14 (37). We have previously found that deac-tivation is achieved mainly through internal cleavage of active KLK3 (37). Here we have shown that exogenous KLK14 could fragment KLK3 ex vivo in a dose-dependent manner, with a pattern similar to the one observed in vitro.
Similarly, consistent with our previous in vitro data, KLK14 seems to activate seminal KLK1. Likewise, KLK2 has recently been identified as another putative activator of pro-KLK1 (50), reinforcing the link between KLK1 and the seminal KLK cascade. Although not fully understood, KLK1 has clinically been shown to enhance sperm motility in asthenospermic patients (58,59). As mentioned previously, semen liquefaction is one of the main postejaculatory determinants of sperm motility. Whether KLK1 functions through regulating coagulation/liquefaction of semen needs to be further explored.
In summary, the present study provides strong evidence for the crucial cascade-mediated function of KLK14 in regulating the coagulation and liquefaction of human semen (Fig. 8). Cascade activation is more likely triggered at the time of semen ejaculation, as a result of mixing of different components of seminal plasma and subsequent redistribution of Zn 2ϩ to Sg proteins. It is conceivable that additional members of the KLK family and/or other proteases participate in this proteolytic cascade. In addition, the complex interplay between proteases and their regulatory checkpoints needs to be further elucidated.
Understanding KLK-mediated proteolytic events in seminal plasma can shed light not only on the physiological role of this important family of enzymes but also on some of the causes of abnormal sperm motility. Accordingly, therapeutic induction of the seminal proteolytic cascade can be utilized to supplement the current clinical treatment of male subfertility. Conversely, targeted inhibition of key components of the cascade may have potential pharmaceutical utility as a novel topical contraceptive strategy.